Surface plasmon resonance-based four-transmembrane protein post-delivery activity evaluation system

Abstract

The application provides a four-transmembrane protein delivery post-activity evaluation system based on surface plasmon resonance. It comprises: pretreatment of a sample to be detected for delivery of a four-transmembrane protein by a chimeric antigen receptor engineered immune cell source modified extracellular vesicle, and fixation of the sample to be detected on a detection chip; in the detection chip, a flow release medium is introduced through microfluidic technology, and signal changes before, during and after release are recorded respectively; based on the principle of surface plasmon resonance, the actual expression amount of the four-transmembrane protein on the cell membrane is detected according to the light signal displacement caused by the change of the interface refractive index; the light signal curve of the whole process of the binding stage, the equilibrium stage and the dissociation stage is obtained, and the detection result is generated; the detection result is compared with the corresponding reference data, the time sequence corresponding relationship between the drug release rate and the light signal is calculated combined with the signal curve, the targeted binding capacity and the overall delivery activity of the targeted delivery system are quantitatively obtained, and the evaluation is completed.

Description

Technical Field

[0002] This application relates to the field of biomedical technology, and in particular to a four-transmembrane protein post-delivery activity evaluation system based on surface plasmon resonance. Background Technology

[0004] The tetraspan protein family is highly specifically overexpressed on the surface of various tumor cells. It can drive tumor progression by activating downstream signaling pathways, promoting tumor metastasis, angiogenesis, and immune escape, making it a highly promising target for anti-tumor drugs. In recent years, chimeric antigen receptor-engineered extracellular vesicle delivery technology derived from immune cells has emerged as a novel anti-tumor delivery strategy targeting tetraspan proteins due to its low immunogenicity and ability to prevent immune clearance. However, this delivery process involves complex biochemical reactions. Current technologies can only predict druggability through logical deduction or indirectly infer delivery effects through in vitro cell proliferation experiments and in vivo animal pharmacodynamic experiments. Tools for directly quantifying the true expression of target proteins and delivery kinetics after delivery are lacking. Furthermore, traditional protein quantification methods require sample pretreatment such as lysis and staining, which damages the native structure of the cell membrane and fails to obtain true expression data of tetraspan proteins on the cell membrane surface, making it difficult to meet the needs of delivery activity assessment. Summary of the Invention

[0005] This application provides a four-transmembrane protein delivery activity evaluation system based on surface plasmon resonance, which aims to solve the problem that traditional protein quantification methods require sample lysis, staining and other pretreatments, which will destroy the original structure of the cell membrane and make it impossible to obtain the true expression data of four-transmembrane proteins on the cell membrane surface, thus failing to meet the needs of delivery activity evaluation.

[0006] In a first aspect, embodiments of this application provide a method for evaluating the activity of a four-transmembrane protein after delivery based on surface plasmon resonance, the method comprising: The sample to be tested, which is a modified extracellular vesicle delivering a tetraspan membrane protein via chimeric antigen receptor-engineered immune cell, is pretreated and then immobilized on the detection chip. In the detection chip, a flowing release medium is introduced through microfluidic technology, and the signal changes before, during and after release are recorded respectively. Based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change in interface refractive index. The optical signal curves of the entire process of binding, equilibrium and dissociation are obtained to generate detection results. The detection results are compared with the corresponding benchmark data, and the temporal correspondence between drug release rate and optical signal is calculated by combining the signal curves. The targeted binding ability and overall delivery activity of the targeted delivery system are obtained by comprehensive quantification to complete the evaluation.

[0007] In some embodiments, the pretreatment of the sample to be tested, which is a modified extracellular vesicle from chimeric antigen receptor-engineered immune cells that delivers a tetraspan membrane protein, includes: removing tissue debris and unbound free proteins from the sample, and sealing the cell membrane sample to prevent non-specific protein binding from interfering with subsequent detection signals.

[0008] In some embodiments, fixing the sample to be detected onto the detection chip includes: uniformly introducing a pretreated cell membrane suspension containing tetraspan membrane proteins into the detection reaction channel of the microfluidic chip, allowing it to stand and incubate to allow the cell membrane to bind and fix to the modified sensing surface of the chip, and discharging unfixed free suspension residue.

[0009] In some embodiments, the introduction of a flow release medium into the detection chip via microfluidic technology includes: controlling a pumping device according to preset parameters to introduce a flow release medium containing an eluent into the detection reaction channel of the microfluidic chip at a constant flow rate, maintaining a stable flow rate and pressure throughout the entire introduction process.

[0010] In some embodiments, recording the signal changes before, during, and after release includes: acquiring the initial light signal of the blank sensing interface as a baseline before release; continuously acquiring resonant light signal intensity data at fixed time intervals during release; continuously acquiring light signal data of the dissociation process after release; and storing all acquired data in chronological order.

[0011] In some embodiments, the method of detecting the actual expression level of a four-span membrane protein on the cell membrane based on the surface plasmon resonance principle and the optical signal displacement caused by the change in interface refractive index includes: identifying the displacement of the resonance peak, calculating the total content of bound four-span membrane proteins based on the correspondence between the change in refractive index and the amount of protein binding, and converting the actual expression level of the four-span membrane protein per unit cell membrane area.

[0012] In some embodiments, acquiring the optical signal curves of the entire process of the binding phase, equilibrium phase, and dissociation phase, and generating detection results, includes: extracting the collected time-series signal data, identifying the binding phase where the signal rises rapidly, the equilibrium phase where the signal remains stable, and the dissociation phase where the signal gradually declines, respectively, fitting the complete response signal curves, and extracting and organizing the characteristic parameters of each phase to generate standardized detection results.

[0013] In some embodiments, the comparison of the detection results with the corresponding benchmark data, the calculation of the temporal correspondence between the drug release rate and the light signal based on the signal curve, and the comprehensive quantitative determination of the targeting binding capacity and overall delivery activity of the targeted delivery system to complete the evaluation include: extracting the expression levels of four transmembrane proteins and signal characteristics at each stage before delivery as benchmark data, calculating the feature differences between the post-delivery detection data and the benchmark data, fitting the signal curve to obtain the drug release rate, inputting the targeting binding capacity and release kinetic parameters into the evaluation model, and outputting the quantitative evaluation results of the overall delivery activity to complete the evaluation.

[0014] In some embodiments, the method further includes: pre-collecting optical signal curves and actual drug efficacy verification results corresponding to multiple sets of samples with different delivery efficiencies, training a deep learning evaluation model, inputting the optical signal curves of the entire process obtained in this detection into the trained model, having the model automatically extract the implicit features of each stage, predicting the expected in vivo anti-tumor effect of the delivery system to be evaluated, and then combining it with a database of differential expression of four-span membrane proteins in different tumors to output a ranking result of drug potential adapted to different tumor types.

[0015] Secondly, this application provides a four-transmembrane protein delivery activity evaluation system based on surface plasmon resonance, the system comprising: The sample processing unit is used to preprocess the sample to be tested, which is a transmembrane protein delivered by extracellular vesicles modified from chimeric antigen receptor-engineered immune cells, and to fix the sample to be tested onto the detection chip. The medium release unit is used to introduce a flowing release medium into the detection chip via microfluidic technology and record the signal changes before, during, and after release. Based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change in interface refractive index. The evaluation completion unit is used to acquire the optical signal curves of the entire process of binding, equilibrium and dissociation, and generate detection results; the detection results are compared with the corresponding benchmark data, and the temporal correspondence between drug release rate and optical signal is calculated by combining the signal curves, so as to comprehensively and quantitatively obtain the targeting binding ability and overall delivery activity of the targeted delivery system, and complete the evaluation.

[0016] This application combines the principle of surface plasmon resonance with microfluidic technology to detect the true expression level of tetraspan proteins on the cell membrane in real time without damaging the original structure of the cell membrane. It can also obtain dynamic signals of the three stages of binding, equilibrium, and dissociation throughout the delivery process. By comparing with the baseline data before delivery, the targeting binding ability and overall delivery activity of the targeted delivery system can be quantitatively obtained. This solves the problem of the lack of quantitative assessment methods for the delivery activity of tetraspan proteins in existing technologies. The assessment results are highly accurate and can provide a reliable quantitative basis for the development of anti-tumor drugs targeting tetraspan proteins, effectively improving the efficiency of drug screening.

[0017] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart illustrating the steps of a method for evaluating the activity of a four-transmembrane protein after delivery based on surface plasmon resonance, provided in an embodiment of this application. Figure 2 This is a schematic diagram illustrating the principle of a method for evaluating the activity of a four-transmembrane protein after delivery based on surface plasmon resonance, provided in an embodiment of this application. Figure 3 This is a schematic block diagram of a four-transmembrane protein delivery activity evaluation system based on surface plasmon resonance, provided in one embodiment of this application. Figure 4 This is a schematic block diagram of the structure of a computer device provided in an embodiment of this application.

[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0023] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0024] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0025] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0026] The tetraspan protein family is highly specifically overexpressed on the surface of various tumor cells. It can drive tumor progression by activating downstream signaling pathways, promoting tumor metastasis, angiogenesis, and immune escape, making it a highly promising target for anti-tumor drugs. In recent years, chimeric antigen receptor-engineered extracellular vesicle delivery technology derived from immune cells has emerged as a novel anti-tumor delivery strategy targeting tetraspan proteins due to its low immunogenicity and ability to prevent immune clearance. However, this delivery process involves complex biochemical reactions. Current technologies can only predict druggability through logical deduction or indirectly infer delivery effects through in vitro cell proliferation experiments and in vivo animal pharmacodynamic experiments. Tools for directly quantifying the true expression of target proteins and delivery kinetics after delivery are lacking. Furthermore, traditional protein quantification methods require sample pretreatment such as lysis and staining, which damages the native structure of the cell membrane and fails to obtain true expression data of tetraspan proteins on the cell membrane surface, making it difficult to meet the needs of delivery activity assessment.

[0027] To solve the above problem, please refer to Figure 1 and Figure 2This application provides a method for evaluating the post-delivery activity of four-transmembrane proteins based on surface plasmon resonance, applicable to computer equipment. The computer equipment can be deployed on a single server or server cluster. It can also be deployed on handheld terminals, laptops, wearable devices, or robots, etc. It should be noted that all information involved in the method provided in this application has been extracted with the authorization of the relevant users and in accordance with relevant regulations, and will not infringe on user privacy.

[0028] The provided method for evaluating the post-delivery activity of four-transmembrane proteins based on surface plasmon resonance includes steps S101 to S103. Details are as follows: Step S101. Preprocess the sample to be tested, which is a modified extracellular vesicle delivering a tetraspan membrane protein via chimeric antigen receptor-engineered immune cell, and fix the sample to be tested onto the detection chip.

[0029] Specifically, this step is the foundation of the entire detection process, and it addresses two key issues: First, it eliminates matrix interference through sample pretreatment to avoid non-specific binding interfering with subsequent SPR optical signal detection and improves the detection signal-to-noise ratio. Second, without damaging the original cell membrane structure, it in situ and stably fixes the delivered intact cell membrane sample on the sensing surface of the SPR detection chip, constructing a solid-phase reaction interface that is highly consistent with the physiological state. This fundamentally avoids the shortcomings of traditional methods that lyse and stain samples, thereby damaging the cell membrane structure and failing to obtain true expression data of four-span membrane proteins on the membrane surface.

[0030] The sample source was confirmed to be a biological sample that has undergone four-transmembrane protein delivery via a targeted delivery system, including but not limited to tumor cell membrane suspensions, primary cell samples, cell mesangial extracts, etc. After collection, the samples were stored in a low-temperature environment of 4°C throughout the process to avoid cell membrane structure rupture and target protein denaturation.

[0031] Sample pretreatment involves first purifying the sample to remove impurities such as tissue fragments, unbound free proteins, and untaken free extracellular vesicles; then, the purified cell membrane sample is sealed to block non-specific binding sites on the membrane surface and at the subsequent chip interface, thereby reducing background interference.

[0032] Chip immobilization involves introducing a pretreated sample suspension containing intact cell membrane structures into the microfluidic detection reaction channel of the SPR chip. Through isothermal incubation, stable covalent binding between the cell membrane and the chip-modified sensing surface is achieved. After incubation, the channel is rinsed with buffer solution to completely remove unfixed free residues and unbound samples, thus completing the chip immobilization of the detection sample and constructing a stable in-situ detection interface for subsequent dynamic detection.

[0033] Step S102. In the detection chip, a flow release medium is introduced through microfluidic technology, and the signal changes before, during and after release are recorded respectively; based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change of interface refractive index.

[0034] Specifically, this step is the core detection step of the method, based on the principle of surface plasmon resonance optics. The refractive index change at the metal sensing interface corresponds linearly to the mass of biomolecules bound at the interface. Target protein binding triggers a change in the interface refractive index, leading to a shift in the resonance light signal. By detecting this shift, the protein content can be accurately quantified. This step combines microfluidic dynamic flow control technology to achieve label-free, in-situ, real-time quantitative detection of four-span membrane proteins on the cell membrane surface. Simultaneously, it comprehensively captures signal changes throughout the entire drug release cycle of the delivery system, addressing the limitations of existing technologies that cannot dynamically track delivery kinetics or quantitatively detect the true expression level of native target proteins on the membrane surface.

[0035] The dynamic fluid environment is constructed by introducing a flow release medium into the chip detection reaction channel at a preset constant flow rate and stable pressure after the sample chip is fixed through a microfluidic pumping system. The medium is a buffer solution that matches the detection system. Eluent can be added as needed to simulate the physiological environment of drug release in vivo. The flow rate and pressure are kept stable throughout the process to avoid signal drift caused by fluid fluctuations.

[0036] Full-cycle signal acquisition is completed in three stages, maintaining a constant temperature environment and consistent acquisition parameters throughout the process. Before release: Collect the initial resonant light signals of the blank sensing interface and the sensing interface after fixing the sample, take the average value of the stable segment as the detection baseline, and subtract system error and background interference; During release: Simultaneously with the introduction of the flowing release medium, time-series acquisition is initiated, and resonance light signal intensity and resonance peak displacement data are continuously and in real time acquired at fixed time intervals to fully record the dynamic signal changes of drug release and target protein binding; After release: switch to blank buffer solution and continuously collect optical signal data during the dissociation phase to fully record the dissociation dynamics of the target protein; All collected data are aligned in timestamp order, denoised, baseline corrected, and then encrypted before storage.

[0037] The quantitative calculation of target protein expression is based on the effective resonance peak displacement and a pre-established standard curve of "refractive index change - protein binding amount". The total content of the four-span membrane protein bound on the sensing interface is calculated. Then, the actual in-situ expression level of the four-span membrane protein per unit cell membrane area is obtained by converting the effective reaction area of ​​the chip and the cell membrane spreading area. This completes the non-destructive quantitative detection of target protein under the native structure of the membrane surface.

[0038] Step S103. Obtain the optical signal curves of the entire process of binding, equilibrium and dissociation phases, and generate detection results; compare the detection results with the corresponding benchmark data, calculate the temporal correspondence between drug release rate and optical signal based on the signal curves, and comprehensively quantify the targeting binding ability and overall delivery activity of the targeted delivery system to complete the evaluation.

[0039] Specifically, this step is the final output of the method. The core is based on the full-cycle SPR dynamic signal curve, extracting multi-dimensional feature parameters, and combining them with benchmark control data to complete the quantitative evaluation of the targeting binding ability, release kinetics, and overall delivery activity of the targeted delivery system. This solves the problem that existing technologies can only indirectly infer the delivery effect through in vitro / in vivo experiments and cannot directly quantify the delivery activity, thus realizing a closed loop from target protein expression quantification to drug-likeness evaluation of the delivery system.

[0040] The dynamic curve fitting and standardization results generation process involves processing the full-time optical signal data acquired by S102, identifying and dividing signal segments into binding, equilibrium, and dissociation stages, and using a standard dynamic model to fit and generate a complete SPR reaction dynamic curve. Characteristic parameters of each stage (including binding rate constant, dissociation rate constant, equilibrium dissociation constant, maximum binding signal value, etc.) are extracted to generate standardized and traceable detection results.

[0041] The benchmark data comparison involves retrieving corresponding benchmark data, including the basic expression level of four-transmembrane proteins in blank cell samples from the same batch that have not been delivered, the non-specific binding signal of the blank delivery system, and the kinetic characteristic parameters of the standard. The detection results are then paired and compared with the benchmark data to calculate core indicators such as the difference in target protein expression levels before and after delivery, the specific binding signal-to-noise ratio, and the deviation of kinetic parameters.

[0042] The comprehensive evaluation of delivery activity combines the temporal changes of signal curves to fit and calculate the temporal correspondence between drug release rate and light signal changes, thereby clarifying the release kinetic characteristics of the delivery system. The target binding ability-related parameters, release kinetic parameters, and target protein expression upregulation factor are input into a preset quantitative evaluation model. After weighted calculation, the overall delivery activity quantitative score is output, and the evaluation results of sub-indicators are also output, thus completing the full-dimensional activity evaluation of the targeted delivery system.

[0043] In some embodiments, the pretreatment of the sample to be tested, which is a modified extracellular vesicle from chimeric antigen receptor-engineered immune cells that delivers a tetraspan membrane protein, includes: removing tissue debris and unbound free proteins from the sample, and sealing the cell membrane sample to prevent non-specific protein binding from interfering with subsequent detection signals.

[0044] This embodiment is a refinement and optimization of the sample pretreatment step in S101. The core purpose is to eliminate interference factors in the sample matrix to the greatest extent through precise sample purification and sealing treatment, avoid interference from non-specific protein binding on the subsequent SPR detection signal, improve the signal-to-noise ratio and quantitative accuracy of the detection, and solve the problem of distorted detection results caused by many impurities and high background signals in biological samples.

[0045] Sample pretreatment involves obtaining tumor cell membrane samples after the delivery of four-transmembrane proteins via a targeted delivery system. The samples are kept at a low temperature of 4°C throughout the collection process to prevent cell membrane rupture and protein denaturation.

[0046] Gradient centrifugation purification: The sample was purified by differential centrifugation. First, the sample was centrifuged at a low speed of 300×g for 5 min to remove tissue fragments and intact cell debris. Then, it was centrifuged at a high speed of 10000×g for 10 min to remove soluble impurities such as free extracellular vesicles and unbound free proteins that were not taken up by cells. After centrifugation, the supernatant was discarded, and the cell membrane precipitate was resuspended in sterile PBS buffer at pH 7.4. The centrifugation and washing were repeated twice to obtain the purified intact cell membrane suspension.

[0047] Non-specific site blocking was achieved by adding bovine serum albumin (BSA) blocking solution to the purified cell membrane suspension at a final concentration of 1% (w / v) and gently incubating at a constant temperature of 37°C for 30 min to block non-specific protein binding sites on the cell membrane surface. After incubation, the sample was centrifuged again at 10000×g for 10 min, the supernatant was discarded, and the sample was resuspended in PBS buffer to obtain the pretreated cell membrane sample to be tested. The entire process, except for the blocking incubation, was carried out at a low temperature of 4°C to avoid degradation of the cell membrane structure.

[0048] In some embodiments, fixing the sample to be detected onto the detection chip includes: uniformly introducing a pretreated cell membrane suspension containing tetraspan membrane proteins into the detection reaction channel of the microfluidic chip, allowing it to stand and incubate to allow the cell membrane to bind and fix to the modified sensing surface of the chip, and discharging unfixed free suspension residue.

[0049] This embodiment is a detailed implementation of the sample chip fixation step in S101. The core objective is to achieve uniform and stable fixation of the intact cell membrane on the SPR chip sensing surface without damaging the original structure of the cell membrane, thereby constructing an in-situ detection interface that is highly consistent with the physiological state. This avoids dynamic detection signal drift caused by uneven fixation and unstable binding, ensuring the accuracy and repeatability of subsequent SPR quantitative detection.

[0050] Chip activation was achieved by using an SPR gold film sensor chip pre-modified with carboxymethyl dextran (CM5) as the detection chip. The chip sensing surface was first activated by introducing a 1:1 mixture of 0.4 mol / L 1-ethyl-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and 0.1 mol / L N-hydroxysuccinimide (NHS) into the microfluidic detection reaction channel of the chip at a flow rate of 20 μL / min for 7 min. This activated the carboxyl groups on the chip sensing surface, providing active sites for cell membrane coupling.

[0051] Sample introduction and incubation fixation: The pretreated cell membrane suspension was diluted with sodium acetate buffer at pH 5.5 to a final protein concentration of 50 μg / mL, and then introduced into the detection reaction channel of the chip at a constant flow rate of 10 μL / min using a microfluidic pump. After the introduction was completed, the pumping was stopped, and the sample was incubated at a constant temperature of 25°C for 60 min to allow the intact cell membrane to be stably bound and fixed to the activated sensing surface of the chip through amino coupling.

[0052] After incubation, PBS buffer at pH 7.4 was introduced at a flow rate of 30 μL / min and rinsed for 5 min to thoroughly remove unfixed free cell membrane residue and unbound samples from the channels. After rinsing, the initial light signal was collected to verify signal stability. The fixed chip was placed in a constant temperature environment at 25°C, keeping the channels filled with PBS buffer to prevent the sensing interface from drying out, and was immediately used for subsequent detection.

[0053] In some embodiments, the introduction of a flow release medium into the detection chip via microfluidic technology includes: controlling a pumping device according to preset parameters to introduce a flow release medium containing an eluent into the detection reaction channel of the microfluidic chip at a constant flow rate, maintaining a stable flow rate and pressure throughout the entire introduction process.

[0054] This embodiment is a detailed implementation of the flow release medium introduction step in S102. The core purpose is to construct a stable, uniform, and repeatable dynamic fluid environment through precise microfluidic parameter control, to simulate the physiological conditions of drug release in the in vivo delivery system, and at the same time ensure the stability of the fluid environment during SPR optical signal acquisition, avoid signal drift caused by flow rate and pressure fluctuations, and ensure the accuracy of release kinetic data.

[0055] After the sample chip is fixed, the SPR detection system, microfluidic pumping device, and constant temperature control module are calibrated together. The system constant temperature environment is set at 37℃ to simulate human physiological temperature, and the temperature fluctuation is kept less than ±0.2℃ throughout the process.

[0056] The fluid release medium was prepared using a PBS buffer containing 0.05% Tween-20 at pH 7.4 as the base medium. Depending on the release characteristics of the delivery system to be evaluated, the appropriate concentration of eluent was added as needed (an acidic buffer or competitive binding reagent that matches the delivery system can be selected). After the medium was prepared, it was filtered through a 0.22 μm filter membrane to degas it, so as to avoid air bubbles entering the microfluidic channel and causing signal interference.

[0057] The constant flow control method controls the pumping device according to preset flow control parameters to introduce the configured flow release medium into the detection reaction channel of the microfluidic chip at a constant flow rate of 50 μL / min. The system's pressure feedback module monitors the pressure in the channel in real time, and the pumping parameters are adjusted in real time through a PID algorithm to keep the flow rate fluctuation less than ±2% and the pressure fluctuation in the channel less than ±5% throughout the entire introduction process, ensuring the stability of the fluid environment. Throughout the entire detection cycle, the constant flow of the medium is maintained until the detection of the binding, equilibrium, and dissociation stages is completed.

[0058] In some embodiments, recording the signal changes before, during, and after release includes: acquiring the initial light signal of the blank sensing interface as a baseline before release; continuously acquiring resonant light signal intensity data at fixed time intervals during release; continuously acquiring light signal data of the dissociation process after release; and storing all acquired data in chronological order.

[0059] This embodiment is a detailed implementation of the signal recording steps before, during, and after release in step S102. The core objective is to capture the changes in SPR optical signals throughout the entire drug release cycle of the delivery system through phased, high-temporal-resolution time-series signal acquisition, and to record the kinetic characteristics of the entire binding and dissociation process without omission. This provides complete and high-fidelity raw data support for subsequent target protein quantification and activity assessment, and solves the pain point that existing technologies cannot dynamically track changes in delivery time.

[0060] Based on a stable microfluidic environment, and in conjunction with the optical acquisition module of the SPR detection system, the full-cycle signal acquisition is completed in three stages, maintaining a constant temperature of 37°C and a constant flow rate throughout the process. Before release, baseline acquisition was performed by first introducing blank PBS buffer into the channel before introducing the flow release medium, and continuously acquiring the initial resonant light signal of the blank sensing interface for 5 minutes. Then, the initial light signal of the sensing interface after sample fixation was acquired, and the average value of the signal stable segment was taken as the detection baseline for subsequent subtraction of background interference and system error.

[0061] During the release process, dynamic signal acquisition is initiated simultaneously with the introduction of the flowing release medium containing eluent. The acquisition frequency is 10 Hz (data is acquired once every 0.1 s), and the intensity of the resonance light signal and the wavelength shift of the resonance peak are continuously acquired. The acquisition duration is set to 30 min according to the release characteristics of the delivery system, and the dynamic signal changes of the entire process of drug release and binding of the four-span membrane protein to the sensing interface are fully recorded. The signal stability is monitored in real time during the acquisition process. If bubble interference occurs, it is immediately marked and supplemented for acquisition.

[0062] After release, dissociation signal acquisition was performed by immediately switching to blank PBS buffer without eluent after the release process was completed, maintaining the same flow rate for 60 minutes, and continuously acquiring optical signal data of the dissociation process at the same acquisition frequency of 10 Hz to fully record the dissociation dynamics of the tetraspan membrane protein.

[0063] Data storage involves aligning, denoising, and baseline-correcting all collected time-series data according to timestamps, then encrypting and storing it in the system database for subsequent quantitative analysis and evaluation.

[0064] In some embodiments, the method of detecting the actual expression level of a four-span membrane protein on the cell membrane based on the surface plasmon resonance principle and the optical signal displacement caused by the change in interface refractive index includes: identifying the displacement of the resonance peak, calculating the total content of bound four-span membrane proteins based on the correspondence between the change in refractive index and the amount of protein binding, and converting the actual expression level of the four-span membrane protein per unit cell membrane area.

[0065] This embodiment is a refined implementation of the target protein quantification step in S102. The core objective is to accurately calculate the actual in-situ expression level of the four-span membrane protein on the cell membrane surface based on the SPR resonance peak shift data and a pre-established standard curve. The entire process does not require destructive treatment of the sample such as lysis or staining, thus achieving non-destructive quantification of the true expression level of the four-span membrane protein under the native structure of the membrane surface, solving the core pain point that traditional methods cannot detect native expression data of membrane proteins.

[0066] Standard curve pre-establishment was achieved by using recombinant four-transmembrane protein standards at gradient concentrations. SPR detection was performed under identical chip fixation, microfluidic environment, and detection parameters to the sample being tested. Resonance peak shifts corresponding to different concentrations of standards were collected and fitted to generate a curve showing "resonance peak shift (unit: RU, resonance units) - protein binding amount (unit: ng / mm²)". 2 The standard curve is used to determine the linear detection range and detection limit, and the goodness of fit R of the standard curve is calculated. 2 It needs to be greater than 0.99.

[0067] Effective displacement extraction involves processing the full-cycle signal data collected in the sample, performing baseline correction and denoising, identifying and extracting the maximum displacement of the resonance peak and the stable displacement during the plateau phase of the sample detection group, and subtracting the non-specific binding displacement of the blank control group to obtain the effective resonance peak displacement caused only by the specific binding of the four-transmembrane protein.

[0068] Protein content calculation and conversion: By substituting the effective displacement into a pre-established standard curve, the total mass of tetraspan proteins bound on the SPR chip sensing interface was calculated; then, by converting the effective reaction area of ​​the chip sensing interface and the fixed total cell membrane spreading area, the actual expression level of tetraspan proteins per unit cell membrane area (unit: ng / mm²) was obtained. 2 Simultaneously, based on the total protein content of the cell membrane, the mass percentage of tetraspan membrane proteins in the total membrane protein can be calculated, thus enabling quantitative detection of the true expression level of tetraspan membrane proteins under the native structure of the cell membrane surface.

[0069] In some embodiments, acquiring the optical signal curves of the entire process of the binding phase, equilibrium phase, and dissociation phase, and generating detection results, includes: extracting the collected time-series signal data, identifying the binding phase where the signal rises rapidly, the equilibrium phase where the signal remains stable, and the dissociation phase where the signal gradually declines, respectively, fitting the complete response signal curves, and extracting and organizing the characteristic parameters of each phase to generate standardized detection results.

[0070] This embodiment is a detailed implementation of the signal curve processing and detection result generation step in S103. The core purpose is to divide the full-time SPR raw signal data into stages, fit the dynamics, and extract the feature parameters to generate standardized and traceable detection results. This provides quantitative feature data support for subsequent delivery activity evaluation and solves the problems of messy raw signal data, unclear features, and inability to be directly used for evaluation.

[0071] Data import and preprocessing involves retrieving and importing full-time SPR optical signal data that has undergone baseline correction and noise reduction into professional dynamic analysis software.

[0072] The automatic segmentation of the reaction phases automatically identifies and divides the reaction into three core phases based on the temporal variation characteristics of the signal: The binding phase identifies segments where the signal rises rapidly, specifically the time interval from the introduction of the flowing release medium to the point where the signal rises to 90% of its plateau phase, corresponding to the specific binding process of the four-span membrane protein to the sensing interface. The equilibrium phase identifies plateau segments where the signal remains stable, i.e. time intervals where the signal fluctuation amplitude is less than ±5%, corresponding to the process of binding and dissociation reaching dynamic equilibrium. The dissociation phase is identified by the segment where the signal gradually decreases, that is, the time interval from the start of switching to blank buffer to the point where the signal drops to 10% of the plateau phase, which corresponds to the dissociation process of the four-transmembrane protein.

[0073] The dynamic curve fitting uses a 1:1 combination dynamic model to globally fit the signal data of the three stages, generating a complete SPR response dynamic curve. The goodness of fit R needs to be greater than 0.99 to ensure the accuracy of the fitted curve. 2 Feature parameter extraction involves extracting standardized feature parameters for each stage from the fitted kinetic curve, including but not limited to: binding rate constant ka, dissociation rate constant kd, equilibrium dissociation constant KD, maximum binding response value Rmax, plateau signal stability, binding half-peak time, and dissociation half-life.

[0074] The standardized test result generation process integrates the original time-series data, the fitted kinetic curves, the extracted feature parameters, and the quantitative results of four-transmembrane proteins, and generates a traceable and repeatable test report according to a preset standardized format, thus completing the generation of standardized test results.

[0075] In some embodiments, the comparison of the detection results with the corresponding benchmark data, the calculation of the temporal correspondence between the drug release rate and the light signal based on the signal curve, and the comprehensive quantitative determination of the targeting binding capacity and overall delivery activity of the targeted delivery system to complete the evaluation include: extracting the expression levels of four transmembrane proteins and signal characteristics at each stage before delivery as benchmark data, calculating the feature differences between the post-delivery detection data and the benchmark data, fitting the signal curve to obtain the drug release rate, inputting the targeting binding capacity and release kinetic parameters into the evaluation model, and outputting the quantitative evaluation results of the overall delivery activity to complete the evaluation.

[0076] This embodiment is a detailed implementation of the comprehensive evaluation step in S103. The core purpose is to achieve a multi-dimensional quantitative evaluation of the targeting binding ability, release kinetics, and overall delivery activity of the targeted delivery system by comparing and analyzing the test results with the benchmark data and combining kinetic characteristic parameters. This solves the problem that existing technologies can only indirectly infer the delivery effect and cannot directly quantify the delivery activity, and provides a direct and accurate quantitative evaluation basis for the drug development screening of delivery systems targeting four-transmembrane proteins.

[0077] The benchmark dataset was constructed by retrieving the corresponding benchmark data, including: ① the basic expression level of tetraspan protein in blank tumor cell samples from the same batch that were not treated with external vesicle delivery, and the corresponding SPR kinetic parameters, as blank benchmarks; ② the detection data of blank external vesicle samples after delivery of unmodified CARs, as non-specific binding benchmarks; ③ the kinetic parameters of tetraspan protein standards, as quantitative benchmarks.

[0078] Benchmark comparison and core indicator calculation: By pairing the generated standardized detection results with the benchmark dataset, the core feature differences are calculated, including the difference in expression levels of four-transmembrane proteins before and after delivery, the signal-to-noise ratio of specific binding signals and non-specific binding signals, and the deviation values ​​of kinetic parameters between the detection group and the standard. At the same time, by combining the fitted SPR kinetic curve, the temporal correspondence between drug release rate and light signal change is calculated through the temporal change of the signal rise segment, clarifying the drug release kinetic parameters of the delivery system, including initial release rate, cumulative release rate, and release half-life.

[0079] The quantitative evaluation model was constructed by means of a delivery activity quantitative evaluation model. The input parameters of the model include target binding capacity related parameters (KD value, Rmax value, signal-to-noise ratio), release kinetic parameters (release rate, cumulative release rate), and target protein expression upregulation factor. Corresponding weight coefficients were set for each input parameter, with target binding capacity accounting for 40%, target protein expression upregulation factor accounting for 30%, and release kinetic characteristics accounting for 30%.

[0080] The comprehensive evaluation output is obtained by inputting the various parameters of this test into the evaluation model, and after weighted calculation, the overall delivery activity quantitative score of 0-100 is output. At the same time, the sub-scores of target binding ability and release kinetics are also output, thus completing the comprehensive quantitative evaluation of this CAR-modified exovesicle targeted four-transmembrane protein delivery system. The higher the score, the better the delivery activity and drug potential.

[0081] In some embodiments, the method further includes: pre-collecting optical signal curves and actual drug efficacy verification results corresponding to multiple sets of samples with different delivery efficiencies, training a deep learning evaluation model, inputting the optical signal curves of the entire process obtained in this detection into the trained model, having the model automatically extract the implicit features of each stage, predicting the expected in vivo anti-tumor effect of the delivery system to be evaluated, and then combining it with a database of differential expression of four-span membrane proteins in different tumors to output a ranking result of drug potential adapted to different tumor types.

[0082] This embodiment is an intelligent extension and advanced application of the core evaluation methods S101-S103. Building upon the completed basic evaluation of SPR in situ quantitative detection and delivery activity, it addresses the industry pain points of existing technologies, such as the disconnect between in vitro activity detection and in vivo efficacy prediction, low efficiency of high-throughput screening for drug development, and insufficient accuracy in indication matching. Based on deep learning technology, this embodiment constructs a quantitative correlation model between SPR temporal kinetic characteristics and in vivo antitumor efficacy, achieving end-to-end accurate prediction from in vitro non-destructive testing data to expected in vivo antitumor effects. Simultaneously, by combining a database of differential expression of four-span membrane proteins in different tumors, it ranks the drug development potential of the delivery system for different tumor indications, providing intelligent support throughout the entire process for early high-throughput screening of drug development and precise indication development of antitumor delivery systems targeting four-span membrane proteins.

[0083] This embodiment is an advanced execution step after the basic delivery activity assessment is completed in step S103. The specific implementation process is divided into 5 core steps, and the entire process complies with the drug development data compliance requirements. All clinical sample data involved have been authorized by relevant users and approved ethically, and there is no risk of privacy infringement.

[0084] Step 1: Construction and Standardization of Multi-Dimensional Training Dataset This step forms the foundation for model training. Its core is to complete the pairing and annotation of SPR detection data and in vivo pharmacodynamic gold standard data, constructing a high-quality, high-coverage training dataset. The specific implementation is as follows: Sample and Data Collection: Samples of engineered extravesicular vesicle delivery systems for targeted four-span membrane protein CARs, covering the full range of delivery efficiencies, were collected in advance. These samples included experimental groups with different CAR modification targets, different extravesicular vesicle drug loadings, different membrane modification types, and different delivery efficiencies. Blank vector, negative control, and marketed positive control drug groups were also included. Two types of core data were collected from each sample group: In vitro detection data: The full-cycle SPR time-series optical signal curve obtained by detection according to the above standardized procedure, as well as the corresponding target protein expression level, kinetic characteristic parameters, and delivery activity quantification score; The gold standard data for drug efficacy are obtained through in vitro cellular efficacy validation data (tumor cell proliferation inhibition rate, apoptosis rate, and target binding rate) and in vivo animal model antitumor efficacy data (tumor growth inhibition rate TGI, survival time, target tissue enrichment, and systemic toxicity data in tumor-bearing mice).

[0085] Data cleaning and quality control involves cleaning the collected raw data to remove invalid samples with signal drift, excessive batch error, or insufficient repeatability of efficacy data. The retained valid samples must meet the following requirements: the coefficient of variation (CV) of three repeated experiments of SPR detection is <10%, the in vivo efficacy data are set up with at least three parallel animals, and the statistical difference of the data is P <0.05.

[0086] Data labeling and dataset partitioning: The SPR time series curves of effective samples are used as input features, and the corresponding core efficacy indicators such as in vivo tumor growth inhibition rate and target organ enrichment efficiency are used as gold standard labels for standardized labeling. The labeled dataset is randomly divided into training set, validation set and test set in a ratio of 7:2:1. The training set is used for model parameter fitting, the validation set is used for hyperparameter tuning, and the test set is used for final model accuracy verification.

[0087] Data standardization and augmentation eliminate systematic errors caused by differences in instrument batches, detection environments, and chips by performing baseline correction and normalization on all SPR time-series signal data; for extreme data with small sample sizes (such as samples with ultra-high / ultra-low delivery efficiency), time-series data augmentation techniques are used to amplify the data and improve the generalization ability of the model.

[0088] Step 2: Construction, training, and validation of the deep learning evaluation model: The core of this step is to construct a deep learning model adapted to the characteristics of SPR time-series signals, and establish a precise correlation between in vitro detection data and in vivo drug efficacy. The specific implementation is as follows: Model architecture design: A CNN-LSTM hybrid deep learning architecture is adopted, with a corresponding attention mechanism module, to meet the feature extraction requirements of SPR time series signals. Input layer: Receives standardized full-cycle SPR timing optical signal data, with the input dimension being the number of timing sampling points × 1; Feature extraction layer: Local nonlinear features at each stage of combination, balancing, and dissociation are extracted through CNN convolutional layers, temporal dynamic latent features of the whole cycle are extracted through LSTM long short-term memory network, and key signal segments with the greatest impact on in vivo drug efficacy are automatically weighted through attention mechanism module to improve model interpretability; Output layer: Outputs the core in vivo efficacy prediction values ​​of the delivery system to be evaluated, including three core indicators: in vivo tumor growth inhibition rate, target tumor tissue enrichment efficiency, and in vivo circulating half-life. It also outputs the 95% confidence interval of the prediction results.

[0089] Model training and tuning: The model was trained under supervision using the training set data. The loss function was set as mean squared error (MSE), the optimizer was set as Adam, the initial learning rate was set to 0.001, and the number of training iterations was set to 500. After each training iteration, the model accuracy was verified using the validation set data. An early stopping mechanism was used to avoid overfitting. The model hyperparameters were tuned using the grid search method to determine the optimal combination of model parameters.

[0090] Model accuracy verification and consolidation: The trained model is verified using independent test set data. The Pearson correlation coefficient R between the model's predicted values ​​and the actual in vivo drug efficacy values ​​is required to be ≤5%, which meets the accuracy requirements for preclinical drug screening. After verification, the model weights and inference logic are consolidated to form a standardized inference model that can be directly called. 2 ≥0.95 Step 3: Prediction of the expected in vivo antitumor effect of the sample to be evaluated: The core of this step is to input the detection results from the aforementioned embodiments into the solidified model to complete the automated prediction of in vivo drug efficacy. The specific implementation is as follows: Data preprocessing: Retrieve the full-cycle standardized SPR optical signal curve of the sample to be evaluated generated in Example 6, and complete the baseline correction and normalization process according to the preprocessing process that is completely consistent with the training set to generate standardized time series data that meets the model input requirements.

[0091] Model inference and feature extraction: The preprocessed time series data is input into the trained and solidified deep learning evaluation model. The model automatically extracts implicit dynamic features that are not human-identifiable in the binding phase, equilibrium phase, and dissociation phase, including the nonlinear rise rate in the binding phase, the signal fluctuation characteristics during the plateau phase, and the biphasic decay characteristics in the dissociation phase.

[0092] Drug efficacy prediction results output: Based on the extracted features, the model outputs the core indicators of the expected anti-tumor effect of the delivery system to be evaluated in vivo from end to end, including the expected tumor growth inhibition rate of the tumor-bearing model, the expected enrichment efficiency of the target tumor tissue, and the expected circulating half-life in vivo. At the same time, it outputs the 95% confidence interval of each indicator, providing a quantitative reference for the design of preclinical in vivo experiments.

[0093] Step 4: Ranking of drug potential based on tumor target expression database: The core of this step is to combine the tumor expression differences of tetraspan proteins to complete the indication suitability assessment and drug potential ranking of the delivery system. The specific implementation is as follows: Construction of a differential tumor expression database for four-span membrane proteins: Data from the TCGA and GEO public tumor genomics databases, as well as compliant and authorized clinical tumor sample testing data, were pre-integrated to construct a standardized differential tumor target expression database. The database covers common solid tumor types such as lung cancer, breast cancer, liver cancer, colorectal cancer, pancreatic cancer, and ovarian cancer, and includes data on the abundance of four-span membrane protein target mRNA expression, positive expression rate of cell membrane surface proteins, correlation between target expression and tumor prognosis, and level of unmet clinical needs for each tumor type.

[0094] Indication suitability score calculation: Core parameters such as delivery system target binding ability, in vivo enrichment efficiency, and target protein upregulation efficiency predicted by the model are matched with target expression data for various tumor types in the database. A weighted algorithm is used to calculate the suitability score. The scoring formula is as follows: Suitability score = Target positive expression rate weight × 40% + Delivery system targeted binding ability weight × 35% + Clinical unmet need level weight × 25% Drug potential ranking results output: Based on the suitability score from high to low, the drug potential ranking results for different tumor types suitable for this delivery system are output. At the same time, the recommended development priority, preclinical trial suggestions, and target differentiation competitive advantage analysis are also output for each indication, providing a decision-making basis for the precise development of indications for this delivery system.

[0095] An incremental iteration mechanism for the model is established, continuously supplementing the training dataset with newly added delivery system SPR detection data, in vivo efficacy verification results, and clinical sample verification data. The model is incrementally trained and its accuracy is optimized once every quarter to continuously improve the model's prediction accuracy, generalization ability, and indication coverage, and to adapt to more types of delivery systems targeting four-span membrane proteins and tumor indication scenarios.

[0096] Please see Figure 3 As shown, Figure 3 This is a schematic diagram of the structure of a four-transmembrane protein post-delivery activity evaluation system 200 based on surface plasmon resonance (SPR) provided in this application embodiment. This SPR-based four-transmembrane protein post-delivery activity evaluation system 200 is used to perform the steps of the SPR-based four-transmembrane protein post-delivery activity evaluation method shown in the above embodiments. The SPR-based four-transmembrane protein post-delivery activity evaluation system 200 can be a single server or a server cluster, or it can be a terminal, such as a handheld terminal, laptop computer, wearable device, or robot.

[0097] like Figure 3 As shown, the four-transmembrane protein delivery activity assessment system 200 based on surface plasmon resonance includes: The sample processing unit 201 is used to preprocess the sample to be tested, which is a chimeric antigen receptor engineered immune cell-derived modified extracellular vesicle delivering a tetraspan membrane protein, and to fix the sample to be tested onto the detection chip. The medium release unit 202 is used to introduce a flowing release medium into the detection chip through microfluidic technology and record the signal changes before, during and after release. Based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change in interface refractive index. Evaluation completion unit 203 is used to acquire the optical signal curves of the entire process of binding, equilibrium and dissociation phases, and generate detection results; compare the detection results with the corresponding benchmark data, calculate the time sequence correspondence between drug release rate and optical signal based on the signal curves, and comprehensively quantify the targeting binding ability and overall delivery activity of the targeted delivery system to complete the evaluation.

[0098] In some embodiments, the pretreatment of the sample to be tested, which is a modified extracellular vesicle from chimeric antigen receptor-engineered immune cells that delivers a tetraspan membrane protein, includes: removing tissue debris and unbound free proteins from the sample, and sealing the cell membrane sample to prevent non-specific protein binding from interfering with subsequent detection signals.

[0099] In some embodiments, fixing the sample to be detected onto the detection chip includes: uniformly introducing a pretreated cell membrane suspension containing tetraspan membrane proteins into the detection reaction channel of the microfluidic chip, allowing it to stand and incubate to allow the cell membrane to bind and fix to the modified sensing surface of the chip, and discharging unfixed free suspension residue.

[0100] In some embodiments, the introduction of a flow release medium into the detection chip via microfluidic technology includes: controlling a pumping device according to preset parameters to introduce a flow release medium containing an eluent into the detection reaction channel of the microfluidic chip at a constant flow rate, maintaining a stable flow rate and pressure throughout the entire introduction process.

[0101] In some embodiments, recording the signal changes before, during, and after release includes: acquiring the initial light signal of the blank sensing interface as a baseline before release; continuously acquiring resonant light signal intensity data at fixed time intervals during release; continuously acquiring light signal data of the dissociation process after release; and storing all acquired data in chronological order.

[0102] In some embodiments, the method of detecting the actual expression level of a four-span membrane protein on the cell membrane based on the surface plasmon resonance principle and the optical signal displacement caused by the change in interface refractive index includes: identifying the displacement of the resonance peak, calculating the total content of bound four-span membrane proteins based on the correspondence between the change in refractive index and the amount of protein binding, and converting the actual expression level of the four-span membrane protein per unit cell membrane area.

[0103] In some embodiments, acquiring the optical signal curves of the entire process of the binding phase, equilibrium phase, and dissociation phase, and generating detection results, includes: extracting the collected time-series signal data, identifying the binding phase where the signal rises rapidly, the equilibrium phase where the signal remains stable, and the dissociation phase where the signal gradually declines, respectively, fitting the complete response signal curves, and extracting and organizing the characteristic parameters of each phase to generate standardized detection results.

[0104] In some embodiments, the comparison of the detection results with the corresponding benchmark data, the calculation of the temporal correspondence between the drug release rate and the light signal based on the signal curve, and the comprehensive quantitative determination of the targeting binding capacity and overall delivery activity of the targeted delivery system to complete the evaluation include: extracting the expression levels of four transmembrane proteins and signal characteristics at each stage before delivery as benchmark data, calculating the feature differences between the post-delivery detection data and the benchmark data, fitting the signal curve to obtain the drug release rate, inputting the targeting binding capacity and release kinetic parameters into the evaluation model, and outputting the quantitative evaluation results of the overall delivery activity to complete the evaluation.

[0105] In some embodiments, the method further includes: pre-collecting optical signal curves and actual drug efficacy verification results corresponding to multiple sets of samples with different delivery efficiencies, training a deep learning evaluation model, inputting the optical signal curves of the entire process obtained in this detection into the trained model, having the model automatically extract the implicit features of each stage, predicting the expected in vivo anti-tumor effect of the delivery system to be evaluated, and then combining it with a database of differential expression of four-span membrane proteins in different tumors to output a ranking result of drug potential adapted to different tumor types.

[0106] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the above-described four-transmembrane protein delivery activity evaluation system and its modules based on surface plasmon resonance can be referred to the corresponding content in the various embodiments of the above-described four-transmembrane protein delivery activity evaluation method based on surface plasmon resonance, and will not be repeated here.

[0107] The aforementioned method for evaluating the post-delivery activity of four-transmembrane proteins based on surface plasmon resonance can be implemented as a computer program, which can be used in applications such as... Figure 3 It runs on the device shown.

[0108] Please see Figure 4 , Figure 4 This is a schematic block diagram of the structure of a computer device provided in an embodiment of this application. The computer device includes a processor, a memory, and a network interface connected via a device bus, wherein the memory may include a storage medium and internal memory.

[0109] The storage medium may store operating devices and computer programs. The computer program includes program instructions that, when executed, cause the processor to perform any method for evaluating the post-delivery activity of a four-transmembrane protein based on surface plasmon resonance.

[0110] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0111] Internal memory provides an environment for the execution of computer programs in non-volatile storage media. When executed by a processor, the computer program enables the processor to perform any method for evaluating the activity of four-transmembrane proteins after delivery based on surface plasmon resonance.

[0112] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the terminal to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0113] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0114] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps: The sample to be tested, which is a modified extracellular vesicle delivering a tetraspan membrane protein via chimeric antigen receptor-engineered immune cell, is pretreated and then immobilized on the detection chip. In the detection chip, a flowing release medium is introduced through microfluidic technology, and the signal changes before, during and after release are recorded respectively. Based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change in interface refractive index. The optical signal curves of the entire process of binding, equilibrium and dissociation are obtained to generate detection results. The detection results are compared with the corresponding benchmark data, and the temporal correspondence between drug release rate and optical signal is calculated by combining the signal curves. The targeted binding ability and overall delivery activity of the targeted delivery system are obtained by comprehensive quantification to complete the evaluation.

[0115] In some embodiments, the pretreatment of the sample to be tested, which is a modified extracellular vesicle from chimeric antigen receptor-engineered immune cells that delivers a tetraspan membrane protein, includes: removing tissue debris and unbound free proteins from the sample, and sealing the cell membrane sample to prevent non-specific protein binding from interfering with subsequent detection signals.

[0116] In some embodiments, fixing the sample to be detected onto the detection chip includes: uniformly introducing a pretreated cell membrane suspension containing tetraspan membrane proteins into the detection reaction channel of the microfluidic chip, allowing it to stand and incubate to allow the cell membrane to bind and fix to the modified sensing surface of the chip, and discharging unfixed free suspension residue.

[0117] In some embodiments, the introduction of a flow release medium into the detection chip via microfluidic technology includes: controlling a pumping device according to preset parameters to introduce a flow release medium containing an eluent into the detection reaction channel of the microfluidic chip at a constant flow rate, maintaining a stable flow rate and pressure throughout the entire introduction process.

[0118] In some embodiments, recording the signal changes before, during, and after release includes: acquiring the initial light signal of the blank sensing interface as a baseline before release; continuously acquiring resonant light signal intensity data at fixed time intervals during release; continuously acquiring light signal data of the dissociation process after release; and storing all acquired data in chronological order.

[0119] In some embodiments, the method of detecting the actual expression level of a four-span membrane protein on the cell membrane based on the surface plasmon resonance principle and the optical signal displacement caused by the change in interface refractive index includes: identifying the displacement of the resonance peak, calculating the total content of bound four-span membrane proteins based on the correspondence between the change in refractive index and the amount of protein binding, and converting the actual expression level of the four-span membrane protein per unit cell membrane area.

[0120] In some embodiments, acquiring the optical signal curves of the entire process of the binding phase, equilibrium phase, and dissociation phase, and generating detection results, includes: extracting the collected time-series signal data, identifying the binding phase where the signal rises rapidly, the equilibrium phase where the signal remains stable, and the dissociation phase where the signal gradually declines, respectively, fitting the complete response signal curves, and extracting and organizing the characteristic parameters of each phase to generate standardized detection results.

[0121] In some embodiments, the comparison of the detection results with the corresponding benchmark data, the calculation of the temporal correspondence between the drug release rate and the light signal based on the signal curve, and the comprehensive quantitative determination of the targeting binding capacity and overall delivery activity of the targeted delivery system to complete the evaluation include: extracting the expression levels of four transmembrane proteins and signal characteristics at each stage before delivery as benchmark data, calculating the feature differences between the post-delivery detection data and the benchmark data, fitting the signal curve to obtain the drug release rate, inputting the targeting binding capacity and release kinetic parameters into the evaluation model, and outputting the quantitative evaluation results of the overall delivery activity to complete the evaluation.

[0122] In some embodiments, the method further includes: pre-collecting optical signal curves and actual drug efficacy verification results corresponding to multiple sets of samples with different delivery efficiencies, training a deep learning evaluation model, inputting the optical signal curves of the entire process obtained in this detection into the trained model, having the model automatically extract the implicit features of each stage, predicting the expected in vivo anti-tumor effect of the delivery system to be evaluated, and then combining it with a database of differential expression of four-span membrane proteins in different tumors to output a ranking result of drug potential adapted to different tumor types.

[0123] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method for evaluating the activity of a four-transmembrane protein based on surface plasmon resonance as provided in any embodiment of this application.

[0124] The computer-readable storage medium can be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device. The above descriptions are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for evaluating the activity of a four-transmembrane protein after delivery based on surface plasmon resonance, characterized in that, include: The sample to be tested, which is a modified extracellular vesicle delivering a tetraspan membrane protein via chimeric antigen receptor-engineered immune cell, is pretreated and then immobilized on the detection chip. In the detection chip, a flowing release medium is introduced through microfluidic technology, and the signal changes before, during and after release are recorded respectively. Based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change in interface refractive index. The optical signal curves of the entire process of binding, equilibrium and dissociation are obtained to generate detection results. The detection results are compared with the corresponding benchmark data, and the temporal correspondence between drug release rate and optical signal is calculated by combining the signal curves. The targeted binding ability and overall delivery activity of the targeted delivery system are obtained by comprehensive quantification to complete the evaluation.

2. The method according to claim 1, characterized in that, The pretreatment of the sample to be tested, which is a modified extracellular vesicle delivering a tetraspan membrane protein derived from chimeric antigen receptor-engineered immune cells, includes: Remove contaminating tissue debris and unbound free proteins from the sample, and seal the cell membrane sample to prevent non-specific protein binding from interfering with subsequent detection signals.

3. The method according to claim 1, characterized in that, The process of fixing the sample to be tested onto the detection chip includes: The pretreated cell membrane suspension containing tetraspan membrane proteins was introduced at a constant rate into the detection reaction channel of the microfluidic chip. After static incubation, the cell membrane was bound and fixed to the modified sensing surface of the chip, and the unfixed free suspension residue was discharged.

4. The method according to claim 1, characterized in that, The process of introducing a flow release medium into the detection chip via microfluidic technology includes: The pumping device is controlled according to preset parameters to introduce the flowing release medium containing eluent into the detection reaction channel of the microfluidic chip at a constant flow rate, maintaining a stable flow rate and pressure throughout the entire introduction process.

5. The method according to claim 4, characterized in that, The recording of signal changes before, during, and after release includes: Before release, the initial optical signal of the blank sensing interface is collected as a baseline. During the release process, the intensity data of the resonant optical signal is continuously collected at fixed time intervals. After the release is completed, the optical signal data of the dissociation process is continuously collected. All collected data are stored in chronological order.

6. The method according to claim 1, characterized in that, The method, based on the principle of surface plasmon resonance and the displacement of light signals caused by changes in interfacial refractive index, detects the actual expression level of a four-transmembrane protein on the cell membrane, including: The displacement of the resonance peak is identified, and the total content of bound tetraspan proteins is calculated based on the correspondence between the refractive index change and the amount of protein binding. The actual expression level of tetraspan proteins per unit cell membrane area is then calculated.

7. The method according to claim 1, characterized in that, The acquisition of optical signal curves throughout the entire process of the binding, equilibrium, and dissociation phases, and the generation of detection results, include: The collected time-series signal data is extracted, and the binding stage (where the signal rises rapidly), the equilibrium stage (where the signal remains stable), and the dissociation stage (where the signal gradually declines) are identified. A complete response signal curve is fitted, and standardized detection results are generated after extracting and organizing the characteristic parameters of each stage.

8. The method according to claim 1, characterized in that, The process of comparing the detection results with corresponding benchmark data, calculating the temporal correspondence between drug release rate and light signal based on signal curves, and comprehensively quantifying the targeting binding ability and overall delivery activity of the targeted delivery system to complete the evaluation includes: The expression levels of four transmembrane proteins and signal characteristics at each stage before delivery were extracted as baseline data. The feature differences between the post-delivery detection data and the baseline data were calculated. The drug release rate was obtained by fitting the signal curve. The targeting binding capacity and release kinetic parameters were input into the evaluation model, and the quantitative evaluation results of the overall delivery activity were output to complete the evaluation.

9. The method according to claim 1, characterized in that, Also includes: The optical signal curves and actual drug efficacy verification results corresponding to multiple sets of samples with different delivery efficiencies were collected in advance to train a deep learning evaluation model. The optical signal curves of the whole process obtained in this test were input into the trained model, and the model automatically extracted the hidden features of each stage to predict the expected in vivo anti-tumor effect of the delivery system to be evaluated. Then, combined with the database of differential expression of four-span membrane proteins of different tumors, the drug potential ranking results adapted to different tumor types were output.

10. A system for evaluating the activity of a four-transmembrane protein after delivery based on surface plasmon resonance, characterized in that, The method applied to any one of claims 1-9 includes: The sample processing unit is used to preprocess the sample to be tested, which is a transmembrane protein delivered by extracellular vesicles modified from chimeric antigen receptor-engineered immune cells, and to fix the sample to be tested onto the detection chip. The medium release unit is used to introduce a flowing release medium into the detection chip via microfluidic technology and record the signal changes before, during, and after release. Based on the principle of surface plasmon resonance, the actual expression level of the four-transmembrane protein on the cell membrane is detected according to the optical signal displacement caused by the change in interface refractive index. The evaluation completion unit is used to acquire the optical signal curves of the entire process of binding, equilibrium and dissociation, and generate detection results; the detection results are compared with the corresponding benchmark data, and the temporal correspondence between drug release rate and optical signal is calculated by combining the signal curves, so as to comprehensively and quantitatively obtain the targeting binding ability and overall delivery activity of the targeted delivery system, and complete the evaluation.

Images