Nanoparticle tracking analysis (NTA) is a widely used method to characterize extracellular vesicles. This paper highlights NTA experimental parameters and controls plus a uniform method of analysis and characterization of samples and diluents necessary to supplement the guidelines proposed by MISEV2018 and EV-TRACK for reproducibility between laboratories.
Nanoparticle tracking analysis (NTA) has been one of several characterization methods used for extracellular vesicle (EV) research since 2006. Many consider that NTA instruments and their software packages can be easily utilized following minimal training and that size calibration is feasible in-house. As both NTA acquisition and software analysis constitute EV characterization, they are addressed in Minimal Information for Studies of Extracellular Vesicles 2018 (MISEV2018). In addition, they have been monitored by Transparent Reporting and Centralizing Knowledge in Extracellular Vesicle Research (EV-TRACK) to improve the robustness of EV experiments (e.g., minimize experimental variation due to uncontrolled factors).
Despite efforts to encourage the reporting of methods and controls, many published research papers fail to report critical settings needed to reproduce the original NTA observations. Few papers report the NTA characterization of negative controls or diluents, evidently assuming that commercially available products, such as phosphate-buffered saline or ultrapure distilled water, are particulate-free. Similarly, positive controls or size standards are seldom reported by researchers to verify particle sizing. The Stokes-Einstein equation incorporates sample viscosity and temperature variables to determine particle displacement. Reporting the stable laser chamber temperature during the entire sample video collection is, therefore, an essential control measure for accurate replication. The filtration of samples or diluents is also not routinely reported, and if so, the specifics of the filter (manufacturer, membrane material, pore size) and storage conditions are seldom included. The International Society for Extracellular Vesicle (ISEV)'s minimal standards of acceptable experimental detail should include a well-documented NTA protocol for the characterization of EVs. The following experiment provides evidence that an NTA analysis protocol needs to be established by the individual researcher and included in the methods of publications that use NTA characterization as one of the options to fulfill MISEV2018 requirements for single vesicle characterization.
Accurate and repeatable analysis of EVs and other nanometer-scaled particles presents numerous challenges across research and industry. Replication of EV research has been difficult, in part, due to the lack of uniformity in reporting necessary parameters associated with data collection. To address these deficiencies, the ISEV proposed industry guidelines as a minimal set of biochemical, biophysical, and functional standards for EV researchers and published them as a position statement, commonly referred to as MISEV20141. The accelerating pace of EV research required an updated guideline, and the "MISEV2018: a position statement of the ISEV" expanded the MISEV2014 guidelines2. The MISEV2018 paper included tables, outlines of suggested protocols, and steps to follow to document specific EV-associated characterization. As a further measure to facilitate interpretation and replication of experiments, EV-TRACK was developed as a crowd-sourcing knowledgebase (http://evtrack.org) to enable more transparent reporting of EV biology and the methodology used for published results3. Despite these recommendations for standardized reporting of methods, the field continues to suffer regarding replicating and confirming published results.
Fitting with the National Institutes of Health's and National Science Foundation's effort for quality assessment tools, this paper suggests that ISEV requires standardized reporting of methods and details so that data assessment tools might be applied with the goal of replicating results between laboratories. Reporting cell sources, cell culture procedures, and EV isolation methods are important factors to define the qualities of the EV population. Among NTA instruments, factors such as detection settings, the refractive index of carrier fluid, heterogeneous particle populations contributing to polydispersity, lack of standardized reporting requirements, and absent intra- and inter-observer measurement results make NTA comparison between labs difficult or impossible.
In use since 2006, NTA is a popular method for nanoparticle size and concentration determination that is currently used by approximately 80% of EV researchers4. The MISEV2018 Guidelines require two forms of single-vesicle analysis, of which NTA is one of the popular options. NTA continues to be in common use for EV characterization due to its wide accessibility, low cost per sample, and its straightforward founding theory (the Stokes-Einstein equation). EV assessment by NTA generates a particle size distribution and concentration estimate using laser light scattering and Brownian motion analysis, with the lower limit of detection determined by the refractive index of the EV. When using a fluid sample of known viscosity and temperature, the trajectories of the EVs are tracked to determine their mean-square displacement in two dimensions. This allows the particle diffusion coefficient to be calculated and converted into a sphere-equivalent hydrodynamic diameter by a modified Stokes-Einstein equation5,6,7. NTA's particle-to-particle analysis has less interference by agglomerates or larger particles in a heterogeneous population of EVs than other methods of characterization7. While a few larger particles have minimal impact on sizing accuracy, the presence of even minute amounts of large, high light-scattering particles results in a notable reduction in the detection of smaller particles due to reduced software EV detection and tracking8. As a measurement technique, NTA is generally considered not to be biased toward larger particles or aggregates of particles but can resolve multiple-sized populations through individual particle analysis9. Because of the use of light-scattering by particles, one of the limitations of NTA analysis is that any particulate such as dust, plastic, or powder with similar refraction and size attributes compared to EVs cannot be differentiated from actual EVs by this method of characterization.
The NanoSight LM10 (nanoparticle size analyzer) and LM14 (laser module) have been sold since 2006, and although newer models of this instrument have been developed, this particular model is found in many core facilities and is considered a reliable workhorse. Training is needed to properly optimize the NTA settings for high-resolution measurements of size and concentration. The two important settings needed for optimum video recordings are (1) the camera level and (2) the detection threshold. These must be set by the operator based on the sample's characteristics. One of the major constraints of NTA analysis is the recommendation of sample concentrations between 107 and 109 particles/mL, to achieve this sample dilution may be required10. Solutions used for dilution, such as phosphate-buffered saline, 0.15 M saline, or ultrapure water, are rarely free of particles less than 220 µm in size, which may affect the NTA measurements. NTA characterization of the solutions used for dilution should be performed at the same camera level and detection threshold as the nanoparticle samples that are being analyzed.The size and concentration of nanoparticles present in diluents used for EV sample dilutions are seldom included in publications involving NTA analysis of EVs.
This protocol uses NTA analysis of synthetic EV-like liposomes evaluated using selected camera levels, detection thresholds, and mechanical filtering of the samples to analyze the systematic effects of camera level, detection threshold, or sample filtration on the NTA dataset. Liposomes were synthesized as described in Supplemental File S1. Synthetic liposomes were used in this experiment because of their size uniformity, physical characteristics, and stability in storage at 4 °C. Although actual samples of EVs could have been used, the heterogenicity and stability of EVs during storage may have complicated this study and its interpretation. Similarities in the NTA reports from (A) liposomes and (B) EVs indicate that the systematic effects revealed for liposomes in this paper will likely also apply to EV characterization (Figure 1). Together, these findings support the notion that complete reporting of critical software settings and the description of sample processing, such as diluent, dilution, and filtration, impact the reproducibility of NTA data.
The purpose of this paper is to demonstrate that varying the NTA settings (temperature, camera level, and detection threshold) and sample preparation changes the results collected: systematic, significant differences in size and concentration were obtained. As NTA is one of the popular options to fulfill the MISEV2018 characterization specification, these results demonstrate the importance of reporting sample preparation and NTA settings to ensure reproducibility.
Figure 1: Representative NTA reports to compare liposomes to EVs. (A) Liposomes: unfiltered sample characterized on NTA on 12 March 2020. (B) EVs: unfiltered sample characterized on NTA on 26 August 2021. Abbreviations: NTA = Nanoparticle tracking analysis; EVs = extracellular vesicles. Please click here to view a larger version of this figure.
1. General protocol guidelines
2. Preparation of 50 nm and 100 nm size calibration standards
NOTE: See the Table of Materials.
3. Cleaning and assembly of the laser module
4. Flushing procedure for the laser module prior to and between samples
5. Placement of the laser module on the microscope stage
Figure 2: Laser module focus alignment guide. Please click here to view a larger version of this figure.
6. Focusing and positioning of the laser module
NOTE: This must be performed with fluid in the chamber.
7. Loading standards/samples/diluent into the laser module for NTA analysis
8. Validation of calibration
NOTE: It is recommended to validate the module calibration using size standards (see section 2) prior to sample analysis. Routine validation is necessary to ensure accurate measurements. In a multiuser laboratory, individual user adjustments of software configuration settings can inadvertently cause inaccurate data collection. For critical data collection, daily validation is a matter of good laboratory practice. The day-to-day reproducibility of validation needs to be included in the reported results. Typically, calibration is set by the technician and is not adjustable by the individual user unless the user has administrator access. This prevents unauthorized reconfiguration by individual users.
9. Optimizing sample concentration for NTA
NOTE: The screen should contain between 50 and 100 measurable particles when the camera level and sample concentration are adjusted properly. If there is any question about whether a sample has an appropriate particle number, a Quick Measurement can be run on the sample at this point (see steps 9.1 to 9.7). It is used to assess the sample characteristics rapidly prior to longer video captures. The Quick Measurement tab is found within the SOP tab in the bottom middle box.
10. Sample NTA
NOTE: The Standard Measurement tab is within the SOP tab in the bottom middle box and is used for routine sample analysis (see steps 10.1 to 10.12).
11. Re-analysis of the current sample at different detection thresholds
NOTE: Immediately following NTA analysis (step 10), the data can be reanalyzed using different Detection Threshold settings. However, Camera Level cannot be modified following capture.
12. Analysis of archived files
NOTE: If previously analyzed experiments have not been saved or additional analysis needs to be done on these samples, the individual files can be reloaded into the NTA software for additional Detection Threshold evaluations. Camera Level changes cannot be modified following capture.
13. Cleaning and disassembly of the laser module
14. Sample analysis protocol
15. Statistical analysis of NTA results
Table 1 contains the results of the NTA videos for the liposome samples (18 filtered and 18 unfiltered) and a representative DPBS diluent. Comparisons across the two groups were completed regardless of the camera level or detection threshold in this paper. Filtered samples had a mean particle diameter of 108.5 nm, a particle mode of 86.2 nm, and a concentration of 7.4 × 108 particles/mL. In contrast, unfiltered samples had a mean particle diameter of 159.1 nm, a particle mode of 105.7 nm, and a concentration of 7.6 × 108 particles/mL. Mean and mode values for the filtered and unfiltered samples, regardless of the camera level or detection threshold, were statistically significant (p < 0.05). Differences in concentration between the filtered and unfiltered samples, regardless of the camera level or detection threshold, were non-significant (p = 0.86).
Comparisons of Filtered vs Unfiltered Samples | ||||||||||||
Sample | Cam Lev | Det Thr | Mean | Mean %CV | St. Dev. | Conc. × 108 | St. Dev. × 107 | |||||
Filtered | Unfiltered | Filtered | Unfiltered | Filtered | Unfiltered | Filtered | Unfiltered | Filtered | Unfiltered | |||
Liposome | 12 | 3 | 138.3 | 162.1 | 55.5 | 47.5 | 76.7 | 77 | 3.2 | 6.48 | 1.74 | 3.16 |
Liposome | 12 | 2 | 126.3 | 153.7 | 58.7 | 49.8 | 74.2 | 76.5 | 4.94 | 8.23 | 2.74 | 3.5 |
Liposome | 12 | 5 | 155 | 172.2 | 51.8 | 45.6 | 80.3 | 78.6 | 1.91 | 5.05 | 1.02 | 2.9 |
Liposome | 13 | 3 | 102.7 | 169 | 43.5 | 51.3 | 44.7 | 86.7 | 6.72 | 7.75 | 1.91 | 1.17 |
Liposome | 13 | 2 | 95.8 | 161.4 | 43.3 | 52.7 | 41.5 | 85 | 10.3 | 9.7 | 1.61 | 1.83 |
Liposome | 13 | 5 | 110.7 | 176.1 | 42.5 | 47.1 | 47 | 83 | 4.16 | 6.28 | 1.83 | 1.07 |
Liposome | 14 | 3 | 98 | 147.6 | 42.4 | 43.8 | 41.6 | 64.6 | 10.6 | 8.56 | 2.4 | 1.66 |
Liposome | 14 | 2 | 92.9 | 142.1 | 42.6 | 45.2 | 39.6 | 64.3 | 14.9 | 10.7 | 2.54 | 1.83 |
Liposome | 14 | 5 | 103.8 | 153.8 | 41.1 | 42.6 | 42.7 | 65.5 | 7.42 | 7.02 | 2.37 | 1.51 |
Liposome | 12 | 3 | 105.6 | 179.4 | 22.2 | 46.7 | 23.4 | 83.7 | 5.2 | 5.81 | 1.06 | 4.28 |
Liposome | 12 | 2 | 100.3 | 170.8 | 24.3 | 49.1 | 24.4 | 83.8 | 7.76 | 7.39 | 1.61 | 4.41 |
Liposome | 12 | 5 | 112 | 187.2 | 20.4 | 42.9 | 22.8 | 80.4 | 3.27 | 4.68 | 0.815 | 3.93 |
Liposome | 13 | 3 | 99.8 | 153.3 | 23.4 | 52.1 | 23.4 | 79.8 | 7.19 | 7.34 | 3.37 | 1.5 |
Liposome | 13 | 2 | 94.3 | 143.6 | 25.8 | 53.2 | 24.3 | 76.4 | 10.8 | 9.53 | 4.3 | 2.46 |
Liposome | 13 | 5 | 106.8 | 162 | 21.3 | 49.4 | 22.7 | 80 | 4.64 | 5.76 | 2.63 | 1.01 |
Liposome | 14 | 3 | 103.4 | 142.3 | 31.7 | 49.6 | 32.8 | 70.6 | 9.91 | 8.66 | 3.29 | 12.5 |
Liposome | 14 | 2 | 97.8 | 136 | 33.3 | 51.4 | 32.6 | 69.9 | 13.8 | 11.5 | 2.98 | 15.7 |
Liposome | 14 | 5 | 109.5 | 151.4 | 31.1 | 49.5 | 34 | 74.9 | 7.27 | 6.76 | 3.42 | 10.1 |
Average | 108.5 | 159.1 | 36.4 | 48.3 | 40.5 | 76.7 | 7.4 | 7.6 | 2.3 | 4.1 |
Table 1: Data table of collected values, standard deviation, and percent coefficient of variation for filtered and unfiltered samples. Abbreviations: Cam Lev = camera level; Det Thr = detection threshold; CV = coefficient of variation; St. Dev. = standard deviation; Conc. = concentration.
When the liposome sample results were parsed by detection threshold (2, 3, 5) and camera level (12, 13, 14), the results were not significant (Figure 3). It should be noted that there were only 3 evaluations (n = 3) at each of these individual levels. This small sample size at each camera level and detection threshold likely contributed to the lack of individual comparisons being significant. However, when detection threshold (2, 3, 5) samples were evaluated regardless of camera level across the filtered and unfiltered samples (n = 3), both the mean size (Figure 3A) and mode size (Figure 3B) were significantly (p < 0.05) different. In contrast, differences between filtered and unfiltered sample concentrations (Figure 3C) were not significantly different.
Figure 3: Effects of camera level and detection threshold on measured particle size and concentration of filtered and unfiltered samples. Mean (A) and Mode (B) particle size at combined camera levels 12, 13, 14 as the detection threshold was increased from 2 to 3 to 5 (n = 3), showing a significant decrease in particle sizes of the filtered samples. Particle concentrations (C) at combined camera levels 12, 13, 14 as the detection threshold was increased from 2 to 3 to 5) (n = 3), showing a concentration decrease as the detection threshold increased with no significant difference between filtered and unfiltered samples. Mean (D) and Mode (E) particle size at combined detection thresholds as the camera level increased from 12 to 14 (n = 3), showing a decrease in particle size of the filtered samples. Particle concentrations (F) at combined detection thresholds as camera levels increased from 12 to 14 (n = 3), showing a concentration increase as the camera level increases with no significant differences between filtered and unfiltered samples. Abbreviation: DT = detection threshold. Please click here to view a larger version of this figure.
When the 3 camera levels (12, 13, 14) were evaluated regardless of detection threshold level (n = 3), both the mean size (Figure 3D) and mode size (Figure 3E) increased in the filtered samples. Sample concentrations showed a tendency to increase as the camera level increased from 12 to 14 (Figure 3F). The differences between filtered and unfiltered sample concentrations evaluated at different camera levels were not significantly different.
Supplemental File 1: Please click here to download this File.
There are several methods available to estimate the size and concentration of nanoparticles11. These include ensemble methods that generate a size estimate from a population, including dynamic light scattering (DLS), centrifugal sedimentation, and single-particle level analysis-electron microscopy, NTA, atomic force microscopy, and tunable resistive pulse sensing. Of these, DLS and NTA are widely used, nondestructive size and concentration measurement methods, based on Brownian movement in an ideal medium. DLS relies on the scattering of light and the intensity is proportional to the square of the particle volume. Thus, DLS is more sensitive than NTA to the presence of large particles, aggregates, or polydisperse populations.
NTA calculates the diffusion coefficient from the path length of individual particles measured in successive video frames. The main limitation of NTA is the narrow range of particle concentration that it can evaluate compared to DLS and other measurement methods, such that individual particle pathlengths must fall within the diffraction limit of the microscope and the tracking software's capabilities. As DLS and NTA depend on Brownian movement, both can be expected to show good size agreement in monodispersed populations; they diverge when evaluating polydisperse populations and those with aggregates. The latter renders DLS useless and increases the NTA particle size estimate significantly12. NTA's best-known limitation is that it requires much lower particle concentration (or greater dilution) than other measurement methods. Despite this, NTA characterization is popular in nanomaterials research. Because NTA size and concentration estimates depend on a more diluted population, with defined temperature, video capture settings, including recording length, camera level, detection threshold, and sample dilution to be highly reproducible, this paper focuses on the need for reporting these to generate reproducible results.
This paper shows that using a standardized protocol enabled replication of results and that utilization of positive controls, such as size standards, provides information about the machine's calibration. Furthermore, these results indicated the importance of reporting laser module chamber temperature, camera levels, detection threshold, and filtration (filter type and size). In contrast, laser module chamber temperature, diluent, and dilution factor are equally important for accurate and reproducible results. Although neither MISEV2018 nor EV-TRACK specifically recommends the inclusion of this information, we suggest that the inclusion of these details enables independent confirmation of published results and adds robustness to the experimental design.
Limitations of using latex size calibration standards for NTA calibration in EV analysis are acknowledged and include the known refractive index differences compared to lipid bi-layer nanoparticles of similar size. In this paper, latex beads were used to confirm machine calibration prior to measurements and not to determine the limits of detection. The liposomes have a membrane similar to those of naturally occurring EVs, and the refractive index will be likewise representative of EVs. The size standards, as well as the liposome samples, are monodispersed populations; therefore, their size distribution will follow a Gaussian or log-normal distribution. Natural EVs are polydisperse, and their size distribution will follow a power-law function13.
Historically, publications using NTA characterization inconsistently report necessary details to duplicate the research results. The ability to reproduce NTA data relies on the ability to duplicate the settings used to capture the original data. Without this information, the reproduction of experimental results using NTA will be extremely difficult. With rigorous adherence to a set protocol and publication of the setting parameters used with the NTA, accurate replication of results can be attained. The following recommendations are made to improve the consistency of nanoparticle characterization of size, concentration, composition, and purity using a nanoparticle size analyzer.
First, always check the calibration of the nanoparticle size analyzer using appropriate size standards, such as latex size standards. This should be done on a regular basis and recorded in the instrument log and prior to the analysis of critical samples. Second, all adjustable parameters, such as laser module chamber temperature, camera levels, and detection thresholds, should be recorded for each sample in the Sample Log file, as should the dilutions and diluents used. These parameters should be reported as they are operator-dependent and impact NTA measurements. Third, diluents used for sample dilution need to be characterized for nanoparticle content and reported. The diluents used for individual nanoparticle samples will need to be evaluated using the same camera level and detection threshold settings as those used for the diluted sample. Fourth, syringe filters should be flushed with two times the dead space volume prior to data collection or sample preparation steps to flush the numerous particulates remaining from the manufacturing process. Fifth, the concentration of the nanoparticles within the sample should be adjusted to within the suggested optimum 1.0 × 107 to 1.0 × 109 per mL.
Acknowledging the above-described limitations in this study, we show that both the size and concentration values obtained by NTA can be affected by NTA parameters, such as camera levels and detection thresholds, and that the size, but not the concentration, can be affected by sample preparation. This drives home the critical importance of reporting these parameters in nanomaterial and EV literature, enabling the production of robust, reproducible literature so that we can systematically investigate the impact of EV source, isolation, and other experimental variables.
The authors have nothing to disclose.
The work was supported by the state of Kansas to the Midwest Institute for Comparative Stem Cell Biology (MICSCB), the Johnson Cancer Research Center to MLW and NIH R21AG066488 to LKC. OLS received GRA support from the MICSCB. The authors thank Dr. Santosh Aryal for providing the liposomes used in this project and the members of the Weiss and Christenson laboratories for helpful conversations and feedback. Dr. Hong He is thanked for technical support. MLW thanks Betti Goren Weiss for her support and counsel.
Automatic Pipetter | |||
Centrifuge Tubes, Conical, Nunc 15 mL | Thermo Sci. | 339650 | |
Kimwipes | |||
Lens Cleaner | |||
Lens Paper | |||
NanoSight LM-10 | Malvern Panalytical | ||
NanoSight LM-14 Laser Module | Malvern Panalytical | ||
Nanosight NTA Software Ver. 3.2 | Malvern Panalytical | ||
Paper Towels | |||
Pipette Tips, 1-200 µL, Filtered, Sterile, Low Binding | BioExpress | P -3243-200X | |
Pipette Tips, 50-1,000 µL, Filtered, Sterile | BioExpress | P-3243-1250 | |
Saline, Dulbecco's Phosphate Buffered (No Ca or Mg) | Gibco | 14190-144 | |
Standards, Latex Transfer- 100 nm (3 mL) | Malvern | NTA4088 | |
Standards, Latex Transfer- 50 nm (3 mL) | Malvern | NTA4087 | |
Syringe Filter, 33 mm, .22 µm, MCE, Sterile | Fisher brand | 09-720-004 | |
Syringe, TB, 1 mL, slip tip | Becton Dickinson | 309659 | |
Waste fluid container |