Application of urine extracellular vesicle protein marker combination in the preparation of early diagnosis products for prostate cancer

Abstract

The present invention discloses the application of a urine extracellular vesicle protein marker combination in the preparation of an early diagnosis product for prostate cancer. The invention is characterized in that the protein marker combination includes CD177, PEPD, DNASE1, CTNNB1, SEMA3D and SCGB1A1, wherein the application of the CD177, PEPD, DNASE1 and CTNNB1 protein combination in the preparation of an early diagnosis product for prostate cancer, and the application of the CD177, PEPD, DNASE1, SEMA3D and SCGB1A1 protein combination in the preparation of an early diagnosis product for gray zone prostate cancer have the advantages of achieving early and accurate diagnosis of prostate cancer and gray zone prostate cancer, and by combining a machine learning algorithm, prostate cancer can be diagnosed quickly and accurately, thereby improving diagnostic efficiency.

Description

Technical Field

[0001] The present invention belongs to the field of medical diagnosis, and in particular relates to the application of a urine extracellular vesicle protein marker combination in the preparation of an early diagnosis product for prostate cancer. Background Art

[0002] Prostate cancer (PCa) is a malignant tumor that occurs in the prostate epithelium. It is the most common malignant tumor of the male urogenital system in China, and its incidence rate currently ranks first among malignant tumors of the male urogenital system in China. The main methods currently used to diagnose prostate cancer are: tumor markers, imaging examinations, and prostate puncture. Among them, blood prostate specific antigen (PSA) detection is currently the most important diagnostic method, but its specificity is low and it cannot achieve early diagnosis and monitor the progression of prostate cancer. In imaging examinations, whether it is transrectal ultrasound, CT examination, or magnetic resonance imaging (MRI), although it has a certain diagnostic effect on prostate cancer, its early diagnostic value is insufficient and cannot be accurately used for tumor progression staging. At present, the "gold standard" for the diagnosis of prostate cancer is prostate puncture. However, puncture biopsy is an invasive procedure with risks such as pain, bleeding, and urinary retention; and due to the multifocality and heterogeneity of prostate cancer histology, even with multi-point puncture biopsy, there is still a high false negative rate. Therefore, puncture biopsy cannot understand the full picture of prostate cancer. Late diagnosis and inaccurate risk progression grading will significantly affect the selection of the correct treatment plan for prostate cancer patients, leading to poor prognosis. In general, the existing prostate cancer screening and diagnosis methods can no longer meet the diagnosis and treatment needs in the context of precision medicine today. A new non-invasive PCa tumor marker with high sensitivity and specificity is urgently needed for early diagnosis and risk grading of PCa.

[0003] Extracellular vesicles (EVs) are an effective liquid biopsy tool with outstanding advantages: 1) They exist in almost all body fluids and have high stability due to lipid bilayer packaging, and can be stored at -80°C for months or even years; 2) They are secreted by living cells, contain biological information from parental cells, and are more tumor-specific than free DNA secreted during necrosis or apoptosis; 3) Extracellular vesicles are clearly identifiable and express specific protein markers, such as CD63, ALIX, TS101, and HSP70, which can be used as markers to effectively distinguish extracellular vesicles from other vesicle subsets, and their specific cup-shaped features make them identifiable by electron microscopy; 4) Compared with circulating tumor cells, which are difficult to collect, extracellular vesicles can be relatively easily obtained from almost any body fluid, such as the classic extraction methods of ultracentrifugation and density gradient centrifugation. In addition, urinary extracellular vesicles (uEVs) are secreted by urothelial cells and originate from the urinary system including the prostate. They have the advantage of being related to urogenital cancer and can directly reflect changes in prostate function more than blood. In summary, uEVs are a non-invasive biomarker for prostate cancer and are used for early diagnosis and monitoring of prostate cancer, which has important clinical application value. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a non-invasive, highly specific and highly sensitive urine extracellular vesicle protein marker combination for use in the preparation of an early diagnosis product for prostate cancer.

[0005] The technical solution adopted by the present invention to solve the above technical problems is:

[0006] The present invention provides an application of a urine extracellular vesicle protein marker combination in the preparation of an early diagnosis product for prostate cancer, wherein the protein marker combination comprises CD177, PEPD, DNASE1, CTNNB1, SEMA3D and SCGB1A1.

[0007] The present invention also provides an application of a urine extracellular vesicle protein marker combination in the preparation of an early diagnosis product for prostate cancer, wherein the protein marker combination includes CD177, PEPD, DNASE1 and CTNNB1.

[0008] The present invention also provides an application of a urine extracellular vesicle protein marker combination in the preparation of an early diagnosis product for gray zone prostate cancer, wherein the protein marker combination comprises CD177, PEPD, DNASE1, SEMA3D and SCGB1A1.

[0009] Compared with the prior art, the advantages of the present invention are: the present invention discloses the application of a combination of extracellular vesicle protein markers in urine in the preparation of an early diagnosis product for prostate cancer, and realizes an early and accurate diagnosis of prostate cancer by analyzing and screening the combination of extracellular vesicle protein markers in urine. This method not only reduces the pain and risk of patients, but also avoids unnecessary biopsies and reduces the possibility of misdiagnosis and missed diagnosis. In addition, the combination of CD177, PEPD, DNASE1, SEMA3D and SCGB1A1 proteins screened and obtained by the present invention is suitable for patients in the "gray zone" with a PSA (PSA index stands for prostate-specific antigen) level of 4-10 ng / mL. Traditional diagnostic methods in this area are difficult to determine the presence or severity of prostate cancer. In summary, the combination of extracellular vesicle protein markers in urine of the present invention has significant advantages in improving the accuracy, efficiency and patient experience of prostate cancer and gray zone diagnosis of prostate cancer, and shows good clinical transformation potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 To identify and analyze the extracellular vesicles extracted using the fully automated extracellular vesicle extraction system, a is a representative image of uEVs obtained by transmission electron microscopy at a scale of 1000 nm, b is a representative image of uEVs obtained by transmission electron microscopy at a scale of 400 nm, c is the particle size distribution of uEVs in prostate cancer patients and the control group measured by nanoparticle tracking analysis, d is a Western blot showing the expression levels of CD9, TSG101, CD63, Flotillin-1, ALIX and Uromodulin in purified uEVs, C is the control group; I is the low- and medium-risk group; H is the high-risk group; M is the metastasis group;

[0011] Figure 2 Volcano plot of all protein markers in urine extracellular vesicles between prostate cancer group and control group;

[0012] Figure 3 The results of the significant differential expression analysis of the six selected characteristic proteins in the training set data; C is the control group, T is the tumor group, Gray_C is the gray area control group and Gray_T is the gray area tumor group;

[0013] Figure 4 In the figure, A is the ROC curve of five prostate cancer diagnosis models constructed by combining four characteristic proteins in the training set data, and B is the ROC curve of five gray zone prostate cancer diagnosis models constructed by combining five characteristic proteins in the training set data;

[0014] Figure 5The results of the differential expression significance analysis of the six selected characteristic proteins in the independent validation set; C is the control group, T is the tumor group, Gray_C is the gray area control group and Gray_T is the gray area tumor group;

[0015] Figure 6 In the figure, A is the ROC curve of five prostate cancer diagnosis models constructed by combining four characteristic proteins in the independent validation set, and B is the ROC curve of five gray zone prostate cancer diagnosis models constructed by combining five characteristic proteins in the independent validation set. DETAILED DESCRIPTION

[0016] The present invention is further described in detail below with reference to the accompanying drawings.

[0017] Specific example 1: Screening of urine extracellular vesicle protein marker combinations.

[0018] 1. Collection and pretreatment of urine samples: All patient samples collected have signed informed consent forms by the patients themselves or their agents, and the research involving human samples has been approved by the Ethics Committee of the First Affiliated Hospital of Ningbo University (Ethics Review Number: 2021-R106). Urine samples from 168 patients with prostate cancer and the control group were collected as training sets, including the control group (Control, n=30) and the prostate cancer group (Tumor, n=138), including the gray control group (Gray Control, n=23) and the gray prostate cancer group (Gray Tumor, n=79). The risk stratification was performed according to the 2021 National Comprehensive Cancer Network (NCCN) guidelines. The selected clinical subjects will be divided into 6 groups, namely the high-risk group (including the high-risk and very high-risk recommended by the guidelines), the low-medium-risk group (including the low-risk and medium-risk recommended by the guidelines), the control group is healthy patients with prostate hyperplasia (abbreviated as BPH), and the metastasis group (patients with lymph node, bone, lung, liver or other tissue metastasis). At the same time, 2 groups of patients in the diagnostic gray zone (PSA level 4-10ng / mL) were added, namely the gray control group and the gray tumor group. The specific grouping criteria are as follows:

[0019] Control group: healthy BPH patients (40-80 years old) excluding PCa, without other tumors and of similar age;

[0020] Low-intermediate risk group: PSA ≤ 20 ng / mL, Gleason score ≤ 7, TNM stage ≤ T2c, all three are met;

[0021] High-risk group: PSA>20ng / mL, Gleason score>7, TNM stage>T2c, at least one of the following is met;

[0022] Metastatic group: patients with metastasis to lymph nodes, bones, lungs, liver, or other tissues;

[0023] Gray area control group: PSA level 4-10 ng / mL, age-matched normal population, that is, PCa patients are excluded. Gray area tumor group: PSA level 4-10 ng / mL, PCa patients.

[0024] The inclusion criteria for PCa patients were as follows: (1) All patients underwent ultrasound-guided perineal prostate puncture biopsy, were diagnosed with prostate adenocarcinoma by pathological histology, and had undergone a washout period of more than two weeks; (2) The diagnosis of PCa was confirmed with the exclusion of other malignant tumors; (3) No radiotherapy, chemotherapy, endocrine therapy, etc. were performed before inclusion; (4) No recent history of urinary tract infection or infectious disease.

[0025] Collection method: (1) All patients collected 50-100 mL of midstream urine without eating between 6 and 8 in the morning, and processed it within 2 hours after collection; (2) First, centrifuge at 4000 rpm, 20°C, for 10 min; (3) Transfer the supernatant to a clean test tube without disturbing the urine sediment layer; (4) Label the supernatant samples and store them in a -80°C refrigerator until subsequent use.

[0026] 2. Extraction and identification of extracellular vesicles in urine: Use the EXODUS® fully automatic extracellular vesicle extraction system to extract extracellular vesicles from the pretreated urine supernatant. Urine extracellular vesicles were identified by transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and protein blotting (WB). The results are as follows: Figure 1 shown.

[0027] like Figure 1 Middle a, Figure 1 (b) A representative image of urine EVs obtained by transmission electron microscopy shows urine EVs with a lipid bilayer and a cup-shaped structure (red arrow), indicating that a large number of urine EVs were obtained by EXODUS®.

[0028] like Figure 1 As shown in (c), the particle size distribution of urinary EVs in prostate cancer patients and controls was measured by nanoparticle tracking analysis, indicating that a large number of urinary EVs were obtained by EXODUS®.

[0029] like Figure 1 As shown in middle d, Western blotting was used to detect the surface marker proteins Alix, Flotilin-1, CD63, CD9, CD81, UMOD, Syntenin, TSG101 and CANX in the extracted urine extracellular vesicles. Western blotting showed the expression levels of CD9, TSG101, CD63, Flotillin-1, ALIX and Uromodulin in the purified uEVs, indicating that a large number of urine EVs were obtained by EXODUS®.

[0030] 3. Determination of protein in extracellular vesicles in urine, the steps are as follows:

[0031] Step 1. Sample pretreatment: All urine extracellular vesicle samples were transferred to a vacuum spin dryer and evaporated to dryness; after being re-dissolved with 30uL lysis buffer (6M urea + 2M thiourea), the protein was dissolved at 32°C, 45min, and 800rpm; 1.5 μL of 200mM thiol reducing agent (TCEP) solution was added to the centrifuge tube, and the tube was placed on a constant temperature oscillator at 800rpm, 45min, and 32°C for reduction reaction; then 1.5 μL of 800mM iodoacetamide (IAA) solution was added to each tube, and the tube was incubated at 800rpm, 45min, and 32°C in the dark for alkylation reaction, and then 120μL ammonium bicarbonate solution, 1μg trypsin and 0.25μg lysine endonuclease were added, and preliminary enzymatic hydrolysis was performed at 33°C, 4h, and 600rpm, and then 1μg trypsin and 0.25μg Lysine, 33°C, 14h, 400 rpm for overnight enzymatic hydrolysis; after the enzymatic hydrolysis is completed, trifluoroacetic acid is added to each well of the well plate to make the final concentration of 1%, and the digestion is terminated and then desalted. Then, the eluate is centrifuged and concentrated and evaporated using a CHRIST vacuum evaporator at 6-7mBar, and about 10uL of the sample is prepared in the loading bottle according to the required concentration.

[0032] Step 2, liquid chromatography-mass spectrometry analysis (abbreviated as LC-MS / MS):

[0033] Liquid chromatography analysis: The sample obtained in step 1 is separated by a 60-minute liquid chromatography gradient from 3% to 28% buffer B; wherein buffer A is composed of 98% by volume water and 2% acetonitrile containing 0.1% formic acid, and buffer B is composed of 98% by volume acetonitrile and 2% aqueous solution containing 0.1% formic acid;

[0034] Mass spectrometry analysis: ORBITRAP ECLIPSE mass spectrometer was used with the optional FAIMS ProTM Interface. The compensation voltage CV was between -45 and -65, switched every 1 s, and the Nanospray FlexTM (NSI) ion source was used. The ion spray voltage was set to 2.0 kV, the ion transfer tube temperature was set to 320°C, the mass spectrometry adopted data-dependent acquisition mode, the full scan range of the mass spectrometry was m / z 350-1500, the primary mass spectrometry resolution was set to 120000 (200 m / z), the AGC was 1.2e6, the C-trap maximum injection time was 50 ms, and the peptide fragmentation collision energy was set to 30%.

[0035] Mass spectrometry data analysis: DIA-NN (version 1.8) software was used to analyze the mass spectrometry files of the experimental samples. The main parameter settings are as follows: (1) Library information: Library-free; (2) Raw files: 204 raws; Homosapiens fasta files; (3) Fixed modifications: carbamidomethylation of cysteine, methionine at the N-terminus; (4) Dynamic modifications: oxidation of methionine (+15.994915); N-terminal-M excision; C carbamate methyl ester; (5) MBR mode was enabled in DIA-NN; (6) Mass accuracy was set to 0; (7) A target false discovery rate of 1% was allowed for precursors and proteins; (8) Peptide length range: 7–30; Precursor ion mass-to-charge ratio (m / z) range: 300–1500, fragment ion mass-to-charge ratio range: 200–1800. After data quality control, 5409 proteins were identified overall. After removing proteins with a missing rate higher than 70%, 4111 proteins with a missing rate lower than 70% were obtained. The mass spectrometry proteomics data have been submitted to the ProteomeXchange consortium with the dataset identifier PXD052569.

[0036] 4. Screening of differentially expressed proteins in urine extracellular vesicles, the steps are as follows:

[0037] Step 1. Statistical analysis of experimental data was performed using GraphPad Prism V 7.0 (GraphPad Software, San Diego, California). Unpaired t-test with unequal variance was used to perform statistical tests on each protein, and the P value of the t-test was corrected for FDR using the "BH" method ("Benjamin-Horgberg method"), and the P value after FDR correction was used to screen differential proteins;

[0038] Step 2: Calculate the fold change between the experimental group and the control group for each protein, and take the fold change as ㏒2;

[0039] Step 3. If the calculation results meet the following conditions: adjusted P value < 0.05 and multiple difference > 1.5, the protein is considered to have statistically significant difference between the experimental group and the control group. Use ggplot2 and ggpubr packages for statistical analysis and draw statistical graphs such as volcano graphs, fitted lines, and box plots.

[0040] Figure 2 Volcano plot of all protein markers in urine extracellular vesicles between the prostate cancer group and the control group. Figure 2 It can be seen that the expression levels of 111 proteins were upregulated, the expression levels of 63 proteins were downregulated, and the remaining proteins had some overlap and no differential expression.

[0041] 5. The machine learning method was used to screen the protein marker combination in urine extracellular vesicles. The steps are as follows:

[0042] Step 1: Screening of characteristic proteins:

[0043] (1) The expression levels of 111 up-regulated proteins and 63 down-regulated proteins in the urine extracellular vesicles of 168 patients were obtained to establish the training set data, and the training set data was divided into "internal training set" and "internal test set" in a ratio of 7:3;

[0044] (2) Next, three different feature screening algorithms were used, namely, recursive feature elimination (RFE); model-based feature selection (Model-based Feature Selection); and sequential feature selector (SFS). For each feature screening algorithm, an XGboost model was constructed, the internal training set was input into the XGboost model, the AUC value of each model classifier was output, the AUC values ​​were arranged from large to small, and the top 30 important feature proteins were selected;

[0045] (3) In order to obtain more stable results, the input training set data is randomly divided into "internal training set" and "internal test set" for a total of 5 times, and step (2) is repeated;

[0046] (4) We counted the frequency of these results and selected the top 30 characteristic proteins with the highest frequency for further screening and analysis. This method made full use of multiple screening and evaluations to ensure that the characteristic protein set we finally selected was the most representative and stable.

[0047] Step 2: Optimal characteristic protein combination for prostate cancer diagnosis model

[0048] Based on the TOP30 characteristic proteins obtained in step 1, according to the results of previous studies on these characteristic proteins, characteristic proteins that are not related to tumor occurrence are excluded, and the relevant characteristic proteins are combined in various ways. The five existing classifier models include: Logistic Regression Classifier, K Nearest Neighbors Classifier, Support Vector Machine Classifier (SVC, Support Vector Machine Classifier), Random Forest Classifier, Gradient Boosting Classifier to build a prostate cancer diagnosis model based on protein expression data, and use the internal test set to evaluate the sensitivity (Sensitivity), specificity (Specificity), accuracy (Accuracy), recall (Recall), F1 score (F1Score), positive predictive value (Positive Predictive Value, PPV), negative predictive value (Negative Predictive Value, NPV), AUC value (Area Under the Curve) of each protein combination, and screen out the four optimal characteristic protein combinations, as shown in Table 1.

[0049] Table 1 Characteristic protein combination of prostate cancer diagnostic model

[0050]

[0051] Step 3: Optimal characteristic protein combination of gray zone prostate cancer diagnosis model

[0052] The expression levels of 111 up-regulated proteins and 63 down-regulated proteins in urine extracellular vesicles of 102 patients in the gray zone control group and the gray zone prostate cancer group were obtained to establish the training set data. The gray zone prostate cancer diagnosis model based on protein expression data was constructed according to the 5-step 1 and step 2 method, and 5 optimal characteristic protein combinations were screened out. The results are shown in Table 2.

[0053] Table 2 Characteristic protein combination of gray zone prostate cancer diagnostic model

[0054]

[0055] Figure 3 is the difference expression significance analysis result of 6 selected characteristic proteins in the training set data. Figure 3It can be seen that there are significant differences in CD177, PEPD, DNASE1 and CTNNB1 proteins between prostate cancer and the control group, and there are significant differences in CD177, PEPD, DNASE1, SEMA3D and SCGB1A1 between gray-zone prostate cancer and gray-zone control group samples.

[0056] Figure 4 A in the figure is the ROC curve of the five prostate cancer diagnosis models constructed using the combination of four selected feature proteins. Figure 4 As shown in Figure A, in the internal test set, all five models achieved high area under the curve (AUC) values ​​(0.92-0.95), showing good diagnostic effects. This indicates that the screened combination of CD177, PEPD, DNASE1 and CTNNB1 can be used as characteristic proteins for the prostate cancer diagnostic model.

[0057] Figure 4 B is the ROC curve of the five gray-zone prostate cancer diagnosis models constructed using the five selected feature proteins. Figure 4 Figure B shows that in the internal test set, all five models achieved high area under the curve (AUC) values ​​(0.85-1.00), with good diagnostic effects. This indicates that the screened combination of CD177, PEPD, DNASE1, SEMA3D and SCGB1A1 can be used as characteristic proteins for the gray zone prostate cancer diagnostic model.

[0058] In summary, we collected urine from 168 patients with prostate cancer and a control group, combined proteomics with machine learning, and screened out six characteristic proteins, CD177, PEPD, DNASE1, CTNNB1, SEMA3D, and SCGB1A1. Four of them were used to construct a prostate cancer diagnostic model, and five of them were used to construct a gray zone prostate cancer diagnostic model. This shows that the model constructed by using these six proteins to identify prostate cancer, especially gray zone prostate cancer, is feasible in clinical applications.

[0059] Specific Example 2: Independent verification of the diagnostic value of the protein marker combination screened in Specific Example 1 for prostate cancer and gray zone prostate cancer.

[0060] We collected another 80 independent urine samples, including a control group (Control, n=20) and a prostate cancer group (Tumor, n=60), as an independent validation set, and used the method 2 in Specific Example 1 to extract extracellular vesicles for ELISA detection to verify

[0061] The diagnostic value of the four characteristic protein combination of CD177, CTNNB1, PEPD and DNASE1 for prostate cancer was verified by further selecting 11 cases from the above control group as the gray zone control group and 34 cases from the above prostate cancer group as the gray zone prostate cancer group to verify the diagnostic value of the five characteristic protein combination of CD177, PEPD, DNASE1, SEMA3D and SCGB1A1 for gray zone prostate cancer.

[0062] Figure 5 To independently verify the results of the differential expression significance analysis of the six selected characteristic proteins in the prostate cancer and control groups, Figure 5 It can be seen that there are significant differences in CD177, PEPD, DNASE1 and CTNNB1 proteins between prostate cancer and the control group, and there are significant differences in CD177, PEPD, DNASE1, SEMA3D and SCGB1A1 between gray-zone prostate cancer and gray-zone control group samples.

[0063] Figure 6 Figure A is the ROC curve of five prostate cancer diagnostic models constructed with four characteristic protein combinations in an independent validation set. The area under the curve (AUC) value of each model reached 0.89-0.97. The above results show that the combination of these four characteristic proteins is closely related to the incidence of prostate cancer and has good diagnostic value. The five models include: Logistic Regression, K Neighbors Classifier, Support Vector Machine Classifier (SVC), Random Forest Classifier, and Gradient Boosting Classifier.

[0064] Figure 6 Figure B shows the ROC curves of the five gray zone prostate cancer diagnostic models constructed using the five characteristic protein combinations in the independent validation set. The area under the curve (AUC) values ​​of each model reached 0.96-1.00. The above results show that the five gray zone characteristic protein combinations have high efficiency in distinguishing gray zone prostate cancer and have high diagnostic value.

[0065] The above description is not intended to limit the present invention, and the present invention is not limited to the above examples. Any changes, modifications, additions or substitutions made by a person skilled in the art within the spirit and scope of the present invention shall also fall within the protection scope of the present invention.

Claims

1. Use of a reagent for detecting a combination of extracellular vesicle protein markers in urine in the preparation of an early diagnosis product for prostate cancer, characterized in that: The protein marker combination includes CD177, PEPD, DNASE1, CTNNB1, SEMA3D and SCGB1A1.

2. Use of a reagent for detecting a combination of extracellular vesicle protein markers in urine in the preparation of an early diagnosis product for prostate cancer, characterized in that: The protein marker combination includes CD177, PEPD, DNASE1 and CTNNB1.

3. Use of a reagent for detecting a combination of extracellular vesicle protein markers in urine in the preparation of a product for early diagnosis of gray zone prostate cancer, characterized in that: The protein marker combination includes CD177, PEPD, DNASE1, SEMA3D and SCGB1A1.

Images