A quick pipeline for the isolation of 3D cell culture-derived extracellular vesicles

Abstract

Recent advances in cell biology research regarding extracellular vesicles have highlighted an increasing demand to obtain 3D cell culture-derived EVs, because they are considered to more accurately represent EVs obtained in vivo. However, there is still a grave need for efficient and tunable methodologies to isolate EVs from 3D cell cultures. Using nanofibrillar cellulose (NFC) scaffold as a 3D cell culture matrix, we developed a pipeline of two different approaches for EV isolation from cancer spheroids. A batch method was created for delivering high EV yield at the end of the culture period, and a harvesting method was created to enable time-dependent collection of EVs to combine EV profiling with spheroid development. Both these methods were easy to set up, quick to perform, and they provided a high EV yield. When compared to scaffold-free 3D spheroid cultures on ultra-low affinity plates, the NFC method resulted in similar EV production/cell, but the NFC method was scalable and easier to perform resulting in high EV yields. In summary, we introduce here an NFC-based, innovative pipeline for acquiring EVs from 3D cancer spheroids, which can be tailored to support the needs of variable EV research objectives.

Keywords: 3D cell culture; cancer spheroids; extracellular vesicles; isolation; nanofibrillar cellulose.

FIGURE 1

Flowchart for batch and harvesting methods to isolate EVs from NFC 3D cultures. For the batch method (i) the NFC is first diluted to the desired NFC concentration (depends on the cell type), a preoptimized number of cells are suspended into the NFC and 300 μl of the suspension is seeded on a 48‐well plate. Medium (300 μl) is added on top of the hydrogel, and the cells are allowed to grow into spheroids with the medium changed every two days by carefully aspirating 150 μl of the medium from top of the NFC and gently adding 150 μl of medium. To release the spheroids and EVs trapped in the NFC, 600 μg of cellulase per mg of NFC is diluted with medium to 300 μl and the mixture is added in the culture well and incubated for 24 h in +37°C. The spheroids are removed by centrifugation at 600 × g for 10 min and the supernatant is collected for EV isolation. To remove cell and NFC debris, the supernatant is first centrifuged at 5000 × g for 15 min at 4°C and the pellet is discarded. The supernatant can then be filtered to increase sample purity. EVs are finally isolated with a method of choice, for example in this case ultracentrifugation at 189,000 × g for 90 min at 4°C. The pellet containing the EV is suspended into PBS. For the harvesting method (ii) both the NFC dilution and cell seeding are done similarly as in the batch method, but the well content is collected every two days and centrifuged at 1000 × g for 10 min. The supernatant is collected and centrifuged at 5000 × g for 30 min at 4°C to remove cell and NFC debris and the pellet is discarded. EVs are isolated from the supernatant with a method of choice, for example, in this case by ultracentrifugation at 110,000 × g for 120 min at 4°C and the pellet containing the EV is suspended into PBS. The NFC pellet from cell cultures is resuspended into fresh conditioned medium (300 μl) and reseeded back to a 48 well plate to continue the culture, or it can be digested with 600 μg of cellulase per mg of NFC (diluted with medium to 300 μl) for 24 h to release the spheroids

FIGURE 2

Quick pipeline for EV isolation from NFC‐cultured 3D spheroids – batch and harvesting method. Briefly, the cancer spheroids were cultured in NFC, and the EV –containing conditioned medium was accessed by (i) NFC digestion with cellulase, aiming for maximal recovery of EVs (batch method); (ii) removing the NFC & spheroid mixture and separating the conditioned medium for EV isolation and subsequently re‐plating the mix of NFC & spheroids every two days for long‐term 3D culture (harvesting method). Confocal 3D maximum intensity projection of a live MCF7 spheroid expressing GFP‐HAS3 grown in NFC is shown in the right panel. Scale bar: 50 μm. Figure (flow chart) created with Biorender.com

FIGURE 3

Effect of NFC concentration, cellulase digestion time, and supernatant filtering on sample purity and particle concentration by NTA. (A) High concentration of NFC increased the number of particles and variation of particle concentrations from the empty NFC scaffolds, n = 6 independent experiments. T‐test, (n.s.). (B) Increasing the time of cellulase digestion decreased the number of empty NFC derived particles, n = 6 independent experiments. One‐way ANOVA with Tukey's post hoc test, (n.s.). (C) Comparison of particle concentrations from the cell culture (EV) and empty NFC (NFC) samples with different filtering steps, n = 4‐6 independent experiments. One‐way ANOVA with Tukey's post hoc test, *** P < 0.001. (D) The non‐filtered and filtered samples were compared with scanning electron microscopy. Cell culture samples contained a high number of spherical particles of variable sizes with vesicle‐like morphology (arrows). Fibre‐like structures found in the non‐filtered samples (arrowheads) were absent in the filtered samples. Scale bars:1 μm

FIGURE 4

Concentration of EVs obtained by the harvesting method. Comparison of the effect of i) refreshing the medium completely and ii) topping up with the fresh medium on particle numbers by NTA. (A‐B) To compare the refreshing with topping up, A2058 cells were seeded in NFC, and the conditioned medium was harvested for EV isolation, and particle concentration (A) and particle size (B) were measured by NTA from 3 to 6 independent experiments Statistics: ANOVA with Tukey's post‐hoc test. Day 3 vs Day 9, Day 3 vs Day 11 top‐up medium: (**) P < 0.01, Day 3 vs Day 11 refreshing medium: (***) P < 0.001

FIGURE 5

Comparison of EV production with spheroid growth, cell number and total protein. (A) Spheroids’ live imaging shows the morphology and size when cultured up to 11 days in aNFC. (B) Number of cells within 3D A2058 melanoma spheroids followed up to 11 days from three independent experiments. (C) Total protein concentration of the 3D spheroids from three independent experiments. (D) Particle concentration of the isolated EV samples from the conditioned medium by NTA from 3 to 6 independent experiments. The spheroids were lysed using RIPA buffer and total protein concentration was determined by the Lowry method from three independent experiments. (E) EV production yield shown as ratio of particles/cell count. Statistics: ANOVA with Tukey's post‐hoc test. (**) P < 0.01; (***) P < 0.001

FIGURE 6

EV characterization by SP‐IRIS. EV characterization by SP‐IRIS from A2058 spheroids with harvesting method (days 3–11) and MCF7 spheroids with batch method (day 7) with secondary labelling of EVs on CD9, CD81 and CD63 antibody capture spots: anti‐CD9 (488A – blue), anti‐CD81 (555 – green), and anti‐CD63 (647 – red). The boxplot shows the total particle count from the correspondent antibody‐spot, and replicates are represented by dot plot. On the bottom of the graphics, fluorescence microscopy representative images of 3D‐derived EVs measured. Particle count and size data is shown in Supplementary File T1 and T2

FIGURE 7

EV production by the harvesting method using ULA plate. Comparison between refreshing the medium completely and topping up with fresh medium. (A–B) A2058 cells were seeded in the ULA plate, and the conditioned medium was harvested for EV isolation, and particle concentration (A) and particle size (B) were measured by NTA from 3 independent experiments. Statistics: ANOVA with Tukey's post‐hoc test. Day 11 top‐up medium vs Day 11 refreshing medium: (*) P < 0.05