Download App Note

Dr. Sascha Raschke
Particle Metrix GmbH

 

INTRODUCTION

Fluorescence NTA (F-NTA) provides a powerful and specific approach for characterizing extracellular vesicles (EVs) in complex biological samples like plasma or serum. In these samples, conventional light-scattering NTA reaches its physical limits because potential lipoprotein co-isolates are detected non-specifically alongside the target vesicles. These contaminants overlap in size and often outnumber EVs by several orders of magnitude, distorting scatter-mode quantification [1]. To improve specificity, universal lipid dyes like CellMask™ are frequently used in fluorescence NTA (F-NTA); however, they lack specificity and selectivity for particular membrane components, as they bind indiscriminately to any accessible hydrophobic structure through non-selective hydrophobic interactions [2]. To ensure accurate EV quantification, the International Society for Extracellular Vesicles (ISEV) recommends targeting EV-enriched membrane proteins. Fluorophore-conjugated antibodies against tetraspanins (e.g. CD9, CD63, and CD81) have therefore become the standard approach for fluorescence-based EV detection [3].

Using pure HDL, LDL, and reference HCT116 EV islolates, this Application Note highlights how our specific F-NTA antibodies enable reliable extracellular vesicle detection, eliminating the false-positive signals associated with lipid-based staining and allowing confident EV characterization.

MATERIALS & METHODS

Samples and Fluorescence Labeling Strategy

Extracellular vesicle isolates derived from the HCT116 cell line were characterized using a panel of antibodies originated from the Particle Metrix Tetraspanin Detection Kits (488-, 520- and 640- conjugation, corresponding to the Particle Metrix article nos. #700384, #700385 and #700386). To rule out artifactual particle formation, antibody-only and dye-only controls were analyzed alone; since no aggregation was observed the data are not shown.

For tetraspanin staining, 1µl HCT116 EVs (≙1.6x10⁸ particles) were incubated with 1µl of a CD9/CD63/CD81 PAN antibody cocktail (pre-diluted 1:10) and 18µl 10% PBS. For membrane labeling, EVs were similarly incubated with CellMask™ Green (CMG, pre-diluted 1:10, #C37608), Orange (CMO, 1:1,000, #C10045), or Deep Red (CMDR, 1:1,000, #C10046), all purchased from Thermo Fisher Scientific. Lipoproteins (Sigma Aldrich, #SAE0053 10MG, #SAE0054 10MG) were pre-diluted in 10% PBS (HDL 1:5; LDL 1:500), and 1µl of diluted HDL (≙1.2x10⁸ particles) or LDL (≙7.4x10⁷ particles) was incubated with 1µl of the respective CellMask™ dyes and 18µl 10% PBS. All staining reactions (20µl total volume) were incubated for 1 hour at room temperature in the dark, then supplemented with 10% PBS to a final measurement volume of 1,000µl.

Nanoparticle Tracking Analysis

Triplicate NTA measurements were conducted on a ZetaView® PMX x35 QUATT (technically equivalent to the x40 Evolution series), with the channels (scatter, 488/500, 520/550, 640/660) automatedly set up via the ZetaSphere software (v1.1 SP4). To prevent photobleaching, a "fluorescence-first" sequence (fluorescence before scatter) was used, combined with automated sample advancement before each video acquisition. Pre-installed EV measurement parameters were fine-tuned to optimize camera settings for specific particle properties. Detailed acquisition parameters are listed below.
Table 1: NTA-acquisition settings for scatter and fluorescence mode.

RESULTS

Lipoprotein Cross-Reactivity and Masking (LDL vs. HDL)

To evaluate generic lipid staining, pure low-density (LDL) and high-density (HDL) lipoproteins were stained with CMG, CMO, and CMDR and analyzed via F-NTA. While scatter mode showed stable baseline concentrations for both populations, the fluorescence channels revealed a stark biophysical divergence. LDL exhibited massive cross-reactivity across all wavelengths, with labeling efficiencies ranging from 55.2% (CMO) to 70.2% (CMDR). This high binding rate is driven by the accessible, fluid lipid monolayer of LDL. In contrast to this, HDL showed a labeling efficiency of 0%, indicating that it remained completely masked. This observation may be attributed to the dense apolipoprotein shell (primarily composed of ApoA-I), which could sterically hinder access to the tiny lipid core.

Figure 1: CellMask™ staining of HDL (left) and LDL (right). While HDL could not be stained, LDL exhibited a clear fluorescence signal in all fluorescence channels after labelling with CMG, CMO and CMDR.

Analytical Specificity Validation and EV:LDL Spiking Control

To evaluate assay selectivity in the presence of co-isolated contaminants, standard HCT116 extracellular vesicle isolates were benchmarked against an intentional 50:50 particle mixture of EVs and isolated low-density lipoproteins (EVs:LDL 50:50). Preparation of the particle mixture was based on particle number in the NTAs field of view. While conventional scatter-mode NTA remained entirely uniform for PAN- and CellMask™ staining across all groups (1.59x1011 to 1.72x1011 particles/ml), switching to the fluorescence channels exposed a fundamental divergence between the two staining strategies.

Figure 2 illustrates, that in the anti-tetraspanin cocktail (CD9/CD63/CD81) F-NTA tracked a uniform baseline concentration of approximately 7.6x10¹⁰ particles/ml (~47% recovery) in the „HCT116 EV“ setup. As expected, the 50:50 EV:LDL particle mixture (50% EVs, 50% LDL) exhibited an approximately 50% lower fluorescent particle concentration across all channels (488, 520, and 640nm) with a mean of around 3.37x10¹⁰ particles/ml. This precise 50% reduction in fluorescent particles proves that the antibodies of the anti-tetraspanin cocktail remain completely inert toward LDL, selectively tracking only true CD9, -63, and -81 positive biological target vesicles.

Figure 2: PAN staining of HCT116 EVs (left) and EV:LDL 50:50 particle mixture (right). While particle concentration in scatter mode remained unchanged, a 50% drop of fluorescent particles in the LDL fraction was clearly observable in comparison to the EV fraction.

Conversely, as demonstrated in Figure 3, the universal membrane dyes (CMG, CMO, and CMDR) yielded a baseline recovery of approximately 1.2x10¹¹ particles/ml (~70% staining efficiency) in the HCT116 EV setup. However, following the 50:50 LDL spike, the fluorescence signals remained virtually unchanged, generating a high apparent staining rate of 1.03x10¹¹ particles/ml. Because these amphiphilic dyes insert blindly into any hydrophobic environment via non-specific interactions, they treat the phospholipid monolayer of the spiked LDL fraction identically to the vesicular bilayer of the EVs.

Figure 3: Universal membrane staining via CellMask™ dyes of HCT116 EVs (left) and EV:LDL particle mixture (right). The fluorescence particle concentration remains virtually unchanged despite the 50% reduction of the EV fraction, showing identical co-staining of the spiked LDL population across all channels.

CONCLUSION

The significant discrepancy between the apparent lipid-positive count (~70%, Fig. 3) and the tetraspanin-positive count (~47%, Fig. 2) within the same HCT116 EV sample indicates that non-vesicular lipoprotein remnants (such as LDL) remain co-isolated. Since CellMask™ dyes lack chemical selectivity for biological bilayers, they unselectively co-stain these low-density lipoprotein contaminants alongside the actual EVs, a phenomenon proven by the 55.2% to 70.2% cross-reactivity rate observed on pure LDL (see Fig. 1). Consequently, the elevated fluorescence counts in the CellMask™-EV group represent a systemic overestimation driven by co-isolated lipid impurities, rather than true vesicles. Conversely, any co-isolated HDL remnants remain inaccessible to CellMask™, potentially due to their tight, proteinaceous ApoA-I shell (see Fig. 1), thereby creating an unpredictable "blind spot" in the assay.

The comparative matrix-spiking kinetics demonstrate that while generic membrane dyes are highly efficient for staining pristine vesicle isolates, they are prone to introducing misleading results in heterogeneous samples containing lipid- or lipoprotein impurities. As shown by the flat signal profile in figure 3, CellMask™ cannot discriminate between biological membranes and persistent LDL co-isolates, driving a severe overestimation of vesicle concentrations and creating an unpredictable artifact profile. Consequently, generic membrane dyes are not recommended for EV quantification in heterogeneous matrices. However, they function as a proxy for total accessible lipid content rather than biological vesicular entities.

Consequently, purely lipid-based fluorescence tracking must be discouraged for complex matrices. In contrast, targeted multiplex immuno-NTA targeting CD9/CD63/CD81 (e.g. by using the Particle Metrix Tetraspanin Detection Kits) validated in figure 2, remains therefore a valid, artifact-free method capable of optically discriminating a specifically identified EV population from persistent size-overlapping lipoprotein contaminants.

REFERENCES

1. Théry, G., et al. (2018).: Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. Journal of Extracellular Vesicles, 7(1), 1535750.

2. Simonsen, J. B. (2017).: What are we looking at? Extracellular vesicles, lipoproteins, or both? Circulation Research, 121(8), 920-922.

3. Welsh, J. A., et al. (2024).: Minimal information for studies of extracellular vesicles 2023 (MISEV2023) guidelines: position statement of the International Society for Extracellular Vesicles. Journal of Extracellular Vesicles, 13(2), e12404.

Disclaimer: While we have taken every care to ensure that the information in this document is correct, nothing herein should be construed as implying any representation or warranty as to its accuracy, correctness or completeness of this information and we shall not be held liable for any errors therein or for any damages in connection with the use of this material. Particle Metrix reserves the right to change the content of this material at any time without notice.

Copyright: © July 2026 Particle Metrix. This publication or parts thereof may not be copied or distributed without our express written permission.