PINK1 deficiency permits the development of Lewy body dementia with coexisting Aβ pathology

Abstract

INTRODUCTION

Dementia with Lewy bodies (DLB), a prevalent neurodegenerative dementia, involves α‐synuclein (α‐syn) aggregates and frequent amyloid beta (Aβ) co‐pathology, but mechanistic drivers remain unclear.

METHODS

We crossed pink1 knockout with APP/PS1 mice, and assessed behavioral and pathological phenotypes of the resulting animals. We also performed biochemical and biophysical characterizations of PTEN‐induced kinase 1 (PINK1) phosphorylation of α‐syn.

RESULTS

DLB brains show PINK1 deficiency alongside α‐syn and Aβ co‐pathology. Mirroring human DLB patients, APP/PS1::pink1‐/‐ mice spontaneously develop Lewy pathology at endogenous α‐syn levels, affecting both central and peripheral nervous systems with heterogeneous phenotypes. Mechanistically, PINK1 phosphorylates α‐syn at Thr44, suppressing Aβ‐induced α‐syn aggregation. Moreover, pT44‐α‐syn levels are correlated with PINK1 expression and activity in human brains.

DISCUSSION

PINK1 deficiency synergizes with Aβ to promote Lewy pathology via loss of protective α‐syn phosphorylation. The APP/PS1::pink1‐/‐ model recapitulates key DLB features without α‐syn overexpression, offering a valuable tool for future mechanistic and therapeutic studies.

Highlights

• PTEN‐induced kinase 1 (PINK1) deficiency, either through reduced expression or impaired activity, is found in human dementia with Lewy bodies (DLB) patients with amyloid beta (Aβ) co‐pathology.

• PINK1 specifically phosphorylates α‐synuclein at Thr44, inhibiting Aβ‐induced aggregation and preventing the development of Lewy pathology.

• The APP/PS1::pink1‐/‐ mouse model recapitulates key features of human DLB, exhibiting widespread Lewy pathology and heterogeneous phenotypes.

• PINK1 alterations emerge as a novel genetic risk factor for DLB, opening new avenues for diagnosis and therapeutic intervention.

1. INTRODUCTION

Dementia with Lewy bodies (DLB) is the second most common neurodegenerative dementia after Alzheimer's disease (AD). ¹ , ² It presents with a wide range of clinical manifestations, including fluctuating cognition, visual hallucinations, and motor symptoms similar to Parkinson's disease (PD). ³ , ⁴ , ⁵ A prodromal phase marked by emotional, behavioral, and cognitive changes often precedes the diagnosis of DLB by years. ⁶ As the disease progresses, symptom complexity increases, further complicating clinical management. This heterogeneity, combined with a limited understanding of underlying mechanisms, hinders effective treatment.

The neuropathological hallmark of DLB is Lewy pathology—widespread presence of Lewy bodies (LBs) and Lewy neurites (LNs), with α‐synuclein (α‐syn) aggregates as the key component. ⁷ , ⁸ , ⁹ Lewy pathology predominantly affects the cortex and hippocampus, regions overlapping with amyloid beta (Aβ) pathology observed in AD. ³ , ¹⁰ , ¹¹ , ¹² This contrasts with PD, in which Lewy pathology is largely confined to dopaminergic neurons of the substantia nigra. ¹³ , ¹⁴ The broader distribution in DLB likely contributes to its distinct clinical presentations.

A key neuropathological feature differentiating DLB from PD is the frequent co‐occurrence of Aβ plaques with α‐syn aggregates, ¹⁵ , ¹⁶ , ¹⁷ seen in more than half of the DLB cases. ¹⁵ , ¹⁸ Evidences suggest Aβ actively promotes cortical α‐syn aggregation and spread. For instance, injecting pre‐formed α‐syn fibrils into transgenic AD mice potentiates LB formation within the brain. ¹⁹ Conversely, reducing α‐syn expression in the mouse model mitigates Aβ‐induced neurodegeneration. ²⁰ These findings strongly suggest a synergistic interaction between Aβ and α‐syn in DLB pathogenesis.

Animal models of Lewy pathology include α‐syn overexpression, ²¹ , ²² , ²³ α‐syn point mutations, ²⁴ , ²⁵ and injection of pre‐formed α‐syn fibrils. ¹⁹ , ²² , ²⁶ , ²⁷ The overexpression model leads to α‐syn aggregation and causes damages to dopaminergic neurons in the substantia nigra, ²¹ , ²³ , ²⁴ , ²⁵ neurons especially vulnerable to mitochondrial stresses and oxidative damages. ²⁸ , ²⁹ Fibril injection models, especially in the hippocampus, mimic the spread of Lewy pathology and address transmissibility. ¹⁹ Yet, pathological outcomes vary depending on fibril source, implying strain differences in α‐syn. ²⁷ Although informative, these models rely on artificial overexpression or external seeding and do not replicate spontaneous α‐syn aggregation from endogenous proteins, particularly in brain regions outside the substantia nigra.

PTEN‐induced kinase 1 (PINK1), a Ser/Thr kinase encoded by the pink1 gene, is essential for mitophagy—the clearance of damaged mitochondria. ³⁰ It activates Parkin, an E3 ligase, by phosphorylating ubiquitin, and plays a neuroprotective role. ³¹ , ³² , ³³ Loss‐of‐function mutations in pink1 have been linked to early‐onset PD. ³⁴ , ³⁵ , ³⁶ Although pink1‐knockout animals do not develop spontaneous α‐syn aggregation or motor symptoms, ³⁷ , ³⁸ they show increased tendency of Lewy pathology when exposed to fibrils or other stress. ³⁹ , ⁴⁰ , ⁴¹ Beyond mitophagy, cytoplasmic PINK1 may regulate additional neuroprotective pathways, though its substrates remain largely unknown. ⁴² Although PINK1 mutations are found in DLB patients, ⁴³ , ⁴⁴ its role in DLB—whether as a risk factor or contributor to pathogenesis—is unclear.

Given the frequent Aβ co‐pathology in DLB, we generated a mouse model by crossing APP/PS1 transgenic mice (a model of Aβ pathology) with pink1‐knockout mice. Remarkably, this animal model spontaneously develops key DLB features: widespread α‐syn aggregation beyond the substantia nigra, including cortical and hippocampal regions, and diverse behavioral phenotypes mirroring human DLB. Notably, this occurs without α‐syn overexpression or fibril injection. Our findings reveal that PINK1 plays a critical role in suppressing Lewy pathology through a mechanism independent of mitophagy, in which PINK1 deficiency creates a permissive environment for Aβ‐induced α‐syn aggregation.

2. METHODS

2.1. Analysis of human brain tissue

2.1.1. Acquisition of brain tissue samples

Fourteen brain tissue samples, including cingulate gyrus (CG) and parietal cortex (PC) from seven donors (two with AD, two with mixed AD/DLB, and three with DLB), were obtained from the Netherlands Brain Bank (NBB), Netherlands Institute for Neuroscience, Amsterdam (Project number 1613). An additional four brain tissues, including CG and PC from two donors without neurological disorders, were obtained from the Chinese Brain Bank (CBB), part of the National Health and Disease Human Brain Tissue Resource Center, Hangzhou, China (Project number CN20240454).

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed the literature using traditional sources and meeting presentations, and have found that dementia with Lewy bodies (DLB), the second most prevalent neurodegenerative dementia, presents significant diagnostic and therapeutic challenges due to its overlapping clinical and pathological features with Alzheimer's disease (AD) and Parkinson's disease (PD). DLB and PD both exhibit Lewy pathology, but with distinct distributions: widespread across cortical and hippocampal neurons in DLB, but largely confined to substantia nigra dopaminergic neurons in PD. Moreover, current animal models primarily rely on α‐synuclein (α‐syn) overexpression, mutation, or exogenous fibril injection, which fail to capture spontaneous and widespread distribution of Lewy pathology.

2. Interpretations: Our findings demonstrate that PTEN‐induced kinase 1 (PINK1) deficiency, whether through reduced expression or impaired kinase activity, facilitates amyloid beta (Aβ)–induced α‐syn aggregation. Conversely, PINK1 exerts a protective role by specifically phosphorylating α‐syn at Thr44, inhibiting its pathological aggregation. Thus, PINK1 is a critical regulator of α‐syn pathology in DLB, therefore representing a novel genetic risk factor for DLB.

3. Future directions: This study provides a transformative platform for investigating DLB pathogenesis and developing novel therapeutics. The identification of PINK1 as a risk factor for DLB opens new avenues for diagnostic innovation, including the integration of PINK1 genetic screening with biomarker analysis, such as S65‐phosphorylated ubiquitin and T44‐phosphorylated α‐syn. Our study also suggests therapeutic potential of PINK1 activators for synucleinopathies. Future studies will validate these findings in larger cohorts and explore the efficacy of PINK1‐targeted therapies in preclinical models.

All tissue samples were collected with written informed consent from the donors or their next of kin, permitting brain autopsy and the use of the tissue and associated clinical information for research purposes through the establishment of the NBB and CBB. Detailed information for all tissue samples is provided in Table S1 in supporting information. All experimental procedures were approved by the ethics committee of the School of Medicine, Zhejiang University (Approval number ZJU2023‐011).

2.1.2. Immunofluorescence staining of human brain tissue samples

Brain samples were incubated at 62°C overnight for dewaxing. The samples were then rehydrated through a graded series of solvents: three washes in xylene (10 minutes each); followed by two washes each in 100%, 95%, and 75% ethanol; and finally two washes in double‐distilled water (5 minutes each). Antigen retrieval was performed by immersing the samples in ethylenediaminetetraacetic acid (EDTA) Antigen Retrieval Solution (Solarbio, C1034) and heating them to 95°C to 100°C for 30 minutes, after which they were cooled to room temperature for 30 minutes.

Samples were washed three times with phosphate‐buffered saline (PBS), then incubated in 3% hydrogen peroxide for 10 minutes to quench endogenous peroxidase activity. After three additional PBS washes, samples were incubated in 0.3% Triton X‐100 in PBS for 30 minutes for permeabilization. After three more PBS washes, the samples were treated with 0.3% Tween‐20 in PBS for 30 minutes, then rinsed three times with PBS to remove residual detergent.

TrueBlack lipofuscin autofluorescence quencher (Cell Signaling, 92401) was prepared by heating the vial at 70°C for 5 minutes and diluting it 1:20 in 70% ethanol. The solution was applied to the samples at room temperature for 30 seconds (up to a maximum of 3 minutes).

Samples were rinsed with PBS and then incubated in blocking solution (5% normal goat serum and 2% bovine serum albumin in Tris‐buffered saline with Tween 20 [TBST] containing 0.1% Tween‐20) for 1 hour at room temperature. Subsequently, samples were incubated overnight at 4°C with the following primary antibodies: rabbit anti‐α‐synuclein (1:1000, Abcam, ab212184), rabbit anti‐PINK1 (1:200, Novus, BC100‐494), rabbit anti‐phosphorylated ubiquitin (pUb; 1:200, Millipore, ABS1513), mouse anti‐Aβ (1:500, BioLegend, SIG‐39320), and rabbit anti‐pT44‐α‐synuclein (1:200, custom, HuaBio).

After three washes with PBS, samples were incubated with secondary antibodies—Cy3‐conjugated anti‐rabbit immunoglobulin G (IgG; 1:200, Jackson ImmunoResearch, 711‐165‐152) and fluorescein isothiocyanate (FITC)‐conjugated anti‐mouse IgG (1:200, Jackson ImmunoResearch, Pennsylvania, USA, 715‐096‐150)—for 2 hours at room temperature. After three additional PBS washes, samples were mounted on slides using ProLong Gold Antifade Mountant with 4′,6‐diamidino‐2‐phenylindole (DAPI; Invitrogen, P36931).

Slides were scanned using a digital virtual slice scanning system (Olympus), and images were acquired with an Olympus FV1000 confocal microscope. Quantification of fluorescence intensities for pUb and pT44‐α‐synuclein, as well as the percentage of PINK1‐positive cells, was performed using ImageJ (v1.53t, National Institutes of Health [NIH]).

2.2. Assessment of prevalent clinical features of DLB patients

Our study was approved by the ethics committee of Peking University Institute of Mental Health (Sixth Hospital). Written informed consent was obtained from all participants and their family caregivers. We analyzed the prevalence of DLB‐related symptoms in 47 individuals diagnosed with probable DLB, who were registered at the Dementia Care and Research Center (DCRC) of the Peking University Sixth Hospital between January 2008 and August 2019. For the purposes of this study, each diagnosis was re‐evaluated by two senior geriatric psychiatrists based on the criteria outlined in the Fourth Consensus Report of the DLB Consortium. ³

A diagnosis of clinically probable DLB required the presence of either (1) two or more core clinical features, with or without indicative biomarkers, or (2) one core clinical feature accompanied by one or more indicative biomarkers. Core clinical features included: fluctuating cognition with pronounced variations in attention and alertness; recurrent, well‐formed visual hallucinations; rapid eye movement (REM) sleep behavior disorder; and one or more spontaneous cardinal features of parkinsonism. Indicative biomarkers included reduced dopamine transporter uptake in the basal ganglia, as demonstrated by positron emission tomography (PET) imaging, and polysomnographic confirmation of REM sleep without atonia.

Only individuals meeting the criteria for clinically probable DLB were included in the analysis. Exclusion criteria comprised a history of schizophrenia, major depressive disorder, or bipolar disorder, as well as the presence of severe medical conditions that precluded clinical interviews or neuropsychological assessments.

Clinical information was retrospectively retrieved from the case registry. In addition to clinical interviews conducted by dementia specialists, diagnoses of DLB were supported by supplementary assessments, including neuropsychological testing with the Mini‐Mental State Examination and Montreal Cognitive Assessment (n = 37), routine blood tests (n = 30), brain magnetic resonance imaging scans (n = 22), polysomnography (PSG; n = 6), dopamine transporter PET imaging (¹³C‐FP‐CIT PET, n = 5), and amyloid PET imaging (¹¹C‐Pittsburgh compound B [PiB] PET, n = 6).

Note that in the memory clinic, symptoms were initially reported spontaneously by informants. Memory specialists then followed standardized procedures to elicit further clinical details, including the initial symptom(s), age at symptom onset, the full range of symptoms from onset to the time of the clinic visit, and the duration of symptoms. In cases in which there was inconsistency between the informant's report and the specialist's assessment, the symptoms identified by the specialist were recorded. For example, an informant may report that the patient sought medical attention due to memory decline over the past 6 months. However, upon thorough inquiry, the memory specialist might uncover that the patient had exhibited clear signs of REM sleep behavior disorder (RBD)—such as falling out of bed during vivid dreams—1 year earlier. Because the informant had not considered this behavior problematic, it was not initially reported. The memory specialist would then determine that RBD, not memory loss, was the initial symptom and document this accordingly.

All patient histories were documented in the hospital's electronic medical record system. For this study, relevant excerpts from these records were extracted and analyzed. From these clinical records, we systematically extracted information about the initial presenting symptom(s) and other symptoms observed during the disease course, including their timing and specific characteristics. Symptom descriptions reflect the terminology used in the original clinical documentation.

2.3. Characterization of APP/PS1::pink1‐/‐ transgenic mice

2.3.1. Animals

Male C57BL/6J mice were purchased from the Zhejiang Academy of Medical Science. APP/PS1 transgenic mice were obtained from Hangzhou Ziyuan Laboratory Animal Technology. The pink1 knockout mice (pink1‐/‐, C57BL/6J background) were bred and acquired from the Transgenic Mouse Laboratory at the Laboratory Animal Center of Zhejiang University. APP/PS1::pink1‐/‐ and APP/PS1::pink1+/‐ mice were generated by crossing APP/PS1 mice with pink1‐/‐ mice.

All animals were housed at the Laboratory Animal Center of Zhejiang University School of Medicine under standard conditions, with ad libitum access to food and water. The animal rooms were maintained at 20°C to 26°C with ≈ 50% relative humidity and a 12 hour light/dark cycle. All procedures involving animals were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (NIH), and experimental protocols were approved by the Ethics Committee for Laboratory Animal Care and Welfare, Zhejiang University School of Medicine (Approval No. ZJU20190138).

2.3.2. Validation of transgenic mouse model

Knockout of the pink1 gene was confirmed by DNA sequencing, quantitative polymerase chain reaction (qPCR), and western blotting. DNA sequencing revealed that pink1‐/‐ mice expressed a truncated mutant form of PINK1, designated as PINK1(1‐373)‐GVHSPFWPQCGN.

For qPCR, total RNA was extracted from the brain cortex using RNAiso PLUS reagent (Takara Biotechnology; Cat. 9109). Primers were synthesized by Sangon Biotech with the following sequences:

• Pink1‐exon5‐F: GCCCCACACCCTAACATCATCCG

• Pink1‐exon5‐R: CCTCCAGCAACTGCAAGGTCA

• Actb‐F: GGCTGTATTCCCCTCCATCG

• Actb‐R: CCAGTTGGTAACAATGCCATGT

qPCR was performed in triplicate using a SYBR Green kit (Accurate Biology; Cat. AG11701) on a LightCycler 480 real‐time PCR system (Roche, LightCycler 480 II). Relative mRNA expression levels were calculated using the 2^–ΔΔCt method, normalized to Actb.

The level of pUb was used as a surrogate marker for PINK1 expression and activity. A plasmid encoding the mutant PINK1(1‐373)‐GVHSPFWPQCGN under the control of a CMV promoter (pEnCMV‐PINK1(1‐373)‐GVHSPFWPQCGN‐SV40‐Neo) was synthesized (MiaoLingPlasmid). Either wildtype or mutant PINK1 was transfected into HEK293 cells. pUb levels were assessed by western blotting in the presence or absence of 10 µM CCCP (MedChemExpress; Cat. HY‐100941) treatment for 2 hours.

2.3.3. Behavioral tests for the transgenic mice

A Y‐maze test was conducted using the Any‐maze 7.4 software (Stoelting Co.) to assess short‐term working memory, as previously described. ⁴⁵ During the training phase, one arm of the maze was closed, and the mouse was gently placed in the designated start arm. The mouse's activity was recorded for 10 minutes. After training, the mouse was returned to its home cage for a 2 hour interval. In the testing phase, the previously closed arm was opened and designated as the novel arm. The mouse was again placed in the start arm, and its activity was recorded for another 10 minutes. The number of entries and time spent in each arm were automatically analyzed by the software.

A fear conditioning test was carried out using the Fear Conditioning System 4.104 (Coulbourn Instruments), as described previously. ⁴⁵ Freezing behavior was measured throughout each phase to assess fear memory. During the training session, the mouse was placed in a conditioning chamber (Box A) without stimulation for 1.5 minutes. A tone (3000 Hz, 85 dB) was then played for 30 seconds, with a 1 mA foot shock delivered during the final 2 seconds. After the shock, the mouse remained in Box A for 30 seconds before being returned to its home cage.

Testing was conducted 24 hours later. The mouse was returned to Box A with no stimulation, and freezing behavior was recorded for 5 minutes to assess contextual memory.

Two hours after contextual testing, the mouse was placed in a novel chamber (Box B). For the first 2 minutes, no stimulation was given, followed by a 3 minute tone presentation (3000 Hz, 85 dB). Freezing behavior before and during the tone in Box B was recorded to evaluate cued memory.

A pole climbing test was used to evaluate motor function. A custom‐made vertical pole (60 cm in height) was used. Each mouse was placed at the top of the pole and allowed to descend freely in five consecutive trials. During the test session, the time taken for the mouse to climb from 20 to 40 cm above the ground was recorded as a measure of motor coordination and balance.

Motor coordination and balance were further assessed using a Rotarod apparatus (Panlab). The test consisted of a training and a testing phase. During training, mice were placed on a rotating rod with the speed gradually increasing from 4 to 40 rpm over 300 seconds. Each mouse received three training trials with a 10 minute inter‐trial interval. For the testing phase, each mouse was tested in five trials, and the latency to fall off the rotating rod was recorded. The average latency was used as a measure of motor performance.

To measure fecal water content, fresh fecal pellets were collected and weighed immediately (W₁). The samples were then dried in an oven at 60°C until a constant weight was achieved (W₂). Water content was calculated using the formula:

$$Watercontent\left(\%\right)=\left(\left(W_1--W_2\right)/W_1\right)\times 100.$$

2.3.4. Mouse tissue collection and processing

Mice were anesthetized with pentobarbital sodium (150 mg/kg, intraperitoneally) prior to sacrifice. For morphological analysis, the brains, ileum, colon, and heart were collected after transcardial perfusion with ice‐cold saline, followed by freshly prepared 4% paraformaldehyde (PFA). The brains were post‐fixed in 4% PFA at 4°C for 24 hours, then cryoprotected in 30% sucrose for 3 days. Brain tissues were sectioned at a thickness of 25 µm using a cryostat (CM1900, Leica). The heart, colon, and ileum samples were fixed in 4% PFA and processed for paraffin embedding and sectioning.

For western blot analysis, brains were collected after transcardial perfusion with ice‐cold saline. The neocortex was dissected, snap‐frozen, and stored at −80°C until further use.

2.3.5. Immunofluorescence staining of mouse tissue samples

Floating‐section immunofluorescence staining was performed on free‐floating brain slices, while paraffin sections were used for staining the heart, colon, and ileum tissues. For paraffin‐embedded samples, slides were placed in a 65°C oven overnight to remove wax. Tissue sections were then rehydrated through a graded ethanol series (xylene, 100%, 95%, and 70%) followed by distilled water. Antigen retrieval was performed in sodium citrate buffer.

Before incubation with primary antibodies, all tissue sections were permeabilized in 0.3% Triton X‐100 in PBS for 30 minutes and then blocked with 5% donkey serum for 2 hours at room temperature. Samples were subsequently incubated overnight at 4°C with one of the following primary antibodies:

• Rabbit anti‐α‐synuclein (1:600, Cell Signaling, 4179S)

• Mouse anti‐α‐synuclein (1:1000, Abcam, ab280377)

• Rabbit anti‐α‐synuclein aggregate (1:1000, Abcam, ab209538)

• Rabbit anti‐phospho‐α‐synuclein (S129) (1:200, Abcam, ab51253)

• Mouse anti‐Aβ (1:500, BioLegend, SIG‐39320)

• Mouse anti‐NeuN (1:800, Millipore, AB377)

• Rabbit anti‐NeuN (1:1000, Cell Signaling, 24307)

• Rabbit anti‐GFAP (1:800, Cell Signaling, 12389S)

• Mouse anti‐GFAP (1:1000, Cell Signaling, 3670S)

• Rabbit anti‐Iba1 (1:800, Wako, 019‐19741)

• Goat anti‐Iba1 (1:500, Abcam, ab5076)

• Rabbit anti‐MAP2 (1:200, Millipore, AB5622)

• Rabbit anti‐Tau (1:200, Cell Signaling, 46687S)

• Rabbit anti‐phospho‐Tau (Thr181) (1:200, Cell Signaling, 12885S)

• Mouse anti‐β3‐tubulin (1:200, Sigma, MAB1637)

• Chicken anti‐NF‐L (1:200, Abbiotec, 253178)

• Rabbit anti‐CaMKII (1:200, ABclonal, A0186)

• Mouse anti‐tyrosine hydroxylase (1:1000, ImmunoStar, 22941)

• Mouse anti‐Chat (1:200, Sigma, AB144P)

• Rabbit anti‐phospho‐α‐synuclein (Thr22, Thr33, Thr81, and Ser87; 1:200, HuaBio, custom)

After three PBS washes, sections were incubated with secondary antibodies for 2 hours at room temperature. The secondary antibodies included:

• Cy3‐conjugated anti‐rabbit IgG (1:200, Jackson ImmunoResearch, 711‐165‐152)

• FITC‐conjugated anti‐mouse IgG (1:200, Jackson ImmunoResearch, 715‐096‐150)

• FITC‐conjugated anti‐rabbit IgG (1:200, Jackson ImmunoResearch, 111‐095‐003)

• Cy3‐conjugated anti‐mouse IgG (1:200, Millipore, AP181C)

• Alexa Fluor 647‐conjugated anti‐goat IgG (1:200, Jackson ImmunoResearch, 109‐605‐088)

After another three PBS washes, sections were mounted using ProLong Gold Antifade Mountant with DAPI (Invitrogen, P36931). Whole‐slide scanning was performed with the digital virtual slice scanning system (Olympus), or images were acquired using an Olympus FV1000 confocal microscope. Image analysis was conducted using ImageJ and MetaMorph software (version 7.8.0.0).

To count the number of LNs with Aβ, LNs without Aβ, and Aβ plaque without LN, the whole mouse brain was sliced with 25 µm thickness. One brain slice was taken from every 200 µm and double immunofluorescent staining using anti‐α‐syn and anti‐Aβ antibodies. Approximately 35 to 40 brain slices were collected for each mouse. The sum of LNs with Aβ, LNs without Aβ, and Aβ plaque without LN was counted, and the percentage of LNs with Aβ was calculated.

2.3.6. RNA in situ hybridization and immunofluorescence

Brain sections were first treated with 2 µg/mL Proteinase K (Invitrogen, 4333793) for 10 minutes, followed by a 10 minute incubation in 0.1 M Triethanolamine (TEA; Sangon Biotech, A600970). Sections were placed in a humidified chamber and pre‐hybridized for 1 hour at room temperature with a hybridization solution containing 50% formamide (Invitrogen, 15515‐026), 5×SSC (Invitrogen, 15557044), 5×Denhardt's solution (Invitrogen, 750018), 0.25 mg/mL yeast RNA (Biotechnology, R0038), and 0.5 mg/mL herring sperm DNA (Invitrogen, 15634‐017). After pre‐hybridization, the sections were incubated overnight at 65°C with a 0.5 ng/µL digoxigenin (DIG)‐labeled RNA probe. The probe, synthesized by Sangon Biotech, corresponds to the complementary transcript for PINK1 residues 247‐523. After hybridization, sections were washed in 0.2× SSC for 20 minutes at 65°C, then cooled to room temperature. They were washed twice for 10 minutes each in Buffer B1 (0.1 M Tris‐HCl, 1.5 M NaCl, pH 7.4). For blocking, sections were incubated in a humidified chamber for 1 hour at room temperature with Buffer B2 (Buffer B1 with 10% heat‐inactivated sheep serum; Sigma, S2263).

The sections were then incubated overnight at 4°C with a sheep anti‐digoxigenin antibody (Roche, 11093274910) diluted in Buffer B2. The next day, sections were washed three times for 20 minutes each in Buffer B1, followed by two 10 minute washes in Buffer B3 (0.1 M Tris‐HCl, 0.1 M NaCl, 0.05 M MgCl₂, 0.5% TWEEN‐20, pH 9.5). The colorimetric reaction was initiated by adding 2% NBT/BCIP solution (Roche, 11681451001) in Buffer 3 for 10 minutes at room temperature, then stopped with ddH₂O. Last, sections were rinsed with methanol for 15 minutes to remove excess stain before being processed for immunofluorescence with NeuN, allograft inflammatory factor 1 (Iba‐1), and glial fibrillary acidic protein (GFAP) antibodies to identify specific cell types.

2.3.7. Histochemistry staining

Free‐floating brain sections were incubated in 0.3% hydrogen peroxide (H₂O₂) for 30 minutes to quench endogenous peroxidase activity, followed by blocking in 5% goat serum for 2 hours at room temperature. The sections were then incubated overnight at 4°C with primary antibodies: mouse anti‐α‐synuclein (1:1000, Abcam, ab280377) and mouse anti‐NeuN (1:800, Millipore, AB377). After washing with PBS, the sections were incubated with horseradish peroxidase (HRP)‐conjugated goat anti‐mouse IgG (1:100, Biotechnology, A0286) or HRP‐conjugated goat anti‐rabbit IgG (1:100, Biotechnology, A0277) for 1 hour at room temperature. After additional PBS washes, sections were treated with SABC solution (SABC‐HRP Kit, Biotechnology, P0603) for 30 minutes, followed by color development with DAB substrate (ZSGB‐Bio, Beijing, China, ZLI‐9018) according to the manufacturer's instructions.

For dehydration and clearing, the sections were sequentially immersed in 70% ethanol (2 minutes), 95% ethanol (2 minutes), 100% ethanol (3 minutes), and xylene (2 minutes), followed by an additional 5 minute incubation in xylene. Sections were mounted using rhamsan gum. Images were acquired with a BX51 microscope (Olympus) and analyzed using ImageJ software.

To assess the resistance of aggregated α‐syn to protease digestion, brain sections were washed three times with PBS, then incubated with 0.2 mg/mL proteinase K (Accurate Biology, AG12004) for 5 minutes at room temperature. After additional PBS washes, sections were processed for immunohistochemistry as described above.

2.3.8. Hematoxylin and eosin staining

Paraffin‐embedded tissue sections were deparaffinized and rehydrated, then stained with hematoxylin for 5 minutes and rinsed in distilled water for 10 minutes. Sections were then briefly differentiated in 1% hydrochloric acid ethanol for 3 seconds and rinsed again in distilled water for 5 minutes. Counterstaining was performed with eosin for 2 minutes. After dehydration and clearing through graded alcohol and xylene, sections were mounted with rhamsan gum. Images were captured using a BX51 microscope (Olympus) and analyzed with ImageJ.

2.3.9. Masson's trichrome staining

After deparaffinization and rehydration, tissue sections were immersed in Ponceau S staining solution for 2 minutes and rinsed with tap water. Sections were then treated with phosphomolybdic acid for 15 to 20 seconds (without rinsing), followed immediately by staining with aniline blue (Sigma, 415049) for 3 minutes. After washing with tap water, differentiation was performed in 0.2% glacial acetic acid for 2 seconds. Sections were dehydrated and cleared sequentially in 95% ethanol (15 minutes), 100% ethanol (10 minutes, twice), and xylene (10 minutes, twice). After air drying, sections were mounted with neutral balsam. Images were acquired with a BX51 microscope (Olympus) and analyzed using ImageJ.

2.3.10. Transmission electron microscopy

Mouse cortical tissues were fixed in 2.5% glutaraldehyde for 24 hours at 4°C. After three PBS washes, samples were post‐fixed in 1% osmium tetroxide for 1.5 hour at room temperature, followed by three washes in distilled water. Sections were then stained with 4% uranyl acetate for 30 minutes, dehydrated in a graded ethanol series (50%, 70%, 90%, and 100%), and then in 100% acetone (twice). Samples were infiltrated in a 1:1 mixture of acetone and EPON812 resin (Electron Microscopy Sciences, 14900) for 2 hours, followed by 100% EPON812 infiltration for an additional 2 hours at room temperature. Polymerization was performed sequentially at 37°C for 24 hours, 45°C for 24 hours, and 60°C for 48 hours.

Ultrathin sections (≈ 120 nm) were cut using an Ultracut microtome (Leica), stained with 4% uranyl acetate for 20 minutes and lead citrate for 5 minutes, rinsed, and dried. Images were obtained using a Philips TECNAI 10 transmission electron microscope (80 keV; Thermo Scientific) equipped with a Gatan 794 CCD camera (Thermo Scientific).

2.4. Cell culture and transfection

HEK293 cells (National Collection of Authenticated Cell Cultures, SCSP‐5500) were cultured in Dulbecco's Modified Eagle Medium (DMEM; ThermoFisher Scientific, C11995500BT) supplemented with 10% heat‐inactivated fetal bovine serum (Zhejiang Tianhang Biotechnology, 11011‐8611). Cells were maintained at 37°C in a humidified incubator with 5% CO₂.

Plasmid transfection was performed using Lipofectamine 2000 (ThermoFisher Scientific, 11668019) according to the manufacturer's instructions. Six hours post‐transfection, the medium containing transfection reagents was replaced with fresh culture medium. Cells were harvested at the indicated time points after transfection by trypsinization.

2.5. Immunoprecipitation and western blot analyses

2.5.1. Preparation of protein samples from mouse brain tissues and cultured cells

Cells for western blot analysis were collected by trypsinization after two washes with PBS. Cell lysates were prepared using radioimmunoprecipitation assay (RIPA) buffer (Beyotime Biotechnology, P0013B) supplemented with protease inhibitor (Beyotime Biotechnology, P1005) and phosphatase inhibitor (Beyotime Biotechnology, P1081) from the same supplier. Mouse brain tissues were homogenized in the same lysis buffer using a pre‐cooled TissuePrep instrument (TP‐24, Gering Instrument Company) for 1 minute at 4°C.

Samples were incubated on ice for 30 minutes and centrifuged at 13000 g for 30 minutes at 4°C. The supernatant was collected as the soluble protein fraction. The remaining pellet was resuspended in sodium dodecyl sulfate (SDS) buffer (2% SDS, 50 mM Tris‐HCl, pH 7.5) and subjected to ultrasonic disruption (Diagenode) at 4°C using 8 cycles of 10 second sonication with 30 second intervals. After complete solubilization, the samples were centrifuged again at 13000 g for 30 minutes at 4°C, and the supernatant was collected as the insoluble protein fraction. Protein concentrations were determined using a BCA Protein Assay Kit (Beyotime Biotechnology, P0009).

2.5.2. Preparation of mitochondrial protein samples

Mitochondria were isolated using a commercial mitochondrial extraction kit (Elabscience, E‐BC‐E001). Approximately 0.5 mL of pre‐chilled Reagent 1 from the kit was added to pre‐weighed mouse hippocampal tissues, which were then homogenized using a pre‐cooled TissuePrep instrument (TP‐24, Gering Instrument Company) for 1 minute at 4°C. The homogenate was transferred to a pre‐chilled 2 mL centrifuge tube and centrifuged at 500 × g for 5 minutes at 4°C. This low‐speed centrifugation step was repeated twice to remove debris.

The final supernatant was transferred to a new pre‐chilled tube and centrifuged at 15,000 × g for 15 minutes at 4°C. The resulting mitochondrial pellet was resuspended in 0.2 mL of Reagent 2 from the kit and centrifuged again at 11,000 × g for 10 minutes. After discarding the supernatant, the purified mitochondrial pellet was lysed in 30 µL of RIPA buffer, incubated on ice for 30 minutes, and centrifuged at 13000g for 30 minutes at 4°C. The supernatant was collected as the mitochondrial protein sample.

2.5.3. Co‐immunoprecipitation

For mouse brain samples, lysates were prepared as described in section 2.5.1 above. Samples were incubated overnight at 4°C with one of the following antibodies: rabbit anti‐PINK1 (1:50, Novus, BC100‐494), rabbit anti‐α‐synuclein (1:50, Cell Signaling, 51510T), or rabbit IgG control (1:50, Proteintech, 30000‐0‐AP). The immune complexes were then incubated with Protein G magnetic beads (ThermoFisher Scientific, 10003D) for 3 hours at 4°C. Bead‐bound complexes were washed three times with NP‐40 buffer (Beyotime Biotechnology, P0013F) and subsequently eluted for western blot analysis.

To detect interactions between PINK1 and α‐syn, HEK293 cells were co‐transfected with either pCMV‐SNCA(human)‐3×Flag‐Neo (MiaoLingPlasmid, P42769) and pCDNA3.1‐PINK1‐Myc (MiaoLingPlasmid, P4941), or pCMV‐SNCA(human)‐Myc‐Neo (MiaoLingPlasmid, P62659) and a custom pIRES2‐PINK1‐3×Flag construct. Lysates were prepared as described in section 2.5.1 and incubated overnight at 4°C with anti‐FLAG M2 magnetic beads (Millipore, M8823‐5ML). After three PBS washes, the immune complexes were eluted for downstream analysis.

2.5.4. Western blot analysis

Equal amounts of protein were loaded for SDS polyacrylamide gel electrophoresis (SDS‐PAGE) and transferred to polyvinylidene fluoride membranes. Membranes were blocked and incubated overnight at 4°C with the following primary antibodies at an appropriate dilution:

• Rabbit anti‐α‐synuclein (1:1000, Cell Signaling, 4179S)

• Mouse anti‐GAPDH (1:5000, Proteintech, 60004‐1‐Ig)

• Mouse anti‐β‐actin (1:5000, Proteintech, 66009‐1‐Ig)

• Rabbit anti‐DYKDDDDK (FLAG) (1:5000, Proteintech, 80010‐1‐RR)

• Mouse anti‐ubiquitin (1:800, Santa Cruz Biotechnology, SC‐8017)

• Rabbit anti‐pUb (1:1000, Millipore, ABS1513)

• Rabbit anti‐TOM20 (1:1000, Cell Signaling, 42406S)

• Rabbit anti‐LAMP1 (1:1000, Abcam, , ab24170)

• Rabbit anti‐p62 (1:1000, Abcam, ab109012)

• Rabbit anti‐pT22, ‐pT44, ‐pT81, ‐pS87 α‐synuclein (HuaBio, , custom)

• Rabbit anti‐LC3 (1:1000, Sigma, L7543)

• Rabbit anti‐TIM23 (1:1000, Proteintech, 11123‐1‐AP)

• Rabbit anti‐PINK1 (1:1000, Novus, BC100‐494)

After three washes with TBST, membranes were incubated with HRP‐conjugated secondary antibodies: goat anti‐mouse IgG (1:10,000, Cell Signaling, 7076S) or goat anti‐rabbit IgG (1:10,000, Jackson ImmunoResearch, 111‐035‐003).

Protein bands were visualized using an ECL detection kit (Potent ECL Kit, MultiSciences Biotech, P1425) and imaged with the GBOX system (LI‐COR, Odyssey‐SA‐GBOX). Densitometric analysis was performed using ImageJ.

2.6. Analysis of published RNA‐seq data

2.6.1. Analysis of bulk RNA‐seq data

Bulk RNA‐seq data were obtained from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/). The dataset GSE150696 includes transcriptomic profiles from the prefrontal cortex of 9 control donors and 12 patients diagnosed with DLB. Based on PINK1 mRNA expression levels, DLB samples were stratified into two subgroups: PINK1‐low (pink1_L ), in which expression was lower than all controls, and PINK1‐high (pink1_H ), in which expression exceeded that of control samples.

Differentially expressed genes (DEGs) were identified using the limma package ⁴⁶ (v3.54.2). DEGs were considered significant if they met the criteria of adjusted p < 0.05 and fold change (FC) > 2 or < 0.5. Results are provided in Data S1 in supporting information. Visualization of common DEGs across comparisons was performed using ggplot2 (v3.4.1), and a Venn diagram was generated with an online tool (https://bioinformatics.psb.ugent.be/webtools/Venn/). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted using clusterProfiler ⁴⁷ (v4.7.1.003).

2.6.2. Analysis of single‐nucleus RNA‐seq data

Single‐nucleus RNA‐seq data were also retrieved from GEO. Dataset GSE178146 contains transcriptomic data from the anterior cingulate cortex of seven control donors and seven DLB patients. Data analysis was performed using Seurat ⁴⁸ (v4.3.0), and heatmaps of gene expression were visualized using ComplexHeatmap ⁴⁹ (v2.14.0).

DEGs between subgroups were identified using Seurat's FindMarkers function, with significance defined as an adjusted p < 0.05. Results are provided in Data S2 in supporting information. Venn diagrams were generated using the same online tool as above, and functional enrichment analyses (GO and KEGG) were performed using clusterProfiler (v4.7.1.003).

2.7. In vitro characterization α‐syn aggregation

2.7.1. Generation of α‐syn/T44E plasmid via PCR‐based mutagenesis

Site‐directed mutagenesis of the pT7‐7‐α‐syn/T44E plasmid was performed using PCR‐based mutagenesis based on pT7‐7 α‐Syn plasmid (Tsingke Biotech, 36046). The reaction mixture (50 µL total volume) contained: 5 µL of 10× KOD‐Plus‐Neo buffer, 5 µL of 2 mM dNTPs, 3 µL of 25 mM MgSO₄, 1 µL of KOD‐Plus‐Neo DNA polymerase (Toyobo; KOD‐401), 3 µL of 10 pM primer mix (forward: 5′‐AGG CTC CAA AGA GAA GGA GGG AGT GGT GCA TGG TGT GGC‐3′; reverse: 5′‐ACT CCC TCC TTC TCT TTG GAG CCT ACA TAG AGA ACA CCC TC‐3′), and 50 ng of pT7‐7‐α‐syn template, adjusted with nuclease‐free water.

PCR was conducted with an initial denaturation at 94°C for 2 minutes, followed by 30 cycles of 98°C for 10 seconds and 68°C for 2 minutes. The amplified product was digested with DpnI (New England Biolabs, R0176S) to eliminate the methylated template, then purified using a commercial PCR purification kit (Omega Bio‐Tek, D6492‐02).

2.7.2. Purification of wildtype and T44E α‐syn proteins

Plasmids encoding wildtype α‐syn or α‐syn/T44E were transformed into Escherichia coli BL21(DE3) cells. Protein expression was induced with 1 mM IPTG (Yuanye Bio‐Technology, R21349) at 37°C for 4 hours. Cells were harvested and lysed using a high‐pressure homogenizer in 10 mM Tris‐HCl (pH 8.0) containing 1 mM phenylmethylsulfonyl fluoride and 1 mM EDTA. The lysate was boiled and centrifuged to remove debris. Nucleic acids were removed by adding streptomycin (10 mg/mL), followed by α‐syn precipitation with ammonium sulfate (360 mg/mL). The protein was then dissolved in 20 mM Tris‐HCl (pH 7.7).

Purification was carried out via anion exchange chromatography on a Source 15Q column (Cytiva) with a linear NaCl gradient (0–1 M) in 20 mM Tris‐HCl (pH 7.7), followed by size‐exclusion chromatography using a HiLoad 16/600 Superdex 75 pg column (Cytiva) in 20 mM 2‐morpholinoethanesulfonic acid (MES), 100 mM NaCl (pH 6.0).

2.7.3. Preparation of PINK1‐phosphorylated α‐syn

The pGEX‐6P‐1‐PhPINK1 (Pediculus humanus, Uniprot entry E0W1I1, cytoplasmic residues 115‐575) plasmid (Tsingke Biotech, customized) was transformed into E. coli BL21(DE3) cells. Expression was induced with 0.2 mM isopropyl β‐d‐1‐thiogalactopyranoside (IPTG) at 20°C for 20 hours. Cells were lysed in 20 mM Tris‐HCl (pH 7.4), 150 mM NaCl, 5% glycerol, and 1 mM dithiothreitol (DTT) using a high‐pressure homogenizer. PhPINK1 was purified using a GSTrap HP column (Cytiva), followed by size‐exclusion chromatography on a HiLoad 16/600 Superdex 75 pg column (Cytiva) in 20 mM HEPES, 150 mM NaCl (pH 7.4).

For in vitro phosphorylation, α‐syn was incubated with PhPINK1 at a 1:20 molar ratio in reaction buffer (20 mM HEPES pH 7.4, 150 mM NaCl, 5 mM MgCl₂, 5 mM ATP, and 10 mM tris[2‐carboxyethyl]phosphine) overnight at 30°C. Phosphorylated α‐syn was purified by Superdex 200 Increase 10/300 GL (Cytiva). SDS‐PAGE followed by western blotting confirmed phosphorylation using rabbit anti‐α‐syn antibody (1:1000, Cell Signaling, 4179S) and site‐specific anti‐phospho‐α‐syn antibodies (Thr22, Thr33, Thr81, Ser87; 1:200, HuaBio, customized).

2.7.4. Analysis of α‐syn fibrilization

Purified α‐syn was buffer‐exchanged into HEPES using the Superdex 75 pg column and adjusted to 50 µM in HEPES buffer. Fibrils were formed by incubation at 37°C with 1000 rpm shaking for up to 7 days in a thermomixer, following an established protocol. ⁵⁰ Fibril formation was confirmed by Thioflavin T fluorescence assay.

Lyophilized Aβ42 (Macklin, A834109) was dissolved in Hexafluoroisopropanol to 1 mM, evaporated overnight, and redissolved in DMSO (5 mM stock). The monomer was freshly prepared by diluting the stock to 0.36 mg/mL in ice‐cold HEPES buffer.

Aβ42 Fibrils were generated by further dilution to 1 mg/mL in ice‐cold HEPES buffer and incubating at 37°C with 1000 rpm shaking for 2 days. Fibril formation was confirmed by Thioflavin T assay.

α‐Syn (27.8 µM), α‐syn/T44E, or phosphorylated α‐syn in 900 µL MES buffer was incubated with 100 µL of HEPES buffer, or pre‐sonicated Aβ42 monomers, fibrils, or α‐syn fibrils (0.36 mg/mL) at 37°C with 1000 rpm shaking for 4 days. At selected time points, 50 µL samples were mixed with 1.25 µL of 2 mM Thioflavin T (Harveybio, DY1206), incubated at 37°C for 15 minutes, and fluorescence was measured (excitation: 440 nm; emission: 480 nm) using a BioTek microplate reader.

Samples were serially diluted, and 6 µL aliquots were spotted onto nitrocellulose membranes. After drying, membranes were blocked in 5% milk/TBST, incubated overnight at 4°C with an anti‐α‐syn aggregate antibody (1:1000, Abcam, ab209538) in 0.1% BSA/TBST, and then with HRP‐conjugated secondary antibody (1:10,000, Cell Signaling, 7076S). Detection was performed using chemiluminescent substrate, and spot intensities were quantified using ImageJ (v1.54j, NIH).

2.7.5. Phosphorylation site identification by mass spectrometry

PINK1‐phosphorylated α‐syn was analyzed using a Waters Vion IMS QTof mass spectrometer coupled to an ACQUITY UPLC I‐Class PLUS system. Separation was performed on a 3 µm, 5 cm Ultimate XB‐C4 column (Welch) with a 3 minute linear gradient from 90% water/0.1% formic acid to 60% acetonitrile/0.1% formic acid at 200 µL/minutes. Source settings included: 3.00 kV capillary voltage, 100°C source temperature, 300°C desolvation temperature, and gas flows of 50 L/h (cone) and 800 L/h (desolvation). Data were processed using UNIFI software (Waters).

For site‐specific identification, proteins were acetone precipitated, reduced with DTT, alkylated with iodoacetamide, and digested overnight with trypsin (1:20, Promega, V5111). Peptides were desalted, dried, and analyzed using an Easy‐nLC 1200 coupled to an Orbitrap Fusion Lumos. Peptides were separated on a PepMap 100 trap column and an RSLC analytical column with a 75 minute gradient from Buffer A (0.1% formic acid in water) to Buffer B (0.1% formic acid in 80% acetonitrile).

MS1 scans were acquired at 120,000 resolution (m/z 200), automated gain control (AGC) target 5.5 × 10⁵, max injection time 90 ms. MS2 spectra were collected in DDA mode using higher energy collisional disassociation (HCD) fragmentation (NCE 30%) with resolution 30,000, AGC target 5 × 10⁴. Raw data were analyzed in Proteome Discoverer v2.2 with Percolator for PSM filtering at 1% false discovery rate (FDR). Statistical analysis was conducted for each phosphorylation site.

2.8. Data‐assisted modeling of PINK1/α‐syn complex structure

2.8.1. Cross‐linking analysis between PINK1 and α‐syn

The cross‐linker BS3 (100 mM in DMSO; ThermoFisher Scientific, 21586) was added to a protein mixture containing 21 µM PhPINK1 and 42 µM α‐syn at a final concentration of 1.73 mM. The reaction was carried out at room temperature for 60 minutes and quenched with 20 mM Tris‐HCl (pH 7.5). All experiments were performed in triplicate.

Cross‐linked proteins were precipitated by adding six volumes of pre‐chilled acetone and incubated at 4°C, followed by centrifugation to remove the supernatant. Pellets were resuspended in 8 M urea/0.1 M Tris‐HCl (pH 8.5), reduced with 5 mM DTT for 10 minutes at 25°C, and alkylated with 10 mM iodoacetamide for 15 minutes in the dark. Samples were then diluted 3‐fold with 0.1 M Tris‐HCl (pH 8.5) containing 1 mM CaCl₂ and 20 mM methylamine, and digested overnight at 37°C with sequencing‐grade modified trypsin at a 1:20 enzyme‐to‐substrate mass ratio. The resulting peptides were desalted and dried under vacuum.

Peptides were reconstituted in 0.1% formic acid and analyzed using an Easy‐nLC 1200 system coupled to an Orbitrap Fusion Lumos mass spectrometer. Peptides were separated on a C18 column using a 60 minute gradient of Buffer A (0.1% formic acid in water) and Buffer B (0.1% formic acid in acetonitrile) as follows: 0% to 5% B in 4 minutes, 5% to 38% B in 41 minutes, 38% to 100% B in 7 minutes, and 100% B for the final 8 minutes.

Data‐dependent acquisition (DDA) was performed with MS1 full scans at 120,000 resolution (at m/z 200), AGC target of 6 × 10⁵, and maximum injection time of 50 ms. MS2 scans were acquired using HCD fragmentation at 15,000 resolution (at m/z 200), with an AGC target of 1 × 10⁵ and normalized collision energy (NCE) of 30%.

Cross‐linked peptides were identified using pLink2 software ⁵¹ with the following parameters: instrument type, HCD; precursor mass tolerance, 10 ppm; fragment ion tolerance, 20 ppm; peptide length, 6 to 60 amino acids per chain; peptide mass, 600 to 6000 Da per chain; variable modifications, carbamidomethyl[C] and oxidation[M]; cross‐linker, BS3 (targeting lysine residues and protein N‐termini; cross‐link mass shift: 138.068 Da; mono‐link mass shift: 156.079 Da). Search results were filtered using a 5% FDR and a spectral‐level Support Vector Machine score cutoff of 0.01.

2.8.2. Structural modeling of the human PINK1–α‐syn complex

The structural model of the human PINK1–α‐syn complex was generated using AlphaFold‐Multimer. ⁵⁰ To minimize non‐specific docking interactions, both the N‐ and C‐terminal regions of α‐syn were truncated, retaining primarily the central non‐amyloid component (NAC) region.

A total of 25 predicted complex structures were generated. Models were first screened for α‐syn placement within the active‐site cleft of PINK1. Docking quality was evaluated using the predicted alignment error (PAE), with lower PAE values (in Å) indicating more reliable inter‐protein positioning. Models were further validated by comparing predicted interface residues to experimentally determined intermolecular cross‐links between PhPINK1 and α‐syn.

2.9. Statistical analysis

All data are presented as mean ± standard deviation (SD). Statistical analyses and graph generation were performed using GraphPad Prism software (version 6.0; GraphPad Software). Outliers were identified using the ROUT test, and data normality was assessed using the D'Agostino–Pearson omnibus normality test. For data that passed the normality test, statistical comparisons were made using either an unpaired t test, one‐way analysis of variance (ANOVA) followed by Tukey multiple comparisons test, or two‐way ANOVA with Tukey post hoc test, as appropriate. A p value of < 0.05 was considered statistically significant.

3. RESULTS

3.1. Aβ and α‐syn co‐pathologies in human brains with DLB

To investigate the relationship between Aβ plaques and α‐syn aggregates in DLB, we analyzed human brain tissue samples from the NBB. These samples were obtained from two patients with AD and two patients with AD/DLB at amyloid stage B, as well as three patients with DLB at amyloid stages O, A, and B, respectively (Table S1). Immunofluorescence staining was performed to visualize the distributions of α‐syn and Aβ in cingulate gyrus and parietal cortex (Figure 1; stages of LNs, LBs, Aβ plaques, and intracellular Aβ were categorized and listed in Table S2 in supporting information). In brain samples from donors A and B with AD, we observed Aβ plaques but only a few α‐syn aggregates (Figure 1A(i), B(i)). Samples from donors C and D with AD/DLB showed similar numbers of Aβ plaques, with some of them in close proximity to α‐syn aggregates that are morphologically consistent with LNs (denoted with asterisks in Figure 1A(ii), B(ii)). In some other instances, α‐syn aggregates were observed overlapping with Aβ deposits and surrounding DAPI‐positive clumps, suggestive of LBs formed around damaged nuclei (denoted with arrows in Figure 1A(ii), B(ii)).

FIGURE 1.

Co‐occurrence of α‐syn and Aβ pathologies were found in the brains of human DLB patients. A, Representative double immunofluorescence staining images from cingulate gyrus. B, Representative images from parietal cortex. The patients include (i) donor A and B with AD, (ii) donor C and D with AD/DLB, and (iii) donor E, G, and F with DLB at various stages. The detailed information of all donors is provided in Table S1 in supporting information. Lewy neurites are indicated with asterisks, and Lewy bodies are indicated with arrows. α‐syn, α‐synuclein; Aβ, amyloid beta; AD, Alzheimer's disease; DAPI, 4′,6‐diamidino‐2‐phenylindole; DLB, dementia with Lewy bodies.

In donor E (amyloid stage O) with DLB, though Aβ signals were weak, thread‐like α‐syn puncta resembling LNs were observed (denoted with asterisks in Figure 1A(iii), B(iii)). In the same donor, α‐syn aggregates were found within several neurons and co‐localized with intracellular Aβ (denoted with arrows in Figure 1A(iii), B(iii)). In donor F (amyloid stage A) and donor G (amyloid stage B) with DLB, α‐syn aggregates were observed surrounding the nuclei, consistent with LBs (denoted with arrows in Figure 1A(iii), B(iii)), with some of the aggregates co‐localized with Aβ plaques, suggestive of LNs (denoted with asterisks in Figure 1A(iii), B(iii)). These observations confirm the frequent Aβ/α‐syn co‐pathologies in brains with DLB, with α‐syn existing in either LBs within neuronal cell bodies and LNs associated with Aβ plaques. ³ , ¹⁵ , ⁵²

The analysis of these human brain tissues, while limited in cohort size, was essential for generating our central hypothesis. The qualitative observations—particularly the spatial relationship between α‐syn aggregates and Aβ plaques in DLB brains—provided the foundation for us to systematically investigate a novel role for PINK1 deficiency in facilitating this pathological cross‐talk with the construction of in vivo and in vitro models.

3.2. PINK1 level and activity decreased in the brain with DLB

Increased PINK1 level and activity have been previously found in AD brains. ⁵³ Given that Ub is a main substrate of PINK1, ³³ the level of pUb was assessed for PINK1 level and activity. Using immunofluorescent staining, we found that the PINK1 and pUb level increased in an age‐dependent manner in the cortex of the APP/PS1 mouse, a well‐established AD mouse model (Figure S1 in supporting information). We then assessed the expression level and activity of PINK1 in diseased human brain samples, as well as age‐matched individuals without morphological neuronal abnormalities (Table S1). The control brains had mild PINK1 expression and pUb level (Figure 2A(i), B(i)). Compared to the control samples, PINK1 and pUb signals were much higher in the two brains with AD. The increased PINK1 and pUb signal were found co‐localized with Aβ within cell bodies or co‐occurring with Aβ plaques outside cell bodies (Figure 2A(ii), B(ii)).

FIGURE 2.

Correlation between PINK1 deficiency with DLB pathogenesis. A, Double immunofluorescence staining of PINK1 and Aβ in human brain samples from cingulate gyrus. B, Double immunofluorescence staining of pUb and Aβ in human brain samples from cingulate gyrus. C and D, Analysis of published bulk transcriptomics data for the parietal brain region (cortex BA9), ⁵⁴ with the differentially expressed genes (DEGs) defined as |logFC| ≥ 1. PINK1 expression was quantitated in control (CTRL) and DLB brains. The DLB samples were stratified into low (pink1 _L) and high (pink1 _H) expression groups relative to the control (C), and the Venn diagram shows the number of DEGs between CTRL, pink1 _L, and pink1 _H groups (D). E and F, Analysis of published single‐nucleus transcriptomic data for the anterior cingulate gyrus. ⁵⁵ DEGs were defined as |logFC| ≥ 1. The percentage of PINK1‐positive neurons (pink1+) and average PINK1 mRNA expression in neuronal nuclei from control (CTRL) and DLB brains (E), and the Venn diagram illustrating DEGs between the pairwise comparisons of neuronal nuclei from CTRL::pink1‐, CTRL::pink1+, DLB::pink1‐, and DLB::pink1+ groups (F). Aβ, amyloid beta; DAPI, 4′,6‐diamidino‐2‐phenylindole; DLB, dementia with Lewy bodies; PINK1, PTEN‐induced kinase 1; pUb, phospho‐ubiquitin.

The PINK1 level and activity were variegated in the two AD/DLB brains. In one sample (donor C), PINK1 and pUb level were moderately increased compared to the control (upper panel in Figure 2A(iii), B(iii)). However, in the other sample (donor D), the PINK1 level was similar to control samples, and pUb level was absent (lower panel in Figure 2A(iii), B(iii)), indicating a deficiency of PINK1 activity.

In the three DLB brains, PINK1 expression was mild in the brain with amyloid stage O (donor E), while pUb was absent in this sample (upper panel in Figure 2A(iv), B(iv)). The two other brains with DLB (donor F and G) showed the same pattern; both PINK1 and pUb were absent (middle and lower panel in Figure 2A(iv), B(iv)). Thus, in contrast to AD brains, these findings demonstrate reduced PINK1 expression and activity in DLB brains.

3.3. PINK1 deficiency is associated with DLB neurodegeneration

To explore the role of PINK1 in DLB, we analyzed the impact of PINK1 gene expression on DLB‐related genes using two publicly available databases. First, bulk transcriptomics data from the parietal cortex of 12 DLB and 9 control samples ⁵⁴ were evaluated. Based on PINK1 expression levels in the controls, the DLB samples were stratified into low (pink1 _L , N = 5) and high (pink1 _H , N = 7) expression subgroups (Figure 2C, Data S1). We identified 11 DEGs between pink1 _H and control, 68 DEGs between pink1 _L and pink1 _H , and 288 DEGs between pink1 _L and control (Figure 2D). The large overlap of 58 DEGs between the pink1 _L versus pink1 _H and pink1 _L versus control comparisons corroborates the stratification of samples into pink1 _L and pink1 _H subgroups. Furthermore, the substantial difference in DEGs between the pink1 _H versus control and pink1 _L versus control comparisons indicates the critical role of PINK1 in DLB pathogenesis. GO and KEGG pathway analyses of the 68 DEGs between pink1 _L and pink1 _H subgroups, as well as the 288 DEGs between pink1 _L and controls subgroups, consistently revealed significant enrichment in pathways related to synaptic structure and function (Figure S2 in supporting information). This suggests that PINK1 deficiency causes disruption of synaptic processes, potentially contributing to DLB pathogenesis.

The second dataset contains single‐nucleus transcriptomic reads from the anterior cingulate gyrus of seven controls and seven DLB patients ⁵⁵ (Data S2). Neurons were categorized based on PINK1 expression into pink1+ (detectable expression, for ≈ 10% of the neurons) and pink1‐ (undetectable expression, for ≈ 90% of the neurons) subgroups. The proportion of pink1+ neurons and pink1 expression level were mostly lower in DLB brains than in the controls (Figure 2E). Comparison of DLB and control groups revealed 761 DEGs between CTRL::pink1‐ and DLB::pink1‐ neurons, and 590 DEGs between CTRL::pink1+ and DLB::pink1+ neurons. Comparison of pink1+ and pink1‐ groups revealed 197 DEGs between DLB::pink1+ and DLB::pink1‐, but only 10 DEGs between CTRL::pink1‐ and CTRL::pink1+ (Figure 2F). The significantly larger number of DEGs in DLB than controls further indicates that PINK1 expression plays a significant role in regulating gene expression and associated pathways in the context of DLB, but only has a limited impact in healthy brains.

GO analysis of the 197 DEGs between DLB::pink1+ and DLB::pink1‐ neurons revealed enrichment in pathways related to synaptic vesicle cycle, synaptic vesicle maturation, oxidative phosphorylation, and mitochondrial processes, consistent with the roles of α‐syn and the established role of PINK1 in mitochondrial function (Figure S3A in supporting information). These DEGs were also enriched in cellular components of distal axons, growth cones, and clathrin‐coated vesicle membranes, suggesting impacts on neuronal structure and function beyond mitochondrial function (Figure S3A). Furthermore, KEGG analysis highlighted pathways related to multiple neurodegenerative diseases, including PD, AD, Huntington's disease, and amyotrophic lateral sclerosis (Figure S3B), despite that DLB has not been included as a distinct entry in the database.

Together, transcriptomic analysis strongly indicates that PINK1 deficiency is linked to the dysregulation of synaptic function and pathways relevant to DLB. Based on these findings, we hypothesize that reduced PINK1 expression or activity plays a key role in promoting α‐syn aggregation and the onset of Lewy pathology, particularly under stress conditions such as Aβ accumulation.

3.4. APP/PS1::pink1‐/‐ mice exhibit motor and cognitive deficits at early age

We generated APP/PS1::pink1‐/‐ mice by crossing Pink1 knockout mice (pink1‐/‐) with APP/PS1 mice. The pink1‐/‐ mice were generated in our laboratory and the successful knockout was confirmed through genotyping, qPCR, and the subsequent assessment of PINK1 catalytic activity (Figure S4A–C in supporting information). Although this knockout effectively eliminates PINK1 expression across all cell types, our analysis of publicly available single‐cell RNA sequencing data and subsequent validation via RNA in situ hybridization with immunofluorescence staining showed that pPink1 mRNA is predominantly expressed in neurons within the cortex and hippocampus of wildtype mice (Figure S5 in supporting information). Consequently, the functional consequences of the global knockout are most pronounced in these neuronal populations, leading to the observed motor and cognitive deficits at early age.

As early as 3 to 4 months old, the APP/PS1::pink1‐/‐ mice started to exhibit motor dysfunction, including increased pole climbing time (Figure 3A) and decreased rotarod latency (Figure 3B). One mouse even experienced an atonic seizure at 1 month old, despite of a normal body weight (Figure 3C). Furthermore, these mice exhibited cognitive impairment, characterized with reduced freezing responses in fear conditioning experiments and deficits in both contextual and cue memory (Figure 3D,E). Additionally, they failed to show a preference to the novel arm in the Y‐maze test, further demonstrating their cognitive deficits (Figure 3F,G). The behavioral test results indicate that the APP/PS1::pink1 ^−/− mice experience pathology beyond dopaminergic neurons, with brain regions associated with learning and memory impacted.

FIGURE 3.

APP/PS1::pink1‐/‐ mice exhibit DLB‐like behavioral and pathological characteristics. A, Pole climbing test demonstrating locomotor dysfunction in 3–4‐month‐old mice. n = 10, mean ± SD, *p < 0.05, one‐way ANOVA. B, Rotarod performance (latency to fall) indicating locomotor dysfunction in 3‐ to 4‐month‐old mice. n = 10, mean ± SD, one‐way ANOVA. C, Mouse #804 exhibited atonia at 1 month old. Freezing percentage in contextual (D) and cue‐dependent (E) memory tasks in the fear conditioning test in 3‐ to 4‐month‐old mice. n = 10, mean ± SD, ***P < 0.001 compared to training (D) or without tone (E), paired t test. ^## P < 0.01, ^### P < 0.001 compared to wild type (one‐way ANOVA). Percentage of time spent (F) and entries (G) into arms in the Y‐maze during training and testing in 3‐ to 4‐month‐old mice. n = 10, mean ± SD, **P < 0.01, *** P < 0.001 among groups (one‐way ANOVA). H, Representative images of α‐syn and Aβ double immunofluorescence staining showing Lewy neurites (LNs) surrounding Aβ plaques. The images from APP/PS1 mice aged 5.5 and 12 months are shown in rows 3 and 4. I, Quantitative analysis of LNs with Aβ, LNs without Aβ, and Aβ without LNs for 6 APP/PS1::pink1‐/‐ mice. The total number of LNs with Aβ, LNs without Aβ, and Aβ without LNs (in parentheses) and the percentage of LNs with Aβ of each mouse were presented above each bar. J, Mapping of the distribution of LNs in brain regions using the mouse brain atlas. K, Representative images of α‐syn and Aβ double immunofluorescence staining in APP/PS1::pink1+/‐ mice at 6, 9, and 12 months old. L, Statistical analysis of the co‐localization of α‐syn puncta and Aβ plaques for the APP/PS1::pink1‐/‐ at 5 to 6 months old and the APP/PS1::pink1+/‐ mice at 6, 9, and 12 months old. N = 6 for APP/PS1::pink1‐/‐ mice and 3 for APP/PS1::pink1+/‐ mice at each age. Mean ± SD, *p < 0.05, ***p < 0.001, compared to APP/PS1::pink1+/‐ mice at 6‐month‐old, ^## P < 0.01, compared to APP/PS1::pink1‐/‐ mice, one‐way ANOVA. Detailed mouse information is available via mouse identification numbers in Table S3 in supporting information. α‐syn, α‐synuclein; Aβ, amyloid beta; ANOVA, analysis of variance; DAPI, 4′,6‐diamidino‐2‐phenylindole; DLB, dementia with Lewy bodies; LN, Lewy neurite; SD, standard deviation.

3.5. Model mice exhibit widespread Lewy pathologies

The pathologies of APP/PS1::pink1‐/‐ mice were assessed at 5 to 6 months of age, with the detailed information of each mouse given in Table S3 in supporting information. Immunohistochemistry staining revealed a lower neuron density in the cortex and hippocampal CA1 region of the model mice compared to wildtype, pink1‐/‐, and APP/PS1 mice (Figure S6 in supporting information). Double immunofluorescent staining for α‐syn and Aβ showed a number of Aβ plaques surrounded by α‐syn–positive puncta, forming a flower‐like LN structure (Figure 3H). In comparison, α‐syn–positive puncta were absent in the brains of APP/PS1 mice, even at 12 months of age, although Aβ pathology was much more severe (Figure 3H).

Assessment of the number of LNs and the co‐occurrence between LNs and Aβ plaques in APP/PS1::pink1‐/‐ brains revealed large variation and heterogeneity (Figure 3I). We analyzed 35 to 40 brain slices for each mouse, obtaining the total number of LNs with Aβ, LNs without Aβ, and Aβ plaque without LN ranged from 784 (mouse #K142) to 7364 (mouse #K412). The percentage of LNs with Aβ ranged from 38.1% (mouse #K408) to 82.7% (mouse #K11). On the other hand, 10% to 50% of the Aβ plaques had no surrounding α‐syn puncta. The LNs without Aβ comprised < 8% of the total area, which may result from imperfect slicing of Aβ plaque from the center of LNs. Mapping LNs to the mouse brain atlas revealed predominant localization in the cortex, followed by hippocampal dentate gyrus (DG) and molecular layer (ML; Figure 3J; Figure S7 in supporting information). Such distribution of Lewy pathology observed in the model mice closely resemble those observed in DLB patients, but differ from those observed in other synucleinopathies. ³ , ¹³ , ¹⁴ , ⁵⁶

High levels of α‐syn staining were also observed in neurons (NeuN positive), which included excitatory (CamKII‐positive), dopaminergic (TH‐positive) neurons, and cholinergic neurons (Chat‐positive), forming intracellular puncta characteristic of LBs in several APP/PS1::pink1‐/‐ mice (Figure S8A in supporting information). Moreover, the mouse #K120, which died at 112 days old, exhibited a large number of neurons containing LBs throughout the cortex and subcortex, primarily in cortical layers III through V (Figure S8B). Thus, the increase in α‐syn pathology is not specific to any single neuronal subtype but is a general feature across different neuronal populations. On the other hand, immunofluorescence staining against GFAP (an astrocyte marker) and Iba1 (a microglia marker) showed heightened glial responses surrounding LNs and Aβ plaques in APP/PS1::pink1‐/‐ mice compared to APP/PS1 mice (Figure S9A–D in supporting information).

To further assess how PINK1 deficiency plays into disease progression, we examined Lewy pathology in APP/PS1::pink1+/‐ heterozygote mice. With one allele of Pink1 intact, LN formation was significantly delayed. Compared to 5‐ to 6‐month‐old APP/PS1::pink1‐/‐ mice, α‐syn–positive puncta surrounding Aβ plaques in age‐matched APP/PS1::pink1+/‐ mice were sparser, only reaching comparable levels at 9 to 12 months of age (Figure 3K). Importantly, mice with this genetic background allowed us to better assess the age‐dependent development of Lewy pathology: the average percentage of structures with α‐syn and Aβ co‐occurrence was 16.4 ± 3.0%, 37.03 ± 1.4%, and 85.4 ± 4.8% at 6, 9, and 12 months old, respectively, in APP/PS1::pink1+/‐ mice, but was already 61.1 ± 14.5% at 5 to 6 months in APP/PS1::pink1‐/‐ mice at 5 to 6 months old (Figure 3L). Therefore, p ink1 heterozygosity delays disease progression, further confirming a negative correlation between PINK1 level and activity and Lewy pathology.

3.6. Further characterization of Lewy pathology in APP/PS1::pink1‐/‐ mice

To assess the physical properties of α‐syn proteins in LBs and LNs, we used an antibody specific for the aggregated form of α‐syn. We observed α‐syn–positive LNs surrounding the damaged neurons (indicated by NeuN‐positive clumps) and α‐syn–positive LBs within neurons, though detailed features varied depending on brain slices and mice (Figure 4A). Furthermore, pretreatment of brain slices with proteinase K prior to immunohistochemistry revealed that the α‐syn signal largely disappeared in wildtype, pink1‐/‐, and APP/PS1 mice, but remained mostly intact in APP/PS1::pink1‐/‐ mice, indicating the presence of proteinase K‐resistant, insoluble α‐syn aggregates (Figure 4B).

FIGURE 4.

The characteristic of Lewy pathology in APP/PS1::pink1‐/‐ mouse brains. A, Double immunofluorescence staining for aggregated α‐syn and neuronal marker NeuN. B, Immunohistochemical staining of α‐syn with and without proteinase K treatment, revealing the aggregated form of α‐syn. C, TEM image of a LN with abundant lysosomes and mitochondria; the right panel shows a higher magnification of the boxed region. D, TEM image of a LN with an enlarged terminal connected to a neurite; magnified images for the boxed regions show the enlarged terminal with abnormal organelles and vesicles (left) and the neurite with abnormal mitochondria, distorted neurofilaments, and proteinaceous materials (right). E, TEM image of a LB with abnormal mitochondria, lysosomes, membrane structures, and tubulovesicular structures, with the right panel showing a higher magnification of the boxed region. F, The legend for the symbols in TEM images in (C)–(E). Brain samples for TEM were taken from the parietal cortex of two APP/PS1::pink1‐/‐ mice (mouse #K556 and #K928), with the presence of LNs and LBs confirmed by immunofluorescence staining against α‐syn and Aβ antibodies (Figure S7 in supporting information). Statistical analysis of mitochondrial length (G) and area (H) outside or inside LNs. n = 50 mitochondria from two APP/PS1::pink1‐/‐ mice. **P < 0.01, t test. I–N, Double (or triple) immunofluorescence staining for Ser129 phosphorylated α‐syn (pS129‐α‐Syn) and Aβ (I), α‐syn, Aβ, and neurofilament light chain, (J), α‐Syn and βIII‐tubulin (K), α‐syn and MAP2 (L), α‐syn and tau (M), and α‐syn and phosphorylated tau (p‐tau, N). Detailed mouse information is available via mouse identification numbers in Table S3 in supporting information. α‐syn, α‐synuclein; Aβ, amyloid beta; DAPI, 4′,6‐diamidino‐2‐phenylindole; LB, Lewy body; LN, Lewy neurite; TEM, transmission electronic microscopy.

LBs and LNs are known to exhibit complex ultrastructure, often containing misfolded proteins and organelle remnants. ⁵⁷ , ⁵⁸ After immunofluorescence confirmation of widespread LBs and LNs in mouse #K556 and #K928 (Figure S10A,B in supporting information), we performed transmission electron microscopy (TEM) to examine the ultrastructural characteristics of LBs and LNs in the neocortex. These LNs and LBs displayed abnormal organellar content and structural disorganization, notably including degenerated mitochondria and densely packed, disordered neurofilaments (Figure 4C‐F). LNs were specifically localized within swollen axons or neurites (Figure 4D). Quantitative analysis revealed that mitochondria within LNs were significantly elongated compared to those outside LNs (Figure 4G), although their cross‐sectional area was similar (Figure 4H). These findings indicate that LBs and LNs in APP/PS1::pink1‐/‐ mice are structurally heterogeneous and contain impaired mitochondria, resembling the ultrastructural features observed in human PD brains. ⁵⁸

S129 phosphorylation of α‐syn has been considered a Lewy pathology marker. ⁵⁹ Using an antibody specific for S129‐phosphorylated α‐syn, we observed numerous puncta surrounding Aβ plaques in Aβ‐positive neurons of the APP/PS1::pink1‐/‐ mouse brains (Figure 4I). In contrast, pS129 signal was barely detectable surrounding Aβ plaques in APP/PS1 brains. Additionally, neither LBs, LNs, Aβ, nor pS129 α‐syn were detectable in wildtype or pink1‐/‐ mice, even at 18 months (Figure 4I; Figure S11 in supporting information).

Further immunofluorescence analysis revealed partial co‐localization between α‐syn and distorted neurofilament light chain (NfL, Figure 4J) and co‐occurrence between α‐syn and β3‐tubulin (Figure 4K), both of which provide structural support for neuronal processes. On the other hand, no co‐localization nor co‐occurrence was observed between α‐syn and MAP2, a dendritic marker (Figure 4L; Figure S12A, B in supporting information). In APP/PS1::pink1‐/‐ mice, α‐syn–positive puncta were found partially co‐localized with tau puncta, an axonal marker, whereas co‐localization was not observed in APP/PS1 mice (Figure 4M). Moreover, a small subset of LNs co‐localized with phosphorylated tau (p‐tau, at residue T181) puncta, suggesting tau‐related pathology (Figure 4N). These findings are consistent with GO analysis with “distal axon,” a top‐ranked term among DEGs between pink1 _L and controls (Figure S3), and supports the established view that Aβ and α‐syn pathology precede tau pathology. ¹⁸

3.7. Model mice have impaired peripheral nervous system

Mirroring the sudden cardiac events often observed in DLB patients, we observed a high rate of sudden death among APP/PS1::pink1‐/‐ mice during breeding. Of the 47 mice monitored over a 150‐day period post‐birth (excluding three sacrificed for experimental purposes), 30 either died suddenly or were euthanized due to severe weight loss (Table S3). This accounts for a 68.2% mortality rate in APP/PS1::pink1‐/‐ mice by 150 days, in stark contrast to the 100% survival of all wild‐type, pink1‐/‐, and APP/PS1 mice at this time point (Figure 5A). Additionally, three APP/PS1::pink1‐/‐ mice developed severe rectal prolapse, and three exhibited severe bladder urine retention discovered upon dissection (Figure 5B). Fecal water content also significantly decreased in APP/PS1::pink1‐/‐ mice, indicating constipation (Figure 5C).

FIGURE 5.

Impaired peripheral nervous system in APP/PS1::pink1‐/‐ mice. A, High death rate of APP/PS1::pink1‐/‐ mice were observed within 150 days, while all other genotypes had survived. B, Images demonstrating urine retention and severe rectal prolapse in APP/PS1::pink1‐/‐ mice. C, Percentage of fecal water content in 5‐month‐old mice, n = 11–15, mean ± SD, *P < 0.05, **P < 0.01 (one‐way analysis of variance). D, Hematoxylin and eosin (H&E) and Masson's trichrome staining of a heart with dilated cardiomyopathy (mouse #K412), with a wild‐type heart as control. E, Double immunofluorescence staining for α‐syn and β3‐tubulin (neurofilament marker), revealing Lewy pathology in the heart. F, H&E staining of ileum from wildtype, pink1‐/‐, APP/PS1, and APP/PS1::pink1‐/‐ mice. G, Double immunofluorescence staining for α‐syn and β3‐tubulin demonstrating Lewy pathology in the ileum. H, H&E staining of colon from wildtype, pink1‐/‐, APP/PS1, and APP/PS1::pink1‐/‐ mice. I, Double immunofluorescence staining for α‐syn and β3‐tubulin revealing Lewy pathology in the colon. Detailed mouse information is available via mouse identification numbers in Table S3 in supporting information. α‐syn, α‐synuclein; DAPI, 4′,6‐diamidino‐2‐phenylindole; SD, standard deviation.

Mouse #K412 exhibited severe weight loss at 4 months old, and was found to have cardiac hypertrophy upon sacrifice, as characterized by enlarged ventricles and thinner ventricular walls compared to the wildtype control (Figure 5D). Masson's trichrome staining revealed fibrosis in the heart tissue (Figure 5D). Double staining for α‐syn and β3‐tubulin showed beaded α‐syn–positive deposits that are co‐localized with β3‐tubulin in the heart tissue. Lewy pathology in the peripheral nervous system likely contributes to cardiac hypertrophy resulted from denervation (Figure 5E).

The reduced fecal water content also likely results from denervation of the intestinal plexus. Hematoxylin and eosin staining revealed high heterogeneity in mucosal thickness, villus length, villus spacing, and crypt depth in APP/PS1::pink1‐/‐ mice, as opposed to the consistent features seen in wildtype, pink1‐/‐, and APP/PS1 mice (Figure 5F). Co‐localization of α‐syn and β3‐tubulin puncta were observed in the intestinal wall of APP/PS1::pink1‐/‐ mice but not in other groups of mice (Figure 5G).

Furthermore, colon morphology was found heterogeneous in APP/PS1::pink1‐/‐ mice, with varying crypt thickness and goblet cell numbers (Figure 5H). In contrast, wild‐type, pink1‐/‐, and APP/PS1 mice had colons with uniform crypt depths and goblet cell linings. Co‐localization of beaded α‐syn and β3‐tubulin puncta was only observed in the colon walls of APP/PS1::pink1‐/‐ mice (Figure 5I).

These findings indicate that Lewy pathology‐induced neurodegeneration also impacts the peripheral nervous system, contributing to the various sympathetic and parasympathetic symptoms, ultimately leading to organ failure and sudden death. Degenerative changes in the peripheral nervous system may account for the heterogeneity of peripheral clinical manifestations of DLB patients. The relationship between central and peripheral Lewy pathology—whether one precedes the other—remains a critical question in the field. Our APP/PS1::Pink1‐/‐ mice model provides a valuable tool for investigating this issue, offering a unique opportunity to study how systemic Lewy pathology develops and contributes to the heterogeneous phenotypes of DLB.

3.8. Model mice exhibit phenotypes similar to clinical presentations of DLB

DLB patients presented with a range of heterogeneous clinical features that can be categorized into six groups: cognitive impairment, sleep problems, motor dysfunction, psychiatric symptoms, sympathetic nervous system dysfunction, and parasympathetic nervous system dysfunction (Table 1). Among the 47 recruited patients with DLB, the most common symptoms within each group were memory decline (91.49%) and fluctuating cognitive impairment (46.81%); insomnia (42.55%) and REM sleep behavior disorder (34.04%); motor slowness (40.43%) and Parkinsonism symptoms (40.43%); hallucinations (70.21%) and delirium (38.30%); constipation (19.15%) and palpitations (17.02%); and urinary incontinence (17.02%) and salivation (12.77%).

TABLE 1.

Phenotype/presentation comparison between DLB patients and APP/PS1::pink1‐/‐ mice.

Clinical manifestation	DLB patients	Prevalence (47 patients)	APP/PS1::pink1−/− mice	Prevalence (47 mice)
Cognition impairment	Memory decline	43 (91.49%)	Short‐term memory	Impairment in Y‐maze test and fear conditioning test
	Fluctuation in cognition impairment	22 (46.81%)
Sleep problems	RBD	16 (34.04%)	NA	NA
	Insomnia	20 (42.55%)
	Circadian rhythm disturbance	3 (6.38%)
Motor dysfunction	Abnormal gait	9 (19.15%)	NA	NA
	Motor slowness	19 (40.43%)	Atonia	1/47 (2.13%)
	Parkinsonism symptoms	19 (40.43%)	Parkinsonism symptoms	Impairment in pole climbing and rotarod tests
	Fall	10 (21.28%)	NA	NA
Psychiatric symptoms	Delirium	18 (38.3%)	NA	NA
	Hallucination	33 (70.21%)	NA	NA
	Delusion	18 (38.3%)	NA	NA
	Capgras syndrome	4 (8.51%)	NA	NA
	Anxiety	8 (17.03%)	NA	NA
	Apathy	7 (14.89%)	NA	NA
	Decreased interest	12 (25.53%)	NA	NA
	Feeling upset	5 (10.64%)	NA	NA
	Irritation	13 (27.66%)	Irritation	1/47 (2.13%)
	Aggressive behavior	7 (14.89%)	Aggressive behavior	1/47 (2.13%)
Sympathetic nervous system dysfunction	Hypertension	6 (12.77%)	Dilated cardiomyopathy	1/20 (5%)
	Palpitation	8 (17.02%)	NA	NA
	Sweating	4 (8.51%)	Skin diseases	10/47 (21.3%)
	Shortness of breath	7 (14.89%)	NA	NA
	Constipation	9 (19.15%)	Constipation	Decreased water contents in feces
	NA	NA	Urine retention	3/20 (15%)
	NA	NA	Prolapse of anus	3/47 (6.38%)
	NA	NA	Splenomegaly	2/20 (10%)
Parasympathetic nervous system dysfunction	Hypotension	1 (2.13%)	NA	NA
	Salivation	6 (12.77%)	NA	NA
	Urinary incontinence	8 (17.02%)	NA	NA
Shared features	✓Cognition impairments ✓Motor dysfunctions ✓Irritation and aggression in psychiatric symptoms ✓Dysfunction of sympathetic nervous system and parasympathetic nervous system, which was observed in gastrointestinal system and cardiovascular system

Abbreviations: DLB, dementia with Lewy bodies; NA, not applicable; RBD, rapid eye movement sleep behavior disorder.

Mirroring the clinical manifestations, APP/PS1::pink1‐/‐ mice exhibited cognitive impairment (as assessed by Y‐maze and fear conditioning), motor dysfunction (pole climbing and rotarod tests), and peripheral nervous system dysfunction (including dilated cardiomyopathy, skin diseases, urine retention, decreased fecal water content, rectal prolapse, and splenomegaly), which is summarized in Table 1. Recapitulating these clinical manifestations, the model mice we have developed represent a valuable tool for investigating the complex pathophysiology of DLB.

3.9. PINK1 directly phosphorylates α‐syn and mitigates α‐syn aggregation

To investigate how PINK1 deficiency contributes to Lewy pathology, we examined protein aggregation, autophagy, and mitophagy in the neocortex of 4‐ to 6‐month‐old mouse brains. Ub levels were significantly elevated in both the soluble and insoluble fractions of APP/PS1::pink1‐/‐ mice compared to wildtype, pink1‐/‐, and APP/PS1 mice (Figure S13A,B in supporting information), indicating increased protein aggregation. In contrast, α‐syn protein levels remained unchanged across all four genotypes in both fractions (Figure S13C,D), suggesting that Lewy pathology is not driven by changes in α‐syn abundance.

To assess autophagic activity, we measured the levels of LAMP1 and LC3‐II, which were comparable among the four groups (Figure S13E,F). Similarly, p62 levels in both soluble and insoluble fractions did not differ significantly (Figure S13G,H), indicating that bulk autophagic flux remains largely intact.

Given that the PINK1‐Parkin pathway mediates mitophagy via ubiquitination of mitochondrial outer membrane proteins, we also evaluated mitochondrial integrity. The levels of mitochondrial markers TOM20 and TIM23 were comparable across all groups (Figure S13I,J), and ubiquitination of mitochondrial proteins in purified mitochondrial fractions, normalized to TOM20, also showed no significant differences (Figure S13K). These findings suggest that while proteostasis is disrupted in APP/PS1::pink1‐/‐ mice, neither autophagy nor mitophagy is notably impaired at this age. This observation stands in contrast to previous studies reporting autophagic impairment after injection of α‐syn preformed fibrils. ¹⁹

We performed in vitro phosphorylation of α‐syn using a recombinantly prepared PINK1 and detected a phosphorylated form of α‐syn with an 80 Da mass increase (Figures 6A, S13A,B). PINK1‐phosphorylated α‐syn showed significantly reduced aggregation propensity, either spontaneously or in the presence of cross‐seeding Aβ monomers, Aβ fibrils, or α‐syn fibrils, as evaluated by thioflavin‐T staining and dot blot analyses (Figure 6B,C).

FIGURE 6.

PINK1 phosphorylates α‐syn and mitigates its aggregation. A, Electrospray ionization mass spectrometry (ESI‐MS) of α‐syn before and after PINK1 phosphorylation at various charge states. Deconvoluted mass spectra are displayed on the right. B, Thioflavin T (ThT) fluorescence‐based assessment of α‐syn and phosphorylated α‐syn (p‐α‐syn) aggregation kinetics, in the absence or presence of Aβ monomer, Aβ fibrils, and α‐syn fibrils. Mean ± SD, n = 3. *p < 0.05, ^** p < 0.01, and ***p < 0.001 compared to time 0; ^# p < 0.05, ^## p < 0.01, and ^### p < 0.001 compared between α‐syn and p‐α‐syn, two‐way ANOVA followed by Tukey post hoc test. C, Dot blot analysis of α‐syn and p‐α‐syn aggregation after 4 days of shaking, using an antibody specific for α‐syn aggregates. mean ± SD, n = 6, *p < 0.05, **p < 0.01, and ***p < 0.001, unpaired t test. D, AlphaFold‐multimer model showing α‐syn docked into PINK1 substrate crevice, with α‐syn T44 positioned near PINK1 active‐site residue D362. Key residues for ATP interaction (PINK1 K219), known disease related mutation (PINK1 M318), and cross‐linked residues (PINK1 K380 and α‐syn K12) are also shown. E, Western blot analysis of co‐immunoprecipitation by anti‐PINK1 antibody, anti‐α‐syn antibody, or anti‐IgG antibody (as a negative control) in APP/PS1 mouse brain sample. F, Western blot analysis of pT44‐α‐syn in the cortex of mouse brains. Mean ± SD, n = 5, *P < 0.05, **P < 0.01, and one‐way ANOVA. G, Immunofluorescence analysis for T44‐phosphorylated α‐syn and Aβ in mouse brains, with representative images and the statistical analysis of the intensities of T44‐phosphorylated α‐syn. mean ± SD, n = 28 from 3 mice, ***P < 0.001, one‐way ANOVA. H, Double immunofluorescence staining for T44‐phosphorylated α‐syn and Aβ in human brain cingulate gyrus samples. I, Correlations between the percentage of PINK1‐positive (PINK1+) neurons, pUb intensity, and pT44‐α‐Syn intensity for immunofluorescent stained human brain samples. Representative images of PINK1 and pUb staining are shown in Figure 2. The R ² and p value are given in each panel. α‐syn, α‐synuclein; Aβ, amyloid beta; ANOVA, analysis of variance; IgG, immunoglobulin G; PINK1, PTEN‐induced kinase 1; pUb, phospho‐ubiquitin; SD, standard deviation.

Mass spectrometry (MS) analysis identified multiple phosphorylation sites in α‐syn, including T22, T44, T81, and S87. To obtain the most relevant PINK1 phosphorylation site, we generated antibodies targeting the variously phosphorylated α‐syn. Western blotting confirmed that PINK1 phosphorylation produced a mixture of phosphorylated forms in vitro (Figure S14C). The pT44‐α‐syn exhibited the highest abundance of matched peptide spectra in the MS analysis (Figure S14D). Accordingly, an AlphaFold complex model showed that the N‐terminal region of α‐syn adopts an anti‐parallel β‐structure that can be accommodated between the two lobes of PINK1 (Figures 6D; S15A in supporting information). This model places the Cα atom of T44 in α‐syn within 8 Å of the active‐site residue D362 in PINK1, which was further validated by a cross‐linking between PINK1 residue K380 (corresponding to residue R408 in human PINK1) and α‐syn residue K12 (Figures 6D, S15B).

Co‐immunoprecipitation using anti‐α‐syn or anti‐PINK1 antibodies, followed by western blotting, further confirmed a direct interaction between α‐syn and PINK1 in the neocortex of APP/PS1 mice (Figure 6E). Nevertheless, we could not reproduce this interaction in HEK293 cells (Figure S16 in supporting information) as reported previously, ⁶⁰ suggesting that PINK1 activity and its interaction with α‐syn may depend on cell type and tissue context.

Western blot analysis revealed a significant reduction in overall pT44 α‐syn levels in both pink1‐/‐ and APP/PS1::pink1‐/‐ mice compared to wild‐type and APP/PS1 mice (Figure 6F). Immunofluorescence staining showed that pT44 α‐syn was enriched around Aβ plaques in APP/PS1 mice but was largely absent in APP/PS1::pink1⁻/⁻ mice (Figure 6G). Additionally, phosphorylation at T22 was elevated in APP/PS1 mice relative to wild type; however, the pT22 α‐syn signal did not co‐localize with Aβ plaques and remained unchanged upon Pink1 knockout (Figure S17A in supporting information). Phosphorylation at T81 was also observed around Aβ plaques in APP/PS1 mice and showed only a modest reduction in APP/PS1::pink1‐/‐ mice, suggesting involvement of other kinases (Figure S17B). Due to a Ser‐to‐Asn substitution at residue 87 in murine α‐syn, immunofluorescence analysis of pS87 was not performed.

To further investigate the role of T44 phosphorylation in α‐syn aggregation, we assessed aggregation kinetics using thioflavin‐T fluorescence assays with wild‐type α‐syn and the phosphomimetic mutant α‐syn/T44E. Introduction of the T44E mutation significantly delayed α‐syn aggregation under spontaneous conditions as well as in the presence of cross‐seeding Aβ monomers, Aβ fibrils, or α‐syn fibrils (Figure S17C).

As a further support to this mechanism, analysis of post mortem human brain tissue revealed the lowest levels of pT44 α‐syn in DLB brains and the highest levels in AD brains (donors F and G; Figure 6H). Notably, across all human samples analyzed, the percentage of PINK1‐positive neurons, the intensity of pUb staining, and the levels of pT44 α‐syn were strongly correlated (Figure 6I). Together, these results suggest that PINK1 exerts a protective effect by phosphorylating α‐syn at T44, thereby attenuating its aggregation in the presence of Aβ pathology.

4. DISCUSSION

Our findings reveal a novel mechanism for DLB pathogenesis: PINK1 levels and activity typically increase in response to Aβ accumulation, leading to specific phosphorylation of α‐syn that counteracts Aβ‐induced cross‐seeding of α‐syn aggregation. Conversely, PINK1 deficiency permits the development of Lewy pathology in DLB, in the presence of Aβ burden. Although the link between PINK1 deficiency and AD pathology, or protein aggregation in a broader sense, has been previously explored, the focus has been on impaired mitochondrial quality control and subsequent neuronal damage or the induction of the nuclear factor κB pathway. ⁶¹ , ⁶² , ⁶³ In contrast, our work provides a more direct mechanistic link, revealing a protective function by which PINK1 directly modifies α‐syn to prevent aggregation.

This hypothesis was supported by multiple lines of evidence. First, previous studies, ³ , ¹⁵ , ¹⁶ along with our current findings, demonstrate the frequent co‐occurrence and co‐localization of Aβ and Lewy pathologies in DLB patients, underscoring Aβ’s role in DLB pathogenesis. Second, Aβ has been shown to cross‐seed α‐syn aggregation, which has been demonstrated in animal models through the injection of pre‐formed fibrils or protein overexpression. ¹⁹ , ²⁷ , ⁶⁴ Third, the accumulation of Aβ plaques can lead to mitochondrial dysfunction and proteasomal inhibition, which then elevate the levels of full‐length PINK1 and its cleaved cytoplasmic form, sPINK1, as shown previously. ³¹ , ⁴² , ⁵³ , ⁶⁵ Fourth, PINK1 deficiency is a recognized risk factor for early‐onset PD, and emerging evidence suggests that its role in Lewy pathology may extend beyond the well‐established function in mitophagy. ³⁹ , ⁴⁰ , ⁴¹ , ⁴² Fifth, pUb levels negatively correlate with Lewy pathology, in which pUb granules were found sparsely localized around mature LBs and did not colocalize with S129‐phosphorylated α‐syn. ⁶⁶ Consistent with these previous reports, our data showed reduced PINK1 expression and activity in DLB patients. Additionally, transcriptomic data analysis revealed a strong association between decreased PINK1 expression and DLB progression (Figure 2). Collectively, this evidence suggest a neuroprotective role for PINK1 in counteracting Aβ‐induced α‐syn aggregation, thereby mitigating the development of DLB.

To validate this hypothesis, we developed the APP/PS1::pink1‐/‐ transgenic mouse model. This model recapitulates key clinical features, including motor dysfunctions and dementia, and exhibits characteristic DLB pathologies, namely, widespread co‐occurrence of Lewy bodies and Aβ plaques. Ultrastructural analysis of LBs and LNs in these mice revealed remarkable similarities to the Lewy pathology observed in PD patients. ⁵⁸ Furthermore, the mice exhibit phenotypes and pathologies associated with autonomic nervous system dysfunction, mirroring the heterogeneous clinical manifestations observed in human DLB patients. In comparison, APP/PS1::pink1+/‐ heterozygote mice exhibited a delayed progression of these symptoms, owing to the partial rescue of PINK1 activity. Taken together, the APP/PS1::pink1‐/‐ mouse serves as a comprehensive model of DLB and, to our knowledge, is the first to fully recapitulate the diverse manifestations of this complex disease. It should also be noted that, unlike previous models of α‐syn aggregation and Lewy pathology, ¹⁹ , ²¹ , ²⁵ , ²⁷ the current animal model was established without altering α‐syn concentration or injection of pre‐formed fibrils. Significantly, α‐syn aggregation emerges spontaneously at endogenous α‐syn levels beyond the substantia nigra, while affecting both the central and peripheral nervous systems.

Phosphorylation, a common post‐translational modification of α‐syn, has been shown to modulate α‐syn aggregation. Although α‐syn contains up to 18 potential phosphorylation sites, ⁶⁷ many of the responsible kinases and the functional consequences of phosphorylation remain unclear. ⁶⁸ Among the different sites, phosphorylation at S129 is the most extensively studied and correlates positively with α‐syn aggregation severity in PD, making it a potential biomarker. ⁵⁹ Yet, the identified kinases for S129 appear to have limited direct impact on α‐syn aggregation. ⁶⁹ , ⁷⁰ Additionally, phosphorylation at S87 has been shown to reduce α‐syn aggregation, ⁵⁹ and phosphorylation at Y39 either promotes or reduces α‐syn aggregation, ⁶⁸ , ⁷¹ highlighting the complexity of α‐syn phosphorylation.

Our study demonstrates that recombinant PINK1 phosphorylates multiple sites on α‐syn in vitro and identifies T44 as the primary phosphorylation site in vivo. Notably, pT44 α‐syn is specifically enriched around Aβ plaques in APP/PS1 mice but absent in APP/PS1::pink1‐/‐ mice. These findings suggest that unphosphorylated α‐syn is more prone to aggregation in response to Aβ, ultimately contributing to the development of Lewy pathology. Consistent with our observations in mice, post mortem human brain samples revealed elevated levels of pT44 α‐syn in AD patients and reduced levels in DLB patients, indicating a potential deficiency in PINK1‐mediated protective phosphorylation in DLB. Indeed, DLB patients harboring missense mutations near PINK1 active sites, such as M318L, ⁴³ potentially impacting PINK1's enzymatic activity, ⁷² , ⁷³ has been reported. It is important to note that the genetic backgrounds underlying Aβ pathology in the human brain samples analyzed here are not fully characterized, though they likely represent sporadic AD cases. Although we used the APP/PS1 transgenic model to ensure experimental reproducibility, the mechanism we have uncovered—Aβ‐induced α‐syn aggregation mitigated by PINK1‐dependent phosphorylation—likely extends beyond familial AD to encompass sporadic cases involving Aβ accumulation and cross‐seeding processes that give rise to Lewy pathology.

Previous studies using acute manipulation have linked α‐syn to autophagy. For example, injecting pre‐formed α‐syn fibrils into the cortex of AD mice reportedly increases p62 levels, suggesting impaired autophagy. ¹⁹ Similarly, overexpression of both PINK1 and α‐syn in HEK293 cells led to elevated LC3 levels, leading authors to propose that PINK1 facilitates α‐syn degradation by promoting autophagy. ⁶⁰ In stark contrast, our study detected no significant changes in LAMP1, p62, TOM20, or TIM23 levels (Figure S13), indicating that basal autophagic flux remains unchanged in our model. Such a difference is likely because we did not artificially manipulate α‐syn levels through overexpression or fibril injection, and may have to do with distinct PINK1 levels in mouse cortex versus cultured cells. These findings underscore the roles of PINK1 beyond canonical autophagy and highlight the advantages of our model in investigating the endogenous mechanisms underlying DLB pathogenesis without the confounding effects of acute, non‐physiological manipulations.

Although PINK1 has primarily been studied in relation to PD, it is important to note that PINK1‐knockout animals typically require additional stressors or pathological insults to develop α‐syn aggregation, PD‐like pathology, or associated phenotypes. ³⁹ , ⁴⁰ , ⁴¹ Our study demonstrates that in the context of PINK1 deficiency, Aβ aggregation drives the development of Lewy pathology, which spatially correlates with Aβ distribution. This suggests that PINK1 deficiency establishes a genetic predisposition to disease. Furthermore, our findings suggest that PINK1's functions extend beyond initiating mitophagy and maintaining mitochondrial homeostasis, as previously suggested. ⁴² , ⁵³ Critically, we show here that PINK1 directly phosphorylates α‐syn, primarily at T44, a modification that prevents α‐syn aggregation and mitigates the formation of Lewy pathology.

Although our study offers valuable insights into DLB pathogenesis, certain limitations should be noted. The observed clinical heterogeneity in DLB patients presents a major challenge, a complexity also mirrored in the phenotypic variability of our genetically identical, co‐housed mice. Several factors likely contribute to this. First, α‐syn aggregation propensity is probably influenced by proteins and factors beyond Aβ, such as tau. ¹⁹ Second, while PINK1 primarily targets T44, it might phosphorylate other sites on α‐syn, though to a lesser extent, and T44 itself could be targeted by other kinases, further contributing to variability. Additionally, PINK1's role in proteasomal degradation ³⁰ , ³³ might affect Aβ and α‐syn clearance, thereby impacting disease progression. Last, the heightened glial response around Aβ plaques in our DLB model mice compared to AD mice, as shown in Figure S9 and prior reports, ⁷⁴ , ⁷⁵ suggests that neuroinflammation could be involved in DLB pathogenesis. Future rescue experiments targeting PINK1, encompassing both cell‐autonomous and non–cell‐autonomous mechanisms, will be critical to clarify neuroinflammation's involvement.

In summary, we have uncovered a novel mechanism central to DLB pathogenesis: the interplay between Aβ pathology and PINK1‐mediated direct phosphorylation of α‐syn. Our findings establish PINK1 deficiency, at either the gene or protein level, as a critical risk factor for DLB. The APP/PS1::pink1‐/‐mouse model not only validates this mechanism but also provides a vital tool for future research into this complex disease. Importantly, our study suggests potential avenues for early diagnosis of DLB, such as screening for PINK1 gene mutations, assessing PINK1 kinase activity, and quantifying α‐syn T44 phosphorylation in patients with coexisting Aβ pathology. Ultimately, this work provides profound insights into DLB pathogenesis and opens new avenues for developing both early diagnostic tools and therapeutic interventions, with PINK1 activators as promising therapeutic candidates.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the supporting information.

CONSENT STATEMENT

The study used post mortem brain tissue collected at the Netherlands Brain Bank (NBB), Netherlands Institute for Neuroscience, Amsterdam (Project number 1613) and the Chinese Brain Bank (CBB), part of the National Health and Disease Human Brain Tissue Resource Center, Hangzhou, China (Project number CN20240454). Consent for autopsy for research purposes was provided at the time of death by the next of kin in accordance with a protocol approved by the Committee for Oversight of Research and Clinical Training Involving Decedents.

The data of prevalent clinical features were de‐identified. The study was approved by the ethics committee of Peking University Institute of Mental Health (Sixth Hospital). Written informed consent was obtained from all participants and their family caregivers.

Supporting information

Supporting information

Supporting information

Supporting information

Gao T‐Y, Wang X‐Z, Xie Y‐H, et al. PINK1 deficiency permits the development of Lewy body dementia with coexisting Aβ pathology. Alzheimer's Dement. 2025;21:e70730. 10.1002/alz.70730