Application of extracellular vesicle membrane proteins in the preparation of early gastric cancer diagnostic kits

Abstract

This invention discloses the application of extracellular vesicle membrane proteins in the preparation of diagnostic kits for early gastric cancer. Using proteomics technology and the EVArray platform, this invention screened a series of novel biomarkers for EV membrane proteins: transforming growth factor β1, annexin A2, tendinin N, phospholipase-like domain protein 8, transferrin, fibroblast activating protein α, and cadherin 5. Clinical trials have demonstrated the potential of this invention for early diagnosis and postoperative monitoring of gastric cancer (GC). Among these, transforming growth factor β1 and tendinin N showed the most significant value in the early diagnosis of GC.

Description

Technical Field

[0001] This invention belongs to the field of medical biological detection, and specifically relates to the application of extracellular vesicle membrane proteins in the preparation of early gastric cancer diagnostic kits. Background Technology

[0002] Gastric cancer (GC) is one of the most common malignant tumors worldwide, ranking among the top in both incidence and mortality. Clinical data shows that the five-year survival rate for early-stage GC patients can reach over 90%, while it drops sharply to below 30% for late-stage patients. This significant difference highlights the crucial role of early diagnosis and treatment. Current screening methods, such as fecal occult blood tests and serum biomarkers (CEA, CA199, etc.), suffer from insufficient sensitivity and limited specificity. Although gastroscopy and biopsy are considered the gold standard for diagnosis, their invasiveness and the possibility of missed diagnoses affect their effectiveness in early screening. Therefore, exploring novel circulating biomarkers with high sensitivity and specificity is of great value in improving the early diagnosis and postoperative monitoring of gastric cancer.

[0003] Extracellular vesicles (EVs) are nanoscale lipid bilayer membrane structures released by cells, containing various bioactive molecules derived from the parent cell, including proteins, nucleic acids (DNA, RNA), and lipids, and can stably exist in bodily fluids such as blood, urine, and saliva. Among the components carried by EVs, functional nucleic acids and bioactive proteins are two types of biomarkers of great interest, especially EV membrane proteins. Studies have shown that these membrane proteins, as key effector molecules, play an important role in the occurrence and development of cancer, and therefore hold promise as novel biomarkers for early cancer diagnosis.

[0004] Traditional strategies for discovering EV membrane protein biomarkers primarily involve genomics and proteomics. While gene expression studies can provide information, they are not always correlated with the abundance and variation of the encoded protein, which is often affected by proteolytic cleavage or post-translational modifications. Proteomics approaches focus on characterizing protein expression and alterations, providing crucial information on protein dynamics, including function, post-translational modifications, interactions with other biomolecules, and responses to environmental factors. This allows for direct, in-depth, and quantitative analysis of the expression levels of various cancer-related proteins and cancer-specific proteins. With the rapid advancements in mass spectrometry (MS) detection techniques and data processing methods, MS-based high-throughput and high-precision proteomics technologies have developed rapidly. Membrane protein biomarkers play a vital role in early tumor diagnosis and treatment efficacy assessment, and research using proteomics to identify early cancer biomarkers is increasingly widespread. Based on whether the target protein is quantified, commonly used proteomics quantitative techniques can be divided into two main categories: non-targeted quantitative proteomics and targeted quantitative proteomics. Non-targeted quantitative proteomics involves the indiscriminate quantitative analysis of all proteins in a sample, aiming to discover and detect more proteins in the sample, thereby generating a list of candidate biomarkers that forms a crucial bridge for subsequent stages of analysis. It is suitable for the early stages of biomarker discovery research. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings and deficiencies of the prior art and to provide the application of extracellular vesicle membrane proteins in the preparation of early gastric cancer diagnostic kits.

[0006] The objective of this invention is achieved through the following technical solution: the application of extracellular vesicle membrane proteins in the preparation of early gastric cancer diagnostic kits, wherein the extracellular vesicle membrane proteins are at least one of transforming growth factor beta1 (TGF-β1), annexin A2 (ANXA2), tenascin N (TNN), pattin-like phospholipase domain protein 8 (PNPLA8), transferrin (TF), fibroblast activation protein alpha (FAP), and cadherin 5 (CDH5); preferably at least one of transforming growth factor beta1 and tenascin N.

[0007] The aforementioned early-stage gastric cancer refers to a tumor that only invades the mucosal layer or submucosa of the stomach, without breaking through the muscular layer of the stomach wall.

[0008] The kit contains reagents for determining the content of extracellular vesicle membrane proteins.

[0009] The reagents used to determine the content of extracellular vesicle membrane proteins include TGF-β1 antibody-oligonucleotide complex, TNN antibody-oligonucleotide complex, and CD9 / CD63 / CD81 functionalized magnetic beads.

[0010] The reagent used to determine the content of extracellular vesicle membrane proteins also contains RNase.

[0011] The RNase is preferably RNase A / T1; more preferably RNase A / T1 at 5000 U / mL.

[0012] The TGF-β1 antibody-oligonucleotide complex is a complex obtained by sequentially linking a TGF-β1 antibody, a linker, a linker-linked DNA, RNA that can be specifically cleaved by RNase, and template DNA for initiating signal amplification.

[0013] The TNN antibody-oligonucleotide complex is a complex obtained by sequentially linking a TNN antibody, a linker, a linker-linked DNA, RNA that can be specifically cleaved by RNase, and template DNA for initiating signal amplification.

[0014] The linker ligation DNA sequence is as follows: 5'-AAGTATT / ACCAGAAA-3'; where the slash is the position where dRep specifically recognizes, cuts, and ligates single-stranded DNA.

[0015] The RNA sequence that can be specifically cleaved by RNase is as follows: 5'-GCUGUG-3'.

[0016] The template DNA used for initiation signal amplification can be arbitrarily designed, preferably without linker-linking DNA and easy to amplify; a preferred nucleic acid sequence is one of the following:

[0017] 5'-TGTTGTAAGGGCCCGTGACTATGTCGAAGCGACCCGGCGATATAATCATTTCCACGCCCGTC-3';

[0018] 5'-CATAGGAGAAACTGAGATGCCAACTGTGATGAATGGGCTTATGGTTTGGTGCATTGAAAATGGAACCTCGCCA-3'.

[0019] The CD9 / CD63 / CD81 functionalized magnetic beads are preferably prepared by the following method:

[0020] 1) Sulfo-NHS-LC-Biotin was incubated with CD9 antibody, CD63 antibody, and CD81 antibody to obtain biotinylated antibody;

[0021] 2) Biotinylated antibodies were conjugated with streptavidin magnetic beads to obtain CD9 / CD63 / CD81 functionalized magnetic beads.

[0022] The biotinylated antibodies described in step 1) can be mixed before biotinylation, or the antibodies can be biotinylated separately and then mixed.

[0023] The amount of Sulfo-NHS-LC-Biotin used in step 1) is preferably calculated as 2.4-2.7 μg antibody to 0.7 nmol; more preferably, it is calculated as 2.6-2.7 μg antibody to 0.7 nmol.

[0024] The preferred incubation operation described in step 1) is as follows: first incubate at 20-30℃ for 20-40 min, then transfer to 2-8℃ for 12-20 h; more preferably, incubate at 24-26℃ for 25-35 min, then transfer to 4℃ for 12-16 h.

[0025] The preferred method for conjugation in step 2) is as follows: dilute the biotinylate antibody with PBS, add streptavidin magnetic beads that have been washed with PBS and incubate; after incubation, wash with PBS and then block with bovine serum albumin; after blocking, wash with PBS and then resuspend in PBS.

[0026] The amount of magnetic beads used in step 2) is preferably calculated based on a ratio of 2.4 to 2.7 μg of antibody to 1 mg of magnetic beads.

[0027] The present invention has the following advantages and effects compared with the prior art:

[0028] This invention utilizes proteomics technology and the EVArray platform to screen a series of novel biomarkers for EV membrane proteins. Clinical validation has confirmed their potential application in the early diagnosis and postoperative monitoring of gastric cancer (GC). Transforming growth factor β1 and tendinin N, in particular, showed the most significant value in the early diagnosis of GC. Attached Figure Description

[0029] Figure 1 The flowchart (A) shows the proteomics analysis of early GC diagnostic markers, and the preliminary validation diagram (B) shows the seven EV membrane protein candidate markers in GC-EVArray detection.

[0030] Figure 2This is a proteomic analysis diagram of early GC diagnostic markers; in A, two Venn diagrams show the intersection of upregulated proteins identified by proteomic analysis with the UniProt public membrane protein database (top) and the intersection of upregulated membrane proteins with the Vesiclepedia EV protein database (bottom); B is a volcano diagram of 48 differentially expressed EV membrane proteins; C is a heatmap showing 48 differentially expressed EV membrane proteins in proteomic analysis; D is a PCA score diagram of EV membrane proteins, showing a clear separation between healthy controls and stage I GC patients.

[0031] Figure 3 shows the KEGG and GO analysis of EV membrane proteins from healthy individuals and patients with stage I GC; where A is the KEGG pathway analysis of differentially expressed EV membrane proteins, listing the top 10 enriched pathways; B is the GO classification of differentially expressed EV membrane proteins, listing the top 10 enriched items in biological processes (BP), cellular components (CC), and molecular functions (MF).

[0032] Figure 4 This is a box plot of the expression levels of seven candidate biomarkers (TGF-β1, TNN, PNPLA8, CDH5, TF, FAP, and ANXA2); where ** indicates P<0.05, ** indicates P<0.01, *** indicates P<0.001, and **** indicates P<0.0001.

[0033] Figure 5 This figure shows the discovery and preliminary validation results of early GC diagnostic biomarkers using GC-EVArray. A represents the representative fluorescence images of one healthy individual and one patient with stage I GC detected by GC-EVArray; B is a heatmap of seven EV membrane proteins detected by GC-EVArray; C–I represent the relative expression levels of TNN (C), TGF-β1 (D), ANXA2 (E), PNPLA8 (F), FAP (G), CDH5 (H), and TF (I) detected by GC-EVArray, respectively; * indicates P < 0.05, ** indicates P < 0.01, and ns indicates no statistically significant difference.

[0034] Figure 6 shows the ROC curves of various markers on EVs used to distinguish between healthy individuals and stage I GC patients; where, in the GC-EVArray, A is TNN (AUC=0.86, 95% CI=69.35%-100%), B is TGF-β1 (AUC=0.82, 95% CI=61.33%-100%), C is ANXA2 (AUC=0.81, 95% CI=60.56%-100%), D is PNPLA8 (AUC=0.78, 95% CI=57.61%-98.39%), E is FAP (AUC=0.73, 95% CI=50.53%-95.47%), F is CDH5 (AUC=0.70, 95% CI=46.38%-93.62%), and G is TF (AUC=0.74, 95% CI=46.38%-93.62%). 95% CI = 51.83% - 69.17%.

[0035] Figure 7 This is a graph showing the diagnostic performance evaluation results of early GC diagnostic markers. A is a heatmap displaying the expression of TGF-β1 and TNN in plasma samples, including healthy controls (n=20), patients with benign disease (n=20), stage I GC (n=30), stage II GC (n=30), and stage III-IV GC (n=30). B–E are scatter plots showing the concentrations of TGF-β1 (B, D) and TNN (C, E) in different groups, based on droplet detection (*P<0.05, **P<0.01, ****P<0.0001, ns indicates no statistically significant difference). F–M are ROC curve analyses evaluating the diagnostic performance of TGF-β1 and TNN in distinguishing healthy individuals and patients with benign disease from patients with different stages of GC, including stage I GC (F, G), stage II GC (H, I), stage III-IV GC (J, K), and all GC patients (L, M).

[0036] Figure 8 These are ROC curves used to distinguish healthy individuals or benign patients from different stages and all GC patients; where A is the ROC curve used to distinguish healthy individuals from stage I GC patients, B is the ROC curve used to distinguish benign patients from stage I GC patients, C is the ROC curve used to distinguish healthy individuals from stage II GC patients, D is the ROC curve used to distinguish benign patients from stage II GC patients, E is the ROC curve used to distinguish healthy individuals from stage III-IV GC patients, F is the ROC curve used to distinguish benign patients from stage III-IV GC patients, G is the ROC curve used to distinguish healthy individuals from all GC patients, and H is the ROC curve used to distinguish benign patients from all GC patients.

[0037] Figure 9 shows the results of single-molecule analysis of EV membrane proteins in postoperative monitoring; where A is the paired connectivity diagram of TGF-β1 before and after surgery in GC patients, and B is the paired connectivity diagram of TNN before and after surgery in GC patients; n = 20, ****p<0.0001, ns indicates no statistically significant difference.

[0038] Figure 10 These are photographs (A) and curves (B) showing tumor growth in tumor-bearing mice at different time points.

[0039] Figure 11 This is a graph showing the application results of single-molecule analysis of EV membrane proteins in tumor burden assessment; where A is a schematic diagram of mouse blood collection; B and E are the concentrations of EGFR (B), MUC1 (C), TGF-β1 (D), and TNN (E) detected in three mice at each time point, respectively. Detailed Implementation

[0040] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0041] Example 1

[0042] 1. Materials and Methods

[0043] 1.1 Materials

[0044] 1.1.1 Research Subjects

[0045] The clinical plasma samples used in this study were approved by the Ethics Committee of Nanfang Hospital, Southern Medical University. Plasma samples from healthy individuals and patients with benign gastritis were collected from the Department of Laboratory Medicine, Nanfang Hospital, Southern Medical University (Guangzhou, China), while plasma samples from patients with gastric globus genital heart disease (GC) were collected from the Department of General Surgery, Nanfang Hospital. The diagnoses of both benign gastritis and GC were determined based on pathological examination of tissue biopsies.

[0046] 1.1.2 Main Reagents and Consumables

[0047] Table 1. Main Reagents and Consumables

[0048]

[0049] Table 2 Summary of Instrument Names and Manufacturers

[0050]

[0051] Table 3. List of nucleic acid sequences required for the experiment

[0052]

[0053] 1.2 Methods

[0054] 1.2.1 Collection and Preservation of Plasma Samples

[0055] Venous blood samples were collected using EDTA blood collection tubes and centrifuged at 2,500g for 15 minutes at room temperature, repeated twice, to remove excess blood cells. The resulting plasma was then aliquoted and stored at -80°C for subsequent analysis.

[0056] 1.2.2 Extraction of plasma EVs

[0057] In this study, mixed plasma from healthy individuals was diluted 1:2 in 0.01M, pH 7.4 phosphate-buffered saline (PBS) and centrifuged at 3,000g for 20 min to remove cell debris. The supernatant was then centrifuged at 16,000g for 30 min to remove microvesicles. The resulting supernatant was further centrifuged twice at 135,000g for 70 min using ultracentrifugation (UC) to separate EVs. The obtained EVs were resuspended in 100 μL PBS. All centrifugation steps were performed at 4°C. The EVs were finally stored at -80°C for subsequent analysis.

[0058] 1.2.3 Protein Sample Preparation Procedure for Mass Spectrometry Analysis

[0059] First, thaw the samples on ice and add the sample solution to an ultrafiltration tube. Add 200 µL of 8M urea solution for protein extraction. After protein extraction, add dithiothreitol (DTT) solution to a final concentration of 10 mM and react at 37°C for 30 minutes. Add iodoacetamide (IAA) solution to a final concentration of 20 mM and react at room temperature in the dark for 30 minutes. Add 200 µL of 50 mM ammonium bicarbonate (NH4HCO3) and centrifuge at 12,000g for 10 minutes. Add 200 µL of 50 mM NH4HCO3 and 1 µg of LysC endopeptide to each sample and incubate at 37°C with shaking for 2 hours. Add 1 µg of trypsin and incubate at 37°C with shaking overnight. Terminate the reaction by adding 10 µL of 10% v / v trifluoroacetic acid (TFA) to the reaction system. Peptide desalting was then performed. 100 µL of methanol was added to a SoLAµ HRP plate and centrifuged at 600g for 1 minute. 100 µL of 80% v / v acetonitrile (ACN) / 0.1% v / v TFA was added and centrifuged at 1,000g for 1 minute. 200 µL of 0.1% v / v TFA was added and centrifuged at 1,000g for 1 minute. The sample was then added to a SoLAµ HRP plate and centrifuged at 1,000g for 2 minutes. This process was repeated once. 200 µL of 0.1% v / v TFA was added and centrifuged at 1,000g for 2 minutes. 100 µL of 80% ACN / 0.1% TFA was added and centrifuged at 1,000g for 3 minutes. The eluent was collected. The eluent was evaporated to dryness at 40°C in a centrifuge and dissolved in 0.1% formic acid (FA) to a concentration of 0.5 µg / µL. 2 µg of the eluent was loaded onto the mass spectrometer for quantitative analysis.

[0060] 1.2.4 Mass Spectrometry Data Processing

[0061] DIA Data Analysis: Data was imported into DIA-NN v1.8.0 and targeted extraction was performed using a predicted human proteomics database. Remaining parameters used default settings to control the false discovery rate (FDR) for peptides and proteins to <1%. Statistical Analysis: The output was protein intensity calculated by DIA-NN, using the mean of the top three peptides. This protein intensity data was imported into Perseus and Metaboanalyst for statistical analysis. Differentially expressed proteins were selected based on a fold change of 2x (p < 0.05). Principal component analysis (PCA), cluster analysis, and correlation analysis were performed in R 4.0.5. GO and KEGG enrichment analyses were performed using Cluster Profile v3.18. Upregulated and downregulated proteins were imported into the String database for protein-protein interaction annotation, using interactions with a confidence level of 0.7. The membrane protein database UniProt and the EV database Vesiclepedia are from https: / / www.uniprot.org / and http: / / www.microvesicles.org / , respectively.

[0062] 1.2.5 GC-EVArray Detection Process

[0063] Each antibody microarray well was first blocked with 100 μL of 3% w / v BSA solution at room temperature for 30 minutes, then the solution was removed. Next, 100 μL of diluted sample was added to the wells and incubated for 60 minutes, followed by washing three times with 200 μL of PBS solution. Then, 100 μL of biotin-labeled detection antibody (diluted with PBS at a volume ratio of 1:1000 to 1:1500) was added and incubated for 30 minutes. Afterward, the wells were washed three times again with 200 μL of PBS, and 100 μL of Ly3-streptavidin (diluted with PBS at a volume ratio of 1:1500) was added. The microarray was incubated at room temperature in the dark for 30 minutes. After incubation, the supernatant was removed, and the wells were washed three times again with 200 μL of PBS. After air drying, the signal was scanned at 532 nm using a laser scanner. The fluorescence intensity was quantified using GenePix software.

[0064] 1.2.6 Single-molecule detection technology for EV membrane proteins based on droplet microfluidics

[0065] 2 μL of Ts1 or Ts2 (100 μM), 10 μL of 2× ssDNA buffer (100 mM HEPES, 300 mM NaCl, 2 mM MgCl2, 2 mM MnCl2, pH 8.0), 4.5 μL of ultrapure water, and 3.5 μL of Linker (1 mg / mL) were mixed and incubated at 37°C for 1 hour to obtain Linker-Ts1 and Linker-Ts2. 5 μL of Linker-Ts1 and 3 μL of TGF-β1 antibody were mixed and incubated at 4°C for 1 hour to obtain the TGF-β1 antibody-Ts1 complex. 5 μL of Linker-Ts2 and 3 μL of TNN antibody were mixed and incubated at 4°C for 1 hour to obtain the TNN antibody-Ts2 complex. The EGFR antibody-Ts2 complex and MUC1 antibody-Ts1 complex were obtained using the same method.

[0066] Sulfo-NHS-LC-Biotin was dissolved in PBS to a concentration of 10 mM, and then further diluted 100-fold with PBS to a final concentration of 0.1 mM. Subsequently, 7 μL of Sulfo-NHS-LC-Biotin was added to 1.65 μL of CD9 antibody (0.53 mg / mL), 1.65 μL of CD63 antibody (0.53 mg / mL), and 1.65 μL of CD81 antibody (0.53 mg / mL). After incubation at room temperature (25°C) for 30 minutes, the mixture was transferred to 4°C for overnight incubation (16 hours) to obtain the biotinylated antibody. Take 100 μL of streptavidin-coated magnetic beads at a concentration of 10 mg / mL and wash twice with 100 μL of PBS; add 100 μL of PBS to the biotinylated antibody solution, and then incubate the magnetic beads with 100 μL of antibody solution at room temperature for 50 minutes; subsequently, wash the magnetic beads three times with 100 μL of PBS, block with 100 μL of 1% bovine serum albumin (BSA) for 30 minutes, wash three times with 100 μL of PBS again, and finally resuspend in 100 μL of PBS to obtain CD9 / CD63 / CD81 functionalized magnetic beads.

[0067] 2 μL of CD9 / CD63 / CD81 functionalized magnetic beads at a concentration of 10 mg / mL were added to 50 μL of plasma. The mixture was incubated at room temperature for 1 hour on a rotary mixer to capture and enrich EVs. The magnetic beads containing the captured EVs were then magnetically retrieved using a magnetic rack. The sample was then washed twice with 50 μL of PBS to remove residual plasma. Next, 2.5 μL of a 60 pM TGF-β1 antibody-Ts1 complex and 2.5 μL of a 60 pM TNN antibody-Ts2 complex were added, and the volume was brought up to 50 μL with blocking buffer (1×PBS containing 5% BSA, 0.05% dextran sulfate, and 0.2 mg / mL salmon sperm DNA). The mixture was incubated at room temperature for 1 hour. After incubation, the sample was washed twice with Wash Buffer 1 (PBS containing 5% BSA and 0.05% dextran sulfate), followed by two washes with Wash Buffer 2 (PBS containing 0.1% salmon extract), 50 μL each time. Then, 0.1 μL of RNase A / T1 was added, and the volume was brought up to 50 μL with PBS. After incubation at room temperature for 20 minutes, the supernatant was collected for ddPCR detection. The gene amplification system was as follows: 10 μL 2× ddPCR universal probe mixture, 1.8 μL forward primer, 1.8 μL reverse primer, 0.6 μL 10× probe, 3.8 μL nuclease-free water, and 2 μL reaction solution containing the template strand (Ts1 used forward primer 1, reverse primer 1, and probe 1-CY5; Ts2 used forward primer 2, reverse primer 2, and probe 2-FAM). The amplification conditions were as follows: initial denaturation at 94°C for 5 minutes; denaturation at 94°C for 30 seconds, annealing at 58°C for 2 minutes, and extension for 45 cycles; finally, slow cooling to 4°C. The amplified droplets were injected into a droplet analysis chip, and the experimental results were read and analyzed using an ExoStar Droplet Reader.

[0068] 1.2.7 Preparation of tumor-bearing mice

[0069] The mice were injected with MKN28 human gastric cancer cells at a dose of 5 × 10⁶ per mouse. 6 Cells were injected subaxillarily into three 5-week-old nude mice, and the same procedure was performed. Every seven days after model establishment, 200 μL of blood was collected from the submandibular region of the mice. The plasma samples were processed and stored using the steps described above, and subsequently analyzed by ddPCR.

[0070] 2. Screening process for EV membrane protein biomarkers

[0071] 2.1 Proteomics Technology

[0072] To explore early biomarkers for gastric cancer, we performed proteomic analysis on plasma-derived EVs from an early GC cohort, which included 10 healthy individuals and 10 patients with stage I GC. Figure 1 (A) Compared with healthy controls, we identified 325 differentially expressed proteins in stage I GC patients, of which 178 were upregulated and 147 were downregulated. By cross-aligning the EV protein dataset with the public membrane protein database Uniprot, we screened out 48 differentially expressed membrane proteins, of which 23 were upregulated. Further cross-alignment of these 23 membrane proteins with the public EV protein database Vesiclepedia revealed that 22 of them were present in the EV protein database (see A). Figure 2 (A in the middle). Figure 2 B in the figure represents a volcano diagram of 48 differentially expressed EV membrane proteins; Figure 2 The clustering heatmap shown in Figure C illustrates the expression of 48 differentially expressed membrane proteins in various samples, indicating that these screened proteins can effectively distinguish between the healthy group and the early GC group. Furthermore, PCA further reveals the differences between the two groups, highlighting the proteomic characteristics of the early GC samples and healthy controls (see Figure C). Figure 2 The D in the text provides a theoretical basis for the subsequent screening of new biomarkers.

[0073] Subsequently, we used bioinformatics methods to perform functional analysis on differentially expressed EV membrane proteins. Analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) indicated that these differentially expressed EV membrane proteins are primarily involved in cancer-related biological processes, such as cell adhesion molecules and neutrophil extracellular traps (see...). Figure 3 (A in the text). Furthermore, Gene Ontology (GO) analysis showed that, compared to the healthy control group, differentially expressed EV membrane proteins in the Phase I GC group exhibited specific signals associated with focal adhesion, cell-matrix adhesion, and cell adhesion regulation (see A in the text). Figure 3 (B in the text). These findings may reflect the functional roles and molecular heterogeneity of differentially expressed EV membrane proteins in GC occurrence and development.

[0074] Among the 22 upregulated EV membrane proteins, we screened seven differentially expressed membrane proteins as candidate biomarkers based on their fold change ranking and annotations from the Human Protein Atlas (HPA) database for further early GC analysis. These seven proteins include transforming growth factor beta1 (TGF-β1), annexin A2 (ANXA2), tenascin N (TNN), pattin-like phospholipase domain protein 8 (PNPLA8), transferrin (TF), fibroblast activation protein alpha (FAP), and cadherin 5 (CDH5). Figure 4 As shown, there were statistically significant differences in the expression levels of these seven EV membrane proteins between healthy donors and GC patients.

[0075] 2.2 EVArray Platform

[0076] Subsequently, in another early GC validation cohort (10 healthy individuals and 10 patients with stage I GC), we conducted preliminary validation of these seven EV membrane protein candidate biomarkers using the EVArray chip. A schematic diagram of the GC-EVArray chip is shown below. Figure 1 As shown in B in the diagram. By calculating the average signal intensity of the three replicates on the GC-EVArray and subtracting the signal from the PBS spot in the same well, we obtained the final expression readings for each membrane protein.

[0077] The layout of the GC-EVArray chip is as follows Figure 5 As shown in A in the diagram. Figure 5 Figure B in the graph is a heatmap showing the expression intensity of the seven biomarkers in the validation queue. The scatter plot further shows the expression distribution of the seven biomarkers in the GC-EVArray validation queue, where TNN, TGF-β1, ANXA2, and PNPLA8 showed statistically significant differences (see Figure B). Figure 5 In addition, we used ROC curve analysis to evaluate the AUC values ​​of seven biomarkers in distinguishing healthy individuals from stage I GC patients (see CI). Figure 6The results showed that the AUC values ​​of TNN (0.86, 95% CI = 69.35%–100%) and TGF-β1 (0.82, 95% CI = 61.33%–100%) were higher than those of the other five biomarkers. Currently, no studies have evaluated the application of EV-expressed TNN and TGF-β1 membrane proteins in the early diagnosis of GC. Therefore, we selected TNN and TGF-β1 as novel biomarkers for the early diagnosis of GC for further validation.

[0078] To evaluate the clinical diagnostic performance of the selected TGF-β1 and TNN, we applied droplet microfluidics-based single-molecule detection technology of EV membrane proteins to the GC cohort (n = 130), which included 20 healthy individuals (Helalthy), 20 patients with benign gastritis (Benigh), and 90 patients with malignant tumors (30 in stage I, 30 in stage II, and 30 in stages III-IV). Figure 7 Figure A shows a heatmap of the expression of the two markers in each sample. Figure 7 The BC diagram shows the concentrations of TGF-β1 and TNN in patients with different GC stages, compared with healthy and benign groups. It shows that although there was no significant difference in expression levels among different GC stages, the concentrations of markers in stage I and II patients were significantly higher than in healthy individuals and patients with benign gastritis. Furthermore, Figure 7 The DE analysis further illustrates the concentration distribution of TGF-β1 and TNN in the healthy group, benign group, and all GC patients, showing that the expression levels of the biomarkers in GC patients were significantly higher than those in the healthy and benign groups. Subsequently, we performed ROC curve analysis on TGF-β1 and TNN to evaluate their diagnostic performance in distinguishing patients with different GC stages from healthy individuals or patients with benign disease (see [link to analysis]). Figure 7 F–M in the model), and compared with traditional markers ( Figure 8When differentiating between healthy individuals and stage I gastric globulin (GC) patients, the AUC values ​​for TGF-β1 and TNN were 0.8883 (95% CI = 0.7937–0.9829) and 0.7550 (95% CI = 0.6117–0.8983), respectively. Combined detection of these two biomarkers increased the AUC value to 0.8900 (95% CI = 0.7958–0.9842). When differentiating between patients with benign gastric lesions and stage I GC patients, the AUC values ​​for TGF-β1 and TNN were 0.9000 (95% CI = 0.8162–0.9838) and 0.7550 (95% CI = 0.6225–0.8875), respectively. Combined detection of these two biomarkers yielded the highest AUC value, reaching 0.9050 (95% CI = 0.8236–0.9864). However, when differentiating between healthy individuals or benign patients and patients at different stages or all GC patients, the AUC values ​​of TGF-β1 alone and the combined detection of TGF-β1 and TNN were similar. This result may indicate that TGF-β1 has a greater potential than TNN in the early diagnosis of GC. Furthermore, in all the analyzed groups, the diagnostic performance of both TGF-β1 and TNN was significantly better than that of traditional markers (CEA and CA-199), further demonstrating the advantages of these two EV membrane proteins in the early diagnosis of GC.

[0079] To further evaluate the clinical application value of the two screened biomarkers, we dynamically monitored two EV membrane proteins (TGF-β1 and TNN) in plasma samples from patients with gastrointestinal tract infections (GC) using EV membrane protein single-molecule detection technology. A total of 20 patients were included in the study, and blood samples were collected 24 hours before surgery and 1 week after surgery. Changes in protein expression levels before and after surgery were compared. Results showed that TGF-β1 expression levels decreased significantly after surgery, with a statistically significant difference. Although TNN showed a decreasing trend in some patient samples, there was no statistically significant difference overall, which may be related to the failure of the TNN biomarker to respond promptly to surgical treatment (see [link to study]. Figure 9 ).

[0080] To evaluate the potential of this method in tumor progression monitoring, we validated it in a tumor-bearing mouse model. Figure 10 Figure A shows typical tumor growth in GC-bearing mice, with tumor volume measured every 3-4 days (see Figure 1). Figure 10 (B in the middle).

[0081] To investigate the expression changes of four EV membrane proteins in plasma, we collected submandibular vein blood samples from mice every 7 days and used this method to detect the concentrations of biomarkers in three mice at different time points (see [link to study]. Figure 11(A in the original text). The results showed that the concentrations of all four biomarkers increased with prolonged tumor bearing time. Furthermore, EGFR and MUC1 were still undetectable on day 7, while TGF-β1 showed an increasing trend in all three experimental mice, and TNN was detected in only one mouse (see [reference to another mouse]). Figure 11 (BE in the study). This result further validates the value of EV membrane proteins (especially TGF-β1) in the early diagnosis of GC. In conclusion, our method has significant application potential in postoperative monitoring and tumor progression assessment.

[0082] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. The application of reagents for detecting extracellular vesicle membrane proteins in the preparation of early gastric cancer diagnostic kits, characterized in that: The extracellular vesicle membrane protein is tendinin N, or a combination of transforming growth factor β1 and tendinin N.

2. The application of the reagent for detecting extracellular vesicle membrane proteins according to claim 1 in the preparation of an early gastric cancer diagnostic kit, characterized in that: The reagents for detecting extracellular vesicle membrane proteins include TGF-β1 antibody-oligonucleotide complex, TNN antibody-oligonucleotide complex, and CD9 / CD63 / CD81 functionalized magnetic beads.

3. The application of the reagent for detecting extracellular vesicle membrane proteins according to claim 2 in the preparation of an early gastric cancer diagnostic kit, characterized in that: The reagent for detecting extracellular vesicle membrane proteins also contains RNase.

4. The application of the reagent for detecting extracellular vesicle membrane proteins according to claim 3 in the preparation of an early gastric cancer diagnostic kit, characterized in that: The TGF-β1 antibody-oligonucleotide complex is a complex obtained by sequentially linking a TGF-β1 antibody, a linker, a linker-linked DNA, RNA that can be specifically cleaved by RNase, and template DNA for initiating signal amplification. The TNN antibody-oligonucleotide complex is a complex obtained by sequentially linking a TNN antibody, a linker, a linker-linked DNA, RNA that can be specifically cleaved by RNase, and template DNA for initiating signal amplification. The linker sequence for linking DNA is as follows: 5'-AAGTATT / ACCAGAAA-3'; The RNA sequence that can be specifically cleaved by RNase is as follows: 5'-GCUGUG-3'.

5. The application of the reagent for detecting extracellular vesicle membrane proteins according to claim 4 in the preparation of an early gastric cancer diagnostic kit, characterized in that: The template DNA used for initiating signal amplification is one of the nucleic acids with the following sequence: 5'-TGTTGTAAGGGCCCGTGACTATGTCGAAGCGACCCGGCGATATAATCATTTCCACGCCCGTC-3'; 5'-CATAGGAGAAACTGAGATGCCAACTGTGATGAATGGGCTTATGGTTTGGTGCATTGAAAATGGAACCTCGCCA-3'.

6. The application of the reagent for detecting extracellular vesicle membrane proteins according to any one of claims 2 to 5 in the preparation of a diagnostic kit for early gastric cancer, characterized in that: The CD9 / CD63 / CD81 functionalized magnetic beads are prepared by the following method: 1) Sulfo-NHS-LC-Biotin was incubated with CD9 antibody, CD63 antibody, and CD81 antibody to obtain biotinylated antibody; 2) Biotinylated antibodies were conjugated with streptavidin magnetic beads to obtain CD9 / CD63 / CD81 functionalized magnetic beads.

7. The application of the reagent for detecting extracellular vesicle membrane proteins according to claim 6 in the preparation of an early gastric cancer diagnostic kit, characterized in that: The dosage of Sulfo-NHS-LC-Biotin mentioned in step 1) is calculated based on 2.4–2.7 μg antibody to 0.7 nmol; The amount of magnetic beads used in step 2) is calculated based on a ratio of 2.4 to 2.7 μg of antibody to 1 mg of magnetic beads.

8. The application of the reagent for detecting extracellular vesicle membrane proteins according to claim 6 in the preparation of an early gastric cancer diagnostic kit, characterized in that: The incubation procedure described in step 1) is as follows: first incubate at 20-30℃ for 20-40 minutes, then transfer to 2-8℃ for 12-20 hours; The conjugation procedure described in step 2) is as follows: Dilute the biotinylate antibody with PBS, add streptavidin magnetic beads that have been washed with PBS, and incubate. After incubation, wash with PBS and then block with bovine serum albumin. After blocking, wash with PBS and then resuspend in PBS.

Images