Marker combination, kit thereof and application of marker combination in prediction of diabetic peripheral neuropathy

Abstract

The invention discloses a marker combination, a kit thereof and application of the marker combination to prediction of diabetic peripheral neuropathy, the marker combination comprises human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1) and human mannan binding lectin serine peptide 1 (MASP1), TMT quantitative proteomics and PRM targeted proteomics are adopted, and the marker combination is used for detecting diabetic peripheral neuropathy. A technology combination strategy of an ELISA experiment is adopted, and key differential expression proteins such as MASP1, VCAM1 and ANG in serum exosomes of patients with diabetic peripheral neuropathy are systematically screened and verified. By establishing a multi-model machine learning diagnosis system based on logistic regression, a support vector machine and a naive Bayes algorithm, simple and rapid identification of diabetic peripheral neuropathy patients is realized, and a new experimental basis is provided for clarification of a molecular regulation network of diabetic peripheral neuropathy; more importantly, a rapid screening tool with clinical application potential is developed.

Description

Technical Field

[0001] The present invention belongs to the field of biological detection, and in particular relates to a marker combination, a kit thereof and an application thereof in predicting diabetic peripheral neuropathy. Background Art

[0002] Type 2 diabetes mellitus (DM) is a chronic metabolic disease that is prevalent worldwide. The global prevalence of diabetes mellitus (DM) has surged in recent years. Among the many complications of diabetes, diabetic peripheral neuropathy (DPN) has attracted considerable attention due to its high disability rate. Diabetic peripheral neuropathy (DPN) is defined as peripheral nerve dysfunction caused by hyperglycemia toxicity, microvascular damage, and oxidative stress, after other etiologies have been excluded. Its typical clinical manifestations include symmetrical pain in the distal limbs, "glove-and-stocking" paresthesias, and autonomic dysfunction. DPN is the leading cause of diabetic foot ulcers and lower limb amputations, severely impacting patients' quality of life and shortening their life expectancy. Currently, the diagnosis of DPN relies on electromyography (EMG). Despite its high accuracy, it has limitations such as time-consuming operation, high cost, and significant patient discomfort, making it difficult to popularize, especially in primary care settings. Therefore, the development of noninvasive and efficient biomarkers for early screening of DPN has become a pressing clinical challenge. Summary of the Invention

[0003] To address the above-mentioned technical problems, the present invention seeks to protect a marker combination, a kit thereof, and its use in predicting diabetic peripheral neuropathy. The marker combination of the present invention includes human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1). The present invention uses multi-omics technology to screen for DPN-specific serum exosome protein markers and constructs a machine learning-assisted diagnostic model. The present invention discovers for the first time that serum exosomes MASP1, VCAM1, and ANG can serve as potential biomarkers for DPN. The machine learning diagnostic model constructed by combining these exosomes with serological indicators provides a new strategy for early screening.

[0004] The first aspect of the present invention protects a marker combination comprising human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1).

[0005] The second aspect of the present invention protects the use of a marker combination as described in the first aspect of the present invention in the preparation of a drug for predicting and / or diagnosing diabetes, wherein the diabetes is diabetic peripheral neuropathy or diabetes without peripheral neuropathy.

[0006] In a preferred embodiment, there is a significant difference in the expression levels of the marker combination between diabetic peripheral neuropathy and diabetic non-peripheral neuropathy.

[0007] In a preferred embodiment, the marker combination is obtained from a serum sample.

[0008] In a specific embodiment, when the level of the marker combination is measured using Tandem Mass Tags (TMT) or Parallel Reaction Monitoring (PRM), and the drug is used to predict and / or diagnose whether diabetes mellitus has peripheral neuropathy, the subject whose expression level of the marker combination is significantly increased is a diabetic peripheral neuropathy patient; and / or, the subject whose expression level of the marker combination is significantly decreased is a diabetic non-peripheral neuropathy patient.

[0009] In a specific embodiment, when the level of the marker combination is measured using ELISA and the medicament is used to predict and / or diagnose the presence or absence of diabetic peripheral neuropathy, the subject whose expression levels of MASP and ANG are significantly increased is a diabetic peripheral neuropathy patient; and / or, the subject whose expression level of VCAM1 is significantly decreased is a diabetic peripheral neuropathy patient; the subject whose expression levels of MASP and ANG are significantly decreased is a diabetic patient without peripheral neuropathy; and / or, the subject whose expression level of VCAM1 is significantly increased is a diabetic patient without peripheral neuropathy.

[0010] The “significant difference” refers to the degree of statistical significance, such as the p-value in hypothesis testing: when the p-value is less than the preset significance level (such as p<0.001, p<0.01, p<0.05), the result is considered to be statistically significant.

[0011] Another aspect of the present invention also protects the use of a reagent for determining a marker combination in the preparation of a kit for determining diabetic peripheral neuropathy in a subject, wherein the marker combination includes human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1).

[0012] In a preferred embodiment, the marker combination is obtained from a serum sample of a subject.

[0013] In a preferred embodiment, the reagent is a reagent for determining the level of a marker.

[0014] In a preferred embodiment, the reagent is a reagent used in parallel reaction monitoring, tandem mass tagging or ELISA.

[0015] Another aspect of the present invention also protects a reagent for detecting a marker combination, wherein the marker combination includes human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1), and the reagent is a reagent for determining the level of the marker.

[0016] In a preferred embodiment, the reagent is a reagent used in parallel reaction monitoring, tandem mass tagging or ELISA.

[0017] In a preferred embodiment, the marker is derived from a serum sample.

[0018] Another aspect of the present invention also protects a diagnostic kit for determining diabetes in a subject, comprising the reagent and control substance as described in the above aspect of the present invention, wherein the diabetes is diabetic peripheral neuropathy or diabetic without peripheral neuropathy.

[0019] Another aspect of the present invention also protects a diabetes diagnosis system, comprising a detection module and an analysis and judgment module; the detection module detects the level of a marker combination in the serum of a subject and transmits the level data to the analysis and judgment module; the analysis and judgment module judges the level data of the marker combination and outputs a diagnosis result as follows: the subject has diabetic peripheral neuropathy or diabetic without peripheral neuropathy; wherein the marker combination includes human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1).

[0020] In a preferred embodiment, the analysis and judgment module judges the level data of the marker combination including: determining whether there is a significant difference in the expression level of the marker combination between diabetic peripheral neuropathy and diabetic non-peripheral neuropathy.

[0021] In a specific embodiment, the analysis and judgment module determines the level data of the marker combination, including: when the level of the marker combination is measured using Tandem Mass Tags (TMT) or Parallel Reaction Monitoring (PRM), and the drug is used to predict and / or diagnose whether diabetes has peripheral neuropathy, the subject whose expression level of the marker combination is significantly increased is a patient with diabetic peripheral neuropathy; and / or, the subject whose expression level of the marker combination is significantly decreased is a patient with diabetes without peripheral neuropathy.

[0022] In a specific embodiment, the analysis and judgment module determines the marker combination level data including: when the marker combination level is measured using ELISA, and the drug is used to predict and / or diagnose whether diabetic peripheral neuropathy exists, the subject whose MASP and ANG expression levels are significantly increased is a diabetic peripheral neuropathy patient; and / or, the subject whose VCAM1 expression level is significantly decreased is a diabetic peripheral neuropathy patient; the subject whose MASP and ANG expression levels are significantly decreased is a diabetic patient without peripheral neuropathy; and / or, the subject whose VCAM1 expression level is significantly increased is a diabetic patient without peripheral neuropathy.

[0023] Another aspect of the present invention further seeks protection for a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the functions of the diagnostic system as described in the above aspects of the present invention.

[0024] Another aspect of the present invention further seeks protection for an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to implement the functions of the diagnostic system as described in the above aspect of the present invention.

[0025] On the basis of conforming to the common sense in this field, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present invention.

[0026] Unless otherwise specified, the reagents and raw materials used in the present invention are commercially available.

[0027] The positive progress effect of the present invention is:

[0028] This study employed a combined strategy of TMT quantitative proteomics, PRM targeted proteomics, and ELISA to systematically screen and validate key differentially expressed proteins, including MASP1, VCAM1, and ANG, in the serum exosomes of patients with diabetic peripheral neuropathy. By establishing a multi-model machine learning diagnostic system based on logistic regression, support vector machines, and a naive Bayesian algorithm, the study enabled rapid and simple identification of patients with diabetic peripheral neuropathy. This not only provides new experimental evidence for elucidating the molecular regulatory network underlying diabetic peripheral neuropathy but, more importantly, develops a rapid screening tool with potential clinical application, enabling the rapid and simple diagnosis of diabetic peripheral neuropathy. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Comparison of serum exosomal MASP1 concentrations between the diabetic group without peripheral neuropathy and the diabetic group with peripheral neuropathy (relative quantification).

[0030] Figure 2 Comparison of serum exosomal VCAM1 concentrations between the diabetic non-peripheral neuropathy group and the diabetic peripheral neuropathy group (relative quantification)

[0031] Figure 3 Comparison of serum exosomal ANG concentrations between the diabetic group without peripheral neuropathy and the diabetic group with peripheral neuropathy (relative quantification).

[0032] Figure 4 Comparison of serum MASP1 concentrations between the diabetic non-peripheral neuropathy group and the diabetic peripheral neuropathy group (pg / ml)

[0033] Figure 5 Comparison of serum VCAM1 concentrations between the diabetic non-peripheral neuropathy group and the diabetic peripheral neuropathy group (ng / mL)

[0034] Figure 6 Comparison of serum ANG concentrations between the diabetic non-peripheral neuropathy group and the diabetic peripheral neuropathy group (pg / mL)

[0035] Figure 7 To demonstrate the results of protein quantification analysis using PRM technology, the volcano plot uses the log2-transformed fold difference (FC) on the horizontal axis and the -log10-transformed statistical significance value on the vertical axis. Red dots in the figure represent proteins with significantly different expression levels (|log2FC| ≥ 0.26 and p < 0.05) between patients with diabetic peripheral neuropathy and those without diabetic peripheral neuropathy, while black dots represent proteins that did not reach the significance threshold.

[0036] Figure 8Figure 3. PRM relative quantitative proteomic cluster analysis of differentially expressed proteins between the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group. The heatmap uses log2 normalization to visualize the quantitative distribution of differentially expressed proteins. In the color coding, red indicates upregulation, blue corresponds to downregulation, and gray areas indicate no quantitative information.

[0037] Figure 9 Receiver operating characteristic (ROC) curve of the prediction model for diabetic peripheral neuropathy constructed for the machine learning method logistic regression.

[0038] Figure 10 ROC curve of the prediction model of diabetic peripheral neuropathy constructed by support vector machine, a machine learning method.

[0039] Figure 11 Receiver operating characteristic (ROC) curve of a prediction model for diabetic peripheral neuropathy constructed using the machine learning method Naive Bayes.

[0040] Figure 12 For feature importance visualization analysis, MASP1: human mannan-binding lectin serine peptide 1, DD dimer: DD dimer, HbA1C: glycated hemoglobin, AST: aspartate aminotransferase, Scr: serum creatinine, TG: serum triglyceride, TC: serum total cholesterol, VCAM1: human vascular endothelial cell adhesion molecule 1, TBA: total bile acid, ANG: human angiotensin. DETAILED DESCRIPTION

[0041] This patent integrates multi-omics technology and machine learning: (1) screening of exosomal proteins related to diabetic peripheral neuropathy through TMT-labeled quantitative proteomics; (2) dual validation using PRM-targeted mass spectrometry and ELISA; (3) construction of a machine learning diagnostic model based on clinical indicators and exosomal proteins. The study used ultracentrifugation combined with 0.22μm filtration to separate serum exosomes, mapped protein expression profiles using TMT and C-MS / MS, combined with PRM / ELISA to verify candidate markers, and finally optimized model performance using logistic regression, support vector machine, and naive Bayes algorithm. This strategy not only provides biomarkers for the early diagnosis of diabetic peripheral neuropathy, but also provides a predictive model with good predictive performance for the simple and rapid diagnosis of diabetic peripheral neuropathy.

[0042] The present invention is further illustrated by way of examples below, but the present invention is not limited to the scope of the examples. Experimental methods in the following examples where specific conditions are not specified were performed according to conventional methods and conditions, or selected according to the product specifications.

[0043] The experimental methods involved in the examples are now described.

[0044] 1. Research subjects

[0045] 1.1 Patient selection

[0046] This study enrolled patients with diabetes who were hospitalized in the Department of Endocrinology and Metabolism at the Shanghai Tenth People's Hospital affiliated with Tongji University between December 2022 and December 2024. Participants were divided into two groups based on electromyography (EMG) test results: a diabetic group without peripheral neuropathy and a diabetic group with peripheral neuropathy. All participants completed neuroelectrophysiological examinations within 48 hours of admission. The diabetic group with peripheral neuropathy met the EMG diagnostic criteria for neuropathy, while the diabetic group without peripheral neuropathy consisted of patients with diabetes whose EMG test did not indicate neuropathy. The diagnostic criteria are formulated in accordance with the "Guidelines for the Prevention and Treatment of Type 2 Diabetes in China (2020 Edition)": ① The diagnostic criteria for diabetes are fasting blood glucose (FPG) ≥7.0mmol / L (126mg / dL) and / or blood glucose ≥11.1mmol / L (200mg / dL) 2 hours after an oral glucose tolerance test; ② The diagnosis of diabetic peripheral neuropathy must meet the following requirements at the same time: the presence of clinical symptoms or signs of nerve damage, abnormal nerve conduction confirmed by electromyography, and the exclusion of other possible causes of neuropathy.

[0047] Exclusion criteria included: ① hospitalization time less than 48 hours; ② age <18 years or >89 years; ③ clinical data missing more than 30%; ④ combined with severe hepatic and renal insufficiency (eGFR <30 mL / min / 1.73 m 2 ), malignant tumors, acute metabolic disorders or central nervous system diseases.

[0048] 1.2 Clinical Data Collection

[0049] The following data are collected through the electronic medical record system:

[0050] Medical history characteristics

[0051] Basic information such as the patient's gender, age, and smoking history was recorded, with a focus on collecting information on metabolic-related complications (hypertension, atherosclerosis) and diabetic complications.

[0052] Laboratory testing

[0053] The test indicators completed within 24 hours of admission were collected: routine blood test (white blood cell count, red blood cell count, platelet count, lymphocyte count, hemoglobin, etc.), coagulation function, fasting blood glucose, glycosylated hemoglobin (HbA1c), lipid profile (TC, TG, LDL-C, HDL-C, bilirubin); blood creatinine, urea nitrogen, estimated glomerular filtration rate (eGFR), liver function (ALT, AST, TBA); 24-hour urine protein quantification, urine microalbumin / creatinine ratio (UACR); tumor markers such as alpha-fetoprotein (AFP) and carcinoembryonic antigen (CEA), and routine urine test.

[0054] Imaging and functional examinations

[0055] Cardiovascular assessment: standard 12-lead electrocardiogram, cardiac color Doppler ultrasound (left ventricular ejection fraction); vascular ultrasound: limb vascular color Doppler ultrasound (plaque location and blood flow parameters); neuroelectrophysiology: limb electromyography (motor nerve conduction velocity and amplitude, sensory nerve conduction velocity and amplitude).

[0056] 1.3 Patient blood sample collection and storage

[0057] All subjects had 10 mL of fasting venous blood collected by a nurse at 7:00 a.m. on the morning of the next day after admission (after a 12-hour fast). Samples were processed according to the following standardized procedures:

[0058] Coagulation and separation: The blood collection tube was placed upright at room temperature for 30 minutes to complete coagulation, and then centrifuged at 3000×g for 10 minutes at room temperature.

[0059] Aliquot and store: Use a sterile pipette to aspirate the upper serum layer and aliquot into 1.5 mL pyrogen-free EP tubes (500 μL per tube) to avoid repeated freezing and thawing.

[0060] Low temperature storage: Transfer the samples to a -80°C ultra-low temperature freezer within 30 minutes after packaging until subsequent testing.

[0061] 2. Experimental steps

[0062] 2.1 Serum exosome extraction

[0063] (1) Thaw the collected serum at room temperature and centrifuge the serum at 1500 g for 10 minutes to remove particulate matter;

[0064] (2) Centrifuge the supernatant at 10,000 g for 30 minutes at 4°C to remove small precipitates and transfer the supernatant to a new centrifuge tube;

[0065] (3) Filter with a 0.22 μm filter to remove residue;

[0066] (4) Centrifugation at 100,000 g for 1 hour at 4°C;

[0067] (5) Discard the supernatant, resuspend the reddish-white precipitate in PBS, and centrifuge at 100,000 g for 1 hour at 4°C.

[0068] (6) Discard the supernatant, add 100-200 μL PBS to the exosome pellet at the bottom, resuspend it, collect it, and store it at -80°C.

[0069] 2.2 Protein extraction and trypsin digestion

[0070] (1) Remove the sample at -80°C and add it to a 3KD ultrafiltration tube. Then add 500 μL of PBS to the 3KD ultrafiltration tube and centrifuge (12000 × g, 30 minutes) until the residual liquid in the ultrafiltration tube is 50 μL.

[0071] (2) Add 450 μL of lysis buffer (6 M urea, 2 M thiourea) to each tube, centrifuge at 12,000 g for 40 min, and transfer 60-70 μL of the remaining liquid in the ultrafiltration tube to a 1.5 mL EP tube;

[0072] (3) Take 10 μL of sample from the PCT tube.

[0073] (4) Denaturation, reduction, and alkylation were performed using the Barocycler high-pressure sample pretreatment system. 30 μL of lysis buffer (6 M urea, 2 M thiourea) was added to each tube, followed by 5 μL of 0.2 M TCEP and 2.5 μL of 0.8 M IAA. Parameters were set as follows: 45 kpsi, 30 seconds of high pressure; 10 seconds of normal pressure, 90 pressure cycles. The temperature was 30°C.

[0074] (5) Enzymatic hydrolysis was performed using the Barocycler high-pressure sample pretreatment system. Equal amounts of protein from each sample were hydrolyzed, and TEAB, trypsin, and Rys-C protease were added. The pH was adjusted to 8. Parameters were set as follows: 20 kpsi, 50 seconds of high pressure; 10 seconds of normal pressure, 120 pressure cycles, and a temperature of 30°C.

[0075] (6) Finally, the samples were transferred to 1.5 mL EP tubes and 15 μL of 10% TFA was added to each tube. The final concentration of TFA was 1% and the pH was adjusted to 2-3.

[0076] 2.3 TMT-based quantitative proteomics analysis

[0077] (1) Desalination was performed by solid phase extraction (SPE) and vacuum freeze drying.

[0078] (2) Peptide labeling was performed using the TMT Kit (Thermo) according to the operating instructions. 7 μg of the enzymatic peptide fragments of each sample were accurately weighed, concentrated and dried by vacuum centrifugation, and then reconstituted. 2.8 μL of pre-activated labeling reagent (concentration 20 μg / μL) was added at a labeling ratio of 1:8. The mixture was placed in a thermomixer and the reaction parameters were set to 25°C, 1200 rpm, and incubated for 90 minutes.

[0079] (3) Add 0.5 μl of 5% hydroxylamine and incubate at 25°C for 15 minutes to terminate the reaction.

[0080] (4) Fractionation was performed using acetonitrile and 10 mM ammonium hydroxide solution (pH = 10.0).

[0081] 2.4 LC-MS analysis of exosomal proteins

[0082] (1) The samples were analyzed using a Vanquish Neo UHPLC system equipped with FAIMS Pro TM (Thermo Scientific TM The samples were analyzed by Orbitrap Exploris 480 mass spectrometer (ThermoScientific, San Jose, USA) interfaced with the instrument for data-dependent acquisition (DDA).

[0083] (2) Mobile phase A: 2% acetonitrile, 98% water, 0.1% formic acid; mobile phase B: 20% water, 80% acetonitrile, 0.1% formic acid.

[0084] (3) The following mass spectrometry parameters were used for the scan: FAIMS voltage was -45 V and -65 V, and the cycle time was 1 s. The primary scan range was 375–1800 m / z, the resolution was 60,000 (at 200 m / z), the normalized AGC target was set to 300%, the maximum ion injection time (maxIT) was 50 ms, and ions with a signal intensity lower than 2e4 were not captured. The maximum ion injection cycle was set to 86 ms, the normalized AGC target was set to 200%, the parent ion selection isolation window was set to 0.7 m / z, the secondary resolution was set to 30,000, the first mass of the scan was set to 110 m / z, and Turbo TMT was set to TMTpro Reagent. The peptides were introduced into the column (3 μm, 100 Å, 20 mm*75 μm id) at a flow rate of 6 μL / min. After entering the analytical column (1.9 μm, 120A, 150 mm*75 μm), analysis was performed using a 60-min HPLC gradient (8% to 35% mobile phase B) at a flow rate of 300 nL / min.

[0085] 2.5PRM Verification Experiment

[0086] The obtained differential protein labeling was verified by the parallel reaction monitoring (PRM) technique. Before starting the parallel reaction monitoring (PRM) analysis, preparations including protein extraction and trypsin digestion as described in the previous procedure were performed.

[0087] (1) Mobile phase A: 2% acetonitrile, 98% water, 0.1% formic acid; Mobile phase B: 20% water, 80% acetonitrile, 0.1% formic acid.

[0088] (2) Mass spectrometry scanning was performed using the following parameters: FAIMS voltages were -45 V and -65 V. The primary scan range was 350–1010 m / z, the resolution was 60,000 (at 200 m / z), the maximum ion injection time (max IT) was 100 ms, the normalized AGC target was set to 300%, the second-order resolving power was 30,000, the standard AGC objective was set to 1000%, and the optimal ionization time was 100 ms.

[0089] (3) The peptide entered the precolumn (3 μm, 100 Å, 20 mm*75 μm) at a flow rate of 6 μL / min. It then entered the analytical column (1.9 μm, 120 Å, 150 mm*75 μm) at a flow rate of 300 nL / min and was analyzed using a 30-min HPLC gradient (8% to 35% mobile phase B).

[0090] 2.6 ELISA method to detect serum ANG, VCAM1, and MASP1 levels

[0091] Take the sample to be tested out of the refrigerator and let it stand at room temperature for about 15-30 minutes to achieve temperature equilibrium.

[0092] Set up standard wells and add standard samples: Add 50 μL of standard sample to each of the five standard wells. The concentrations in each well are as follows: MASP1 ELISA kit: 480 pg / mL, 240 pg / mL, 120 pg / mL, 60 pg / mL, 30 pg / mL, and 0 pg / mL; VCAM1 ELISA kit: 480 ng / mL, 240 ng / mL, 120 ng / mL, 60 ng / mL, 30 ng / mL, and 0 ng / mL; ANG ELISA kit: 160 pg / mL, 80 pg / mL, 40 pg / mL, 20 pg / mL, 10 pg / mL, and 0 pg / mL.

[0093] Set up standard and sample wells: Add 50 μL of the gradient concentration standard solution to each standard well. Using a two-step loading process, first add 10 μL of the test serum sample to the experimental well, followed by 40 μL of sample dilution buffer (PBS containing 0.1% BSA). Leave blank wells unloaded.

[0094] Enzyme addition and incubation: Use a multichannel pipette to quantitatively add 100 μL of horseradish peroxidase-labeled secondary antibody (HRP) to all non-blank wells, cover with anti-evaporation sealing film, and incubate in a constant temperature incubator at 37°C for 60 minutes with humidity maintained at 80%.

[0095] Washing buffer preparation: Mix 20x concentrated washing buffer with ultrapure water at a volume ratio of 1:19 and use immediately.

[0096] Washing: ① Pour the liquid in the wells and tap the plate rack vertically three times on the sterile filter paper; ② Fill each well with diluted washing solution and let it stand for 60±5 seconds; ③ Repeat the pouring and tapping operation to complete five complete cycles.

[0097] Color development: sequentially add 50 μL of color developer A, followed by 50 μL of color developer B, and transfer to a light-proof constant temperature shaker at 37°C for reaction for 15 minutes.

[0098] Stop: Add 50 μL of stop solution to each well, immediately place in the microplate reader, and complete absorbance acquisition within 15 minutes.

[0099] Determination: Baseline calibration was performed using the blank well as a reference, and the absorbance value of each well was read at a dominant wavelength of 450 nm. The data collection interval was controlled within 30 seconds.

[0100] Calculate sample concentration: Use the results of the standard to draw a standard curve and calculate the sample concentration.

[0101] The ELISA detection procedures for VCAM1 and ANG in serum are the same as those for MASP1 detection described above, with only the specific capture antibodies being replaced.

[0102] 3. Statistical methods

[0103] 1) Data were analyzed using SPSS 26 statistical software. Kolmogorov-Smirnov tests (significance level α = 0.05) were performed for all measured data (e.g., HbA1c, eGFR, and protein expression). If the P value was > 0.05, the data were considered normally distributed and presented as mean ± standard deviation (mean ± SD); otherwise, the median (interquartile range) was used. For non-normally distributed data, outliers (exceeding 1.5 times the interquartile range) were removed using boxplot analysis. For data between the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group, independent sample t-tests (α = 0.05) were used to compare differences between groups for variables that were normally distributed; non-normal variables were compared using the Mann-Whitney U test. P values ​​were calculated, and statistically significant differences were indicated (P < 0.05).

[0104] 2) Calculate the protein expression ratio (Fold Change, FC) between the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group, and take the base 2 logarithm (Log2 FC). If the above calculation results meet both: P < 0.05 and fold change > 1.2, the protein is considered to have a statistically significant difference between the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group.

[0105] If the proteomic analysis met both of the following criteria: the fold difference was greater than 1.2-fold and the P value of the hypothesis test was lower than the significance threshold of 0.05, the protein was considered to be statistically different between the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group.

[0106] Skyline (22.2.0.527) was used to analyze the mass spectrometry data of 16 samples, and a peptide matrix was derived from which protein, peptide, and other information was extracted. The peptide matrix was converted into a protein matrix (the total area value of each peptide was summed), and the protein matrix was subsequently used for correlation analysis.

[0107] 4. Machine learning methods for building diagnostic models

[0108] The missing data percentage in this study was less than 15%. Therefore, missing values ​​in the dataset were filled using mean imputation. The specific rules are as follows: For continuous variables (e.g., HbA1c, eGFR), the mean of the training set was calculated to impute missing values ​​to avoid information leakage in the test set. For categorical variables (e.g., smoking history, medication use), the mode was imputed to preserve the original distribution characteristics. Pycharm 2022.1.4 was used for feature selection, data preparation, balancing, modeling, and evaluation. The dataset was randomly divided into a training set and a validation set with a 7:3 ratio. The training set was used for model construction, hyperparameter tuning, and cross-validation, while the validation set was used for final model performance evaluation to avoid overfitting. Predictive models were constructed using the machine learning algorithms logistic regression, support vector machine, and naive Bayesian algorithm. After constructing the diabetic peripheral neuropathy prediction model using these algorithms, 5-fold cross-validation was used, and model performance was evaluated using accuracy, precision, recall, F1 score, and area under the receiver operating characteristic curve (AUC). Finally, permutation importance is used to select features for disease prediction model.

[0109] Example 1 Patient Condition

[0110] In the TMT labeling quantification experiment and PRM targeted quantification experiment, a total of 30 patients were included, of which 5 patients met the exclusion criteria, 5 patients refused to sign the informed consent form, and 4 patients failed to collect blood samples. A total of 16 patients were included in the study analysis, including 11 patients with diabetic peripheral neuropathy and 5 patients with diabetes without peripheral neuropathy. In the ELISA validation experiment, a total of 66 patients were included, of which 3 patients met the exclusion criteria, 5 patients refused to sign the informed consent form, and 2 patients failed to collect blood samples. A total of 56 patients were included in the study analysis, including 29 patients with diabetic peripheral neuropathy and 27 patients with diabetes without peripheral neuropathy.

[0111] Example 2 Clinical data of selected patients

[0112] Among the 16 patients enrolled, 5 were diabetic and 11 were diabetic neuropathy patients. Gender, albumin, bilirubin Among the 16 patients enrolled, 11 were diabetic peripheral neuropathy patients and 5 were diabetic without peripheral neuropathy patients. There was no statistically significant difference in gender, albumin, bilirubin, glycated hemoglobin, alanine aminotransferase, aspartate aminotransferase, blood creatinine, blood urea nitrogen, fasting blood glucose, total cholesterol, triglycerides, C-reactive protein and other parameters between the groups (P>0.05). The demographic characteristics and biomarker characteristics of the diabetic peripheral neuropathy group and the diabetic without peripheral neuropathy group are shown in Table 1. Age, low-density lipoprotein, platelets, common peroneal nerve motor fiber conduction amplitude, superficial peroneal nerve sensory fiber conduction velocity, and superficial peroneal nerve sensory fiber conduction amplitude were statistically significant (P<0.05). The PRM targeted quantitative experiment performed a relative quantitative analysis of the differentially expressed proteins in serum exosomes of patients with diabetic peripheral neuropathy and patients with diabetic without peripheral neuropathy. The results showed that serum MASP1( Figure 1 )、VCAM1( Figure 2 )、ANG( Figure 3 ) concentrations in diabetic patients without peripheral neuropathy were significantly higher than those in diabetic patients without peripheral neuropathy (P<0.05).

[0113] Table 1 Basic characteristics of the participants

[0114]

[0115]

[0116] Note: Unless otherwise stated, all parameters are expressed as mean ± standard deviation. Indicates P < 0.05, P < 0.05 is the significance level, ANG, VCAM1, and MASP1 are PRM relative quantitative values.

[0117] Of the 56 patients enrolled, 29 had diabetic peripheral neuropathy and 27 had diabetic non-peripheral neuropathy. There were no statistically significant differences in age, diastolic blood pressure, albumin, bilirubin, bile acid, alanine aminotransferase, aspartate aminotransferase, serum creatinine, and blood urea nitrogen between the two groups (P>0.05). The demographic characteristics and biomarker characteristics of the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group are shown in Table 2. DD dimer, MASP1, VCAM1, and ANG were statistically significant (P<0.05). Serum MASP1 ( Figure 4 ),VCAM1( Figure 5 ) and ANG( Figure 6 The average concentrations of MASP1 in the diabetic group without peripheral neuropathy were 60.21±8.46pg / mL, 35.25±3.59ng / mL, and 29.27±1.91pg / mL, respectively. Figure 4 )、VCAM1( Figure 5 ) and ANG( Figure 6 The mean concentrations of MASP1, VCAM1 and ANG in serum were 38.99±15.60pg / mL, 40.88±12.23ng / mL and 25.88±5.02pg / mL respectively. The serum concentrations of MASP1, VCAM1 and ANG were statistically significant in the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group (P<0.05).

[0118] Table 2 Basic characteristics of the participants

[0119]

[0120]

[0121] Note: Unless otherwise specified, all parameters are expressed as mean ± standard deviation. P < 0.05 was considered significant. Indicates P<0.05.

[0122] Example 3 Analysis of relative quantitative results of target protein PRM

[0123] Results from a quantitative proteomics analysis (TMT) revealed 49 differentially expressed proteins between the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group, with 45 differentially expressed proteins upregulated and 4 differentially expressed proteins downregulated. To further validate the proteins identified by quantitative proteomics, we conducted a literature search. Finally, we selected 20 differentially expressed proteins closely associated with diabetic peripheral neuropathy, neurodegenerative diseases, and neuroregenerative disorders for PRM validation experiments. These proteins included VCAM1, CAT, ANG, SERPINE1, RAB10, MASP1, CAP1, YWHAE, CFL1, HLA-B, HSP90AB1, YWHAH, PPIA, YWHAB, ARPC3, COTL1, PSMA3, ARP2, CORO1A, and RAB14. Quantitative validation of the PRM method showed good correlation with previous TMT results. Differentially expressed proteins between the diabetic peripheral neuropathy group and the non-diabetic peripheral neuropathy group were selected based on the following criteria: log2|fold-change|≥1.5; ≤0.67; and p<0.05. The results showed that in the PRM targeted quantitative proteomics experiment, a total of 16 proteins were identified, including: HSD17B10, YWHAH, PPIA, YWHAE, RAB14, RAB10, CAPZA1, MASP1, CORO1A, PSMA3, CFL1, VCAM1, SERPINE1, ANG, FGG, ARPC3. These 16 proteins showed an up-regulation or down-regulation trend similar to that of the TMT quantitative proteomics experiment. In the diabetic peripheral neuropathy group and the diabetic non-peripheral neuropathy group, 14 proteins were up-regulated and 2 proteins were down-regulated. The protein quantitative results were statistically displayed in a volcano plot, see Figure 7 Hierarchical cluster analysis was applied to obtain a dynamic map of differentially expressed proteins, which showed good parallelism both within and between groups ( Figure 8 Subsequent database searches revealed that all 16 differentially expressed proteins validated by PRM were associated with diabetic peripheral neuropathy. Furthermore, three differentially expressed proteins, MASP1, VCAM1, and ANG, were selected for further statistical analysis based on clinical practicality, fold change, protein function, the number of literature related to neuropathy, molecular weight, whether previous literature supported ELISA experiments, and serum concentrations.

[0124] Example 4 Machine Learning Model Establishment and Analysis

[0125] Three differentially expressed proteins, MASP1, VCAM1, and ANG, were selected as markers, and a prediction model for diabetic peripheral neuropathy was established using logistic regression, support vector machine, and naive Bayes. The optimal parameter modeling was determined by grid search and cross-validation methods, and then feature importance screening was performed. The accuracy of the logistic regression model was 0.83, precision was 0.90, recall was 0.82, F1 value was 0.86, and AUC value was 0.79. The accuracy of the support vector machine was 0.83, precision was 0.79, recall was 0.78, F1 value was 0.88, and AUC value was 0.78. The accuracy of the naive Bayes was 0.83, precision was 0.83, recall was 0.91, F1 value was 0.87, and AUC value was 0.87 ( Figure 9-11 ). Feature importance screening performed a visual analysis of important variables, including MASP1, DD dimer, HbA1c, AST, Scr, TG, TC, VCAM1, TBA, ANG ( Figure 12 ).

Claims

1. A marker combination comprising human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1).

2. Use of the marker combination according to claim 1 in the preparation of a drug for predicting and / or diagnosing diabetes, wherein the diabetes is diabetic peripheral neuropathy or diabetic without peripheral neuropathy.

3. The use according to claim 2, characterized in that There is a significant difference in the expression level of the marker combination between diabetic peripheral neuropathy and diabetic non-peripheral neuropathy.

4. Use of a reagent for determining a marker combination in the preparation of a kit for determining diabetic peripheral neuropathy in a subject, wherein: The marker combination includes human angiotensin (ANG), human vascular endothelial cell adhesion molecule 1 (VCAM1), and human mannan-binding lectin serine peptide 1 (MASP1).

5. The use according to claim 4, characterized in that The marker combination is obtained by obtaining a serum sample from a subject.

6. The use according to claim 5, characterized in that The reagent is a reagent for determining the level of a marker; for example, the reagent is a reagent for use in parallel reaction monitoring, tandem mass tags, or ELISA.

7. A reagent for detecting a marker combination, characterized in that The marker combination is as defined in claim 1, and the reagent is a reagent for determining the level of the marker, such as a reagent used in parallel reaction monitoring, tandem mass tag or ELISA; preferably, the marker is from a serum sample.

8. A diagnostic kit for determining whether a subject has diabetes, comprising the reagent according to claim 7 and a control substance, wherein the diabetes is diabetic peripheral neuropathy or diabetic without peripheral neuropathy.

9. A diabetes diagnosis system, characterized in that: The diagnostic system includes a detection module and an analysis and judgment module; the detection module detects the level of a marker combination in the serum of a subject and transmits the level data to the analysis and judgment module; the analysis and judgment module judges the level data of the marker combination and outputs a diagnosis result: the subject has diabetic peripheral neuropathy or diabetic without peripheral neuropathy; wherein, the marker combination is as defined in claim 1.

10. The diagnostic system according to claim 9, wherein: The analyzing and judging module judging the level data of the marker combination includes: determining whether there is a significant difference in the expression level of the marker combination between diabetic peripheral neuropathy and diabetic non-peripheral neuropathy.

11. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the functions of the diagnostic system according to claim 9 or 10 can be realized.

12. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: The processor is configured to execute the computer program to implement the functions of the diagnostic system according to claim 9 or 10.

Images