Irisin-integrin αV/β5 coupling of α-synuclein phagocytosis and clearance

Abstract

Parkinson’s disease-associated cognitive impairment (PD-CI) is closely linked to α-synuclein (α-syn) accumulation and synaptic dysfunction, yet effective disease-modifying strategies remain limited. Irisin is an exercise-inducible myokine with neuroprotective potential, but its receptor mechanisms and its role in α-syn clearance in PD-CI are poorly defined. Here, we observed that aerobic exercise markedly increased circulating irisin levels, reduced serum α-syn levels, and improved cognitive performance in a cohort of 21 PD patients. In addition, irisin signals through integrin αV/β5 to enhance microglial α-syn clearance, resulting in reduced α-syn burden and improved PD-CI. Mechanistically, irisin activates integrin αV/β5-FAK axis to promotes microglial phagocytic uptake of α-syn, while concurrently stabilizing HMGB1 to facilitate autophagy-lysosome mediated degradation of internalized α-syn, thereby coupling phagocytic uptake to efficient degradation. In summary, these results highlight a dual-module irisin-integrin αV/β5 mechanism that couples microglial phagocytosis and autophagy-lysosome clearance to reduce α-syn burden and ameliorate PD-CI.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12974-026-03882-4.

Introduction

Parkinson’s disease (PD) is traditionally characterized by motor symptoms [1]; however, cognitive impairment is increasingly recognized as one of its most disabling non-motor manifestations [2]. PD-associated cognitive impairment (PD-CI) spans deficits in executive function, attention, working memory, and visuospatial processing, and can progress to PD dementia [3, 4]. Current therapeutic strategies largely target dopaminergic motor symptoms, whereas effective strategies for PD-CI remain limited [5].

Pathological aggregation and propagation of α-synuclein (α-syn) are central features of PD pathogenesis and are closely associated with cognitive decline [6, 7]. Beyond the nigrostriatal system, α-syn pathology accumulates in cognition-relevant regions such as the hippocampus and prefrontal cortex, where it disrupts synaptic structure and plasticity and impairs neural circuit function. Growing evidence supports the concept that synaptic dysfunction is an early and direct driver of cognitive deficits in PD [8, 9]. Pathological α-syn can compromise synaptic transmission, trigger neuroinflammation, and facilitate prion-like spreading, thereby amplifying synaptic injury and accelerating cognitive impairment [8]. Thus, enhancing the effective clearance of pathological α-syn and limiting its accumulation represent promising avenues to mitigate PD-CI.

Microglia, the resident immune cells of the central nervous system, play a pivotal role in maintaining proteostasis and synaptic homeostasis through recognition, phagocytosis, and degradation of aberrant proteins and debris[10, 11]. In the context of α-syn pathology, microglia can internalize neuron-released α-syn and engage selective autophagy programs to promote its degradation, a process termed “synucleinphagy”. However, persistent exposure to fibrillar α-syn can also induce lysosomal stress and dysfunctional clearance, contributing to chronic neuroinflammation and exacerbated neuropathology [12, 13]. These observations imply that therapeutic benefit may require restoring effective clearance-enhancing both α-syn uptake and downstream autophagy-lysosome degradation, rather than increasing internalization alone. However, the upstream receptors and signaling hubs that coordinate microglial α-syn phagocytosis and degradation remain incompletely defined.

Irisin, a cleavage product of fibronectin type III domain-containing protein 5 (FNDC5), is a myokine induced by physical exercise and has been implicated in neuroprotection, synaptic plasticity, and cognitive enhancement [14, 15]. Recent studies propose integrin αV/β5 as a functional binding receptor for irisin, and integrin αV/β5 is known to regulate adhesion, cytoskeletal remodeling, migration, and phagocytic programs in immune cells [16]. Importantly, integrin-dependent signaling in microglia has been linked to the regulation of phagocytosis, inflammatory polarization, and lysosome-associated clearance pathways [17, 18]. These insights raise the possibility that the irisin-integrin αV/β5 axis may reprogram microglial phagocytosis-to-degradation capacity toward more efficient α-syn clearance, thereby mitigating α-syn-linked synaptic injury and improving cognitive outcomes in PD.

In the present study, our found that exercise training increases serum irisin levels in patients with PD, which is positively associated with cognitive improvement, and that exogenous irisin exerts neuroprotective effects in PD models by preventing α-syn accumulation. Mechanistically, irisin activates the integrin αV/β5–FAK axis to enhance microglial phagocytic uptake of α-syn, while stabilizing HMGB1 to promote the degradation of α-syn, thereby strengthening the coupling between phagocytosis and degradation. These findings link exercise-induced irisin to productive microglial α-syn clearance and support the irisin-integrin αV/β5 axis as a therapeutic target for PD-CI. These findings establish a receptor-driven signaling framework that connects exercise-elevated irisin to microglial α-syn clearance, and they provide a mechanistic basis for targeting the irisin-integrin αV/β5 axis to reduce α-syn burden and alleviate PD-CI.

Methods

Human tissues samples

In this study, we included 26 patients with PD who were recruited from the Department of Neurology and Parkinson’s Center of the Affiliated Hospital of Xuzhou Medical University. The studies involving humans were approved by the Affiliated Hospital of Xuzhou Medical University Ethics Committee (Approval number: XYFY2023-KL141-01). Of the 26 enrolled participants, 1 withdrew from further exercise training and follow-up because of low back pain, and 4 were lost to follow-up for personal reasons. Consequently, 21 participants completed the exercise intervention and all scheduled assessments, with no adverse events reported. The inclusion criteria were as follows: (i) patients with PD were diagnosed by specialists according to the 2015 International Movement Disorder Society clinical diagnostic criteria for Parkinson’s disease [19]; (ii) patients were aged 30–75 years and were classified as Hoehn and Yahr stage 1–3; (iii) patients had no contraindications to exercise and signed informed consent before participation; (iv) all participants completed an 8-week moderate-intensity aerobic exercise intervention on a stationary home exercise bike for 30–50 min at least 3 times per week; (v) the outcomes included the Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MOCA) and so on; (vi) all assessments were completed at baseline and after the 8-week intervention period; and (vii) fasting blood samples were collected according to standard procedures before the intervention and after completion of the 8-week training program, and plasma samples were isolated immediately and stored at − 80 °C for subsequent analysis.

Elisa

Plasma levels of α-syn and irisin were quantified using commercially available ELISA kits according to the manufacturer’s instructions (Jiang lai Biotechnology). Plasma samples were obtained by centrifugation of anticoagulated blood and stored at − 80 °C until analysis. Before measurement, samples were thawed on ice and diluted as required. Standards and plasma samples were added to 96-well plates in duplicate, followed by incubation with the corresponding detection antibody and HRP-conjugated secondary antibody. After color development with TMB substrate, absorbance was measured at 450 nm using a microplate reader. The concentrations of α-syn and irisin were determined based on standard curves generated in parallel.

Animal models

Animal experiments were conducted in accordance with the National Institutes of Health’s “Guidelines for the Care and Use of Laboratory Animals” and approved by the Ethics Committee of Xuzhou Medical University (Ethics Approval No.:202507T008). The acquisition and use of mice complied with relevant regulatory guidelines and were facilitated by the Laboratory Animal Center of Xuzhou Medical University (License No.: SYXK[Su]2020-0048). Male C57BL/6 mice (6–8 weeks old, weighing 20–30 g, 6 per group) were randomly assigned to the following groups: Control, PD, PD + Exercise (PD + Ex), PD + Exercise + Irisin neutralizing antibody group (PD + Ex + NE-Irisin), PD + Ex + PLX, PD + Irisin, PD + Irisin + ITGB5 inhibitor (PD + Irisin + Cilengitide), and PD + Irisin + FAK inhibitor (VS-4718). To induce the PD model, mice received intraperitoneal injections of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP, APExBIO, A3634) at 30 mg/kg/day, twice per week for 5 weeks, whereas control mice received an equal volume of saline. Mice in the exercise groups underwent treadmill training for 8 weeks, 5 days per week, using a progressive moderate-intensity protocol. After a short adaptation period, mice were trained at 10 m/min for 30 min/day during week 1, 12 m/min for 40 min/day during weeks 2–3, and 15 m/min for 60 min/day during weeks 4–8. For irisin neutralization antibody (Phoenix Pharmaceuticals, G-067-17), mice in the PD + Ex + Irisin group received tail-vein injections of irisin-neutralizing antibody at 10 µg/mouse during weeks 6, 7, and 8 of treadmill exercise. Mice in the corresponding control group received the same volume of control IgG according to the same schedule. For exogenous irisin (abclonal, RP01164) treatment, recombinant irisin was administered intraperitoneally at a dose of 200 µg/kg, three times per week, starting 3–7 days before the first MPTP injection and continuing throughout the 5-week modeling period. Mice in the corresponding control groups received an equal volume of saline, and, where appropriate, matched control IgG was administered in parallel as the antibody control for subsequent intervention experiments. For microglial depletion, mice in the PD + Irisin + PLX group received PLX3397 (Med Chem Express, HY-16749), at 40 mg/kg/day for 21 consecutive days. This regimen has been reported to effectively deplete brain microglia in mice [20]. For integrin inhibition, cilengitide (a selective integrin αv/β5 inhibitor, Selleck, USA) was administered intraperitoneally at 10 mg/kg three times per week, starting 3–7 days before the first MPTP injection and continuing throughout the 5-week modeling period, with each dose given 30 min to 2 h before irisin administration. This dose was selected based on previous in vivo studies using cilengitide as an integrin inhibitor. For FAK inhibition, mice in the PD + Irisin + FAK inhibitor group received VS-4718 (Selleck, S7653) at 50 mg/kg/day during the irisin intervention period.

Behavioral tests

Morris water maze test

The Morris water maze test was performed to evaluate spatial learning and memory. The apparatus consisted of a blue circular pool 120 cm in diameter and 30 cm in height, filled with water maintained at 24 °C. The pool was divided into four quadrants: north, east, south, and west. During the acquisition phase, mice underwent four training trials per day for five consecutive days. In each trial, mice were allowed 60 s to locate a hidden platform. After reaching the platform, mice remained on it for 10 s to ensure equal exposure to spatial cues. If a mouse failed to locate the platform within 60 s, it was gently guided to the platform and allowed to stay there for 10 s. Escape latency and swimming path length were recorded during the training period. A probe trial was conducted 24 h after the final training session. During the probe trial, the platform was removed, and each mouse was released from the same starting quadrant and allowed to swim freely for 60 s.

Open-field test

The open-field test was used to assess spontaneous locomotor activity and exploratory behavior. The apparatus consisted of a white plastic chamber measuring 50 cm × 50 cm × 50 cm, with a camera positioned above the arena to record mouse movement. After habituation to the testing environment, each mouse was placed in the center of the arena and allowed to explore freely for 5 min under normal lighting conditions. The arena was cleaned with 75% ethanol after each trial to eliminate olfactory cues.

Novel object recognition test

The novel object recognition test was performed to assess recognition memory, as previously described. The test consisted of habituation, sample, and test phases. During the habituation phase, each mouse was placed in the empty area for 10 min per day for two consecutive days. During the sample phase, two identical objects were placed in the central area of the arena, and each mouse was allowed to explore the arena freely for 5 min. During the test phase, one familiar object was replaced with a novel object, and each mouse was allowed to explore the two objects for 5 min. Exploration time was recorded using ANY-maze software (version 7.1, Stoelting, Wood Dale, IL, USA).

Cell culture and treatmet

Mouse BV2 microglial cells were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). BV2 cells were cultured in medium composed of DMEM-H supplemented with 10% FBS and 1% penicillin/streptomycin (Hycyte, TCM-G718) in a humidified incubator at 37 °C with 5% CO₂. On the second day after passaging, BV2 cells were treated with 20 nM irisin and 2 µg/mL α-syn preformed fibrils (PFFs) for 24 h. For irisin preparation, 10 mg irisin was diluted to a stock concentration of 0.1 mg/mL. For cell treatment, irisin was used at a final concentration of 20 nM for 24 h. For cycloheximide treatment, the protein synthesis inhibitor cycloheximide (CHX; Sigma Aldrich, R750107) was used to assess the stability of HMGB1. The degradation pathway of HMGB1 was further examined using the proteasome inhibitor MG132 (APEx BIO, A2585, 10 µM) and chloroquine (CQ; Aladdin, C19834, 100 µM). PFFs were purchased from Mabioway (MAG-7494-FITC). The product was supplied as sterile, filtered preformed fibrils at a concentration of 1 mg/mL. For cell treatment, PFFs were added to BV2 cells at a final concentration of 2 µg/mL. The fibrillar morphology and authenticity of the PFFs preparation were further confirmed by transmission electron microscopy.

Mouse primary microglia cells

Primary microglia cell cultures were generated from the cerebral cortices of five to eight neonatal mice (1 to 2 days old, mixed sex). Following previously published protocols [21–23], cortical tissues were first digested in 0.25% trypsin-EDTA for 30 min at 37 °C. The digested tissues were then mechanically triturated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (NEST Biotechnology, Cat# 209111). The resulting mixed cell suspension was filtered through a 70-µm cell strainer, after which the cells were seeded into poly-d-lysine–coated T25 flasks (BaiDi Biotechnology Co., Ltd., Cat#H804000) containing DMEM supplemented with 1% penicillin-streptomycin and 10% fetal bovine serum. When the mixed glial cultures reached confluency after about 10 days in vitro, microglia were harvested by shaking the flasks at 180 rpm for 1 h at 37 °C. Because microglia are only loosely attached to the surface layer of the mixed cells, this shaking step allowed them to detach from the flask surface. The supernatant was then transferred and the detached microglia were replated for 1 to 2 h to permit adherence. After attachment, the cells were washed twice with warm PBS to remove cell aggregates and debris, and fresh culture medium was subsequently added. Cells were collected for experiments when microglial confluency reached 60 to 70%.

Cell transfection

MYC-HMGB1, HA-UB, Flag-ITGB5, HA-FAK and WT, K43R, K112R, K114R, K128R, K146R, K6O, K110, K27O, K29O, K33O, K48O, K63O plasmids were purchased from Miaoling Bioscience. Plasmids were prepared using the abclonal plasmid extraction kit (RK30102) according to the manufacturer’s instructions. When BV2 cells reached 40%–60% confluence, plasmid transfection was performed using Lipofectamine 3000 (APExBIO, K2705). After 24 h of transfection, cells were subjected to the indicated treatments for an additional 24 h. Transfection efficiency was verified by western blotting. siITGB5 and siFAK were purchased from Hycyte and transfected into cells using Lipofectamine 3000 according to the manufacturer’s instructions. The ITGB5 siRNA sequence used was GGACAAUUGUAACUGCUCATT. The siFAK sequence used was AAGAUUACCAAUGCCUCCAAA.

RNA extraction and quantitative real-time polymerase chain reaction (PCR) analysis

For RNA extraction, culture medium was removed from BV2 cells, followed by cell lysis with TRIzol reagent, and total RNA was isolated according to the manufacturer’s instructions (Gene Better, R013-50). RNA concentration and purity were determined by measuring absorbance at 260 nm and calculating the A260/A280 ratio. First-strand cDNA was synthesized from 1 µg of total RNA in a 20 µL reaction system using the AB Script Neo RT Master Mix for qPCR with gDNA Remover kit (abclonal, RK20433), in accordance with the manufacturer’s protocol.

qPCR was performed using Bright Cycle Universal SYBR Green qPCR Mix with UDG (Abclonal, RK21219) on a Light Cycler 480 real-time PCR system. β-actin was used as the internal reference gene. The mouse-specific primer sequences were as follows: β-actin, Forward: GGCTGTATTCCCCTCCATCG and Reverse: CCAGTTGGTAACAATGCCATGT; HMGB1, Forward Primer: GGCGAGCATCCTGGCTTATC and Reverse Primer: GGCTGCTTGTCATCTGCTG; ITGB5, Forward Primer: GAAGTGCCACCTCGTGTGAA and Reverse Primer : GGACCGTGGATTGCCAAAGT. Relative mRNA expression levels were calculated using the 2^−ΔΔCt method. Each sample was analyzed in duplicate.

RNA sequencing

RNA sequencing and downstream bioinformatic analyses were performed by LC Bio (Zhejiang, China) using samples from the α-syn + Irisin and α-syn + Irisin + siITGB5 groups. Libraries were constructed and sequenced on the Illumina NovaSeq 6000 platform. Genes meeting the criteria of fold change ≥ 2 and P < 0.05 were considered significantly differentially expressed genes (DEGs). Functional annotation of these DEGs was then carried out by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses using the Metascape online platform (http://metascape.org).

Cell wound-healing and migration assays

For the wound-healing assay, BV2 cells were seeded into six-well plates and cultured until they reached 80%–90% confluence. A scratch was then made in the center of each well using a 200 µL pipette tip. The culture medium was subsequently replaced with serum-free medium, and the cells were further incubated for 24 h. The wound area was observed under a microscope, and the wound area in five different fields was measured using ImageJ software. The wound-healing rate was calculated as the ratio of the residual wound area to the initial wound area.

For the Transwell migration assay, 5 × 10⁴ BV2 cells were seeded into the upper chamber (NEST, 725301) in 200 µL serum-free medium, while the lower chamber was filled with medium containing 15% FBS. After 24 h of incubation, the cells that had migrated to the lower surface of the membrane were fixed with 4% paraformaldehyde and stained with 1% crystal violet (APExBIO K1184), followed by counting under a microscope. Five randomly selected fields were counted for each sample, and all experiments were performed in triplicate.

Latex beads phagocytosis

BV2 cells were seeded into 24-well plates at a density of 2 × 10⁵ cells per well and cultured for 24 h. Fluorescent latex beads (Sigma, L4655) were diluted 1:100 in phosphate-buffered saline (PBS), sonicated, and centrifuged at 1000 rpm for 10 min. After removal of the supernatant, the beads were resuspended in sterile PBS, and this washing procedure was repeated three times. The processed latex beads were then added to BV2 cells at a dilution ratio of 1:1000–1:2000 and incubated at 37 °C for 2–4 h. At the end of incubation, the culture medium and non-phagocytosed beads were removed, and the cells were washed three times with pre-cooled 1× PBS for 10 min each. The cells were then fixed with 4% paraformaldehyde at room temperature for 15 min, followed by three washes with PBS. Subsequently, the cells were incubated with Dil (Beyotime, C1036) for 10 min at room temperature in the dark, washed three times with PBS, and observed under a fluorescence microscope.

Western blotting

Protein expression in BV2 cells and mouse hippocampal tissues was analyzed by western blotting. Samples were lysed in RIPA buffer containing PMSF and phosphatase inhibitor cocktail (Beyotime, China), followed by sonication and centrifugation at 4 °C. Protein concentrations were determined using a BCA assay kit (Cat No.20020, Yeasen, Shanghai, China). Equal amounts of protein (20 µg) were separated by SDS–PAGE and transferred onto 0.22 μm PVDF membranes (Millipore, USA). After blocking with 5% nonfat milk for 2 h at room temperature. The membranes were then incubated overnight at 4 °C with the corresponding primary antibodies, including anti-CD68 (1:2000, HUABIO, HA722285), anti-α-syn (1:3000, HUABIO, HA751248), anti-phospho-FAK (1:500, HUABIO, ET1610-34), anti-FAK (1:5000, Immunoway, YM8154), anti-ubiquitin (1:4000, proteintech, 10201-2-AP), anti-LAMP1 (1:1000, HUABIO, HA722302), anti-HA (1:5000, proteintech,51064-2-AP), anti-LC3B (1:2000, abclonal, A19665), anti-HMGB1 (1:10000, proteintech, 10829-1-AP), anti-Flag (1:10000, proteintech, 66008-4-lg), anti-MYC (1:10000, proteintech, 10828-1-AP), and anti-β-actin (1:50000, HUABIO, HA750898), anti-ITGB5 (1:1000, HUABIO, ER63768). Following primary antibody incubation, the membranes were washed three times in washing buffer for 10 min each and then incubated with HRP-labeled goat anti-rabbit IgG secondary antibody (1:2000, Yamei, LF102) for 2 h at room temperature. Next, immunoreactive bands were visualized using an enhanced chemiluminescence substrate (APExBIO, K1231) and captured with a chemiluminescence imaging system (Tanon, China). Signal intensity was quantified using ImageJ software. Each assay was performed in three independent experiments.

CO-IP

Whole-cell lysates were prepared at 4 °C using IP lysis buffer (Beyotime, P0013J) containing 1.0% NP-40, 0.2 mM EDTA, 20 mM Tris-HCl (pH 8.0), and 180 mM NaCl, supplemented with protease inhibitor cocktail (APExBIO, K1007). The lysates were then centrifuged at 12,000 rpm for 10 min at 4 °C, and the supernatants were collected for endogenous or exogenous co-immunoprecipitation assays. For endogenous IP, the supernatants were incubated overnight at 4 °C with the indicated antibodies and Protein A/G Magnetic Beads (APExBIO, K1305). For exogenous IP, the supernatants were incubated overnight at 4 °C with anti-Flag or anti-HA, anti-MYC antibodies together with Protein A/G Magnetic Beads. The immune complexes were subsequently separated using a magnetic stand, denatured in SDS loading buffer, and then subjected to western blotting or mass spectrometry analysis.

Immunofluorescence

For brain immunofluorescence (IF), whole brains were collected from mice and post-fixed in 4% paraformaldehyde for 24 h, followed by cryoprotection in 30% sucrose for 48 h. Hippocampal Sect.  (20 μm) were prepared using a freezing microtome. Free-floating sections were rinsed three times in PBS for 5 min each and blocked with 5% bovine serum albumin for 2 h at room temperature. Sections were then incubated overnight at 4 °C with primary antibodies against α-syn, CD68, vinculin (Servicebio, GB150068), GFAP( Servicebio, GB15096), or Iba1(Servicebio,GB153502). After three washes in PBS, sections were incubated with the appropriate fluorophore-conjugated secondary antibodies. Nuclei were counterstained with DAPI (ABSLAB, R3074). BV2 cells were fixed with 4% paraformaldehyde and processed for staining with primary antibodies against Iba1, α-syn, CD68, or LAMP1 at 4 °C overnight. After washing three times with PBS, cells were incubated with the secondary antibodies and counterstained with DAPI. Fluorescence images were acquired using an immunofluorescence microscope (ZEISS LSM880). The fluorescence intensity was analyzed using ImageJ software.

Statistical analysis

All statistical analyses and figure preparation were performed using GraphPad Prism 8.0 and SPSS 20.0. Comparisons between two groups were conducted using a two-tailed Student’s t-test (paired or unpaired, as appropriate), whereas the Wilcoxon rank-sum test was applied when the data did not follow a normal distribution. Differences among multiple groups were analyzed by one-way analysis of variance (ANOVA), followed by Tukey’s post hoc multiple-comparison test. Correlation analysis was performed using Pearson’s correlation test, whereas Spearman’s rank correlation test was used when the data were not normally distributed. Statistical significance was defined as ns, p > 0.05; * p < 0.05; ** p < 0.01; and *** p < 0.001.

Results

Aerobic exercise elevates circulating irisin in PD patients and correlates with cognitive improvement

To assess the association between exercise-induced irisin and cognitive outcomes in PD, we included 21 patients with PD and measured circulating biomarkers and cognition before and after 8 weeks of aerobic exercise (Tables 1 and 2). Compared with pre-training, serum irisin levels were significantly increased, whereas serum α-syn levels were significantly decreased after training (Fig. 1A-B). In parallel, cognitive performance improved, as indicated by higher MMSE and MoCA scores relative to baseline (Fig. 1C-D). Correlation analyses further supported coordinated relationships among these measures: plasma α-syn was inversely associated with MMSE and MoCA, while circulating irisin showed a negative correlation with plasma α-syn and positive correlations with MMSE and MoCA (Fig. 1E-I).

Table 1.

Baseline characteristics of participants

Characteristic	Participants (n = 26)
Gender, n (%)
Male	14 (54)
Female	12 (46)
Age (y), mean (SD)	61 (9.5)
Height(cm), mean (SD)	164 (8.8)
Weight(kg), mean (SD)	68 (12)
BMI, mean (SD)	25 (3.5)
Disease duration(y), median (IQR)	5.0 (5.0)
H-Y (%)
1  1.5  2  2.5  3	5.0 (19) 8.0 (31) 9.0 (35) 2.0 (7.7) 2.0 (7.7)
Education years(y), median (IQR)	11 (4.7)
LEDD (mg), mean (SD	624 (309)

BMI Body mass index, H-Y Hoehn-Yahr staging, LEDD Levodopa daily dose

Table 2.

Cognitive assessment scores, and laboratory results of patients with Parkinson’s disease

Scales	Before Exercise	After Exercise	P-value
MDS-UPDRS1	10.67 ± 5.98	9.05 ± 5.30	0.021
MDS-UPDRS2	12.05 ± 4.50	10.67 ± 3.88	0.005
MDS-UPDRS3OFF	36.76 ± 15.97	32.48 ± 12.81	0.005
MDS-UPDRS3ON	28 ± 15.70	21.38 ± 13.52	0.001
MDS-UPDRS4	1.86 ± 3.40	1.81 ± 3.46	0.771
PSQI	8.05 ± 3.93	7.1 ± 4.7	0.236
α-syn (ng/ml)	134.4 ± 34.85	78.71 ± 18.03	<0.001
Irisin (ng/ml)	12.22 ± 2.372	32.41 ± 3.523	<0.001
MoCA	22.67 ± 4.33	25.05 ± 2.73	0.002
MMSE	27.27 ± 1.23	28.76 ± 1.24	0.007

MDS-UPDRS Movement Disorders Society Unified Parkinson’s Disease Rating Scale, MMSE Mini-Mental State Examination, MoCA Montreal Cognitive Assessment, PSQI Pittsburgh Sleep Quality Index

Fig. 1.

Aerobic exercise elevates serum irisin in PD patients and correlates with cognitive improvement. A Comparison of serum irisin levels in PD patients before and after aerobic exercise. B Comparison of serum α-syn levels in PD patients before and after aerobic exercise. C Comparison of MoCA scores in PD patients before and after aerobic exercise. D Comparison of MMSE scores in PD patients before and after aerobic exercise. E Correlation analysis between serum α-syn level and MMSE score in PD patients after aerobic exercise. F Correlation analysis between serum α-syn level and MoCA score in PD patients after aerobic exercise. G Correlation analysis between serum irisin level and MoCA score in PD patients after aerobic exercise. H Correlation analysis between serum irisin level and MMSE score in PD patients after aerobic exercise. I Correlation analysis between serum irisin level and serum α-syn level in PD patients after aerobic exercise. **P < 0.01, ****P < 0.0001

Exercise-derived irisin is critical for enhancing microglial α-syn clearance and alleviates PD-associated cognitive deficits

Since irisin is an exercise-inducible hormone, we sought to directly test whether irisin is required for the cognitive benefits of exercise in PD-CI. To this end, we administered an irisin-neutralizing antibody during the exercise training paradigm to specifically block irisin signaling (Fig. 2A). In the open-field test, irisin neutralization reversed the exercise-induced increases in total distance traveled and the number of entries into the center zone (Fig. 2B-C). Moreover, behavioral analyses showed that irisin neutralization significantly weakened the exercise-associated improvement in cognitive performance. In the Morris water maze, compared with exercise plus control IgG, mice receiving the irisin-neutralizing antibody displayed longer escape latency during training and fewer platform crossings in the probe trial (Fig. 2D-E). In addition, in the novel object recognition task, the PD + Ex+NE-Irisin group exhibited a lower recognition index (Fig. 2F). Together, these results indicate that exercise-derived irisin is required for the amelioration of PD-associated cognitive deficits. α-syn accumulation is a key pathological contributor to PD-associated cognitive deficits, and glial-mediated clearance is considered a major route for limiting α-syn burden. We therefore asked whether exercise-derived irisin modulates α-syn accumulation in PD-CI. Double-immunofluorescence staining showed that α-syn colocalized predominantly with microglia, but not with astrocytes, indicating that microglia are the primary glial population involved in α-syn handling in our model (Fig. 2G-J). To further determine whether microglia are required for the cognitive benefits of exercise in PD-CI, we performed a microglial depletion experiment using PLX treatment. Effective depletion of microglia abolished the exercise-associated improvement in cognitive performance. Specifically, in PLX-treated mice, exercise training no longer improved spatial learning and memory in the Morris water maze or recognition memory in the novel object recognition task, indicating a loss of exercise-induced cognitive rescue (Figure S1A-C). In addition, following PLX intervention, immunofluorescence analyses revealed a marked reduction in CD68 and Iba1 co-localization, indicating diminished microglial phagocytic activation (Figure S1D-E). Consistently, the co-localization of α-syn and Iba1, and the protein of α-syn was also significantly reduced, suggesting impaired microglial engagement with α-syn after microglial depletion (Figure S1F-I). In line with this observation, exercise training reduced α-syn levels, whereas treatment with an irisin-neutralizing antibody abrogated the exercise-induced reduction (Fig. 2K-L). We further investigated whether irisin enhances microglial clearance of α-syn using BV2 microglial cells as an in vitro model (Figure S2A). Transmission electron microscopy verified the successful formation of fibrillar α-syn (Figure S2B). Notably, 20 nM irisin was sufficient to effectively suppress α-syn accumulation in BV2 microglia (Figure S2C-F). Immunofluorescence staining showed that irisin treatment markedly reduced intracellular α-syn signals in BV2 cells, accompanied by increased CD68 activity (Fig. 2M-P). Western blot analysis also confirmed a significant decrease in α-syn protein and increased CD68 levels following irisin exposure (Figure S2G-H). Together, these data indicated exercise-derived irisin is required for the suppression of α-syn accumulation and implicating microglial clearance as a principal downstream effector mechanism.

Fig. 2.

Irisin is required for microglial α-syn clearance and alleviates PD-associated cognitive deficits. A Schematic diagram of the experimental design, time points and procedures (n = 6). B, C Representative images of the moving paths of mice in each group in the open-field test and relative quantitative analysis of entries into the center zone and total distance traveled. D, E Representative images of the swimming paths of mice in each group in the Morris water maze test and relative quantitative analysis of escape latency during training, time spent in the target quadrant and platform crossing frequency in the probe trial. F Schematic diagram of the novel object recognition test and relative quantitative analysis of the recognition index for the familiar object and the novel object. G, H Representative immunofluorescence images and relative quantitative analysis showing α-syn colocalized with Iba1-positive microglia in the hippocampus from mice in each group(n = 3). Scale bar, 10 μm. I, J Representative immunofluorescence images and relative quantitative analysis showing α-syn colocalized with GFAP-positive astrocytes in brain sections from mice in each group. Scale bar, 10 μm. K, L Western blot showing α-syn expression in the hippocampus in each group and relative quantitative analysis. M, N Relative quantitative analysis and representative immunofluorescence images showing α-syn within Iba1-positive BV2 cells under different treatments (n = 10). Scale bar,10 μm. O, P Relative quantitative analysis and representative immunofluorescence images showing CD68 within Iba1-positive BV2 cells under different treatments. Scale bar,10 μm. All data are presented as mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. ***P < 0.001, ****P < 0.0001

Integrin αV/β5 mediates the irisin-driven clearance of α-syn and cognitive improvement

To determine the mechanism underlying the clearance-promoting effect of irisin, we next focused on integrin αV/β5. We pharmacologically inhibited αV/β5 signaling using cilengitide during exogenous irisin administration. In the open-field test, cilengitide treatment reversed the irisin-induced increases in exploratory locomotion, including total distance traveled and center entries (Fig. 3A-B). In the Morris water maze, the irisin-driven improvement in spatial learning and memory was abolished by cilengitide, as evidenced by prolonged escape latency during training and reduced platform crossings in the probe trial (Fig. 3C). Similarly, in the novel object recognition task, cilengitide prevented the irisin-induced enhancement of recognition memory, resulting in a reduced recognition index compared with the irisin-treated group (Fig. 3D). Immunofluorescence analyses in the hippocampus further indicated that αV/β5 inhibition compromised microglial phagocytic activation and engagement with α-syn. Specifically, cilengitide treatment reduced CD68-Iba1 colocalization, consistent with diminished microglial phagolysosomal activity, and decreased α-syn-Iba1 colocalization (Fig. 3E-H). Western blotting showed that cilengitide increased α-syn levels while decreasing CD68 expression (Fig. 3I-J). Similar results were observed in vitro. In ITGB5-KD microglia, immunofluorescence staining demonstrated increased intracellular α-syn accumulation accompanied by reduced CD68 signals (Fig. 3K-N). Functionally, ITGB5 knockdown also significantly attenuated the ability of irisin to enhance phagocytosis, as evidenced by reduced latex bead uptake in response to irisin stimulation (Fig. 3O-P). Western blot analyses further confirmed that ITGB5 knockdown elevated α-syn levels while decreasing CD68 expression (Fig. 3Q-R). Similarly, in primary microglia, immunofluorescence analyses showed that ITGB5 knockdown markedly attenuated the irisin-induced reduction of α-syn and suppressed phagocytic activity. Consistent with these findings, western blot analysis revealed that compared to the PFFs + irisin group, ITGB5 deficiency increased α-syn levels and reduced CD68 expression (Figure S3A–B). Therefore, these findings indicate that integrin αV/β5 is required for irisin to activate microglial phagocytic and clearance programs and to restrain α-syn accumulation.

Fig. 3.

Integrin αV/β5 mediates irisin-driven α-syn clearance and cognitive improvement. A, B Representative images of the moving paths of mice in each group in the open-field test and relative quantitative analysis of entries into the center zone and total distance traveled (n = 6). C Representative images of the swimming paths of mice in each group in the Morris water maze test and relative quantitative analysis of escape latency during training, time spent in the target quadrant and platform crossing frequency in the probe trial. D Relative quantitative analysis of the recognition index for the familiar object and the novel object in each group in the novel object recognition test. E, F Representative immunofluorescence images and relative quantitative analysis showing CD68 signal within Iba1-positive microglia in the hippocampus from mice in each group. Enlarged views of boxed regions are shown. G, H Representative immunofluorescence images and relative quantitative analysis showing α-syn colocalized with Iba1-positive microglia in the hippocampus in each group. Scale bar,10 μm. I, J Western blot showing ITGB5, α-syn and CD68 expression in brain tissues from mice in each group and relative quantitative analysis (n = 3). K, L Representative immunofluorescence images and relative quantitative analysis showing α-syn colocalized with Iba1-positive BV2 cells under different treatments. M, N Representative immunofluorescence images and relative quantitative analysis showing CD68 staining in Iba1-positive BV2 cells under different treatments. Scale bar,10 μm. O, P Representative images and relative quantitative analysis of phagocytosis assay using latex beads in BV2 cells under different treatments (n = 10). Scale bar,10 μm. Q, R Western blot showing ITGB5, α-syn and CD68 expression in BV2 cells transfected with si-ITGB5 under different treatments and relative quantitative analysis (n = 3). All data are presented as mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. ***P < 0.001, ****P < 0.0001

FAK-dependent adhesion is required for irisin-integrin αV/β5 axis-driven microglial phagocytosis of α-syn

To elucidate the mechanisms by which irisin–integrin αV/β5 activation promotes microglial α-syn clearance, we performed RNA-seq in BV2 cells exposed to α-syn and irisin, comparing ITGB5-deficient and control conditions (Fig. 4A). The volcano plot displayed both upregulated and downregulated genes (Fig. 4B). Notably, Gene set enrichment analysis (GSEA) revealed a pronounced suppression of cell adhesion-related programs following ITGB5 knockdown, indicating that integrin αV/β5 is required to sustain an adhesion signaling state in response to irisin (Fig. 4C). Given that adhesion signaling governs cytoskeletal remodeling and provides the mechanical basis for microglial motility and target engagement, these processes that are prerequisite for efficient phagocytosis, thus, we next tested whether ITGB5 knockdown suppressed microglial migration. Accordingly, scratch-wound and Transwell assays were performed, and both assays demonstrated that knockdown of ITGB5 attenuated microglial migratory capacity, consistent with the transcriptomic signature of impaired adhesion signaling (Fig. 4D-G). In the hippocampus of PD mice treated with irisin, we observed increased colocalization of vinculin with Iba1, indicating enhanced focal-adhesion signaling in microglia. In contrast, cilengitide treatment markedly reduced vinculin-Iba1 colocalization (Fig. 4H-I). Together, these results indicated that irisin–integrin αV/β5 promotes a microglia adhesion. Adhesion signaling downstream of integrins is largely coordinated by focal adhesion kinase (FAK). We therefore conducted a series of functional experiments to determine how FAK contributes to irisin-integrin αV/β5-driven microglial clearance of α-syn. As shown in (Fig. 5A-D), FAK knockdown significantly impaired the pro-migratory effect observed in the Irisin + PFFs condition, indicating that FAK is required for microglial motility in this context. Moreover, immunofluorescence analyses revealed that FAK knockdown led to increased intracellular α-syn accumulation and reduced CD68 signals in BV2 cells, suggesting attenuation of the phagolysosomal clearance (Fig. 5E-H). Functionally, FAK knockdown also diminished irisin-enhanced phagocytosis, as evidenced by reduced latex bead uptake upon irisin stimulation (Fig. 5I, J). Consistent with these findings, western blotting confirmed that, compared with the PFFs + irisin group, FAK knockdown reduced CD68 expression but did not significantly affect α-syn levels (Fig. 5K-L). Similarly, we also observed that FAK knockdown reduced CD68 expression without significantly altering α-syn levels in primary microglia (Figure S4A-C). However, more intriguingly, compared with the ITGB5-knockdown + PFFs group, combined ITGB5 knockdown and FAK overexpression increased microglial uptake of α-syn, yet the internalized α-syn accumulated within the cells rather than being efficiently cleared (Figure S4D-K). This dissociation between enhanced uptake and increased burden indicates that the ITGB5-FAK signal primarily governs the phagocytic entry step of α-syn. Further, we pharmacologically inhibited FAK in vivo to determine whether FAK is required for the irisin-mediated cognitive benefits in PD mice. As shown, FAK inhibition abolished the behavioral improvements conferred by exogenous irisin, reversing its effects in the open-field test, Morris water maze, and novel object recognition assays (Fig. 6A-F). Immunofluorescence analyses of the hippocampus showed a marked reduction in CD68-Iba1 colocalization after inhibitor treatment (Fig. 6G-H). In parallel, α-syn-Iba1 colocalization was also decreased (Fig. 6I-J). Western blotting revealed increased α-syn levels together with reduced CD68 expression following FAK inhibitor administration (Fig. 6K-L). In addition, cilengitide treatment reduced downstream FAK activation in the hippocampus, as evidenced by decreased level of phosphorylated-FAK (Fig. 6M-N), further supporting that αV/β5 blockade suppresses the αV/β5-FAK signaling cascade in vivo. These results indicate that integrin αV/β5-FAK signaling is necessary for irisin to ameliorate α-syn accumulation and PD-associated cognitive deficits.

Fig. 4.

Transcriptomic profiling identifies adhesion-related molecules. A Schematic diagram of the RNA-sequencing workflow (n = 3). B Volcano plot showing differentially expressed genes in irisin and α-syn treated samples compared with the ITGB5 knockdown group. C Gene set enrichment analysis showing enrichment of the positive regulation of cell adhesion mediated by integrin pathway. D, E Representative images of wound-healing assay and relative quantitative analysis of migration rate in BV2 cells under different treatments at 0 h and 24 h. F, G Representative images of Transwell migration assay and relative quantitative analysis of the number of migrated BV2 cells under different treatments at 24 h (n = 3). Scale bar, 50 μm. H, I Representative immunofluorescence images and relative quantitative analysis showing vinculin colocalized with Iba1-positive microglia in the hippocampus in each group (n = 3). Enlarged views of boxed regions are shown. Scale bar, 10 μm. All data are presented as mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 5.

Irisin promotes microglial adhesion and enhances α-syn phagocytosis through FAK signaling. A, B Representative images of wound-healing assay and relative quantitative analysis of migration rate in BV2 cells under different treatments at 0 h and 24 h (n = 3). C, D Relative quantitative analysis of the number of migrated cells and representative images of Transwell migration assay in BV2 cells under different treatments at 24 h. Scale bar, 50 μm. E, F Representative immunofluorescence images and relative quantitative analysis showing CD68 signal in BV2 cells under different treatments. G, H Relative quantitative analysis and representative immunofluorescence images showing α-syn signal in Iba1-positive BV2 cells under different treatments. I, J Representative images and relative quantitative analysis of phagocytosis assay using latex beads in BV2 cells under different treatments. K, L Western blot showing α-syn, CD68 and FAK expression in BV2 cells transfected with siFAK under different treatments and relative quantitative analysis. All data are presented as mean ± SEM (n = 3). Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. ***P < 0.001, ****P < 0.0001

Fig. 6.

Irisin enhances microglial α-syn phagocytosis in MPTP mice through FAK signaling. A, B Representative images of the moving paths of mice in each group in the open-field test and relative quantitative analysis of total distance traveled and entries into the center zone (n = 6). C, D Representative images of the swimming paths of mice in each group in the Morris water maze test and relative quantitative analysis of escape latency during training, platform crossing frequency and time spent in the target quadrant in the probe trial. E, F Relative quantitative analysis of the recognition index for the familiar object and the novel object in each group in the novel object recognition test. G, H Representative immunofluorescence images and relative quantitative analysis showing CD68 signal within Iba1-positive microglia in the hippocampus in each group (n = 3). I, J Representative immunofluorescence images and relative quantitative analysis showing α-syn colocalized with Iba1-positive microglia in the hippocampus in each group. Enlarged views of boxed regions are shown (n = 3). Scale bar, 10 μm. K, L Western blot and relative quantitative analysis showing FAK, α-syn and CD68 expression in the hippocampus of MPTP treated mice with irisin in the presence or absence FAK inhibition (n = 3). M, N Western blot and relative quantitative analysis showing FAK and phosphorylated FAK expression in the hippocampus of MPTP treated mice receiving treatment with irisin in the presence or absence of cilengitide (n = 3). All data are presented as mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. ns, not significant. ***P < 0.001, ****P < 0.0001

ITGB5 interacts with HMGB1 to facilitate degradative clearance of α-syn

Above data indicate that the ITGB5-FAK axis primarily regulates the phagocytic entry step of α-syn; however, the mechanism by which irisin promotes α-syn clearance beyond uptake remained unclear. To identify pathways potentially responsible for post-phagocytic clearance, RNA-seq found that lysosome pathway was prominently enriched, implicating the autophagy-lysosome system as a key downstream component of irisin-αV/β5 (Fig. 7A). Based on this, we hypothesized that activation of integrin αV/β5 by irisin may enable ITGB5 to recruit molecules functionally linked to autophagy-lysosome. To test this hypothesis, we conducted IP-MS to identify ITGB5-associated protein complexes in microglia (Fig. 7B). Among the proteins identified, HMGB1 emerged as an autophagy related ITGB5 interacting candidate, and its identity was further supported by the corresponding MS/MS spectrum (Fig. 7C). CO-immunoprecipitation (CO-IP) analyses confirmed a physical interaction between endogenous HMGB1 and ITGB5, as well as between overexpressed Flag-tagged ITGB5 and Flag-tagged HMGB1 (Fig. 7D-G). In addition, HMGB1 knockdown (Fig. 7H-K) or ITGB5 knockdown (Figure S5A-C) significantly suppressed the expression of LC3-II and LAMP1, accompanied by an increase in α-syn levels, indicating impaired autophagy–lysosome activity and reduced degradative capacity. Conversely, compared with the PFFs + irisin + siITGB5 + FAK overexpression group, HMGB1 overexpression markedly increased LC3-II and LAMP1 expression and promoted α-syn clearance (Fig. 7L-O). In primary microglia, both western blot and immunofluorescence demonstrated that, compared with the PFFs + irisin + siITGB5 group, HMGB1 overexpression markedly increased the expression of LC3-II and LAMP1, while reducing α-syn levels (Figure S5D-F). Together, these results further indicate a functional segregation downstream of irisin–integrin αV/β5 signaling: FAK mainly drives the uptake step, whereas HMGB1 supports the subsequent degradative clearance, by enhancing autophagy-lysosome to facilitate efficient disposal of internalized α-syn.

Fig. 7.

Irisin regulates the interaction between ITGB5 and HMGB1 to promote lysosomal degradation of α-syn. A KEGG pathway enrichment analysis of differentially expressed genes identified from transcriptomic profiling. B Schematic diagram of the experimental workflow for proteomic analysis, including sample collection and LC-MS/MS analysis. C Representative mass spectrometry spectrum showing HMGB1 peptide fragments identified in proteomic analysis. D, E Co-immunoprecipitation assays showing the interaction between ITGB5 and HMGB1 in 293T cells transfected with ITGB5-Flag or HMGB1-Flag. F, G Immunoblot analysis showing endogenous interaction between ITGB5 and HMGB1 in BV2 cells. H, I Western blot analysis and relative quantitative analysis of LAMP1, LC3II and α-syn expression in BV2 cells treated with PFFs and irisin in the presence or absence of HMGB1 knockdown. J, K Representative immunofluorescence images and relative quantitative analysis showing α-syn and LAMP1 mean fluorescence intensity under different treatments. L, M Western blot analysis and relative quantitative analysis of LAMP1, α-syn and LC3II expression in BV2 cells with ITGB5 knockdown, FAK overexpression or HMGB1 overexpression under irisin and PFFs treatment. N, O Representative immunofluorescence images and relative quantitative analysis showing α-syn and LAMP1 mean fluorescence intensity in primary microglia treated with LAMP1 and ITGB5 knockdown, FAK overexpression or HMGB1 overexpression under irisin and PFFs treatment. Scale bars, as indicated in the panels. All data are presented as mean ± SEM (n = 3). Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. ns, not significant. ***P < 0.001, ****P < 0.0001

ITGB5 inhibits K48-linked deubiquitination of HMGB1 at K114 and prevents proteasomal degradation

Given that HMGB1 interaction with ITGB5, we next assessed whether ITGB5 regulates HMGB1 expression and/or stability. Knockdown of ITGB5 did not alter HMGB1 mRNA levels (Fig. 8A), but markedly reduced HMGB1 protein abundance (Fig. 8B-C), indicating that ITGB5 regulates HMGB1 primarily at the post-transcriptional level. To directly assess protein stability, we performed cycloheximide (CHX) chase assays and found that ITGB5 silencing significantly accelerated HMGB1 decay, supporting a role for ITGB5 in stabilizing HMGB1 protein (Fig. 8D-E). In addition, treatment with the proteasome inhibitor MG132, but not chloroquine (CQ), rescued the reduction of HMGB1 protein caused by ITGB5 deficiency, indicating that ITGB5 primarily protects HMGB1 from proteasome-dependent degradation (Fig. 8F-G). Consistent with this, IP assays demonstrated that ITGB5 overexpression reduced overall HMGB1 ubiquitination (Fig. 8H-I). Next, we examined ubiquitin linkage specificity. Our data indicate that ITGB5 suppresses K48-linked ubiquitination of HMGB1 (Fig. 8J); Notably, when a K48R ubiquitin mutant was used, ITGB5 no longer reduced HMGB1 ubiquitination, indicating that the inhibitory effect of ITGB5 is specific to K48-linked ubiquitin chains (Fig. 8K). The conserved HMGB1 sites from lysine (K) to arginine (R) were mutated to exclude ubiquitylation in BV2 cells. IP analyses showed that only the K114R mutation disrupted the interaction between ITGB5 and HMGB1. Moreover, CHX chase assays revealed that the K114R mutation prolonged the half-life of HMGB1, indicating enhanced protein stability (Fig. 8L-O). These data indicate that ITGB5 stabilizes HMGB1 expression by limiting K48-linked ubiquitination K48-linked deubiquitination of HMGB1 at K114.

Fig. 8.

ITGB5 stabilizes HMGB1 by inhibiting K48-linked ubiquitination at K114 and limiting proteasomal degradation. A Relative mRNA expression of HMGB1 and ITGB5 in BV2 cells under the indicated treatments (n = 3). B, C Western blot and relative quantitative analysis of HMGB1 and ITGB5 protein expression in BV2 cells with ITGB5 knockdown or HA-ITGB5 overexpression under the indicated treatments (n = 3). D, E CHX chase assays were performed to detect the protein stability of HMGB1 in BV2 cells with ITGB5 knockdown under PFFs and Irisin treatment, and the relative HMGB1 expression at the indicated time points was quantified. F, G The protein expression of HMGB1 and ITGB5 was measured by western blot and quantified in BV2 cells with ITGB5 knockdown combined with MG132 or CQ treatment under PFFs and Irisin treatment (n = 3). H The ubiquitination of endogenous HMGB1 in the presence or absence of Flag-ITGB5 was analyzed by Co-IP assay with MG132 treatment. I MYC-HMGB1, Flag-ITGB5 and HA-Ub were transfected into 293T cells, followed by Co-IP assay to detect the ubiquitination of exogenous HMGB1 in the presence or absence of ITGB5 with MG132 treatment. J MYC-HMGB1, Flag-ITGB5 and HA-tagged ubiquitin mutants were co-transfected into 293T cells, followed by Co-IP assay to determine the ubiquitin linkage type involved in HMGB1 ubiquitination under MG132 treatment. K HA-Ub (K48R), MYC-HMGB1 and Flag-ITGB5 were co-transfected into 293T cells, followed by Co-IP assay to examine the effect of ITGB5 on HMGB1 ubiquitination in the presence of K48R ubiquitin mutant under MG132 treatment. L Wild-type HMGB1 and lysine-mutant HMGB1 constructs were transfected into cells, and the expression of each HMGB1 mutant was measured by western blot. M, N CHX chase assays were performed to compare the protein stability of HMGB1-WT and HMGB1-K114R, and the relative HMGB1 expression at the indicated time points was quantified. O Flag-HMGB1-WT or Flag-HMGB1-K114R was co-transfected with HA-Ub, followed by Co-IP assay to examine the ubiquitination level of HMGB1-WT and HMGB1-K114R. All data are presented as mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Tukey’s post hoc test. ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Discussion

PD-CI is common, clinically heterogeneous, and mechanistically multifactorial, with synaptic dysfunction, α-syn pathology, and neuroimmune dysregulation increasingly viewed as convergent drivers of decline [24, 25]. In this context, our work contributes two linked advances. First, our clinical data show that exercise training increases circulating irisin and improves cognitive performance in patients with PD, which consistent with previous meta-analysis results that physical exercise yields modest gains in global cognition in PD [26]. Second, we define a receptor-driven mechanism: irisin through integrin αV/β5 to promote α-syn engulfment via a FAK-dependent program while simultaneously facilitating lysosomal degradation through regulating HMGB1 stabilization, thereby enhancing microglial clearance of pathological protein rather than uptake alone.

Most PD-focused studies of irisin have emphasized neuronal bioenergetics and mitochondrial protection as the dominant mechanisms [27, 28]. For example, exercise increased serum irisin and linked irisin to neuroprotection through mitigation of mitochondrial damage in PD models [29]. More recent reviews have similarly emphasized mitochondria-centered and broadly cytoprotective pathways, including reduction of oxidative stress, engagement of neurotrophic signaling, and generalized anti-inflammatory effects [30, 31]. Our findings are aligned with these observations but extend the mechanistic landscape. Our findings suggested that irisin could operate through an additional route, enhancing microglial proteostasis to reduce α-syn burden. Moreover, although integrin αV/β5 is the primary receptor for irisin [32]; however, the downstream signal transduction involved in receptor activation in the central nervous system, particularly in microglia and in the context of proteinopathy, have remained insufficiently defined. Our data extends this receptor biology into an α-syn pathology framework, proposing that integrin αV/β5 is not a passive binder but an actionable target that can be leveraged to pathological protein clearance efficiency in disease-relevant contexts.

A consistent insight from α-syn biology is that uptake and clearance are not equivalent. Microglia can ingest extracellular α-syn [33]; however, when post-endocytic trafficking and autophagy–lysosome delivery are inefficient, internalized α-syn can accumulate, and amplify inflammatory signaling that may ultimately exacerbate pathology [34]. In this regard, the concept of microglial “synucleinphagy”, selective autophagic degradation of neuron-released α-syn, showed that effective clearance requires coordinated engagement of degradation machinery, not merely phagocytosis [13, 35, 36]. Our data support this distinction by showing that irisin strengthens the coupling between engulfment and degradation. This is important for cognition because cortical and hippocampal circuits are vulnerable to chronic pathological protein stress [37]; interventions that increase uptake without ensuring disposal risk converting microglia into reservoirs of α-syn rather than resolvers of it.

Integrin-dependent phagocytosis is fundamentally driven by adhesion-dependent signaling and actin remodeling, with FAK serving as a canonical integrin effector that coordinates the cytoskeletal dynamics required for engulfment [38]. Although much of the foundational αV/β5–FAK phagocytosis literature was developed outside microglia (e.g., retinal pigment epithelium) [39, 40], the signaling logic is conserved: αV/β5 promotes FAK activation and provides the mechanical and signaling prerequisites for efficient internalization. Consistent with this, our transcriptomic profiling shows that genetic ablation of αV/β5 is accompanied by a broad attenuation of adhesion-associated signaling, supporting the notion that αV/β5 sits upstream of an adhesion-cytoskeleton module required for phagocytic competence. Pharmacological inhibition of FAK in both in vitro and in vivo revealed that markedly reduced microglial uptake of α-syn, directly linking the αV/β5–FAK axis to the phagocytic. By positioning FAK downstream of integrin αV/β5, we align the biology of irisin action with a well-defined, canonical phagocytic mechanism. Moreover, this emphasis on cytoskeletal control is cognitively relevant, as recent study demonstrates that perturbations in microglial cytoskeletal regulation can drive synaptic abnormalities and cognitive deficits, and that restoring actin-dependent control of microglial phagocytosis can rescue synapse function and memory performance[41, 42].

HMGB1 is a well-established regulator of autophagy, in part through its ability to interact with Beclin 1 and modulate autophagosome formation and autophagic flux under stress. Endogenous HMGB1 has been reported to facilitate autophagy-dependent degradation of wild-type and mutant α-syn in an ATG5-dependent manner [43]. Conversely, α-syn overexpression can perturb HMGB1-associated autophagic competence, highlighting bidirectional crosstalk between α-syn burden and HMGB1-regulated proteostasis [42]. Notably, HMGB1 biology is highly context- and compartment-dependent, whereas stress-induced intracellular/cytosolic HMGB1 can participate in autophagy regulation by interacting with autophagy-related proteins, including Beclin1 [44, 45]. In contrast, extracellular HMGB1, released passively from damaged cells or actively secreted by activated immune cells, can act as a damage-associated molecular pattern and promote inflammatory signaling through receptors such as TLR4 and RAGE [46, 47]. However, The HMGB1-stabilizing effect observed in our study should be interpreted primarily as an intracellular mechanism that supports autophagy-lysosome-mediated degradation of internalized α-syn in microglia, rather than as evidence for enhanced extracellular HMGB1 signaling. Our results refine HMGB1’s role by placing it within a αV/β5 axis and by emphasizing functional outcome, enhanced coupling between phagocytosis and clearance. Using IP-MS, we identify HMGB1 as an αV/β5-interacting protein, and further show that irisin activation of αV/β5 enhances the αV/β5–HMGB1 association. Functionally, irisin suppresses K48-linked ubiquitination of HMGB1, stabilizing HMGB1 protein levels and thereby promoting downstream autophagy–lysosome activity, accelerating α-syn degradation and improving cognitive performance in PD models. This indicates that does not simply “turn up” autophagy; instead, it establishes an αV/β5–HMGB1 regulatory state that improves the degradation of internalized α-syn, mechanistically distinct from the αV/β5-FAK arm that governs phagocytic entry.

Several limitations should be considered. First, the observed association between circulating irisin levels and cognitive improvement in our PD cohort is correlational and therefore does not establish causality. Second, while our data support integrin αV/β5 as a key receptor node and implicate FAK and HMGB1 in distinct route, receptor or pathway specificity would be strengthened by specific genetic mice approaches to exclude contributions from other αV-class integrins or non-microglial cell types. Third, regulates HMGB1 ubiquitination, particularly the E3 ubiquitin ligases or deubiquitinases responsible for K48-linked ubiquitination remains undefined in our study. Fourth, our mechanistic conclusions are based primarily on murine models and mouse microglia cells; therefore, further validation in human iPSC-derived microglia would strengthen the conclusions and provide more direct support for the irisin–integrin αV/β5–HMGB1 axis in a human cellular context.

Fifth, PLX depletion experiments cannot fully exclude contributions from peripheral macrophages or infiltrating myeloid cells. Because PLX compounds systemically inhibit CSF1R, they may affect CNS-resident microglia as well as peripheral monocytes, tissue-resident macrophages, border-associated macrophages, and other CSF1R-dependent myeloid populations. Thus, our results should be interpreted as supporting a major microglia-dependent contribution of the irisin–integrin αV/β5 pathway to α-syn clearance. Future studies using microglia-specific genetic models, or flow cytometric separation of resident microglia and infiltrating macrophages are needed to clarify their relative roles. Finally, our in vivo experiments were mainly performed in the MPTP model, which primarily reflects dopaminergic neurotoxicity rather than progressive α-synucleinopathy. Therefore, our findings should be interpreted as evidence that irisin reduces α-syn burden and enhances microglial clearance. Nevertheless, MPTP can affect α-syn dynamics through mitochondrial complex I inhibition, oxidative stress, neuroinflammation, and disruption of protein degradation pathways, which may collectively impair α-syn homeostasis [48]. Future studies using α-syn preformed fibril or AAV-α-syn models will be required to determine whether the irisin-integrin αV/β5 pathway also suppresses seeded α-syn aggregation, intercellular propagation, and Lewy-like pathology [49].

Conclusions

In summary, our study demonstrates that irisin couples α-syn phagocytosis and clearance through integrin αV/β5 by engaging two coordinated pathways: an αV/β5–FAK axis that promotes the phagocytosis of α-syn and an αV/β5–HMGB1 axis that facilitates degradation. This integrated framework provides a mechanistic basis linking exercise-elevated irisin to reduced α-syn burden and supports targeting the irisin-integrin αV/β5 axis as a potential strategy to alleviate PD-CI (Fig. 9).

Fig. 9.

Mechanism flowchart

Abbreviations

PD

Parkinson’s disease

α-syn

α-synuclein

PD-CI

Parkinson’s disease-associated cognitive impairment

CHX

cycloheximide

CO-IP

Co-immunoprecipitation

CQ

chloroquine

FAK

focal adhesion kinase

GSEA

Gene set enrichment analysis

IF

immunofluorescence

ANOVA

one-way analysis of variance

DEGs

differentially expressed genes

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

MMSE‌

‌Mini-Mental State Examination

MoCA

Montreal Cognitive Assessment

Authors’ contributions

YHL: Writing – review & editing, Validation, Methodology, Investigation, Formal analysis, Conceptualization. WW: Writing – review & editing, Methodology, Investigation, Formal analysis, Data curation, Funding acquisition. MXW: Writing – review & editing, Validation, Methodology. JX: Validation, Supervision, Resources, Funding acquisition, Project administration. JYZ: Methodology, Investigation, Formal analysis, Data curation. SSJ: Methodology, Investigation, Formal analysis. MYH: Methodology, Investigation, Formal analysis. TTZ: Methodology, Investigation.

Data availability

The data supporting the conclusions of this article are included within the article and its additional files. The data are available from the corresponding authors on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.