Molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection as well as preparation method and application of molecularly imprinted gold nanopore array substrate

Abstract

The invention discloses a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection as well as a preparation method and application of the molecularly imprinted gold nanopore array substrate, and belongs to the technical field of biological detection. The molecularly imprinted gold nanopore array substrate structurally comprises a base material and a nanogold layer from bottom to top, the nanogold layer comprises sunken gold nanopores which are distributed in an array mode, the bottoms of the gold nanopores are provided with modified gold nanoparticles, the gold nanopores comprise imprinted cavities matched with exosomes, the size of each gold nanopore is matched with the size of a single exosome, and the size of each imprinted cavity is matched with the size of the corresponding exosome. The gold nanopore channels distributed in the array form a gap between single exosomes, and the gap is greater than or equal to 1 mu m. According to the invention, a gold nanopore channel can be accurately regulated and controlled, and an exosome molecular imprinting structure is modified in the pore channel, so that capture and detection of a single exosome in body fluid are realized; a good adsorption effect on urine exosomes is achieved, and the adsorbed exosomes can be efficiently captured through the nano pore channels; and the nanopore enhancement factor is as high as 7.27 * 10 < 5 >.

Description

Technical Field

[0001] The present invention belongs to the technical field of biological detection, and particularly relates to a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection, a preparation method thereof, and an application thereof. Background Art

[0002] Exosomes are nanoscale vesicles (30 - 200 nm) secreted by cells and are widely present in various body fluids. Tumor-derived exosomes usually carry and transport molecular cargos related to tumorigenesis to neighboring or distant cells, thereby affecting many physiological and pathological processes. Therefore, the analysis of exosomes can provide comprehensive and accurate information related to tumorigenesis and can uncover potential diagnostic biomarkers. Currently, liquid biopsy methods based on exosomes usually take exosome populations as the research object and it is difficult to take into account the heterogeneity among individual exosomes. Chinese Patent CN112269023A discloses a microfluidic Raman chip and a method for detecting exosomes in blood based on the microfluidic Raman chip. In actual operation, the use and storage conditions of antibodies are relatively harsh, and the cost of antibodies is high, which causes difficulties in actual popularization and use; it can only reflect the number of exosomes and the content information of EpCAM on the exosome surface, and it is difficult to provide information on all disease-related substances inside and outside exosomes, so it has certain limitations in disease diagnosis applications; based on the analysis of exosome populations rather than single exosomes, it is impossible to further explore other potential disease biomarker information of exosomes. Since tumor information is usually only reflected in exosomes secreted by tumor cells and their microenvironment cells, this population analysis method often drowns out the information presented by tumor-derived exosomes. Chinese Patent CN113194820A discloses a method and system for providing cancer diagnosis information using liquid biopsy based on exosomes, which relies on additional exosome separation means and the operation is more complex and cumbersome; it can only collect data on the separated exosome populations and fuzzily analyze the differences between different exosome populations through statistical methods, and it is impossible to further study the information carried by a single exosome.

[0003] Single exosome detection can obtain exosome heterogeneity information. The heterogeneity among individuals can precisely reflect the differences and characteristics of their parent cells, further highlighting the significance of single exosome research. Therefore, developing a method that can efficiently isolate and in-situ analyze single exosomes, efficiently and specifically detect single exosomes, and fully explore exosome heterogeneity information is of great significance for realizing early cancer diagnosis based on exosomes.

[0004] At present, single-exosome analysis mostly relies on advanced technical means, such as optical tweezers, nano-flow cytometry, nano-infrared, etc. These methods rely on expensive instruments and have high technical thresholds. Most importantly, most single-exosome analysis methods require cumbersome and time-consuming pretreatment steps to achieve pre-separation of exosomes; these factors severely restrict their large-scale application in clinical practice. In addition, single-exosome analysis is mostly limited to the laboratory research stage, and the analysis objects are mostly exosomes secreted by in vitro cultured cell lines. There has been no report on using single-exosome analysis methods to assist in the diagnosis of clinical diseases. The reason is that there is a lack of supporting intelligent data analysis means for spectral differences at the single-exosome level, making it impossible to effectively analyze the more heterogeneous clinical sample information. Chinese Patent CN119510391A discloses a preparation method and application of a concave structure SERS substrate for prostate cancer exosome detection. Although this method prepares a gold nanoparticle concave trap array with a size matching that of exosomes, realizes the mutual isolation of exosomes, and combines machine learning algorithms to achieve high-sensitivity detection and high-accuracy typing of prostate cancer exosomes, the distance between its densely arranged concave traps is relatively close, making it difficult to achieve single-exosome analysis; this concave array is not modified with a molecularly imprinted layer. Therefore, it is impossible to directly separate exosomes from the sample solution and requires an additional exosome separation means, which is more complex to operate and difficult to achieve the integration of exosome separation and analysis.

[0005] Although exosome analysis methods based on Raman detection have been widely used in the field of disease diagnosis, immunolabeled Raman detection obtains differential exosome information by quantifying surface protein markers of exosomes to achieve disease diagnosis. This strategy relies on known protein markers and lacks the ability to discover unknown disease markers; label-free Raman detection can obtain complete spectral information of exosome components, but the complexity of Raman spectra and the heterogeneity of clinical samples pose high requirements for machine learning means; at the same time, most methods require pre-separation of exosomes by ultra-high-speed centrifugation. These methods generally require professional equipment and expensive biological reagents, and it is difficult to balance information extraction and target separation. There is still much room for improvement in the application of exosome separation and analysis. Summary of the Invention

[0006] Aiming at the problems in the prior art, the purpose of the present invention is to provide a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, its preparation method and application. As a surface-enhanced Raman spectroscopy (SERS) substrate, the molecularly imprinted inorganic material of this molecularly imprinted gold nanopore array substrate can perfectly fit the SERS substrate. Through precise regulation of the surface morphology of the enhanced substrate, efficient capture and separation of single exosomes are achieved, and at the same time, the Raman signal enhancement effect is obtained. Then, combined with machine learning algorithms to classify and process the spectral data, in-situ Raman detection of single exosomes is realized, which can further assist in the early diagnosis of cancer and improve the accuracy of diagnosis.

[0007] The present invention is realized through the following technical solutions:

[0008] The present invention provides a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection (hereinafter referred to as the molecularly imprinted gold nanopore array substrate). The structure of the molecularly imprinted gold nanopore array substrate from bottom to top is a substrate and a gold nanoparticle layer. The gold nanoparticle layer contains recessed gold nanopores distributed in an array. The bottom of the gold nanopores has modified gold nanoparticles. The gold nanopores contain imprinted cavities that match exosomes. The size of each gold nanopore matches the size of a single exosome, so that each gold nanopore can only capture a single exosome. The array-distributed gold nanopore channels form a spacing between single exosomes, and the spacing ≥ 1 μm.

[0009] Furthermore, the substrate is a glass slide, a quartz slide, a silicon wafer or a polydimethylsiloxane sheet;

[0010] The bottom of the gold nanopores has modified gold nanoparticles, which means that 3-aminophenylboronic acid (3-APBA) is modified on the surface of the gold nanoparticles at the bottom of the gold nanopores. It is used to anchor the template exosomes used to synthesize the exosome imprinted cavity at the bottom of the gold nanopores and act as an internal standard for Raman detection.

[0011] The present invention provides a preparation method for a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, including the following steps:

[0012] Step 1: Preparation of nanoparticle mask templates and gold nanoparticle seed solutions;

[0013] Step 2: Preparation of nanoparticle mask template array substrates;

[0014] Step 3: Preparation of gold nanoparticle-gold nanopore array substrates;

[0015] Step 4: Modifying the exosome imprinted layer on the gold nanopore channels.

[0016] Furthermore, in Step 1, the nanoparticle mask template is an isopropylacrylamide nanoparticle suspension or a polystyrene microsphere suspension;

[0017] Synthesize a nano-gold seed solution by the citric acid reduction method.

[0018] Furthermore, the preparation of the nanoparticle mask template array substrate in step 2 includes the following steps:

[0019] (1) Place the substrate in a piranha solution for surface treatment, then rinse it repeatedly with deionized water, and dry it with an inert gas stream.

[0020] (2) Drop the nanoparticle mask template on the dried substrate, then quickly add absolute ethanol, and after drying with an inert gas stream at room temperature, form a 2D array substrate of the nanoparticle mask template on the top of the substrate.

[0021] (3) Place the 2D array of the nanoparticle mask template in a plasma cleaner, and obtain a nanoparticle mask template array substrate with intervals after etching.

[0022] Even further, for the preparation of the nanoparticle array mask template in step 2, the time for surface treatment of the substrate in step (1) is 4 h - 12 h;

[0023] The added amount of absolute ethanol in step (2) is 1 to 10 times the volume of the nanoparticles;

[0024] The etching time in step (3) is 90 s - 150 s.

[0025] Furthermore, the preparation of the nano-gold - gold nanopore array substrate in step 3 includes the following steps:

[0026] (1) Place the nanoparticle mask template array substrate and 3-aminopropyltriethoxysilane together in a dryer, evacuate and heat to perform amino modification on the exposed part of the substrate not covered by the nanoparticle mask template.

[0027] (2) Add the nano-gold seed solution on the top of the amino-functionalized nanoparticle mask template array substrate for incubation. The electrostatic adsorption of nano-gold occurs on the amino-modified exposed substrate part to obtain a nanoparticle - nano-gold hybrid array substrate, and then rinse it repeatedly with deionized water.

[0028] (3) Continue to perform ultrasonic treatment in the template cleaning solution to remove the nanoparticle mask template and obtain a nano-gold array substrate.

[0029] (4) Place the nano-gold array substrate in a gold growth solution containing hydroxylamine hydrochloride and chloroauric acid trihydrate for oscillating incubation to in-situ grow a nano-gold layer and obtain a gold nanopore array substrate, rinse it with deionized water, and dry it in an inert gas stream.

[0030] (5) Incubate the gold nanopore array substrate in a gold nanoparticle seed solution. Gold nanoparticles are deposited by electrostatic interaction in the substrate area at the bottom of the gold nanopores, and then it is incubated in a mixed solution of 3-aminophenylboronic acid and a coupling agent, washed repeatedly with a washing buffer, and then incubated in a blocking buffer to obtain a gold nanoparticle-gold nanopore array substrate; The surface of the gold nanoparticles at the bottom of the gold nanopores is modified with 3-aminophenylboronic acid by the coupling agent method. Since the gold layer constituting the outer wall of the gold nanopores is grown by the chemical reduction method of hydroxylamine hydrochloride, only the gold nanoparticles at the bottom of the gold nanopores have carboxyl groups that can react with 3-aminophenylboronic acid; Incubation in the blocking buffer can block non-specific binding sites.

[0031] Furthermore, in the preparation of the gold nanoparticle-gold nanopore array substrate in step 3, the heating temperature in step (1) is 60°C - 80°C, and the heating time is ≥2 h;

[0032] The incubation time in step (2) is ≥4 h;

[0033] The template cleaning solution in step (3) is an aqueous methanol solution or tetrahydrofuran. The volume ratio of methanol to water in the aqueous methanol solution is (1 - 20):1, and the ultrasonic treatment time is 2 min - 5 min;

[0034] The concentration of hydroxylamine hydrochloride in step (4) is 0.4 mM, the mass ratio of chloroauric acid to water in chloroauric acid trihydrate is 0.5% - 1%, and the incubation time is 2 h;

[0035] The concentration of 3-aminophenylboronic acid in step (5) is 3 mg / mL - 6 mg / mL;

[0036] The coupling agent is 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysuccinimide. The concentrations of both 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysuccinimide are 20 mg / mL - 50 mg / mL, and the incubation time in the mixed solution of the coupling agent and 3-aminophenylboronic acid is 60 min;

[0037] The washing buffer is PBS buffer, with a concentration of 5 mM - 10 mM and a pH value of 7.4;

[0038] The blocking buffer is a 5 mM - 10 mM PBS solution and contains 1 wt.% bovine serum albumin, with a pH value of 7.4 and a temperature of 37°C. The incubation time in the blocking buffer is 20 min - 60 min.

[0039] Further, step 4 for modifying the exosome imprinted layer on the gold nanopores includes the following steps:

[0040] (1) Add the exosome template to the top of the gold nanoparticle-gold nanopore array substrate for incubation to anchor the exosome template at the bottom of the pores;

[0041] (2) Immerse the substrate anchored with the exosome template into the tris(hydroxymethyl)aminomethane-hydrochloric acid buffer containing dopamine hydrochloride, and carry out a polymerization reaction under stirring to form a gold nanoparticle-gold nanopore array substrate with a polydopamine exosome imprinted layer;

[0042] (3) Immerse the gold nanoparticle-gold nanopore array substrate with exosome imprinting into PBS buffer for ultrasonic treatment to remove the polydopamine nanoparticles adsorbed on the surface, then soak it in a mixed solution of acetic acid and acetonitrile to remove the exosome template, and obtain an exosome imprinted gold nanopore array substrate. Then incubate it in a blocking buffer to obtain a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection, and store it in PBS buffer at 4 °C for later use.

[0043] Furthermore, in step 4, the gold nanopores are modified with an exosome imprinted layer, and the incubation time in step (1) is 4 h - 12 h;

[0044] In step (2), the concentration of dopamine hydrochloride is 2 mg / mL - 8 mg / mL, the concentration of tris(hydroxymethyl)aminomethane-hydrochloric acid buffer is 5 mM - 10 mM, and the pH value is 8.5; the polymerization reaction time is 2 h - 12 h;

[0045] In step (3), the concentration of PBS buffer is 5 mM - 10 mM, and the ultrasonic treatment time is 10 s - 120 s;

[0046] The volume ratio of acetic acid to acetonitrile is 1:(9 - 49), and the soaking time is ≥10 min;

[0047] The blocking buffer is a 5 mM - 10 mM PBS solution, and 1 wt.% bovine serum albumin is added, the pH value is 7.4, the temperature is 37 °C, and the incubation time in the blocking buffer is 30 min - 60 min.

[0048] The present invention provides an application of a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection in a digital diagnosis system for early diagnosis of bladder cancer. The system includes a standard data acquisition module, a data to be detected acquisition module, a data screening module, a calculation module, and a diagnosis module;

[0049] The data acquisition module uses the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection to capture standard exosomes, and then uses a Raman spectrometer to perform surface enhanced Raman spectroscopy detection to collect standard exosome spectral data, wherein the standard exosomes are bladder cancer cell-derived exosomes and healthy-derived exosomes obtained by ultra-high-speed centrifugation, and the bladder cancer cell-derived exosomes include exosomes from three common bladder cancer cell lines, T24, BIU-87 and 5637 cells, and each sample is subjected to at least 200 surface enhanced Raman spectroscopy detections, which are defined as standard cancer cell exosome spectra; the healthy-derived exosomes are obtained from the urine of healthy volunteers, and each sample is subjected to at least 200 surface enhanced Raman spectroscopy detections, which are defined as standard healthy exosome spectra; at the same time, the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection is subjected to surface enhanced Raman spectroscopy detection, and spectra of more than 1,000 blank nanopores are collected, which are defined as standard blank spectra;

[0050] The test data acquisition module is to adjust the pH value of the collected urine sample to 8.5, use a sterile syringe to draw 2 mL of urine sample and inject it into a 0.22 μm filter membrane to filter out large impurities, immerse the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection in the filtrate, incubate at 37° C. for 1 hour with shaking, capture urine exosomes through the exosome imprinting structure on the surface, and then use a Raman spectrometer to perform surface enhanced Raman spectroscopy detection to collect exosome spectral data of the sample to be tested, and perform surface enhanced Raman spectroscopy detection for each sample at least 200 times;

[0051] The data screening module performs feature pre-screening on invalid data in the standard spectral data of exosomes from bladder cancer cells, standard spectral data of exosomes from healthy sources, and spectral data of exosomes from samples to be tested, and uses 998 cm -1 The average Raman intensity I 998cm -1 As a benchmark to identify invalid spectral data, I 998cm -1 <0.9AverageI 998cm -1 The data was determined to be invalid data, and the invalid data was deleted through the automatic screening program written in Python;

[0052] The calculation module marks the standard spectrum data of exosomes derived from bladder cancer cells and the standard spectrum data of exosomes derived from healthy people after being processed by the data screening module, and marks the standard blank spectrum obtained by the data acquisition module as the standard blank;

[0053] The standard spectral data of exosomes derived from bladder cancer cells were clustered using K-Means and divided into two clusters. Then, the Pearson correlation coefficient P between the spectra and the standard blank spectrum was compared. The spectra with significantly lower P values were determined as the spectra of exosomes from bladder cancer cells and labeled as "standard taEV".

[0054] The standard spectral data of exosomes from healthy sources were clustered and divided into two clusters. Then, the correlation coefficient P between the spectra and the standard blank spectrum was compared. The spectra with significantly lower P values were determined as the spectra of healthy exosomes and labeled as "standard nEV".

[0055] The exosome spectra of bladder cancer patient subjects were clustered and divided into three clusters. The P values between the spectra and the standard blank, standard taEV, and standard nEV were compared separately and divided into three categories, namely blank, nEV, and taEV spectra.

[0056] After labeling the spectra as blank, nEV, and taEV spectra, a convolutional neural network (CNN) model was further constructed for exosome spectral classification. The CNN architecture was determined using a grid search algorithm, and class weights were used during model training.

[0057] The diagnostic module used the CNN spectral classification model to obtain the exosome digital count matrices of healthy volunteers and BCa patients belonging to the training samples. From this, the counts of the blank spectrum (ε1), nEV spectrum (ε2), and taEV (ε3) were derived. The total exosome count was calculated as ξ1 = ε2 + ε3, and the proportion of taEV in the total exosomes was ξ2 = ε3 / ξ1. All thresholds were statistically determined based on the training data.

[0058] First, the threshold δ1 was set to the third quartile ξ1 of the healthy samples. If ξ1 < δ1, the individual was classified as healthy; otherwise, the person moved on to the next discrimination. Then, the threshold δ2 was set based on the lower limit after removing outliers from ε3 of the early patients. If ε3 < δ2, the individual was considered non-cancerous; otherwise, the individual was diagnosed as a bladder cancer patient and moved on to the next discrimination. Finally, the threshold δ3 was set to the third quartile ξ2 of the samples of patients with early-stage bladder cancer. If ξ2 < δ3, the individual was regarded as a patient with early-stage bladder cancer; otherwise, the individual was diagnosed as a patient with advanced bladder cancer.

[0059] Compared with the existing methods, the present invention has the following beneficial effects:

[0060] 1. Based on the gold nanopore array, the present invention can precisely regulate the gold nanopores and modify the exosome molecular imprint structure in the pores, realizing the capture and detection of single exosomes in body fluids.

[0061] 2. The molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection in the present invention has a good adsorption effect on urinary exosomes, and the nanopores can efficiently capture the adsorbed exosomes;

[0062] 3. The nanopore enhancement factor of the present invention is as high as 7.27×10 5 ;

[0063] 4. The machine learning model trained in the present invention can well distinguish the blank spectra, spectra of exosomes from healthy sources, and spectra of cancer-related exosomes, with accuracies of 97.85%, 97.85%, and 99.38% respectively;

[0064] 5. The present invention conducts statistics based on the classification results of different samples and designs a digital diagnosis system for early bladder cancer diagnosis based on the number of spectra of different categories. This system can effectively distinguish healthy samples, early-stage bladder cancer patients, and middle- and late-stage bladder cancer patients, with an overall accuracy of over 90%. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 is a schematic diagram of the preparation process of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection in Example 1 of the present invention, where EI-AuNH corresponds to the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection;

[0066] Figure 2 is a scanning electron microscope image (SEM) corresponding to part of the preparation stage in Example 1 of the present invention, where: (a) is the isopropylacrylamide nanoparticle mask template array substrate, (b) AuNH is the gold nanopore array substrate, (c) AuNP-AuNH is the gold nanoparticle-gold nanopore array substrate, and (d) EI-AuNH is the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection;

[0067] Figure 3 is a detection graph of the Raman enhancement effect of the substrate under different modification conditions during the preparation of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection in Example 1 of the present invention;

[0068] Figure 4 is a scanning electron microscope image of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection in Example 1 of the present invention for capturing a single exosome;

[0069] Figure 5 is a flowchart of the bladder cancer diagnosis system in Application Example 1 of the present invention;

[0070] Figure 6This is an example of a digital counting matrix for three typical samples in Application Example 1 of the present invention, where: HV corresponds to healthy volunteers, ES corresponds to patients with early-stage bladder cancer, and TS corresponds to patients with mid- to late-stage bladder cancer.

[0071] Figure 7 This is the confusion matrix of the diagnostic module in Application Example 1 of the present invention for outputting the diagnostic result after analyzing the spectral data of the test sample. Detailed implementation manners

[0072] The present invention will be described in detail below in conjunction with the embodiments.

[0073] In each embodiment of the present invention, a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection can be prepared. The substrate structure from bottom to top is a substrate and a gold nanoparticle layer. The gold nanoparticle layer contains recessed gold nanopores distributed in an array. The bottom of the gold nanopores has modified gold nanoparticles. The surface of the gold nanoparticles is modified with 3-aminophenylboronic acid. The gold nanopores contain imprinted cavities for modified exosomes. The size of each gold nanopore is equivalent to that of a single exosome. Each gold nanopore channel captures a single exosome. The array-distributed gold nanopore channels form a spacing between single exosomes, and the spacing is ≥1 μm; the substrate is a glass slide, a quartz slide, a silicon wafer, or a polydimethylsiloxane sheet; from the perspective of cost savings, the substrate in each embodiment is a coverslip.

[0074] Example 1

[0075] A preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, and the preparation process is as Figure 1 shown, including the following steps:

[0076] Step 1, preparation of isopropylacrylamide nanoparticle suspension and gold nanoparticle seed solution:

[0077] Preparation of isopropylacrylamide nanoparticle suspension: Add 24 mL of deionized water to a round-bottom flask. While stirring, add 0.54 g of isopropylacrylamide (NIPAM) and 0.02 g of N,N'-methylenebisacrylamide (BIS) thereto. Under an argon atmosphere, heat the mixture to 70 °C, inject 1 mL of potassium persulfate (KPS, 8 mg / mL) to initiate the free radical polymerization reaction, react at 70 °C for 4 h, and after centrifuging at 7000 rpm for 30 min, purify the obtained nanoparticle suspension by repeating washing with deionized water 5 times;

[0078] Preparation of gold nanoparticle seed solution: Add 24 mL of 0.0125% (w / w) chloroauric acid trihydrate (HAuCl4·3H2O) into a round-bottom flask and heat to boiling. Add 1 mL of sodium citrate solution (1%, w / w) into the mixture, stir and heat to boiling for 15 min, then cool the solution to room temperature and store it at 4 °C for later use.

[0079] Step 2: Preparation of the isopropylacrylamide nanoparticle mask template array substrate, including the following steps:

[0080] (1) Place the coverslip in the piranha solution H2SO4 / H2O2 (7:3, v / v) for 12 h, then rinse repeatedly with deionized water and dry in a nitrogen stream;

[0081] (2) Drop 10 μL of the isopropylacrylamide nanoparticle suspension on the dried coverslip, then quickly add 50 μL of absolute ethanol. After drying with nitrogen at room temperature, a quasi-hexagonal 2D array substrate of isopropylacrylamide nanoparticles is formed on the top of the coverslip;

[0082] (3) Place the quasi-hexagonal 2D array substrate of isopropylacrylamide nanoparticles in a plasma cleaner. After etching for 90 s, the exposed glass part between the isopropylacrylamide nanoparticles is enlarged, as shown in (a), to obtain the isopropylacrylamide nanoparticle mask template array substrate. Figure 2 (a)

[0083] Step 3: Preparation of the gold nanoparticle-gold nanopore array substrate, including the following steps:

[0084] (1) Place the isopropylacrylamide nanoparticle mask template substrate and 3-aminopropyltriethoxysilane together in a desiccator, evacuate to 0.3 mbar, and heat at 80 °C for 2 h to perform amino modification on the exposed glass part;

[0085] (2) Add 100 μL of gold nanoparticle seeds on the top of the amino-functionalized isopropylacrylamide nanoparticle mask template and incubate for 4 h. The amino-modified exposed substrate part electrostatically adsorbs the gold nanoparticles to obtain the isopropylacrylamide nanoparticle-gold nanoparticle hybrid array substrate, and then rinse repeatedly with deionized water;

[0086] (3) Continue to ultrasonically treat in the H2O:CH3OH (1:20, v / v) solution for 5 min to remove the isopropylacrylamide nanoparticles on the substrate and obtain the gold nanoparticle array substrate;

[0087] (4) Under the agitation of the eddy current oscillator, the nano-gold array substrate was incubated in a gold growth solution containing 0.4 mM hydroxylamine hydrochloride and 0.5% (w / v) chloroauric acid trihydrate for 2 h to in-situ grow a nano-gold layer, so as to in-situ grow nano-gold pores in the area covered with nano-gold seeds, as Figure 2 (b) shown, to obtain a gold nano-pore array substrate (AuNH array), which was rinsed with deionized water and dried in a nitrogen stream;

[0088] (5) The gold nano-pore array substrate was incubated in the nano-gold seed solution for 4 h, and gold nanoparticles were deposited on the substrate area at the bottom of the gold nano-pores by electrostatic interaction. Then it was placed in a mixed solution of 1 mL 3-aminophenylboronic acid (6 mg / mL), 1 mL 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (50 mg / mL), and 1 mL N-hydroxysuccinimide (50 mg / mL) and incubated for 60 min, and then repeatedly washed with PBS buffer (5 mM, pH = 7.4). Then it was placed in 1 mL blocking buffer (10 mM PBS, 1% BSA, pH = 7.4, 37 °C) and incubated for 30 min to block non-specific binding sites, as Figure 2 (c) shown, to obtain a nano-gold - gold nano-pore array substrate (AuNP-AuNH array);

[0089] Step 4: Modifying the exosome imprinting layer on the gold nano-pores, including the following steps:

[0090] (1) At 4 °C, 10 μL of exosome template was dropped onto the top of the nano-gold - gold nano-pore array substrate and incubated for 4 h to anchor the exosome template;

[0091] (2) The nano-gold - gold nano-pore array substrate with the exosome template was immersed in 10 mL of tris(hydroxymethyl)aminomethane - hydrochloric acid buffer (10 mM, pH = 8.5) containing 20 mg of dopamine hydrochloride, and a polymerization reaction was carried out under oscillating agitation for 12 h to form a nano-gold - gold nano-pore array substrate with a polydopamine exosome imprinting layer;

[0092] (3) The nano-gold - gold nano-pore array substrate with the exosome imprinted polydopamine layer was immersed in PBS buffer (5 mM, pH = 7.4) and ultrasonically treated for 10 s to remove the polydopamine nanoparticles adsorbed on the surface. Then it was treated with a CH3COOH:C2H3N (1:49, v / v) mixture for 30 min to remove the exosome template, to obtain an exosome imprinted gold nano-pore array substrate, and then blocked with blocking buffer (10 mM PBS, 1 wt.% BSA, pH = 7.4) at 37 °C for 60 min, as Figure 2As shown in (d), a molecularly imprinted gold nanopore array substrate (EI-AuNH array) for single exosome capture-in situ Raman detection was obtained and stored at 4 °C in PBS buffer for later use.

[0093] Performance test:

[0094] (1) Detection of the Raman enhancement effect of the substrate:

[0095] During the preparation of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, the Raman enhancement effects of the substrates under four different modification conditions of blank substrates (cover glasses), gold nanoparticle arrays, gold nanopore arrays, and gold nanoparticle-gold nanopore arrays were detected. As Figure 3 shown, the gold nanoparticle-gold nanopore array substrate exhibited the highest signal intensity, and its enhancement factor was as high as 7.27×10 5 .

[0096] (2) Capture of single exosomes by the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection:

[0097] On the surface of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection prepared in Example 1, an exosome solution with a concentration of 10 9 per mL was added dropwise, incubated at 20 °C for 4 h, and the surface of the substrate was characterized using a scanning electron microscope. As Figure 4 shown by the results, single exosomes can be successfully captured.

[0098] Example 2

[0099] A preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, comprising the following steps:

[0100] Step 1, preparation of isopropylacrylamide nanoparticle suspension and gold nanoparticle seed solution:

[0101] Preparation of isopropylacrylamide nanoparticle suspension: Add 24 mL of deionized water to a round-bottom flask, while stirring, add 0.54 g of isopropylacrylamide (NIPAM) and 0.02 g of N,N'-methylenebisacrylamide (BIS) thereto, heat the mixture to 70 °C under an argon atmosphere, inject 1 mL of potassium persulfate (KPS, 8 mg / mL) to initiate the free radical polymerization reaction, react at 70 °C for 4 h, after centrifuging at 7000 rpm for 30 min, the obtained nanoparticle suspension was purified by repeating washing with deionized water 5 times;

[0102] Preparation of gold nanoparticle seed solution: Add 24 mL of 0.0125% (w / w) chloroauric acid trihydrate (HAuCl4·3H2O) into a round-bottom flask and heat it to boiling. Then add 1 mL of sodium citrate solution (1%, w / w) into the mixture, stir and heat it to boiling for 15 min. After that, cool the solution to room temperature and store it at 4 °C for later use.

[0103] Step 2: Preparation of an isopropylacrylamide nanoparticle mask template array substrate, including the following steps:

[0104] (1) Immerse a cover glass in piranha solution H2SO4 / H2O2 (7:3, v / v) for 12 h, then rinse it repeatedly with deionized water and dry it in a nitrogen stream.

[0105] (2) Drop 10 μL of isopropylacrylamide nanoparticle suspension on the dried cover glass, then quickly add 100 μL of absolute ethanol. After drying with nitrogen at room temperature, a quasi-hexagonal 2D array substrate of isopropylacrylamide nanoparticles is formed on the top of the cover glass.

[0106] (3) Place the quasi-hexagonal 2D array substrate of isopropylacrylamide nanoparticles in a plasma cleaner. After etching for 120 s, the exposed glass part between isopropylacrylamide nanoparticles is enlarged to obtain an isopropylacrylamide nanoparticle mask template array substrate.

[0107] Step 3: Preparation of a gold nanoparticle - gold nanopore array substrate, including the following steps:

[0108] (1) Place the isopropylacrylamide nanoparticle mask template substrate and 3-aminopropyltriethoxysilane together in a desiccator, evacuate to 0.3 mbar, and heat at 70 °C for 3 h to perform amino modification on the exposed glass part.

[0109] (2) Add 100 μL of gold nanoparticle seeds on the top of the amino-functionalized isopropylacrylamide nanoparticle mask template and incubate for 8 h. The amino-modified exposed substrate part electrostatically adsorbs gold nanoparticles to obtain an isopropylacrylamide nanoparticle - gold nanoparticle hybrid array substrate, and then rinse it repeatedly with deionized water.

[0110] (3) Continue to ultrasonically treat in an H2O:CH3OH (1:20, v / v) solution for 4 min to remove the isopropylacrylamide nanoparticles on the substrate and obtain a gold nanoparticle array substrate.

[0111] (4) Under the agitation of the eddy current oscillator, the nano-gold array substrate was incubated in a gold growth solution containing 0.4 mM hydroxylamine hydrochloride and 0.5% (w / v) chloroauric acid trihydrate for 2 h to in-situ grow a nano-gold layer, so as to in-situ grow nano-gold pore channels in the area covered with nano-gold seeds, obtaining a gold nano-pore array substrate (AuNH array), rinsed with deionized water, and dried in a nitrogen gas stream;

[0112] (5) The gold nano-pore array substrate was incubated in the nano-gold seed solution for 4 h, and gold nanoparticles were deposited by electrostatic interaction in the substrate area at the bottom of the gold nano-pores, and then placed in a solution of 1 mL 3-aminophenylboronic acid (6 mg / mL), 1 mL 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (40 mg / mL) and 1 mL N-hydroxysuccinimide (40 mg / mL) and incubated for 60 min, then repeatedly washed with PBS buffer (5 mM, pH = 7.4), and then placed in 1 mL of blocking buffer (10 mM PBS, 1% BSA, pH = 7.4, 37 °C) and incubated for 30 min to block non-specific binding sites, obtaining a nano-gold - gold nano-pore array substrate (AuNP - AuNH array);

[0113] Step 4: Modifying the exosome imprinting layer on the gold nano-pore channels, including the following steps:

[0114] (1) At 4 °C, 10 μL of exosome template was dropped onto the top of the nano-gold - gold nano-pore array substrate and incubated for 4 h to anchor the exosome template;

[0115] (2) The nano-gold - gold nano-pore array substrate with the exosome template was immersed in 10 mL of tris(hydroxymethyl)aminomethane - hydrochloric acid buffer (10 mM, pH = 8.5) containing 60 mg of dopamine hydrochloride, and polymerized under oscillating agitation for 6 h to form a nano-gold - gold nano-pore array substrate with a polydopamine exosome imprinting layer;

[0116] (3) The nano-gold - gold nano-pore array substrate with the exosome imprinted polydopamine layer was immersed in PBS buffer (5 mM, pH = 7.4) and sonicated for 10 s to remove the polydopamine nanoparticles adsorbed on the surface, and then treated with a CH3COOH:C2H3N (1:49, v / v) mixture for 30 min to remove the exosome template, obtaining an exosome imprinted gold nano-pore array substrate, and then blocked with blocking buffer (10 mM PBS, 1 wt.% BSA, pH = 7.4) at 37 °C for 60 min to obtain a molecularly imprinted gold nano-pore array substrate (EI - AuNH array) for single exosome capture - in-situ Raman detection, and stored at 4 °C in PBS buffer for later use.

[0117] Example 3

[0118] A preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection, where the nanoparticle mask template uses a commercially available polystyrene microsphere suspension, including the following steps:

[0119] Step 1, Preparation of gold nanoparticle seed solution:

[0120] Preparation of gold nanoparticle seed solution: Add 24 mL of 0.0125% (w / w) chloroauric acid trihydrate (HAuCl4·3H2O) to a round-bottom flask and heat to boiling. Add 1 mL of sodium citrate solution (1%, w / w) to the mixture, stir and heat to boiling for 15 min, then cool the solution to room temperature and store at 4 °C for later use.

[0121] Step 2, Preparation of a polystyrene microsphere nanoparticle mask template array substrate, including the following steps:

[0122] (1) Place a coverslip in piranha solution H2SO4 / H2O2 (7:3, v / v) for 12 h, then rinse repeatedly with deionized water and dry in a nitrogen stream;

[0123] (2) Drop 20 μL of polystyrene microsphere suspension on the dried coverslip, then quickly add 20 μL of absolute ethanol, dry at room temperature with nitrogen, and form a 2D array substrate with close-packed polystyrene microspheres on the top of the coverslip;

[0124] (3) Place the 2D array substrate with close-packed polystyrene microspheres in a plasma cleaner, etch for 150 s, and expand the exposed glass part between the polystyrene microspheres to obtain a polystyrene microsphere nanoparticle mask template array substrate.

[0125] Step 3, Preparation of a gold nanoparticle-gold nanopore array substrate, including the following steps:

[0126] (1) Place the polystyrene microsphere nanoparticle mask template substrate and 3-aminopropyltriethoxysilane in a desiccator, evacuate to 0.3 mbar, and heat at 60 °C for 4 h to perform amino modification on the exposed glass part;

[0127] (2) Add 100 μL of gold nanoparticle seeds on the top of the amino-functionalized polystyrene microsphere nanoparticle mask template and incubate for 6 h. The electrostatic adsorption of gold nanoparticles occurs on the exposed substrate part modified with amino groups to obtain a polystyrene microsphere-gold nanoparticle hybrid array substrate, and then rinse repeatedly with deionized water;

[0128] (3) Continue to ultrasonically treat in tetrahydrofuran for 2 min to remove the polystyrene microsphere nanoparticles on the substrate and obtain a gold nanoparticle array substrate;

[0129] (4) Under the stirring of the eddy current oscillator, the nano-gold array substrate was incubated in a gold growth solution containing 0.4 mM hydroxylamine hydrochloride and 0.5% (w / v) chloroauric acid trihydrate for 2 h to in-situ grow a nano-gold layer, so as to in-situ grow nano-gold pores in the area covered with nano-gold seeds, obtaining a gold nano-pore array substrate (AuNH array), rinsed with deionized water, and dried in a nitrogen stream;

[0130] (5) The gold nano-pore array substrate was incubated in the nano-gold seed solution for 4 h, and gold nanoparticles were deposited by electrostatic interaction in the substrate area at the bottom of the gold nano-pores, and then placed in a solution of 1 mL 3-aminophenylboronic acid (6 mg / mL), 1 mL 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (20 mg / mL) and 1 mL N-hydroxysuccinimide (20 mg / mL) and incubated for 60 min, and then repeatedly washed with PBS buffer (5 mM, pH = 7.4), and then placed in 1 mL blocking buffer (10 mM PBS, 1% BSA, pH = 7.4, 37 °C) and incubated for 30 min to block non-specific binding sites, obtaining a nano-gold - gold nano-pore array substrate (AuNP-AuNH array);

[0131] Step 4: Modifying the exosome imprinting layer on the gold nano-pores, including the following steps:

[0132] (1) At 4 °C, 10 μL of exosome template was added dropwise to the top of the nano-gold - gold nano-pore array substrate and incubated for 4 h to anchor the exosome template;

[0133] (2) The nano-gold - gold nano-pore array substrate with the exosome template was immersed in 10 mL of tris(hydroxymethyl)aminomethane-hydrochloric acid buffer (10 mM, pH = 8.5) containing 80 mg of dopamine hydrochloride, and a polymerization reaction was carried out under oscillating stirring for 2 h to form a nano-gold - gold nano-pore array substrate with a polydopamine exosome imprinting layer;

[0134] (3) The nano-gold - gold nano-pore array substrate with the exosome imprinted polydopamine layer was immersed in PBS buffer (5 mM, pH = 7.4) and ultrasonically treated for 10 s to remove the polydopamine nanoparticles adsorbed on the surface, and then treated with a CH3COOH:C2H3N (1:49, v / v) mixture for 30 min to remove the exosome template, obtaining an exosome imprinted gold nano-pore array substrate, and then blocked with blocking buffer (10 mM PBS, 1 wt.% BSA, pH = 7.4) at 37 °C for 60 min, obtaining a molecularly imprinted gold nano-pore array substrate (EI-AuNH array) for single exosome capture-in-situ Raman detection, and stored at 4 °C in PBS buffer for later use.

[0135] Application Example 1

[0136] The molecularly imprinted gold nanopore array substrate prepared in Example 1 for single exosome capture-in situ Raman detection is used in an early diagnosis system for bladder cancer. The diagnosis system includes a standard data acquisition module, a data to be tested acquisition module, a data screening module, a calculation module, and a diagnosis module.

[0137] (1) Standard data acquisition module:

[0138] Collection of spectral data of exosomes derived from standard bladder cancer cells: Exosomes from three common bladder cancer cell lines (T24, BIU-87, and 5637 cells) were collected, a total of 6 batches of culture media (2 for each cell line). Exosomes derived from bladder cancer cells obtained by ultra-high-speed centrifugation were dropped onto the surface of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, and surface-enhanced Raman spectroscopy (SERS) detection was performed using a Raman spectrometer. The objective magnification of the Raman spectrometer was 100×, the laser wavelength was 785 nm, the power was 5 W, and the integration time was 10 s; each sample was subjected to no less than 200 SERS acquisitions, and a total of 1623 sets of spectra were collected as the standard spectra of exosomes derived from cancer cells.

[0139] Collection of spectral data of exosomes from standard healthy sources: Exosomes from 21 healthy volunteers were collected. Exosomes from healthy sources obtained by ultra-high-speed centrifugation were dropped onto the surface of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, and SERS detection was performed using a Raman spectrometer. The objective magnification of the Raman spectrometer was 100×, the laser wavelength was 785 nm, the power was 5 W, and the integration time was 10 s; each sample was subjected to no less than 200 SERS acquisitions, and a total of 4321 sets of spectra were collected as the standard spectra of exosomes from healthy sources.

[0140] Collection of spectral data of the array substrate: The molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection was subjected to SERS detection using a Raman spectrometer. The objective magnification of the Raman spectrometer was 100×, the laser wavelength was 785 nm, the power was 5 W, and the integration time was 10 s, and the standard blank spectra of 1100 blank nanopores were collected.

[0141] (2) Data to be tested acquisition module:

[0142] Capture and data collection of urinary exosomes: Adjust the pH value of the collected urine sample to 8.5. Use a sterile syringe to aspirate 2 mL of the urine sample and filter out larger impurities through a 0.22 μm filter membrane. Immerse the material in the filtrate and incubate it with shaking at 37 °C for 1 h. Capture urinary exosomes through the exosome imprinting structure on the surface of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection. Use a Raman spectrometer to perform SERS detection on the captured exosomes. The objective magnification of the Raman spectrometer is 100×, the laser wavelength is 785 nm, the power is 5 W, and the integration time is 10 s; at least 200 SERS acquisitions are performed for each sample.

[0143] (3) Data screening module:

[0144] Spectral data analysis by machine learning: Use the average Raman intensity (I998cm -1 ) at 998 cm -1 of the collected Raman spectra as a benchmark to identify invalid spectral data. Data with I998cm -1 < 0.9AverageI998cm -1 is determined as invalid data and eliminated through an automatic screening program written in Python. Normalize all valid data using the Raman signal intensity at 998 cm -1 as an internal standard, and then perform feature pre-screening by eliminating irrelevant data from the substrate or internal standard (400 cm -1 - 427 cm -1 , 503 cm -1 - 553 cm -1 and 966 cm -1 - 1009 cm -1 ).

[0145] (4) Calculation module:

[0146] Exosome spectral classification: Use K-Means to divide the spectra collected from bladder cancer cell exosomes into two clusters, and then compare the Pearson correlation coefficient (P) between the exosome spectra of these samples and the blank spectra. Those spectra with significantly lower P values are determined as bladder cancer cell exosome spectra and labeled as "standard taEV";

[0147] Cluster the spectra of 21 healthy volunteers and divide them into two clusters, and further compare the correlation coefficient (P) between the exosome spectra of these healthy volunteers and the blank spectra. Those spectra with significantly lower P values are determined as exosome spectra from healthy sources, thus determining "standard nEV";

[0148] By clustering the spectra of 55 bladder cancer patients into three clusters and comparing the P-values between the exosome spectra of these bladder cancer patients and the standard spectra (standard blank, standard taEV, standard nEV), the exosome spectra of all 55 bladder cancer patients can be divided into three categories, namely blank, nEV, and taEV spectra;

[0149] After labeling the spectra into three categories, a convolutional neural network (CNN) model was further constructed for exosome spectrum classification, and the CNN architecture was determined using a grid search algorithm. The first two convolutional layers used filters of different sizes: 32 filters with a kernel size of (3,5) and 64 filters with a kernel size of (2,2). After passing through the max-pooling layer, we connected to the third convolutional layer, which used 128 filters with a kernel size of (2,2); all convolutional operations used a stride of (1,1). Then the resulting features were flattened and fed into a fully connected layer containing 100 neurons; to avoid overfitting, a dropout layer with a dropout rate p = 0.5 was added; finally, a fully connected layer with 3 neurons was used to generate the final scores for each class label; then these scores were passed through the SoftMax activation function to calculate the final class probabilities. The complete model included 1,654,867 trainable parameters, and the classification results of each group of spectra were presented in the form of a 10×17 array using a digital counting method.

[0150] To address the class imbalance problem in the training data, we used class weights during model training, which improved the classification accuracy of the model while enhancing the overall model robustness.

[0151] The following metrics were used to evaluate the performance of the CNN model:

[0152] 1) Classification accuracy: The proportion of all samples correctly classified by the model.

[0153]

[0154] where TP is the number of positive samples correctly classified, TN is the number of negative samples correctly classified. FP is the number of negative samples misclassified as positive, and FN is the number of positive samples misclassified as negative.

[0155] 2) Sensitivity: The ability of the model to identify all positive samples, representing the proportion of actual positive cases correctly identified.

[0156]

[0157] 3) Precision: The proportion of positive samples correctly classified as positive by the model.

[0158]

[0159] The machine learning model trained in Application Example 1 can well distinguish the blank spectrum, the spectrum of exosomes from healthy sources, and the spectrum of cancer-related exosomes, with accuracies of 97.85%, 97.85%, and 99.38% respectively.

[0160] (5) Diagnostic module:

[0161] The diagnostic process is as Figure 5 shown. By using the CNN spectral classification model, an exosome digital count matrix of 76 participants (healthy volunteers and BCa patients) belonging to the training samples was obtained, and typical sample cases are as Figure 6 shown; from this, the counts of the blank spectrum (ε1), nEV spectrum (ε2), and taEV (ε3) were deduced, the count of total exosomes was calculated as ξ1 = ε2 + ε3, and the proportion of taEV in total exosomes was ξ2 = ε3 / ξ1. All thresholds were statistically determined based on the training data.

[0162] First, the threshold (δ1) was set to 53 (the third quartile ξ1 of healthy samples). If ξ1 < δ1, the individual was classified as healthy; otherwise, the person moved on to the next discrimination. Then, the threshold (δ2) was set to 10 (determined based on the lower limit after removing outliers from ε3 of early patients). If ε3 < δ2, the individual was considered non-cancerous; otherwise, the individual was diagnosed as a bladder cancer patient and moved on to the next discrimination. Finally, the threshold (δ3) was set to 0.38 (the third quartile ξ2 of early bladder cancer patient samples). If ξ2 < δ3, the individual was regarded as an early bladder cancer patient; otherwise, the individual was diagnosed as a late-stage bladder cancer patient.

[0163] On this basis, the diagnostic process of Application Example 1 was used to diagnose 11 healthy volunteers and 27 bladder cancer patient samples, and the results are as Figure 7 shown. It can be seen that the diagnostic system of Application Example 1 has extremely high comprehensive accuracy, with a precision of up to 97.37%.

Claims

1. A molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, characterized in that, The structure of the molecularly imprinted gold nanopore array substrate from bottom to top is a substrate and a gold nanoparticle layer. The gold nanoparticle layer contains recessed gold nanopores distributed in an array. The bottom of the gold nanopores has modified gold nanoparticles. The gold nanopores contain imprinted cavities matching exosomes. The size of each gold nanopore matches the size of a single exosome. The array-distributed gold nanopore channels form a spacing between single exosomes, and the spacing is ≥1 μm.

2. The molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 1, wherein, The substrate is a glass slide, a quartz slide, a silicon wafer, or a polydimethylsiloxane sheet; The bottom of the gold nanopores has modified gold nanoparticles, and 3-aminophenylboronic acid is modified on the surface of the gold nanoparticles at the bottom of the gold nanopores.

3. A preparation method of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 1, characterized in that, It includes the following steps: Step 1: Preparation of nanoparticle mask templates and gold nanoparticle seed solutions; Step 2: Preparation of nanoparticle mask template array substrates; Step 3: Preparation of nano-gold - gold nanopore array substrates; Step 4: Modification of exosome imprinted layers on gold nanopore channels.

4. The preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 3, characterized in that, The preparation of the nanoparticle mask template array substrate in Step 2 includes the following steps: (1) Place the substrate in piranha solution for surface treatment, then rinse repeatedly with deionized water, and then dry with an inert gas stream; (2) Drop the nanoparticle mask template on the dried substrate, then quickly add absolute ethanol, and dry with an inert gas stream at room temperature to form a 2D array substrate of nanoparticle mask templates on the top of the substrate; (3) Place the 2D array of nanoparticle mask templates in a plasma cleaner, and after etching, obtain a nanoparticle mask template array substrate with spacings.

5. The preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 4, wherein The time for substrate surface treatment in Step (1) is 4 h - 12 h; The added amount of absolute ethanol in Step (2) is 1 to 10 times the volume of the nanoparticles; The etching time in Step (3) is 90 s - 150 s.

6. The preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 3, characterized in that, The preparation of the nano-gold - gold nanopore array substrate in Step 3 includes the following steps: (1) Place the nanoparticle mask template array substrate and 3-aminopropyltriethoxysilane together in a dryer, evacuate and heat to perform amino modification on the exposed part of the substrate not covered by the nanoparticle mask template; (2) Add a gold nanoparticle seed solution on the top of the amino-functionalized nanoparticle mask template array substrate for incubation. The electrostatic adsorption of nano-gold occurs on the amino-modified exposed substrate part to obtain a nanoparticle - nano-gold hybrid array substrate, and then rinse repeatedly with deionized water; (3) Continue to perform ultrasonic treatment in the template cleaning solution to remove the nanoparticle mask template to obtain a nano-gold array substrate; (4) Place the nano-gold array substrate in a gold growth solution containing hydroxylamine hydrochloride and chloroauric acid trihydrate for oscillating incubation to in-situ grow a gold nanoparticle layer to obtain a gold nanopore array substrate, rinse with deionized water, and dry in an inert gas stream; Place the gold nanopore array substrate in a gold nanoparticle seed solution for incubation, deposit gold nanoparticles on the substrate area at the bottom of the gold nanopores through electrostatic interaction, then place it in a mixed solution of 3-aminophenylboronic acid and a coupling agent for incubation, then wash repeatedly with a washing buffer solution, and then place it in a blocking buffer solution for incubation to obtain a nano-gold - gold nanopore array substrate.

7. The preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 6, characterized in that, The heating temperature in Step (1) is 60 °C - 80 °C, and the heating time is ≥2 h; The incubation time in Step (2) is ≥4 h; In step (3), the template cleaning solution is an aqueous methanol solution or tetrahydrofuran, and the ultrasonic treatment time is 2 min - 5 min; In step (4), the concentration of hydroxylamine hydrochloride is 0.4 mM, the mass ratio of chloroauric acid to water in chloroauric acid trihydrate is 0.5% - 1%, and the incubation time is 2 h; In step (5), the concentration of 3-aminophenylboronic acid is 3 mg / mL - 6 mg / mL; The coupling agents are 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysuccinimide, and the concentrations of both 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysuccinimide are 20 mg / mL - 50 mg / mL. The incubation time in the mixed solution of the coupling agent and 3-aminophenylboronic acid is 60 min; The washing buffer is PBS buffer with a concentration of 5 mM - 10 mM and a pH value of 7.4; The blocking buffer is a PBS solution with a concentration of 5 mM - 10 mM, and 1 wt.% bovine serum albumin is added. The pH value is 7.4, the temperature is 37 °C, and the incubation time in the blocking buffer is 20 min - 60 min.

8. The preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection according to claim 3, wherein, Step 4: Modifying the exosome imprinting layer on the gold nanopore channels, including the following steps: (1) Adding the exosome template to the top of the gold nanoparticle - gold nanopore array substrate for incubation to anchor the exosome template at the bottom of the pores; (2) Immersing the substrate anchored with the exosome template into a tris(hydroxymethyl)aminomethane - hydrochloric acid buffer containing dopamine hydrochloride, and carrying out a polymerization reaction under stirring to form a gold nanoparticle - gold nanopore array substrate with a polydopamine exosome imprinting layer; (3) Immersing the gold nanoparticle - gold nanopore array substrate with exosome imprinting into PBS buffer for ultrasonic treatment, then soaking it in a mixed solution of acetic acid and acetonitrile to remove the exosome template, obtaining an exosome - imprinted gold nanopore array substrate, and then incubating it in the blocking buffer to obtain a molecularly imprinted gold nanopore array substrate for single exosome capture - in-situ Raman detection, and storing it at 4 °C in PBS buffer for later use.

9. The preparation method of a molecularly imprinted gold nanopore array substrate for single exosome capture-in-situ Raman detection according to claim 8, characterized in that, In step (1), the incubation time is 4 h - 12 h; In step (2), the concentration of dopamine hydrochloride is 2 mg / mL - 8 mg / mL, the concentration of the tris(hydroxymethyl)aminomethane - hydrochloric acid buffer is 5 mM - 10 mM, and the pH value is 8.5; the polymerization reaction time is 2 h - 12 h; In step (3), the concentration of PBS buffer is 5 mM - 10 mM, and the ultrasonic treatment time is 10 s - 120 s; The volume ratio of acetic acid to acetonitrile is 1:(9 - 49), and the soaking time is ≥10 min; The blocking buffer is a PBS solution with a concentration of 5 mM - 10 mM, and 1 wt.% bovine serum albumin is added. The pH value is 7.4, the temperature is 37 °C, and the incubation time in the blocking buffer is 30 min - 60 min.

10. Use of the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection according to claim 1, characterized in that, A digital diagnosis system for the early diagnosis of bladder cancer, the system includes a standard data acquisition module, a data to be detected acquisition module, a data screening module, a calculation module, and a diagnosis module; The data acquisition module uses the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection to capture standard exosomes, and then uses a Raman spectrometer to perform surface-enhanced Raman spectroscopy detection to collect standard exosome spectral data. The standard exosomes are exosomes derived from bladder cancer cells and healthy source exosomes obtained by ultra-high-speed centrifugation. The exosomes derived from bladder cancer cells include exosomes from three common bladder cancer cell lines, T24, BIU-87, and 5637 cells. At least 200 surface-enhanced Raman spectroscopy detections are performed on each sample, which is defined as the standard cancer cell exosome spectrum. The healthy source exosomes are taken from the urine of healthy volunteers, and at least 200 surface-enhanced Raman spectroscopy detections are performed on each sample, which is defined as the standard healthy exosome spectrum. At the same time, surface-enhanced Raman spectroscopy detection is performed on the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection, and the spectra of more than 1000 blank nanopores are collected, which is defined as the standard blank spectrum. The data acquisition module for the sample to be tested adjusts the pH value of the collected urine sample to be tested to 8.5, uses a sterile syringe to aspirate 2 mL of the urine sample, filters out larger impurities through a 0.22 μm filter membrane, immerses the molecularly imprinted gold nanopore array substrate for single exosome capture-in situ Raman detection in the filtrate, oscillates and incubates at 37 °C for 1 h, captures urine exosomes through the exosome imprinting structure on the surface, and then uses a Raman spectrometer to perform surface-enhanced Raman spectroscopy detection to collect exosome spectral data of the sample to be tested. At least 200 surface-enhanced Raman spectroscopy detections are performed on each sample. The data screening module pre-screened the invalid data in the standard spectral data of exosomes derived from bladder cancer cells, the standard spectral data of exosomes from healthy sources, and the spectral data of exosomes in the sample to be tested obtained by the data acquisition module. By using the average Raman intensity I -1 at 998 cm 998cm -1 as a benchmark to identify the invalid spectral data, the data with I 998cm -1 <0.9AverageI 998cm -1 was determined as invalid data, and the invalid data was deleted by an automatic screening program written in Python; The calculation module labels the standard spectral data of exosomes derived from bladder cancer cells and the standard spectral data of exosomes from healthy sources processed by the data screening module, and labels the standard blank spectrum obtained by the data acquisition module as the standard blank. Use K-Means to cluster the standard spectral data of exosomes derived from bladder cancer cells into two clusters, and then compare the Pearson correlation coefficient P with the standard blank spectrum. The spectrum with a significantly lower P value is determined as the exosome spectrum of bladder cancer cells and labeled as "standard taEV". Cluster the standard spectral data of exosomes from healthy sources into two clusters, and then compare the correlation coefficient P with the standard blank spectrum. The spectrum with a significantly lower P value is determined as the exosome spectrum of healthy sources and labeled as "standard nEV". Cluster the exosome spectra of bladder cancer patient subjects into three clusters, and compare the P values with the standard blank, standard taEV, and standard nEV respectively, and divide them into three categories, namely blank, nEV, and taEV spectra. After labeling the spectra as three categories: blank, nEV, and taEV spectra, a convolutional neural network model is further constructed for exosome spectrum classification. The CNN architecture is determined using a grid search algorithm, and class weights are used during model training. The described diagnostic module uses a CNN spectral classification model to obtain a digital count matrix of exosomes from healthy volunteers and BCa patients belonging to the training samples, from which the counts of blank spectra (ε1), nEV spectra (ε2), and taEV (ε3) are derived. The count of total exosomes is calculated as ξ1 = ε2 + ε3, and the proportion of taEV in total exosomes is ξ2 = ε3 / ξ1. All thresholds are statistically determined based on the training data; First, the threshold δ1 is set to the third quartile ξ1 of the healthy samples. If ξ1 < δ1, then the individual is classified as healthy; Otherwise, the person proceeds to the next discrimination. Then, the threshold δ2 is set based on the lower limit after removing outliers from ε3 of early patients. If ε3 < δ2, then the individual is considered non-cancerous; otherwise, the individual is diagnosed as a BCa patient and proceeds to the next discrimination. Finally, the threshold δ3 is set to the third quartile ξ2 of the samples of patients with advanced BCa. If ξ2 < δ3, then the individual is regarded as a patient with early-stage BCa; otherwise, the individual is diagnosed as a patient with advanced-stage BCa.

Images