Preparation method of honeycomb melamine-formaldehyde resin microspheres and application thereof in extracellular vesicle capture and typing detection

Abstract

The application discloses a preparation method of honeycomb melamine-formaldehyde resin microspheres and application of the honeycomb melamine-formaldehyde resin microspheres in cell extracellular vesicle capture and typing detection, and is based on honeycomb melamine-formaldehyde resin microspheres with a large pore structure and a surface easy modification characteristic to construct a cell extracellular vesicle capture and typing detection system, realizes efficient capture, and utilizes iron-doped carbon dot nanoscale enzymes with high catalytic activity and easy functionalization characteristics to construct a multiple signal amplification system, combines economy and flexibility of aptamers and data processing capacity of machine learning, and successfully realizes a leap from'multi-parameter detection' to 'intelligent phenotype typing'. Not only can different cell sources of cell extracellular vesicles be distinguished, but also early tumor patients and healthy controls can be identified in nanoliter-level clinical plasma samples, a new strategy with clinical transformation potential is provided for early non-invasive diagnosis of cancer, and new methodological support is provided for disease typing and precision medicine research based on EVs.

Description

Technical Field

[0001] This invention relates to the field of extracellular vesicle detection technology, specifically to a method for preparing honeycomb-shaped melamine-formaldehyde resin microspheres and their application in extracellular vesicle capture and typing detection. Background Technology

[0002] Extracellular vesicles (EVs) are nanoscale vesicles secreted by cells and are widely distributed in bodily fluids such as peripheral blood and urine. These tiny vesicles carry a wealth of bioinformatic molecules, including proteins and nucleic acids, and play a crucial role in the occurrence, development, and metastasis of tumors. As a highly promising liquid biopsy biomarker, tumor-associated EVs can dynamically reflect the real-time biological state of tumors through the expression types and levels of their membrane proteins, providing a minimally invasive yet sensitive observation window for early cancer detection and treatment monitoring. Compared to circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA), EVs exhibit unique advantages in practical applications: their abundance in plasma is extremely high, far exceeding that of CTCs, and they carry a richer array of biomolecular information; simultaneously, their lipid bilayer membrane structure provides excellent protection for their contents, exhibiting significantly better stability than ctDNA. Of particular note is that EV membrane proteins are directly derived from the surface of the blastocyst, and their expression profiles are closely correlated with tumor type, malignancy, and clinical stage, making them ideal targets for precise identification and subtyping. Therefore, establishing a technical system capable of performing multiplex, highly sensitive, and highly specific analysis of EV membrane proteins is of great significance for promoting precision diagnosis and treatment of cancer.

[0003] However, the nanoscale size, high heterogeneity, and interference from complex biological matrices of EVs pose multiple challenges to existing analytical methods. Current mainstream technologies can be divided into two categories: one is the stepwise strategy of "separation first, detection later," which is often cumbersome, has limited throughput, and is costly, making it difficult to meet the needs of multiplex analysis; the other is the integrated strategy (such as nanoflow cytometry), which simplifies the process but still relies on complex antibody labeling and expensive instruments. A deeper problem is that the heterogeneity of EVs makes reliable typing (such as differentiating between benign and malignant samples or different cancer subtypes) difficult to achieve using a single or a few biomarkers. This requires new technologies to generate high-dimensional data suitable for pattern recognition. In recent years, research has been moving towards integration and intelligence, such as enzyme logic gates based on metal-organic frameworks, biomimetic microfluidic chips, and nanozyme-assisted immunoassays without enzyme-labeled secondary antibodies. Although these systems have excellent performance, they mostly rely on expensive antibodies as recognition elements, limiting their scalability and cost control; while sensing schemes based on inexpensive aptamers often have simple structures and lack multiplex detection and intelligent data analysis capabilities. Therefore, how to deeply integrate the economic advantages of aptamers with advanced array sensing, nanozyme signal amplification, and machine learning typing to build an efficient and practical integrated platform has become a direction worth exploring. Summary of the Invention

[0004] This invention provides a method for preparing honeycomb-shaped melamine-formaldehyde resin microspheres and their application in the capture and typing of extracellular vesicles (EVs).

[0005] In view of this, the solution of the present invention is as follows: The first aspect of the present invention is to provide honeycomb-shaped melamine-formaldehyde resin microspheres, which are obtained by polymerization reaction of melamine, formaldehyde and acidic catalyst under the condition of macroporous microspheres as template, and then etching to remove the template; the pore size of the macroporous microspheres is in the range of 0.5-1 μm.

[0006] Furthermore, the honeycomb-shaped melamine-formaldehyde resin microspheres have a pore size of 0.5-1 μm.

[0007] And / or, the macroporous microspheres are silica microspheres, and the etching process for removing the template uses an alkaline solution.

[0008] Furthermore, the molar ratio of melamine to formaldehyde is 1:(5-20). And / or, the acidic catalyst is an inorganic acid; And / or, the polymerization reaction temperature is 30-80℃, and the reaction time is 0.5-15 h; And / or, the polymerization reaction is further complicated by the addition of polyethylene glycol, in an amount of 0.5-1% of the molar amount of melamine.

[0009] A second aspect of the present invention is to provide an EVs-capturing microsphere, said capturing microsphere being obtained by modifying a CD63 nucleic acid aptamer with honeycomb melamine-formaldehyde resin microspheres as described in the first aspect.

[0010] Furthermore, the nucleotide sequence of the CD63 aptamer is shown in SEQ ID NO:1; And / or, the CD63 nucleic acid aptamer is modified with carboxyl groups, and the modification process of the honeycomb melamine-formaldehyde resin microspheres involves incubating the CD63 nucleic acid aptamer with the honeycomb melamine-formaldehyde resin microspheres at room temperature to undergo a condensation reaction. Specifically, the amino groups on the surface of the honeycomb melamine-formaldehyde resin microspheres condense with the carboxyl groups of the nucleic acid aptamer to form amide bonds, thereby achieving the modification of the CD63 nucleic acid aptamer on the surface of the honeycomb melamine-formaldehyde resin microspheres.

[0011] A third aspect of the invention is to provide the use of the EVs-capturing microspheres described in the second aspect in the preparation of EVs-capturing or detection products.

[0012] A fourth aspect of the present invention is to provide an EVs genotyping detection kit, comprising the EVs capturing microspheres described in the second aspect, and a detection probe targeting a target protein, said target protein including MUC1, HER2, PD-L1, and CD63.

[0013] Furthermore, any of the detection probes is a nucleic acid aptamer corresponding to the target protein; And / or, the detection probe is modified with carbon dot nanozymes or their metal dopants; And / or, the kit may also include TMB chromogenic solution, peroxide, and buffer.

[0014] A fifth aspect of the present invention is to provide an EVs genotyping detection method for non-diagnostic purposes, comprising: S1. Use the capturing microspheres described in the second aspect to capture EVs in the sample; S2. Add detection probes and co-incubate with captured EVs; the detection probes include probes targeting MUC1, HER2, PD-L1 and CD63 proteins, and any of the probes includes a nucleic acid aptamer of the target protein and is modified with iron-doped carbon dot nanozymes; S3. Add colorimetric reagent, obtain colorimetric signals from different channels, and calculate the typing results.

[0015] Furthermore, the sample is a plasma sample; And / or, the acquired four-channel signals are constructed into a feature array and used for pattern detection through a linear discriminant analysis algorithm.

[0016] Preferably, when the aforementioned captured microspheres are used for genotyping and diagnostic purposes, the genotyping detection specifically involves: simultaneously detecting the absorbance values ​​of four channels at specific wavelengths to obtain the multi-protein expression spectrum on the surface of EVs; then constructing a feature vector from the obtained four-dimensional signal and inputting it into a classification model trained based on the LDA algorithm; automatically extracting the most discriminative feature combination and performing dimensionality reduction projection to complete intelligent genotyping and visual classification of EV sources, thereby enabling accurate differentiation between cancer patients and healthy individuals. The cancers mentioned include, for example, breast cancer or pancreatic cancer.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The honeycomb-shaped melamine-formaldehyde resin microspheres of this invention are formed by using macroporous microspheres with specific pore sizes as templates, with melamine-formaldehyde resin as the backbone. The surface is naturally rich in amino groups, which greatly facilitates the efficient covalent modification of CD63 aptamers and ensures capture specificity. In addition, the honeycomb structure with specific pore sizes provides a guarantee for the efficient capture of EVs.

[0018] The EVs capturing microspheres of this invention are based on honeycomb melamine-formaldehyde resin microspheres modified with CD63 aptamers, which meet the requirements of specific and efficient EVs capture and provide a basis for EVs typing and detection.

[0019] The genotyping detection kit of the present invention achieves capture through EVs-capturing microspheres and amplifies the signal through detection probes modified with carbon nanozymes or their metal dopants, which can effectively improve detection quality and detection rate.

[0020] The EV genotyping detection method described in this invention integrates the economic efficiency of aptamers, the efficient signal amplification of nanozymes, and the design concept of arrayed sensing, constructing a practical EV phenotyping platform. Furthermore, it verifies the feasibility of the "functional materials-biorecognition-data intelligence" collaborative strategy through practice, providing new ideas for the development of cancer liquid biopsy and precision medicine. Attached Figure Description

[0021] Figure 1 This is a SEM image of the SiO2 microspheres used in the HMFMs preparation process described in Example 1 of the present invention.

[0022] Figure 2 The above are the characterization results of the HMFMs described in Example 1 and the aptamer modified in Example 2 of this invention.

[0023] Figure 3 This is a comparison of the EV capture performance of different HMFMs in Example 3 of the present invention.

[0024] Figure 4The above are the characterization results of the FeCDzymes and Apt-FeCDzymes composite probes described in Example 4 of this invention.

[0025] Figure 5 This is the verification result of the FeCDzymes catalytic activity described in Example 4 of the present invention.

[0026] Figure 6 The results show the FeCDzymes catalytic mechanism and the performance changes after Apt modification as described in Example 4 of this invention.

[0027] Figure 7 This is a comparison and characterization result of EVs of different concentrations before and after treatment with different materials as described in Example 5 of the present invention.

[0028] Figure 8 This is the result of differential expression of four membrane proteins on the surface of five CM-EVs in Example 5 of the present invention.

[0029] Figure 9 In Embodiment 6 of the present invention, five different CM-EVs are respectively connected to four sensing elements (Apt). CD63 Apt HER2 Apt MUC1 Apt PD-L1 A linear relationship between the colorimetric signal generated during the interaction and its EV concentration.

[0030] Figure 10 This is a classification efficiency diagram of the phenotypic analysis system for EVs from different sources under different EV concentration conditions in Embodiment 6 of the present invention.

[0031] Figure 11 This is the performance verification result of the clinical sample detection and phenotypic analysis system in Embodiment 6 of the present invention.

[0032] Figure 12 The comparison results of EV concentration and membrane protein detection in the plasma of cancer patients and healthy individuals in Example 6 of this invention. Detailed Implementation

[0033] The technical solution of the present invention will now be clearly and completely described in conjunction with preferred embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] In one embodiment of the present invention, honeycomb-like melamine-formaldehyde microspheres (HMFMs) with interconnected honeycomb-like channels were prepared using a sacrificial template method. Compared with ordinary mesoporous materials, their micron-sized interconnected channels provide unobstructed mass transfer pathways and high-loading interfaces for EVs (EVs), fundamentally improving the problem of insufficient capture efficiency of traditional magnetic beads or mesoporous materials. Specifically, the above-mentioned HMFMs are obtained by polymerization of melamine, formaldehyde, and an acidic catalyst under the condition of macroporous microspheres as templates, followed by etching to remove the template; the pore size of the macroporous microspheres ranges from 0.5 to 1 μm.

[0035] In the above embodiments, the surface of HMFMs is naturally rich in amino groups, which greatly facilitates the efficient covalent modification of CD63 aptamers. While ensuring capture specificity, it avoids the cumbersome activation and modification steps required for carriers such as silica gel.

[0036] In a second embodiment of the present invention, an EV-capturing microsphere is proposed, which is made by modifying the surface of HMFMs described in the above embodiment with a CD63 aptamer. The carboxyl-modified CD63 aptamer is incubated with the HMFMs, causing the amino groups abundant on the surface of the HMFMs to undergo a condensation reaction with the carboxyl groups to form amide bonds. The HMFM surface modification with CD63 aptamer ensures efficient and specific EV capture, facilitating subsequent genotyping and detection.

[0037] In some embodiments, the pore size of the EVs capturing microspheres is 0.5-1 μm, and the capture efficiency for EVs with a diameter of 30-150 nm reaches 100%. However, when the pore size of HMFMs increases, the contact sites between HMFMs and EVs are reduced, resulting in a decrease in the capture efficiency of EVs.

[0038] In the third embodiment, to achieve multiple and sensitive signal conversion, iron-doped carbon dot nanozymes (FeCDzymes) were designed and synthesized as the signal amplification core. These nanozymes exhibit excellent peroxidase-like activity, converting target recognition events into visually visible, low-value instrument-readable colorimetric signals. Their abundant carboxyl groups also facilitate modular assembly with specific aptamers of different EV membrane proteins, thereby constructing a flexibly configurable "recognition-signal" probe library.

[0039] In the fourth embodiment, a phenotypic analysis system for aptamer-engineered EVs was constructed. The entire process encompasses three seamlessly connected stages: (1) Efficient capture: HMFMs modified with CD63 aptamer specifically enrich EVs from the sample; (2) Multiple signal conversion: Four aptamer-FeCDzymes probes were used to simultaneously recognize MUC1, HER2, PD-L1 and CD63 proteins and generate corresponding colorimetric signals; (3) Intelligent typing: The acquired four-channel signals are constructed into a feature array, and automatic classification is achieved through the linear discriminant analysis (LDA) algorithm. The system not only successfully distinguished EVs from eight different cell types (HPNE, PANC-1, MCF-10A, MCF-7, MDA-MB-231, A549, HGC-27, HCT-15) (covering a variety of cancer cells and normal cells), but also demonstrated good clinical applicability: without relying on ultracentrifugation pre-separation, the system can differentiate between healthy individuals and early-stage cancer patients by capturing and phenotyping EVs in nanoliter plasma samples.

[0040] In the above embodiments, the working principle of the EVs phenotypic analysis system is as follows: First, amino-rich HMFMs were used as a solid-phase support, whose unique interconnected micron-sized pore structure provided an interface with low mass transfer resistance and high loading capacity. Then, an aptamer targeting the universal biomarker CD63 was covalently modified onto its surface to form an Apt. CD63 -HMFMs enable efficient and specific enrichment from complex biological samples. After capture, the system enters a multi-parallel recognition and signal conversion stage: four modular "recognition-signal" probes are constructed, each covalently linked to an aptamer targeting MUC1, HER2, PD-L1, and CD63 membrane proteins with FeCDzymes; these probes are then used in conjunction with the captured Apt... CD63 After co-incubation with HMFMs, the FeCDzymes bind to corresponding target proteins on the surface of EVs through the specific recognition of aptamers, thereby precisely anchoring the peroxidase-like FeCDzymes to the EV surface. Subsequently, H2O2 and the chromogenic substrate TMB are added. The FeCDzymes catalyze the oxidation of colorless TMB to blue oxTMB, generating a colorimetric signal proportional to the abundance of each target protein. By simultaneously detecting the absorbance values ​​of four channels at 652 nm, the multi-protein expression profile on the EV surface can be obtained in one step. Finally, the system constructs a feature vector from the obtained four-dimensional signal and inputs it into a classification model trained based on the LDA algorithm. By automatically extracting the most discriminative feature combinations and performing dimensionality reduction projection, intelligent typing and visual classification of EV-originating factors are completed, enabling accurate differentiation between breast cancer or pancreatic cancer patients and healthy individuals.

[0041] In this invention, the amino or carboxyl-modified aptamers (Apt) targeting EV membrane proteins were synthesized by Shanghai Sangon Biotech Co., Ltd. The aptamer nucleotide sequences involved are as follows: Apt CD63 :CACCCCACCTCGCTCCCGTGACACTAATGCTA (SEQ ID NO: 1); Apt MUC1 :GCAGTTGATCCTTTGGATACCCTGG (SEQ ID NO: 2); Apt HER2 :GGGGTGTGGCGACG (SEQ ID NO: 3); Apt PD-L1 : TACAGGTTCTGGGGGGTGGGTGGGGAACCTGTT (SEQ ID NO: 4).

[0042] In the following examples, the silica microspheres were synthesized according to the previously reported method (Li J, Li LS, Xu L. Hierarchically macro / mesoporous silica sphere: A high efficient carrier for enzyme immobilization[J]. Microporous & Mesoporous Materials, 2016), and the particle size was controllable.

[0043] Example 1: Preparation and characterization of honeycomb-shaped melamine-formaldehyde resin microspheres (HMFMs)

[0044] 1) Preparation

[0045] Macroporous HMFMs materials were synthesized using a sacrificial template method. The specific steps are as follows: 1.25 g of Melamine and 0.6 g of PEG 10000 were weighed and dissolved in 8 mL of formaldehyde aqueous solution (37%). Then, 0.5 mL of 1 M HCl was added, and the mixture was magnetically stirred in a 50°C water bath until clear and transparent. 200 μL of the above solution was mixed with 100 mg of SiO2 microsphere template, and the mixture was reacted in a 50°C air bath for 12 h. After the reaction, the solid product was collected by centrifugation and washed repeatedly with methanol and pure water. Subsequently, the obtained material was placed in NaOH solution (2 M) and shaken for 1 h to etch away the SiO2 template. After etching, the material was collected by centrifugation again and washed several times with methanol and deionized water until the pH of the supernatant was neutral. The HMFMs material was then dried to obtain the final product.

[0046] The SiO2 microspheres used in this preparation example have a porous structure, as shown in the SEM image below. Figure 1 As shown.

[0047] The SiO2 microspheres used in this preparation example include two pore sizes, 1-3 μm and 0.5-1 μm, and the materials prepared are named HMFMs-1 and HMFMs-2 respectively.

[0048] 2) Characterization

[0049] The characterization results of the prepared HMFMs-2 material are as follows: Figure 2 As shown. Figure 2 The SEM image in A shows that it perfectly replicates the macroporous structure of the through-flow SiO2 template, with a morphology similar to a honeycomb. Figure 2 The porosity test results of the medium-pressure mercury method (BJH method) show that the pore size is mainly distributed between 500 nm and 1 μm, a structural feature crucial for the capture of EVs. Traditional magnetic beads or mesoporous materials cannot fully utilize the inner surface of the material due to pore channel limitations, while HMFMs with micron-scale interconnected channels provide an unobstructed mass transfer path and a high-loading interface. Nitrogen adsorption-desorption isotherms and mesopore size distribution calculated based on the BJH model are also presented. Figure 2 C) shows that its BET specific surface area is approximately 3.3 m². 2 / g. Figure 2D shows the FT-IR spectrum of the prepared HMFMs-2 material. The spectrum is located at 3355 cm⁻¹. -1 The broad absorption peak at 2933 cm⁻¹ is attributed to the stretching vibrations of -NH₂ and -OH groups in the material, and may also include contributions from trace amounts of residual moisture. -1 1486 cm -1 and 1202 cm -1 The absorption bands observed at [location] correspond to the stretching, bending, and rocking vibrations of the methylene bridge (-CH2-), respectively, collectively confirming the methylene cross-linked structure formed by formaldehyde participation in the reaction. The core aromatic heterocyclic skeleton of the material was confirmed by the characteristic signal of the triazine ring: 1552 cm⁻¹ -1 The strong peak at 910 cm⁻¹ is attributed to the skeletal stretching vibration of the triazine ring, while the peak at 910 cm⁻¹ is attributed to the skeletal stretching vibration of the triazine ring. -1 With 815 cm -1 The absorption peaks at 1150 cm⁻¹ all originate from the out-of-plane bending vibration of the triazine ring. Furthermore, the absorption peak at 1150 cm⁻¹... -1 The signal at this location is clearly attributed to the stretching vibration of the CN bond. (At 950-1050 cm⁻¹) -1 The composite signal appearing in the region should be attributed to the overlap of the stretching vibrations of the ether bond (COC) and the CO in the hydroxymethyl group (-CH2OH). The characteristic peaks match the signals of MF-type polymer materials reported in the literature, fully demonstrating the successful synthesis of the target product HMFMs. Figure 2 E presents the thermogravimetric analysis (TGA) results of HMFMs-2. Below 200°C, the material exhibits only slight mass loss, mainly due to the desorption of adsorbed water, indicating that its framework maintains excellent thermal stability within this temperature range. As the temperature rises to 800°C, the polymer framework undergoes complete pyrolysis, with a final weight loss of 94.37%, leaving approximately 5% carbonaceous residue. This residue rate is perfectly consistent with the typical pyrolysis behavior of highly cross-linked melamine-formaldehyde resin in an inert atmosphere, further confirming that the SiO2 template was completely removed during the synthesis process, and the obtained material is a pure organic porous polymer.

[0050] Example 2: Adaptor modification of HMFMs materials (Apt-HMFMs)

[0051] To prepare Apt-HMFMs, 50 μL of COOH-Apt was added to 2 mL of EDC / NHS (prepared with MES buffer at pH 6.0, 40 mM: 20 mM). CD63 The carboxyl groups were activated by incubation at room temperature for 30 minutes in a 100 μM solution. Subsequently, 60 mg of HMFMs were added to the solution, and the mixture was incubated overnight in a shaker at room temperature. After the reaction, the material was washed sequentially with PBS and ultrapure water to remove unbound Apt. The resulting product, abbreviated as Apt-HMFMs, was resuspended in 1 mL of PBS to prepare a 60 mg / mL stock solution, which was stored at 4°C for later use.

[0052] HMFMs via Apt CD63 The change in zeta-potential before and after modification is as follows: Figure 2 As shown in F. In pure water, the amino groups on the surface of HMFMs are fully protonated, thus exhibiting a positive charge; in modified Apt... CD63 Subsequently, the positively charged amino groups on the material surface are consumed through amidation, while a nucleic acid aptamer rich in a negatively charged phosphate backbone is successfully grafted, leading to a reversal of the surface potential to -14.1 mV. This potential reversal confirms the Apt... CD63 The successful immobilization on the surface of HMFMs lays the material foundation for the subsequent specific recognition and capture of EVs.

[0053] Example 3: Enrichment of EVs

[0054] To verify the feasibility of the EVs capturing microspheres of the present invention, we conducted two verification experiments.

[0055] 1. The efficient and specific capture of EVs by Apt-HMFMs was evaluated through a series of adsorption experiments and NTA detection: 900 μL of HMFMs (2 mg / mL) and HMFMs blocked with 5% BSA (2 mg / mL) were mixed with 450 μL of EV suspension to prepare a final concentration of 1×10⁻⁶. 8 A mixture of EVs with particles per mL; 2. Take 900 μL of Apt-HMFMs (2 mg / mL) blocked with 5% BSA and mix it with 450 μL of EV suspensions of different concentrations to prepare solutions with a final concentration of 1×10⁻⁶. 8 Particles / mL and 4×10 7 A mixture of EVs with particles per mL.

[0056] All systems were incubated overnight at 37°C, followed by centrifugation to collect the supernatant. The concentration of residual EVs in the supernatant was determined by NTA and compared with the initial EV concentration in the mixed system to evaluate the EV capture efficiency and specificity of Apt-HMFMs and HMFMs. Simultaneously, Cryo-TEM was used to observe the morphology of Apt-HMFMs after binding with EVs.

[0057] In the embodiments, we adjusted the pore size of the through-flow SiO2 microsphere template to obtain HMFMs with different pore sizes (such as...). Figure 3 As shown in the figure, HMFMs-2 (average pore size 500-1000 nm) was used in this study, while HMFMs-1 had an average pore size of 1-3 μm. After modification with AptCD63, the EVs capture performance of HMFMs-1 and HMFMs-2 was compared (initial EV concentration was 1×10⁻⁶). 7 The particle concentration of remaining EVs was determined using NTA after capture (particles / mL). Calculations showed that the capture efficiencies of HMFMs-1 and HMFMs-2 were 61.2% and 100%, respectively. This may be related to the excessively large pore size of HMFMs-2, leading to a significant decrease in specific surface area and thus reducing the contact sites between HMFMs and EVs. (The specific surface area of ​​HMFMs-1 is 3.3 m² / mL). 2 The specific surface area of ​​HMFMs-2 is 0.8 m² / g. 2 Therefore, when using HMFMs for EV capture, the pore size needs to be appropriate to ensure the permeability of EVs with molecular size of around 100 nm, while also maintaining sufficient specific surface area to ensure a sufficient number of interaction sites.

[0058] Example 4: Preparation, characterization, and performance testing of FeCDzymes and Apt-FeCDzymes composite probes

[0059] 1. FeCDzymes were prepared using a one-step hydrothermal method. The specific steps are as follows: 68 mg of FeCl3 6H₂O and 40 mg of L-Try were dissolved in 10 mL of ultrapure water and placed in a 25 mL polytetrafluoroethylene-lined reactor. After sonication to ensure complete dissolution of the raw materials, the reactor was placed in an oven and reacted at 160 °C for 10 h. After the reaction, the system was cooled to room temperature and filtered through a 0.22 μm filter membrane to remove unreacted large particulate impurities. Subsequently, the filtrate was placed in a dialysis bag and dialyzed in deionized water for 48 h (with water changed every 6 h) to remove small molecule impurities. Finally, the solution was freeze-dried to obtain a brown FeCDzymes solid powder. To obtain FeCDzymes with optimal enzyme activity, FeCl₃ was used during the FeCDzymes synthesis process. The key synthesis conditions, including the feed ratio of 6H₂O:L-Try (1:4, 1:2, 3:4, 1:1, 5:4, 3:2), hydrothermal temperature (140–180 °C), and reaction time (6–16 h), were systematically optimized. The morphology, structure, and chemical composition of the FeCDzymes prepared under the optimal conditions were characterized by TEM, STEM-EDS mapping, XRD, FTIR, and XPS, respectively.

[0060] The results showed that when FeCl3 The FeCDzymes obtained exhibited the highest catalytic activity when the molar ratio of 6H2O to L-Try was 5:4, the hydrothermal temperature was 160℃, and the reaction time was 10h. Subsequent characterization was performed based on samples prepared under these optimal conditions.

[0061] 2. The characterization results of FeCDzymes are as follows: Under TEM, FeCDzymes appeared as well-dispersed spherical or near-spherical particles with an average particle size of 3.25 ± 0.14 nm. Figure 4 A), which conforms to the morphological characteristics of typical carbon dots. The high-resolution TEM image (inset) shows clear lattice fringes with a plane spacing of approximately 0.22 nm, corresponding to sp... 2 The (100) crystal plane of graphite carbon. XRD pattern ( Figure 4 B) A broadened diffraction peak is observed at approximately 23.7°, further confirming its amorphous / graphite composite structure of graphitized carbon. Scanning transmission electron microscopy-energy dispersive spectroscopy (STEM-EDS mapping) shows that the elemental signals of C, O, N, and Fe are uniformly distributed within the nanoparticles and highly consistent with the morphology. Figure 4 C). FT-IR spectrum ( Figure 4 In D), 3200-3400 cm -1The broad absorption peak at 2930 cm⁻¹ is attributed to the stretching vibrations of O–H and N–H; -1 The absorption peak at 1632 cm⁻¹ corresponds to the C–H stretching vibration in the alkyl chain, indicating that the skeletal structure of the organic precursor tryptophan is partially preserved. -1 The sharp, strong absorption peak observed at 1300-1400 cm⁻¹ is attributed to the characteristic stretching vibration peak of the C=O bond in amide I or carboxylic acids. -1 Between and 1095 cm -1 The signals at these locations can be attributed to the stretching vibrations of the C–N and C–O bonds, respectively. Of particular importance is the signal at 1040 cm⁻¹. -1 The characteristic peaks appearing nearby can be attributed to the vibrations of the Fe–N bond, indicating that iron has been successfully doped into the carbon dot matrix via chemical bonding. (XPS full-scan spectrum) Figure 4 E) further confirmed the presence of C, N, O, and Fe. High-resolution spectrum ( Figure 4 F) Analysis shows that the C1s spectrum can be unconvolved into two main peaks of 284.8 eV and 288.8 eV, which are attributed to C–C / C=C and O–C=O bonds, respectively; the peak of 531.9 eV in the O1s spectrum can be identified as a C=O bond; the N1s spectrum shows two characteristic peaks of 399.9 eV (pyridinic N) and 401.7 eV (pyrrolic N); the Fe 2p spectrum shows a peak of 710.7 eV (Fe 2p). 2+ 2p 3 / 2 ), 724.5 eV (Fe 2+ 2p 1 / 2 ) and 714.5 eV (Fe 3+ 2p 3 / 2 ), 727.9 eV (Fe 3+ 2p 1 / 2 Two sets of spin orbital bimodal peaks confirm the presence of Fe in the material. 2+ and Fe 3+ The coexistence also indicates that FeCDzymes affect some Fe 3 + It has a reducing effect. In addition, the satellite peaks observed near 718.1 eV and 731.0 eV further support the formation of Fe–N and Fe–O coordination structures.

[0062] The characterization results of the above system fully confirm the successful preparation of FeCDzymes with clear structural features and chemical composition.

[0063] 3. FeCDzymes catalytic activity test

[0064] To evaluate the application potential of FeCDzymes in practical detection systems, we further investigated the dependence of its peroxidase-like activity on reaction conditions and its catalytic mechanism. The evaluation or testing methods are as follows: The POD-like activity of FeCDzymes was assessed using a catalytic TMB-H2O2 colorimetric system: 10 μL of FeCDzymes suspension (1 mg / mL) was added to 170 μL of NaAc-HAc buffer (0.2 M, pH 4), followed by the addition of 10 μL of TMB solution (10 mM) and 10 μL of H2O2 solution (10 mM). After incubating the mixture at 25 °C for 10 min, the absorbance at 652 nm was measured using a Varioskan LUX multi-mode microplate reader. 652 nm To investigate the effect of pH on the POD-like activity of FeCDzymes, the pH of the NaAc-HAc buffer was adjusted to (3–5.5), while other conditions remained constant, and the above experiments were performed. Furthermore, we studied the enzymatic properties of FeCDzymes through steady-state kinetic experiments. The specific procedures were as follows: keeping the total volume of the reaction system and other conditions constant, the concentration range of TMB (0.3–2.5 mM) or H₂O₂ (0.1–2 mM) in the system was adjusted, and A was recorded. 652nm Plot the Michaelis-Menten curve and the double reciprocal Lineweaver-Burk plot with respect to the initial rate of change over time (v=ΔA / Δt), and the Michaelis constants of TMB and H2O2. K m ) and enzyme-catalyzed reactions reach their maximum rate ( V max The result is calculated based on the Lineweaver-Burk model (Equation 1).

[0065] (Equation 1)

[0066] The test results are as follows: The effect of buffer pH on its catalytic activity is as follows: Figure 5 As shown in Figure A, FeCDzymes exhibited high activity in the pH range of 3.0-4.0, with activity gradually decreasing as the pH rose to 5.5. To avoid the potential impact of a strongly acidic environment on stability, subsequent experiments were conducted in a buffer system at pH 4.0. Steady-state kinetic analysis revealed that, within a certain concentration range, the reaction rate V increased with increasing TMB or H2O2 concentration and gradually approached saturation. Figure 5B, 5C). The Lineweaver-Burk double reciprocal plot in the illustration shows a good linear relationship, indicating that the catalytic process conforms to a typical Michaelis kinetic model. The calculated enzyme kinetic parameters show the Michaelis constant of FeCDzymes with respect to H2O2. K m The maximum reaction rate was 117.45 μM. V max It is 0.098 μM·s -1 For TMB K m 826.2 μM, V max It is 0.18 μM·s -1 These nanozymes generally exhibit superior H2O2 affinity to natural enzymes or other substrates through the specific binding of the Fe active center to the substrate.

[0067] To elucidate its catalytic mechanism, we conducted free radical quenching experiments, such as... Figure 6 As shown in Figure A, when adding ·OH quencher (MeOH, 10 mM) or ·O2... - After adding the quencher (PBQ, 10 mM), the absorbance of the system at 652 nm decreased significantly, while the addition of... 1 The O2 quencher (FA) did not cause significant changes, indicating that ·OH and ·O2 are mainly produced during the FeCDzymes-catalyzed oxidation of TMB. - Two types of reactive free radicals. We further verified this using EPR characterization. Figure 6 B) In systems using DMPO as the trapping agent, the differences between DMPO / ·OH and DMPO / ·O2 can be clearly observed. - The typical signal peaks further confirm the formation of these two free radicals during the reaction. In summary, FeCDzymes can efficiently catalyze the decomposition of H₂O₂ to produce ·OH and ·O₂ under acidic conditions. - This process oxidizes TMB to generate the blue product oxTMB, providing a reliable enzymatic basis for its application as a signal amplification unit in sensing.

[0068] 4. Preparation of Apt-FeCDzymes composite probe

[0069] To prepare Apt-FeCDzymes, 2 mg of FeCDzymes was dispersed in 2 mL of EDC / NHS (40 mM: 20 mM) activation buffer and activated at room temperature for 30 minutes. Then, 25 μL of each of the four NH2-modified Apt solutions (CD63, MUC1, PD-L1, HER2, all at 100 μM) were added, and the mixture was reacted overnight at 4°C. After the reaction, the mixture was purified by ultrafiltration centrifugation (3 kDa) to remove free Apt and FeCDzymes. The retained fraction was resuspended in PBS to obtain a stock solution of Apt-FeCDzymes complex with a concentration of approximately 1 mg / mL, which was stored at 4°C for later use. Under fixed reaction conditions, different Apt dosages (in Apt...) were measured... CD63 For example, the POD-like activity of the prepared Apt-FeCDzymes was measured under conditions of 0.5, 1.25, 2.5, 3.75 and 5 μM. The optimal feed concentration of Apt was optimized to balance the catalytic performance of Apt-FeCDzymes with the preparation cost.

[0070] 5. Characterization of the Apt-FeCDzymes composite probe

[0071] FeCDzymes inherits a large number of carboxyl groups from the precursor L-Try structure. Under the activation system of EDC / NHS, it can covalently couple with amino-modified aptamers through amide condensation reaction. Subsequently, unbound free FeCDzymes are removed by ultrafiltration purification to obtain the target product Apt-FeCDzymes composite probe.

[0072] Compare the IR spectra of FeCDzymes before and after modification with the aptamer ( Figure 4 D) As can be seen, the signal of the newly formed amide bond in Apt-FeCDzymes is difficult to distinguish directly because it overlaps with the amide bond contained in FeCDzymes itself. However, the successful introduction of the aptamer still caused several identifiable spectral changes: 1230 cm⁻¹ -1 The enhancement and broadening of the nearby absorption peaks can be attributed to the asymmetric stretching vibration of P=O in the aptamer's phosphate skeleton; 1000~1100 cm⁻¹ -1 The signal in the region is significantly enhanced and broadened, mainly due to the superposition of PO vibrations in phosphate esters with nucleic acid sugar rings and residual CO vibrations; 1600~1700 cm⁻¹ -1 The broadening of the absorption band at 2920 cm⁻¹ may be related to the ring vibration of the aptamer base contributing to the amide I band; furthermore, at 2920 cm⁻¹... -1 With 2950 cm -1The faint absorption peak appearing nearby likely originates from the -CH2 stretching vibration of the alkyl chain in the C6 amino modifier introduced at the 5′ end of the aptamer. Further analysis of surface charge changes using zeta potentials... Figure 6 C), FeCDzymes exhibit negative charge (-8.2 mV) in a neutral aqueous phase. Despite the presence of residual α-amino and indole groups from the precursor tryptophan, and trace amounts of Fe... 3+ While the surface may contribute some positive charge, the deprotonation (-COO-) of numerous carboxyl groups under neutral conditions still dominates its negative charge properties, providing a chemical basis for its subsequent amidation coupling with amino-modified aptamers. After aptamer coupling, the surface potential further decreased to -35.8 mV. This change is mainly attributed to the high-density negative charge carried by the phosphate backbone of the nucleic acid aptamer covering the material surface, thereby further enhancing the negative charge of the system; simultaneously, the potential influence of the EDC / NHS reagent on surface groups during coupling and the changes in interfacial hydration structure caused by the aptamer layer may also have an auxiliary influence on the potential measurement. In summary, the above experimental evidence collectively demonstrates the successful preparation of Apt-FeCDzymes. Furthermore, under the condition of fixed FeCDzymes feed amount, we investigated the effect of different aptamer concentrations on the catalytic activity of the obtained complex. Figure 6 (D) When the aptamer concentration increased from 0.5 μM to 1.25 μM, the absorbance at 652 nm increased significantly; further increasing the aptamer concentration did not result in a further increase in absorbance, indicating that the effective binding sites on the FeCDzymes surface were approaching saturation. Considering the cost-effectiveness of practical applications, and the possibility that excessively high aptamer concentrations might inhibit the accessibility of the catalytic interface due to steric hindrance or electrostatic effects, 1.25 μM was ultimately selected as the optimal concentration for FeCDzymes aptamer modification.

[0073] Based on this, this embodiment successfully prepared and systematically characterized FeCDzymes with high peroxidase-like activity, clarified its structural characteristics, catalytic mechanism and aptamer modification process, and it can be used as a high-efficiency signal amplification unit for EVs phenotypic analysis system.

[0074] Example 5: Phenotypic Analysis System for EVs Based on Aptamer Engineering

[0075] The EVs phenotypic analysis system constructed in this embodiment is based on the aptamer-functionalized materials Apt-HMFMs and Apt-FeCDzymes prepared in the above embodiments.

[0076] Specifically, the construction process is as follows: 1) Incubate 100 μL of Apt-HMFMs (2 mg / mL) with 50 μL of EVs dispersion at 37℃ for 30 min. After the EVs are specifically captured, centrifuge and discard the supernatant.

[0077] 2) Add 40 μL of Apt-FeCDzymes solution (1 mg / mL) corresponding to the four membrane proteins (CD63, MUC1, PD-L1, HER2) to the precipitate, incubate at 37℃ for 20 min to achieve binding of EVs and signal probes, centrifuge and wash with PBS to remove non-specifically adsorbed probes.

[0078] 3) Resuspend the precipitate in 180 μL of HAc-NaAc buffer (0.2 M, pH 4.0), then add 10 μL of TMB (10 mM) and 10 μL of H2O2 (10 mM) sequentially to initiate the colorimetric reaction. After reacting at 37℃ for 20 min, measure A. 652 nm The absorbance of the blank control was subtracted. The absorbance of the four channels was simultaneously detected. 652 nm The algorithm acquires the expression profiles of multiple proteins on the surface of EVs in a single step, constructs the resulting four-dimensional signals into feature vectors, and inputs them into a classification model based on a linear discriminant analysis algorithm. This model achieves dimensionality reduction and projection of the data by calculating a discriminant function, ultimately completing intelligent typing and visual classification of EV origins.

[0079] To verify the feasibility of the EV Phenotyping System constructed in this embodiment, we conducted two verification experiments.

[0080] 1. The efficient and specific capture of EVs by Apt-HMFMs was evaluated through a series of adsorption experiments and NTA detection: 900 μL of HMFMs (2 mg / mL) and HMFMs blocked with 5% BSA (2 mg / mL) were mixed with 450 μL of EV suspension to prepare a final concentration of 1×10⁻⁶. 8 EVs mixture system with particle number / mL; separately, 900 μL of Apt-HMFMs (2 mg / mL) blocked with 5% BSA was mixed with 450 μL of EVs suspensions of different concentrations to prepare a final concentration of 1×10⁻⁶. 8 Particle count / mL and 4×10 7 EVs mixture systems with particle counts / mL were established. All systems were incubated overnight at 37°C, followed by centrifugation to collect the supernatant. The residual EV concentration in the supernatant was determined by NTA and compared with the initial EV concentration in the mixture system to evaluate the EV capture efficiency and specificity of Apt-HMFMs and HMFMs. The morphology of Apt-HMFMs bound to EVs was observed using Cryo-TEM.

[0081] 2. To verify the applicability of the four selected protein biomarkers (CD63, HER2, PD-L1, and MUC1) in the construction of the sensing system, the expression levels of HER2, PD-L1, MUC1, and CD63 in EVs from five cell lines (HPNE, PANC-1, MCF-10A, MCF-7, and MDA-MB-231) were detected using nanoflow cytometry. The simplified procedure was as follows: Appropriately diluted EVs were incubated with fluorescently labeled antibodies (PE anti-human CD63, APC anti-human PD-L1, Alexa Fluor 488 Biosimilar anti-human HER2, and BV421 Mouse anti-human MUC1) at 37°C in the dark for 60 min. Subsequently, 1 mL of pre-chilled PBS was added for dilution, and the mixture was ultracentrifuged at 120,000 ×g for 70 min at 4°C to remove free antibodies. After resuspending the precipitate in PBS, it was centrifuged and washed again under the same conditions, and finally resuspended in an appropriate amount of PBS. The precipitate was then analyzed using a nanoflow cytometer equipped with a four-color laser (405 nm, 488 nm, 561 nm, 638 nm). Standard nanoparticle calibration was performed before detection, a forward scattering threshold was set to eliminate background noise, and at least 10,000 valid particle events were collected for subsequent analysis.

[0082] 2. Verification Results

[0083] 1) The principle of Apt-HMFMs' specific capture of EVs was verified using NTA and Cryo-TEM. The changes in particle concentration before and after treatment with different materials for two different concentrations of EVs were measured using NTA. The results are as follows: Figure 7 As shown in Figure A. For an initial concentration of 1×10⁻⁶... 8 HMFMs with particle / mL EVs (a) and without bovine serum albumin (BSA) blocking exhibit significant nonspecific adsorption (b), which is likely due to the inherent affinity of abundant hydrophilic groups such as amino and hydroxyl groups on the material surface for proteins. Such nonspecific adsorption may lead to the co-enrichment of impurities such as lipoproteins and cell debris in complex biological samples, resulting in quantitative deviations. After BSA blocking, the HMFMs essentially lost their ability to capture EVs (c), indicating that BSA effectively occupied the nonspecific binding sites on the material surface. In contrast, HMFMs covalently modified with Apt... CD63 After being sealed with BSA, the material can achieve specific capture (d) of EVs, demonstrating that Apt CD63 It plays a crucial role in mediating the specific recognition of EVs; however, at this concentration, approximately 40% of EVs were not captured, suggesting that the capture capacity or affinity can be further optimized. When the initial EV concentration was reduced to 4 × 10⁻⁶...7 When particles / mL (e), Apt blocked by BSA CD63 -HMFMs can achieve near-complete capture (f), indicating that the capture system is highly efficient at appropriate target concentrations.

[0084] 2) The morphology of Apt-HMFMs after EV capture was observed using Cryo-TEM. Figure 7 (B) Spherical EVs can be clearly observed adhering to the outer surface of the microspheres. It should be noted that, since the particle size of HMFMs is on the order of tens of micrometers, which is much larger than that of nano-sized EVs, and the electron beam has difficulty penetrating the interior of the microspheres, the image can only show the local structure of the edge region of HMFMs, but it still directly confirms the specific binding of EVs on the surface of the functionalized material.

[0085] 3) Four membrane proteins (CD63, HER-2, MUC1, and PD-L1) on the surface of five CM-EVs were characterized using nanoflow cytometry. The results are as follows: Figure 8 As shown in the results, the expression of the four membrane proteins on the surface of the five CM-EVs varied to varying degrees. This is the biological basis for our selection of these four membrane proteins as target recognition objects and the construction of sensor arrays to achieve phenotypic analysis. Specifically, except for MDA-MB-231, CD63 expression was the highest among the other proteins, while PD-L1 expression was the lowest. Meanwhile, HER2, MUC1, and PD-L1, which are associated with cancer, were detected in multiple cell lines, including normal and cancerous cells, and their expression abundance showed significant cell type-dependent changes, reflecting inherent biological heterogeneity. Further analysis showed that the expression of HER-2, MUC1, and PD-L1 in pancreatic ductal carcinoma cells PANC-1 was higher than that in normal pancreatic cells HPNE, indicating that the upregulation of these proteins may have the potential to indicate pancreatic cancer risk. In breast cell lines, although HER-2, MUC1, and PD-L1 did not show a consistent upregulation trend in the two breast cancer cell lines MCF-7 and MDA-MB-231 compared to normal MCF-10A, significant expression differences still existed, laying a biological foundation for distinguishing these three breast cell origins. These results confirm the existence of these four target membrane proteins in the samples, and their differential expression profiles provide direct theoretical basis and data support for subsequent construction of multi-parameter features and the use of pattern recognition algorithms for phenotyping. They also validate the rationality and necessity of the intelligent phenotyping strategy combining aptamer arrays and machine learning in this study from the perspective of membrane proteins.

[0086] Example 6: Validation and Clinical Application of the EV Phenotypic Analysis System

[0087] 1) To evaluate the performance of the phenotypic analysis system established in this study for the quantitative analysis of EV membrane proteins, 50 μL of the same source at different concentrations (gradient concentration: 1×10⁻⁶) were respectively tested. 3 1×10 4 1×10 5 1×10 6 1×10 7 A suspension of CM-EVs (particles / mL) was added to the system constructed in Example 5, and the C... EVs -A 652 nm Linear fitting was performed. To examine the system's ability to classify EVs from different cell lines, 50 μL of equal concentrations (1×10⁻⁶) were taken. 7 1×10 6 1×10 5 CM-EV suspensions from different sources (particles / mL) were analyzed to obtain A4-channel data. 652 nm Dimensionality reduction and classification are performed using LDA.

[0088] 2) Clinical validation

[0089] All plasma samples (10 healthy controls, 5 patients with early-stage breast cancer, and 5 patients with early-stage pancreatic cancer) and their corresponding clinical data were obtained with the approval of the Ethics Committee of Xiangyang Central Hospital (Ethics Approval No.: 2025-188). The collected EDTA-anticoagulated blood samples were centrifuged at 3000 ×g for 10 min at 4℃, and the supernatant was centrifuged again to obtain plasma samples. These samples were then processed and tested using the two methods described below.

[0090] ① To evaluate the suitability of the phenotypic analysis system, plasma samples were diluted with PBS at a ratio of 1:1000 (v / v), and 50 μL of the diluted sample was analyzed according to the procedure described in Example 5. Based on the A... 652 nm Heatmaps were generated and principal component analysis (PCA) was performed to differentiate between breast cancer and pancreatic cancer patients and healthy controls. Furthermore, clinical tumor marker (CEA, CA15-3, and CA19-9) detection data were collected from these patients, receiver operating characteristic (ROC) curves were plotted, and the diagnostic performance of the sensor developed in this study was evaluated by comparing the area under the curve (AUC) with that of traditional tumor markers.

[0091] ② To investigate the reason why the overall expression of membrane proteins in plasma EVs from cancer patients was higher than that in healthy controls, 0.5 mL of pretreated plasma samples were centrifuged sequentially at 10000 ×g / 30 min and 120000 ×g / 2 h. The precipitate was resuspended in 100 μL PBS and washed once. The obtained EVs were analyzed for particle concentration using NTA and uniformly diluted to 1 × 10⁻⁶ ppm with PBS.7 Particles / mL, then 50 μL was taken and analyzed according to the procedure described in Example 5, based on A in four channels. 652 nm And draw a heat map.

[0092] result: 1) Five different cell-derived CM-EVs were measured and compared with four sensing elements (Apt). CD63 Apt HER2 Apt MUC1 Apt PD-L1 The linear relationship between the colorimetric signal generated during the interaction and its EV concentration is shown in the following figure. Figure 9 As shown. Overall, this sensor is effective for individual membrane proteins at EV concentrations of 1 × 10⁻⁶. 3 Up to 1×10 7 It exhibits good linear response (R0) across the particle / mL range. 2 > 0.99), the linear equations of each group are shown in Table 1. They have a wide linear range and high sensitivity, which can meet the detection requirements of EVs in routine biological fluids such as plasma or urine.

[0093] Table 1:

[0094] Through calculation, material Apt CD63 - The capture efficiency of HMFMs for EVs is approximately 2.5 × 10⁻⁶. 9 Compared to existing technologies, this method achieves a particle / mg ratio with significant advantages. At the nanoscale, we utilize FeCDzymes with excellent POD-like activity as signal output units. Each FeCDzyme can bind to multiple FeCDzymes through multiple aptamers, achieving signal amplification from "one EVs to multiple FeCDzymes." Compared to the fluorescence signals used in previous work, this enzyme-catalyzed colorimetric signal output strategy maintains similar sensitivity while further reducing the requirements and cost of detection instruments. It should be noted that the upper limit of the linear range is related to Apt... CD63 The amount of material used in HMFMs is closely related. In practical applications, if it is necessary to expand the detection range to a higher concentration, it can be achieved by appropriately increasing the amount of material or by moderately diluting the sample.

[0095] 2) Under high concentration conditions (1×10⁻⁶) 7We tested a four-protein sensing combination consisting of MUC1, CD63, HER2, and PD-L1. The results showed that this system could clearly separate eight different EVs (adding three tumor cell-derived EVs—HCT-15, A549, and HGC-27—to the original five EVs) in linear discriminant analysis (LDA) space, demonstrating excellent class discrimination ability. Figure 10 A). Even after removing MUC1 and retaining only the three markers CD63, HER2, and PD-L1, the system can still effectively classify eight EVs. Figure 10 B). This indicates that at high concentrations, the general capture signal provided by CD63, together with the supplementary information from HER2 and PD-L1, constitutes sufficient classification dimensions, reducing the system's dependence on MUC1 and demonstrating a certain degree of stability. To better understand the weight of each biomarker, we further conducted an LDA contribution analysis ( Figure 10 C). The first two discriminant functions explained a total of 84.6% of the variance. LD1 contributed 48.5% of the variance, with the dominant factor being the universal marker CD63 (absolute coefficient 53.70), significantly higher than other markers. This aligns with the widespread expression of CD63 in EVs, providing a fundamental dimension for classification. In LD2 (contributing 36.1% of the variance), the contribution distribution of markers changed significantly: the contribution of MUC1 increased dramatically (absolute coefficient 108.28), becoming the dominant marker. As a highly glycosylated and tissue-specific transmembrane mucin, MUC1 exhibits significant differences in expression and modification states on the surface of different cell sources (especially different tumor subtypes), making it crucial for finely distinguishing cell identity in LD2. The shift from "CD63-dominated" to "MUC1-dominated" reveals the functional differentiation and synergistic relationship of markers during classification: CD63 lays the foundation for basic subtyping dimensions across cell types, while MUC1 provides high-resolution subtype identification information.

[0096] 3) As the concentration decreases to 1×10 6 With particles / mL, the system's classification ability was tested, and the weighting of each marker was adjusted accordingly. The four-protein combination could still effectively separate five particles (…). Figure 10 D), and after removing MUC1, the overall discrimination ability of the system is significantly weakened ( Figure 10E). Not only did the distance between each category in the LDA plot generally decrease, but more importantly, significant overlap occurred between two breast cancer cell lines with different biological characteristics (MCF-7 and MDA-MB-231), making them difficult to distinguish. This result directly reveals the crucial discriminative role of MUC1 under low signal conditions: when the concentration decreases, the common signal differences provided by universal markers such as CD63 weaken, and the subtle differences provided by specific markers like MUC1 become the core basis for distinguishing cell subtypes. The concentration was further reduced to 1×10⁻⁶. 5 particles / mL ( Figure 10 (F) The system's classification ability is approaching its limit. At this point, there is still a certain distance between normal cells (MCF-10A) and cancer cell populations, and they can still be distinguished; however, the boundaries between different cancer subtypes (such as MCF-7 and MDA-MB-231) are very blurred and overlap; other categories (such as PANC-1 and HPNE) can only be barely separated. This indicates that at extremely low concentrations, although markers such as CD63 can still maintain basic capture and macroscopic classification functions, helping to distinguish between normal and cancerous states, the weak signals supported by specific markers such as MUC1, which are used for fine subtype differentiation, are close to the lower limit of system detection, resulting in a significant decrease in discrimination ability.

[0097] 4) Clinical sample test results, such as Figure 11 As shown.

[0098] ① The expression heatmap after Z-score normalization is as follows Figure 11 As shown in A (pancreatic cancer group) and 11B (breast cancer group), under the same sampling volume (50 μL, diluted only without ultracentrifugation pre-enrichment), the overall membrane protein signal of plasma EVs in the cancer patient group was higher than that in the healthy control group. This trend, while not entirely consistent with the expression pattern in the aforementioned cell line-derived CM-EVs, is as expected: plasma EVs are widely sourced and complex in composition, while CM-EVs originate from a single cell line, and the difference in their membrane protein expression profiles reflects the overall changes in the composition and molecules of EVs in cancer patients. These results preliminarily indicate that this system can distinguish between cancer patients and healthy individuals with only a small amount of plasma, and does not rely on ultracentrifugation, demonstrating good clinical operability.

[0099] ② Principal component analysis (PCA) based on the original four-dimensional membrane protein expression data further confirmed the classification ability of this system: such as Figure 11 As shown in C and 11D, patients with early-stage breast or pancreatic cancer and healthy controls showed significant separation in the PCA space, indicating that the membrane protein features extracted by this system have the potential for early cancer identification. For comparison, we also simultaneously analyzed the levels of commonly used clinical tumor markers in each group of samples (…). Figure 11 E, Figure 11F). In early-stage breast cancer patients, the positive rate of CA15-3 was only 20% (1 / 5), and in early-stage pancreatic cancer patients, the positive rate of CA19-9 was also only 20%, while CEA did not exceed the clinical cutoff value in all samples. These results indicate that traditional single-protein biomarkers have limited sensitivity in early cancer detection. Further evaluation of diagnostic performance was conducted using receiver operating characteristic (ROC) curves: such as... Figure 11 As shown in G (breast cancer) and 11H (pancreatic cancer), the area under the curve of the diagnostic model constructed based on the four-dimensional EVs membrane protein characteristics of this system is significantly higher than the AUC when CEA, CA15-3 or CA19-9 are used alone.

[0100] ③ Equal volumes of plasma from cancer patients and healthy individuals were processed by ultracentrifugation. The concentration of extracellular energy (EVs) was measured using NTA and standardized to the same level before testing. Results showed that the concentration of EVs in the plasma of cancer patients was generally higher than that in healthy individuals. Figure 12 A), and at the same EV concentration, its membrane protein expression was still significantly higher ( Figure 12 B. Figure 12 C). This illustrates... Figure 11 The enhanced abundance signal of plasma EV membrane proteins observed in cancer patients in A and 11B may be the result of the combined effect of increased EV concentration and upregulation of surface membrane protein expression of individual EVs.

[0101] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A honeycomb-shaped melamine-formaldehyde resin microsphere, characterized in that, The product is obtained by polymerization of melamine, formaldehyde, and an acidic catalyst under conditions where macroporous microspheres serve as templates, followed by etching to remove the templates; the macroporous microspheres have a pore size range of 0.5-1 μm.

2. The honeycomb-shaped melamine-formaldehyde resin microspheres according to claim 1, characterized in that, The honeycomb-shaped melamine-formaldehyde resin microspheres have a pore size of 0.5-1 μm; And / or, the macroporous microspheres are silica microspheres, and the etching process for removing the template uses an alkaline solution.

3. The honeycomb-shaped melamine-formaldehyde resin microspheres according to claim 1, characterized in that, The molar ratio of melamine to formaldehyde is 1:(5-20). And / or, the acidic catalyst is an inorganic acid; And / or, the polymerization reaction temperature is 30-80℃, and the reaction time is 0.5-15 h; And / or, the polymerization reaction is further complicated by the addition of polyethylene glycol, in an amount of 0.5-1% of the molar amount of melamine.

4. An extracellular vesicle-capturing microsphere, characterized in that, The capturing microspheres are obtained by modifying CD63 nucleic acid aptamers with honeycomb melamine-formaldehyde resin microspheres as described in any one of claims 1-3.

5. The extracellular vesicle capturing microspheres according to claim 4, characterized in that, The nucleotide sequence of the CD63 aptamer is shown in SEQ ID NO: 1; And / or, the CD63 nucleic acid aptamer is modified with a carboxyl group, and the process of modifying the CD63 nucleic acid aptamer with the honeycomb melamine-formaldehyde resin microspheres is to incubate the CD63 nucleic acid aptamer and the honeycomb melamine-formaldehyde resin microspheres at room temperature to undergo a condensation reaction.

6. The use of the extracellular vesicle capturing microspheres of claim 4 or 5 in the preparation of extracellular vesicle capturing or detection products.

7. An extracellular vesicle typing detection kit, characterized in that, It includes the extracellular vesicle capturing microspheres as described in claim 4 or 5, and a detection probe targeting a target protein, including MUC1, HER2, PD-L1, and CD63.

8. The genotyping kit according to claim 7, characterized in that, Any of the detection probes is a nucleic acid aptamer corresponding to the target protein; And / or, the detection probe is modified with carbon dot nanozymes or their metal dopants; And / or, the kit may also include TMB chromogenic solution, peroxide, and buffer.

9. A method for detecting extracellular vesicle typing, for non-diagnostic purposes, characterized in that, include: S1. Using the capturing microspheres according to claim 4 or 5 to capture extracellular vesicles in a sample; S2. Add detection probes and co-incubate with captured extracellular vesicles; the detection probes include probes targeting MUC1, HER2, PD-L1 and CD63 proteins, and any of the probes includes a nucleic acid aptamer of the target protein and is modified with iron-doped carbon dot nanozymes; S3. Add colorimetric reagent, obtain colorimetric signals from different channels, and calculate the typing results.

10. The typing detection method according to claim 9, characterized in that, The sample was a plasma sample; And / or, the acquired four-channel signals are constructed into a feature array and used for pattern detection through a linear discriminant analysis algorithm.

Images