Experiments were conducted on male Sprague–Dawley (SD) rats, with an average weight of 250 g. This study was approved by the Ethics Committee for Animal Experiments at the University of Nanchang (RRID: RGD_728193; Permit Number: 2010–0002). As is common with most other experiments on cocaine addiction, we excluded female rats to ensure stability in the results of behavioral testing. All experimental procedures adhered to the US National Institutes of Health Guide for the Care and Use of Laboratory Animals. Efforts were made to alleviate discomfort, suffering, and pain in experimental animals. Rats that were not within the age range of 2-months ± 2 weeks were excluded. The nestin-Cre mice have been introduced in prior publications (Tronche et al., 1999). With loxP sites flanking exons 4 and 5 of the Rab10 gene, Rab10 ^{Floxed/Floxed} (Rab10 ^F/F) mice were utilized in this study (Biocytogen Company, Beijing, China). The primers used for genotyping conditional knockout mice are listed as follows: cre, 5′-TCG​ATG​CAA​CGA​GTG​ATG​AG-3’ (P1) and 5’-TCC​ATG​AGT​GAA​CGA​ACC​TG-3’ (P2); Rab10, 5′-CAA​CCT​AAT​CAT​ACT​AAT TAGATTGGT-3’ (P3) and 5′-GAT​GCC​TAT​TGT​TAG​TGA​TAC​TAC-3’ (P4). All mice involved in this study were on a C57BL/6J background. GAD ₆₇ -GFP knock-in mice (GAD ₆₇ ⁺ mice) were generously provided by Dr Shujia Zhu (Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences, Shanghai 200031, China).

The experimental procedures within this study adhered to pseudo-randomized and repetitive-measurement rules. For scientific grouping, a computer-based random number generator was employed. The in vivo experiments, involving six rats per group, were repeated three times. 1) In three repeated experiments, 36 male SD rats were randomly assigned to two groups (cocaine and saline, 15 mg/kg), with each group comprising six rats. 2) The AAV-GFP or the AAV-Rab10-siRNA-GFP construct was injected into the bilateral NAc of 72 rats. Subsequently, these rats were injected with either cocaine or saline (15 mg/kg, i.p.). Based on different treatments, these rats were randomly allocated into four groups, with each group comprising six rats. No animals were sacrificed during the experiments.

Gene expression profiles from an independent dataset were downloaded from the Gene Expression Omnibus (GEO). The dataset (Accession Number: GSE18751) was used to investigate the differential expression levels of candidate genes of interest and perform pathway enrichment analysis. Collected using the Illumina MouseWG-6 v2.0 Expression Beadchip platform, this dataset consists of 12 whole-genome expression profiles of NAc samples from C57BL/6J mice treated with acute cocaine (20 mg/kg, i.p.) or saline. Based on the Rab10 expression levels, two groups of samples were selected to perform differential expression analysis. The top quartile was defined as the Rab10 high-expression group, and the bottom quartile was defined as the low-expression group. Genes with a false discovery rate (FDR) of p < 0.05 and |fold change| > 1.5 were considered highly differentially expressed. Pathway enrichment analysis was carried out using the DAVID tool (version 6.8), and the R package limma was used to perform differential expression analysis.

Antibodies used in this study were obtained from the following companies: Proteintech (Rab10-11808-1-AP for Western blot and immunostaining); R&D Systems (GABA_B2R-AF1188 for Western blot and immunostaining); Sigma (pan-Cadherin-C1821 for Western blot); and Millipore (GABA_B1R-MABN492 for Western blot and immunostaining). DAPI (4′,6-diamidino-2-phenylindole) was procured from Beyotime Biotechnology. For immunostaining, the secondary antibodies were obtained from Invitrogen. Horseradish peroxidase-conjugated secondary antibodies were obtained from Millipore. In this work, GABA_B1R-pHluorin constructs were developed. GCaMP6 and MYC-GABA_B1R were provided by Dr Shujia Zhu (Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences, Shanghai 200031, China).

The cells or tissues of NAc were homogenized in lysis buffer that includes 1% Nonidet P-40, 50 mM Tris-HCl, pH 7.5, 0.5% sodium deoxycholate, 150 mM NaCl, and protease inhibitors (Cocktail set III, 539134, Merck/Millipore). Membrane proteins of cultured primary neurons or NAc brain regions of rats were prepared using the membrane protein extraction kit (k268-50, BioVision), followed by immunoblotting experiments using specific antibodies, as detailed in the Reagents and Antibodies section.

The protein was separated on a 10% SDS-PAGE gel and subsequently transferred onto a polyvinylidene fluoride (PVDF) membrane. The membrane was subsequently blocked in Tris-buffered saline containing 5% skimmed milk powder for 2 h at room temperature, followed by incubation at 4°C with the primary antibodies at the desired concentration overnight. Primary antibodies used were as follows: Rab10 (1:1,000), pan-Cadherin (1:1,000), GABA_B1R (1:1,000), and GABA_B2R (1:1,000). Thereafter, the membrane was incubated with appropriate secondary antibodies coupled with HRP for 2 h at room temperature. The bands’ chemiluminescence signals on the film were detected using a chemiluminescence system, and the intensity of these bands was quantified using ImageJ software.

Our preliminary experiments (Meng et al., 2018), along with a previous study (Sano et al., 2016), demonstrated that sodium pentobarbital functions as a short-acting anesthetic that exerts fewer side effects on the blood–brain barrier and cardiovascular function than those observed with other anesthetics, such as isoflurane and ketamine/xylazine. Rats were thus anesthetized prior to surgery with sodium pentobarbital (dissolved in saline, 42 mg/kg, i.p.; Sigma Co., St. Louis, MO). Placed in the stereotaxic frame (Neurostar Co., Sindelfingen, Germany), rats were surgically implanted with bilateral stainless-steel guide cannula assemblies (22 gauge; Plastics One, Roanoke, VA) directly into the regions of bilateral NAc shells during an aseptic stereotaxic procedure. The guide cannula was positioned using stereotaxic coordinates relative to bregma: M/L+ 1.6 mm, A/P +1.7 mm, and D/V −7.5 mm. Using dental cement and two fixed small stainless-steel screws, the guide cannula assemblies were securely affixed to the skull. The screws were carefully inserted to a deliberate shallow depth to prevent potential damage to the brain. A dummy cannula was positioned to prevent blockage. A dust cap was also screwed onto the top of the cannula assemblies. Following the surgical procedure, rats were placed in a heated chamber for approximately 1 h to recover from the anesthesia.

The health condition of the rats was closely monitored after surgery. Prior to the start of the experiments, at least 10 days of recovery was allowed. To perform the infusions, stainless-steel injector cannulas (28 gauge; Plastics One) fixed to 5-μL syringes (Neurostar Co., Sindelfingen, Germany) and polyethylene-10 tubing were used. Viral supernatant (1 μL/side) was injected into the bilateral NAc shells over a period of 120 s. Following the completion of the infusion, the injection needle was maintained in place for 10 min.

Constructed and encapsulated by Shanghai GeneChem Co., Ltd, replication-deficient adeno-associated virus (AAV) carrying the gene for Rab10 was used in this study. The mean titer of the viral stocks, derived from transfected HEK 293 cells, was measured at 7.0 × 10⁹ infectious units/ml. For viral injection, 1 µL of the concentrated AAV solution was injected into the rat NAc shell. Polybrene (5 μg/mL) was included in the viral solutions to improve infection.

One day before experiments, rats were introduced to an activity cage (100 × 100 × 40 cm) for a 30-min habituation. Right after the injection of cocaine (15 mg/kg, i.p.), the locomotor activity of the rats was assessed in activity cages (Panlab Harvard Apparatus, Spain). Open-field tests (OFTs) were conducted to assess locomotor activity for 7 successive days after treatment. Cameras were fitted right above each cage and linked to a computer that had SMART software installed (version 3.0.01, Panlab S.L.U). Rats were allowed to perambulate in the activity cage under illuminated conditions for 10 min. As rats broke the infrared beam path, their movement was tracked, and the connected computer recorded the total distance traveled.

Adhering to the National Academy of Sciences’ guidelines for the care and use of laboratory animals, adult pregnant female mice were euthanized by asphyxiation in a CO₂ chamber. Consistent with prior research, NAc cells harvested from the brains of embryonic day 18 mice were rinsed in cold HBSS (Shi and Rayport, 1994). Incubation of the cells in 2 mL HBSS was followed by their dissociation using 2 mL of trypsin-EDTA (Sigma) for 12 min at 37°C. To terminate the dissociation, 2 mL of inoculation medium was applied, containing neurobasal growth media, 1% B27, 5% fetal bovine serum (FBS), and 1% glutamic acid. A 5-min centrifugation process was performed at 1,000 r/min. For plating, NAc cells were adjusted to a density ranging from 10⁵ to 10⁶ in the growth medium. The growth medium, which contains the components neurobasal growth media, 1% B27, and 1% glutamic acid, was replaced every 3 days. Cells were utilized after 14 days of growth.

Derived from GAD67 ⁺ mice at embryonic day 18, 48 NAc cultures were measured after exposure to different treatments: 1 μM cocaine for 5 or 10 min, 100 μM baclofen for 30 min, and saline. Each group consisted of four cultures and was replicated three times. The 48 NAc cultures obtained from wild-type (WT, Nestin⁺; Rab10 ^+/+) and Rab10-deficient (Nestin⁺; Rab10 ^F/F) mice underwent measurement subsequent to saline or cocaine (1 μM, 5 min post-treatment) treatment, with four cultures per group across three replicates. For membrane immunofluorescent labeling, live neurons were subjected to 15-min incubation at 37°C with antibodies against GABA_B1R, GABA_B2R, and Rab10. After being exposed to cocaine or baclofen at 37°C, the cell slices underwent a 10-min fixation in 4% paraformaldehyde in PBS, pH 7.4 with no permeabilization. Next, they were washed twice in PBS for 5 min. Subsequently, a 10-min incubation at room temperature was applied to the cell slices using a 3% solution of hydrogen peroxide (H₂O₂), followed by two rounds of washing in PBS for 5 min each. Fluorescence-conjugated secondary antibodies (Alexa 647 and Alexa 568, 1:400, Invitrogen) were applied in an appropriate amount at room temperature for 120 min, followed by 5 min of PBS washing.

To determine the fluorescence intensity of the membrane, confocal microscopy was utilized with a ×20 objective. Pinhole, gain, and laser were set to the same levels across all images. Using a NIKON TiE or A1R Laser Scanning Confocal Microscope, 24 images were captured from the NAc cell cultures of three different animals, all at an identical focal level. Captured at 1-μm intervals, images were reconstructed into three dimensions (3D) that incorporated 40 to 50 Z-stacks. The receptors’ membrane fluorescence images were captured from the monolayer cell membrane at peak intensity, and the images were then processed using Fiji.

Rab10 membrane expression was measured in NAc cells from GAD₆₇ ⁺ mice with the same treatment as immunofluorescence. Using trypsin-EDTA (Sigma), the cell disassociation solution for flow cytometry analysis was prepared, with cell counts conducted at 37°C after treatment. Subsequently, the cells underwent fixation in a solution of 0.5% paraformaldehyde in PBS and were incubated with antibodies specific for Rab10 and GFP at 4°C for overnight. Incubation of the cells was carried out for 2 h at room temperature with fluorescence-conjugated secondary antibodies (Alexa 568 and Alexa 488, 1:400, Invitrogen). Cell analysis was conducted using an LSR II Flow Cytometer from BD Biosciences (San Jose, CA) with FACSDiva software.

Immunoendocytosis analysis was conducted as previously described (Terunuma et al., 2010; Meng et al., 2018). In accordance with methods described in other research, NAc cells derived from wild-type and Rab10-deficient mice were subjected to transfection of MYC-GABA _B1R with the help of calcium phosphate (Jiang and Chen, 2006). The living cells were incubated with MYC antibodies in the culture medium for 15 min at 37°C. They were then washed and placed into a hypotonic solution (a 1:1 mix of Dulbecco’s modified Eagle’s medium and water) for 5 min at 37°C. Next, they were placed in an isotonic medium devoid of KCl for half an hour at 37°C. The living cells were subjected to no treatment or a 5-min exposure to cocaine. Then, the cells underwent fixation and were processed for immunofluorescence. A 10% solution of goat serum in PBS was applied for 60 min to prevent non-specific binding. Prior to permeabilization, GABA_B1R’s membrane expression was visualized by MYC and secondary antibodies labeled with Alexa 568 (Invitrogen) at a 1:400 dilution. After permeabilization, the internalized receptors were visualized using secondary antibodies labeled with Alexa 488 (Invitrogen) at a 1:400 dilution. Employing the same method as immunofluorescence, confocal images were acquired to quantify fluorescence intensities. The maximum-intensity images from several stacks were acquired and processed using Fiji software.

NAc cells from Rab10-deficient and wild-type mice were transfected with GABA_B1R-pHluorin using calcium phosphate (Fu et al., 2020). For the elimination of fluorescent spots, the somas of neurons expressing pHluorin were subjected to 470 nm pulsed blue light stimuli at 2 Hz, with each pulse lasting 5 ms. Each experiment consisted of 10 pulses with 1-min intervals between them. Then, imaging analysis was immediately performed. Using the NIKON FN1 Confocal Microscope (NIR Apo 40x DIC Water N.A. 0.8), the fluorescence dynamics of GABA _B1R-pHluorin (excitation 488 nm; EM: 500–550 nm) were captured, with images recorded at 1-s intervals and processed using Fiji software.

Cultured NAc cells were moved to a recording chamber for continuous perfusion at a rate of 2–4 mL/min with oxygenated ACSF, with a composition of (in mM) 125 NaCl, 2.5 KCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 1.3 MgCl₂, 2.5 CaCl₂, and 11 glucose at 30°C. For visualization, a ×40 water-immersion objective on an upright fluorescence microscope (BX51WI; Olympus), complete with infrared-differential interference contrast video microscopy and epifluorescence (Olympus), was utilized. The patch pipettes, exhibiting a resistance of 4–6 M Ω, were crafted from G150TF-4 borosilicate glass (Warner Instruments). These pipettes were loaded with an internal solution comprising (in mM) 1 EGTA, 130 CsCl, 0.2 NaGTP, 2 MgATP, 10 HEPES, and 0.1% neurobiotin, pH 7.35, with an osmolarity ranging from 270 to 285 mOsm.

The internal solution for the measurement of mIPSCs included (in mM) 10 BAPTA, 115 potassium methylsulfate, 20 NaCl, 1.5 MgCl₂, 10 sodium phosphocreatine, 0.4 NaGTP, and 4 MgATP, at pH 7.35 and osmolarity of 285 mOsm. To block AMPA, GABA_A, and NMDA receptors, respectively, recordings were made in the presence of 20 µM CNQX (6-cyano-7-nitroquinoxaline-2,3-dione, MCE), 50 µM picrotoxin (Sigma), and 50 µM D-AP5. To isolate monosynaptic inputs, 1 μM tetrodotoxin (TTX, MCE) was added to the bath solution for inhibiting voltage-gated sodium channels (Petreanu et al., 2009). Recordings of electrophysiological signals were conducted at a temperature of 32°C using a MultiClamp 700B Amplifier. The signals were digitized using a Digidata 1440A Digitizer, sampled at a frequency of 10 kHz and filtered at 2 kHz. Data acquisition was carried out using pCLAMP software from Molecular Devices. mIPSCs were recorded at −70 mV. Regarding the pharmacological experiments, drugs were applied in the bath for durations ranging from 5 to 10 min. This included 50 μM picrotoxin for the inhibition of GABA_A receptor-mediated currents and 100 μM baclofen from Sigma to activate GABA_B receptors. Prior to baseline recording, cells were given 3 min to equilibrate after achieving the whole-cell configuration. Baseline responses were recorded for 5 min. Clampfit (Molecular Devices) and Mini Analysis Program (Synaptosoft) were used to analyze electrophysiological data. Access resistance was <25 MΩ, and cells exhibiting excessive fluctuations in access resistance (>20%) were not included in the analysis. The detailed parameter settings are consistent with those reported in the study conducted by Yang et al. (2018). Each recording lasted for 5 min, and a random 60-s segment was extracted from each for data analysis. Each experiment yielded six valid data points, and the same experiment was conducted three times. Given that the GAD₆₇-GFP knock-in mice in this study, together with the previously published literature (Meredith, 1999), have demonstrated that over 90% of the NAc cells are GABAergic neurons, no steps were taken to identify the cell type in the electrophysiological experiments.

After isolation and plating, the NAc cells were prepared for Ca²⁺ analysis with the GCaMP system through calcium phosphate-mediated transfection, a method described by Jiang and Chen (2006). After NAc cells were treated with saline, cocaine (1 μM, 5 min post treatments), and/or baclofen (100 μM, 30 min post-treatments), cells in 48-well plates were live-imaged after 50 mM KCl treatments using a Nikon FN1 Confocal Microscope (NIR Apo 40× DIC Water, N.A. 0.8). High K⁺ elicited Ca²⁺ influx by depolarizing the cells’ membrane potential. The images were analyzed for the fluorescence dynamic of GCaMP (excitation 488 nm; EM: 500–550 nm). The images underwent processing using Rainbow RGB of Fiji software to create pseudo-color images. Relative fluorescence intensity is represented as ΔF/F, with F indicating the initial fluorescence after treatments with high K⁺ (50 mM KCl) and ΔF, indicating the fluorescence change after treatments with high K⁺ combined with cocaine and/or baclofen.

SPSS 20.0 and GraphPad Prism 6 (GraphPad software) were used for statistical analysis. To evaluate statistical differences, parametric statistical tests, including unpaired student’s t-tests, one-way analysis of variance (ANOVA), two-way ANOVA, or mixed-effects models, were used. Once a substantial main effect was identified via ANOVA, Bonferroni’s post hoc test or Dunnett’s post hoc analysis was employed. Dunnett’s multiple comparison test was employed to compare multiple experimental groups with a single control group, and Bonferroni’s multiple comparison test was applied for pairwise comparisons among multiple groups. Statistical significance was set as p < 0.05. Data are presented as the mean ± SEM.

In order to understand how Rab10 in the NAc regions may participate in response to cocaine treatment, we first sought to determine differences in its expression patterns with or without cocaine treatment. We therefore examined previously generated RNAseq data from mouse NAc treated with cocaine or saline (GEO dataset GSE18751; unpaired t-tests; false discovery rate (FDR) p < 0.05 and |fold change| > 1.5; Figure 1A) and found that Rab10 was transcribed at significantly higher levels in the NAc of the cocaine-group compared to the control group (Figure 1B, p = 0.032). Further examination of the RNAseq data to identify genes that are more likely to affect cocaine addiction revealed 367 differentially expressed genes between Rab10 high- and low-expression groups (unpaired t-tests; FDR p < 0.05 and |fold change| > 1.5; Figure 1C). By performing functional- and pathway-based enrichment analysis for those genes, we found 56 significantly enriched pathways (Figure 1D). Among these pathways, membrane-bounded organelle and cation-binding function suggested that membrane binding could contribute, along with Rab10, to the role of cocaine-associated behavioral effects.

The repeated cocaine administration at a dosage of 15 mg/kg led to hyperlocomotion on day 1, induced behavioral sensitization on day 3, and peaked on day 5 (Figure 1E), which was consistent with previously published findings (Peng et al., 2014; Hu et al., 2015; Xiong et al., 2018). Based on the results of Rab10 induction under cocaine treatment, we next used Western blot analysis to measure the expression of Rab10 on the plasma membrane of rat NAc tissue sampled on days 0, 1, 3, 5, and 7 of cocaine treatment. We found that Rab10 exhibited significant, time-dependent changes in its membrane expression levels during exposure to cocaine (Figure 1F). The expression levels of Rab10 were generally elevated within day 1 of initiating the cocaine treatment regimen, which was consistent with the results of differential expression analysis. However, Rab10 levels subsequently decreased in the following time points, suggesting that it may play different roles in early (1 day and earlier) and late (1 day later) responses (i.e., acute versus chronic responses) to cocaine treatment.

To determine whether Rab10 is related to behavioral effects induced by cocaine, we generated the AAV-Rab10-siRNA-GFP construct, which was microinjected into rat NAc regions 7 days before the administration of saline or cocaine (Figure 2A). With the extremely high density of Nissl bodies observed around the NAc shell, the position of the implanted cannula within the NAc was inferred through Nissl staining (Supplementary Figure S1A). Membrane expression of Rab10 decreased for 7, 10, and 14 days, with its lowest expression at 7 days (Supplementary Figure S1B).

After infection, data collected in the OFT for 7 successive days indicated that after saline treatment, rats expressing AAV-Rab10-siRNA showed similar locomotor activity compared to AAV-vehicle rats (Figure 2B, Table 1), suggesting that Rab10 deficiency alone did not significantly impact locomotor activity under saline-treated physiological circumstances. In rats treated with AAV vehicles, cocaine administration significantly elevated locomotor activity and induced behavioral sensitization since day 3 (Figure 2B, Table 1). Moreover, in the cocaine-administered groups, the expression of AAV-Rab10-siRNA markedly reduced cocaine-elevated locomotor activity, and the development of behavioral sensitization was effectively blocked (Figure 2B, Table 1), indicating that Rab10 deficiency could inhibit cocaine-elicited hyperlocomotion and behavioral sensitization.

Our earlier findings indicated that cocaine inhibited the membrane expression of GABA_BR in NAc regions by enhancing the phosphorylated calcium/calmodulin-dependent protein kinase II (pCaMKII–GABA_B1R interaction) (Lu et al., 2023). Thus, after 5 days of cocaine treatment, we collected samples for Western blot analysis and analyzed the membrane expression of GABA_BR, CaMKII, and pCaMKII in NAc neurons transfected with AAV-Rab10-siRNA. We found significantly lower membrane protein levels of GABA_B1R and GABA_B2R and the significantly higher protein level of pCaMKII induced in the NAc neurons transfected with AAV-Rab10-siRNA than those in AAV-vehicle neurons (Figures 2C, D). However, during cocaine treatments, the membrane expression of GABA_B1R and GABA_B2R was significantly increased, along with the decrease in pCaMKII, in the neurons transfected with AAV-Rab10-siRNA compared to that in neurons transfected with the AAV vector control (Figures 2C, D). These differential responses suggested that Rab10 could modulate the GABA_BR membrane expression of NAc neurons during cocaine-induced hyperlocomotion and behavioral sensitization.

In vivo experiments have shown that cocaine treatment can affect the expression of Rab10, and the effect varies depending on the time of treatment. Next, we explored the effect of cocaine treatment on the expression of Rab10 in NAc GAD ₆₇ ⁺ neurons in vitro. We isolated GAD ₆₇ ⁺ neurons from the NAc of GAD₆₇-GFP knock-in mice and treated them in culture with cocaine (1 μM). We then examined Rab10 expression at 5 and 10 min of cocaine exposure. Immunofluorescent labeling with confocal microscopy (Figures 3A, B) and flow cytometers (Figures 3C, D) and Western blot analysis (Figures 3E, F) revealed significantly decreased Rab10 membrane fluorescence, as well as decreased protein levels of Rab10. At the same time, the treatment of the GABA_BR agonist baclofen led to a high induction of Rab10 expression. Based on these data, we suggested that Rab10 likely regulates the membrane expression of GABA_BR.

We next investigated whether and how Rab10 contributes to the membrane expression of GABA_BR. Western blot (Figures 4A, B) and immunostaining (Figures 4C, D) showed that membrane levels of GABA_B1R and GABA_B2R were both significantly lower in the NAc neurons in Rab10-deficient mice than in WT mice. Immunoendocytosis showed that GABA_BR appeared to accumulate throughout the cells of the WT mice, including the plasma membrane and perinuclear regions (Figure 4E). Moreover, in Rab10-deficient mice, GABA_B1R internalization was observed at the plasma membrane, while the internalized GABA_BR is located in the cytoplasmic edges (Figure 4E). Furthermore, a decrease in the GABA_B1R-pHluorin signal (Figures 4F, G) could be observed in the NAc neurons of Rab10-deficient mice compared to that in WT mice. These findings indicated that Rab10 could serve as a regulator of GABA_BR membrane expression.

In the cocaine-treated group, although the fluorescence levels of GABA_B1R and GABA_B2R were not significantly higher in the NAc neurons of Rab10-deficient mice compared to WT mice (Figures 4C, D), an obvious increase in the expression of GABA_B1R and GABA_B2R protein was identified by Western blot analysis (Figures 4A, B). Immunoendocytosis data showed that under cocaine treatment, the WT mice showed GABA_B1R internalization at the plasma membrane, with the majority of internalized GABA_BR remaining spatially restricted in the proximity of the plasma membrane (Figure 4E). However, in Rab10-deficient mice, GABA_BR appeared to accumulate throughout the whole cells, including the plasma membrane and perinuclear regions (Figure 4E). Moreover, in cocaine-treated NAc neurons of Rab10-deficient mice, an increase in GABA_B1R–pHluorin signals could be observed (Figures 4F, G). These results indicated that Rab10 could inhibit the membrane expression of GABA_BR in cocaine-treated neurons and further suggested differences in the NAc neurons’ response to saline-associated homeostasis or cocaine-associated allostasis.

We recorded mIPSCs to measure GABA_BR function in NAc from wild-type (WT, Nestin⁺; Rab10 ^+/+) and Rab10-deficient (Nestin⁺; Rab10^F/F) mice. NAc cells were categorized into four groups based on the absence or presence of Rab10 knockout and cocaine treatment. The frequency of mIPSCs was significantly lower in Rab10-deficient mice than in WT mice (Figures 5A, C), but Rab10 deficiency did not affect the amplitude and dynamic properties of mIPSCs (Figures 5B, D, E). To further characterize the effects of Rab10 deficiency on GABA_BR in vitro, we conducted patch-clamp recordings of mIPSCs in NAc from WT and Rab10-deficient mice. The results indicated that the amplitude and decay constant of mIPSCs were significantly higher in the Rab10-deficient group than those of the control group (Figures 5F, G, J), indicating that under cocaine treatment, the deficiency of Rab10 leads to an enhancement of neural inhibition at the postsynaptic area, which is consistent with the increased membrane expression of GABA_BR observed in our Western blotting and immunoendocytosis. The frequency and 10%–90% rise time showed no significant change (Figures 5H, I).

To further determine the impact of Rab10 deficiency under cocaine treatment on the function of GABA_BR, we measured changes in GCaMP6 fluorescence in these WT and Rab10-deficient neurons treated with cocaine combined with baclofen. The results showed that cocaine could increase Ca²⁺-influx evoked by high K⁺ in both WT and Rab10-deficient mouse NAc, and no significant difference was observed between the two groups (Figures 5K, L). These experiments thus demonstrated that cocaine amplifies the Ca²⁺ influx. Furthermore, baclofen-treated Rab10-deficient neurons exposed to cocaine exhibited a substantially lower Ca²⁺ influx than WT neurons (Figures 5K, M). These results thus demonstrated that Rab10 is a key positive regulator of the baclofen-amplified Ca²⁺-influx process in cells treated with cocaine.