In vitro TME cell reprogramming by prostate cancer-derived large extracellular vesicles (L-EVs) and development of potential fluorescent L-EV tracking models

Prostate cancer (PCa) is the second most frequently diagnosed cancer in men worldwide1, ranging from indolent to lethal metastatic disease2. Paracrine and systemic signaling are major drivers in metastatic dissemination, with extracellular vesicles (EVs) emerging as key mediators of tumor microenvironment (TME) remodeling. Previous studies have shown that both large and small EVs (L- and S-EVs) released from PCa cell lines promote angiogenesis3–5, induce the transition of fibroblasts into cancer-associated fibroblasts (CAFs)5, and support immune evasion6–9. However, no comprehensive study has yet determined whether EVs of different sizes released by PCa cells representing different disease stages elicit differential stromal responses. The current study focuses on how L- and S-EVs released by PCa cell models in vitro, mirroring different stages of the disease from adenocarcinoma to androgen receptor negative neuroendocrine prostate cancer (NEPC), influence endothelial and fibroblast behaviors, and whether these effects relate to EV size. Through a series of molecular, cytofluorimetric, and imaging-based experimental approaches, we show that PCa-derived L-EVs are a major driver of both endothelial and fibroblast responses, while S-EVs induce similar but consistently weaker effects under the tested experimental conditions. In particular, L-EVs markedly enhanced tube formation capacity of HUVECs and promoted the conversion of WPMY-1 fibroblasts into inflammatory CAFs (iCAFs). Moreover, while the induction of the iCAF response by L-EVs appeared independent of the PCa cell of origin phenotype, L-EVs released by NE-like PCa cell models stimulated angiogenesis significantly more than L-EVs released by the other employed PCa cells. Given the prominent role of L-EVs in TME modulation and the limited molecular tools available for their study, DU145 cells were engineered to produce fluorescently labeled L-EVs for tracking and functional characterization. To this end, VDAC1, a mitochondrial outer-membrane protein recently identified as highly enriched in L-EVs from multiple cancers types including PCa10, was selected and genetically engineered through two approaches, specifically via i) lentiviral-mediated overexpression of recombinant VDAC1 fused to a fluorescent protein, and ii) CRISPR/Cas9-based knock-in of a fluorescent reporter at the endogenous VDAC1 locus. While both strategies yielded suboptimal levels of fluorescently labeled L-EVs, possibly due to VDAC1 presence restricted to specific L-EV subpopulations, the overexpression of the recombinant VDAC1 protein produced fluorescent L-EVs whose uptake could be successfully tracked in the recipient cells. Albeit the CRISPR/Cas9-based editing provided a more physiological labeling strategy, as confirmed by the localization of the edited fluorescent VDAC1 protein in donor cells, it resulted in a markedly lower L-EV fluorescence intensity, thereby hindering downstream vesicle tracking in recipient cells. Despite these differences, we show that L-EVs obtained through both approaches were successfully uptaken, delivered the recombinant VDAC1 protein inside the recipient cells and mediated comparable biological effects in WPMY-1 fibroblasts. Finally, our results highlight the need to combine multiple L-EV markers for more sensitive and reliable L-EV tracking using high-efficiency fluorophores. Overall, this work enhances current understanding of how PCa-derived EVs contribute to TME remodeling and introduces novel knowledge for generating a tool for the successful in vitro tracking of L-EVs.