Fast characterisation of cell-derived extracellular vesicles by nanoparticles tracking analysis, cryo-electron microscopy, and Raman tweezers microspectroscopy

Abstract

The joint use of 3 complementary techniques, namely, nanoparticle tracking analysis (NTA), cryo-electron microscopy (Cryo-EM) and Raman tweezers microspectroscopy (RTM), is proposed for a rapid characterisation of extracellular vesicles (EVs) of various origins. NTA is valuable for studying the size distribution and concentration, Cryo-EM is outstanding for the morphological characterisation, including observation of vesicle heterogeneity, while RTM provides the global chemical composition without using any exogenous label. The capabilities of this approach are evaluated on the example of cell-derived vesicles of Dictyostelium discoideum, a convenient general model for eukaryotic EVs. At least 2 separate species differing in chemical composition (relative amounts of DNA, lipids and proteins, presence of carotenoids) were found for each of the 2 physiological states of this non-pathogenic microorganism, that is, cell growth and starvation-induced aggregation. These findings demonstrate the specific potency of RTM. In addition, the first Raman spectra of human urinary exosomes are reported, presumably constituting the primary step towards Raman characterisation of EVs for the purpose of human diseases diagnoses.

Keywords: Dictyostelium discoideum; Raman tweezers microspectroscopy; cryo-electron microscopy; extracellular vesicles; nanoparticle tracking analysis; urinary exosomes.

Fig. 1

Size distribution and concentration of D. discoideum EVs samples (Table I), as measured by NTA. EVs were prepared from growth medium after 48 hours of cell growth, either with the usual conditions for the 12,000×g centrifugation (in Eppendorf tubes) (a), or with the 12,000×g centrifugation performed in 30 ml tubes (see Materials and methods) (b). The curve (c) corresponds to EVs remaining in the 12,000×g supernatant, which relates to sample (b) EVs. The 12,000×g pellets (a and b) are concentrated in PBS by factors of 20 and 100, respectively, whereas the 12,000×g supernatant (c) is as obtained after centrifugation.

Fig. 2

Observation of D. discoideum EVs by Cryo-EM. (A) EVs obtained after 24 hours of cell growth; (B) EVs obtained after 48 hours of cell growth. (C) Some rare EVs configurations, conserved in the vitreous ice environment, observed during growth, showing broken vesicles (a), big EVs inserted inside larger vesicles (b, d) or small EVs into multivesicular bodies-like vesicles (c, e) and EVs prone to fuse (c, e, f). Figure 2D shows 5 different images of EVs obtained from the starvation medium of D. discoideum cells (see Materials and methods). Bars: 100 nm for A and B, 50 nm (a–e) or 100 nm (f) for C and 50 nm for D.

Fig. 3

Raman spectra of optically trapped vesicles released from D. discoideum cells in the starvation phase. (A) The time sequence of raw Raman spectra recorded every 3 seconds, for one particular vesicle set. The event of vesicle(s) trapping is seen within the time window from 6 to 9 seconds, by the appearance of characteristic Raman bands of nucleic acids (NA, ~783 cm⁻¹), phenylalanine (Phe, ~1,005 cm⁻¹) and the intensity increase of the lipids/proteins marker of CH₂ groups deformations (LP, ~1,450 cm⁻¹). Note also the appearance and gradual increase of the prominent carotenoids Raman marker at ~1,533 cm⁻¹. (B) Raman spectra for 2 different vesicle sets (a, b) and their difference spectrum (c = a − b×0.88) corresponding to the characteristic Raman spectrum of carotenoids whose major bands [ν₁, ν₂, δ(C = CH)] are indicated at trace (c). Raman spectra (a) and (b) have been obtained from raw spectra after subtraction of the PBS contribution. The factor 0.88 was found empirically to emphasise the carotenoids’ contribution in spectrum (c).

Fig. 4

Comparison of the Raman spectra of D. discoideum growth and starvation EVs. (A) Proposed interpretation of Raman spectrum originated from vesicles released from D. discoideum in the starvation phase. This spectrum was obtained by averaging 15 different sets of vesicles, with a total accumulation time of 14 minutes, then subtraction of the PBS contribution, and smoothing by 2 adjacent pixels. (B) Variability of Raman spectra of D. discoideum vesicles during growth (a, b) as compared to starvation phase (c, d). Within the same phase, one can qualitatively distinguish at least 2 different species: during growth, species (b) is characterised by an increased amount of lipids as compared to species (a), and during starvation, species (d) contains an increased amount of nucleic acids and carotenoids as compared to species (c). The prominent spectral differences are highlighted by dashed oval curves.

Fig. 5

Raman spectra of optically trapped urinary exosomes from healthy human samples (see Table II). (A) Similarity/variability of spectra from 6 different vesicles sets, for the same sample “P2_11_M” (a), each averaged over ~50 raw Raman spectra. (B) Raman spectra for 4 different exosome samples: “P2_11_M” (a) (averaging of 303 raw spectra), “P2_10_F” (b) (266 raw spectra), “P2_15_F” (c) (287 raw spectra) and “P2_1_M” (d), (140 raw spectra). All spectra were corrected for the PBS contribution, the slowly changing background from Rayleigh scattering and smoothed by 3 adjacent pixels (see Materials and methods).