Male CaMkIIα-cre mice (Jax lab, stock number 028867) were housed in standard laboratory cages with four to five mice per cage. The mice were kept on a 12-h light/dark cycle with lights on at 8:00 a.m. and in a temperature-controlled room with a range of 21–25°C. At the time of rAAV infusion, the mice were between 7 and 8 weeks old. Ad libitum food and water were provided. All animal procedures were approved by the Southern Medical University Animal Ethics Committee and conducted in compliance with the guidelines of the Chinese Council on Animal Care to minimize animal suffering and reduce the number of animals used.

The rAAV vector AAV5-EF1α-DIO-mCherry-WPRE (hereinafter referred to as rAAV5-mCherry) was packaged, purified, and tittered by Sunbio Medical Biotechnology (Shanghai, China). The double floxed inverted open-reading-frame (DIO) strategy makes mCherry expression in a Cre-responsive manner (Atasoy et al., 2008). The titter of rAAV5-mCherry was determined by real-time PCR as previously described (Rohr et al., 2002). Briefly, primers were designed to amplify the WPRE. The forward and reverse primers were 5′-CCTTTCCGGGACTTTCGCTTT-3′ and 5′-GACTAATACGTAGATGTACTGCCAAGTAGG-3′, respectively. Real-time PCR was carried out using an ABI 7500 Real Time PCR System (Applied Biosystems, USA). The PCR program was as follows: 95°C for 10 min followed by 40 cycles of 95°C for 10 s and 60°C for 35 s. AAV genome equivalents ranging from 10⁴ to 10⁷ were used to set up a DNA plasmid standard curve. The average titter of rAAV5-mCherry was 5.77 × 10¹³ vector genomes (v.g.)/ml. Serial dilutions were made to obtain the following doses: 2.88 × 10⁸, 2.88 × 10⁹, and 2.88 × 10¹⁰ v.g. per 0.5-μl (Ortinski et al., 2010).

Mice were anesthetized with 1% pentobarbital sodium for all surgeries and placed in a stereotaxic frame (RWD, CN), which was conducted under aseptic conditions and guided by stereotaxic techniques. The skull was exposed, and four holes were drilled using a small steel burr at the following coordinates: −2 mm anteroposterior (AP), ± 1.6 mm mediolateral (ML), and −1.5 mm dorsoventral (DV) (dorsal hippocampus), and −2.9 mm AP, ±3.7 mm ML, and −3.8 mm DV (ventral hippocampus). All coordinates were relative to bregma. A 5 μl Hamilton syringe with a microinjector pump (KDS, Stoelting, USA) was used to inject the virus into the bilateral dorsal and ventral hippocampi at a rate of 0.05 μl/min, and the total volume of the 4 injections was 2 μl, with 0.5 μl being injected into each site. Once the injection was finished, the needle was lifted by 0.1 mm and left in position for an additional 10 min to facilitate virus diffusion at the injection site before being gradually removed. There were 15 mice in each group. Immunofluorescent staining, cell sorting experiments, and behavioral tests were conducted 3 weeks after virus injections when transgene expression reached asymptotic levels (Klein et al., 2006).

To isolate mCherry-labeled pyramidal neurons, we employed a modified version of a previously reported procedure (Brewer and Torricelli, 2007; Hempel et al., 2007; Zeisel et al., 2018). Mice were anesthetized with 1% pentobarbital sodium and then perfused through the left ventricle with ice-cold artificial cerebrospinal fluid (ACSF). ACSF consisting of 20 mM glucose, 75 mM sucrose, 1.25 mM NaH₂PO₄, 2.5 mM KCl, 26 mM NaHCO₃, 87 mM NaCl, 7 mM MgSO₄, 1 mM CaCl2, and 0.1 μM tetrodotoxin was equilibrated in 95% O₂/5% CO₂ before use. Tetrodotoxin (0.1 μM) was added to ACSF to inhibit neuronal action potentials and improve cell viability. The brain was removed, and 300 μm sections were collected in ice-cold ACSF using a Vibroslice microtome (VT 1000S; Leica). The hippocampus was microdissected and transferred to a 100-mm dish containing 7 ml digestion solution (2 mg/ml papain (Worthington, LS003119), 0.5 mM EDTA-2Na, and 2.5 ml/L Glutamax 100X (ThermoFisher) in Hibernate A (ThermoFisher, A1247501) and incubated for 30–40 min for enzymatic digestion at 30°C. The pH of the digestion solution was adjusted to 7.4 using neurobasal A (ThermoFisher, 10888022). After incubation, the tissue was delicately moved to a 15 ml tube filled with 5 ml ice-cold HABG [60 ml Hibernate A; 20 ml/L B₂₇ 50X, (ThermoFisher), and 2.5 ml/L Glutamax 100X]. The tissue pieces were gently triturated through Pasteur pipettes with polished tip openings of 600, 300, and 150 μm diameter. The supernatant was collected in a 15 ml tube after settling. The sediment was resuspended in 2 ml HABG, and the supernatant after settling was transferred. These procedures were repeated until the sediment was dissociated. The cells were enriched by centrifugation (800 × g, 5 min, 4°C), resuspended in 1 ml HABG containing 1 mg/ml DNase I (Roche, 11284932001), and filtered with a 40 μm cell strainer (Falcon). To label dead cells, 0.1 μg/ml DAPI (Invitrogen) was added to the single-cell suspension. Then, the cells were immediately loaded into the Beckman MoFlo XDP Cell Sorter system for fluorescence-activated cell sorting (FACS). Cells with high mCherry and low DAPI fluorescence were isolated in 1.5 ml tubes containing 5.4 μl sample buffer, 0.6 μl RNase inhibitor, and 4 μl RNase-free water (Discover-sc WTA Kit V2, N711, Vazyme). For cell-sorting experiments, individual biological samples were produced by isolating 500–1,000 cells from 2 to 3 mice. For each group, 3–5 biological replicates were analyzed according to previous reports (Cahoy et al., 2008; Farmer et al., 2016; Wheeler et al., 2023).

We used the Discover-scTM WTA Kit V2 for full-length poly(A)-RNA reverse transcription and cDNA amplification. All steps were carried out according to the manufacturer's instructions. Two controls were used: 1 ng standard total RNA was used as a positive control, and sheath fluid was used as a negative control to monitor the specificity of amplification and the possibility of contamination. The cDNA concentration was quantified by a Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, CA, USA) and Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Samples with a fragment distribution between 400 and 10,000 bp, a peak location of ~2,000 bp, and a total content of more than 2 ng were used for sequencing. After generating cDNA (1 ng), sequencing libraries were constructed using the TruePrepTM DNA Library Prep Kit V2 for Illumina^® (Vazyme, #TD503).

RNA sequencing was conducted by GENEWIZ (Suzhou, China), and the libraries were constructed and sequenced on an Illumina HiSeq X Ten instrument, following the manufacturer's instructions for 2 × 150 bp paired-end sequencing (Illumina, San Diego, CA, USA).

The procedures were modified from the previous reports (Chen et al., 2021). Mice were anesthetized with 1% pentobarbital sodium and intracardially perfused with saline followed by 4% polyformaldehyde. The brains were removed, postfixed in 4% polyformaldehyde for 12 h, and then equilibrated in 30% sucrose. Coronal sections of 40-μm thickness containing the hippocampus were collected. Brain slices were blocked for 1 h with 5% bovine serum albumin and 1% Triton X-100 at room temperature and then incubated with primary antibodies (rabbit anti-NeuN 1:300, Cell Signaling Technology, Cat# 12943; mouse anti-GFAP 1:500, Cell Signaling Technology, Cat# 3670 and rabbit anti-Iba1 1:500, Cat# 01919741) overnight at 4°C. The next day, the slices were incubated with secondary antibodies (Alexa 488-conjugated goat anti-rabbit and Alexa 647-conjugated goat anti-mouse, 1:500, Invitrogen) for 2 h at room temperature after washing three times with PBS and then incubated with DAPI (Invitrogen) for 10 min and washed for another three times. The slices were mounted on microscope slides. Fluorescence images were captured using a confocal microscope (A1R, Nikon). Cell counting and fluorescent area calculation were performed by ImageJ.

Droplet digital PCR was conducted using the QX200 ddPCR EvaGreen Supermix kit (Bio-Rad, Pleasanton, CA). The sample concentrations were adjusted to 0.2 ng/μl. The total reaction volume was 20 μl, including 100 nM forward primer and 100 nM reverse primer for the target gene (Table 1), 10 μl ddPCR EvaGreen Supermix, cDNA template, and double distilled water. Three biological replicates were analyzed for each group. The 20-μl reaction mixture was loaded with 70 μl droplet generation oil (Bio-Rad, USA) into the wells in the middle row of a DG8™ cartridge to generate the droplet using the QX200 Droplet Generator. After processing, the droplets were transferred to the top wells of a 96-well PCR plate. PCR amplification was conducted on a T100 thermal cycler (Bio-Rad, USA) using a protocol as follows: 1 cycle of 95°C for 5 min, 40 cycles of 95°C for 30 s and 60°C for 60 s, 1 cycle of 4°C for 5 min, 1 cycle of 90°C for 5 min, and a hold at 4°C. The ddPCR data were analyzed with QuantaSoft™ analysis software. Positive droplets containing amplification products were discriminated from negative droplets by applying a fluorescence amplitude threshold, which was set to the highest value of the no template control (NTC) in QuantaSoft software.

To ensure high-quality data and eliminate technical artifacts, such as adapters, PCR primers, fragments, and low-quality reads (with a Phred score < 20), raw data in fastq format were subjected to processing with Trimmomatic (v 0.30, RRID: SCR_011848) (Bolger et al., 2014). FastQC (v 0.11.8, RRID: SCR_014583) was used to determine sequencing quality. All libraries were aligned to UCSC mm10 using Hisat2 (v 2.0.1, RRID:SCR_015530) (Kim et al., 2015). The clean data were monitored based on quality control (QC) criteria. QC-qualified samples from each group were included (Q30 > 85%, total mapped > 70%). Aligned reads were used for quantification by HTSeq (v0.6.1, RRID:SCR_005514). Fragments per kilobase of exon model per million reads mapped (FPKM) is used for the estimation of gene expression based on the read count. The analysis procedure is described below. To reduce noise, genes obeying read counts above 10 in each biological sample were included in the analysis. Differential gene expression was analyzed across all three groups by the likelihood ratio test (LRT) in the DESeq2 (RRID:SCR_015687) Bioconductor package in the R (3.6.0) environment (RRID:SCR_001905) and an adjusted p-value of < 0.05 was used to indicate differential expression among the three groups (Love et al., 2014). Short time-series expression miner (STEM, v1.3.11, RRID:SCR_005016), a clustering method, was used to define a set of genes with similar expression patterns (Ernst and Bar-Joseph, 2006). We used this method to cluster differentially expressed genes (DEGs). DEG subsets with the same expression pattern, as determined by the STEM method, were subjected to functional enrichment analysis using the R package clusterProfiler (v 3.10.1, RRID:SCR_016884) (Yu et al., 2012), and a p-value of < 0.05 was considered to indicate significant enrichments. Heatmaps were constructed by using the complexHeatmap R package (RRID:SCR_017270) (Gu et al., 2016). Principal component analysis was conducted by using the plotPCA function in the R (3.6.0) environment.

A 3-day acclimation period was provided for all mice before conducting the behavioral assays, with each mouse handled for 5 min per day to minimize stress due to contact with the experimenter. In addition, mice were allowed to adjust to the testing room for 1 h before the experiments. All the experiments were performed from 13:00 to 18:00 unless otherwise indicated. Two different batches of mice were used for the experiment with the following sequence: OFT, EPM, fear conditioning test; Y maze test, NOR. Experiments were separated by at least 4 days when the same groups of mice were used in different behavioral tests. Behavioral tests were conducted as per the previous reports (Tian et al., 2017; Chen et al., 2021). There were 11–12 mice for each group.

The hippocampal tissue samples were collected from the animals 3 weeks after virus injection, homogenized, and centrifuged to obtain the supernatant. The protein concentration in the supernatant was determined using a BCA protein assay kit.

The concentrations of Il-1β and Csf1 were determined using ELISA kits obtained from Elabscience Biotechnology Co., Ltd. (E-EL-M0037c and E-EL-M2445c). The manufacturers' protocols were followed carefully to ensure that accurate measurements were obtained for each of the proteins of interest.

The experimental groups were not randomly assigned, and the ideal sample size was not predetermined. The normality of data distribution was assessed by the Shapiro–Wilk test. Statistical analysis of normally distributed data was performed using one-way ANOVA, followed by either the post-hoc Tukey's multiple comparison test or post-hoc Bonferroni analysis. Data were presented as mean ± SEM. Significance levels were denoted as ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. The statistical software used for analysis was SPSS.

To specifically express mCherry in pyramidal neurons, we infused the Cre-dependent rAAV vector rAAV5-EF1α-DIO-mCherry (abbreviated as rAAV5-mCherry) into the bilateral hippocampus of CamkIIα-Cre mice, which expressed the Cre recombinase under the pyramidal neuronal specific promoter CamkIIα. Three different doses were employed, among which 2.88 × 10⁸ genome copies per injection were defined as low dose, 2.88 × 10⁹ as medium dose, and 2.88 × 10¹⁰ as high dose according to previous reports (Ortinski et al., 2010). The morphology of the cell bodies expressing mCherry looked like neurons, and the colocalization of mCherry with the neuronal marker NeuN (Figures 1A–C) suggested specific expression of neurons by the virus. Consistent with previous research (Shemiakina et al., 2012), our data showed that pyramidal neurons expressing mCherry expressed apparent fluorescent puncta (Supplementary Figure 1), which are thought to be FP aggregates and precipitates in the cell (Yanushevich et al., 2002). The transduced volumes were dose-dependent, as the dose increased, neurons were efficiently transduced in CA3 and DG with rAAV5- mCherry in addition to CA1 (Figure 1A). We also evaluated mCherry expression levels by FACS-ddPCR. First, the hippocampal slices were sectioned 3 weeks after rAAV5-mCherry injection, and single-cell suspension was generated. Subsequently, mCherry-positive cells were isolated by FACS (Supplementary Figure 2A). Then, full-length poly(A)-RNA was reverse transcribed into cDNA and amplified with high quality cDNA (Supplementary Figure 2B). ddPCR results showed that the expression of mCherry increased with increasing dose of rAAV5- mCherry [Figures 1D, E; one-way ANOVA followed by post-hoc Tukey's multiple comparisons test, F_{(2, 10)} = 728.9, p < 0.0001; L vs. M, p < 0.0001; L vs. H, p < 0.0001; M vs. H, p < 0.0001].

To explore the effects of rAAV5-mCherry expression on neurons without bias, we generated RNA sequencing (NA-seq) data of the three different doses of rAAV5-mCherry. The low dose of rAAV5- mCherry was used as a default control. mCherry-positive cells were isolated by FACS followed by reverse transcription. High-quality samples were selected for fragment and library construction and sequenced by HiSeq X Ten. The analysis process is shown in Supplementary Figure 3. The reference genome (UCSC mm10) was used to map the clean data. Data accessibility and 12,345 genes with read counts above 10 in each sample were selected for follow-up analysis. Principal component analysis (PCA) and Pearson correlation matrix, both of which were based on FPKM values, demonstrated good reproducibility of samples within the groups (Figures 2A, B).

We analyzed the differentially expressed genes (DEGs) among the three groups by the likelihood rate test (LRT) using DESeq2. A total of 4,598 genes were found to be significantly different (p < 0.05) between groups, as shown in Supplementary Table 1. The software short time-series expression miner (STEM) was utilized to group genes exhibiting similar expression patterns based on their temporal changes, and a summary plot was generated to visualize the trends in gene expression. Using STEM analysis, we found that 2,102, 1,608, 450, 171, 108, and 21 DEGs were assigned to profile 9 (Pro 9), profile 8 (Pro 8), profile 7 (Pro 7), profile 5 (Pro 5), profile 6 (Pro 6), and profile 4 (Pro 4) (similarity ≥ 0.7 and p < 0.05), respectively, according to their expression patterns in the three groups (Figure 2C; Supplementary Table 2).

As the profile with the most genetic changes, the Pro 9 module was mainly mapped to the biological processes related to neuronal death and immune response (Figure 2D; Supplementary Table 3) and enriched in the following cellular components: the lysosome, the Golgi apparatus, the endoplasmic reticulum membrane, and the endosome according to the major gene ontology (GO) pathways analysis (Figure 2E; Supplementary Table 3). Notably, as the dose of rAAV5-mCherry increased, the expression of the genes related to neuron death, oxidative stress, antigen presentation and process, and positive regulation cytokine produce parallelly increased (Figures 2F–I). For example, genes involved in apoptotic programs, such as Caspases 2, 6, and 9 (Figure 2F); the antioxidative response-related gene Hmox1, Tnfaip3, and Ucp2 (Richard et al., 2001; Su et al., 2017) (Figure 2G); the proinflammatory cytokines Il-1β, Il-18 (Julien and Wells, 2017; Man et al., 2017), colony-stimulating factor 1 (Csf1) (Guan et al., 2016), and Ccl2 (Liu et al., 2013) (Figure 2F); the immune response-related gene C1qa (Yasuda et al., 2014) (Figure 2F); genes encoding proteins with MHC for antigen presentation and processing, such as H2-DMa, H2-T22, B2m, H2-K1, H2-T23, H2-M3, and H2-D1 (Figure 2H); genes involved in inflammatory and infection processes or apoptosis, such as Lgals9, Ccl4, Fcer1g, Stat1, and Cd83 (Kashio et al., 2003; Huang et al., 2005; Bellac et al., 2007; Szczucinski and Losy, 2007; Glezer et al., 2009) (Figure 2I), were significantly upregulated in high-dose groups.

The DEGs in Pro 8 were determined to be mainly involved in ATP metabolic processes and neurotransmitter transport and to be enriched in several cellular components, such as the mitochondrial protein complex, synaptic vesicles, distal axons, and the synaptic membrane (Figures 2J, K; Supplementary Table 3). In addition, the DEGs in Pro 7 were determined to be mainly involved in the postsynaptic membrane potential, neuronal projections, and apoptosis and to be enriched in actin filaments (Figures 2J, K; Supplementary Table 3). The enrichment analyses suggested that mitochondrial function, synaptic plasticity, and transmitter release are likely to be dampened in a dose-dependent manner (Figures 2L–Q). Specifically, genes encoding mitochondrial ribosome proteins (e.g., Mrpl28, Mrpl37, Mrps11, and Mrps12) and involved in ATP synthase (e.g., Atp5a1, Atp5b, Atp5d, and Atp5g1) were downregulated. Mitochondria are critical for ATP generation, programmed cell death, and signal transduction (Friedman and Nunnari, 2014), and mitochondrial DNA (mt-DNA) released from impaired mitochondria would trigger proinflammatory and type I interferon responses (West et al., 2015; West and Shadel, 2017). Our data also showed that mt-DNAs, such as mt-Rnr2, mt-Nd1, mt-Nd2, mt-Co1, mt-Co2, and mt-Co3 in Pro9, were upregulated dose-dependently (Supplementary Figure 4). Taken together, these results suggest high-dose rAAV5-mCherry induced diminished mitochondrial function characterized by reduced ATP supply and proinflammation response.

Genes involved in synaptic plasticity, such as Cplx1, Camkv, Cpne6, and Cap43, were decreased (Figures 2O–Q) in a high-dose group. Among them, postsynaptic Cplx1 regulates AMPA receptor delivery to synapses during long-term potentiation (LTP) (Ahmad et al., 2012), and LTP is widely considered one of the major cellular mechanisms that underlie learning and memory (Bliss and Collingridge, 1993). Mice with Camkv knockdown from the CA1 pyramidal neurons exhibited marked synaptic transmission and plasticity impairments (Liang et al., 2016). Cpne6, an isoform of copine that is expressed at extremely high levels in mouse hippocampal pyramidal neurons, is required for memory and synaptic plasticity as well as LTP (Reinhard et al., 2016). Gap43, a neuronal-specific protein in the postsynaptic density gene set, is required for proper learning and memory formation and neuronal repair after injury (Hulo et al., 2002; Rekart et al., 2005; Holahan and Routtenberg, 2008). Altogether, these findings suggest high-dose rAAV5-mCherry induced altered synaptic plasticity underlying learning and memory.

Last but not least, the genes in the Pro 6 module were enriched in the biological processes of protein acylation, postsynaptic organization, and regulation of the apoptotic signaling pathway. The genes in the Pro 5 module were mapped to the GO biological processes of the ERK1 and ERK2 cascade and negative regulation of proteolysis. The genes in the Pro 4 module were enriched in the negative regulation of immune system processes and regulation of myeloid cell differentiation (Supplementary Table 3).

As genes involved in neuronal death and oxidative stress upregulated parallelly with a dose of rAAV5 increased, we then performed immunohistochemistry staining neuronal marker NeuN to compare the cellular alterations in the hippocampus of mice transduced with the three different doses of rAAV5-mCherry. We did not observe any significant distortions in the distribution of hippocampal neurons or the structure of the hippocampus among the three groups (Figure 1A). NeuN⁺ quantification revealed that there is no neuronal loss as NeuN⁺ counting was comparable among groups [Figure 3A; one-way ANOVA, F_{(3, 33)} = 0.8890, p = 0.4570]. Both astrocytes and microglia respond to, and contribute to, neuroinflammatory cues in the CNS (Sofroniew, 2015; Wolf et al., 2017). To explore whether astrocytes and microglia were upregulated after virus injection, sections were stained for astrocytic marker glia fibrillary acidic protein (GFAP) and microglia marker Iba1 (Figure 3B). Quantification of GFAP⁺ cell number and GFAP⁺ area revealed that the medial and high- doses rAAV5-mCherry did not induce astrogliosis when compared with the low-dose rAAV5-mCherry [Figure 3C, one-way ANOVA, F_{(3, 35)} = 1.2890, p = 0.2935; Figure 3D, one-way ANOVA, F_{(3, 35)} = 1.7000, p = 0.1848]. Similarly, the quantification of Iba1⁺ cell number and Iba1⁺ area also revealed that the medial and high-doses of rAAV5-mCherry did not induce microgliosis when compared with the low-dose rAAV5-mCherry [Figure 3E, one-way ANOVA, F_{(3, 35)} = 2.4820, p = 0.0770; Figure 3F, one-way ANOVA, F_{(3, 35)} = 2.5760, p = 0.0649]. In a similar study, we found that more reactive astrocytes were induced by targeting transgene expression to astrocytes with high-dose rAAV5-mCherry in Aldh1l1-creER^T2 mice than by targeting transgene expression to astrocytes with a low-dose group (Supplementary Figure 5). In addition, the proinflammatory cytokines, Il-1β and CSF, whose genes showed upregulation in GO analysis (Figure 2F) were tested. We found that a high dose of rAAV5-mCherry significantly increased the concentrations of Il-1β when compared with the buffer (PBS)-injected group [Figure 4A, one-way ANOVA, F_{(8, 27)} = 2.969, P = 0.0161, H vs. PBS: P = 0.0158], while had no effect on CSF [Figure 4B, one-way ANOVA, F_{(8, 27)} = 1.612, P = 0.1680], suggesting that high-dose of rAAV5-mCherry expression may induce an immune response.

We next validated six DEGs (Cplx1, Camkv, Cpne6, Gap43, Fos, and B2m) that are involved in synaptic plasticity underlying learning and memory (Figure 2). mCherry⁺ cells from three different doses of rAAV5-mCherry were sorted by FACS and proceeded for ddPCR. The expression of B2m and Fos were decreased specifically in the high-dose group, while Camkv, Cplx, 1Cpne6, and Gap43 were upregulated in both medium-dose and high-dose groups compared with the low-dose group [Figures 4C, D, one-way ANOVA, Tukey's post-hoc multiple comparisons test; Camkv, F_{(2, 6)} = 103.0, p < 0.0001, M vs. L: P < 0.0001, H vs. L: p < 0.0001, H vs. M: P = 0.1800; Cplx1, F_{(2, 6)} = 464.7, p < 0.0001, M vs. L: p = 0.0001, H vs. L: p < 0.0001, H vs. M: p = 0.0970; Cpne6, F_{(2, 6)} = 279.4, p < 0.0001, M vs. L: p < 0.0001, H vs. L: < 0.0001, H vs. M: p = 0.1075; Gap43, F_{(2, 6)} = 1231, p < 0.0001, M v.s L: p < 0.0001, H vs. L: p < 0.0001, H vs. M: p = 0.4425; B2m, F_{(2, 6)} = 18005, p < 0.0001, M vs. L: p = 0.8557, H vs. L: p < 0.0001, H vs. M: p < 0.0001; Fos, F_{(2, 6)} = 625.5, p < 0.0001, M vs. L: p = 0.6148, H vs. L: p < 0.0001, H vs. M: p < 0.0001], suggesting that medium-dose and high-dose of rAAV5-mCherry expression may alter neuronal synaptic plasticity.

Recent studies have indicated that the hippocampus may play a role in emotional behavior, as both human and animal research has linked hippocampal dysfunction to mood and anxiety disorders (Fanselow and Dong, 2010; Bannerman et al., 2014). To investigate whether different doses of rAAV5- mCherry anxiety-related behaviors, we employed two well-validated anxiety assays, the open field test (OFT) and the elevated plus maze test (EPM) (Carola et al., 2002). We first evaluated the effects of efficient gene transduction by rAAV5- mCherry on anxiety-related behaviors and locomotor activity in the OFT. Mice that were subjected to the OFT 3 weeks after rAAV5-mCherry injection with different doses spent similar time in the central area [Figure 5A, one-way ANOVA; F_{(3, 44)} = 0.1828, p = 0.9075]. Meanwhile, the rAAV5-mCherry expression did not affect the overall activity of the mice as indicated by total path length [Figure 5B, one-way ANOVA; F_{(3, 44)} = 0.2709, p = 0.8460] and speed [Figure 5C, one-way ANOVA; F_{(3, 44)} = 0.8349, p = 0.4819] in the OFT. Additionally, the rAAV5-mCherry expression had no effect on anxiety-related behaviors in the EPM either [Figures 5D–F, one-way ANOVA; D, F_{(3, 44)} = 0.9396, p = 0.4297; E, F_{(3, 44)} = 0.4489, p = 0.7124; F, F_{(3, 44)} = 0.0546, p = 0.9830].

Numerous studies have demonstrated an essential role for the hippocampus in several forms of memory formation (e.g., spatial learning, working memory, recognition memory, or contextual fear) (Fanselow and Dong, 2010; Bannerman et al., 2014). Our transcriptomes and ddPCR results showed that synaptic plasticity underlying learning and memory may get altered after medium-dose or high-dose of rAAV5-mCherry transduction, raising the question that whether medium-dose or high-dose of rAAV5-mCherry transduction would affect learning and memory. To this end, we employed a battery of learning and memory paradigms: the novel object recognition test, the contextual fear conditioning test, and the Y-maze spontaneous alternation test. In the Y-maze spontaneous alternation test, a decrease in the percentage of alternation was observed following high-dose rAAV5-mCherry transduction, while no change was found by medium- or low-dose rAAV5-mCherry transduction compared with the control group [Figure 5G, one-way ANOVA, Tukey's post-hoc multiple comparisons test; F_{(3, 44)} = 7.7960, p = 0.0003, C vs. L: P = 0.8656; C vs. M: P = 0.7313; C vs. H: P = 0.0115; L vs. M: P = 0.9939, L vs. H: p = 0.0012, M vs. H: P = 0.0006]. No significant difference was observed in the number of entries among the groups in the Y-maze spontaneous alternation test. [Figure 5H, one-way ANOVA; F_{(3, 44)} = 0.9231, p = 0.4376]. In contrast, in the object recognition test, medium- and high-dose rAAV5-mCherry transduction showed similar performance with the low-dose group [Figure 5I, one-way ANOVA; F_{(3, 44)} = 0.9615, p = 0.4194]. However, in the contextual fear conditioning test, high-dose rAAV5-mCherry transduction exhibited impaired contextual memory retrieval, while no impairment was found in the medium- or low-dose group compared with the control group [Figure 5J, one-way ANOVA, Tukey's post-hoc multiple comparisons test; F_{(3, 43)} = 8.0990, p = 0.0002, C vs. L: P = 0.7604; C vs. M: P = 0.9078; C vs. H: P = 0.0003; L vs. M: P = 0.9893, L vs. H: p = 0.0049, M vs. H: P = 0.0020], with similar fear learning [Figure 5K, two-way ANOVA; F_{time × group(9, 129)} = 0.1575, p = 0.9975] and tone-cued retrieval [Figure 5L, one-way ANOVA; F_{(3, 43)} = 0.8156, p = 0.4923] among groups.