Characterization of extracellular vesicle-associated DNA and proteins derived from organotropic metastatic breast cancer cells

Abstract

Background: While primary breast cancer (BC) is often effectively managed, metastasis remains the primary cause of BC-related fatalities. Gaps remain in our understanding of the mechanisms regulating cancer cell organotropism with predilection to specific organs. Unraveling mediators of site-specific metastasis could enhance early detection and enable more tailored interventions. Liquid biopsy represents an innovative approach in cancer involving the analysis of biological materials such as circulating tumor DNA and tumor-derived extracellular vesicles (EV) found in body fluids like blood or urine. This offers valuable insights for characterizing and monitoring tumor genomes to advance personalized medicine in metastatic cancers.

Methods: We performed in-depth analyses of EV cargo associated with BC metastasis using eight murine cell line models with distinct metastatic potentials and organotropism to the lung, the bone, the liver, and the brain. We characterized the secretome of these cells to identify unique biomarkers specific to metastatic sites.

Results: Small EVs isolated from all cell lines were quantified and validated for established EV markers. Tracking analysis and electron microscopy revealed EV secretion patterns that differed according to cell line. Cell-free (cf)DNA and EV-associated DNA (EV-DNA) were detected from all cell lines with varying concentrations. We detected a TP53 mutation in both EV-DNA and cfDNA. Mass spectrometry-based proteomics analyses identified 698 EV-associated proteins, which clustered according to metastatic site. This analysis highlighted both common EV signatures and proteins involved in cancer progression and organotropism unique to metastatic cell lines. Among these, 327 significantly differentially enriched proteins were quantified with high confidence levels across BC and metastatic BC cells. We found enrichment of specific integrin receptors in metastatic cancer EVs compared to EVs secreted from non-transformed epithelial cells and matched tumorigenic non-metastatic cells. Pathway analyses revealed that EVs derived from parental cancer cells display a cell adhesion signature and are enriched with proteins involved in cancer signaling pathways.

Conclusion: Taken together, the characterization of EV cargo in a unique model of BC organotropism demonstrated that EV-DNA and EV proteomes were informative of normal and cancer states. This work could help to identify BC biomarkers associated with site-specific metastasis and new therapeutic targets.

Keywords: Biomarkers; DNA; Extracellular vesicles; Metastatic breast cancer; Organotropism; Proteomics.

Fig. 1

Parental and derivative metastatic cell lines with distinct organotropism. Created in BioRender. Burnier, J. (2025) https://BioRender.com/zhjiic3

Fig. 2

Characterization of EVs derived from cell cultures. A Size distribution, B average sizes, and (C) concentration of EVs as measured by NTA. Data were normalized to cell counts. D Micrographs of EV preparations by TEM. Scale bar = 100 nm. E Western blot analysis of specific proteins isolated from cell-derived EVs. In B and C, data are presented as mean ± SD (n = 3 independent experiments, *P < 0.05 and ***P < 0.001)

Fig. 3

Scanning electron microscopy imaging of cell cultures displaying microvesicles-like structures lining the cell membranes. Representative SEM images showing microvilli and budding of EVs on their surface (red arrows) in (A) non-transformed NMuMG cells, B non-metastatic 67NR primary BC cells, C metastatic 4T1 parental cells, D lung-MBC 4T1-533 cells, E bone-MBC 4T1-593 cells, F liver-MBC 4T1-2776 cells and (G) brain-MBC 4T1-BP cells. The box insert shows spherical vesicles on the cell surface. Magnification = 6500X

Fig. 4

Analysis of cfDNA and EV-DNA isolated from primary BC, parental MBC and organotropic MBC cells. A cfDNA levels normalized to CCM volume. B EV-DNA levels normalized to total cell count. C Two-dimensional ddPCR plots representing the detection of the Trp53 P31X mutation in EV-DNA and cfDNA isolated from the BC cell lines. Blue dots represent mutant-positive droplets; green dots represent wild type-positive droplets. D Concentration of mutant EV-DNA and cfDNA from BC and site-specific MBC cell lines. In A and B, data are presented as mean ± SEM (n = 3 independent experiments, **P: 0.0051 (NMuMG vs 4T1-2776), ***P: 0.0008 (NMuMG vs 4T1-2776), ****P < 0.0001). In D, data are presented as mean ± SD (n = 2 independent experiments)

Fig. 5

Proteomes of EVs derived from primary BC and site-specific MBC cell lines. A Venn diagram analyses. Sample datasets were compared for shared proteins between EVs isolated from the different cell lines. In the insert are shown samples datasets compared to EV protein cargo reported in the Vesiclepedia database (see Supplementary Table 1 for the full list of proteins, and Supplementary Table 2 for novel reported proteins). B Differential enrichment of EV-associated proteins derived from primary BC and MBC cell lines. The 100 most differentially abundant proteins in the cell-derived EVs identified by ANOVA (p-value < 0.05) are shown. (C and D) Scatter plot of the PCA results with the PC1 and PC2 scores assigned to each spectrum. This analysis includes 698 identified proteins (n = 3 independent EV preparations for each cell line). C 2D PCA scores plot of the 11 EV experimental groups. D PCA correlation loading plot. PC1 and PC2 represent the first and second principal components derived from PCA, respectively, capturing the directions of maximum and subsequent variance in the dataset. Data was normalized (Scaffold) and adjusted with Pareto scaling (MarkerView)

Fig. 6

BC EVs are enriched in proteins involved in oncogenic and metastatic processes. A Total protein counts from EVs isolated from each cell line. The analysis includes a total of 698 identified proteins. Data was retrieved from normalized TSC. Data are presented as mean ± SD (n = 3 independent experiments (Histograms)). Note that dots represented values for the three replicates. B Venn diagram analyses. Samples datasets were compared for shared proteins between EVs isolated from NMuMG and 67NR cell lines (see Supplementary Table 3 for the list of proteins). C Heatmap chart depicting the relative expression levels of proteins linked to tumorigenesis. Note that both datasets clustered differently one from the other. D GO classification of proteomic data for the differentially expressed proteins. The most enriched categories in Biological Process are shown (see Supplementary Table 4 for the list of proteins). E Venn diagram analyses. Samples datasets were compared for shared proteins between EVs isolated from non-malignant (NMuMG) and cancerous (67NR and 4T1) cells (see Supplementary Table 5 for the list of proteins). F Heatmap chart depicting the relative expression levels of proteins linked to tumorigenesis. Note that datasets clustered differently one from one other. G-I GO classification of proteomic data for the differentially expressed proteins. The most enriched categories in (G) Biological Process, H Molecular Function and (I) KEGG are shown (see Supplementary Table 4 for the list of proteins)

Fig. 7

Metastatic BC EVs are enriched in proteins involved in metastasis regulation and metastatic niche organization. A Venn diagram analyses. Sample datasets were compared for protein differential expression between EVs isolated from 67NR and 4T1 cell lines (see Supplementary Table 6 for the list of proteins). B Heatmap chart depicting the relative expression levels of proteins linked to metastasis. Note that both datasets clustered differently one from the other. C-F GO classification of proteomic data for the differentially expressed proteins. The most enriched categories in (C) Biological Process, D Cellular Component, E Molecular Function and (F) KEGG are shown (see Supplementary Table 7 for the list of proteins)

Fig. 8

Metastatic BC EVs carry proteins involved in organotropic homing. A Venn diagram analyses. Sample datasets were compared for protein differential expression between EVs isolated from MBC cell lines (see Supplementary Table 9 for the list of proteins). The graph was drawn in https://bioinformatics.psb.ugent.be/webtools/Venn/. B Heatmap chart depicting differentially expressed proteins between 67NR-EVs and MBC-EVs. C-E Heatmap charts showing differentially expressed proteins between 67NR-EVs and MBC-EVs in the categories of (C) integrins, and (D) DNA-binding proteins

Fig. 9

Overview of metastatic BC organotropism determinants and potential therapeutic targets. EV-associated proteins were differentially enriched in metastatic BC cells according to organotropism. Integrins (e.g., ITGAV/B5 for lung metastases, ITGA8/A2 for bone metastases) and annexins (e.g. ANXA3/ANXA6 in lung metastases, ANXA11/ANXA7 in liver metastases) were identified as potential biomarkers and therapeutic targets. These findings suggest that EV cargo could inform disease staging and patient stratification while providing insights into tissue-specific therapeutic strategies. Created in BioRender. Burnier, J. (2025) https://BioRender.com/nnnamw1