About the characterization of EV
Estimating the number of EVs, determining their presence, and evaluating the proportion and impact of non EV components in EV preparation all require characterization of EVs. The characterization of EVs faces challenges such as small particle size, size heterogeneity, molecular heterogeneity, lack of universal EV recognition methods, and non EV specificity of many measurement techniques.No single measurement or method can meet the requirements for all EV characterizations, so it is recommended to use orthogonal methods (i.e. methods without the same measurement limitations) to validate each otherThat is to say, the characterization parameters of EV are recommended as key quality attributes (CQA) to use two or more different principles to detect the same indicator.
The overall composition of EVs (the content and proportion of proteins, lipids, nucleic acids, and other biomolecules) varies depending on the source of EVs. Although measurements of these individual molecular categories can be used to estimate EV abundance, these values may not necessarily be entirely correlated with EV concentration, and there is no universality among source materials; Therefore,They should not be used as a measure of EV concentration.Just as there is no single molecule detection that can quantify all EVs, there is also no universal molecular marker for EVs or EV subtypes. Specific markers must be selected based on evidence of source specificity and type specificity. At present,There is no known universal marker that can identify all EVs, regardless of their source.The universality of these markers for so-called exosomes is still unclear or not accepted. Please note that the affinity based approach involving the four transmembrane proteins CD9, CD63, and CD81 is not specific for extracellular vesicles as EV subtypes; The EV population enriched with antibodies targeting one of these four transmembrane proteins cannot fully cover all molecular compositions. In addition, not all EVs express these protein markers, so four transmembrane protein enrichments cannot capture all EVs.Orthogonal method for detecting the same characteristic parameter(Methods based on more than one principle)It is unlikely to have the same deviation; For example, simultaneously deriving and detecting the diameter of EVs through both optical and non optical methods.The use of orthogonal methods to characterize EV samples is also crucial for demonstrating the independence of co isolates from biomarkers or functional discoveries.Due to the fact that many EV characterization methods are not specific to EVs or cannot detect all EVs, it is necessary to transparently disclose the methods and results to ensure the reproducibility of EV data.
1、 Quantification of particle concentration
The quantity and volume measurement of EVs together define the number concentration (in particles/mL), which is a widely reported and used indicator. However, it is often unreliable as many technologies lack specificity for EVs and sensitivity to all EVs.The ISEV Strict and Standardized EV Reference Materials Working Group recently outlined the considerations for measurement techniques and the challenges faced by the field in developing and reporting well characterized EV reference materials for traceable measurement. A key highlight of this work is the need to disclose the LOD of the detection method, allowing others to validate the results regardless of sensitivity limitations.The reliability of EV concentration measurement can be improved by using orthogonal method，Each method has a definite LOD, such as detecting light scattering intensity(such as NTA)Fluorescence intensity(such as FCM)And physical size(such as RPS),Because orthogonal methods do not have the same measurement limitationsFor example, resistance pulse induction (RPS) technology can use particle size to represent LOD. For optical techniques such as flow cytometry, LOD can be reported using diameters derived from light scattering optical models or equivalent soluble fluorophore (MESF) molecules derived from fluorescence intensity. These methods can generate consistent data across different instruments and sensitivities. Due to the large number of variables involved in particle detectionAt present, there are no methods for nanoparticle tracking analysis (NTA) or dynamic light scattering (DLS))Or imaging flow cytometry to obtain traceable LOD.For technologies that cannot distinguish EVs from other potential co separates or pollutants, regardless of the upstream separation steps, it is recommended to report the concentration as "particle concentration or EP concentration" rather than "EV concentration".
2、 Quantification of particle size
The measurement of EV size (in nanoradius or diameter units) depends on assumptions such as sphericity or mobility, and the output results may be influenced by upstream variables. Common high-throughput methods include flow cytometry NTA、RPS、 Both multi angle light scattering and dynamic light scattering (DLS) assume that EVs are spherical. Although "size" and "diameter" are often used interchangeably between measurement methods, their derivation can also lead to consistent differences between measurement techniques. For example,Technologies that rely on particle Brownian motion, such as NTA or DLS, measure fluid dynamic diameter, which can lead to overestimation of particle size compared to imaging methods such as cryo electron microscopy.However, there are few methods that can accurately measure the size of EVs within the entire possible diameter range (from tens of nanometers to micrometers). For example, although high-resolution imaging of cryo electron microscopy is one of the most accurate methods, its detection flux is relatively low, and many larger EVs often have small quantities that may not be quantifiable. The ability to quantify low contrast EVs below 100 nm may also be a limiting factor.
As more and more researchers use specialized single particle analysis techniques with higher sensitivity, it is becoming increasingly clear that many EV formulations exhibit asymmetric right skewed distributions, such as log normal distributions, with most EVs having diameters<100 nm. Most single particle analysis techniques are unable to analyze the entire EV population, thereforeResearchers should share the detected EV diameter distribution，Not just summary indicators such as average size, concentration size, or median sizeThese indicators are easily skewed due to LOD and asymmetric size distribution. Please noteFitting size statistics, such as measuring low refractive index particles through NTA, may be closer to the instrument's LOD than the true modal diameter of the EV population.The use of software with proprietary algorithms to determine particle size may also result in differences between software versions or platforms, therefore software and its versions should also be provided.The technique of deriving size from the assumption of refractive index may vary due to different sample compositions and loadsThe size derived from fluorescent probes (such as membrane embedded dyes) may vary due to different dye embeddings based on different membrane lipid compositions. For technologies that cannot distinguish between EV and co separated substances/pollutants, regardless of the upstream separation steps, it is recommended to describe the diameter report as "particle" or "EP" diameter instead of "EV diameter".
3、 Precautions for EV characterization technology specific reports
With the use of EV detection experiments and instruments, as well as the expansion of expert knowledge, the determination of the scale of relevant characterization reports is also constantly increasing to ensure the reliability and repeatability of data. The MISEV2023 guide provides a detailed list of precautions for minimum experiments and instrument specific reports. Regardless of the experimental design, these are usually applicable. The technologies listed here are not exhaustive, and many detection techniques are currently under development or actively researched. However, the listed technologies are commercially available and have published literature from multiple researchers. (Due to space limitations, the author only lists descriptions and main viewpoints of several commonly used EV particle size and concentration characterization tools.)Flow cytometry based on microbeadsISEV suggests that the control group should include isotype antibodies as detection antibodies, or isotype conjugated capture microbeads, as well as capture microbeads containing only detection antibodies (for antibody coated capture beads); Multiple EV sample concentrations should be used to demonstrate signal titration; If preparing microbeads, their reagents and chemical composition should be disclosed; The catalog and batch number of commercial capture microbead reagents should also be disclosed; Publish the median fluorescence intensity value of standardized microbeads; Publish data and median fluorescence intensity statistics of equivalent soluble fluorophore (MESF) molecules from singlet delineated microbeads (similar to single EV flow cytometry); A complete and detailed methodology should be published.In 2023, the Three Society Working Group (EV Flow Cytometry Working Group) consisting of the International Society for Extracellular Vesicles (ISEV), the International Society for the Advancement of Cell Counting, and the International Society for Thrombosis and Hemostasis initiated the publication of the "Outline of Single EV Flow Cytometry" to comprehensively address the considerations for developing single EV flow cytometry detection.The calibration of sample volume, fluorescence, and light scattering parameters is crucial for interpreting and replicating single EV flow cytometry results. If using a single EV flow cytometer to report particle concentration, please determine the upper and lower limits of LOD to allow for the use of orthogonal methods or others to replicate and validate the results.At present, imaging cell analyzers use dynamic triggering methods, which make it difficult to define and standardize the determination of lower LOD.DLS， Also known as Photon Correlation Spectroscopy (PCS) and Quasi Elastic Light Scattering (QELS), it is a technique that can determine the hydrodynamic diameter of sufficiently monodisperse particles in a diluted aqueous solution dispersion. DLS measures the autocorrelation function of laser intensity scattered by multiple particles in a solution. The autocorrelation function carries information about the particle diffusion coefficient, which is related to the hydrodynamic diameter through the Stokes Einstein Brownian motion theory.Multiple algorithms can be used to derive diffusion coefficients from the measured autocorrelation function. The most common method is cumulative analysis, butIt assumes that the samples have a monodisperse size distribution, which EV samples do not possessFor the multi dispersed size distribution of EV samples, the derivation of the coefficient distribution of the autocorrelation function for diffusion becomes an ill posed mathematical problem. This meansDLS should not be used to determine the quantitative properties of EV samplesUnless DLS is applied to a certain part of EV with monodisperse size. On the other hand, DLS can be used to qualitatively confirm the presence of submicron particles and possible aggregates in EV samples.Various electron microscopes (EM) are one of the few technologies capable of detecting EVs (regardless of size). However, the flux of electron microscopy also means that larger EVs are statistically underestimated compared to smaller EVs. SEM, TEM, and Cryo EM are high-resolution EV characterization methods that do not require switching between each other or can provide images of comparable quality. For example,Cryo EM can clearly display the lipid bilayer structure, maintaining the EV morphology better than TEM under dehydration conditions for sample fixation, and may be more accurate in quantifying (size),Because all particles within a certain volume can be imaged, not just those adhered to the surface of the carrier network.The standardization research on EM methods to determine low reporting requirements is very limited. For TEM, three main parameters should be disclosed: fixation, adsorption, and negative staining methods. Fixation includes: the fixative used, its concentration, and incubation time; Adsorption includes the carrier material, carrier size, film type, coating, incubation time, and cleaning details; The detailed information of negative staining should include substrate, concentration, and incubation time. And low magnification and high magnification images, as well as the selection criteria for images, should be disclosed.Nano particle tracking analysisNTA， It is also considered as a single particle tracking technique, which is widely used in the EV field as an optical technology for estimating particle size and concentration. NTA typically uses algorithms to reduce changes in diameter distribution, by measuring the diffusion coefficient of particles to obtain the hydrodynamic diameter. It is worth noting that,The FTLA algorithm used on some NTA devices is developed to better represent monodisperse mixtures, while EV is not, which may result in multiple artificial fitting distributionsMoreover, there is currently no method to determine or disclose the LOD set by NTA. Many scholars have conducted several standardized studies to compare the results between users and instruments.The use of NTA to measure the diameter distribution and concentration of complex biological fluid samples should be interpreted with cautionBecause it counts co isolates such as lipoproteins and large protein complexes, andNTA finds it difficult to quantify EVs with diameters exceeding a few hundred nanometers.For NTA's published information, it should include instrument model, camera type, camera settings, laser wavelength, laser power, software version, analysis settings, and particles per frame. If known, thenThe algorithm used to generate the diameter distribution should be disclosed, as different results may be generated depending on the algorithm used.In the case of light scattering or fluorescence detection mode, it is recommended to use blank buffer as a control. For fluorescent NTA, report the total number of particles in light scattering mode and the number of labeled particles in fluorescence mode, as well as the label removal method and buffer/reagent control to evaluate the false color of the label.If a fluid sampling device is used, its settings should also be disclosed.Resistance pulse inductionThis is the first time that MISEV has recommended resistance pulse induction (RPS) technology as a commonly used EV particle characterization method alongside flow based methods (including bead based flow and single EV flow) and microscope based methods (including atomic force microscopy, electron microscopy, DLS, NTA, super-resolution microscopy, etc.). It is also one of the few non optical EV characterization techniques.RPS utilizes the Coulter principle to determine the concentration and diameter of particles, and some devices also have zeta potential functionality. At present, the diameter LOD of RPS is as low as about 50 nm. According to literature reports,RPSThe measurement results are indeed consistent withTEMThe data has very high consistency.When reporting RPS data, it is recommended to disclose the instrument model, aperture, calibrated bead diameter and source, as well as software version. As previously mentioned (quantification of particle size), for EV data,It is recommended to publish the diameter distribution of RPS instead of a single diameter statistical data. And it is recommended to include a control of blank buffer to identify the background, as well as samples cracked with the same concentration of detergent to determine labeling events. Due to the susceptibility of RPS technology to clogging by larger particles, pre analysis steps such as centrifugation or filtration can be used to remove larger particles. As these methods may alter the EV population being analyzed and affect comparisons with orthogonal methods, any pre-processing procedures for analysis should be clearly stated.Detailed information on protein enrichment and quantification is required; If possible, positive and negative controls for antigens should be included; If it is claimed that proteins are related to EVs, purity measurements of EV formulations need to be included. Report all details of standardized sample loading, gel electrophoresis, membrane transfer method, probe and imaging/analysis (including but not limited to antibody information, sample denaturation and reduction conditions, membrane transfer method, membrane type, buffer solution, imaging equipment and parameters); Provide all uncut images of WB (as supplementary information if published in a journal).This guide discusses different methods for EV characterization in the above chapters and provides specific recommendations. Regardless of the method used, the overall recommendations for characterization are summarized as follows:
- Each EV formulation should be defined by quantitative measurements of the EV source (e.g. number of secreting cells, volume of biological fluids, mass of tissue, etc.).
- The abundance of EVs should be estimated, including particle count, protein and/or lipid content.
- EV formulations should be tested for the presence of components related to EV subtypes or EVs, depending on the specificity required for experimental purposes.
- Determine the degree of presence of non vesicular and co segregated components.
- When using quantitative indicators to characterize EV, the detection limit (LOD) of the instrument or method should be provided.
