An exosome large-scale purification method based on tangential flow filtration combined with size exclusion chromatography

Abstract

The application discloses an exosome large-scale purification method based on tangential flow filtration and size exclusion chromatography. In view of the technical problems that the membrane flux decays due to concentration polarization when large-volume biological fluid is concentrated in the existing process, and the separation efficiency of the downstream size exclusion chromatography is seriously limited by the high-viscosity feed liquid after high-ratio concentration, the application strictly controls the fluid mechanics coupling of the shear rate of the inner wall of the hollow fiber membrane and the radial transmembrane pressure, and cooperates with the periodic micro-pulse backwashing mechanism to inhibit the formation of the polarization gel layer from the root, and then the rheological reconstruction of the feed liquid is completed through low-ratio isometric constant-volume washing and filtration, so that the downstream size exclusion chromatography column can be stably eluted and separated at a high linear speed far beyond the normal value. The application realizes high-purity large-scale preparation of exosomes with a recovery rate of more than 83%, a vesicle structure integrity rate of more than 95%, a particle-to-protein ratio of more than 10 9 particles / µg, and a host protein removal rate of more than 99.8%, and meets the needs of clinical-grade exosome industrial production.

Description

Technical Field

[0001] This invention relates to the field of biomedical separation and purification engineering technology, and in particular to a method for large-scale purification of exosomes based on tangential flow filtration coupled with size exclusion chromatography. Background Technology

[0002] Exosomes are nanoscale membrane vesicles secreted by cells, possessing immense application value in disease diagnosis and targeted drug delivery. Currently, the core bottleneck in transitioning exosome applications from laboratory research to industrial-scale commercial production lies in downstream large-scale separation and purification processes. Traditional ultracentrifugation relies on extreme centrifugal gravitational fields and immense mechanical stress, easily causing vesicle rupture and exhibiting extremely low equipment processing capacity, making it unsuitable for the large-scale processing demands of industrial-grade biological culture supernatants. Furthermore, due to the fragile structure of exosomes and the complex composition of biological fluids, traditional filtration separation technologies also face significant engineering challenges during processing.

[0003] Size exclusion chromatography (SUC) is widely recognized as an effective method for obtaining high-purity exosomes due to its mild separation conditions and non-destructive physicochemical processes. However, SUC suffers from insurmountable physical volume limitations; the single sample loading volume cannot exceed a very small proportion of the column bed volume. Directly processing large volumes of industrial-grade samples would require extremely large and impractical equipment investments and turnaround times. To overcome these volume limitations, the industry has attempted to combine tangential flow filtration as a pre-concentration step with SUC. However, in existing combined processes, due to the lack of a systematic rheological and hydrodynamic adaptation mechanism, when tangential flow filtration processes complex biological fluids containing high concentrations of free proteins and cell debris, free proteins and impurity particles rapidly accumulate and solidify on the filter membrane surface, leading to severe concentration polarization. This flux decline caused by membrane fouling not only forces the system to abnormally increase transmembrane pressure, directly crushing exosomes, but also results in highly concentrated feed solutions exhibiting extremely high non-Newtonian viscosity without special treatment. Once such high-viscosity concentrate is pumped into a downstream size exclusion chromatography column, it will cause severe fluid channeling and osmotic shock, which will not only cause the compression and collapse of the chromatography resin bed, but also force the elution flow rate to be maintained at an extremely low level, which will seriously reduce the theoretical plate number and process efficiency of the separation system.

[0004] Therefore, how to completely suppress the concentration polarization phenomenon at the membrane interface while achieving high-rate concentration of large-volume feed solutions, and how to accurately match the hydrodynamic state of the concentrated feed solution with the tolerance limit of the size exclusion chromatography matrix, so as to achieve high-throughput large-scale high-purity separation without damaging the exosome activity, is a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] To address the problems of membrane concentration polarization and clogging caused by large-volume feed concentration in existing combined separation processes, and the technical defects of downstream size exclusion chromatography separation flow rate limitation and resin bed instability caused by high-volume feed concentration, this invention provides a method for large-scale purification of exosomes based on the combination of tangential flow filtration and size exclusion chromatography.

[0006] This invention provides a method for large-scale purification of exosomes based on tangential flow filtration coupled with size exclusion chromatography, comprising the following sequential steps: The raw biological fluid containing exosomes is centrifuged and clarified to remove cellular and micron-sized particulate impurities, obtaining a primary supernatant. The primary supernatant is then introduced into a tangential flow filtration system containing a hollow fiber membrane module for cyclic concentration, wherein the molecular weight cutoff of the hollow fiber membrane module is limited to 500 kDa to 750 kDa. During concentration, the wall shear rate of the liquid on the inner wall of the hollow fibers is controlled to be constant at 4,000 s⁻¹. -1 Up to 12,000s -1 Between these points, the radial transmembrane pressure is maintained between 0.5 psi and 3.5 psi by adjusting the back pressure on the retentate outlet side, compressing the total feed volume by 10 to 100 times to obtain a concentrated exosome retention solution. Chromatography buffer is continuously added to the obtained concentrated exosome retention solution at a rate equal to the permeate discharge rate for constant-volume rinsing, with a total rinsing buffer volume of 1 to 2 rinsing volumes. During the concentration and rinsing process, forward permeation is periodically paused, and micro-pulse backwashing is applied to the hollow fiber membrane cavity from the permeate side using the pre-filled chromatography buffer as backwashing fluid to strip the polarization layer on the membrane surface, obtaining an exosome loading solution that has undergone viscosity reduction and rheological reconstruction. The obtained exosome loading solution was injected into a size exclusion chromatography column filled with macroporous agarose chromatography matrix. The single loading volume was controlled to be 5% to 10% of the size exclusion chromatography column bed volume. Subsequently, the chromatographic mobile phase was pumped in at a constant linear velocity of 34 cm / h to 300 cm / h for elution. The chromatographic peak in the first elution size exclusion zone was collected to obtain purified exosomes.

[0007] As a preferred embodiment of the present invention, the hollow fiber membrane module is made of hydrophilic modified polyethersulfone or regenerated cellulose, and the hollow fiber membrane module has open straight channels inside, with the inner diameter of each capillary fiber being 0.5 mm to 1.0 mm. This smooth tubular channel ensures that the fluid cross-section exhibits a laminar parabolic distribution, effectively avoiding mechanical damage to the target vesicles caused by structural dead corners.

[0008] As a preferred embodiment of the present invention, the specific parameters of the micro-pulse backwashing operation are limited to the following: after every 20 minutes of continuous concentration or filtration cycle, the system first pauses the forward permeation operation to bring the transmembrane pressure to zero instantaneously, and then performs a momentary reverse backwash lasting 10 to 15 seconds. The applied backwash fluid flow rate is 5% to 10% of the initial steady-state forward membrane flux. The initial steady-state forward membrane flux refers to the steady-state permeation flux value measured at the beginning of system operation under the set transmembrane pressure conditions, before a significant polarization layer has formed on the membrane surface. With the forward permeation driving force terminated, this small reverse flow rate is sufficient to establish an effective reverse pressure gradient, capable of resuspending hydrophobic impurities that have crossed the critical adhesion point.

[0009] As a preferred embodiment of the present invention, the chromatography buffer is selected from phosphate buffered saline (PBS solution) with a pH of 7.4 or hydroxyethylpiperazine ethanesulfonic acid isotonic saline (HEPES isotonic saline solution). After washing and filtration, the system osmotic pressure of the exosome loading solution is stabilized between 280 mOsm / kg and 320 mOsm / kg.

[0010] As a preferred embodiment of the present invention, the centrifugal clarification process includes multi-stage refrigerated differential centrifugation, which sequentially performs centrifugation at 300 × g for 10 minutes at 4°C to collect the first-stage supernatant, and then centrifuges the first-stage supernatant at 2,000 × g for 20 minutes at 4°C to collect the primary supernatant.

[0011] As a preferred embodiment of the present invention, when the original biofluid is a body fluid with a high lipid content, the supernatant obtained by centrifugation at 2,000 × g is further centrifuged at 10,000 × g for 30 minutes, or pre-filtered through a microfiltration membrane with a pore size of 0.45 μm, based on the above-mentioned multi-stage cryogenic differential centrifugation. The resulting product is used as the primary supernatant. This physically eliminates the possibility of large-volume emulsified lipid clusters clogging the pores.

[0012] As a preferred technical solution of the present invention, the concentration and washing / filtration replacement steps are performed in a system equipped with a pipeline cooling jacket and the system operating temperature is maintained at 4°C throughout. When the system is running at room temperature, the duration of a single full-flow path processing operation of tangential flow filtration must be limited to within 4 hours.

[0013] As a preferred embodiment of the present invention, when the original biofluid is a complex culture medium with high initial viscosity, the radial transmembrane pressure is maintained between 1.5 psi and 3.5 psi.

[0014] As a preferred technical solution of the present invention, when collecting the chromatographic peaks in the exclusion zone, online real-time monitoring is performed using a series of multi-wavelength ultraviolet detectors and multi-angle dynamic light scattering detectors. When the light scattering detector captures a light scattering signal with a particle size distribution concentrated in the range of 50 nm to 150 nm and simultaneously corresponds to the low ultraviolet absorption characteristics, an automated fraction collector is triggered to collect the peaks.

[0015] As a preferred embodiment of the present invention, after chromatographic elution and collection, the method further includes a formulation-grade sterile microfiltration step, in which the collected purified exosomes are combined and pushed through a polyethersulfone terminal sterile filter membrane with a pore size of 0.22 μm under the condition that the transmembrane pressure difference does not exceed 10 psi, and then aliquoted and transferred to -80°C for deep cryogenic storage.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves a tight boundary fluid dynamics coupling between extremely low transmembrane pressure and a high specific wall shear rate, and intermittently intervenes with a micro-pulse backwashing mechanism during the operating cycle. This fundamentally destroys and inhibits the formation of the concentrated polarization layer and solidified filter cake at the filter pore interface. This kinetic energy cleaning mechanism not only ensures that the hollow fiber separation channel maintains a sufficient and stable filtration flux when processing large volumes of high-concentration liquids, but more importantly, it effectively avoids the extreme hydrostatic pressure compression and fluid tearing stress caused by the attempt to force water permeation in traditional technologies. This ensures the integrity of the brittle vesicle morphology under external force stress, and achieves non-destructive, high-ratio continuous compression of large volumes of materials.

[0017] This invention, before reaching the critical point of extremely high density through tangential flow filtration concentration, performs an isotonic constant-volume elution filtration operation with specific fluid volume parameters to effectively replace the liquid phase matrix and reconstruct the rheology of the viscous retaining solution. This operation smoothly reduces the macroscopic flow resistance of the enriched macromolecular liquid and stabilizes the osmotic pressure of the buffer system, ensuring that the reconstructed feed solution, when pumped into the extremely dense size exclusion chromatography stationary phase matrix, does not induce destructive local high-pressure eddy diffusion or resin skeleton shrinkage. This rheologically adaptive process transition allows the downstream chromatography column to operate at extremely efficient elution rates with confidence and safety, significantly reducing the single-cycle separation time. This method of coupling the preceding flow control model with the subsequent micropore size exclusion completely solves the volume and viscosity deadlock in traditional processing techniques, achieving cross-scale stepwise removal of host proteins and meeting the demand for efficient and stable industrial production from crude culture medium to extremely high-purity clinical-grade exosome products. Detailed Implementation

[0018] The exosome large-scale purification method provided by this invention, based on the coupling of tangential flow filtration and size exclusion chromatography, is fundamentally based on the dynamic coupling control of the wall shear rate and radial transmembrane pressure of the fluid within a hollow fiber tube. This, combined with an intermittent physical regeneration mechanism of periodic micro-pulse backwashing, inhibits the formation and solidification of a concentrated polarized gel layer at the membrane interface. This allows for high-rate continuous volume compression of large-volume biofluids without damaging the integrity of the nanoscale exosome vesicle structure. Furthermore, low-rate isotonic constant-volume washing filtration completes the liquid-phase matrix replacement and rheological reconstruction of the concentrated retention solution. This adjusts the apparent viscosity and osmotic pressure of the high-concentration intermediate solution to precisely match the tolerance limits of the downstream size exclusion chromatography stationary phase matrix, thereby endowing the chromatographic separation system with the ability to operate stably at linear velocities far exceeding conventional limits. Ultimately, this achieves integrated, efficient, and stable industrial production from large-volume crude biological culture medium to extremely high-purity clinical-grade exosome products.

[0019] The physicochemical mechanism of this method is based on the synergistic coupling of the boundary kinetic energy sweeping effect in continuum fluid dynamics and the thermodynamics of diffusion exclusion resistance within the gel micropores. During the tangential flow filtration concentration stage, when the complex biofluid containing exosomes circulates at high speed within the hollow fiber microducts, the transmembrane pressure provides a radial permeation driving force perpendicular to the fiber axis, prompting water molecules and low-molecular-weight proteins with sizes much smaller than the membrane pore size to permeate and be expelled from the fiber outer wall. According to the concentration polarization membrane model, this convective permeation inevitably entrains substances larger than the membrane pores towards the inner surface of the filter membrane for mass transfer. Without external intervention, these trapped substances will rapidly accumulate at the solid-liquid interface, forming a polarized gel layer with an extremely steep concentration gradient. The mass transfer resistance increases logarithmically, leading to a sharp decline in filtration flux until it is completely lost. This invention utilizes the strong pipe wall shear rate generated by the continuous fluid kinetic energy delivered by the main feed pump to induce intense viscous friction and tangential shear stress between fluid layers. This sweeping effect can stimulate shear-induced diffusion flux, causing substances pulled radially towards the membrane surface to be re-entrained and carried into the main liquid flow at the center of the channel just before deposition. The shear rate is strictly limited to 4000 s⁻¹. -1 Up to 12000 s -1Within the window, an extremely mild, ultra-low transmembrane pressure is simultaneously matched radially. This ensures, from an engineering control perspective, that the combined force of the normal compressive stress exerted by the fluid pressure on the vesicle surface and the axial tearing shear force generated by fluid layer friction on the vesicle edge is far lower than the elastic modulus and rupture yield limit of the phospholipid bilayer constituting the exosome shell. During the size exclusion chromatography polishing stage, because the exosomes maintain a physical size much larger than the micropore opening diameter of the chromatographic resin gel network due to their rigid spherical phospholipid shell structure, they are geometrically and dynamically excluded from the internal channels of the resin beads. They can only travel with the mobile phase through the gaps between the macroscopic resin particles and are preferentially collected at the elution front of the dead volume with the fastest migration speed. Conversely, smaller residual contaminants, due to their three-dimensional dimensions being smaller than or exactly equal to the upper limit of the pore size of the gel resin, continuously penetrate into the microporous labyrinth system inside the resin particles under the combined action of molecular Brownian motion and concentration gradient diffusion. Their microscopic journey within the chromatographic column is extremely prolonged, resulting in significant retention and delay at the macroscopic level. They are only slowly eluted in subsequent volume fractions, completely separating from the leading exosomes.

[0020] The key raw materials and equipment specifications involved in this invention are as follows. The hollow fiber membrane module is a tubular module made of hydrophilic modified polyethersulfone or regenerated cellulose, with a molecular weight cutoff between 500 kDa and 750 kDa. The inner diameter of a single capillary fiber ranges from 0.5 mm to 1.0 mm, and it has an open, straight-channel structure internally. A typical module integrates approximately 320 parallel fibers. This straight-channel structure ensures that the shear stress on the fluid cross-section is distributed in a laminar parabolic manner, eliminating the risk of mechanical damage to the target vesicles caused by non-streamlined dead angles. The hollow fiber membrane module can be a hydrophilic modified polyethersulfone hollow fiber filter from Sartorius or a regenerated cellulose hollow fiber module from Sitervan. Size exclusion columns are packed with macroporous agarose matrix. High-rigidity cross-linked agarose size exclusion resin from Topfan or similar industrial-grade rigid macroporous agarose matrix can be used. The column specifications are selected based on the processing scale, with a matching inner diameter and bed height combination. For laboratory-scale operations, a pre-packed column with an inner diameter of 16 mm and a bed height of 15–20 cm can be used, while for industrial-scale operations, a large-diameter pre-packed column with an inner diameter of 100–200 mm and a bed height of 20–30 cm is used. The chromatography buffer is either pH 7.4 phosphate-buffered saline (PBS) or sodium chloride-containing hydroxyethylpiperazine ethanesulfonic acid buffer (HEPES), sterilized by 0.22 µm membrane filtration before use. The tangential flow filtration system is equipped with a low-shear peristaltic pump, an electronic balance or level sensor, a precision back pressure regulating valve, and an osmotic micro-injection pump. The micro-injection pump is pre-filled with 0.22 µm membrane-filtered sterilized chromatography buffer as backwash solution. The micro-pulse backwashing operation sequence is as follows: At each set backwashing cycle node, the system first closes the permeate-side drain valve and pauses the forward permeate drive, causing the transmembrane pressure to drop instantaneously to near zero. Subsequently, the permeate-side micro-injection pump immediately applies a reverse flow pulse at a set flow rate. With the forward permeate force already stopped, this tiny reverse flow rate can generate a sufficient reverse pressure gradient at the membrane interface, stripping and resuspending hydrophobic debris and impurities that have crossed the critical adhesion point from the membrane pore surface and the shallow depths. The automated chromatography distribution station is equipped with a precision pulseless peristaltic pump, a multi-wavelength ultraviolet array detector, and a multi-angle dynamic light scattering detector, and is connected to an automated fraction collector.

[0021] It should be noted that the constant-volume washing operation of 1 to 2 washing volumes specified in this invention is not primarily intended to completely remove the original culture medium components—according to the exponential decay kinetic model of constant-volume replacement, 1 to 2 washing volumes correspond to component replacement rates of approximately 63% and 86%, respectively—but rather to perform macroscopic-level buffer system replacement and rheological viscosity reduction on the concentrated retention solution, effectively regulating the osmotic pressure from the culture medium level to the physiological isotonic range of 280–320 mOsm / kg, ensuring that the hydrodynamic state of the feed solution meets the safe injection requirements of the downstream size exclusion chromatography matrix. Residual trace amounts of culture medium components and impurities of the same molecular weight that could not be removed by dialysis are further removed to extremely low levels in the subsequent high-resolution size exclusion chromatography polishing step. This process design, which functionally separates the cost-effective viscosity reduction function of the front-end washing filtration with the fine polishing and separation function of the back-end chromatography, significantly reduces the washing liquid consumption and processing time while ensuring the purity of the final product.

[0022] The performance testing methods and standards involved in the following examples and comparative examples are as follows. Exosome particle concentration and particle size distribution were determined using a nanoparticle tracking analyzer. Before testing, samples were diluted with phosphate-buffered saline to the concentration range recommended by the instrument. At least three 60-second video tracks were collected for each sample, and the main peak particle size, particle size distribution width, and polydispersity index were reported. Exosome structural integrity was morphologically assessed using cryogenic transmission electron microscopy. After preparing cryopreserved samples under liquid nitrogen quenching conditions, at least 10 fields of view were randomly selected under cryo-electron microscopy, and at least 200 vesicles were counted. The percentage of vesicles maintaining a full and intact double-layered boundary membrane morphology was statistically analyzed. Free host protein concentration was determined using the dicaprinic acid total protein quantification method, with bovine serum albumin as the standard to plot a standard curve. The particle-to-protein ratio was calculated by dividing the particle concentration determined by the nanoparticle tracking analyzer by the total protein concentration determined by the dicaprinic acid method. Exosome surface markers were detected using Western blotting. Positive markers included CD63, CD81, and TSG101, while negative markers included the endoplasmic reticulum protein Calnexin. Membrane flux retention was defined as the ratio of osmotic flux to pure water flux at steady-state operation, expressed as a percentage. Sample osmolality was measured using a freezing point depression osmometer. Exosome recovery was calculated as the ratio of the total number of particles in the purified final product to the total number of particles in the primary supernatant.

[0023] Example 1

[0024] The original biofluid processed in this embodiment was the conditioned medium harvested from the human embryonic kidney 293T cell line after 72 h of serum-free culture in a bioreactor, with a total volume of 5 L. The freshly harvested conditioned medium was centrifuged at 300 × g for 10 min at 4 °C, and the first-stage supernatant was collected and the cell pellet was discarded. Subsequently, the first-stage supernatant was centrifuged at 2000 × g for 20 min at 4 °C to obtain approximately 4.8 L of clear primary supernatant, which was transferred to a stainless steel circulating feed tank equipped with a cooling jacket. The system was maintained at 4 °C throughout the process.

[0025] The primary supernatant was introduced into a tangential flow filtration system equipped with a hydrophilic modified polyethersulfone hollow fiber membrane module with a molecular weight cutoff of 500 kDa. The inner diameter of each fiber in this module is 1.0 mm. The wall shear rate of the liquid on the inner wall of the hollow fiber was kept constant at 8000 s⁻¹ by adjusting the feed pump speed. -1 The radial transmembrane pressure was stabilized at 1.0 psi by using a back pressure regulating valve on the retentate outlet side. Under this parameter combination, the system performed closed-loop concentration, with real-time monitoring of the reservoir mass changes via an electronic balance. During concentration, after every 20 minutes of continuous concentration cycles, the system first closed the permeate-side drain valve to pause forward permeation, then applied a 12-second instantaneous reverse backwash via a micro-injection pump on the permeate side. The backwash fluid flow rate was set to 8% of the initial steady-state forward membrane flux. After approximately 120 minutes of continuous concentration operation, the total feed volume was compressed to approximately 96 mL, with an actual concentration factor of 50 times, yielding the exosome concentrated retentate.

[0026] The system was then switched to constant volume rinsing mode. Phosphate-buffered saline (pH 7.4) was continuously added to the circulating reservoir at the same rate as the osmotic end drain, with a total rinsing buffer volume of 1.5 rinsing volumes, approximately 144 mL. During the rinsing process, the same micro-pulse backwashing parameters as in the concentration phase were maintained. After rinsing, approximately 96 mL of the exosome sample solution, having undergone viscosity reduction and rheological reconstruction, was collected from the drain line. Its osmotic pressure was measured to be 295 mOsm / kg using an osmometer.

[0027] Two mL of the sample solution was injected into a size exclusion column filled with a macroporous agarose chromatography matrix. The column had an inner diameter of 16 mm, a bed height of 15 cm, and a bed volume of 30 mL, with the single sample loading volume accounting for 6.7% of the bed volume. Phosphate-buffered saline was then pumped in at a constant linear velocity of 150 cm / h as the mobile phase for elution. When an online multi-angle dynamic light scattering detector captured a light scattering signal with a particle size distribution concentrated in the 50 nm to 150 nm range and simultaneously corresponded to a low UV absorption characteristic, an automated fraction collector was triggered to collect the chromatographic peak in this size exclusion range, obtaining purified exosomes. The collected purified exosomes were combined and pushed through a 0.22 µm polyethersulfone terminal sterile filter membrane under a transmembrane pressure difference not exceeding 10 psi. After aliquoting, the exosomes were transferred to a cryogenic storage facility at -80°C.

[0028] Example 2

[0029] The original biofluid processed in this example was the conditioned medium harvested after mesenchymal stem cells were cultured in a medium containing 5% exosome-free fetal bovine serum for 48 h, with a total volume of 10 L. The pretreatment stage was the same as in Example 1, and approximately 9.5 L of primary supernatant was obtained after two-stage differential centrifugation.

[0030] The tangential flow filtration system is equipped with a regenerated cellulose hollow fiber membrane module with a molecular weight cutoff of 750 kDa, and each fiber has an inner diameter of 0.5 mm. The wall shear rate is controlled to be constant at 12000 s⁻¹. -1 The radial transmembrane pressure was maintained at 0.5 psi. During concentration, a micro-pulse backwash lasting 10 s was performed every 20 min, with the backwash fluid flow rate being 5% of the initial steady-state forward membrane flux. The total feed volume was compressed 100 times to approximately 95 mL to obtain the exosome concentrate. During the constant-volume washing and filtration stage, isotonic saline solution with pH 7.4 (hydroxyethylpiperazine ethanesulfonic acid) was used as the replacement buffer, and the washing volume was twice the washing volume, approximately 190 mL. After washing and filtration, approximately 95 mL of the exosome sample was collected, and the osmotic pressure was measured to be 310 mOsm / kg. The entire process was conducted at 4°C.

[0031] 1.5 mL of the sample solution was injected into a size exclusion column with an inner diameter of 16 mm, a bed height of 15 cm, and a column bed volume of 30 mL, accounting for 5% of the column bed volume. Elution was performed at a constant linear velocity of 34 cm / h. After collecting the size exclusion peak, the sample was sterilized by 0.22 µm filtration, aliquoted, and stored at -80℃.

[0032] Example 3

[0033] In this embodiment, the original biological fluid processed was the conditioned medium harvested after dendritic cells were cultured in serum-free medium for 96 h, with a total volume of 2 L. Approximately 1.9 L of primary supernatant was obtained after two-stage differential centrifugation.

[0034] The tangential flow filtration system is equipped with a hydrophilic modified polyethersulfone hollow fiber membrane module with a molecular weight cutoff of 600 kDa, and each fiber has an inner diameter of 0.75 mm. The wall shear rate is controlled to be constant at 4000 s⁻¹. -1 The radial transmembrane pressure was maintained at 2.0 psi. A micro-pulse backwash was performed every 20 min for 15 s, with the backwash fluid flow rate being 10% of the initial steady-state forward membrane flux. The feed solution was compressed 10 times to approximately 190 mL to obtain a concentrated exosome buffer. Constant-volume washing and filtration were performed with phosphate-buffered saline at pH 7.4, with a washing volume of one wash volume. After washing and filtration, approximately 190 mL of the exosome loading solution was collected, with an osmotic pressure of 285 mOsm / kg. The entire process was conducted at 4°C.

[0035] 3 mL of the sample solution was injected into a size exclusion column with an inner diameter of 16 mm, a bed height of 15 cm, and a column bed volume of 30 mL, accounting for 10% of the column bed volume. Elution was performed at a constant linear velocity of 300 cm / h. The size exclusion peaks were collected, sterilized, filtered, and stored at low temperature.

[0036] Example 4

[0037] In this embodiment, the original biological fluid processed was human whey sample, which is a body fluid with high lipid content and a total volume of 3L. During the pretreatment stage, after two stages of refrigerated differential centrifugation (300 × g for 10 min and 2000 × g for 20 min), the resulting supernatant was further centrifuged at 10000 × g for 30 min to remove large emulsified lipid clusters. The resulting product was used as the primary supernatant, with a volume of approximately 2.6 L.

[0038] The tangential flow filtration system was equipped with a hydrophilic modified polyethersulfone hollow fiber membrane module with a molecular weight cutoff of 500 kDa, and each fiber had an inner diameter of 1.0 mm. Given that the sample was a complex matrix with high initial viscosity, the wall shear rate was controlled to be constant at 10,000 s⁻¹. -1 The radial transmembrane pressure was maintained at 1.5 psi. A micro-pulse backwash was performed every 20 min for 12 s, with the backwash fluid flow rate being 7% of the initial steady-state forward membrane flux. The feed solution was compressed 30 times to approximately 87 mL to obtain a concentrated exosome buffer. Constant-volume washing and filtration were performed with phosphate-buffered saline, with a washing volume of two wash volumes. After washing and filtration, approximately 87 mL of the exosome sample solution was collected, with an osmotic pressure of 305 mOsm / kg. The entire process was conducted in a 4°C cooled jacket system.

[0039] 2.5 mL of the sample solution was injected into a size exclusion column with an inner diameter of 16 mm, a bed height of 15 cm, and a column bed volume of 30 mL, accounting for 8.3% of the column bed volume. Elution was performed at a constant linear velocity of 80 cm / h. The size exclusion peaks were collected, sterilized, filtered, and stored at low temperature.

[0040] Example 5

[0041] The original biofluid processed in this embodiment was the conditioned medium harvested from human umbilical cord mesenchymal stem cells after expansion culture in a 50 L bioreactor using serum-free medium, with a total volume of 50 L. Approximately 47 L of primary supernatant was obtained after two-stage differential centrifugation at freezing speed.

[0042] The tangential flow filtration system is equipped with a regenerated cellulose hollow fiber membrane module with a molecular weight cutoff of 700 kDa, and each fiber has an inner diameter of 0.5 mm. The wall shear rate is controlled to be constant at 6000 s⁻¹. -1 The radial transmembrane pressure was maintained at 1.2 psi. A micro-pulse backwash was performed every 20 min for 13 s, with the backwash fluid flow rate being 6% of the initial steady-state forward membrane flux. The feed solution was compressed 50 times to approximately 940 mL to obtain the concentrated exosome retention solution. The entire process was conducted at 4°C. Constant-volume washing and filtration were performed with phosphate-buffered saline, with a washing volume of 1.5 washing volumes. After washing and filtration, approximately 940 mL of the exosome loading solution was collected, with an osmotic pressure of 292 mOsm / kg.

[0043] This embodiment uses an industrial-grade, large-bore size exclusion column for purification. The column has an inner diameter of 200 mm, a bed height of 30 cm, and a bed volume of approximately 9.4 L. 470 mL of the sample solution was injected into the column, representing 5% of the bed volume. Elution was performed by pumping phosphate-buffered saline at a constant linear velocity of 200 cm / h, and the size exclusion peaks were collected. All 940 mL of sample solution was injected and eluted in two batches. The products from each batch were combined, sterilized by 0.22 µm filtration, aliquoted, and stored at -80°C.

[0044] Example 6

[0045] This embodiment verifies the feasibility of the process under room temperature operating conditions. The original biofluid used for processing was conditioned medium for human embryonic kidney 293T cell line, with a total volume of 3 L. Approximately 2.85 L of primary supernatant was obtained after two-stage differential centrifugation at freezing speed.

[0046] The tangential flow filtration system operates at room temperature without using the tubing cooling jacket. A hydrophilic modified polyethersulfone hollow fiber membrane module with a molecular weight cutoff of 650 kDa and an inner diameter of 0.75 mm per fiber is installed. The wall shear rate is controlled to be constant at 9000 s⁻¹. -1The radial transmembrane pressure was maintained at 1.0 psi. A micro-pulse backwash was performed every 20 min for 11 s, with the backwash fluid flow rate being 8% of the initial steady-state forward membrane flux. The duration of each full-flow-path processing operation was strictly controlled to within 4 h. The feed solution was compressed 20 times to approximately 143 mL to obtain the exosome concentrate. Constant-volume washing and filtration were performed with phosphate-buffered saline, with a washing volume of 1.5 washing volumes. After washing and filtration, approximately 143 mL of the exosome sample solution was collected, with an osmotic pressure of 298 mOsm / kg.

[0047] 2 mL of the sample solution was injected into a size exclusion column with an inner diameter of 16 mm, a bed height of 15 cm, and a column bed volume of 30 mL, accounting for 6.7% of the column bed volume. Elution was performed at a constant linear velocity of 120 cm / h. The size exclusion peaks were collected, sterilized, filtered, and stored at low temperature.

[0048] Comparative Example 1 The original biological fluid, pretreatment method, and tangential flow filtration system configuration in this comparative example were exactly the same as in Example 1, processing 5 L of human embryonic kidney 293T cell line conditioned medium. The only difference was the elimination of the micro-pulse backwashing operation during concentration and washing; that is, the system did not perform any reverse backwashing during the entire tangential flow filtration cycle. The wall shear rate was kept constant at 8000 s⁻¹. -1 The transmembrane pressure was maintained at 1.0 psi, and the target concentration factor was 50-fold. After approximately 70 minutes of concentration operation, the permeate flow sensor detected a decrease in permeate flux of more than 50% compared to the initial pure water flux, forcing the system to terminate the concentration operation. The actual volume of the concentrated retaining liquid obtained was approximately 230 mL, and the actual concentration factor was only about 21-fold. Subsequent washing, filtration, and size exclusion chromatography procedures were the same as in Example 1.

[0049] Comparative Example 2 This comparative example uses conventional dead-end ultrafiltration instead of the tangential flow filtration concentration step of this invention. The original biological fluid used was the same as in Example 1, namely 5 L of human embryonic kidney 293T cell line conditioned medium. After obtaining the primary supernatant through the same two-stage refrigerated differential centrifugation, the primary supernatant was loaded into a polyethersulfone ultrafiltration centrifuge tube with a molecular weight cutoff of 500 kDa and concentrated by centrifugation at 4000 × g at 4°C, aiming for a 50-fold concentration. Due to the vertical pressure of the feed solution onto the filter membrane surface, the filter membrane became severely clogged when the concentration was about 15 times, and the flux decreased sharply to near zero. The concentration operation was barely completed after multiple filter membrane replacements, with a total operation time exceeding 8 hours. The concentrated retainer was not washed or replaced. 2 mL was directly injected into a size exclusion column of the same specifications as in Example 1, and elution was attempted at a linear velocity of 150 cm / h. Due to the extremely high viscosity of the injection solution, the back pressure of the chromatographic system exceeded the safety threshold after about 10 minutes of operation, forcing the linear velocity to be reduced to 30 cm / h to maintain operation. After collecting the chromatographic peaks of the exclusion zone, performance testing was performed.

[0050] Comparative Example 3 In this comparative example, the constant-volume washing and filtration replacement operation was omitted after the tangential flow filtration concentration step, and the high-concentration retained solution was directly sent to the downstream size exclusion chromatography system. The original biological fluid and pretreatment method were the same as in Example 1. The tangential flow filtration system parameters were also set in the same way as in Example 1, including a wall shear rate of 8000 s⁻¹. -1 The process involved a transmembrane pressure of 1.0 psi and micro-pulse backwashing every 20 min. After compressing the feed solution 50 times to approximately 96 mL to obtain a concentrated retainer, 2 mL was injected directly into the size exclusion column without buffer filtration. Because the osmotic pressure of the unfiltered concentrated retainer was over 460 mOsm / kg and its apparent viscosity was significantly high, the system back pressure rapidly increased after starting elution at a set linear velocity of 150 cm / h. To prevent the resin bed from collapsing, the linear velocity was reduced to 50 cm / h, and severe broadening and tailing of the size exclusion peaks were observed in the elution chromatogram, with significant baseline overlap with subsequent protein peaks.

[0051] Comparative Example 4 This comparative example uses a conventional combination of high transmembrane pressure and low shear rate parameters for tangential flow filtration concentration. The original biofluid and pretreatment method are the same as in Example 1. The tangential flow filtration system is equipped with a hollow fiber membrane module of the same specifications as in Example 1, but the wall shear rate is set to only 2000 s⁻¹. -1The transmembrane pressure was increased to 5.0 psi to accelerate concentration through high driving force, without performing micropulse backwashing. Because the low shear rate was insufficient to generate an effective boundary kinetic energy sweep effect, the combined high transmembrane pressure compacted the polarization layer into a dense filter cake, causing a precipitous decline in membrane flux after approximately 40 minutes of operation. The system could not continue operating after concentration to approximately 18 times. The obtained concentrated retentate was washed and filtered before being fed into a size exclusion column, with subsequent steps identical to those in Example 1.

[0052] The performance test results of the above embodiments and comparative examples are summarized in Tables 1 and 2.

[0053] Table 1. Tangential flow filtration process performance and exosome product quality indicators of each embodiment and comparative example. Table 2. Size exclusion chromatography operating parameters and final purity indices for each embodiment and comparative example. It should be noted that Examples 1 to 4 and Example 6 used laboratory-scale chromatographic columns (16 mm inner diameter, 30 mL bed volume). Their design aimed to verify the influence of key process parameters of the present invention on the purity and recovery rate of exosomes. The single sample loading volume was limited to 5%–10% of the column bed volume, and the entire batch of concentrate required multiple cyclic injections to complete purification. In actual industrial production, linear scaling up to a large-diameter chromatographic column matching the volume of the concentrate can significantly reduce the number of injection batches and shorten the total processing time. Example 5 validates this scaling-up strategy: using an industrial-grade chromatographic column with an inner diameter of 200 mm and a bed volume of 9.4 L, 940 mL of concentrate required only two injections to complete the purification, with the entire batch chromatography taking approximately 80 minutes.

[0054] The data in Tables 1 and 2 clearly identify the decisive impact of each technical feature on the final product quality and process efficiency. Examples 1 to 6, covering the parameter ranges defined in the claims of this invention, consistently achieved exosome recovery rates of over 83%, structural integrity rates of over 95%, and breakthroughs in 10... 9 At the particle / µg level, the particle-to-protein ratio and host protein removal rate both exceeded 99.8%, and surface markers showed strong positive results while negative markers were all below the detection limit. Regarding the trend of parameter gradient changes, Example 3 used 4000 s... -1 The lower limit of the wall shear rate and a mild concentration factor of 10 times, although the concentration factor is the lowest, result in the highest membrane flux retention of 95% and the best exosome structure integrity of 98.5%. Simultaneously, size exclusion chromatography can be safely run at a maximum linear velocity of 300 cm / h, with a single chromatography cycle taking only 28 min. Example 2 uses 12000 s...-1 Despite a slight decrease in membrane flux retention to 88% and structural integrity to 96.1% under extreme concentration conditions, the upper limit of the shear rate and a 100-fold limiting concentration factor still maintained a high recovery rate of 85%, demonstrating that the parameter window of this invention still has sufficient safety margin at the limiting operating point. Example 4, for human whey samples with high lipid content, involved an additional 10000 × g high-speed centrifugation step after two stages of conventional centrifugation to pre-remove emulsified lipid clusters. Combined with a transmembrane pressure control of 1.5 psi, an 83% recovery rate and over 2.1 × 10⁻⁶ were also achieved on this challenging sample. 9 The purity level was measured at particles / µg. Example 5 scaled up the process to 50 L using an industrial-grade large-bore column. The entire chromatographic purification was completed in approximately 80 minutes with only two batches of 940 mL concentrate injected, fully demonstrating the high-efficiency scale-up performance of the method at industrial-grade volumes. Example 6, with the operation time strictly controlled to within 4 hours at room temperature, also yielded satisfactory product quality.

[0055] In stark contrast, in Comparative Example 1, after the micro-pulse backwashing operation was eliminated, the polarized layer on the membrane surface continued to accumulate and solidify due to the lack of periodic physical jet stripping. The permeation flux had decreased by more than half after 70 minutes, forcing the premature termination of the concentration operation. The actual concentration factor reached only 21 times, far from achieving the target concentration ratio. The exosome recovery rate dropped to 68%, and the particle-to-protein ratio plummeted to 8.5 × 10⁻⁶. 8The particle / µg concentration indicates that a large amount of contaminating proteins were not effectively removed due to insufficient concentration, and the detection signal of exosome surface markers in the resulting product was also significantly weakened. This result, conversely, proves the irreplaceable role of the micro-pulse backwashing mechanism in maintaining the long-term stable operation of the membrane separation channel. It should be noted that the actual concentration factor of Comparative Example 1 after being forced to terminate concentration due to membrane fouling was only 21 times, far lower than the 50 times of Example 1. This resulted in differences in the feed volume, contaminating protein content, and rheological state of the feed solution entering the subsequent washing and chromatographic steps compared to Example 1. Therefore, the decrease in the final purity of Comparative Example 1 was actually superimposed by the synergistic factor of incomplete removal of contaminating proteins due to insufficient concentration. However, this insufficient concentration factor is precisely the direct cascade consequence of the sharp decline in membrane flux after the micro-pulse backwashing operation was canceled. It is a chain reaction of the same technical defect in the process chain—the lack of a backwashing mechanism not only causes flux collapse in the membrane separation stage, but its consequences are also amplified downstream along the process flow, ultimately forming multiple superimposed negative effects on product purity. Comparative Example 2, which replaced tangential flow filtration with conventional dead-end ultrafiltration, saw a sharp drop in membrane flux retention to 15%, with severe irreversible clogging of the filter membrane. Multiple membrane replacements were required to barely complete concentration, resulting in a total operation time exceeding 8 hours. More seriously, the exosome structural integrity was only 78.3%, with an extremely broad particle size distribution ranging from 55 to 210 nm and a PDI as high as 0.35, indicating that a large number of vesicles underwent irreversible mechanical rupture and aggregation during high-pressure dead-end filtration. Weak bands were detected by negative markers, suggesting the contamination of organelle debris. The unwashed, high-viscosity concentrate entering the chromatographic system forced a reduction in linear velocity to 30 cm / h due to excessive back pressure, resulting in a single chromatography cycle exceeding 4 hours and extremely low process efficiency. Comparative Example 3, with parameter control identical to Example 1 during the tangential flow filtration stage, maintained a membrane flux retention of 91%, demonstrating that this stage itself operated normally. However, due to the omission of the constant-volume washing and filtration step, the residual components of the culture medium in the concentrated retention solution were not replaced and removed, resulting in an osmotic pressure exceeding 460 mOsm / kg. The high-viscosity feed solution, when pumped into the column, triggered severe osmotic pressure gradient shocks and localized fluid channeling, forcing the linear velocity to decrease from the set 150 cm / h to 50 cm / h. The exclusion zone peaks exhibited severe broadening and tailing, overlapping with the baseline of other protein peaks. The host protein removal rate dropped to 97.1%, and the particle-to-protein ratio was only 5.6 × 10⁻⁶. 8 The particle / µg concentration is significantly lower than that of the embodiments of this invention. This comparison precisely reveals the crucial role of the rheological reconstruction achieved by the constant-volume elution step in the efficient operation of downstream size exclusion chromatography. Comparative Example 4 uses a concentration of 2000 s⁻¹, far below the lower limit of this invention. -1The wall shear rate, coupled with a transmembrane pressure of 5.0 psi, far exceeding the upper limit of this invention, represents the conventional high-pressure, low-shear operating mode in the prior art. The hydrodynamic kinetic energy generated by the low shear rate is insufficient to overcome the van der Waals attraction between protein molecules and the hydrophobic interaction forces on the membrane surface, failing to continuously strip away impurities attempting to deposit and form a gel layer. Simultaneously, the high transmembrane pressure not only compacts the polarization layer into a hard, dense, irreversible filter cake but also crushes fragile exosome vesicles under extreme hydrostatic pressure. The comparative example exhibits a structural integrity of only 81.7%, an extremely broad particle size distribution ranging from 50 to 230 nm, a PDI as high as 0.38, and a particle-to-protein ratio of only 3.8 × 10⁻⁶. 8 The presence of particles / µg and weak bands as negative markers fully confirms the devastating damage to the exosome vesicle structure caused by an unreasonable combination of fluid dynamic parameters.

[0056] A thorough analysis of the significant differences presented in the experimental data from the perspective of microstructure and physicochemical mechanisms reveals that the various embodiments of this invention can maintain extremely high membrane flux stability, vesicle structural integrity, and downstream chromatographic separation efficiency simultaneously under large-volume, high-rate concentration conditions. The fundamental reason lies in the precise boundary fluid dynamics equilibrium established between the strictly defined wall shear rate and the controlled transmembrane pressure. In this equilibrium state, the radial permeation driving force provided by the transmembrane pressure is precisely controlled to the lowest thermodynamic level sufficient to drive radial mass transfer of water molecules and low-molecular-weight proteins. Meanwhile, the kinetic energy corresponding to the axial shear stress generated by the fluid on the inner wall of the hollow fiber is just sufficient to overcome the intermolecular van der Waals attraction and hydrophobic interaction forces at the membrane surface, continuously stripping away and resuspending the macromolecules and nanoparticles carried radially to the membrane interface to the mainstream region at the center of the flow channel. The vector resultant force of the normal compressive stress and axial shear tearing stress borne by the exosome vesicles composed of the phospholipid bilayer within this parameter window is far below their yield limit; therefore, a high structural integrity rate of over 95% was observed in each embodiment. The operating logic of the micro-pulse backwashing mechanism lies in the fact that at each set backwashing cycle node, the forward permeation driving force is first paused to instantly reduce the transmembrane pressure to zero. Subsequently, a small amount of reverse liquid flow pulse is applied, which can establish an effective reverse mass transfer gradient under near-zero baseline pressure difference conditions. This physical jet method peels off and resuspends hydrophobic debris and impurities that have crossed the critical adhesion point from the membrane pore surface. In the time dimension, this blocks the pathway of the polarized layer evolving from the reversible concentration gradient enrichment stage to the irreversible gel solidification phase transition stage. This is the direct reason why the membrane flux in Comparative Example 1 decreased sharply within 70 minutes after this operation was canceled. The core function of the constant volume washing filtration step is to effectively replace the liquid phase matrix and reconstruct the rheological viscosity of the retained liquid, which has exhibited non-Newtonian fluid high viscosity characteristics after high-rate concentration, rather than pursuing the complete removal of the original components. By replacing approximately 63%–86% of the matrix components with isotonic chromatography buffer within an economical operating range of 1 to 2 wash volumes, the osmotic pressure can be effectively controlled from the culture medium level to the target range of 280–320 mOsm / kg. This ensures that the hydrodynamic state of the reconstituted feed solution matches the tolerance limit of the size exclusion chromatography matrix gel skeleton. The experimental data of Comparative Example 3 directly confirms that if this step is skipped, the high-viscosity, high-osmotic-pressure concentrate without rheological adjustment will induce a huge osmotic pressure gradient at the gel micropore interface, causing the resin particles to shrink due to local dehydration. This results in a loss of uniformity in the flow resistance distribution within the column bed, leading to macroscopic fluid channeling effects and peak broadening and tailing, and a sharp deterioration in separation resolution.It is precisely this invention that deeply couples the intermediate feed solution with ideal rheological characteristics output from the upstream anti-fouling concentration process with the high flow rate tolerance of the downstream macroporous rigid agarose size exclusion matrix at the parameter level, enabling the size exclusion chromatography system to operate stably and safely within a wide range of high-efficiency linear velocities of 34~300 cm / h, compressing the chromatography cycle of several hours in traditional processes to tens of minutes, while always maintaining a breakthrough of 10. 9 The extreme particle-to-protein ratio at the particle / µg level and the host protein removal rate of over 99.8% achieve a synergistic balance between high throughput and high separation purity in large-scale exosome purification.

[0057] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for large-scale purification of exosomes based on tangential flow filtration coupled with size exclusion chromatography, characterized in that, Includes the following sequential steps: S1. Pretreatment: The original biological fluid containing exosomes is centrifuged and clarified to remove cells and micron-sized particulate impurities, and a primary supernatant is obtained. S2. Tangential Flow Filtration Concentration: The primary supernatant is introduced into a tangential flow filtration system containing a hollow fiber membrane module for circulation and concentration; the molecular weight cutoff of the hollow fiber membrane module is limited to 500 kDa to 750 kDa; during concentration, the wall shear rate of the liquid on the inner wall of the hollow fiber is controlled to be constant at 4,000 s⁻¹. -1 Up to 12,000 s -1 Between these points, the radial transmembrane pressure is maintained between 0.5 psi and 3.5 psi by adjusting the back pressure on the retentate outlet side; the total volume of the feed solution is compressed by 10 to 100 times to obtain the exosome concentrated retention solution; S3, Washing, Filtration and Physical Regeneration: To the exosome concentrate and retention solution obtained in step S2, continuously add chromatography buffer at the same rate as the osmotic end draining solution for constant volume washing and filtration; the total volume of the washing buffer is 1 to 2 washing volumes; during the concentration in step S2 and the washing and filtration operation in this step, periodically pause forward osmosis and apply micro-pulse backwashing operation from the osmotic side using the pre-filled chromatography buffer as backwashing solution to the hollow fiber membrane cavity to strip the polarization layer on the membrane surface; The exosome loading solution after viscosity reduction and rheological reconstruction was obtained; S4. Rheological adaptive chromatography separation: The exosome loading solution obtained in step S3 is injected into a size exclusion chromatography column filled with macroporous agarose chromatography matrix; the single loading volume is controlled to be 5% to 10% of the size exclusion chromatography column bed volume; then the chromatographic mobile phase is pumped in at a constant linear velocity of 34 cm / h to 300 cm / h for elution, and the chromatographic peak in the first elution size exclusion zone is collected to obtain purified exosomes.

2. The method according to claim 1, characterized in that, In step S2, the hollow fiber membrane module is made of hydrophilic modified polyethersulfone or regenerated cellulose; the hollow fiber membrane module has an open linear channel inside, and the inner diameter of a single capillary fiber is 0.5 mm to 1.0 mm.

3. The method according to claim 1, characterized in that, In step S3, the specific parameters of the micro-pulse backwashing operation are limited as follows: every 20 minutes of continuous concentration or filtration cycle, the system first pauses the forward permeation operation to reduce the transmembrane pressure to near zero, and then performs a momentary reverse backwash lasting 10 to 15 seconds. The applied backwash fluid flow rate is 5% to 10% of the initial steady-state forward membrane flux. The initial steady-state forward membrane flux refers to the steady-state permeation flux value measured at the beginning of system operation under the set transmembrane pressure conditions.

4. The method according to claim 1, characterized in that, In step S3, the chromatography buffer is selected from phosphate buffer saline with a pH of 7.4 or isotonic saline with hydroxyethylpiperazine ethanesulfonic acid; after washing and filtration, the system osmotic pressure of the exosome loading solution is stabilized between 280 mOsm / kg and 320 mOsm / kg.

5. The method according to claim 1, characterized in that, In step S1, the centrifugal clarification process includes multi-stage refrigerated differential centrifugation, performed sequentially: First, centrifuge at 300 × g for 10 minutes at 4°C and collect the first-stage supernatant. Subsequently, the first-stage supernatant was centrifuged at 2,000 × g for 20 minutes at 4°C to collect the primary supernatant.

6. The method according to claim 5, characterized in that, When the original biological fluid is a body fluid with high lipid content, the supernatant obtained by centrifugation at 2,000 × g is further centrifuged at 10,000 × g for 30 minutes, or pre-filtered through a microfiltration membrane with a pore size of 0.45 μm, and the resulting product is used as the primary supernatant.

7. The method according to claim 1, characterized in that, Steps S2 and S3 are performed in a system equipped with a pipe cooling jacket, and the system operating temperature is maintained at 4°C throughout the process; when the system is running at room temperature, the duration of a single full-flow path processing operation for tangential flow filtration must be limited to within 4 hours.

8. The method according to claim 1, characterized in that, When the original biofluid is a complex culture medium with high initial viscosity, in step S2, the radial transmembrane pressure is maintained between 1.5 psi and 3.5 psi.

9. The method according to claim 1, characterized in that, In step S4, when collecting the chromatographic peaks in the exclusion zone, online real-time monitoring is performed using a series of multi-wavelength ultraviolet detectors and multi-angle dynamic light scattering detectors; when the light scattering detector captures a light scattering signal with a particle size distribution concentrated in the range of 50 nm to 150 nm, and simultaneously corresponds to a low ultraviolet absorption characteristic, an automated fraction collector is triggered to collect the peaks.

10. The method according to claim 1, characterized in that, Following step S4, the method further includes a formulation-grade sterile microfiltration step: the collected purified exosomes are combined and pushed through a 0.22 μm pore size polyethersulfone terminal sterile filter membrane under a transmembrane pressure difference of no more than 10 psi, and then aliquoted and transferred to -80°C for deep cryogenic storage.

Images