Ginsenoside Rg1 ameliorates stress-exacerbated Parkinson’s disease in mice by eliminating RTP801 and α-synuclein autophagic degradation obstacle

Abstract

Emerging evidence shows that psychological stress promotes the progression of Parkinson’s disease (PD) and the onset of dyskinesia in non-PD individuals, highlighting a potential avenue for therapeutic intervention. We previously reported that chronic restraint-induced psychological stress precipitated the onset of parkinsonism in 10-month-old transgenic mice expressing mutant human α-synuclein (αSyn) (hαSyn A53T). We refer to these as chronic stress-genetic susceptibility (CSGS) PD model mice. In this study we investigated whether ginsenoside Rg1, a principal compound in ginseng notable for soothing the mind, could alleviate PD deterioration induced by psychological stress. Ten-month-old transgenic hαSyn A53T mice were subjected to 4 weeks’ restraint stress to simulate chronic stress conditions that worsen PD, meanwhile the mice were treated with Rg1 (40 mg· kg⁻¹ ·d⁻¹, i.g.), and followed by functional magnetic resonance imaging (fMRI) and a variety of neurobehavioral tests. We showed that treatment with Rg1 significantly alleviated both motor and non-motor symptoms associated with PD. Functional MRI revealed that Rg1 treatment enhanced connectivity between brain regions implicated in PD, and in vivo multi-channel electrophysiological assay showed improvements in dyskinesia-related electrical activity. In addition, Rg1 treatment significantly attenuated the degeneration of dopaminergic neurons and reduced the pathological aggregation of αSyn in the striatum and SNc. We revealed that Rg1 treatment selectively reduced the level of the stress-sensitive protein RTP801 in SNc under chronic stress conditions, without impacting the acute stress response. HPLC-MS/MS analysis coupled with site-directed mutation showed that Rg1 promoted the ubiquitination and subsequent degradation of RTP801 at residues K188 and K218, a process mediated by the Parkin RING2 domain. Utilizing αSyn A53T⁺; RTP801^−/− mice, we confirmed the critical role of RTP801 in stress-aggravated PD and its necessity for Rg1’s protective effects. Moreover, Rg1 alleviated obstacles in αSyn autophagic degradation by ameliorating the RTP801-TXNIP-mediated deficiency of ATP13A2. Collectively, our results suggest that ginsenoside Rg1 holds promise as a therapeutic choice for treating PD-sensitive individuals who especially experience high levels of stress and self-imposed expectations.

Introduction

Parkinson’s disease (PD) is a prevalent neurodegenerative disorder affecting millions of people worldwide [1, 2]. Aggregated α-synuclein (αSyn), a pivotal component of Lewy bodies, represents the central pathological hallmark of PD. Disease-modifying strategies targeting αSyn, such as anti-αSyn antibodies, have been extensively explored for use in PD therapy [3, 4]. However, the disappointing clinical outcomes highlight the urgent need for alternative therapeutic interventions. Emerging evidence suggests that psychological stress contributes to the onset of dyskinesia in non-PD individuals [5] and to PD progression [6, 7]. A study of 358 PD patients revealed a significant correlation between increased COVID-19-related emotional distress and increased severity of PD symptoms [8]. This finding underscores the clinical need to address the impact of psychological stress on the exacerbation of PD symptoms [9, 10]. In previous research, we demonstrated that chronic restraint-induced psychological stress precipitated the onset of parkinsonism in 10-month-old transgenic mice expressing mutant human αSyn (hαSyn A53T) [11]. We refer to these mice as chronic stress-genetic susceptibility (CSGS) PD model mice, and these mice were utilized in the present study to explore treatment strategies and elucidate the underlying mechanisms responsible for the exacerbation of PD symptoms induced by psychological stress.

The hypothalamic‒pituitary‒adrenal (HPA) axis orchestrates the stress response in rodents. Elevated corticosterone (CORT) levels have been observed in mice with psychological stress-induced PD exacerbation [12], along with the activation of stress-responsive proteins such as FKBP5, ERRFI1, and RTP801. These proteins are pivotal in maintaining cellular equilibrium under stress. RTP801, encoded by the stress-responsive gene DDIT4, is inducible by glucocorticoids or stress and acts as a glucocorticoid receptor target. Mice deficient in RTP801 exhibit resistance to numerous stress-induced pathological conditions, and the adverse effects of glucocorticoids are mitigated in these animals [13]. Remarkably, RTP801 is significantly upregulated in the dopaminergic neurons of individuals with PD [14]. However, the mechanisms that trigger this upregulation, the pathways through which it contributes to neuronal injury, and the temporal dynamics of its detrimental effects remain poorly characterized. The removal of accumulated αSyn is a critical process in curtailing the progression of PD pathology [15–17]. Our previous research revealed that RTP801 exacerbates PD progression under psychological stress by impeding autophagy-mediated αSyn clearance [11]. Interestingly, RTP801 has been reported to initiate autophagy via the inhibition of mTOR signaling [18], the specific underlying mechanisms require further investigation.

Ginseng, derived from the root and rhizome of Panax ginseng C.A. Mey, has been esteemed for millennia in traditional Chinese medicine for its ability to soothe the mind and enhance cognitive function [19]. Ginsenoside Rg1, a principal active component of ginseng, has demonstrated efficacy in treating mental disorders, including PTSD-like behaviors [20] and depression-like behavior [21]. Additionally, Rg1 attenuates motor impairment and neuroinflammation in MPTP- and probenecid-induced parkinsonism mouse models by targeting abnormal αSyn in the substantia nigra pars compacta (SNc) [22]. These findings collectively support that Rg1 might represent an effective intervention for psychological stress-exacerbated PD.

In this study, using a combination of behavioral, neuroimaging, electrophysiological, and genetic analyses, we provided evidence that Rg1 significantly ameliorated both motor and nonmotor symptoms in a psychological stress-induced PD mouse model, accompanied by improved brain connectivity and neural activity. Mechanistically, we discovered that psychological stress leads to an increase in RTP801 due to insufficient ubiquitin-mediated degradation, which elevates thioredoxin-interacting protein (TXNIP) levels and results in the formation of the RTP801-TXNIP complex. This complex contributes to ATP13A2 deficiency, disrupting the autophagic clearance of αSyn aggregates. The accumulation of αSyn creates a feedback loop, maintaining high RTP801 levels. Overexpressing RTP801 in the SNc mimics the damage observed in psychological stress-induced PD. Rg1 treatment disrupts this cycle by promoting RTP801 degradation via the Parkin E3 ligase and increasing autophagic flux, suggesting its potential as a disease-modifying therapy for psychological stress-related PD exacerbation.

Materials and methods

Animals

Human αSyn A53T-overexpressing mice [B6; C3-Tg (Prnp-SNCA *A53T)83Vle/JNju] were obtained from the Model Animal Research Center of Nanjing University. RTP801 knockout (RTP801^−/−) mice were generated with CRISPR-Cas9 technology by the Shanghai Model Organisms Center, Inc. Specifically, Cas9 mRNA and guide RNA (gRNA) were synthesized in vitro and subsequently microinjected into C57BL/6J mouse oocytes to generate founder (F0) mice. The breeding of F1 progeny yielded homozygous RTP801^−/− offspring. To generate double mutants, αSyn A53T mice were crossbred with RTP801^−/− mice, resulting in αSyn A53T⁺; RTP801^−/− progeny (Fig. 5c). These mice, along with their C57BL/6J WT counterparts, were housed at SPF Biotechnology Co., Ltd. until they reached the age of 10 months. All the mice were housed in a controlled environment at a stable temperature of 22 ± 1 °C, with the relative humidity maintained at 55% ± 10% under a 12 h light/dark cycle. The animals had ad libitum access to food and water. Rg1 (B21057, Shanghai Yuanye Bio-Technology, HPLC ≥ 98%) was administered orally at a dosage of 40 mg/kg daily for 4 weeks. To evaluate autophagic flux, chloroquine (CQ, S6999, Selleck Chemicals, 60 mg/kg) was administered intraperitoneally 24 h prior to euthanasia.

Fig. 5. RTP801 is a pivotal factor in stress-induced PD susceptibility and the therapeutic potential of Rg1.

a SH-SY5Y-αSyn A53T cells transfected with LV-RTP801-shRNA or control LV-NC-shRNA were treated with or without Rg1 following CORT exposure. Quantification of RTP801 and αSyn accumulation (n = 4). b Quantification of αSyn in soluble and insoluble fractions (n = 4). c, d Generation of RTP801^−/− mice and αSyn A53T⁺; RTP801^−/− mice, detailing the genetic strategy and experimental protocols. e‒h Motor function assessment via the rotarod, pole, beam walking, and grip strength tests (n = 6). i IF depicting TH and RTP801 colocalization (upper panel; scale bar: 50 μm) and IHC of RTP801 in the SNc (lower panel; scale bar: 200 μm). j Western blotting analysis of RTP801, TH, and 5G4, displaying quantitative data (n = 4). k ThS staining and quantitative analysis of ThS⁺ aggregates in the SNc (n = 4). Scale bar: 20 μm. Data are presented as mean ± SEM, ***P < 0.001 vs. CORT/CSGS, ns not significant.

All animal care and experimental procedures adhered to the ethical guidelines established by the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by the Animal Care and Use Committee of the Peking Union Medical College and the Chinese Academy of Medical Sciences (Ethical inspection No. 00009278).

Chronic stress-genetic susceptibility (CSGS)-stimulated PD model

We utilized 10-month-old transgenic mice expressing human αSyn A53T to model PD with chronic stress. The chronic restraint stress (CRS) procedure was adapted from a previously described protocol [11, 23]. The mice were confined in size-adjusted restraints for 4 h daily, excluding Sunday, for 4 weeks. The number of restraint sessions varied randomly between 08:00 and 20:00 to prevent habituation. Care was taken to avoid animal pain or distress during confinement.

Functional magnetic resonance imaging (fMRI) data acquisition and processing in mice

Data acquisition

Mice were anaesthetized with 1.5% isoflurane in 70%/30% N₂/O₂ before undergoing fMRI. Their heads were immobilized using a dedicated body restrainer equipped with tooth and ear bars. Scans were performed with a PharmaScan 70/16 MRI spectrometer (Bruker, Germany) using a mouse head surface coil. T2*-weighted gradient-echo EPI sequences were captured using the following parameters: 20 × 20 mm² FOV, 2000 ms TR/15 ms TE, 64 × 64 matrix, and 40 slices 0.35 mm in thickness spaced 0.05 mm apart. Data were collected via ParaVision_6.0.1 (Bruker BioSpin GmbH).

Data processing

Functional brain images were converted to NIFTI format via dcm2niix and upsampled by a factor of 10 using DPABI software [24]. The central slice was used for temporal slice-timing correction. Head motion correction was performed by normalizing each subject’s volume to the mean volume across 300 realigned images. Spatial normalization to standard space was conducted with the ANT toolkit (v1.9.2) using a two-step registration to the Turone Mouse Brain Atlas via SPM12 [25]. Gaussian smoothing (6 mm kernel) and temporal filtering (0.01–0.08 Hz) were applied using RESTplus V1.2.8-130615. Following covariate regression and detrending, fast Fourier transformation was applied to the preprocessed time series to compute the power spectrum. The amplitude of low-frequency fluctuations (ALFF) was calculated for each voxel. Group differences in ALFFs were assessed using a two-sample t test, whereas regional homogeneity (Reho) was determined using Kendall’s coefficient of concordance (KCC) as a local coherence metric of the BOLD signal without smoothing of any neighboring voxels [26]. A threshold of P < 0.05 (minimum cluster size: 10 voxels) was set for statistical significance.

Functional connectivity analysis

Region of interest (ROI)-based functional connectivity was analyzed by averaging voxel wise correlations within each ROI, following Fisher’s z transformation. Thirty ROIs (15 per hemisphere) were defined according to the Turone Mouse Brain Atlas. ROI-ROI connectivity significance was evaluated via a two-sample t-test with FDR correction (P < 0.05) via the GRETNA toolbox [27]. Statistical analyses were conducted in MATLAB R2013b (The MathWorks, Inc.).

In vivo electrophysiological recording and analysis

The mice were initially anaesthetized with a continuous flow of 1.5% isoflurane in a 70%/30% N₂/O₂ mixture and subsequently positioned within a stereotaxic apparatus (68002, RWD, Shenzhen, China). To safeguard the eyes during the procedure, a thin layer of lubricating ointment was applied. A skin flap of approximately 1.5 cm² was delicately retracted from the cranial surface, followed by careful removal of the periosteum using fine scissors. The exposed skull was thoroughly cleaned and dried with sterile cotton swabs to prepare for electrode placement. A 32-channel microelectrode array (Bio-Signal Technologies, Jiangsu, China) was inserted perpendicularly into the dorsal striatum at coordinates. The recording coordinates were +0.6 mm AP to +0.7 mm AP and ±1.5 mm ML to ±2.0 mm ML from bregma and −2.5 mm DV to −4.0 mm DV from the top of the brain. The electrode tip was coated with Dil cell labeling solution (Beyotime, Shanghai, China) to enable subsequent histological confirmation of the electrode trajectory, as indicated by red fluorescence (Fig. S3a). Reference and ground electrodes were secured around a screw embedded in the skull, and dental acrylic was applied to anchor the head plate firmly. Following a 7-day recovery period, electrophysiological recordings commenced. The signals were amplified and filtered with local field potential (LFP) signals sampled at 2 kHz (low-pass filter at 250 Hz) and spiking activity sampled at 30 kHz. Spike detection employs bandpass filtering between 250 and 5000 Hz with a threshold of −40 μV to discriminate neuronal spikes from background noise.

Neuronal populations, specifically putative medium spiny projecting neurons (MSNs) and fast-spiking interneurons (FSIs) were distinguished and isolated (Figs. 1n and S3b). Neuronal activity was recorded over a continuous 10-min duration. A total of 120 neurons (60 MSNs and 60 FSIs) from six mice were recorded. The curated spike data were analyzed using NeuroExplorer (Plexon Inc., Texas, USA) to assess alterations in firing rates, with the results presented in rate histograms and burst analysis. The LFP signals were analyzed using power spectral densities (PSDs) and spectrograms.

Fig. 1. Rg1 confers protection against psychological stress-induced Parkinson’s disease.

a Schematic of the experimental procedure. b Olfactory dysfunction was assessed using BFPT. c A modified open field test with nuts or paprika was used to quantify olfactory acuity. Motor coordination and balance were evaluated via the rotarod (d), beam walking (e), and pole (f) tests. g Forelimb grip strength was gauged using the grip strength test. h The OFT appraised anxiety-related behaviors, tracking total distance traveled, and center zone activity (n = 8 for the CSGS group, n = 7 for the CSGS + Rg1 group). i Rg1 administration led to significant increases in the ALFF and Reho values compared with those of the CSGS group (P < 0.05, FDR at the cluster level, cluster size = 10 voxels, n = 8 for the CSGS group, n = 7 for the CSGS + Rg1 group). Enhanced brain regions are color-coded based on t values. j Correlations between the ALFF or Reho in the CP and pole test. k Correlation between ALFF or Reho in the SNc and pole test. l Network visualizations displaying the connected ROIs. m Radar chart illustrating the functional connectivity variations between the CP_R and additional regions across the CSGS and CSGS + Rg1 groups (n = 5). n Representative traces of MSNs and FSIs. o Quantitative analysis of MSN firing activity including firing rate, number of spikes in bursts and burst duration (n = 60 neurons from 6 mice). p Quantitative analysis of FSI firing activity including firing rate, number of spikes in bursts and burst duration (n = 60 neurons from 6 mice). q Representative power spectrum from an individual mouse. Population power spectra from the striatum. Statistical examination of the striatal power spectra revealed significant changes in both the β and γ bands across 18 sessions from 6 mice. Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001 vs. CSGS.

Olfactory function assessment in mice

Buried food pellet test (BFPT)

To evaluate olfactory acuity, the mice were fasted for 18 h with water ad libitum and habituated to a clean cage for 5 min. In the BFPT, a 1.0 g food pellet was randomly buried 0.5 cm deep in bedding. The mice were placed in the test enclosure, and the time to locate the pellet was recorded for 5 min, and the process was repeated daily for 3 days. For specificity, a control test (visible BFPT) was conducted with the pellet on the bedding surface to distinguish the olfactory ability from the visual ability.

Modified open field test

Olfactory function was further assessed using a modified open field test with a 50 cm cubic arena and a 3 cm³ vessel in each corner. After a 5-min exploration period, a test substance (nut or paprika) was placed in one vessel, and water was added to the other vessel. The time spent in each corner was recorded, and a preference index was calculated as the time spent in the test substance corner over the total test duration.

Rotarod test

To assess balance and motor coordination, the mice were subjected to the rotarod test. The apparatus rotation speed linearly increased from 5 to 30 rpm over a 300 s duration. Mice were tested individually, and their latency to remain on the rod was recorded. The rotarod was cleaned at regular intervals to ensure consistent performance. The latency from the start of the test to when the mouse fell off the rod was recorded. Each mouse underwent a series of three trials, with an intertrial rest period of 30 min. The average latency across trials was computed for each subject for subsequent statistical evaluation.

Pole test

The pole descent task was conducted to further assess mouse motor coordination and balance. The subjects were acclimated to the apparatus, which consisted of a 50 cm vertical pole (1 cm diameter) topped with a wooden ball and wrapped in gauze for grip enhancement. Training sessions ensured that the mice could proficiently navigate from the apex to the base of the pole. The time taken for each mouse to descend and achieve a stable landing on all four paws was meticulously timed. Following acclimatization, each mouse was subject to three trials, spaced at 30-min intervals. The average descent time was calculated for each mouse for analytical purposes.

Beam walking test

A beam walking assay was used to evaluate the proprioceptive ability and motor coordination of the mice. The mice were trained to traverse a 100 cm long, 1.5 cm wide wooden beam to reach an enclosed black box, which provided a sense of safety. Positioned 60 cm above the floor, the beam presented a challenging yet manageable task for the mice. The assay comprised three consecutive trials per mouse, with performance averages computed and used for statistical interpretation.

Grip strength

Forelimb grip strength was quantified using a digital grip strength meter (YLS-13A, Jinan, China). The mice were induced to grasp the pull bar of the apparatus using both forelimbs, and a consistent posterior pull was applied until grip release. Peak tension readings were captured in grams (g). To ensure reliability, three separate attempts were recorded per mouse, with the mean grip strength calculated for each individual.

Open field test (OFT)

The OFT was implemented to gauge anxiety-related behaviors in mice. Each subject was introduced to a 50 cm³ open field arena with a uniform white floor for a 5-min acclimatization period. Movements were captured via a Sony overhead camera and subsequently analyzed with SMART system software v2.5.3 (Panlab SL, Barcelona, Spain). Behavioral metrics, including total distance traveled, central zone activity, duration within the central zone, and fecal boli count, were quantified post experiment.

Stereotaxic adeno-associated virus (AAV) microinjection

Bilateral microinjections of AAVs were administered into the SNc at the following coordinates: AP, 3.16 mm; ML, ±1.2 mm; and DV, −4.5 mm. The viral constructs utilized in this study included pAAV-CMV-MCS-bGHpolyA-EF1A-mCherry-3FLAG (AAV-NC, titer 3.31 × 10¹³ vg/mL), pAAV-CMV-RTP801-HA-bGHpolyA-EF1A-mCherry-3FLAG (AAV-RTP801, titer 2.86 × 10¹³ vg/mL), pAKD-CMV-bGlobin-mCherry-H1-shRNA (RTP801) (AAV-RTP801 shRNA, titer 3.39 × 10¹³ vg/mL), and pAKD-CMV-bGlobin-mCherry-H1-shRNA (AAV-NC shRNA, titer 2.31 × 10¹³ vg/mL). These AAVs were sourced from Obio Technology Co., Ltd. (Shanghai, China). Briefly, 0.5 μL of the appropriate AAV solution was injected via a Hamilton syringe over a 10-min duration. To allow adequate viral diffusion, the syringe was maintained in position for an additional 10 min postinjection. The injection rate was meticulously regulated using a microinfusion pump system (TFD01, Lead Fluid, Hebei, China) affixed to the stereotaxic frame.

Cell culture and treatment

SH-SY5Y cells expressing the αSyn A53T mutation, with or without mCherry, were cultured in DMEM (TransGen, Beijing, China) supplemented with 10% FBS (Gibco, California, USA) and 10 µg/mL puromycin (Beyotime, Shanghai, China) at 37 °C in 5% CO₂. Chronic stress was simulated by treating cells with 10 µM CORT (27840, Sigma-Aldrich) for 48 h [11]. The cells were also treated with Rg1 (0.1, 1, or 10 µM) in the presence of CORT for 48 h to assess neuroprotection.

PC12-Vehicle cells and PC12-RTP801 stable cells were grown in RPMI-1640 (TransGen, Beijing, China) supplemented with 10% FBS, 100 U/mL penicillin and streptomycin (TransGen, Beijing, China), and 1 µg/mL puromycin under the same conditions. These cells received 10 µM Rg1 for 48 h.

Lentivirus-mediated gene knockdown

Lentiviral vectors carrying short hairpin RNAs (shRNAs) targeting RTP801, ATP13A2, and TXNIP were obtained from Obio Technology. These lentiviruses were used to achieve gene silencing of RTP801, ATP13A2, and TXNIP in αSyn A53T-expressing SH-SY5Y cells. The target sequences used for the knockdown were as follows: RTP801, 5′-TGATGCCTAGCCAGTTGGTAA-3′; ATP13A2, 5′-GCCCGCGTCAGCCCTGCTCCA-3′; TXNIP, 5′-CTCAAGACAGCCCTATCTTTA-3′; and negative control (NC), 5′-TTCTCCGAACGTGTCACGT-3′.

Immunohistochemistry (IHC) and immunofluorescence (IF)

For tissue sections

Frozen sections (20 μm) were subjected to antigen retrieval in 0.01 M citrate buffer (10 min). The sections were then permeabilized with 0.5% Triton X-100 in PBS (15 min) following a PBS wash. For IHC, endogenous peroxidase was blocked with 3% H₂O₂ (10 min). The sections were blocked with 5% BSA (30 min) and incubated with primary antibodies overnight at 4 °C. For IHC detection, after primary antibody incubation, the sections were incubated with horseradish peroxidase-conjugated secondary antibodies (KPL, Maryland, USA, 2 h, room temperature) and developed with DAB, followed by light microscopy (Olympus, Tokyo, Japan). For IF detection, the sections were incubated with Alexa Fluor-conjugated secondary antibodies (Life Technologies, California, USA, 2 h, dark) and imaged using a Cytation C10 confocal imaging reader (BioTek, Vermont, USA).

For cells

Cells on poly-L-lysine-coated coverslips were fixed with 4% paraformaldehyde (20 min), permeabilized with 0.5% Triton X-100 (15 min), blocked with 5% BSA (30 min), and incubated with primary antibodies overnight at 4 °C. The same secondary antibodies and detection protocols for IF were applied as described for the tissue sections.

Thioflavin S (ThS) staining

Tissue sections or cells were incubated with 0.01% ThS (T1892, Sigma-Aldrich, dissolved in PBS) for 10 min at 37 °C [28], followed by washing with PBS. The samples were subsequently incubated with Hoechst 33342 (1:1000) for 10 min at 37 °C and mounted with ProLong antifade reagent, and the accumulated αSyn was observed via a Cytation C10 confocal imaging reader.

Transmission electron microscopy (TEM)

The brain tissues were sectioned to ~1 mm³ and fixed in 2% paraformaldehyde with 2.5% glutaraldehyde. The cells received a similar treatment, with an additional postfixation in 1% osmium tetroxide. After dehydration, the samples were embedded in Epon resin. Finally, ultrathin sections (50–60 nm) of the SNc tissues or cell pellets were cut and stained with uranyl acetate and lead citrate. The ultrastructures were observed using a Hitachi H-7650 transmission electron microscope (Tokyo, Japan).

Preparation of soluble and insoluble fractions

The soluble and insoluble protein fractions were extracted using a commercially available soluble and insoluble protein extraction kit (Sangon Biotech, Shanghai, China). Tissues and cell homogenates were lysed in hypotonic buffer with protease/phosphatase inhibitors and DTT, followed by centrifugation at 22,800 × g for 60 min at 4 °C to obtain the soluble fraction. The pellet was subsequently resuspended and centrifuged at 22,800 × g for 10 min at 4 °C. The resulting supernatant was designated the insoluble fraction.

CORT assay

The mice were initially anaesthetized with a continuous flow of 1.5% isoflurane in a 70%/30% N₂/O₂ mixture. Blood was collected via ocular puncture between 10:00 a.m. and immediately chilled on ice. After centrifugation at 1000 × g for 10 min at 4 °C, the serum was stored at −80 °C. Serum CORT levels were quantified via ELISA kits (Henghuibio, Beijing).

Coimmunoprecipitation (Co-IP) assay

Co-IP was conducted according to established protocols [29]. First, anti-RTP801 or TXNIP antibodies were incubated with the cell homogenates at 4 °C with gentle agitation at 1000 rpm for 4 h. Subsequently, protein A Sepharose beads (GE Healthcare Bio-Sciences, Uppsala, Sweden) were added to the mixture, which was further incubated overnight at 4 °C with the same agitation. The beads were then washed three times with lysis buffer and pelleted by centrifugation at 1000 × g for 5 min at 4 °C. Loading buffer was added to the collected beads, and the protein complexes were denatured at 100 °C. Protein expression was subsequently analyzed using Western blotting. For detection, the blots were probed with antibodies specific to ubiquitin, TXNIP or RTP801.

Western blotting

Western blotting was performed as previously described [30]. Tissues and cells were lysed in RIPA buffer (Applygen, Beijing, China) containing protease and phosphatase inhibitors. Protein concentrations were measured via BCA assay kits (Beyotime, Shanghai, China). The samples were separated via 12% SDS‒PAGE and transferred to PVDF membranes. The membranes were blocked with 5% BSA in TBST and incubated with the following primary antibodies: anti-β-actin (1:10,000, AC026, ABclonal), anti-TH (1:500, sc-25269, Santa Cruz), anti-5G4 (1:1000, MABN389, Sigma-Aldrich), anti-αSyn (1:1000, ab138501, Abcam), anti-ubiquitin (1:1000, 3936, Cell Signaling Technology), anti-RTP801 (1:1000, NBP1-77321, NOVUS), anti-ERRFI1 (1:1000, A13099, ABclonal), anti-FKBP5 (1:1000, A9090, ABclonal), anti-FOXO1 (1:1000, A2934, ABclonal), anti-KLF15 (1:1000, A7194, ABclonal), anti-PARK2/Parkin (1:2000, 66674-1-lg, Proteintech), anti-NEDD4 (1:1000, 5344, Cell Signaling Technology), anti-p-AMPK (1:1000, ab133448, Abcam), anti-AMPK (1:1000, ab131512, Abcam), anti-Beclin 1 (1:2000, ab207612, Abcam), anti-LC3 (1:1000, 3868, Cell Signaling Technology), anti-SQSTM1/p62 (1:1000, 18420-1-AP, Proteintech), anti-ATP13A2 (1:1000, A13083, ABclonal) and anti-TXNIP (1:1000, A9342, ABclonal). The membranes were then incubated with HRP-conjugated secondary antibodies at room temperature for 2 h. The protein bands were visualized via an Image Quant LAS 4000 mini (GE, Boston, USA), and the band intensities were quantified via ImageJ software (Maryland, USA).

Immunoprecipitation and detection of RTP801 ubiquitination using HPLC-MS/MS

To assess the ubiquitination of RTP801, SH-SY5Y cells expressing αSyn A53T were treated with CORT with or without Rg1. Following treatment, the cells were harvested and lysed in RIPA buffer containing a protease inhibitor cocktail for 20 min on ice. The lysates were subsequently centrifuged at 20,000 × g for 10 min at 4 °C, after which the supernatants were collected. The protein concentrations in the supernatants were quantified using a BCA assay kit (Thermo Scientific). Equal amounts of total protein from each lysate were incubated with an anti-RTP801 antibody at 4 °C with gentle rotation at 1000 rpm for 4 h. Protein A Sepharose beads were then added to the antibody-lysate mixture and incubated overnight at 4 °C with gentle rotation. The beads were subsequently washed three times with lysis buffer and pelleted by centrifugation at 1000 × g for 5 min at 4 °C. The collected beads were subjected to SDS‒PAGE followed by Coomassie blue staining. The bands were excised directly from the gel and subjected to HPLC‒MS/MS, a service provided by Beijing Qinglian Biotech Co., Ltd., China. The resulting mass spectra were analyzed via Proteome Discoverer 2.4 software, referencing the UniProtKB/Swiss-Prot human protein database for identifying diglycine (GG)-modified RTP801 peptides.

Tandem fluorescent mCherry-EGFP-LC3 adenovirus transduction

PC12-RTP801 stable cells were transduced with a tandem fluorescent mCherry-EGFP-LC3 adenovirus (HANBIO, Shanghai, China) for 6 h to evaluate autophagic flux, as detailed in a previous study [31]. Following transduction, the cells were returned to regular culture medium and treated with Rg1 (10 μM). After 24 h, the cells were subjected to additional treatment with Rg1, rapamycin (Rap, S1039, Selleck Chemicals) (100 nM), or bafilomycin A1 (BafA1, S1413, Selleck Chemicals) (10 nM) for an additional 24 h. Cellular images were acquired using a Cytation C10 confocal imaging reader.

Real‑time quantitative PCR (qPCR)

Total RNA was extracted from cells using TRIzol reagent (ABclonal, Wuhan, China) in accordance with the manufacturer’s instructions. The RNA was then reverse-transcribed into cDNA via the One-Step gDNA Removal and cDNA Synthesis SuperMix Kit (TransGen, Beijing, China). Primers targeting mouse lysosome-associated genes, as listed in Table S2, were used for qPCR. The qPCRs were conducted using the TransStart Tip Green qPCR Supermix (TransGen, Beijing, China) on a LineGene 9600 Quantitative PCR System (Bioer, Hangzhou, China). The mRNA levels were quantified and normalized via the comparative CT (2^−△△CT) method, and the values were subsequently calculated using the formulas below to obtain the results shown in Fig. 7a:

$$\mathrm{CSGS}:\mathrm{value}\mathrm{CSGS}/{\left(\mathrm{value}\mathrm{CSGS}\times \mathrm{value}\mathrm{CSGS}+\mathrm{Rg}1\right)}^{1/2};\mathrm{CSGS}+\mathrm{Rg}1:\mathrm{value}\mathrm{CSGS}+\mathrm{Rg}1/{\left(\mathrm{value}\mathrm{CSGS}\times \mathrm{value}\mathrm{CSGS}+\mathrm{Rg}1\right)}^{1/2}.$$

Fig. 7. Abnormal RTP801-TXNIP-ATP13A2 pathway was mitigated by Rg1 to rescue autolysosomal dysfunction.

a Heatmap illustrating the levels of lysosome-related genes in CSGS model mice. The color bar ranges from 0.5–1.8. b Western blotting analysis of ATP13A2 (n = 4). c SH-SY5Y-αSyn A53T cells were transfected with LV-ATP13A2-shRNA or LV-NC-shRNA and treated with or without Rg1 after CORT treatment. Quantification of ATP13A2 and αSyn accumulation (n = 4). d Colocalization of LAMP1 with αSyn in SH-SY5Y-αSyn A53T cells transfected with or without LV-ATP13A2-shRNA and colocalization analysis (n = 40 cells from 4 mice). Scale bar: 5 μm. e Venn diagram illustrating the common interactors of RTP801 and ATP13A2. f Colocalization of RTP801 with TXNIP in the SNc of RTP801 OE mice and colocalization analysis (n = 40 cells from 4 mice). Scale bar: 10 μm. g Co-IP analysis demonstrating the interaction between RTP801 and TXNIP in RTP801-OE PC12 cells (n = 4). h, i Western blotting analysis of SH-SY5Y-αSyn A53T cells transfected with LV-TXNIP-shRNA or LV-NC-shRNA and treated with Rg1 post-CORT, including TXNIP, ATP13A2, and αSyn accumulation (n = 4). Data are presented as mean ± SEM, ^###P < 0.001 vs. vector; **P < 0.01, ***P < 0.001 vs. CSGS/CORT/RTP801 OE/NC, ns not significant.

Statistical analysis

The data are presented as the means ± SEM. All the statistical analyses were conducted via GraphPad Prism 8.0 (GraphPad Software Inc., La Jolla, CA, USA). Differences between two groups were analyzed via Student’s t test, whereas differences among multiple groups were analyzed via one-way ANOVA followed by Dunnett’s test or two-way ANOVA followed by Tukey’s test for multiple comparisons. P < 0.05 was considered statistically significant.

Results

Ginsenoside Rg1 confers protection in psychological stress-exacerbated PD

We previously established a CSGS model by subjecting 10-month-old human αSyn A53T-expressing transgenic mice to 4 h of CRS daily over a 4-week period. This regimen resulted in the exacerbation of PD phenotypes relative to those observed in unstressed transgenic mice [11]. To evaluate the effect of Rg1 on CSGS-induced PD, we administered Rg1 intragastrically at a dosage of 40 mg/kg per day for 28 days during the chronic stress period (Fig. 1a). Olfactory dysfunction, a typical PD symptom, was first assessed using the BFPT. Compared with the CSGS group, the Rg1 treatment significantly reduced the time taken to locate the buried food pellet. The visible BFPT excluded the possibility that the increased movement speed induced by Rg1 administration contributed to this improvement (Figs. 1b and S1a). A modified open field test with attractive or irritating substances was performed to assess olfactory discrimination. The Rg1 group presented an increased preference for an attractive substance (nuts) and a reduced preference for an irritating substance (paprika), suggesting that Rg1 improved olfactory function in the CSGS model (Fig. 1c). The sucrose preference test was employed to exclude the influence of olfactory preference abnormalities and to confirm that the stressor did not induce depression-like symptoms (Fig. S1b).

Motor dysfunction, a hallmark of PD, was evaluated via the rotarod test, beam walking test and pole test. Compared with the CSGS group, the Rg1-treated group presented improved motor coordination, as evidenced by increased rotarod endurance (Fig. 1d), expedited beam traversal (Fig. 1e), and quicker pole descent (Fig. 1f). Additionally, an increase in forelimb grip strength following Rg1 administration suggested the amelioration of PD-associated tremors and grip weakness (Fig. 1g). In addition to motor dysfunction, PD patients often experience nonmotor symptoms, including anxiety and autonomic dysfunction. The OFT revealed that Rg1 treatment reduced anxiety-like behavior induced by CSGS, as evidenced by increases in the total distance traveled, the distance traveled in the center zone, and the time spent in the center zone (Fig. 1h). Notably, an increased turn count in the OFT suggested that Rg1 enhances turning agility, a significant challenge for PD patients, which is consistent with the enhanced motor function observed in other assessments. In this OFT assessment, Rg1 treatment did not significantly alter the number of fecal pellets produced by the mice (Fig. S1c). Although OFT-induced anxiety in rodents can stimulate gastrointestinal responses, it is important to note that constipation is a common symptom in PD patients, which could confound the measurement of constipation. Collectively, these results indicate that Rg1 treatment significantly ameliorated both motor and nonmotor symptoms related to PD.

To discern whether Rg1 ameliorates brain activity rather than exerting merely peripheral effects, we employed resting-state functional magnetic resonance imaging (rs-fMRI) [32] to analyze brain activity and functional connectivity, the clinical hallmarks of PD [33]. The ALFF quantifies regional spontaneous neural activity, whereas Reho assesses the coherence and centrality of regional brain activity [34]. We initially confirmed a positive correlation between PD-related motor behaviors and ALFF/Reho in the caudate putamen (CP, dorsal striatum) and the SNc, the primary pathological sites in PD patients (Figs. 1j, k and S2a–d). We subsequently examined ALFF and Reho values to determine the influence of Rg1 on global brain activity. Notable enhancements in ALFF and Reho were detected in the olfactory areas, striatal regions and midbrain areas (Fig. 1i, brain regions shown in Table S1). Functional connectivity is crucial for understanding PD motor symptoms and progression [35]. We investigated functional connectivity between the left and right hemispheres in the above regions, via a Z-score matrix visualized as a heatmap. Elevated Z scores in the Rg1 group indicated stronger functional connectivity within PD-related networks (Fig. S2e), as illustrated via BrainNet Viewer (Fig. 1l). Statistically significant enhancements in the functional network were represented by P value matrices (Fig. S2f), as shown in Fig. S2g. Notably, with Rg1 treatment, the connectivity between the CP and olfactory regions, as well as between SNr and SNc, was significantly altered, as illustrated in the histogram (Fig. S2h, i) and radar chart (Fig. 1m). These neuroimaging results provide evidence for the role of Rg1 in improving behavioral outcomes in PD patients. We further explored other brain regions associated with the nonmotor of PD, including cognitive impairment, sleep disturbances, depression, anxiety, and pain [36–42]. In the Rg1-treated group, the connectivity strength matrix significantly increased (Fig. S2j–m), which was consistent with the alleviation of nonmotor symptoms by Rg1.

The loss of dopaminergic neurons in PD animal models directly affects the firing rate and neuronal synchrony, leading to motor dysfunctions [43]. Given the impact of Rg1 on neuronal activity in the striatum, we investigated alterations in electrical activity with a multichannel in vivo recording system. The MSNs and FSIs of the striatum play pivotal roles in motor regulation (Figs. 1n and S3a, b). The loss or inhibition of MSNs or FSIs within the striatum is postulated to precipitate dyskinesia in PD models [44, 45]. Notably, following Rg1 administration, both MSNs and FSIs in the CP presented significant increases in the firing rate, number of spikes in bursts, and burst duration (Figs. 1o, p and S3c, d). Furthermore, studies on LFP oscillations have shown that beta (β)-band oscillations (12–30 Hz) intensify during impaired locomotion, suggesting their potential as biomarkers for PD, whereas gamma (γ)-band oscillations (30–80 Hz) are implicated in movement initiation [43, 46]. Our analysis of LFPs in the striatum of Rg1-treated mice revealed a decrease in β-band activity coupled with an increase in γ-band activity relative to those of CSGS mice (Fig. 1q). These improvements in dyskinesia-related electrical activity lend further credence to the pro-motor effects of Rg1 in the CSGS model.

Rg1 ameliorates degeneration of dopaminergic neurons and pathological aggregation of αSyn

Disease-modifying treatments predominantly aim to slow the degeneration of dopaminergic neurons in PD. Rg1 effectively mitigates the degeneration of these neurons, as evidenced by increased levels of tyrosine hydroxylase (TH), a marker of dopaminergic neurons, in both the striatum and SNc (Fig. 2a, b). This observation was further supported by the TH content in the midbrain (Fig. 2c). Ultrastructural examination of the SNc using TEM revealed that Rg1 ameliorated the damage observed in the CSGS model, including nucleolus shrinkage, nuclear membrane disruption, and damage to mitochondrial cristae and myelin (Fig. 2d). Abnormal αSyn aggregation in the SNc, a hallmark of PD pathology, was significantly reduced by Rg1, as detected by thioflavin S (ThS) staining. Notable reductions in ThS density, the proportion of ThS-positive cells, and the number of cells with more than 10 aggregates in the SNc of the Rg1-treated group underscored the capacity of Rg1 to diminish αSyn aggregation (Fig. 2e). To evaluate the direct effect of Rg1 on αSyn aggregation induced by stress, we treated CORT-exposed SH-SY5Y-αSyn A53T-mCherry cells with Rg1. Colocalization analysis between ThS and αSyn revealed that Rg1 dose-dependently reduced αSyn aggregation, with 10 μM Rg1 exerting the most potent inhibitory effect (Fig. S4). The content of aggregated αSyn (5G4) in the midbrain further suggested that Rg1 alleviated αSyn aggregation (Fig. 2f). Moreover, a decrease in αSyn levels in both the soluble and insoluble fractions of the midbrain was observed in Rg1-treated mice, confirming the aforementioned result (Fig. 2g). These findings suggest that Rg1 ameliorates stress-induced degeneration of dopaminergic neurons and the pathological aggregation of αSyn.

Fig. 2. Rg1 ameliorates dopaminergic neuron degeneration and αSyn accumulation.

a Representative images of TH immunostaining in the striatum and SNc. Scale bar: 100 μm. b Quantitative assessment of TH signal intensity, measured as IOD in the striatum, and number of TH⁺ neurons in the SNc (n = 5). c Representative blots of TH expression in the midbrain and quantitative analysis (n = 5). d Representative TEM images of neurons (i, scale bar: 2 μm), mitochondria (ii, scale bar: 200 nm), and myelin (iii, scale bar: 1 μm) in the SNc, which are indicated by the red arrows. e ThS staining was used to quantify the number of aggregates in the SNc, including ThS density and the proportion of cells with ThS⁺ puncta (n = 5). Scale bar: 20 μm. f Representative blots of αSyn aggregation using the 5G4 antibody, with quantitative analysis (n = 5). g Representative blots of αSyn in both soluble and insoluble fractions and quantitative data (n = 5). Data are presented as mean ± SEM, ***P < 0.001 vs. CSGS.

Rg1 reversed the PD-like pathological alterations induced by the stress-sensitive protein RTP801

To elucidate the mechanism by which Rg1 prevents PD progression under CRS, our initial focus was on the HPA axis, a critical mediator of the stress response in rodents that is often associated with alterations in body weight. Over a 4-week period, we monitored body weight changes and observed a decrease in the Rg1-treated animals. Nevertheless, this reduction was less marked than that in the CSGS group (Fig. 3a). To ascertain whether the effect of Rg1 was dependent on stress response attenuation, we evaluated serum CORT levels at Days 1, 7, 14, and 28 in the CRS protocol. A decrease in the CORT concentration was observed only in the initial week after Rg1 treatment (Fig. 3b). These findings imply that the ameliorative effect of Rg1 on CSGS-induced PD-like symptoms is not exclusively due to reduced CORT levels, suggesting the involvement of a CORT-independent mechanism underlying the therapeutic effects of Rg1.

Fig. 3. Rg1 reverses RTP801 upregulation in CSGS model.

a Relative change in body weight over a 4-week period (n = 7 for Control, n = 8 for CSGS, n = 7 for CSGS + Rg1). b Serum CORT levels were measured on Days 1, 7, 14, and 28 under CRS (n = 4). c Western blotting analysis of the stress-responsive proteins ERRFI1, FKBP5, FOXO1, KLF15, and RTP801 (n = 5). d Colocalization of RTP801 with TH, and quantification of RTP801⁺ density with a Pearson correlation analysis between RTP801 and TH (n = 5). Scale bar: 50 μm. e Temporal profile of RTP801 protein levels in the midbrain before and after CRS at designated time points (n = 4). f RTP801 protein levels in WT mice and αSyn A53T mice treated with or without CRS or Rg1 on Day 1 or Day 28 (n = 4). g Schematic of the RTP801 OE mouse experiment. h–k Analysis of behavioral performance in the rotarod test, pole test, beam walking test, and grip strength test (n = 6). l Colocalization of RTP801 (mCherry) with TH. Scale bar: 50 μm. IHC staining of RTP801. Scale bar: 200 μm. m Western blotting analysis of RTP801, TH, and αSyn aggregates (5G4) (n = 4). n Western blotting analysis of αSyn in soluble and insoluble fractions (n = 4). Data are presented as mean ± SEM, ^###P < 0.001 vs. Control/NC, *P < 0.05, **P < 0.01, ***P < 0.001 vs. CSGS/RTP801 OE, ^&&&P < 0.001 vs. CRS, ns not significant.

Subsequent investigations focused on assessing the modulation of stress-sensitive proteins, including ERRFI1, FKBP5, FOXO1, KLF15, and RTP801 [47–49]. Notably, Rg1 treatment significantly downregulated FKBP5 and RTP801 levels in the midbrain of the mice compared with those in the CSGS group, with the changes in RTP801 being especially notable (Fig. 3c). The temporal dynamics of the impact of Rg1 on RTP801 levels were examined at various points during stress exposure, specifically prior to the onset of CRS (d 0) and on Days 1, 7, 14, and 28 of CRS. In the initial acute phase (d 1), no significant difference in RTP801 levels was noted between the Rg1-treated and control groups. However, during the prolonged chronic phase (d 7, 14, and 28), RTP801 expression was markedly reduced in the presence of Rg1 (Fig. 3e). These changes suggest that Rg1 selectively modulates RTP801 levels under sustained stress conditions without altering the acute stress response. Additionally, RTP801 levels in αSyn A53T transgenic mice surpassed those in WT mice during both the acute (d 1) and chronic (d 28) phases (Fig. 3f), highlighting the critical role of RTP801 in the interaction between genetic factors and stress, which exacerbates PD progression. Elevated RTP801 levels in the SNc were associated with impaired morphology of TH⁺ neurons. Rg1 treatment significantly decreased RTP801 expression and increased the number of TH⁺ neurons. A negative correlation was found between RTP801 and TH levels (r = −0.9083, P = 0.0329) (Fig. 3d).

To mimic the pathological conditions of increased RTP801, adeno-associated virus-mediated RTP801 overexpression (AAV-RTP801-mCherry) was bilaterally injected into the SNc of 10-month-old WT mice (Figs. 3g and S5). Remarkably, RTP801 overexpression (OE) mice exhibited PD-like motor dysfunction, which was effectively mitigated by Rg1 (Fig. 3h–k). RTP801 OE disrupted the morphological integrity of TH⁺ neurons, reduced TH expression and increased αSyn accumulation. Rg1 treatment in these mice downregulated RTP801 in dopaminergic neurons, increased TH expression, and decreased αSyn accumulation (Fig. 3l–n). RTP801 OE in the SNc was sufficient to mimic PD-like motor dysfunction, neural degeneration, and pathological αSyn aggregation in CSGS model mice, and these effects were mitigated following Rg1 treatment, underscoring that RTP801 may be involved in the anti-PD effect of Rg1.

Rg1 promotes RTP801 degradation through ubiquitination at K188 and K218

To elucidate the underlying mechanism by which Rg1 reduces RTP801 protein levels, we established an in vitro model by simulating CORT-induced RTP801 elevation in SY5Y-αSyn A53T cells (Fig. 4a). Using cycloheximide (CHX), a protein synthesis inhibitor, we observed that the application of CORT led to delayed RTP801 degradation. The half-life of RTP801 was shortened by Rg1 treatment, suggesting that Rg1 might enhance RTP801 degradation (Fig. 4b). The proteasome inhibitor MG132 reversed the downregulatory effect of Rg1 on RTP801 (Fig. 4c), confirming that the Rg1-mediated RTP801 downregulation depends on the proteasome pathway.

Fig. 4. Rg1 promotes RTP801 degradation via the E3 ligase Parkin.

a Western blotting analysis of RTP801 expression in SH-SY5Y-αSyn A53T cells subjected to CORT and Rg1 treatment (n = 4). b Western blotting analysis of RTP801 in SH-SY5Y-αSyn A53T cells treated with CHX for various durations (0‒120 min) (n = 4). c RTP801 levels with or without MG132 (n = 4). d Ubiquitinated sites of RTP801 assessed using immunoprecipitation and HPLC‒MS/MS. MS fragmentation spectrum of RTP801 peptides displaying GG modifications at lysine residues K188 and K218 of RTP801 in the Rg1-treated group. e Co-IP of WT, 2KR, and K155R RTP801 groups for RTP801 and ubiquitin to confirm ubiquitinated RTP801 (n = 4). f Western blotting analysis of RTP801 expression following CHX treatment (n = 4). g Western blotting analysis of Parkin and NEDD4 in SH-SY5Y-αSyn A53T cells (n = 4). h Co-IP of WT and ΔR2 Parkin groups for RTP801 and ubiquitin to confirm ubiquitinated RTP801 (n = 4). Data are presented as mean ± SEM, ^##P < 0.01, ^###P < 0.001 vs. Control/CORT, *P < 0.05, **P < 0.01, ***P < 0.001 vs. CORT/Vehicle, ns not significant.

To identify the specific ubiquitinated lysine sites of RTP801 induced by Rg1, we immunoprecipitated RTP801 from the lysates of CORT-treated αSyn A53T-expressing SH-SY5Y cells and performed mass spectrometry. Rg1 treatment resulted in the ubiquitination of K188 and K218, which were not observed in the CORT group (Fig. 4d). Site-directed mutagenesis of these lysine residues to arginine (2KR) abolished the effect of Rg1 on RTP801 ubiquitination. We also mutated K155 (K155R), which was not ubiquitinated in either group, serving as a negative control; the K155R mutation did not reverse the effect of Rg1 (Fig. 4e). The effect of the 2KR mutation was further confirmed using CHX (Fig. 4f). RTP801 has been reported to be polyubiquitinated by the E3 ligases Parkin and NEDD4. The changes in Parkin expression, but not NEDD4 expression, were consistent with those observed in the PD-sensitive model in the presence and absence of Rg1 (Fig. 4g). Furthermore, the introduction of an inactive Parkin variant (Parkin ΔR2) negated the effect of Rg1 on RTP801 ubiquitination (Fig. 4h). These findings indicate that Rg1 promotes RTP801 degradation through ubiquitination at K188 and K218, and this process is mediated by the RING2 domain of Parkin.

RTP801 is critical for the protective effect of Rg1 under stress

To determine the role of RTP801 in PD progression induced by stress, we generated RTP801-knockdown (KD) SH-SY5Y-αSyn A53T cells via LV-RTP801-shRNA (Fig. S6a). RTP801 KD eliminated CORT-induced αSyn accumulation; interestingly, Rg1 failed to further inhibit αSyn accumulation in RTP801-KD cells (Fig. 5a, b), suggesting that RTP801 plays a crucial role in the anti-PD effect of Rg1. We subsequently generated RTP801^−/− mice and crossed them with αSyn A53T mice to produce αSyn A53T⁺; RTP801^−/− mice. These mice were subjected to CRS to explore the role of RTP801 in stress-induced PD exacerbation in vivo (Figs. 5c, d and S6b). Compared with αSyn A53T⁺; RTP801^+/+ (αSyn A53T) mice, αSyn A53T⁺; RTP801^−/− mice exhibited significantly enhanced motor performance in the rotarod, pole, and beam walking tests (Fig. 5e–g). However, the results of the forelimb grip strength test revealed no significant changes (Fig. 5h). We hypothesized that this may be due to the involvement of RTP801 in regulating peripheral muscle function. To test this hypothesis, we injected AAV-RTP801 shRNA or control virus into the SNc of αSyn A53T mice and compared their grip strength under CRS (Fig. S6c). Compared with control mice (NC CSGS), αSyn A53T mice with SNc-specific RTP801 KD presented significantly improved grip strength, confirming our hypothesis (Fig. S6d). Furthermore, we confirmed the detrimental effect of RTP801 on dopaminergic neurons, as the number of TH⁺ cells increased and αSyn accumulation decreased in the SNc of αSyn A53T⁺; RTP801^−/− mice. Notably, the protective effect of Rg1 was less apparent in RTP801^−/− mice within the CSGS model (Fig. 5i–k). These findings affirm that RTP801 is vital for PD exacerbation induced by stress and is indispensable for the disease-modifying effect of Rg1.

Rg1 mitigates αSyn aggregation caused by RTP801-impaired autophagic degradation

Previous studies have demonstrated that αSyn aggregates are degraded through the autophagy-lysosomal or ubiquitin-proteasome system [50], whereas stress inhibits aggregate degradation [11]. To determine whether αSyn serves as a cargo for autophagic degradation under stress, we observed the colocalization of αSyn with the autophagy markers LC3, SQSTM1 and ubiquitin in CORT-treated SH-SY5Y-αSyn A53T cells. Rg1 reduced αSyn colocalization with LC3 and SQSTM1, whereas αSyn colocalization with ubiquitin remained unchanged, confirming that Rg1-induced αSyn degradation predominantly occurs via the autophagy-lysosomal pathway rather than the ubiquitin-proteasome system (Fig. S7a).

To explore the role of RTP801 in this process, we assessed AMPK, Beclin 1, LC3, and SQSTM1 levels in RTP801 OE mice. Compared with those in the control group (NC), all these proteins were upregulated, indicating that RTP801 promotes autophagy induction but may impede autophagic flux. Rg1 did not significantly alter AMPK and Beclin 1 levels but markedly downregulated LC3-II and SQSTM1, indicating that Rg1 mitigates the impairment of autophagic degradation induced by RTP801 OE (Fig. 6a). In RTP801 OE PC12 cells, which were transfected with mCherry-EGFP-LC3 adenovirus to monitor autophagic flux, increased numbers of yellow puncta (autophagosomes) and red puncta (autolysosomes, due to quenching of EGFP fluorescence in acidic lysosomes) were quantified [51]. The presence of more yellow dots indicated autophagic flux dysfunction. Autophagy induction by rapamycin significantly increased the number of red dots, confirming enhanced autophagic flux. The addition of BafA1, an inhibitor of autolysosome formation, increased the number of yellow dots in each group, negating the beneficial effect of Rg1 (Fig. 6b–d). Additionally, diminished colocalization of LC3 and SQSTM1 with the lysosomal marker LAMP1 was observed in the SNc of RTP801 OE mice, suggesting that RTP801 inhibits the fusion of autophagosomes with lysosomes, a process that was ameliorated by Rg1 (Figs. 6e and S7b). Compared with the CSGS group, the Rg1-treated group presented increased colocalization of αSyn with lysosomes and fewer αSyn⁺ particles, supporting the role of Rg1 in promoting functional autolysosome formation and enhancing αSyn clearance (Fig. 6f). Using TEM, we noted a significant increase in the number of autophagosomes and morphological alterations in the RTP801 OE group. Conversely, the Rg1 group presented a reduced number of autophagosomes, which appeared more normal in size and morphology. The number of autolysosomes significantly increased, indicating that Rg1 ameliorated the impairment of autophagy induced by RTP801 OE and promoted the formation of functional autolysosomes (Fig. 6g). This effect was reversed by CQ, an inhibitor of autophagic flux, thus demonstrating the efficacy of Rg1 in alleviating RTP801-induced disruption of functional autolysosome formation in vivo (Fig. 6h).

Fig. 6. Rg1 mitigates RTP801-induced autophagic degradation impairment.

a Analysis of p-AMPK, AMPK, Beclin 1, LC3, and SQSTM1 protein levels in RTP801 OE mice (n = 4). b‒d Autophagic flux assessment in RTP801-expressing and vehicle-treated PC12 cells via mCherry-EGFP-LC3 adenovirus. Visualization of autophagosomes (yellow puncta) and autolysosomes (red puncta) with or without Rg1 treatment. Quantification of autophagosomes and autolysosomes is presented in stacked columns (n = 15‒20 cells from 3 independent experiments). Scale bar: 10 μm. e LAMP1 colocalization with LC3 and SQSTM1 and their overlap colocalization (n = 40 cells from 4 mice). Scale bar: 5 μm. f Representative images of colocalization of LAMP1 with αSyn in the SNc and colocalization analysis (n = 40 cells from 4 mice). Scale bar: 5 μm. Fluorescence intensity profiles along a 10 µm line across the cell body confirm colocalization (the subsequent intensity analysis was the same). g Representative TEM images of autophagosomes and autolysosomes induced by Rg1 in RTP801-OE PC12 cells. Autophagosomes (red arrows) and autolysosomes (yellow arrows) are indicated. Scale bar: upper, 2 μm; lower, 0.5 μm. h Western blotting analysis of LC3 and SQSTM1 in RTP801 OE mice with or without CQ treatment (n = 4). Data are presented as mean ± SEM, ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. NC/vector, *P < 0.05, **P < 0.01, ***P < 0.001 vs. RTP801 OE/CSGS, ns not significant.

Rg1 alleviates degradation impediment of αSyn by modulating RTP801-TXNIP-mediated ATP13A2 deficiency

To investigate the mechanism underlying the enhancement of functional autolysosome formation by Rg1, we quantified the expression of genes involved in lysosomal biogenesis, acidification, and autophagosome-lysosome fusion in the SNc. Notably, genes regulating autophagosome-lysosome fusion, particularly Atp13a2, were significantly upregulated following Rg1 treatment (Fig. 7a). Atp13a2 encodes a P-type ATPase family protein localized to the lysosomal membrane [52, 53]. Rg1 treatment enhanced ATP13A2 expression in the midbrain of CSGS mice, but this effect was not obvious in αSyn A53T⁺; RTP801^−/− mice (Fig. 7b). These findings suggest that RTP801 is involved in the Rg1-mediated regulation of ATP13A2. Atp13a2 knockdown in SH-SY5Y-αSyn A53T cells (Fig. S8a) negated the beneficial effects of Rg1 on αSyn aggregation (Fig. 7c). Moreover, even with Rg1 treatment, Atp13a2 KD significantly reduced the lysosomal localization of αSyn compared with that of the negative control, underscoring the critical role of ATP13A2 in Rg1-mediated αSyn degradation (Figs. 7d and S8b). To explore the interaction between ATP13A2 and RTP801, we employed STRING network analysis to delineate the relationships between RTP801- and ATP13A2-associated proteins. A Venn diagram revealed two proteins, TXNIP and thioredoxin-2 (TXN2), as potential interactors with both RTP801 and ATP13A2 (Fig. 7e). Colocalization studies of the SNc of RTP801 OE mice demonstrated significant colocalization of TXNIP with RTP801, which decreased following Rg1 treatment (Fig. 7f). In contrast, TXN2 did not significantly colocalize with RTP801 in the SNc (Fig. S8c). Co-IP assays further confirmed that Rg1 treatment reduces the RTP801-TXNIP interaction (Fig. 7g). In SH-SY5Y-αSyn A53T cells transfected with LV-TXNIP shRNA (TXNIP KD), Rg1 treatment failed to further increase ATP13A2 levels (Figs. S8d and 7h) or reduce αSyn accumulation (Fig. 7i). Our results demonstrate that Rg1 disrupts RTP801-TXNIP complex formation, thereby preserving ATP13A2 levels, which are essential for autolysosome functionality and enhanced αSyn autophagic degradation in psychological stress-induced PD models.

Discussion

The relationship between psychological stress and PD has attracted increasing scrutiny in recent years. A survey of 5000 PD patients indicated that they experienced elevated stress levels than non-afflicted individuals did, which aggravated both their motor and nonmotor symptoms [54]. The body of evidence has grown even more in the wake of the COVID-19 pandemic. The psychological stress caused by COVID-19 exacerbates PD-like motor symptoms and disease progression [55]. However, no effective treatment strategies available are currently available. In this study, we utilized a transgenic PD mouse model in which CRS was used to induce psychological stress. We found that the oral administration of Rg1 significantly ameliorated the PD conditions exacerbated by psychological stress. Notably, it reduced the accumulation of αSyn and the loss of dopaminergic neurons in the SNc. This pathological improvement led to enhanced functional connectivity of the nigrostriatal neural pathway and increased excitability of MSNs and FSIs in the striatum, thereby ameliorating PD-like motor symptoms. Additionally, Rg1 enhanced brain activity related to the olfactory and other regions, improving the typical nonmotor symptoms of PD, such as olfactory dysfunction. In summary, Rg1 has beneficial therapeutic effects against the exacerbation of PD caused by psychological stress.

The loss of dopaminergic neurons in PD animal models directly affects the firing rate and neuronal synchrony. MSNs are the main output neurons of the striatum and are regulated by exogenous afferents and local circuit inputs. MSN activation could rescue deficits in freezing, bradykinesia and locomotor initiation in the PD model. FSIs are critical for maintaining balanced firing between direct and indirect pathway neurons of the striatum, which are also crucial for the regulation of motor control [43, 44, 56]. This dyskinesia-related electrical activity demonstrated the improvement in the pro-motor effects of Rg1 in the CSGS model. We also recorded and analyzed the spike data in the striatum before and after Rg1 administration on the first day. The results revealed no significant changes in the rate histograms or burst analysis results following a single treatment (Fig. S3e, f), excluding the possibility that the observed alterations in electrical activity (Figs. 1n–q and S3a–d) were due to direct stimulation caused by Rg1. Instead, the improvement in electrical activity under pathological conditions after 4 weeks of Rg1 treatment is attributable to an overall enhancement in striatal function.

Currently, clinical treatments for PD primarily focus on symptomatic relief and fail to effectively slow disease progression. Aggregated αSyn is a fundamental component of Lewy bodies and plays a critical role in the pathogenesis of PD. Until 2022, anti-αSyn antibodies were extensively investigated as potential disease-modifying therapies for PD. However, their clinical efficacy is comparable to that of placebo, and their administration is often associated with adverse events such as headache, nasopharyngitis, and falls [3, 4]. The safety of Rg1, a primary component of ginseng, a traditional Chinese medicine that has been used for centuries to promote longevity, has been extensively validated over time. Toxicological assessments have confirmed its safety profile, with no observed adverse effects or potential toxicity [19, 57]. Our current study demonstrated that Rg1 facilitates the cellular clearance of accumulated αSyn in the SNc, thereby impeding further pathological advancement and improving both the motor and nonmotor symptoms of PD. Our research offers significant preventive and therapeutic benefits for individuals susceptible to PD, especially those experiencing prolonged stress and those with a familial history, effectively slowing disease progression. In our CSGS model, we only observed a significant presence of Lewy body pathology in the substantia nigra, which was not particularly evident in other regions. Therefore, whether Rg1 has a beneficial effect on all synucleinopathies still needs further validation.

Stress is a physiological regulatory phenomenon that organisms exhibit in response to changes in their external environment. This response is characterized not only by hyperactivation of the HPA axis, but also by the rapid production of stress-sensitive proteins. In our CSGS model, serum CORT levels remained persistently elevated. However, treatment with Rg1 effectively suppressed the CORT levels observed only during the first week. Our analysis after 28 days revealed that the behavioral and pathological effects of Rg1 were HPA axis independent (Fig. 3b), suggesting that alternative mechanisms may be involved in the effects of Rg1. Consequently, we screened for stress-sensitive proteins and detected a significant reduction in RTP801 and FKBP5 protein levels following Rg1 treatment. RTP801, a protein that exhibits significantly elevated expression in neuromelanin-containing neurons of the SNc in PD patients [14], was selected for in-depth investigation of the therapeutic effects of Rg1. Our previous research demonstrated that abnormally enhanced RTP801 leads to a blockade of αSyn clearance [11]. In the present study, Rg1 treatment significantly reduced RTP801 levels in SNc dopaminergic neurons and increased TH levels, indicating a significant negative correlation. These findings suggest that the reduction in RTP801 might underlie the mechanism through which Rg1 prevents the loss of dopaminergic neurons. To simulate the PD pathological conditions induced by elevated RTP801 levels, we injected a virus overexpressing RTP801 into the substantia nigra. We observed that increased RTP801 in the substantia nigra led to significant loss of dopaminergic neurons, PD-like behavioral impairments, and αSyn aggregation. Rg1 treatment reduced RTP801 levels and ameliorated these pathological conditions, indicating that RTP801 might be a critical component in the pharmacological effect of Rg1.

As RTP801 forcibly expressed through lentivirus transfection, the mechanism by which Rg1 reduces RTP801 protein levels remains unclear. To address this question, we initially employed a protein synthesis inhibitor to block protein production and found that CORT significantly prolonged the degradation half-life of RTP801, an effect that was significantly reversed by Rg1 treatment. Using immunoprecipitation followed by HPLC‒MS/MS, we discovered that Rg1 enhances the ubiquitination of RTP801 at lysine residues K188 and K218, a modification not observed in the CORT group. Additionally, Rg1 enhanced the interaction of RTP801 with Parkin and reversed the accumulation of RTP801 caused by stress. Mutations at both K188 and K218 blocked the ability of Rg1 to promote RTP801 degradation, and a mutation in the RING2 domain of Parkin yielded the same result.

To elucidate the role of RTP801 in the above effect of Rg1, we constructed dual-transgenic mice (αSyn A53T⁺; RTP801^−/−). We found that the absence of RTP801 reversed the progression of PD exacerbated by psychological stress and abolished the effect of Rg1, which confirms that reducing RTP801 levels is critical for the protective effect of Rg1 under stress conditions. This finding led us to investigate how Rg1 reduces αSyn aggregation through RTP801. Our findings indicate that Rg1 treatment significantly promotes the colocalization of αSyn with lysosomes rather than with ubiquitin, suggesting that the autophagy-lysosome pathway, but not the ubiquitin pathway, is involved in the reduction in αSyn aggregation in Rg1 (Fig. S7a). The rapid increase in the levels of the stress-sensitive protein RTP801 increases intracellular autophagy levels, facilitating the recycling of cellular metabolites. However, prolonged RTP801 accumulation leads to autophagy impairment in αSyn-expressing neurons. By employing an mCherry-EGFP-LC3 adenovirus for autophagy flux assays, we found that RTP801 OE impaired fusion between autophagosomes and lysosomes. In RTP801 OE mice, a significant reduction in the colocalization of the autophagosome marker protein LC3-II and the substrate protein SQSTM1 with lysosomes and an accumulation of autophagosomes within neurons were observed, fully illustrating the impairment of fusion between autophagosomes and lysosomes. Rg1 treatment promotes the colocalization of autophagy marker proteins with lysosomes, reducing the accumulation of autophagosomes in neurons. Interestingly, LC3-II levels in the Rg1-treated group were significantly lower than that in the model group possibly because Rg1 reduces the RTP801 level, thus effectively reducing abnormal autophagy initiation and promoting autophagic flux to increase LC3-II protein degradation, resulting in less accumulation. To further understand how Rg1 ameliorates the impairment of autophagy-lysosome fusion, we analyzed genes related to lysosomal biogenesis and acidification in the SNc and found that ATP13A2 was significantly upregulated by Rg1 treatment. This gene encodes a protein involved in autophagy-lysosome fusion. Atp13a2 knockdown abolished the reduction in αSyn autophagic aggregation induced by Rg1. STRING network analysis suggested that TXNIP may be the optimal linker between RTP801 and ATP13A2, and our results demonstrated that TXNIP could be immunoprecipitated by an RTP801 antibody. TXNIP knockdown abolished the Rg1-induced improvements in ATP13A2 and degradation of αSyn aggregates. The increase in the RTP801-TXNIP-ATP13A2 axis induced by stress was mitigated by Rg1 to rescue autophagy-lysosome fusion.

In summary, this study revealed that chronic psychological stress leads to a reduction in the degradation rate of the stress-sensitive protein RTP801, resulting in its abnormal accumulation. Elevated levels of RTP801 interact with TXNIP, thus inhibiting ATP13A2 levels and impairing autophagosome-lysosome fusion. This impairment leads to a degradation deficiency in αSyn, exacerbating αSyn aggregation and PD progression. Rg1 promotes the connection between RTP801 and Parkin, thereby enhancing RTP801 ubiquitination at residues K188 and K218, thus rescuing the impairment of RTP801 degradation. This research not only highlights Rg1 as a potentially promising therapeutic agent for delaying PD but also highlights its ability to target RTP801 as a viable strategy in the development of anti-PD drugs (Fig. 8).

Fig. 8. Ginsenoside Rg1 alleviates stress-exacerbated Parkinson’s disease by promoting RTP801 degradation.

Ginsenoside Rg1 mitigates stress-exacerbated PD by improving symptoms, brain connectivity, neuronal activity and reducing αSyn pathology. Mechanistically, chronic psychological stress inhibits RTP801 degradation in the midbrain, leading to RTP801 accumulation and interaction with TXNIP. This interaction decreases ATP13A2 levels, disrupting autophagosome-lysosome fusion, which is essential for αSyn clearance. Rg1 administration enhances RTP801 ubiquitination at K188 and K218 via Parkin, promoting its degradation and restoring autophagic function, thereby reducing αSyn aggregation and slowing PD progression.

Author contributions

NHC, ZZ, SFC, and SSW conceived and designed the study. SSW, YP, PLF, JRY, WYM and XY performed the experiments. JRY, QLW, HYW and YJT participated in data analysis. SSW and ZZ wrote the manuscript. NHC, SFC, ZZ, YP and WBH revised the manuscript. All the authors have read and approved the final version of the manuscript.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41401-024-01374-w.