Targeting microglial inflammation in Parkinson’s disease: irisin activates PAFAH1B1-RAGE ubiquitination and TFEB-dependent autophagy to alleviate neurodegeneration

Abstract

Investigate irisin’s therapeutic potential in Parkinson’s disease (PD). Clinical data from 120 PD patients and 120 controls were analyzed. MPTP-induced PD mice and LPS-stimulated BV2 microglia models were used. In vivo, mice were divided into control, PD, and PD + Irisin groups for behavioral and histological assessments. In vitro, LPS-stimulated BV2 cells were treated with irisin or PBS. RNA sequencing, immunohistochemistry, and Western blot evaluated autophagy, inflammation, and ubiquitination pathways. PD patients exhibited increased TNF-α and IL-1β but decreased irisin levels. In PD mouse models, irisin improved motor deficits, increased nigrostriatal neuron numbers, restored tyrosine hydroxylase expression, and reduced α-synuclein aggregation. It also suppressed microglial inflammation and promoted anti-inflammatory polarization. Mechanistically, irisin enhanced autophagic flux, regulated RAGE ubiquitination mediated by PAFAH1B1, and inhibited neuroinflammation via the TFEB-NLRP3 axis. Specifically, PAFAH1B1 regulated RAGE expression through K61 and K169 sites on K48-linked polyubiquitin chains. Additionally, irisin restored lysosomal function by promoting TFEB nuclear translocation, enhancing NLRP3 inflammasome degradation, and reducing inflammatory factor secretion, thus alleviating neuroinflammation. Irisin alleviates PD pathology by modulating autophagy and ubiquitination pathways, suggesting its potential as a novel immunomodulatory target for PD.

Irisin attenuates microglial inflammation in PD via ubiquitination-autophagy pathways: PAFAH1B1-RAGE ubiquitination and TFEB-NLRP3 modulation.

Introduction

Parkinson’s disease (PD) is a progressive neurodegenerative disorder distinguished by hallmark motor symptoms, including resting tremor, rigidity, and bradykinesia, alongside a range of non-motor manifestations, such as hyposmia, constipation, sleep disturbances, autonomic dysfunctions, and mood disorders like anxiety and depression¹. Currently, PD ranks as the second most prevalent neurodegenerative disorder, following Alzheimer’s disease (AD). According to a global burden of disease study, the prevalence of PD among individuals aged 60–69 is estimated to be between 0.5% and 1%, while in those aged 80 and older, it ranges from 1% to 3%. Given the rapid aging of the population, projections indicate that by 2050, ~23.9–26.9% of the global population will be over the age of 65, leading to a continued increase in PD cases². Consequently, a comprehensive investigation into the molecular mechanisms underlying PD, along with the identification of biomarkers for early diagnosis and the development of effective preventive and therapeutic agents, is essential for enhancing the quality of life in elderly populations.

Recent studies have gradually revealed the important role of immune activation and neuroinflammatory responses in the pathogenesis of PD³. Numerous studies have documented significant inflammatory responses in the brain tissue of patients with PD, particularly highlighting the activation of microglia and astrocytes⁴. Our preliminary investigations into the pathogenesis of PD have also revealed that immune-related molecules play a crucial role in both the prediction and progression of its underlying mechanisms⁵. Therefore, immune activation is considered one of the core mechanisms of PD, and the immune response it triggers, including immune cell infiltration, inflammatory factor release, and the cascade amplification of immune responses, may ultimately lead to the death of dopamine neurons (DA)⁶. Currently, some therapeutic strategies targeting inflammation and immune regulation have entered clinical application. For example, some nonsteroidal anti-inflammatory drugs and resveratrol have shown potential in clinical trials to partially alleviate PD symptoms⁷; there are also exploratory studies using α-synuclein vaccines to alleviate PD symptoms⁸, among others. While these approaches provide new hope for the treatment of PD, they are still subject to certain limitations; specifically, nonsteroidal anti-inflammatory drugs may alleviate symptoms and mitigate systemic inflammation, yet they do not directly address neuronal damage and death within the central nervous system⁸. Consequently, an in-depth investigation of the pathogenesis of PD, aimed at elucidating critical mechanisms involved in its development, as well as the exploration and development of more specific, effective, and less adverse treatments for PD, remains a pivotal focus in research on PD prevention and management^9,10.

Irisin is a myokine released into the bloodstream by skeletal muscle. Current research indicates that irisin exerts anti-inflammatory and anti-apoptotic effects on cardiomyocytes and hepatocytes, thereby protecting these cells from ischemia-reperfusion injury¹¹. A recent study indicates that physical exercise can promote the expression of brain-derived neurotrophic factor (BDNF) and glial cell line-derived neurotrophic factor (GDNF), neurotransmitter DA, and irisin in normal individuals, by downregulating inflammatory cytokines tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6) in the hippocampus, thereby reducing stress and anxiety and improving insulin resistance¹². Another AD research suggests that irisin may have neuroprotective effects by improving brain BDNF levels through its anti-inflammatory and antioxidant actions¹³. Nevertheless, there is a paucity of research regarding the relationship between irisin and PD, particularly concerning the role and mechanisms by which irisin regulates neuroinflammation and immune responses within the PD-affected nervous system. Consequently, this study seeks to investigate the molecular mechanisms through which irisin ameliorates PD using both in vivo and in vitro models, complemented by sequencing and molecular biology techniques, thereby providing a foundation for the development of therapeutic strategies for PD.

Materials and methods

Clinical sample

In this study, we collected data from a total of 120 PD patients treated at the Affiliated Hospital of Guizhou Medical University between August 2022 and September 2023. The diagnosis of PD was confirmed for all patients by two senior neurologists in the Diagnostic Criteria for Parkinson’s Disease in China (2016 Edition). Additionally, 120 healthy individuals who underwent physical examinations at the Physical Examination Center of the same hospital were selected as a control group. Sample size calculations were conducted following Chow S’s clinical sample size estimation method, utilizing α = 0.05, 1−β = 0.80, and kappa = 1; to account for potential follow-up losses, an increase of 20% was applied, resulting in 120 cases per group. Statistical analysis revealed no significant differences between the two groups regarding age, gender, and other general characteristics, thus indicating good comparability between them. All relevant clinical sample information is presented in Supplementary Table 1. This research protocol received approval from the Ethics Committee of the Affiliated Hospital of Guizhou Medical University (No. 2022089K).

Animal and PD model

SPF-grade male C57BL/6 mice, aged 8–10 weeks and weighing 20–25 g, were obtained from the Laboratory Animal Research Center of Guizhou Medical University. Mice were housed under controlled conditions: temperature 23 ± 2 °C, humidity 40–50%, and a 12-h light/dark cycle. Using a random number table, mice were divided into two groups: 10 in the control group and 20 in the model group. After successful modeling, 20 PD mice were further divided into two groups (n = 10 each): the PD group and the PD + Irisin group.

Treatment protocols were as follows: 1. Control group: Mice received intraperitoneal injections of 0.9% normal saline (10 mL/kg/day) for 7 days (days 1–7), followed by tail vein injections of 0.9% normal saline (5 mL/kg/day) for another 7 days (days 9–15). 2. PD group: Mice received intraperitoneal injections of MPTP (20 mg/kg/day) for 7 days (days 1–7), followed by tail vein injections of 0.9% normal saline (5 mL/kg/day) for 7 days (days 9–15). 3. PD + Irisin group: Mice received intraperitoneal injections of MPTP (20 mg/kg/day) for 7 days (days 1–7), followed by tail vein injections of irisin (60 μg/kg/day) for 7 days (days 9–15). Behavioral evaluations were conducted on days 8 and 16. Mice were weighed daily during the treatment period, and data were recorded. On day 8, the success of the model was evaluated using swimming and pole climbing tests. On day 16, after behavioral evaluation, mice were euthanized with excessive anesthetics. Blood was collected via cardiac puncture, serum was separated, and stored at −80 °C. Brain tissues were either fixed in paraformaldehyde or frozen at −80 °C for subsequent experiments. The experimental design is illustrated in Fig. 2A. All animal procedures were approved by the Experimental Animal Ethics Committee of Guizhou Medical University (No. 2200545).

Fig. 2. Improvement of PD-like symptoms and brain region distribution of irisin in mice model.

A Schematic illustration of the construction of a PD model using MPTP. Results of swimming score (B) and pole-climbing test (C and D) in the control and PD groups. Control vs. PD, *p < 0.05. E Dynamic monitoring of body weight in the control, PD, and PD + Irisin groups. F–H Effects of irisin treatment on PD-like behaviors. Irisin treatment significantly improves PD-like behaviors, including increased swimming scores and decreased head-down time and pole-climbing time. Control vs. PD, *p < 0.05; PD + Irisin vs PD, *p < 0.05. I: Levels of irisin in the tissue homogenates of the cortex, cerebellum, and substantia nigra striatum of each group of mice detected by ELISA. Cortex and cerebellum: Control vs. PD, not significant; PD + Irisin vs. PD, not significant. Substantia nigra striatum: Control vs. PD, *p < 0.05; PD  Irisin vs. PD, *p < 0.05. J Detection of irisin level in peripheral blood in the PD group. The irisin level returns to normal after intervention. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05.

Cell culture

The mouse microglia cell line BV2 (CL-0493, Procell, Wuhan, China) was cultured and passaged in DMEM (KeGene Bio, China) medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. In this study, five groups were established: (1) the PBS control group; (2) the LPS group, treated with a complete medium containing 100 ng/mL LPS for 12 h; (3) the LPS + Irisin group, initially treated with LPS-containing medium for 12 h followed by treatment with complete medium containing 100 ng/mL irisin for an additional 12 h; (4) the LPS + Irisin + BafA1 group, which received 100 nM BafA1 for 6 h after the aforementioned treatments; and (5) the LPS + Rap group, treated with 40 nM Rapamycin for 6 h following LPS exposure.

RNA sequencing (RNA-seq) and differentially expressed genes (DEGs) screening

Liquid nitrogen snap-freezing method: The brain samples from each group of mice were rinsed with 0.9% normal saline, blotted dry, and processed on an ice pack. Specifically, 50 mg of midbrain substantia nigra tissue was rapidly frozen in liquid nitrogen, transferred to a 1.5 mL EP tube, and transported under dry ice conditions. Subsequently, the samples were sent to BGI for next-generation sequencing.

The raw counts were loaded into R software (http://www.R-project.org/) and analyzed using the edgeR package. For normalization and differential expression analysis, the screening threshold was set to p-value < 0.05 and fold change > 1. The related packages were then used to generate the differential expression heat map and volcano plot.

Cell transfection

After dissolving 1 OD of RNA oligo in 125 μL DEPC water, the transfection reagent was brought to room temperature and gently mixed before use. For transfection preparation, 200 μL of serum-free medium (for a six-well plate) and 5 μL of Hieff TransTM Liposomal Transfection Reagent were added to a sterile 1.5 mL tube, gently mixed using a pipettor, and incubated for 5 min. Simultaneously, 200 μL of serum-free medium and 10 μL of RNA oligo were added to another sterile 1.5 mL tube, gently mixed with a pipettor, and incubated for 5 min. The transfection reagent mixture was then slowly added dropwise to the RNA oligo solution and gently mixed using a pipette. During transfection, the medium in the six-well plate was replaced with 1600 μL of pre-warmed fresh medium per well. A total of 400 μL of the transfection mixture was added to each well, resulting in a final volume of 2000 μL per well. The plate was gently shaken to ensure even distribution, and cells were cultured under standard conditions. Transfection sequences are provided in Supplementary Table 2.

Behavioral analysis

In this study, a circular water basin with a diameter of 40 cm and a water depth of 10 cm was selected, with the water temperature maintained at 23 ± 2 °C. Three days before the behavioral assessment, each mouse began daily adaptive swimming training. The swimming scoring criteria were as follows: mice that could swim continuously for 1 min were given 3.0 points; those that swam for most of the time but occasionally floated were given 2.5 points; those that floated for more than 50% of the test time were given 2.0 points; those that swam occasionally were given 1.5 points; and those that occasionally used their hind limbs to swim and floated on one side of the water tank were given 1.0 point. Each test was conducted with a 1-min interval, and the average score of five tests was taken.

We used a custom-made wooden pole (0.8 cm diameter, 60 cm height) with a gauze-covered ball at the top to prevent slipping. The experiment had two stages: adaptation and testing. During adaptation, mice were placed head-down on the ball and allowed to climb down naturally for three days before behavioral evaluation. In the testing stage, mice were placed head-down on the pole, and timing started when they began climbing freely. If a mouse turned back, it was retested after a 10-min rest. After each trial, 75% alcohol was sprayed on the pole to eliminate scent interference. Two times were recorded: (1) from placing the mouse head-up to complete head-down orientation and (2) from head-down orientation to hind limbs reaching the base. Each test was repeated five times with 1-min intervals, and the average time was calculated.

mRFP-GFP-LC3 assay

First, prepare a cell suspension at a density of 3 × 10⁴ cells/mL and add 2 mL per well to a six-well plate. Incubate the plate in a CO₂ incubator for 24 h. For virus infection, thaw the virus on ice after retrieving it from −80 °C storage. Prepare a viral solution by mixing 30 μL of virus with 270 μL of DMEM medium to achieve a titer of 1 × 10⁹ PFU/mL. Remove 1 mL of the existing medium from each well and add the virus according to an MOI of 100 (virus volume = MOI × cell number/virus titer). Incubate the plate at 37 °C for 6 h, then replace the medium with 2 mL of complete medium and continue incubation for another 36 h. Once GFP and RFP expression is observed, wash the wells with PBS and fix the cells with 4% paraformaldehyde at room temperature for 30 min. Wash three times with PBS (5 min each), permeabilize with 0.5% Triton X-100 in PBS for 20 min, and wash three times again with PBS (5 min each). Stain nuclei with DAPI at room temperature in the dark for 5 min, followed by three more PBS washes (5 min each). Finally, observe and photograph the samples using a laser confocal microscope, and quantify the results using ImageJ software.

Transmission electron microscopy (TEM)

First, prepare the samples. Digest the cells with trypsin, then centrifuge at 1000 rpm/min for 4 min and discard the supernatant. Wash twice with pre-cooled PBS. Add 2.5% glutaraldehyde to the centrifuge tube and fix it for 24 h. After fixation, wash three times with PBS for 5 minutes each. Then, replace with 2% osmium acid and fix for 2 h. Wash three times with PBS for 5 min each. Subsequently, replace with 50%, 70%, 80%, and 90% ethanol gradients for dehydration for 20 min each, and then dehydrate three times with absolute ethanol for 20 min each. Soak in epoxy resin overnight, then place the samples in embedding plates containing embedding agents and polymerize at 65 °C for 48 h. Use a slicer to cut ultrathin sections of 60 nm in thickness and fix them on the grid. Finally, stain with 3% uranyl acetate-lead citrate.

Hematoxylin and eosin (HE) staining

The brain tissue paraffin sections were incubated at 60 °C for 3 h. Subsequently, the sections underwent hematoxylin staining for 2 min and were then thoroughly rinsed. Eosin staining was carried out at ambient temperature for 1 min. After staining, the sections were subjected to dehydration using a series of graded alcohols, followed by clearing with xylene. They were then mounted using a neutral mounting medium. Lastly, the prepared samples were examined under a microscope and imaged.

Nissl staining

Paraffin sections were dewaxed in water using xylene and a graded alcohol series, followed by a brief soak in distilled water. The sections were then stained with Nissl solution for 10–20 min at room temperature. After staining, the excess stain was quickly rinsed off with distilled water. Differentiation was performed by placing the sections in a solution of 70% ethanol and glacial acetic acid, gently agitating for several seconds to one minute to enhance contrast between nuclei and cytoplasm. Following differentiation, the sections were thoroughly rinsed with distilled water and mounted using a neutral medium. Finally, the morphological structure of the brain tissue was observed under a microscope.

Immunohistochemical (IHC) staining

The immunohistochemical staining was conducted following the instructions provided with the immunohistochemistry kit. To assess the stainings, quantitative imaging techniques were employed to record both the percentage and intensity of immunostaining.

Flow cytometry

To explore the expression of cell surface markers after stimulation, flow cytometry was employed in this study. Microglia were collected after 24 h of stimulation in six-well plates and washed three times with PBS to remove impurities. Cell suspensions containing 1 × 10⁶ M0-polarized or M1-polarized cells were transferred to 1.5 mL EP tubes and incubated with blocking antibody CD16/32 (Thermo Fisher, 14-0161-82) on ice for 10 minutes to block nonspecific binding sites and reduce background fluorescence. After incubation, cells were washed twice with PBS to remove unbound blocking antibody. Subsequently, cells were mixed with PBS containing Intrapore permeabilization reagent and PE-anti-mouse CD86 (Biolegend, 105007) and APC-anti-mouse CD206 antibodies (Biolegend, 141707) and incubated in the dark on ice for 30 min to allow antibodies to bind to the corresponding antigens on the cell surface. Following incubation, cells were washed twice with PBS to remove unbound antibodies. Finally, cells were resuspended in 500 μL PBS containing 3% fetal bovine serum and analyzed using a flow cytometer to obtain data on the expression levels of cell surface markers.

Enzyme-linked immunosorbent assay (ELISA)

The brain tissues of each group of mice were separated into the cerebral cortex, striatum, midbrain, and cerebellar regions on ice. 50–100 mg of brain tissue from each group was added to 1000 μL of pre-cooled RIPA protein lysis buffer, with steel beads added and placed in a tissue homogenizer (60 Hz for 60 s) for homogenization. The supernatant was then subjected to ultrasonic disruption and lysis. After centrifugation at 4 °C and 12,000 rpm for 25 min, the supernatant was separated into EP tubes. Protein quantification was performed, followed by ELISA detection. The detection indicators and related reagent kits are irisin (No. E-EL-H0149c, Elabscience), TNF-α (No. EMC102a, NeoBioscience), TNF-α (No. EMC109a, NeoBioscience), IL-1β (No. EMC001b, NeoBioscience), IL-4 (No. EMC003, NeoBioscience), and IL-10 (No. EMC005, NeoBioscience), respectively.

Quantitative real-time-PCR (QRT-PCR)

TRIzol was employed to extract total RNA. The PrimeScript TM RT Reagent Kit with gDNA Eraser (RR047B, Takara, Shandong, China) was utilized for cDNA synthesis to eliminate genomic DNA. Quantification of RNA levels was performed using the SYBR Green Pro Taq HS qPCR Mix (AG11701-S, Accurate Biology). Supplementary Table 3 was used to present the gene primer sequences.

Co-immunoprecipitation

For the co-immunoprecipitation experiment, 500 µg of protein samples were mixed with agarose beads and antibodies, and the volume was adjusted with RIPA buffer. The mixture was rotated at 4 °C overnight. Afterward, the mixture was centrifuged at 4 °C to separate the supernatant and the precipitate, and the supernatant was discarded. The precipitate was washed with RIPA buffer and centrifuged again. Subsequently, protein loading buffer (P0015L, Beyotime) was added, and the mixture was heated and centrifuged rapidly. Finally, the supernatant was collected for subsequent protein analysis.

Cell and tissue protein extraction

For cell protein extraction, cells were observed to calculate the required volume of lysis buffer. After discarding the culture medium, the lysis buffer was added and kept at a low temperature. Cells were scraped into centrifuge tubes with a cell scraper and kept on ice for 30 min, with shaking for 30 s every 5 min. After centrifugation, the supernatant was collected, and the precipitate was discarded. The protein concentration was measured with a BCA protein assay kit (P0010S, Beyotime), and the sample concentration was adjusted with RIPA and protein loading buffer for subsequent experiments.

For tissue protein extraction (a mouse brain), fresh tissues were washed with PBS to remove impurities and placed in centrifuge tubes. Homogenizing solution was added, and the tissues were homogenized with an electric homogenizer on ice for 10 times, 5 s each. The homogenizer probes were washed with PBS and dried with filter paper between different samples. After homogenization, the samples were kept on ice for 20 min, with shaking for 30 s every 5 min, and then centrifuged. The supernatant containing tissue cytoplasmic proteins was collected, and the precipitate was discarded. The protein concentration was measured with a BCA protein assay kit, and the samples were adjusted with RIPA and protein loading buffer for subsequent experiments.

Western blot (WB)

Cells were harvested and lysed with an SDS loading buffer. In vivo assays were conducted to isolate certain proteins from mice. The resulting proteins were separated by SDS–PAGE electrophoresis and transferred onto a PVDF membrane, which was subsequently blocked with 5% non-fat milk. The PVDF membrane was then incubated with specific primary antibodies (Anti-LC3, 1:2000, 382687, ZEN-BIOSCIENCE; Anti-P62, 1:1500, HA721171, HUABIO; Anti-RAGE, 1:1500, 16346-1-AP, Proteintech Group, Inc; Anti-PAFAH1B1, 1:1500, k2217, Proteintech Group, Inc; Anti-β-actin, 1:4000, AC026, ABclonal). After incubation with secondary antibodies, the proteins were observed by chemiluminescence. Image J software was used for protein gray value analysis. The original, unedited Western blot images have been included in the Supplementary Information PDF and are labeled as Supplementary Fig. 8.

Protein half-life assay

si-NC and si-PAFAH1B1-2 were transfected into BV2 cells, respectively. Cells were then exposed to cycloheximide (CHX, MCE) at a final concentration of 50 μg/mL for the indicated time points (0, 2, 4, 6, and 8 h). The expression levels of RAGE protein in each group were detected by Western blot (WB), and the half-life curve of RAGE protein was plotted.

Bioinformatics analysis

We used UbiBrowser (http://ubibrowser.ncpsb.org) to predict the interactions between ubiquitin ligases and their substrates. This platform identifies potential E3-substrate pairs by integrating structural features, domain interactions, and co-expression data. We used the default parameters for prediction and selected high-confidence prediction results for subsequent experimental validation.

Statistical analysis

Excel 2023 software was used to establish the database, and SPSS 22.0 and Graphpad Prism 10.1.2 software were employed for data analysis. Data were expressed as mean ± standard deviation (mean ± SD) based on normal distribution assumptions. Comparisons between the control group and the model group in vitro were conducted using an independent samples t-test. For comparisons among three groups, one-way ANOVA and two-way ANOVA were performed, followed by pairwise comparisons using the LSD method. A p-value < 0.05 was considered statistically significant. In addition, Pearson correlation analysis was used to explore the linear correlation between variable X and variable Y. The correlation coefficient (r) and p-value were calculated. A p-value < 0.05 was considered statistically significant.

All the original data from the experiments have been placed in Supplementary Data 1.

Ethical and legal declarations

This study and the included experimental procedures were approved by Guizhou Medical University. All animal experiments were approved by the Animal Care and Use Committee of the Ethical Institution of Guizhou Medical University.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Results

The expression levels of inflammatory factors and irisin in patients with PD

Our previous study found that the inflammatory response significantly impacts the progression of PD⁵. To verify the inflammatory immune status in PD patients, we quantitatively detected the expression levels of TNF-α and IL-1β in the blood serum of PD patients. The results showed that the levels of TNF-α and IL-1β in the peripheral blood serum of PD patients were significantly higher than those in the control group (p < 0.05) (Fig. 1A and B). We further measured serum irisin levels in the peripheral blood of PD patients. The results indicated that serum irisin levels in PD patients were significantly lower than those in controls (p < 0.05) (Fig. 1C). The correlation analysis indicated that the serum irisin level was negatively correlated with the UPDRS score (r = -0.3532, p < 0.0001) (Fig. 1D).

Fig. 1. The expression levels of inflammatory factors and irisin in patients with PD.

Comparison of IL-1β (A) and TNF-α (B) levels in the peripheral blood serum between PD patients and the control group. Control vs PD, *p < 0.05. C Comparison of serum Irisin levels in the peripheral blood between PD patients and the control group. Control vs. PD, *p < 0.05. D The correlation between irisin level and UPDRS score.

Improvement of PD-like symptoms and brain region distribution of irisin in mice model

To verify the therapeutic effect of irisin on PD mice, we first constructed a PD model using MPTP (Fig. 2A). The successfully established PD model mice were randomly divided into two groups: the PD group and the PD + Irisin group. Body weight was dynamically monitored in the control, PD, and PD + Irisin groups. The results showed that there were no significant changes in body weight among the three groups throughout the experiment, suggesting that intravenous administration of irisin (60 μg/kg/day) did not affect body weight during treatment (Fig. 2C). Irisin treatment significantly improved PD-like behaviors, including increased swimming scores and decreased head-down time and pole-climbing time (Fig. 2D and E). These findings indicate that irisin can effectively enhance the behavioral performance of PD mice.

To evaluate whether exogenous irisin injected via the tail vein of mice could effectively penetrate the blood–brain barrier and the changes in its content in different brain regions after tail vein injection, the levels of irisin in the tissue homogenates of the cortex, cerebellum, and substantia nigra striatum of each group of mice were detected by ELISA. It was found that compared with the control group, there was no significant difference in the levels of irisin in the cortical and cerebellar homogenates of PD group mice (p > 0.05), but the level of irisin in the substantia nigra striatum of the PD group mice was significantly lower than that of the control group mice (p < 0.05) (Fig. 2F); while compared with the PD group mice, the level of irisin in the cortex of PD + Irisin group mice increased, but there was no statistical significance (p > 0.05), the level of irisin in the cerebellum did not change significantly (p > 0.05), and the level of irisin in the substantia nigra striatum increased significantly (p < 0.05) (Fig. 2F). Finally, the irisin level in peripheral blood was detected in the PD group and returned to a normal level after intervention (Fig. 2G).

Effects of irisin on substantia nigra and striatum of midbrain in mice

First, we observed the substantia nigra and striatum of the midbrain in each group using HE staining. Compared with control mice, PD mice exhibited disordered neuronal cell arrangement and a reduced number of neurons in both regions (Fig. 3A). In contrast, the PD + Irisin group showed an increased number of neurons and partial improvement in cell arrangement (Fig. 3A). Next, Nissl staining was used to examine neurons in the substantia nigra and striatum. Compared with the control group, PD mice had lighter neurons and fewer Nissl bodies, with a more significant reduction in the overall number of neurons in the striatum and the substantia nigra zona compacta (SNpc) (Fig. 3B and C). The PD + Irisin group showed partial recovery in neuron numbers and darker staining compared to the PD group, although not fully restored to control levels (Fig. 3B and C). Tyrosine hydroxylase (TH), a key rate-limiting enzyme in dopamine biosynthesis, was assessed via immunohistochemistry. The TH expression in the substantia nigra and striatum was primarily localized in the cytoplasm of neurons and synapses. Compared with controls, PD mice had significantly reduced TH content and disordered neuronal arrangement (Fig. 3D–F). The PD + Irisin group exhibited a significant increase in TH content compared to the PD group (Fig. 3D–F).

Fig. 3. Effects of irisin on substantia nigra and striatum of midbrain in mice.

A HE staining results of the substantia nigra and striatum of midbrain in each group. Scale bars: 50 µm. B and C Nissl staining results of neurons in the substantia nigra and striatum. Scale bars: 50 µm. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. D–F Immunohistochemistry results of tyrosine hydroxylase (TH) in the substantia nigra and striatum. Scale bars: 50 µm. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05.

Additionally, irisin treatment significantly suppressed α-synuclein expression in the substantia nigra and striatum (Supplementary Fig. 1A and B), further supporting its potential therapeutic role in PD. To evaluate the safety of irisin treatment, we examined major organs in vivo and found no significant changes in body weight (Fig. 2E) or other important organs, confirming its safety at the therapeutic dose (Supplementary Fig. 2).

Inhibition of inflammatory response by irisin in PD mice

Compared with the control mice, the serum levels of TNF-α, interleukin-1α (IL-1α), and IL-1β in PD mice were significantly elevated. In the PD + Irisin group, there was a significant reduction in pro-inflammatory factors in peripheral blood compared to the PD group (p < 0.05) (Fig. 4A). The serum levels of anti-inflammatory cytokines interleukin-4 (IL-4) and interleukin-10 (IL-10) were measured in all groups. Compared with the control group, the PD group exhibited significantly lower levels of these anti-inflammatory cytokines (p < 0.05) (Fig. 4B). Conversely, the PD + Irisin group showed a significant increase in the expression of anti-inflammatory factors in the serum compared to the PD group (p < 0.05) (Fig. 4B).

Fig. 4. Inhibition of inflammatory response by irisin in PD mice.

A Comparison of serum levels of TNF-α, IL-1α, and IL-1β among the control, PD, and PD + Irisin groups. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. B Comparison of serum levels of anti-inflammatory cytokines IL-4 and IL-10 among the control, PD, and PD + Irisin groups. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. C Proportions of CD86⁺/Iba1⁺ cells in the substantia nigra and striatum among the control, PD, and PD + Irisin groups. Control vs. PD, *p < 0.05; PD + Irisin vs PD, *p < 0.05. Scale bars: 20 µm. D Proportions of CD206⁺/Iba1⁺ cells in the substantia nigra and striatum among the control, PD, and PD + Irisin groups. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. Scale bars: 20 µm. E CD206+ and CD86+ were measured and analyzed by flow cytometry. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05.

Using Iba1 immunofluorescence, microglia in the substantia nigra and striatum were labeled. Red fluorescence from Iba1 marked microglia, green fluorescence from CD86 indicated pro-inflammatory microglia, and green fluorescence from CD206 indicated anti-inflammatory microglia. Observations revealed that red Iba1 fluorescence intensity in the substantia nigra and striatum of PD mice was significantly higher than in the control group (p < 0.05), with a significantly higher proportion of CD86+/Iba1+ cells (p < 0.05) (Fig. 4C). This suggests an increased presence of pro-inflammatory microglia in the substantia nigra and striatum of PD mice. After irisin treatment, the proportion of CD86+/Iba1+ cells in the PD + Irisin group was significantly lower than in the PD group (p < 0.05) (Fig. 4C). Additionally, the proportion of CD206+/Iba1+ cells in the substantia nigra and striatum of PD mice was significantly lower than in the control group (p < 0.05), indicating a reduced presence of anti-inflammatory microglia (Fig. 4D). Following irisin treatment, the proportion of CD206+/Iba1+cells in the PD + Irisin group was significantly higher than in the PD group (p < 0.05) (Fig. 4D). Meanwhile, flow cytometry analysis showed that after irisin treatment, cells were able to reduce the elevated CD86+/CD206+ cell ratio in PD mice, indicating that irisin has an impact on the polarization transformation of microglia (Fig. 4E).

Irisin inhibits microglial inflammation by activating autophagic flux

As a critical immunomodulatory cell population in the nervous system, microglia play a pivotal role in regulating the progression of PD^14,15. In vivo assays demonstrated that irisin significantly attenuated the inflammatory response of microglia in the substantia nigra and striatum of PD mice. To further elucidate the underlying mechanisms, BV2 cells were utilized as an in vitro model. Our findings revealed that treatment with 100 ng/mL LPS for 12 h-induced morphological changes in BV2 cells, including swelling, increased proliferation, and transformation into round or rod-shaped phenotypes, accompanied by the disappearance of cellular processes, indicating effective activation of BV2 cells (Supplementary Fig. 3A). Subsequent treatment of LPS-activated BV2 cells with 100 ng/mL irisin for 12 h resulted in a marked reduction in the number of activated cells and alleviation of cell swelling. Moreover, irisin treatment significantly suppressed the inflammatory response of microglial cells (Fig. 5A–D). These data collectively suggest that irisin may mitigate PD progression by inhibiting microglial inflammation.

Fig. 5. Irisin inhibits microglial inflammation by activating autophagic flux.

A–D Effects of irisin on the inflammatory response of LPS-activated BV2 cells. Treatment with 100 ng/mL irisin for 12 h reduces the number of activated cells, alleviates cell swelling, and suppresses the inflammatory response of microglial cells. PBS vs LPS, *p < 0.05; LPS + Irisin vs. LPS, *p < 0.05. E and F Expression of LC3II/I and P62 in PD mice and BV2 cells. Both the PD group and the LPS-treated group exhibit decreased LC3II/I expression and increased P62 expression. E Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. F PBS vs. LPS, *p < 0.05; LPS + Irisin vs, LPS, *p < 0.05. G Restoration of autophagic flux in BV2 cells by irisin after labeling autophagosomes with RFP-GFP-LC3. Scale bars: 10 µm. PBS vs. LPS, *p < 0.05; LPS + Irisin vs LPS, *p < 0.05. H TEM results showing the number of autolysosomes in the LPS group and the restoration effect of irisin treatment. PBS vs LPS, *p < 0.05; LPS + Irisin vs. LPS, *p < 0.05.

To further investigate the therapeutic effects of irisin on PD, we employed RNA-seq to analyze the substantia nigra samples from both PD and irisin-treated groups. The screening criteria for DEGs were set at a p-value < 0.05 and fold change > 1. A total of 803 genes were up-regulated and 725 genes were down-regulated in the irisin-treated group compared to the PD group (Supplementary Fig. 4). GO and KEGG pathway enrichment analyses were conducted on these DEGs. GO analysis revealed that, in Biological Process (BP), the DEGs were primarily enriched in nervous system development, ion transmembrane transport, and ion transport pathways. In molecular function (MF), the DEGs were mainly enriched in protein binding, ion channel activity, and voltage-gated ion channel activity. For the cellular component (CC), the DEGs were predominantly enriched in synapse formation, cell junctions, and glutamatergic synapses (Supplementary Fig. 5A–C). KEGG pathway enrichment analysis indicated that the DEGs were significantly enriched in inflammatory response pathways, AGE-RAGE signaling pathways, and mTOR signaling pathways. Further functional analysis of the pathway enrichment results showed that autophagy-related processes were prominently enriched (Supplementary Fig. 5D and E).

To test this hypothesis, we first examined the expression of LC3II/I and P62 in PD mice and BV2 cells. We observed that both the PD group and the LPS-treated group exhibited decreased LC3II/I expression and increased P62 expression (Fig. 5E and F). This paradoxical phenomenon suggests that, in addition to reduced autophagy levels, autophagic flux is impaired in both PD and LPS groups. After labeling autophagosomes with RFP-GFP-LC3, we found that irisin significantly restored autophagic flux in BV2 cells (Fig. 5G). Further analysis of TEM results revealed a reduction in the number of autolysosomes in the LPS group, which was restored by irisin treatment (Fig. 5H). These findings confirm that autophagy may play a crucial role in the therapeutic effects of irisin.

To verify that irisin-mediated inhibition of microglial inflammation is mediated by autophagy, BV2 cells were treated with Rapamycin (Rap), an autophagy activator¹⁶, and Bafilomycin A1 (BafA1), which inhibits autophagic flux by blocking the fusion of autophagosomes with lysosomes^17,18. Firstly, we examined the levels of LC3 and P62 proteins to confirm the effects of these treatments on autophagy (Supplementary Fig. 3B). BafA1 reversed the inhibitory effect of irisin on both pro-inflammatory and anti-inflammatory markers, while Rapamycin exhibited similar effects to irisin in reducing inflammation (Supplementary Fig. 3C and D). This phenomenon was consistently observed in subsequent assays of inflammatory factors (Supplementary Fig. 3E). These findings suggest that autophagic flux plays a crucial role in irisin-mediated inhibition of the inflammatory response in BV2 cells.

Effect of irisin on RAGE mRNA and protein expression in PD mice

KEGG pathway enrichment analysis in the substantia nigra of PD mice and PD + Irisin mice revealed that differentially expressed genes were significantly enriched in the AGE-RAGE signaling pathway (Supplementary Fig. 5D). Previous studies indicate that the receptor for advanced glycation end products (RAGE) plays a critical role in neuroinflammatory immunity and PD. Based on this, we hypothesized that RAGE protein is involved in irisin’s regulation of the inflammatory response in the nervous system. To test this hypothesis, we examined RAGE mRNA and protein expression in the substantia nigra of PD mice. Compared to controls, PD mice exhibited a significant increase in RAGE protein levels (p < 0.05) (Fig. 6A and B). However, in PD + Irisin mice, this increase was significantly attenuated (p < 0.05). Notably, there were no significant changes in RAGE mRNA expression across all groups (p > 0.05) (Fig. 6A and B). In BV2 microglial cells, RAGE protein expression was significantly elevated in the LPS-stimulated group compared to the PBS control group (p < 0.05) (Fig. 6C and D). Treatment with irisin significantly inhibited this elevation (p < 0.05). Consistent with our findings in vivo, there were no significant differences in RAGE mRNA expression between the groups (p > 0.05) (Fig. 6C and D). These results align with our sequencing data, suggesting that post-translational modifications of the RAGE protein may play a crucial role in irisin’s protective effects against PD.

Fig. 6. Effect of irisin on RAGE and PAFAH1B1 mRNA and protein expression in PD mice.

Expression of RAGE mRNA and protein in the substantia nigra of PD mice, PD + Irisin mice, and the control group. A Control vs. PD, not significant; PD + Irisin vs. PD, not significant. B Control vs. PD, *p < 0.05; PD +  Irisin vs. PD, *p < 0.05. Expression of RAGE mRNA and protein in BV2 microglial cells in the LPS-stimulated group, PBS control group, and irisin-treated group. C PBS vs. LPS, not significant; LPS + Irisin vs. LPS, not significant. D PBS vs. LPS, *p < 0.05; LPS + Irisin vs. LPS, *p < 0.05. E Prediction of 35 candidate ubiquitin E3 ligases using the Ubibrowser database. F Venn diagram screening the intersection of ubiquitination E3 ligases with interacting and differentially expressed mRNAs. Expression of PAFAH1B1 mRNA and protein in the substantia nigra of PD mice (G) and the LPS-induced BV2 microglial model (H). G Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. H PBS vs. LPS, *p < 0.05; LPS + Irisin vs. LPS, *p < 0.05.

In vitro and in vivo data were used to verify the effect of irisin on the expression level of PAFAH1B1

Our functional enrichment analysis indicated that ubiquitination may play a role in irisin regulation, as it is an important post-translational modification of proteins (Supplementary Fig. 5B). To identify potential RAGE-binding ubiquitin E3 ligases, we utilized the Ubibrowser database, which predicted 35 candidate ubiquitin E3 ligases (Fig. 6E). By intersecting these predictions with significantly up-regulated differentially expressed genes from our irisin group, Venn diagram analysis revealed that PAFAH1B1 might be a key ubiquitination regulatory molecule for RAGE protein (Fig. 6F). We found that PAFAH1B1 mRNA and protein expression was significantly down-regulated in the substantia nigra of PD mice compared to the control group (p < 0.05) (Fig. 6G). This down-regulation was reversed by irisin treatment in the PD + Irisin group, with statistically significant differences (p < 0.05) (Fig. 6G). Similarly, in the LPS-induced BV2 microglial model, PAFAH1B1 mRNA and protein expression levels were significantly decreased compared to the PBS group (p < 0.05) (Fig. 6H). However, irisin treatment significantly increased PAFAH1B1 expression levels, with statistically significant differences (p < 0.05) (Fig. 6H).

Irisin inhibits inflammation by regulating RAGE ubiquitination through PAFAH1B1

To demonstrate the potential role of PAFAH1B1 as a ubiquitinated E3 ligase, molecular docking of PAFAH1B1 (PDB: P43034) and UB was performed from the EMBL-EBI database, and a protein-binding free energy of −10.7 kcal/mol was found between them (Supplementary Fig. 6A). Moreover, overexpression of PAFAH1B1 significantly decreased the expression level of RAGE protein, which was reversed by the addition of MG132 (Fig. 7A). This further demonstrated that RAGE inhibition by PAFAH1B1 was mediated through the proteasome pathway. The binding relationship between PAFAH1B1 and RAGE was also analyzed, and the binding energy of PAFAH1B1 was −7.7 kcal/mol (Fig. 7B). Co-IP and immunolocalization results confirmed their binding interaction (Fig. 7C, D and Supplementary Fig. 6B).

Fig. 7. Irisin inhibits inflammation by regulating RAGE ubiquitination through PAFAH1B1.

A Effect of overexpression of PAFAH1B1 on RAGE protein expression and the reversal effect of MG132. NC vs. oe-PAFAH1B1, *p < 0.05; oe-PAFAH1B1 + MG132 vs. oe-PAFAH1B1, *p < 0.05. B Binding energy analysis of PAFAH1B1 and RAGE. C and D Co-IP and immunolocalization results confirming the binding interaction between PAFAH1B1 and RAGE. Scale bars: 10 µm. E Protein half-life analysis of RAGE when PAFAH1B1 is inhibited. F: Ubiquitination level of RAGE upon co-treatment with si-PAFAH1B1 and MG132. G Increase in the ubiquitination level of RAGE in BV2 cells after irisin addition. H Reversal of irisin-induced ubiquitination of RAGE by combining irisin with si-PAFAH1B1.

To confirm the role of ubiquitination regulation in the inhibition of RAGE expression by PAFAH1B1, we first constructed a si-PAFAH1B1 cell model and selected si-PAFAH1B1-2 for subsequent studies based on its inhibitory effect (Supplementary Fig. 6C). Protein half-life analysis revealed that the degradation rate of RAGE was significantly reduced when PAFAH1B1 was inhibited (Fig. 7E). More direct evidence showed that the ubiquitination level of RAGE decreased upon co-treatment with si-PAFAH1B1 and MG132 (Fig. 7F). In BV2 cells, irisin addition increased the ubiquitination level of RAGE (Fig. 7G). Finally, we found that combining irisin with si-PAFAH1B1 reversed the irisin-induced ubiquitination of RAGE (Fig. 7H). These results support our initial hypothesis.

PAFAH1B1 regulates RAGE expression through ubiquitination at the K61 and K169 sites on the K48-linked polyubiquitin chain

To explore the ubiquitination modification sites of RAGE, the ubiquitination modification sites of RAGE were first predicted using the database GPS-uber. The results showed that there might be ubiquitination modification at K61 and K169 sites (Fig. 8A and Supplementary Fig. 6D, E). K61 and K169 were found to be highly conserved among the three species (Fig. 8B and Supplementary Fig. 6D). Next, we constructed the single-point mutation plasmids of RAGE-K61R and K169R and the 2KR plasmid, respectively, to examine the ubiquitination of RAGE. We found that K61R and K169R partially attenuated the ubiquitination level of RAGE, while 2KR attenuated almost all of the ubiquitination (Fig. 8C). This result implies that PAFAH1B1 mainly regulates the ubiquitination of RAGE through two sites, K61 and K169. Further experiments also demonstrated that PAFAH1B1 regulates RAGE ubiquitination through the K48 chain (Fig. 8D).

Fig. 8. PAFAH1B1 regulates RAGE expression through ubiquitination at the K61 and K169 sites on the K48-linked polyubiquitin chain.

A and B Prediction of ubiquitination modification sites of RAGE using the database GPS-uber, showing potential ubiquitination modification at K61 and K169 sites, which are highly conserved among the three species. C Effects of single-point mutation plasmids (RAGE-K61R, K169R) and the 2KR plasmid on the ubiquitination level of RAGE. K61R and K169R partially attenuate the ubiquitination level, while 2KR attenuates almost all of it. D Evidence that PAFAH1B1 regulates RAGE ubiquitination through the K48 chain. E–H Effects of PAFAH1B1 and RAGE on autophagy and inflammation in BV2 cells. si-PAFAH1B1 reverses the irisin-induced enhancement of autophagic flux and the inhibitory effect on pro-inflammatory markers, while overexpression of PAFAH1B1 mimics the effects of irisin treatment. LPS vs. LPS + Irisin, *p < 0.05; LPS + Irisin vs. LPS + Irisin + si-PAFA, *p < 0.05; LPS + Irisin + si-PAFA vs. LPS + oe-PAFA, *p < 0.05; LPS vs. LPS + oe-PAFA, *p < 0.05; LPS + Irisin vs. LPS + oe-PAFA + oe-RAGE, *p < 0.05.

The effects of PAFAH1B1 and RAGE on autophagy and inflammation in BV2 cells were further confirmed. si-PAFAH1B1 reversed the irisin-induced enhancement of autophagic flux, while overexpression of PAFAH1B1 mimicked the effects of irisin treatment (Fig. 8E). Similarly, si-PAFAH1B1 treatment reversed the inhibitory effect of irisin on pro-inflammatory markers, whereas overexpression of PAFAH1B1 achieved the same inflammatory suppression effect, which was subsequently reversed by RAGE overexpression (Fig. 8F). This trend was consistently observed in the detection of both pro-inflammatory and anti-inflammatory markers (Fig. 8G), as well as in the measurement of inflammatory cytokines (Fig. 8H). These findings further suggest that PAFAH1B1 is a critical regulatory molecule for irisin to suppress the inflammatory response by inhibiting RAGE expression.

TFEB-NLRP3 was the effector pathway of irisin in inhibiting microglial inflammatory response

This study demonstrated that irisin could improve the progression of PD by inhibiting microglial inflammation by restoring autophagic flux. To further elucidate the underlying mechanisms, we first investigated the cause of autophagy blockage and found that the expression of lysosomal markers TFEB and LAMP1 was significantly reduced in both PD and BV2 models (Fig. 9A and Supplementary Fig. 7A). Treatment with irisin restored their expression, suggesting impaired lysosomal function as a potential mechanism. Furthermore, we examined TFEB expression in vitro and observed that irisin promoted the nuclear translocation of TFEB (Fig. 9B). When TFEB activity was inhibited, its regulation of NLRP3 via the autophagy pathway was also compromised^19,20. We detected increased NLRP3 protein levels in the striatum and substantia nigra of PD models, which were attenuated by irisin treatment (Fig. 9C). Notably, NLRP3 mRNA levels remained unchanged, indicating selective modulation at the protein level (Supplementary Fig. 7B and C). Further analysis of pyroptosis markers revealed that irisin could inhibit pyroptosis in both PD and BV2 cell inflammation models (Fig. 9D and E). IHC staining showed swollen and indistinct nuclei in microglia from the PD group, effects that were reversed by irisin treatment, further confirming its role in inhibiting microglial pyroptosis (Fig. 9F). To explore why NLRP3 changes only occurred at the protein level, we used IF to co-localize NLRP3 and LAMP1, revealing significant co-localization (Fig. 9G). This suggests that the autophagy pathway is an effective means for irisin to inhibit NLRP3. Additionally, treatment of BV2 cells with MG132 and 3-MA confirmed that NLRP3 expression was restored only when 3-MA was added, supporting our hypothesis (Fig. 9H).

Fig. 9. TFEB-NLRP3 pathway mediates the anti-inflammatory effects of irisin in PD models.

A Expression of lysosomal markers TFEB and LAMP1 in the substantia nigra of PD mice and LPS-induced BV2 microglial cells. Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. B The nuclear and cytoplasmic expression of TFEB in BV2 cells treated with LPS and LPS combined with irisin was detected by WB. PBS vs. LPS, *p < 0.05; LPS + Irisin vs. LPS, *p < 0.05. C Immunofluorescence was used to detect the expression level of NLRP3 in the brain tissue of PD. Scale bars: 20 µm. Western blot analysis of NLRP3 inflammasome components (NLRP3, ASC, cleaved Caspase-1) in PD mice (D) and LPS-treated BV2 cells (E). E Control vs. PD, *p < 0.05; PD + Irisin vs. PD, *p < 0.05. F PBS vs. LPS, *p < 0.05; LPS + Irisin vs. LPS, *p < 0.05. F The morphology and expression of microglial marker IBA-1 in the brain tissue of the PD model were examined by IHC staining. G Immunofluorescence was used to detect the subcellular localization of LAMP1 and NLRP3 after LPS and irisin treatment. Scale bars: 20 µm. H The effect of autophagy intervention on the expression level of NLRP3 after LPS and irisin treatment was detected by WB. LPS vs. LPS + Irisin + MG132, *p < 0.05; LPS + Irisin + MG132 vs. LPS + Irisin + 3-MA, *p < 0.05; LPS + Irisin + 3-MA vs. LPS + Irisin, *p < 0.05; LPS vs. LPS + Irisin, *p < 0.05. I Behavioral assays were conducted to detect the behavioral changes in the PD model after overexpression of irisin and NLRP3. PD vs. PD + Irisin, *p < 0.05; PD + Irisin vs. PD + Irisin + NLRP3, *p < 0.05. J ELISA was used to detect the expression levels of inflammatory factors in each treatment group. PD vs. PD + Irisin, *p < 0.05; PD + Irisin vs. PD + Irisin + NLRP3, *p < 0.05. K The molecular mechanism diagram of irisin regulating autophagolysosomal damage recovery through PAFAH1B1-mediated ubiquitination of RAGE and improving the progression of PD.

Finally, to investigate the function of NLRP3, an NLRP3 overexpression model was constructed in BV2 cells. Initially, autophagy flux analysis revealed that NLRP3 overexpression did not affect irisin-induced recovery of autophagy levels (Supplementary Fig. 7D and E) but reversed the inhibitory effect of irisin on inflammatory activation in BV2 cells (Supplementary Fig. 7F and G). Further in vivo assays confirmed that the therapeutic effects of irisin on PD-like behaviors were attenuated when NLRP3 was overexpressed, with associated inflammatory markers also being restored (Fig. 9I and J). These findings support the conclusion that irisin improves PD progression via the TFEB–NLRP3 pathway.

Discussion

This study systematically investigated the therapeutic potential of irisin in PD and its underlying molecular mechanisms. Through clinical sample analysis, animal model experiments, and in vitro assays, we found that irisin significantly improved motor dysfunction and inhibited neuroinflammation in PD mice, exerting neuroprotection by regulating autophagy and ubiquitination pathways. Specifically, irisin can penetrate the blood-brain barrier and enter brain tissue, where it restores microglial autophagic flux through PAFAH1B1-mediated RAGE ubiquitination, leading to the degradation of NLRP3 (Fig. 9K). This process inhibits the pro-inflammatory response of microglia and mitigates the progression of PD.

Irisin, a peptide produced by exercise, is involved in the regulation of energy metabolism in the body^21,22. Existing studies have found that irisin can cross the blood-brain barrier and participate in a variety of pathophysiological processes in the central nervous system²³. More and more studies have found that irisin is closely related to the occurrence of cognitive dysfunction, and have begun to use irisin as a potential therapeutic target for neurodegenerative diseases^24,25. However, there are few studies on the relationship between irisin and the pathogenesis of PD, especially on the changes in irisin levels in PD patients. To this end, we further examined the peripheral serum levels of irisin in 120 PD patients and found that the peripheral serum levels of irisin in PD patients were significantly lower than those in healthy controls, which was inversely proportional to the elevation of inflammatory factors in the peripheral serum of PD patients. We hypothesized that irisin, as an actin-associated myokine with anti-inflammatory and anti-apoptotic effects, may play a role in the development of PD by inhibiting immune-inflammatory responses. Further study on this aspect may provide important clues for us to further elucidate the pathogenesis of PD and discover its potential therapeutic application. In the MPTP-induced PD mouse model, exogenous irisin intervention not only improved motor behavior (swimming score and pole climbing time) but also restored TH expression and reduced abnormal aggregation of α-synuclein in the nigrostriatal. These results are consistent with previous studies showing that irisin protects neurons through anti-inflammatory and antioxidant effects^23,26. Notably, irisin can specifically cross the blood–brain barrier into the substantia nigra region, indicating its targeting in the central nervous system, which provides an important basis for its clinical application. This study exclusively employed male C57BL/6 mice to circumvent the potential interference of the estrous cycle in female mice on inflammatory markers and the irisin signaling pathway, thereby ensuring model stability. Epidemiological data indicate that the incidence of Parkinson’s disease is significantly higher in males than in females²⁷, suggesting that the gonadal hormone milieu may be implicated in disease progression^28,29. Future studies should examine irisin’s regulation of the PAFAH1B1–RAGE–TFEB–NLRP3 pathway in female and OVX models to clarify estrogen’s role, crucial for developing sex-specific PD treatments.

In addition, the results of this study indicate that irisin exerts its anti-inflammatory effects through the synergistic regulation of the pro-inflammatory and anti-inflammatory cytokine network. The experimental data show that irisin significantly inhibits the expression of pro-inflammatory cytokines, such as TNF-α, IL-1α, and IL-1β, while maintaining the levels of anti-inflammatory cytokines, such as IL-10 and TGF-β. This dual regulatory mechanism is highly consistent with the findings in various disease models. In neuroinflammatory models, irisin regulates the polarization balance of microglia, suppressing the pro-inflammatory M1 phenotype and promoting the transformation of the anti-inflammatory M2 phenotype³⁰. In acute lung injury, irisin downregulates the expression of TNF-α and IL-6 while upregulating the expression of IL-10³¹. This “biphasic regulation” enables irisin to control the inflammatory response process more comprehensively: on the one hand, it reduces tissue damage by inhibiting pro-inflammatory cytokines, and on the other hand, it promotes tissue repair by maintaining anti-inflammatory cytokines. This synergistic mode of action may be the key mechanism underlying the effective therapeutic outcomes of irisin in different inflammatory diseases, providing an important theoretical basis for the development of novel anti-inflammatory drugs.

We focus on the potential impact of irisin on microglial polarization and neuroglial cells. Current studies have demonstrated that irisin can promote the transformation of microglia from the pro-inflammatory M1 phenotype to the anti-inflammatory M2 phenotype. Moreover, irisin has an impact on astrocytes. It can protect them from Aβ-induced damage, inhibit the release of inflammatory^32,33. factors, downregulate the expression of COX-2, and reduce Aβ pathology by increasing the release of NEP^34,35. These findings suggest that irisin has great potential in modulating the function of neuroglial cells, suppressing neuroinflammation, and exerting neuroprotective effects. In future research, we aspire to validate this mechanism in other types of glial cells, such as astrocytes, and to delve into its potential clinical application value.

RAGE is a transmembrane immunoglobulin multiligand receptor. Extensive studies have demonstrated that RAGE can bind to various extracellular ligands, leading to the production of ROS and inflammatory factors. The AGE–RAGE pathway is not only involved in inflammatory responses but is also implicated in the development of diabetes, coronary heart disease, and other conditions^36–38. Recent research has identified RAGE as one of the pattern recognition receptors (PRRs) in microglia^39,40. Under physiological conditions, RAGE expression is relatively low in most tissues; however, it is upregulated in pathological states such as neurodegenerative diseases, chronic inflammation, or diabetes^41,42. We further examined the expression of RAGE protein and mRNA in the substantia nigra of PD mice. Our results showed that RAGE protein expression in the substantia nigra of PD mice was significantly higher than in control mice, while irisin treatment significantly reduced RAGE protein levels compared to untreated PD mice. Notably, no differences were observed in RAGE mRNA expression levels among the three groups of mice. This suggests that the regulation of RAGE protein may involve post-translational modifications. These findings indicate that RAGE may play a crucial role in the progression of PD and the improvement of PD symptoms by irisin, potentially through the inhibition of microglial polarization and reduction of the inflammatory response.

In recent years, significant progress has been made in the research of protein ubiquitination in nervous system inflammation and microglia^43–45. As an important protein post-translational modification process, ubiquitination regulates the stability, localization, and function of proteins by covalently linking ubiquitin molecules, and then plays a key role in the development, function maintenance, and disease occurrence of the nervous system. The GPS-Uber database was used to analyze the ubiquitination sites of RAGE, and it was found that there were potential ubiquitination binding sites in RAGE, which further confirmed our previous hypothesis that ubiquitination regulation may be an important pathway for the post-translational modification of RAGE. Subsequently, intersection analysis of the significantly highly expressed mRNA in the nigra of irisin group mice and 35 ubiquitinated E3 ligases predicted to be related to RAGE by Ubibrowser database using the Venn diagram showed that PAFAH1B1 may be a ubiquitinated regulatory molecule of RAGE protein.

PAFAH1B1 (also known as Lis1) plays a crucial regulatory role in the acetylhydrolase complex. Recent studies have shown that PAFAH1B1 is essential for nuclear movement during neuronal migration^46,47. Transcriptome data from multiple human brain regions have identified PAFAH1B1 as a ubiquitylated E3 ligase in more than three brain regions of AD patients, suggesting its potential importance in the pathogenesis of AD⁴⁸. We found that 803 genes were upregulated in the substantia nigra of PD mice treated with irisin compared to untreated PD mice, and PAFAH1B1 was among them. However, there are currently no reports on the involvement of PAFAH1B1 in regulating RAGE or its role in the pathogenesis of PD. Therefore, based on previous findings, we hypothesized that PAFAH1B1 may be involved in neuroimmune regulation in PD by modulating RAGE protein expression through ubiquitination. To investigate this hypothesis, we obtained 3D structural models of PAFAH1B1 and RAGE from the RCSB PDB and SWISS-MODEL databases. The PAFAH1B1–RAGE complex model generated using the ClusPro database revealed interconnected domains between the two proteins, providing evidence for their direct interaction. To verify the binding of PAFAH1B1 and RAGE, we observed numerous yellow binding spots in cultured BV2 microglial cells via immunofluorescence, indicating co-localization. This finding was further confirmed by Co-IP assays, which demonstrated direct binding between PAFAH1B1 and RAGE in BV2 cells. Subsequent ubiquitination assays also confirmed that PAFAH1B1 regulates RAGE ubiquitination.

Our results demonstrated that LPS induced autophagy impairment and autophagic flux block in BV2 cells, which were significantly ameliorated by irisin treatment^49–51. This finding is consistent with previous reports indicating that restoring autophagic flux in microglia can improve their polarization in nervous system injury. Studies have shown that ketamine can regulate LPS-induced autophagic flux in microglia by modulating the HMGB1/RAGE axis and microglial polarization^42,52. Other studies have reported that inhibiting HMGB1/RAGE axis activation increases stress-induced mitophagic flux, thereby reducing sympathetic vasoconstriction and microglia-mediated neuroinflammation in the ventrolateral rostral medulla (RVLM). These findings highlight the role of RAGE in regulating the inflammatory response of microglia through the autophagy pathway⁵³. In this study, we found that irisin enhanced NLRP3 inflammasome degradation via autophagy by promoting TFEB nuclear translocation to restore lysosomal function (increased LAMP1 expression). The reduction in NLRP3 protein levels (but not mRNA) and its colocalization with LAMP1 suggest that the autophagy-lysosomal pathway plays a central role in irisin’s inhibition of pyroptosis and neuroinflammation. This result aligns with the study by Chen et al., which showed that TFEB activation improves PD pathology by regulating NLRP3¹⁹. Notably, overexpression of NLRP3 reversed the therapeutic effects of irisin, underscoring the critical role of this pathway in PD.

Despite revealing multiple mechanisms of irisin action, this study has limitations. The clinical sample size was limited. Animal experiments assessed only short-term efficacy, lacking evaluation of long-term toxicity or dose-response relationships. Furthermore, in vivo validation of the PAFAH1B1–RAGE axis remains insufficient, necessitating further study using conditional knockout (cKO) models.

This study demonstrates irisin’s inhibition of neuroinflammation via dual pathways: PAFAH1B1-mediated RAGE ubiquitination/degradation and TFEB-NLRP3 modulation. These findings identify a new target for PD immunomodulation. Combined with its favorable BBB penetration and safety, irisin is a promising clinical PD candidate.

Author contributions

Lijun Cai, Yin Liu, Shuang Tang, and Song Deng made contributions to the conception and design; Li Zhang, Yiping Wang, Bei Zhang, Bing Han, Rujia Xie, and Xin Liao analyzed and interpreted data; Lijun Cai drafted the article; Qin Yang revised it critically for important intellectual content; All authors approved the final version to be published.

Peer review

Peer review information

Communications Biology thanks Honoka Fujimori, Yasuhiro Kosuge and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Ibrahim Javed and Benjamin Bessieres. A peer review file is available.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. The data generated or analyzed during this study are included in this article and its supplementary information files (Supplementary Information and Supplementary Data 1). The RNA-seq data have been stored in the database. https://pan.baidu.com/s/1ny1Yx8PGhLlPdQPB5IRNqg?pwd=ca8t#list/path = %2F.

Competing interests

The authors declare no competing interests.

Participant consent

We confirm that all participants provided informed consent for participation in the study and for the publication of results.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-025-09389-7.