Molecular subtypes and differentiation programmes of glioma stem cells as determinants of extracellular vesicle profiles and endothelial cell-stimulating activities

Abstract

We have previously uncovered the impact of oncogenic and differentiation processes on extracellular vesicles (EVs) in cancer. This is of interested in the context of glioma stem cells (GSC) that are responsible for recurrent nature of glioblastoma multiforme (GBM), while retaining the potential to undergo differentiation and self renewal. GSCs reside in vascular niches where they interact with endothelial cells through a number of mediators including bioactive cargo of EVs. GSCs can be classified as proneural (PN) or mesenchymal (MES) subtypes on the basis of their gene expression profiles and distinct biological characteristics. In the present study we investigated how GSC diversity and differentiation programmes influence their EV-mediated communication potentials. Indeed, molecular subtypes of GBMs and GSCs differ with respect to their expression of EV-related genes (vesiculome) and GSCs with PN or MES phenotypes produce EVs with markedly different characteristics, marker profiles, proteomes and endothelial stimulating activities. For example, while EVs of PN GSC are largely devoid of exosomal markers their counterparts from MES GSCs express ample CD9, CD63 and CD81 tetraspanins. In both GSC subtypes serum-induced differentiation results in profound, but distinct changes of cellular phenotypes including the enhanced EV production, reconfiguration of their proteomes and the related functional pathways. Notably, the EV uptake was a function of both subtype and differentiation state of donor cells. Thus, while, EVs produced by differentiated MES GSCs were internalized less efficiently than those from undifferentiated cells they exhibited an increased stimulatory potential for human brain endothelial cells. Such stimulating activity was also observed for EVs derived from differentiated PN GSCs, despite their even weaker uptake by endothelial cells. These findings suggest that the role of EVs as biological mediators and biomarkers in GBM may depend on the molecular subtype and functional state of donor cancer cells, including cancer stem cells. Abbreviations: CryoTEM: cryo-transmission electron microscopy; DIFF: differentiated GSCs; EGF: epidermal growth factor; DUC: differential ultracentrifugation; EV: extracellular vesicle; FGF: fibroblast growth factor; GBM: glioblastoma multiforme; GFAP: glial fibrillary acidic protein; GO: gene ontology; GSC: glioma stem cells; HBEC-5i: human brain endothelial cells; MES: mesenchymal cells; MTS - [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; PMT1: proneural-to-mesenchyman transition cell line 1; PN: proneural cells; TEM: transmission electron microscopy; WB: western blotting.

Keywords: Glioblastoma; angiogenesis; extracellular vesicles; glioma stem cells; proteome.

Figure 1.

Expression of vesiculation-related genes in datasets of glioma cell lines and stem cell lines. Representation of 90 genes of which 87 are known to regulate EV biogenesis, or contribute to EV cargo is shown in Table S1. The remaining genes are elements of the GBM landscape. This information was extracted from transcriptomes of 15 glioma cell lines, including 10 glioma stem cell isolates (GSC) (6 PN, 4 MES) and 5 established GBM cell lines (GBM). The hierarchical clustering shows differential expression of genes based on stem cells subtypes. Blue arrows represent upregulated genes in PN GSC, while red arrows represent upregulated genes in MES GSC.

Figure 2.

Transmission electron microscopy (TEM) of cultured GSC spheres. Morphological differences in three-dimensional growth pattern between PN (a-b) and MES GSC populations (c-d).

Figure 3.

Cellular characteristics and phenotypic differentiation of proneural (PN) and mesenchymal (MES) GSCs in culture. Isolation and differentiation of proneural (PN) and mesenchymal (MES) glioma stem cells. (a) Schematic representation of GSCs differentiating into more committed cells in the presence of serum and after removal of stem cell growth factors (B27, FGF and EGF). After 30 days, PN differentiated cells (PN DIFF) enter senescence programme, while MES differentiated cells (MES DIFF) assume a more spindle-shaped phenotype, but continue proliferating. (b) Monitoring cell senescence by β-galactosidase assay (β-gal). While differentiated MES cells (top panel) remain β-gal-negative, their PN counterparts stain blue for β-gal (bottom panel), as described earlier [37].

Figure 4.

Quantitative PCR analysis of vesiculation-related genes reveals differences associated with GSC subtype and differentiation state. Customized RT2 targeted PCR arrays were used to evaluate the expression levels of vesiculation-related genes (vesiculome) in glioma stem cells (PN GSC and MES GSC) and their counterparts differentiated in the presence of serum (PN DIFF and MES DIFF). Upregulation and downregulation of notable genes involved in vesiculation were denoted by blue arrows for PN and red arrows for MES subtypes of GSC and DIFF cells.

Figure 5.

Characteristics of extracellular vesicles released by GSCs vary as a function of subtype and differentiation. (a) NTA analysis of EVs size distribution indicates that differentiated MES cells (MES DIFF) exhibit a significantly increased release of EVs smaller than 150 nm. (b) NTA profiles of the total particle numbers released by the GSC cell populations. Differentiation significantly increases the number of particles released by these glioma cells. (c) Cryo-TEM of EVs isolated from GSC and their differentiated counterparts illustrating the heterogeneity in EV size and CD63 immunogold labelling (dark particles). The images represent PN 10K pellet (left and middle panels) and MES DIFF 100K fraction (right panel), additional images are shown in Figure S3.

Figure 6.

Distinct profiles of marker and cargo proteins reflect the impact of donor cell subtype and differentiation status on the biogenesis of extracellular vesicles. Immunoblotting for indicated proteins was performed using 2 μg of the protein preparation from 100K EV fraction isolated from conditioned media of the respective PN and MES GSCs and their corresponding differentiated progeny (DIFF). Of note are differences in expression of CD63, Syntenin-1, Flot1 and TSG101 (see text and supplementary figures).

Figure 7.

The impact of GSC subtype and differentiation status on the EV proteome. (a) Protein quantification in EV isolates from comparable GSC and DIFF cultures of PN and MES glioma cell lines. MES derived EV preparations contain higher total protein content than that of PN EVs. (b) Venn diagram of common and unique EV proteins identified by mass spectrometry in preparations of conditioned media from the indicated cell lines. (c-d) Differential expression of proteins expressed in GSC versus DIFF cells within PN and MES cell subtypes, respectively. MES glioma cells contain a greater number of proteins relative to PN cells. Coordinates: x axis = log2(fold-change) (GSC/DIFF), y axis = −log10(P value). The horizontal line indicates P value = 0.05. Data represent results of three independent experiments pooled together.

Figure 8.

Gene ontology analysis of extracellular vesicle proteomes across subtypes and differentiation states of donor GSCs. GO analysis was performed using FUNRICH database. EV cargos of the indicated GSC populations differ with regards to their cellular compartment (a) and molecular functions (b).

Figure 9.

Cytoscape representation of changes in functional clusters of extracellular vesicle-associated proteins released from mesenchymal and proneural GSCs in their undifferentiated and differentiated states. GO analysis was performed using geneMANIA software. Network of proteins in PN (a) and MES (b) GSC-derived EVs and those detected in PN (c) and MES (d) DIFF-derived EVs. The clusters and their constituent proteins dramatically changed between subtypes and differentiation states of glioma cell subsets.

Figure 10.

Differential uptake and endothelial stimulating activity of EVs as a function of donor GSCs subtype and differentiation status. Endothelial cell (HBEC-5i) uptake of, and proliferative response to EVs, as measured by FACS and MTS, respectively. (a) PKH26-labelled EVs were incubated with HBEC-5i recipients and the fluorescence transfer was quantified by flow cytometry. HBEC-5i preferentially uptake EVs from MES GSC. (b) The 100K fractions of indicated EV isolates were used to stimulate HBEC-5i cell proliferation (MTS). The cellular responses are particularly notable in the case of EVs from DIFF glioma cells, especially of the MES subtype.