Many cells shed small vesicles in a regulated way; such vesicles have a key role in intercellular communication. In general there are three main types of vesicles: apoptotic bodies (500 nm–3 μm in diameter) 1 released by cells undergoing apoptosis, microvesicles (100 nm–1 μm), which directly bud from the plasma membrane, 2 and nanovesicles (30 nm–100 nm), which include exosomes released via exocytosis from multivesicular bodies of the endosome. 3 All are involved in cell signaling. The properties of cellular vesicles have been reviewed extensively elsewhere, 3-5 Briefly, (i) they carry diverse membrane and cytosolic proteins as well as messenger and microRNAs; (ii) they can affect the physiology of their target cells in various ways, from inducing intracellular signaling following binding to receptors, to conferring new properties after the acquisition of new receptors, enzymes, or genetic material by fusion or endocytosis 3-5 ; (iii) they participate in physiological processes including hemostasis and thrombosis, inflammation, immune interactions, and angiogenesis 4,6 ; and (iv) circulating levels are elevated in various disorders, including atherosclerosis and coronary artery disease, pre-eclampsia, hematological and inflammatory diseases, diabetes, and cancer. 4 They may therefore be useful as prognostic and diagnostic biomarkers for early detection of a wide variety of diseases and have a potential role in monitoring treatment. Relevant research is constrained by the limitations of current methods of measurement. 7 Electron microscopy demonstrates the presence of microvesicles and exosomes in ultracentrifuge pellets of biological fluids (e.g., cell culture medium, plasma, or urine) but either is not quantitative and requires extensive sample preparation. The composition of pelleted material can be determined by western blotting, but this gives no information on the number of particles or their sizes. Enzyme-linked immunosorbent assays do not capture all the vesicles present, cannot discriminate between microvesicles and exosomes, and also detect soluble antigens. Conventional flow cytometry, using analogue or first-generation digital instruments, is unable to detect vesicles of 2 = K B T t s 3 π η d h where 2 is the mean squared displacement, KB is Boltzmann's constant, T is the temperature of the solvent in Kelvin, ts is the sampling time (i.e., 1/30 fpsec = 33 msec), η is the viscosity, and dh is the hydrodynamic diameter. A range of parameters can be adjusted both in video capture (such as camera gain and shutter speed) and in analysis (such as filter settings, background subtraction, minimum required track length, and frame-to-frame search area), thus allowing the user to optimize particle identification and tracking for a particular sample. The range of sizes that can be analyzed by NTA depends on the particle type—that is, whether it has a high (e.g., colloidal gold) or a low refractive index (e.g., cellular vesicles). The lower limit is a function of the signal-to-noise ratio of the image and is thus strongly affected by the amount of light scattered. The particles in consideration here (i.e., at the limit of detection of the system) are in the Rayleigh scattering regime. In this regime the amount of light scattering is given by: σ s = 2 π 5 3 d 6 λ 4 ( n 2 − 1 n 2 + 2 ) 2 where d is the particle diameter, λ is the wavelength of light, and n is the ratio of the particle refractive index to the solvent refractive index. Generally speaking, cellular vesicles have a low refractive index, and the smallest detectable size using the NTA system is in the order of 50 nm. At ∼1 μm the brownian motion of a particle becomes too limited to track accurately. To demonstrate the ability of NTA to discriminate particles of different sizes in a polydisperse sample, a 5:1 mixture of 100 nm and 300 nm polystyrene beads (NIST beads; Thermo Scientific, Fremont, California) was analyzed. To demonstrate how the particle concentration can be estimated, 100 nm beads at a range of concentrations between 2 × 108 and 20 × 108 particles per milliliter were analyzed. Placental vesicles, stored in PBS at −80°C, were thawed at room temperature (18−25°C) before NTA analysis. Each sample was diluted in PBS over a range of concentrations between 2 × 108 and 8 × 108 vesicles per milliliter. The samples were mixed before introduction into the sample chamber and a video recording, typically 1 minute, initiated. A combination of high shutter speed and gain followed by manual focusing enables optimum visualization of a maximum number of vesicles. To accurately track the vesicles they must be visualized as single points of light. The NTA software is unable to effectively track very large vesicles or those with confounding Newton rings. In this case the shutter speed and gain are reduced accordingly (see Supplementary Videos S1-S3, which can be found in the online version of this article). For highly polydisperse samples, multiple analyses at a range of settings may be necessary. The samples were advanced between each recording to perform replicate measurements. NTA post-acquistion settings were optimized and kept constant between samples, and each video was then analyzed to give the mean, mode, and median vesicle size together with an estimate of the concentration. An Excel spreadsheet (Microsoft Corp., Redmond, Washington) was also automatically generated, showing the concentration at each vesicle size. Immunofluorescence labeling with NDOG2 antibody conjugated to quantum dots Quantum dots were conjugated to NDOG2 antibody with a Qdot 605 Antibody Conjugation Kit (Invitrogen, Paisley, United Kingdom) coated with polyethylene glycol amine (molecular weight 2000) according to the manufacturer's instructions. Briefly, quantum dots were activated with the cross-linker 4-(maleimidomethyl)-1-cyclohexanecarboxylic acid N-hydroxysuccinimide ester (SMCC), yielding a maleimide-nanocrystal surface. Excess SMCC was removed by size exclusion chromatography. The antibody was then reduced by dithiothreitol to expose free sulfhydryl groups, and excess dithiothreitol was removed by size exclusion chromatography. The activated quantum dots were covalently coupled with reduced antibody and the reaction quenched with mercaptoethanol. Conjugates were concentrated by ultrafiltration and purified by size exclusion chromatography. Mouse immunoglobulin G was labeled in the same way for use as the isotype control. Placental vesicle suspension (see above) was diluted 1:100 and NDOG2-conjugated quantum dots were added to give a final quantum dot concentration of 10 nM. The molar ratio of antibody to quantum dots was 3.5:1. After a 15 minute incubation at room temperature, the labeled vesicles were diluted in PBS for analysis by NTA. Flow cytometry Placental vesicles were analyzed using a Becton Dickinson LSRII flow cytometer (Becton Dickinson, Oxford, United Kingdom) equipped with a 405 nm (violet) and 488 nm (blue) laser. A vesicle gate was determined using calibration microspheres of various sizes (200 nm, 290 nm, 390 nm, 590 nm; Duke Scientific Corporation, Palo Alto, California) and 1 μm Fluoresbrite Plain YG beads (Polysciences Inc, Warrington, Pennsylvania). Using a side-scatter (SSC) threshold of 200 arbitrary units the lower sensitivity of the instrument was established and the SSC and forward scatter (FSC) voltages were set. There is overlap in the 200 nm bead population and the instrument noise (determined by running PBS filtered through a 100 nm filter). Therefore, a gate was set to include vesicles of 300 nm-1 μm in diameter. Vesicles were enumerated by running Becton Dickinson TruCount Tubes (three tubes per experiment) to establish the flow rate of the system. All samples were run on the “LO” setting with a flow rate of 12 μL per minute. Placental vesicles were diluted 1:50 in PBS and incubated with 10 μL of FcR receptor block at 4°C for 10 minutes. They were then incubated with NDOG2-Qdot605 or IgG1-Qdot605 at 10 μg/mL and incubated for 15 minutes at room temperature. Samples were then diluted to 1 mL with PBS and analyzed by flow cytometry. Fluorescence NTA The NanoSight NS500 instrument (NanoSight Ltd.) is a development of the NanoSight LM10 that allows the detection of microvesicles and nanovesicles labeled with stable fluorophores. It uses a 405 nm (violet) laser diode to excite suitable fluorophores whose fluorescence can then be determined using a matched 430 nm long-pass filter. Measurements are made in fluorescence mode with the long-pass filter in place so that only fluorescently emitted light is measured. The fluorescent particles are individually tracked in real time, from which labeled particle size and concentration can be determined. Under light scatter mode, the total number of particles can be measured and subsequently compared to the concentration of labeled particles. The calibration beads used were Nanosphere Size Standards 200 nm [1.7% coefficient of variation (CV); Thermo Scientific] and Fluoresbrite Plain YG 100 nm Microspheres (5.8% CV) (Polysciences Inc.). Labeling of plasma cellular vesicles with fluorescent QTracker dye Platelet-poor plasma was prepared from blood collected into 0.106 mM trisodium citrate, by double centrifugation at 2500 g for 15 minutes. Cellular vesicles were either labeled directly in the plasma or after ultracentifugation of platelet-poor plasma at 100,000 g for 1 hour at 4°C. Vesicles were labeled using the QTracker cell labelling kit (Invitrogen, Carlsbad, California), which uses a targeting peptide to deliver QDot nanocrystals across the plasma membrane into the cytoplasm. 15 Briefly, 2 μL of a 0.625 nM solution of QTracker quantum dots were added to 200 μL of plasma or plasma vesicle pellet resuspended in PBS and incubated at 37°C for 1 hour. NTA analysis was performed using the NanoSight NS500 instrument as described above. Results NTA measurement of particle size and concentration Figure 2, A shows that NTA successfully addresses one of the key problems with DLS (i.e., polydispersity), in that it can resolve and accurately measure different-size particles (here 100 nm and 300 nm polystyrene beads) simultaneously within the same solution. Although the distributions partly overlap, the two peaks of each bead subtype at 100 nm and 300 nm peaks are clearly discriminated. This is an essential prerequisite for analyzing cellular microvesicles and nanovesicles in biological fluids. The overlap of the peaks in Figure 2, A is due to the inherent limitation of measuring a stochastic process (brownian motion) by sampling over a finite time period (the time for which each particle can be tracked). For analysis of beads where the particle size distribution is well defined, the NTA software can correct for this and produce tighter peaks. However, for analyzing cellular vesicles where we cannot make assumptions about the size distribution, we chose not to use this correction, resulting in the apparent poorer resolution seen here. NTA also allows the particle concentration to be estimated directly, with good linearity being seen between the actual concentration and that measured by NTA in the range of 1 × 108-8 × 108 beads per milliliter for monodisperse 100 nm beads (Figure 2, B). Concentration measurement is less precise in polydisperse samples. This is due to the need to use instrument settings (camera shutter speed and gain) that are optimal for smaller (e.g., 100 nm) particles to ensure that they are included in the analysis. However, this leads to an overestimate of the concentration of the larger particles (e.g., 300 nm), because at these settings they may scatter multiple points of light, which are interpreted by the software as individual particles (see Supplementary Video S2). Comparison of NTA, electron microscopy, and flow cytometry analysis of placental vesicles The size distribution of placental vesicles was determined by measuring their diameters directly from electron micrographs of ultracentrifuge pellets. This showed a polydisperse sample with vesicles ranging in size from 20 nm to 600 nm with a peak size between 100 and 140 nm (Figure 3, A). Figure 3, B is a screenshot of the same sample analysed on the NanoSight LM10 showing a range of vesicle sizes. NTA gave a similar vesicle size distribution from 40 to 600 nm with a peak around 250 nm (Figure 3, C). The peak of smaller material (20−60 nm) seen by electron microscopy but not by NTA is partly due to the lower sensitivity of the LM10 instrument in this size range, but shrinkage artifacts during fixation for electron microscopy can also lead to undersizing of the vesicles. 16,17 We then demonstrated that NTA can measure vesicles that cannot be detected by flow cytometry. Figure 3, D shows a series of beads of standard sizes, analyzed by flow cytometry on a FSC (size) versus SSC (granularity) plot. On our instrument (Becton Dickinson LSRII), FSC can resolve beads of 1 μm from 590 nm diameter, but nothing smaller. By adding the SSC measurement, 200 nm beads can be detected but appear in the same region as instrument “noise.” On this basis, a 300 nm cutoff is a more realistic limit for practical use. The placental vesicle preparation was then analyzed on these settings by flow cytometry (Figure 3, E). The majority (>90%) of the vesicles were below 1 μm in diameter, with the main population at around 300-400 nm, cut off to the 300 nm detection limit. The vesicle count was 1.6 × 109 per milliliter. When the same sample was analyzed by NTA, ∼70% of the vesicles were 500 nm) alone (e.g., apoptotic bodies) may be better carried out by flow cytometry than by NTA. DLS can also detect nanovesicles, but it cannot accurately resolve heterogeneous mixtures of vesicles. 22 This is because a single detection element collects light from all particles simultaneously; meaning that the estimate of particle size and size distribution is biased to a larger particle size. In contrast, NTA simultaneously measures particle size and scattering intensity on individual particles, thus allowing heterogeneous particle mixtures to be resolved. Furthermore, the ability of NTA to see particles directly and individually allow the particle concentration to be estimated from extrapolating the number of particles seen at any given instant to a particle concentration per unit volume through knowing the scattering volume. This facility, which is unobtainable by conventional DLS methods, is invaluable for studying particles in biological fluids, because it allows the measurement of changes in the concentration of particles of different sizes between normal and disease states. Thus, the NanoSight instrument allows a quantitative estimation of sample size, size distribution, and concentration. The disadvantage of NTA in its original form was that it was unable to determine the phenotype of the vesicles. Biological fluids such as plasma and urine will inevitably contain mixtures of vesicles derived from many different cell types. It is therefore crucial to be able to determine the cellular origin of the vesicles and, to understand their biological function, the molecules that they express on their surface. Here we have successfully adapted and developed this technology by adding a fluorescence capability and demonstrated the proof of principle by measuring placental vesicles labeled with a specific antibody conjugated to quantum dots and cellular vesicles in human plasma using a quantum dot-labeled fluorescent cell tracker peptide. Labeling with the cell tracker peptide clearly shows that a large proportion of the vesicles in plasma are not derived from cells but are probably lipid vesicles, because they were removed by ultracentrifugation and the cellular vesicles pelleted. Experiments are now under way to label plasma vesicles with a panel of antibodies to platelet, erythrocyte, leukocyte, and endothelial cell markers to determine their composition. In summary, NTA in effect extends the power of flow cytometry downward by nearly one order of magnitude in terms of particle size and opens up new possibilities for research into microvesicles and nanovesicles.
