Biomarker for Parkinson's disease diagnosis and application thereof

Abstract

The invention relates to a biomarker for Parkinson's disease diagnosis and application thereof, and belongs to the technical field of biological detection. The biomarkers comprise one or more of NDUFC1 (Non-Dulbecco Fluorescent Receptor 1), CHGA (Chrysanthemum Glycolic Acid), CCL5 and ST3GAL6. The biomarker can play an important role in mechanism research of Parkinson's disease. According to the invention, a group of Parkinson's disease diagnosis biomarkers are found in neuron-derived exosomes and planetary glial cell-derived exosomes of a Parkinson's disease clinical queue by combining machine learning with a proteomics method, and the AUC value reaches 0.967.

Description

Technical Field

[0001] The present invention relates to a biomarker for the diagnosis of Parkinson's disease and its application, belonging to the technical field of biological detection. Background Art

[0002] Parkinson's disease is a chronic and progressive neurodegenerative disease characterized by the degeneration and death of dopaminergic neurons in the substantia nigra pars compacta of the midbrain and the appearance of Lewy bodies formed by the misfolding of α-synuclein in the cytoplasm of the remaining dopaminergic neurons in the brain, which will lead to a decrease in dopamine levels in the patient's brain and induce a series of neurological lesions. Parkinson's disease has strong heterogeneity in terms of onset age, clinical manifestations, progression speed, and treatment response. For patients with a typical medical history, typical asymmetric motor symptoms, clear and significantly effective dopaminergic drug treatment, no atypical features, and a clear etiology, the diagnosis of Parkinson's disease is relatively simple; however, in actual clinical diagnosis, the accuracy rate of Parkinson's disease diagnosis is only about 80%. Especially in the early stage of the disease, the clinical features of patients are not typical, and the response of some patients to dopaminergic drug treatment is also unclear. Neurons have already suffered irreversible damage when clinical symptoms appear. Therefore, finding biomarkers that can reflect the core elements in the occurrence process of Parkinson's disease and using them as tools for early disease identification and differential diagnosis is of great significance for the prevention and treatment of Parkinson's disease, improving the quality of life of patients, and reducing the social medical burden.

[0003] As a type of extracellular vesicles generated by the endosomal pathway, exosomes can participate in processes such as cell communication, immune response, and cell migration in the body and are widely present in various body fluids. At the same time, because exosomes contain various active components derived from the source cells and can cross the blood-brain barrier bidirectionally, after recognizing specific receptors on the surface of exosomes in various body fluids, exosomes from different cell sources can be effectively isolated. Especially in the central nervous system, neurons, astrocytes, oligodendrocytes, etc. can produce and release exosomes, namely neurogenic exosomes, which can be captured by specific antibodies. Therefore, as a novel research object, neurogenic exosomes have been proven to play a key role in the pathology of various neurodegenerative diseases, and multiple teams have carried out research on biomarkers of neurodegenerative diseases around them. In 2019, Wang et al. used the ELISA method to simultaneously analyze the levels of DJ-1 and α-Synuclein in the blood and neuron-derived exosomes in the blood of 39 Parkinson's disease patients and found that compared with the control group, the levels of DJ-1 and α-Synuclein in neuron-derived exosomes in the Parkinson's disease group were significantly increased, but no such change was found in the blood. In 2020, Jiang et al. carried out a large clinical cohort study on neuron-derived exosomes in the blood of Parkinson's disease patients using mass spectrometry technology and electrochemiluminescence method and found that the area under the receiver operating characteristic curve of the level of α-Synuclein in neuron-derived exosomes reached 0.86 when distinguishing Parkinson's disease patients from healthy people, the area under the curve of using α-Synuclein combined with clusterin to identify other synucleinopathies reached 0.98, and the area under the curve of identifying multiple system atrophy reached 0.96. The above results indicate that neurogenic exosomes, as a new research object, have great research value and application prospects in the discovery and screening of potential biomarkers for neurodegenerative diseases.

[0004] Proteomics technology has played an important role in the screening of biomarkers for Parkinson's disease. In 2021, Federica Anastasi et al. used proteomics methods to analyze the proteome of neuron-derived exosomes in the blood of Parkinson's disease patients, and on average identified 349 ± 38 proteins and analyzed the correlation between these proteins and Parkinson's disease, but no potential biomarkers were found.

[0005] The research on discovering Parkinson's disease biomarkers based on proteins in neurogenic exosomes in the blood is in a rapid development stage, but there are still several deficiencies at present:

[0006] (1) Currently, there is no protein database for neurogenic exosomes in Parkinson's disease patients. The Vesiclepedia website updated the vesicle protein database in September 2023. After filtering, it included a total of 3,578 human proteins, but there is no dedicated protein database for neurogenic exosomes.

[0007] (2) Most of the existing studies on Parkinson's disease diagnostic biomarkers in neurogenic exosomes are conducted through immunological methods. There are few studies on discovering Parkinson's disease progression biomarkers through neurogenic exosome proteomics analysis, and no effective clinical diagnostic biomarkers have been found yet. Summary of the Invention

[0008] In view of this, the purpose of the present invention is to provide a biomarker for the diagnosis of Parkinson's disease and its application.

[0009] To achieve the above purpose, the technical solution of the present invention is as follows.

[0010] A biomarker for the diagnosis of Parkinson's disease, the biomarker includes one or more of NDUFC1, CHGA, CCL5, and ST3GAL6.

[0011] Further, the Parkinson's disease is early Parkinson's disease.

[0012] Further, the biomarkers all originate from neurogenic exosomes; furthermore, NDUFC1 and CHGA originate from neuronal exosomes; CCL5 and ST3GAL6 originate from astrocyte exosomes.

[0013] An application of the biomarker of the present invention in the preparation of a product for the diagnosis of Parkinson's disease.

[0014] A reagent for detecting the biomarker of the present invention, the reagent is used to detect the protein expression level of the biomarker; and / or, the reagent is a reagent that specifically binds to the biomarker.

[0015] A kit, including the reagent of the present invention.

[0016] A method for constructing a protein database of neurogenic exosomes related to Parkinson's disease, the method steps include:

[0017] (1) Separating neurogenic exosome proteins from blood samples of early Parkinson's disease patients, mid - to - late Parkinson's disease patients, and healthy people respectively, and subjecting them to enzymatic hydrolysis to obtain peptide solutions respectively;

[0018] (2) Mix the obtained peptide solution to obtain a mixed peptide solution; fractionate the mixed peptide solution using a high-performance liquid chromatograph and a high pH value reversed-phase chromatographic column, and collect the fractions;

[0019] (3) Perform high-performance liquid chromatography-tandem mass spectrometry analysis on the fractions to obtain a mass spectrometry data file;

[0020] (4) Process the mass spectrometry data file to construct a protein database of neurogenic exosomes related to Parkinson's disease.

[0021] A screening method for biomarkers for the diagnosis of Parkinson's disease, the method steps include:

[0022] (1) Mass spectrometry detection: Resolubilize the peptides of neurogenic exosome proteins from early Parkinson's disease patients and healthy individuals with formic acid aqueous solution, perform high-performance liquid chromatography-tandem mass spectrometry analysis, and the mass spectrometry analysis adopts the data-independent acquisition mode (DIA) to collect the original mass spectrometry data file;

[0023] (2) Data retrieval: Import the original mass spectrometry data file into Spectronaut 18 software, retrieve the DIA data of neurogenic exosomes using the database constructed by the present invention, set Trypsin digestion, fixed modification: Carbamidomethylation (C), variable modification: Oxidation (M), and the remaining parameters are default;

[0024] (3) Bioinformatics analysis: Normalize the protein quantification results of neurogenic exosomes from Parkinson's disease patients, and perform missing value filtering and filling; use Perseus software to calculate the fold change (FC) and significance of difference (P-value) of proteins between the early Parkinson's disease patient group and the healthy individual group, and screen for differentially expressed proteins with P < 0.05 and |Log2FC| > 0.5 as the threshold;

[0025] (4) Screening for Parkinson's disease diagnostic biomarkers by machine learning: Screening for Parkinson's disease diagnostic biomarkers in neuron-derived exosomes and astrocyte-derived exosomes through machine learning methods; First, import the quantitative results of differentially expressed proteins in the healthy control group and the early Parkinson's disease group as feature data, set the number of proteins in the biomarker combination to n, where n is 2-5, and loop through all combinations; Subsequently, import the data of each combination into a logistic regression model and perform five-fold cross-validation, and execute this step 10 times in a loop to calculate 5×10 accuracy (Accuracy, ACC) values and area under the curve (Area Under Curve, AUC), and take the average of the 50 results as the ACC value and AUC value of this biomarker combination; Finally, calculate the AUC values of all biomarker combinations, sort them, and select the protein combination with the highest AUC value, which is the screened Parkinson's disease diagnostic biomarker combination.

[0026] Further, in step (1), the chromatographic gradient is 130 min, the sample loading amount is 1 μg, mobile phase A is an aqueous solution of 0.1% formic acid by volume, mobile phase B is an acetonitrile / water solution of 0.1% formic acid by volume, the volume ratio of acetonitrile to water is 20:80, and the chromatographic gradient changes as follows: 0 min - 5 min, the linear gradient of mobile phase B is from 2% - 8%; 5 min - 90 min, the linear gradient of mobile phase B is from 8% - 24%; 90 min - 110 min, the linear gradient of mobile phase B is from 24% - 32%; 110 min - 115 min, the linear gradient of mobile phase B is from 32% - 90%; 115 min - 125 min, the linear gradient of mobile phase B is from 90% - 5%; 125 min - 130 min, mobile phase B is maintained at 5%; the flow rate is 300 nL / min; during mass spectrometry analysis, the first-order mass spectrometry scanning range is 350 - 1250 m / z, the resolution is 120000, the maximum gain control is 3e6, and the maximum injection time is 60 ms; the second-order mass spectrometry resolution is 30000, the maximum gain control is 1e6, the maximum injection time is 55 ms, and the collision energy is 25.5 eV, 27.0 eV, 30.0 eV.

[0027] Further, in step (3), log2(x) is used for normalization; and / or, filter out missing values under the condition that the missing value in each group is <40%, and use a K-nearest neighbor algorithm considering truncated distribution (Truncation KNN) for missing value filling; and / or, a total of 40 differentially expressed proteins are identified in neuron-derived exosomes, including 13 up-regulated proteins and 27 down-regulated proteins; a total of 57 differentially expressed proteins are identified in astrocyte-derived exosomes, including 5 up-regulated proteins and 52 down-regulated proteins.

[0028] Beneficial effects

[0029] The present invention provides a new set of biomarkers for the clinical diagnosis of Parkinson's disease, and these biomarkers can play an important role in the mechanism study of Parkinson's disease.

[0030] The present invention has established two protein databases of neuron-derived exosomes and astrocyte-derived exosomes from patients with Parkinson's disease, including 2,247 and 1,722 proteins respectively. The establishment of these two databases provides a reference database for subsequent research on neurogenic exosomes and proteomic analysis of Parkinson's disease.

[0031] The present invention has discovered a set of diagnostic biomarkers for Parkinson's disease in neuron-derived exosomes and astrocyte-derived exosomes of a clinical cohort of Parkinson's disease by combining machine learning and proteomic methods, and the AUC value reaches 0.967. Description of the Drawings

[0032] Figure 1 It is a Venn diagram of the protein databases of neuron-derived exosomes and astrocyte-derived exosomes from patients with Parkinson's disease established in Example 1 of the present invention and the human protein database in the Vesiclepedia vesicle database.

[0033] Figure 2 It is a flow chart of machine learning for screening diagnostic biomarkers for Parkinson's disease in neurogenic exosomes in Example 2 of the present invention.

[0034] Figure 3A It is an ROC curve of four proteins, NDUFC1, CHGA, CCL5, and ST3GAL6, screened in Example 2 of the present invention for independently differentiating early Parkinson's disease and healthy people.

[0035] Figure 3B It is an ROC curve of four proteins, NDUFC1, CHGA, CCL5, and ST3GAL6, screened in Example 2 of the present invention for jointly differentiating early Parkinson's disease and healthy people.

[0036] Figure 4A It is an ROC curve of four proteins, NDUFC1, CHGA, CCL5, and ST3GAL6, screened in Example 2 of the present invention for independently differentiating early Parkinson's disease and atypical Parkinson's syndrome.

[0037] Figure 4B It is an ROC curve of four proteins, NDUFC1, CHGA, CCL5, and ST3GAL6, screened in Example 2 of the present invention for jointly differentiating early Parkinson's disease and atypical Parkinson's syndrome.

[0038] Figure 5AIn Example 3 of the present invention, the diagnostic ROC curves for independently differentiating Parkinson's disease and healthy individuals of four proteins, namely NDUFC1, CHGA, CCL5, and ST3GAL6, were verified in an independent cohort.

[0039] Figure 5B In Example 3 of the present invention, the diagnostic ROC curves for jointly differentiating Parkinson's disease and healthy individuals of four proteins, namely NDUFC1, CHGA, CCL5, and ST3GAL6, were verified in an independent cohort.

[0040] Figure 6A In Example 3 of the present invention, the diagnostic ROC curves for independently differentiating Parkinson's disease and atypical Parkinson's syndrome of four proteins, namely NDUFC1, CHGA, CCL5, and ST3GAL6, were verified in an independent cohort.

[0041] Figure 6B In Example 3 of the present invention, the diagnostic ROC curves for jointly differentiating Parkinson's disease and atypical Parkinson's syndrome of four proteins, namely NDUFC1, CHGA, CCL5, and ST3GAL6, were verified in an independent cohort. Detailed implementation manners

[0042] The present invention will be further described in detail below in conjunction with specific embodiments.

[0043] A biomarker for the diagnosis of Parkinson's disease, the biomarker comprising one or more of NDUFC1, CHGA, CCL5, and ST3GAL6.

[0044] Further, the Parkinson's disease is early-stage Parkinson's disease.

[0045] Further, the biomarkers are all derived from neurogenic exosomes; more specifically, NDUFC1 and CHGA are derived from neuronal exosomes; CCL5 and ST3GAL6 are derived from astrocyte exosomes.

[0046] An application of the biomarker of the present invention in the preparation of a product for the diagnosis of Parkinson's disease.

[0047] A reagent for detecting the biomarker of the present invention, the reagent being used to detect the protein expression level of the biomarker; and / or, the reagent is a reagent specifically binding to the biomarker.

[0048] A kit comprising the reagent of the present invention.

[0049] A method for constructing a protein database of neurogenic exosomes related to Parkinson's disease, the method steps comprising:

[0050] (1) Isolate neurogenic exosome proteins from blood samples of early-stage Parkinson's disease patients, mid- to late-stage Parkinson's disease patients, and healthy individuals respectively, and perform enzymatic hydrolysis treatment to obtain peptide solutions respectively;

[0051] (2) Mix the obtained peptide solutions to obtain a peptide mixed solution; fractionate the peptide mixed solution using a high-performance liquid chromatograph and a high-pH reversed-phase chromatographic column, and collect the fractions;

[0052] (3) Perform high-performance liquid chromatography-tandem mass spectrometry analysis on the fractions to obtain a mass spectrometry data file;

[0053] (4) Process the mass spectrometry data file to construct a protein database of neurogenic exosomes related to Parkinson's disease.

[0054] Further, in step (1), isolate the exosome precipitate in plasma or serum, disperse the exosome precipitate in PBS solution, add a biotinylated antibody of a neurocyte exosome marker, incubate to allow the antibody to fully bind to the exosomes of neurocytes, then add streptavidin resin beads, incubate to allow streptavidin to fully bind to biotin, form an exosome-antibody-bead complex in the solution, centrifuge and separate, wash the precipitate with PBS solution, centrifuge and separate, and collect the precipitate to obtain the exosome-antibody-bead complex; the neurocytes include neurons and glial cells: add a sodium dodecyl sulfate (SDS) solution to the exosome-antibody-bead complex, boil at 90-100 °C for 5-10 min, centrifuge and separate, collect the supernatant, which is the SDS solution of the neurocyte-derived exosome protein, and detect the protein content;

[0055] Add the SDS solution of the neurocyte-derived exosome protein to an ultrafiltration tube, centrifuge to remove SDS, the upper layer of the ultrafiltration tube retains the neurocyte-derived exosome protein, add a mixed solution of urea (UA) and dithiothreitol (DTT) to the neurocyte-derived exosome protein, incubate at 37 °C for 30-60 minutes to reduce the disulfide bonds in the protein; then add a thiol-reactive alkylating reagent for alkylation modification, then wash and centrifuge successively with urea solution and ammonium bicarbonate (NH4HCO3) solution to replace the solution system, after replacing the collection tube of the ultrafiltration tube, add trypsin or a trypsin / Lys-C protease mixture solution, perform enzymatic hydrolysis and denaturation reaction at 37 °C for 12-16 hours, supplement the trypsin or trypsin / Lys-C protease mixture solution, perform denaturation reaction at 37 °C for 4-8 hours, centrifuge, and collect the solution in the collection tube after washing with ammonium bicarbonate solution, which is the peptide solution of the neurogenic exosome protein.

[0056] Further, in step (2), 200 μg of the peptide mixture solution was taken, vacuum centrifugally concentrated and then redissolved in an aqueous solution of 0.1% formic acid with a redissolution volume of 100 μl; fractionation was carried out using a high-performance liquid chromatograph and a high pH value reversed-phase chromatographic column. Mobile phase A was an aqueous solution of 10 mM ammonium formate, and mobile phase B was an 80% acetonitrile and 10 mM ammonium formate solution. Ammonia water was used to adjust the pH of mobile phases A and B to pH = 10; the mixed peptide solution was injected into the chromatographic column and then eluted. The chromatographic gradient changed as follows: from 0 min to 5 min, the linear gradient of mobile phase B was from 6% to 10%; from 5 min to 30 min, the linear gradient of mobile phase B was from 10% to 27%; from 30 min to 40 min, the linear gradient of mobile phase B was from 27% to 44%; from 40 min to 45 min, the linear gradient of mobile phase B was from 44% to 98.5%; from 45 min to 50 min, mobile phase B was maintained at 98.5%; from 50 min to 51 min, the linear gradient of mobile phase B was from 98.5% to 6%; from 51 min to 60 min, mobile phase B was maintained at 6%; the flow rate was 0.2 mL / min, and detection was carried out at UV 214 nm; 1 tube of sample was collected every 2 minutes, and a total of 30 fractions were collected in 0 - 60 minutes.

[0057] Further, in step (3), the peptide solutions were combined into 10 portions with reference to the peak intensities of the chromatogram, vacuum centrifugally concentrated and then redissolved in an aqueous solution of 0.1% formic acid for high-performance liquid chromatography-tandem mass spectrometry analysis; preferably, the chromatographic gradient was 130 min, the sample loading amount was 1 μg, mobile phase A was an aqueous solution of 0.1% formic acid by volume fraction, mobile phase B was an acetonitrile / water solution of 0.1% formic acid by volume fraction, and the volume ratio of acetonitrile to water was 20:80. The chromatographic gradient changed as follows: from 0 min to 5 min, the linear gradient of mobile phase B was from 2% to 8%; from 5 min to 90 min, the linear gradient of mobile phase B was from 8% to 24%; from 90 min to 110 min, the linear gradient of mobile phase B was from 24% to 32%; from 110 min to 115 min, the linear gradient of mobile phase B was from 32% to 90%; from 115 min to 125 min, the linear gradient of mobile phase B was from 90% to 5%; from 125 min to 130 min, mobile phase B was maintained at 5%; the flow rate was 300 nL / min; preferably, the mass spectrometry analysis was carried out in the data-dependent acquisition mode (DDA). The primary mass spectrometry scanning range was 300 - 1550 m / z, the resolution was 120000, the maximum gain control was 3e6, and the maximum injection time was 20 ms; the top 25 primary parent ions were selected for fragmentation and the secondary spectrum was collected. The secondary mass spectrometry resolution was 15000, the maximum gain control was 2e4, the maximum injection time was 30 ms, and the collision energy was 27.0 eV.

[0058] Further, in step (4), the collected original data file is imported into Spectronaut 18 software to construct a protein (spectrum) database. The protein database uses the Human Reviewed (Swiss-Prot) database downloaded from the Uniprot website, which contains 20,423 proteins. The remaining parameter settings are as follows: Missed Cleavage: 2, Min Peptide Length: 7, Max Peptide Length: 52, Toggle N-terminal M: True, Digest Rule: Trypsin / P. The above database construction is performed on exosomes derived from neurons and astrocytes of Parkinson's disease patients respectively, obtaining a protein database of exosomes derived from neurons of Parkinson's disease patients, which includes 2,247 proteins and 15,056 peptide segments; obtaining a protein database of exosomes derived from astrocytes of Parkinson's disease patients, which includes 1,722 proteins and 15,487 peptide segments.

[0059] A screening method for biomarkers for the diagnosis of Parkinson's disease, the method steps include:

[0060] (1) Mass spectrometry detection: The peptide segments of proteins in neurogenic exosomes of early Parkinson's disease patients and healthy individuals are redissolved with formic acid aqueous solution, and high-performance liquid chromatography-tandem mass spectrometry analysis is performed. The mass spectrometry analysis adopts the data-independent acquisition mode (DIA) to collect the original mass spectrometry data file.

[0061] (2) Data retrieval: The original mass spectrometry data file is imported into Spectronaut 18 software, and the DIA data of neurogenic exosomes is retrieved using the database constructed by the present invention. Trypsin digestion is set, fixed modification: Carbamidomethylation (C), variable modification: Oxidation (M), and the remaining parameters are default.

[0062] (3) Bioinformatics analysis: Normalize the protein quantification results of neurogenic exosomes of Parkinson's disease patients, and perform missing value filtering and filling; use Perseus software to calculate the fold change (FC) and significance of difference (P-value) of proteins between the early Parkinson's disease patient group and the healthy group, and screen differentially expressed proteins with P < 0.05 and |Log2FC| > 0.5 as the thresholds.

[0063] (4) Screening for Parkinson's disease diagnostic biomarkers by machine learning: Screening for Parkinson's disease diagnostic biomarkers in neuron-derived exosomes and astrocyte-derived exosomes by machine learning methods; First, import the quantitative results of differentially expressed proteins in the healthy control group and the early Parkinson's disease group as feature data, set the number of proteins in the biomarker combination to n, where n is 2-5, and loop through all combinations; Subsequently, import the data of each combination into a logistic regression model and perform five-fold cross-validation, and execute this step 10 times in a loop to calculate 5×10 accuracy (Accuracy, ACC) values and area under the curve (Area Under Curve, AUC), and take the average of the 50 results as the ACC value and AUC value of this biomarker combination; Finally, calculate the AUC values of all biomarker combinations, sort them, and select the protein combination with the highest AUC value, which is the screened Parkinson's disease diagnostic biomarker combination.

[0064] Further, in step (1), the chromatographic gradient is 130 min, the sample loading amount is 1 μg, mobile phase A is an aqueous solution of 0.1% formic acid by volume, mobile phase B is an acetonitrile / water solution of 0.1% formic acid by volume, the volume ratio of acetonitrile to water is 20:80, and the chromatographic gradient changes are as follows: 0 min - 5 min, the linear gradient of mobile phase B is from 2% - 8%; 5 min - 90 min, the linear gradient of mobile phase B is from 8% - 24%; 90 min - 110 min, the linear gradient of mobile phase B is from 24% - 32%; 110 min - 115 min, the linear gradient of mobile phase B is from 32% - 90%; 115 min - 125 min, the linear gradient of mobile phase B is from 90% - 5%; 125 min - 130 min, mobile phase B is maintained at 5%; the flow rate is 300 nL / min; during mass spectrometry analysis, the primary mass spectrometry scan range is 350 - 1250 m / z, the resolution is 120000, the maximum gain control is 3e6, and the maximum injection time is 60 ms; the secondary mass spectrometry resolution is 30000, the maximum gain control is 1e6, the maximum injection time is 55 ms, and the collision energy is 25.5 eV, 27.0 eV, 30.0 eV.

[0065] Further, in step (3), log2(x) is used for normalization; and / or, filter out missing values under the condition that the missing values in each group are <40%, and use a Truncation K - nearest neighbor algorithm (Truncation KNN) considering the truncated distribution for missing value imputation; and / or, a total of 40 differentially expressed proteins are identified in neuron-derived exosomes, including 13 up-regulated proteins and 27 down-regulated proteins; a total of 57 differentially expressed proteins are identified in astrocyte-derived exosomes, including 5 up-regulated proteins and 52 down-regulated proteins.

[0066] Example 1: Establishment of a Protein Database of Neurogenic Exosomes from Parkinson's Disease Patients

[0067] (1) Forty-eight patients with primary Parkinson's disease who met the 2015 diagnostic criteria of the Movement Disorder Society (MDS) for Parkinson's disease were enrolled. According to the severity of the disease, they were staged into early-stage (Hoehn-Yahr stage 1.0 - 2.5) and mid-late-stage (Hoehn-Yahr stage 3 - 5) Parkinson's disease. Among them, there were 23 early-stage Parkinson's disease patients and 25 mid-late-stage Parkinson's disease patients. At the same time, 24 healthy individuals without relevant medical history and confirmed to have no neurological diseases through physical examination and cognitive function assessment were enrolled.

[0068] (2) Isolate neuron-derived exosomes and astrocyte-derived exosome proteins from the blood of the participants. Take 0.5 ml of plasma samples from 72 participants, add 5 μl of thrombin to the samples, incubate at 25°C for 5 minutes, then centrifuge at 10,000 g for 20 minutes at 25°C, and take the supernatant. Filter the supernatant using a 0.22 μm pore size filter membrane, add an appropriate amount of ExoQuick TM reagent (126 μl), incubate at 4°C for 1 hour, then centrifuge at 1500 g for 30 minutes at 4°C to remove the supernatant. Add 0.5 ml of PBS solution containing protease inhibitors and phosphatase inhibitors to resuspend the exosome precipitate, and pipette repeatedly to evenly distribute the exosome particles in the solution system. Then divide the solution into two equal parts, add 2 μl of L1CAM biotinylated antibody to one part and 2 μl of GLAST biotinylated antibody to the other part, and incubate at 4°C for 1 hour. Add 12.5 μl of streptavidin resin beads to each part respectively, and incubate with a rotary mixer at 4°C for 1.5 hours. Centrifuge at 4000 g for 15 minutes at 4°C to remove the supernatant, add 0.3 ml of PBS solution to wash the beads, centrifuge at 4000 g for 15 minutes at 4°C to remove the supernatant. Add 100 μl of 0.1% SDS, boil for 5 minutes. Centrifuge at 4000×g for 20 minutes at 4°C, and collect the supernatant, which is the neuron-derived exosomes and astrocyte-derived exosome proteins in the blood of the participants.

[0069] (3) Neurogenic exosome proteolysis. Add 100 μl of 8 M urea to a 10 KDa ultrafiltration tube and rinse it, then centrifuge at 14,000 g for 20 minutes at 25 °C. Subsequently, transfer the above exosome protein sample into the 10 KDa ultrafiltration tube and centrifuge at 14,000 g for 20 minutes at 25 °C. Add 200 μl of a mixed solution of 8 M urea (UA) and 10 mM dithiothreitol (DTT) to the 10 KDa ultrafiltration tube, place it in an incubator at 37 °C for 60 minutes to reduce the disulfide bonds in the protein. Add 500 mM chloroacetamide (CAA) to the 10 KDa ultrafiltration tube to make its final concentration 50 mM, and let it stand in the dark at 25 °C for 30 minutes to alkylate the protein molecules. Then centrifuge at 14,000 g for 20 minutes at 25 °C. Add 200 μl of 8 M urea to the 10 KDa ultrafiltration tube and centrifuge at 14,000 g for 20 minutes at 25 °C, repeat twice. Add 200 μl of 50 mM ammonium bicarbonate (NH4HCO3) to the 10 KDa ultrafiltration tube and centrifuge at 14,000 g for 20 minutes at 25 °C, repeat three times. Replace the bottom collection tube of the 10 KDa ultrafiltration tube, add 100 μl of 50 mM ammonium bicarbonate to the 10 KDa ultrafiltration tube, and add trypsin according to the mass ratio of exosome protein to trypsin of 50:1. Place it in an incubator at 37 °C for a denaturation reaction for 14 hours, then supplement trypsin according to the mass ratio of exosome protein to trypsin of 100:1 and continue to place it in an incubator at 37 °C for a denaturation reaction for 4 hours. Centrifuge at 14,000 g for 20 minutes at 25 °C, then supplement 50 μl of 50 mM ammonium bicarbonate and centrifuge at 14,000 g for 20 minutes at 25 °C. Retain the liquid in the collection tube after two centrifugations, which is the peptide solution of neurogenic exosome protein.

[0070] (4) Establish a protein database of exosomes derived from neurons and astrocytes of Parkinson's disease patients.

[0071] ① Take out 10 μl of solution from the exosomes derived from neurons of 72 participants and mix them to obtain a mixed solution of peptide fragments of exosomes derived from neurons of Parkinson's disease patients; take out 10 μl of solution from the exosomes derived from astrocytes of 72 participants and mix them to obtain a mixed solution of peptide fragments of exosomes derived from astrocytes of Parkinson's disease patients. The remaining peptide solutions of exosomes derived from neurons and astrocytes of the 72 participants are used in Example 2.

[0072] ②Vacuum centrifugally concentrate the mixed solution of neuron-derived exosome peptides and the mixed solution of astrocyte-derived exosome peptides, redissolve with 0.1% formic acid aqueous solution, and fractionate using a high-performance liquid chromatograph (Thermo Scientific Vanquish) and a high pH reversed-phase chromatographic column (Waters XBridge Peptide BEH C18). Mobile phase A is 10 mM ammonium formate aqueous solution, mobile phase B is 80% acetonitrile, 10 mM ammonium formate solution, and the pH of mobile phases A and B is adjusted to pH = 10 using ammonia water. After injecting the mixed peptide solution into the chromatographic column, elution is carried out, and the chromatographic gradient changes are as follows: 0 min - 5 min, the linear gradient of mobile phase B is from 6% - 10%; 5 min - 30 min, the linear gradient of mobile phase B is from 10% - 27%; 30 min - 40 min, the linear gradient of mobile phase B is from 27% - 44%; 40 min - 45 min, the linear gradient of mobile phase B is from 44% - 98.5%; 45 min - 50 min, mobile phase B is maintained at 98.5%; 50 min - 51 min, the linear gradient of mobile phase B is from 98.5% - 6%; 51 min - 60 min, mobile phase B is maintained at 6%; the flow rate is 0.2 mL / min, and detection is carried out at UV 214 nm. Collect 1 tube of sample every 2 minutes, and a total of 30 fractions are collected in 0 - 60 minutes. Combine the neuron-derived exosome peptide fractions and the astrocyte-derived exosome peptide fractions into 10 peptide solutions according to the numbers 1, 11, 21, 2, 12, 22, ……, 10, 20, 30, vacuum centrifugally concentrate and redissolve in 0.1% formic acid aqueous solution.

[0073] ③Perform high-performance liquid chromatography-tandem mass spectrometry analysis on 10 peptide solutions of neuron-derived exosomes and astrocyte-derived exosomes, respectively. The chromatographic gradient is 130 min, the sample loading amount is 1 μg, mobile phase A is an aqueous solution of 0.1% formic acid by volume, mobile phase B is an acetonitrile / water solution of 0.1% formic acid by volume, and the volume ratio of acetonitrile to water is 20:80. The chromatographic gradient changes are as follows: 0 min - 5 min, the linear gradient of mobile phase B is from 2% - 8%; 5 min - 90 min, the linear gradient of mobile phase B is from 8% - 24%; 90 min - 110 min, the linear gradient of mobile phase B is from 24% - 32%; 110 min - 115 min, the linear gradient of mobile phase B is from 32% - 90%; 115 min - 125 min, the linear gradient of mobile phase B is from 90% - 5%; 125 min - 130 min, mobile phase B is maintained at 5%; the flow rate is 300 nL / min. Mass spectrometry analysis is performed using the data-dependent acquisition mode (DDA). The primary mass spectrometry scan range is 300 - 1550 m / z, the resolution is 120000, the maximum gain control is 3e6, and the maximum injection time is 20 ms; select the top 25 primary parent ions for fragmentation and collect the secondary spectra. The secondary mass spectrometry resolution is 15000, the maximum gain control is 2e4, the maximum injection time is 30 ms, and the collision energy is 27.0 eV.

[0074] ④Import the original data files collected in the DDA mode into Spectronaut 18 software to construct a spectral library with default parameters and settings as follows: Missed Cleavage: 2, Min Peptide Length: 7, Max Peptide Length: 52, Toggle N-terminal M: True, Digest Rule: Trypsin / P. Perform the above library construction on the neuron-derived exosomes and astrocyte-derived exosomes of Parkinson's disease patients respectively to obtain a protein database of neuron-derived exosomes of Parkinson's disease patients, which includes 2247 proteins and 15056 peptides; obtain a protein database of astrocyte-derived exosomes of Parkinson's disease patients, which includes 1722 proteins and 15487 peptides. Figure 1 For the Venn diagram of the protein databases of neuron-derived exosomes and astrocyte-derived exosomes of Parkinson's disease patients established in Example 1 and the human protein database in the Vesiclepedia vesicle database, it can be seen that there are 1356 common proteins between the protein database of neuron-derived exosomes and the protein database of astrocyte-derived exosomes; compared with the human protein database in the Vesiclepedia vesicle database, 1685 proteins were newly identified.

[0075] Example 2: Screening of Biomarkers for Parkinson's Disease Diagnosis

[0076] (1) Proteomic Analysis of Neurogenic Exosomes in the Parkinson's Disease Clinical Cohort: The peptide solutions of neuron-derived exosomes and astrocyte-derived exosomes remaining in the healthy control group and the early Parkinson's disease group were concentrated by vacuum centrifugation and reconstituted with 0.1% formic acid aqueous solution. At the same time, 23 patients with atypical Parkinson's syndrome were included, and sample preparation was carried out by the same method, followed by high performance liquid chromatography-tandem mass spectrometry analysis. The chromatographic gradient was 130 min, the sample loading amount was 1 μg, mobile phase A was 0.1% formic acid aqueous solution by volume fraction, mobile phase B was acetonitrile / water solution with 0.1% formic acid by volume fraction, and the volume ratio of acetonitrile to water was 20:80. The chromatographic gradient changed as follows: 0 min - 5 min, the linear gradient of mobile phase B was from 2% - 8%; 5 min - 90 min, the linear gradient of mobile phase B was from 8% - 24%; 90 min - 110 min, the linear gradient of mobile phase B was from 24% - 32%; 110 min - 115 min, the linear gradient of mobile phase B was from 32% - 90%; 115 min - 125 min, the linear gradient of mobile phase B was from 90% - 5%; 125 min - 130 min, mobile phase B was maintained at 5%; the flow rate was 300 nL / min. The mass spectrometry analysis was performed in data-independent acquisition mode (DIA). The primary mass spectrometry scan range was 350 - 1250 m / z, the resolution was 120000, the maximum gain control was 3e6, and the maximum injection time was 60 ms; the secondary mass spectrometry resolution was 30000, the maximum gain control was 1e6, the maximum injection time was 55 ms, and the collision energy was 25.5 eV, 27.0 eV, 30.0 eV.

[0077] (2) Data Retrieval: The original data files collected in DIA mode were imported into Spectronaut 18 software. The DIA data of neuron-derived exosomes were retrieved using the constructed protein database of neuron-derived exosomes from Parkinson's disease patients, and the DIA data of astrocyte-derived exosomes were retrieved using the constructed protein database of astrocyte-derived exosomes from Parkinson's disease patients. Trypsin digestion was set, fixed modification: Carbamidomethylation (C), variable modification: Oxidation (M), and the remaining parameters were default.

[0078] (3) Bioinformatics analysis: Normalize the protein quantification results of neurogenic exosomes (preferably, use log2(x) for normalization), and perform missing value filtering and imputation (preferably, filter missing values under the condition that the missing value of each group < 40%, and use a Truncation K-Nearest Neighbor algorithm to impute missing values). Use Perseus software to calculate the fold change (FC) and significance of difference (P-value) of proteins between the early Parkinson's disease patient group and the healthy group. Screen differentially expressed proteins with P < 0.05 and |Log2FC| > 0.5 as the threshold. A total of 40 differentially expressed proteins were identified in neuron-derived exosomes, including 13 up-regulated proteins and 27 down-regulated proteins; 57 differentially expressed proteins were identified in astrocyte-derived exosomes, including 5 up-regulated proteins and 52 down-regulated proteins.

[0079] (4) Screening Parkinson's disease diagnostic biomarkers by machine learning: Screen Parkinson's disease diagnostic biomarkers in neuron-derived exosomes and astrocyte-derived exosomes by machine learning methods. First, import the quantification results of differentially expressed proteins in the healthy control group and the early Parkinson's disease group as feature data, and set the number of proteins in the biomarker combination to 4, and loop through all combinations. Subsequently, import the data of each combination into the logistic regression model and perform five-fold cross-validation, and execute this step 10 times in a loop to calculate 5×10 accuracy (ACC) values and area under the curve (AUC). Take the average of the 50 results as the ACC value and AUC value of this biomarker combination. Finally, calculate the AUC values of all biomarker combinations, sort them, and select the protein combination with the highest AUC value, which is the screened Parkinson's disease diagnostic biomarker combination; as Figure 2 shown. The machine learning results are shown in Table 1.

[0080] Table 1 Screening results of biomarkers in differentially expressed proteins of neurogenic exosomes (n = 4)

[0081]

[0082] The results show that when the four proteins NDUFC1, CHGA in neuron-derived exosomes and CCL5, ST3GAL6 in astrocyte-derived exosomes are used as a Parkinson's disease diagnostic biomarker combination, the best ACC value and AUC value reach 0.888 and 0.95 respectively, which is a set of effective Parkinson's disease diagnostic biomarkers.

[0083] The selected biomarker combinations were subjected to ROC curve analysis. First, the sensitivity, specificity, accuracy, AUC value, and confidence interval for each protein in differentiating the healthy control group and the early Parkinson's disease group were calculated. Subsequently, a logistic regression model was constructed with the protein combinations, and the diagnostic efficacy of the combined index was calculated. The ROC curves of the four proteins, NDUFC1, CHGA, CCL5, and ST3GAL6, in independently differentiating early Parkinson's disease and healthy individuals are as Figure 3A shown. The AUC values of NDUFC1 and CHGA in neuron-derived exosomes were 0.741 and 0.703, respectively, and the AUC values of CCL5 and ST3GAL6 in astrocyte-derived exosomes were 0.783 and 0.725, respectively. The ROC curve of the four proteins in jointly differentiating early Parkinson's disease and healthy individuals is as Figure 3B shown, with the AUC value reaching 0.967 and the confidence interval being 0.927 - 1.000. The supplementary results of the ROC curves of the biomarker combinations are shown in Table 2.

[0084] Table 2 Supplementary Results of ROC Curves of Biomarker Combinations for Early Diagnosis of Parkinson's Disease

[0085]

[0086]

[0087] Using the proteomic data of atypical Parkinson's syndrome, the differential diagnostic ability of the selected biomarker combinations was further evaluated. The ROC curves of the four proteins, NDUFC1, CHGA, CCL5, and ST3GAL6, in independently differentiating early Parkinson's disease and atypical Parkinson's syndrome are as Figure 4A shown. The AUC values of NDUFC1 and CHGA in neuron-derived exosomes were 0.662 and 0.698, respectively, and the AUC values of CCL5 and ST3GAL6 in astrocyte-derived exosomes were 0.474 and 0.611, respectively. The ROC curve of the four proteins in jointly differentiating early Parkinson's disease and healthy individuals is as Figure 4B shown, with the AUC value reaching 0.766 and the confidence interval being 0.625 - 0.907.

[0088] Example 3: Validation of Biomarkers for Parkinson's Disease Diagnosis

[0089] In an independent research cohort, the selected diagnostic biomarkers were validated by ELISA. NDUFC1 and CHGA in neuron-derived exosomes and CCL5 and ST3GAL6 in astrocyte-derived exosomes were selected for validation.

[0090] (1) Isolation of neurogenic exosomal proteins: In the validation study cohort, 15 patients with early Parkinson's disease, 15 patients with atypical Parkinsonian syndromes, and 15 healthy individuals were included, and neuronal-derived exosomes and astrocyte-derived exosomal proteins in the blood of the participants were isolated. 0.5 ml of plasma samples from 45 participants were taken, 5 μl of thrombin was added to the samples, and after incubation at 25 °C for 5 minutes, centrifugation was performed at 10,000 g at 25 °C for 20 minutes, and the supernatant was taken. The supernatant was filtered using a filter membrane with a pore size of 0.22 μm, and an appropriate amount of ExoQuick TM reagent (126 μl) was added to the sample, incubated at 4 °C for 1 hour, and then centrifuged at 1500 g at 4 °C for 30 minutes to remove the supernatant. 0.5 ml of PBS solution containing protease inhibitors and phosphatase inhibitors was added to resuspend the exosomal precipitate, and the exosomal particles were evenly distributed in the solution system by repeated pipetting. Subsequently, the solution was divided into two equal parts, 2 μl of L1CAM biotinylated antibody was added to one part, and 2 μl of GLAST biotinylated antibody was added to the other part, and incubated at 4 °C for 1 hour. 12.5 μl of streptavidin resin beads were added respectively, and incubated on a 4 °C rotary mixer for 1.5 hours. Centrifugation was performed at 4000 g at 4 °C for 15 minutes to remove the supernatant, 0.3 ml of PBS solution was added to wash the beads, and centrifugation was performed at 4000 g at 4 °C for 15 minutes to remove the supernatant. 100 μl of high-efficiency RIPA solution premixed with protease and phosphatase inhibitors was added, and lysed on ice for 15 minutes. Centrifugation was performed at 4000×g at 4 °C for 20 minutes, and the supernatant was collected, which was the neuronal-derived exosomes and astrocyte-derived exosomal proteins in the blood of the participants.

[0091] (2) ELISA test: The selected Parkinson's disease diagnostic biomarkers NDUFC1, CHGA, CCL5 and ST3GAL6 were verified by ELISA method. Experimental method: 1) Prepare the standard according to the instructions of the ELISA test kit. The kit numbers used are: Fankewei F0296-HA, Fankewei F0290-HA, Jianglai Bio JL11689, and Fankewei F10879-A. Add 100 μl of samples or standards of different concentrations to the corresponding wells, and add 100 μL of universal diluent to the blank wells. Cover with the sealing film and incubate at 37°C for 1 hour. 2) Remove the ELISA plate, discard the liquid, and do not wash. Add 100 μL of biotinylated antibody working solution directly to each well, cover with the sealing film and incubate at 37°C for 1 hour. 3) Discard the liquid, add 300μL 1x washing solution to each well, let stand for 1 minute, shake off the washing solution, pat dry on absorbent paper, and repeat the plate washing process 3 times. 4) Add 100μL of enzyme conjugate working solution to each well, cover with sealing film and incubate at 37℃ for 30 minutes. 5) Discard the liquid, add 300μL 1x washing solution to each well, let stand for 1 minute, shake off the washing solution, pat dry on absorbent paper, and repeat the plate washing process 5 times. 6) 10. Add 90μL of substrate (TMB) to each well, cover with sealing film, and incubate at 37℃ in the dark for 15 minutes. 7) Take out the ELISA plate, directly add 50μL of stop solution to each well, and immediately measure the OD value of each well at a wavelength of 450nm.

[0092] (3) Data processing: Take the average value of the parallel group data and subtract the OD value of the blank well; draw a standard curve based on the concentration of the standard and fit the equation to calculate the concentration of the target protein in the sample, and calculate the diagnostic ROC curve of the four proteins as a combined indicator.

[0093] In the validation cohort, the ROC curves of the four proteins NDUFC1, CHGA, CCL5, and ST3GAL6 independently distinguished early Parkinson's disease from healthy subjects. Figure 5A As shown in Figure 2, the AUC values ​​of NDUFC1 and CHGA in neuron-derived exosomes were 0.820 and 0.758, respectively, and the AUC values ​​of CCL5 and ST3GAL6 in astrocyte-derived exosomes were 0.657 and 0.856, respectively. The ROC curve of the four proteins combined to distinguish early Parkinson's disease from healthy people is shown in Figure 2. Figure 5B As shown, the AUC value reached 0.859 with a confidence interval of 0.727-0.992.

[0094] The differential diagnostic ability of the biomarker combination in the validation study cohort was further evaluated. The ROC curves of the four proteins NDUFC1, CHGA, CCL5 and ST3GAL6 independently distinguished early Parkinson's disease from atypical Parkinson's syndrome. Figure 6AAs shown, the AUC values of NDUFC1 and CHGA in neuron-derived exosomes were 0.489 and 0.859, respectively, and the AUC values of CCL5 and ST3GAL6 in astrocyte-derived exosomes were 0.681 and 0.869, respectively. The ROC curve for the four proteins combined to distinguish early Parkinson's disease from healthy individuals is as Figure 6B shown, with an AUC value reaching 0.985 and a confidence interval of 0.957 - 1.000.

[0095] In summary, the invention includes but is not limited to the above embodiments. Any equivalent replacement or partial improvement made under the spirit and principles of the present invention shall be regarded as within the protection scope of the present invention.

Claims

1. A biomarker for diagnosing Parkinson's disease, characterized in that: The biomarkers include one or more of NDUFC1, CHGA, CCL5 and ST3GAL6.

2. A biomarker for diagnosing Parkinson's disease according to claim 1, characterized in that: The Parkinson's disease is early-stage Parkinson's disease.

3. A biomarker for diagnosing Parkinson's disease according to claim 1, characterized in that: The biomarkers are all derived from neurogenic exosomes; further, the NDUFC1 and CHGA are derived from neuronal exosomes; and CCL5 and ST3GAL6 are derived from astrocyte exosomes.

4. Use of the biomarker according to any one of claims 1 to 3 in the preparation of a product for diagnosing Parkinson's disease.

5. A reagent for detecting the biomarker according to any one of claims 1 to 3, characterized in that: The reagent is used to detect the protein expression level of the biomarker; and / or, the reagent is a reagent that specifically binds to the biomarker.

6. A kit, characterized in that Comprising the reagent according to claim 5.

7. A method for constructing a protein database of neurogenic exosomes associated with Parkinson's disease, characterized in that: The method steps include: (1) Neurogenic exosome proteins were isolated from blood samples of patients with early Parkinson's disease, patients with mid-to-late Parkinson's disease, and healthy subjects, and enzymatically treated to obtain peptide solutions; (2) mixing the obtained peptide solutions to obtain a peptide mixed solution; fractionating the peptide mixed solution using a high performance liquid chromatograph and a high pH reverse phase chromatographic column, and collecting fractions; (3) performing high performance liquid chromatography-tandem mass spectrometry analysis on the fraction to obtain a mass spectrometry data file; (4) Processing the mass spectrometry data file to construct a protein database of neurogenic exosomes related to Parkinson's disease.

8. A method for screening biomarkers for diagnosing Parkinson's disease, characterized in that: The method steps include: (1) Mass spectrometry: The peptides of neurogenic exosome proteins from patients with early Parkinson's disease and healthy subjects were re-dissolved in formic acid aqueous solution and analyzed by high performance liquid chromatography-tandem mass spectrometry. The mass spectrometry analysis adopted the data-independent acquisition mode (DIA) to collect the original files of mass spectrometry data; (2) Data retrieval: The original mass spectrometry data file was imported into Spectronaut 18 software, and the DIA data of neurogenic exosomes were retrieved using the database constructed in claim 7, and the trypsin enzyme digestion, fixed modification: carbamidomethylation (C), variable modification: oxidation (M), and the other parameters were set as default; (3) Bioinformatics analysis: The protein quantification results of neurogenic exosomes from Parkinson's disease patients were normalized, and missing values ​​were filtered and filled. The Perseus software was used to calculate the fold difference FC and the significance P of the difference between the early Parkinson's disease patient group and the healthy control group, and the differentially expressed proteins were screened with P < 0.05 and |Log2FC| > 0.5 as the thresholds. (4) Machine learning screening of diagnostic biomarkers for Parkinson's disease: Machine learning methods were used to screen diagnostic biomarkers for Parkinson's disease in neuronal exosomes and astrocyte-derived exosomes. First, the quantitative results of differentially expressed proteins in the healthy control group and the early Parkinson's disease group were imported as feature data, and the number of proteins in the biomarker combination was set to n, with n ranging from 2 to 5, and all combinations were looped over. Subsequently, the data of each combination was imported into the logistic regression model and a five-fold cross-validation was performed. This step was repeated 10 times to calculate 5×10 accuracy ACC values ​​and area under the curve AUC values. The average of the 50 results was taken as the ACC value and AUC value of this biomarker combination. Finally, the AUC values ​​of all biomarker combinations were calculated, and the protein combination with the highest AUC value was selected after sorting, which was the screened Parkinson's disease diagnostic biomarker combination.

9. A method for screening biomarkers for diagnosing Parkinson's disease according to claim 8, characterized in that: In step (1), the chromatographic gradient is 130min, the sample load is 1μg, the mobile phase A is a 0.1% formic acid aqueous solution by volume, the mobile phase B is an acetonitrile / water solution with a 0.1% formic acid by volume fraction, the volume ratio of acetonitrile to water is 20:80, and the chromatographic gradient changes as follows: 0min-5min, the linear gradient of mobile phase B is from 2% to 8%; 5min-90min, the linear gradient of mobile phase B is from 8% to 24%; 90min-110min, the linear gradient of mobile phase B is from 24% to 32%; 110min-115min, the linear gradient of mobile phase B is from 24% to 32% From 32%-90%; 115min-125min, the linear gradient of mobile phase B is from 90%-5%; 125min-130min, the mobile phase B is maintained at 5%; the flow rate is 300nL / min; during mass spectrometry analysis, the primary mass spectrometry scanning range is 350-1250m / z, the resolution is 120000, the maximum gain control is 3e6, and the maximum injection time is 60ms; the secondary mass spectrometry resolution is 30000, the maximum gain control is 1e6, the maximum injection time is 55ms, and the collision energy is 25.5eV, 27.0eV, and 30.0eV.

10. The method for screening biomarkers for diagnosing Parkinson's disease according to claim 8, characterized in that: In step (3), log2(x) was used for normalization; and / or, missing values ​​were filtered with the condition that the missing values ​​of each group were <40%, and a K-nearest neighbor algorithm considering truncated distribution was used to fill missing values; and / or, a total of 40 differential proteins were identified in neuronal exosomes, including 13 up-regulated proteins and 27 down-regulated proteins; a total of 57 differential proteins were identified in astrocyte exosomes, including 5 up-regulated proteins and 52 down-regulated proteins.

Images