Evaluation of extracellular vesicle isolation and detection repeatability from cell cultures, biofluids, and mammalian tissues

Abstract

Unaccounted variations in biological marker sample handling, preparation, and data collection can result in inaccurate or unreliable data sets. Herein, we assess the repeatability (test‐retest reliability) of extracellular vesicle (EV) size and concentration measurements by dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA) methods. We evaluate the contribution of pre‐analytical variation and analytical imprecision factors toward the total experimental variability in routine preparation and quantitative analysis of crude large (microvesicle, MV) or small (exosome, EXO) EV particle‐enriched fractions following differential velocity centrifugation. This is performed on EV‐enriched cell culture media, human biofluids, and mammalian tissues, which represent an increasing degree of analyte complexity and heterogeneity. Incremental increase in NTA camera level (CL) setting (shutter and gain) was associated with a proportional detection of larger small‐sized particle counts. There was a strong linear relationship between the raw sample volume or wet tissue weight input, and EV output. At optimal non‐oversaturating particle detection conditions, NTA and DLS techniques delivered low (< 10%) inter‐sample coefficient of variability (CV%) when estimating EV size or number in technical replicates (TRs) prepared by using a 2‐step analyte dilution and sampling strategy. However, according to Bland‐Altman analysis, the repeatability of intra‐sample EV concentration measurements, but not size measurements by NTA, was prone to a significant run order bias within the first data capture cycle (DCC). Although analyte origin (random anatomical sub‐locations or different batches) did not represent a substantial source of error in EV analysis, a parallel EV isolation in TRs derived from a single supply, and a successive sample pooling with a potential preparation of second‐tier serially diluted TRs, is advised. To reduce the amount of analytical error in NTA analytes with the particle per frame (PPF) number between 40 and 100, and the polydispersity index (PDI) of < 30%, no less than three 60 sec‐long or six 30 sec‐long repeated intra‐sample measurements by at least two successive DCCs should be recorded. In DLS, ten 30 sec‐long repeated measurements should be captured in each of three independent runs.