Proteomic profiling of NCI-60 extracellular vesicles uncovers common protein cargo and cancer type-specific biomarkers

Abstract

Packed with biological information, extracellular vesicles (EVs) offer exciting promise for biomarker discovery and applications in therapeutics and non-invasive diagnostics. Currently, our understanding of EV contents is confined by the limited cells from which vesicles have been characterized utilizing the same enrichment method. Using sixty cell lines from the National Cancer Institute (NCI-60), here we provide the largest proteomic profile of EVs in a single study, identifying 6,071 proteins with 213 common to all isolates. Proteins included established EV markers, and vesicular trafficking proteins such as Rab GTPases and tetraspanins. Differentially-expressed proteins offer potential for cancer diagnosis and prognosis. Network analysis of vesicle quantity and proteomes identified EV components associated with vesicle secretion, including CD81, CD63, syntenin-1, VAMP3, Rab GTPases, and integrins. Integration of vesicle proteomes with whole-cell molecular profiles revealed similarities, suggesting EVs provide a reliable reflection of their progenitor cell content, and are therefore excellent indicators of disease.

Keywords: biomarkers; co-inertia; exosomes; microvesicles; proteomics.

Figure 1. Proteomic analysis of extracellular vesicles secreted by the NCI-60 cells

A. Centrifugation protocol and general workflow of EV enrichment for LC-MS/MS analysis. B. Venn diagram of proteins identified in EV samples in the NCI-60 and NCI-60]_stringent datasets compared to the Vesiclepedia database of proteins. See also Supplementary Table S2. C. Average spectral counts per cell line across tissue types. Parentheses indicate the number of cell lines represented in each tissue type. Data are represented as mean ± STD. D. Total proteins including common proteome (dotted line) and tissue-specific proteins identified across the nine histological origins represented. See also Supplementary Table S3 and S4.

Figure 2. Enrichment analysis of EV proteins identified

Proteins identified in the [NCI-60]_stringent dataset were used for enrichment analyses. A. Functional and B. pathway enrichment analysis of EV proteins using the DAVID database (GOTERM_BP_FAT and KEGG_PATHWAY). All terms were significant (p < 0.05) following Benjamini correction. C. Subcellular localization enrichment using FunRich. All terms were significant (p < 0.001) following Benjamini correction. D. Spectral counts of common vesicle protein markers measured across all NCI-60 EV samples.

Figure 3. Differential expression of proteins found in NCI-60 cell-derived extracellular vesicles

A. PCA plot based on variant vesicle proteins across tissue types. B. Unsupervised hierarchal clustering of cell lines based on EV proteomic profiles. Spectral count-based differential expression of C. Agrin, D. Intercellular adhesion molecule (ICAM) 3, E. Premelanosome protein (PMEL), and F. Tenascin-XB. See also Supplementary Figure 1 and Supplementary Table S6.

Figure 4. Network analysis of protein content with vesicle secretion

A. Network heatmap plot of topological overlap depicting protein dendrogram and module assignment. Targeted yellow module is highlighted. See also Supplementary Figure 3. B. Heatmap of module-trait correlation containing correlation and relevant p-values of modules detected. See also Supplementary Table S7 and S8. C. Scatter plot of membership in the yellow module and protein significance corresponding to vesicle secretion quantities across the NCI-60. A significant positive correlation between protein significance and module membership was determined (p = 0.006). D. GOTERM_BP_FAT analysis of biological processes enriched in the yellow module protein dataset.

Figure 5. Comparison of vesicular proteome with cellular proteome and transcriptome

A. Co-inertia analysis of cellular proteome and transcriptome with vesicular proteome across the NCI-60 panel. Vectors represent the proportional divergence between data sets within each cell line. B. Coordinates of the proteins or RNA from each data set plotted in the same orientation as the co-inertia analysis. C. Histogram of eigenvalues. Blue bars represent the absolute values of eigenvalues retained in the analysis. Black dots represent the proportion of variance found within each eigenvector. D. Plot of pseudo-eigenvalues space, indicating the variance of eigenvalues contributed by each of the three datasets. Comparison of expression levels of core vesicular proteins with whole cell protein expression from representative cells lines: E. HCC-2993 (colon); F. NCI-H322M (lung); G. IGROV1 (ovary); and H. HOP-62 (lung). See also Supplementary Table S9.