Cancer cell invasion alters the protein profile of extracellular vesicles

Abstract

Extracellular vesicles (EVs) are important mediators of intercellular communication involved in local and long-range signalling of cancer metastasis. The onset of invasion is the key step of the metastatic cascade, but the secretion of EVs has remained unexplored at that stage due to technical challenges. In this study, we present a platform to track EVs over the course of invasive development of human prostate cancer cell (PC3) tumoroids utilizing in vivo-mimicking extracellular matrix-based 3D cultures. Using this EV production method, combined with proteomic profiling, we show that PC3 tumoroids secrete EVs with previously undefined protein cargo. Intriguingly, an increase in EV amounts and extensive changes in the EV protein composition were detected upon invasive transition of the tumoroids. The changes in EV protein cargo were counteracted by chemical inhibition of invasion. These results reveal the impact of the tumoroids' invasive status on EV secretion and cargo, and highlight the necessity of in vivo-mimicking conditions for uncovering novel cancer-derived EV components.

Keywords: 3D cultures; EV heterogeneity; EV proteome; PC3 cells; extracellular vesicles; invasion; prostate cancer.

FIGURE 1

Isolation of EVs from ECM‐based 3D cultures undergoing invasive transition. Graphical workflow of EV isolation from the ECM‐based 3D cell culture. PC3 cells were seeded on top of coverslips in‐between two layers of Matrigel (3.5 mg/mL) and allowed to grow and form tumoroids during 12 days. Conditioned media of the tumoroid cultures was collected every two days, from which EVs were isolated by differential centrifugation into 10 and 100K pellets. Shown are representative brightfield images of the tumoroids. Scale bars 200 µm.

FIGURE 2

EV secretion increases substantially at the time of invasive transition of PC3 tumoroids. (a) Particle concentration of the 10 and 100K pellets obtained by nanoparticle tracking analysis (NTA). Significant differences were assessed by one‐way ANOVA with a Tukey post hoc test, *< 0.05, +SEM, n = 4. (b) EV size distribution of the 10 and 100K pellets detected by NTA shown as mean values of four independent experiments. (c) Immunoblot analysis of EV markers CD81, CD63 and TSG101 in 100K pellets. APOA1, Lamin A/C and GM130 were used as purity controls for the EV samples (d4‐d12). WCL, whole cell lysate. (d) ZETA potential measurement of the 100K samples. Significant differences were assessed by one‐way ANOVA with a Tukey post hoc test, *< 0.05, +SEM, n = 4. (e) Number of PC3 cells grown per well as described in Figure 1. The culture was started (day 0) with 6000 cells per well. Mean values of five independent repeats are shown +SEM. (f) Immunoblot analysis of PARP1 in PC3 cells grown as in Figure 1. Bortezomib (BTZ) treatment (300 nM, 22 h) of 2D grown PC3 cells was used as a positive control for PARP1 cleavage. HSC70 was used as a loading control.

FIGURE 3

High‐resolution gradient purification of EVs from non‐invasive and invasive PC3 tumoroid cultures. (a) EVs were isolated from the conditioned media of non‐invasive PC3 3D cultures at days 2–8 (3D d2‐d8), invasive cultures at days 10–14 (3D d10‐d14), dissolved Matrigel at day 14 (3D Dissolved Matrigel), and conditioned media of 2D cultures (2D). OptiPrep density gradient centrifugation was employed for these samples and the resulting 11 fractions were screened for EVs (b). Fractions 5, 6 and 7 were pooled together and used for subsequent assays (Figures 4 and 5). (b) Immunoblot analyses of the 11 fractions with EV markers CD81, CD63 and TSG101. Lamin A/C was used as an EV purity marker and PC3 whole cell lysate (WCL) as a positive control for immunoblotting.

FIGURE 4

Non‐invasive and invasive PC3 tumoroid cultures secrete EVs of similar size and morphology. (a) Size distribution of EVs from non‐invasive (3D d2‐d8) and invasive (3D d10‐d14) 3D cultures, dissolved Matrigel (3D Dissolved Matrigel), and from 2D cultures (2D). Mean values of three independent experiments conducted with NTA are shown. (b) Morphology of the EVs from 3D and 2D cultures as analyzed by TEM. Representative images with scale bar 500 nm for the full sized and 100 nm for the zoomed in images.

FIGURE 5

EV protein content changes upon invasive transition of 3D cultured PC3 cells. LC‐MS/MS analyses of EV proteins from non‐invasive PC3 3D cultures at days 2–8 (3D d2‐d8), invasive cultures at days 10–14 (3D d10‐d14), dissolved Matrigel at day 14 (3D Dissolved Matrigel), and 2D cultures (2D), n = 2. (a and b) Venn diagram of EV proteins from 2D and 3D cultures and with EV proteins deposited in the Vesiclepedia database for PC3 cells. (c) Venn diagrams of EV proteins from non‐invasive and invasive 3D cultures, and (d) EV proteins retained by the Matrigel (3D Dissolved Matrigel). (e) Venn diagram of the identified EV proteins. (f) Immunoblot analysis of the EV samples with selected MS identified proteins; VCL, ITGB4, GAPDH, FGB and CD59. DM, dissolved Matrigel; WCL, whole cell lysate of 3D grown cells.

FIGURE 6

Inhibition of invasion changes the EV cargo content. (a) IPA‐3 treatment of PC3 cells grown in 3D. The media was harvested and replenished every 48 h. From day 4 onwards, 15 µM of IPA‐3 or the equivalent volume of DMSO was added to the media. Harvested media from days 2–8 and 10–14 were pooled together. (b) Representative brightfield images of the DMSO or IPA‐3 treated cells. Scale bars 200 µm. (c) Immunoblot analysis of EVs from DMSO or IPA‐3 treated PC3 cells with VCL, ITGB4, CD59 and CD73. WCL from DMSO and IPA‐3 treated tumoroids were harvested at day 14.