TGF-β induces cholesterol accumulation to regulate the secretion of tumor-derived extracellular vesicles

Rodrigues-Junior, Dorival Mendes; Tsirigoti, Chrysoula; Psatha, Konstantina; Kletsas, Dimitris; Aivaliotis, Michalis; Heldin, Carl-Henrik; Moustakas, Aristidis

doi:10.1186/s13046-025-03291-0

Research
Open access
Published: 06 February 2025

TGF-β induces cholesterol accumulation to regulate the secretion of tumor-derived extracellular vesicles

Dorival Mendes Rodrigues-Junior¹,
Chrysoula Tsirigoti^1,2,
Konstantina Psatha³,
Dimitris Kletsas⁴,
Michalis Aivaliotis³,
Carl-Henrik Heldin¹ &
…
Aristidis Moustakas¹

Journal of Experimental & Clinical Cancer Research volume 44, Article number: 42 (2025) Cite this article

1429 Accesses
2 Altmetric
Metrics details

Abstract

Background

Cancer cells are avid extracellular vesicle (EV) producers. EVs transport transforming growth factor-β (TGF-β), which is commonly activated under late stages of cancer progression. Nevertheless, whether TGF-β signaling coordinates EV biogenesis is a relevant topic that remains minimally explored.

Method

We sought after specific TGF-β pathway mediators that could regulate EV release. To this end, we used a large number of cancer cell models, coupled to EV cell biological assays, unbiased proteomic and transcriptomic screens, followed by signaling and cancer biology analyses, including drug resistance assays.

Results

We report that TGF-β, by activating its type I receptor and MEK-ERK1/2 signaling, increased the numbers of EVs released by human cancer cells. Upon examining cholesterol as a mediator of EV biogenesis, we delineated a pathway whereby ERK1/2 acted by phosphorylating sterol regulatory element-binding protein-2 that transcriptionally induced 7-dehydrocholesterol reductase expression, thus raising cholesterol abundance at both cellular and EV levels. Notably, inhibition of MEK or cholesterol synthesis, which impaired TGF-β-induced EV secretion, sensitized cancer cells to chemotherapeutic drugs. Furthermore, proteomic profiling of two distinct EV populations revealed that EVs secreted by TGF-β-stimulated cells were either depleted or enriched for different sets of cargo proteins. Among these, latent-TGF-β1 present in the EVs was not affected by TGF-β signaling, while TGF-β pathway-related molecules (e.g., matrix metalloproteinases, including MMP9) were either uniquely enriched on EVs or strongly enhanced after TGF-β stimulation. EV-associated latent-TGF-β1 activated SMAD signaling, even when EV uptake was blocked by heparin, indicating competent signaling capacity from target cell surface receptors. MMP inhibitor or proteinase treatment blocked EV-mediated SMAD signaling, suggesting that EVs require MMP activity to release the active TGF-β from its latent complex, a function also linked to the EV-mediated transfer of pro-migratory potential and ability of cancer cells to survive in the presence of cytotoxic drugs.

Conclusion

Hence, we delineated a novel signaling cascade that leads to high rates of EV generation by cancer cells in response to TGF-β, with cholesterol being a key intermediate step in this mechanism.

Graphical Abstract

• TGF-β increases EV release by activating a MEK-ERK1/2-SREBP2-DHCR7 signaling and transcriptional pathway.
• TGF-β-induced DHCR7 expression raises cholesterol abundance that promotes EV release.
• EVs carry surface latent TGF-β and MMP9 that can activate TGF-β receptor signaling on the surface of recipient cells.

Blurb

The growth factor TGF-β orchestrates transcriptional signaling that enhances cholesterol biosynthesis and mediates release of extracellular vesicles. EV cargo further activate TGF-β signaling in recipient cells and promote tumor cell invasion and chemoresistance.

Background

Extracellular vesicles (EVs) are cell-derived particles confined by a lipid bilayer that mediate intercellular communication by trafficking bioactive molecules (proteins, nucleic acids, lipids and carbohydrates) from secreting to recipient cells [1]. Major attention has been given to tumor-derived EVs due to the communication between cancer cells and cells in their microenvironment, as this is crucial for local tumor progression and the development of pre-metastatic niches at distant sites [2]. Whether blockage of EV-mediated tumorigenic processes may improve future treatment of cancer patients remains to be evaluated clinically [3]. Yet, multiple reports raise the utility of EVs and their molecular cargo as biomarkers with predictive and diagnostic potential that can guide cancer treatment decisions by molecular subclassification of tumors [4].

Among molecules carried by EVs, transforming growth factor β (TGF-β) has been steadily featured in cancer EV biology [5, 6]. TGF-β and its family members signal by binding to serine/threonine kinase receptors known as type I (TGFβRI) and type II (TGFβRII) receptors, which phosphorylate SMAD proteins (SMAD2 and SMAD3) and activate non-SMAD pathways, inducing MAP-kinases, Src and phosphatidylinositol-3´kinase/Akt, to mediate cytoplasmic signaling and regulate the expression of several genes [7]. The contextual responses leading to the dual role of TGF-β signaling in breast, colorectal, pancreatic, liver and lung adenocarcinoma is exemplified by the cell growth arrest and induction of cell death mediated by apoptosis in early-stage epithelial tumors, while the cancer cells that escape TGF-β suppression respond to TGF-β with tumorigenic immunosuppression, invasion and metastasis, thus promoting tumor progression [8]. TGF-β is implicated in epithelial-mesenchymal transition (EMT) and the maintenance of stem cell populations, which play an important role in anticancer therapy resistance [8]. Moreover, TGF-β enhances tumor angiogenesis and the development of cancer-associated fibroblasts and suppresses the immune system [8]. High numbers of circulating EVs carrying TGF-β were linked to the poor prognosis of cancer patients [6]. In different disease contexts, non-senescent fibroblasts stimulated with TGF-β1 secreted higher numbers of EVs [9], whereas, TGF-β1 stimulation did not affect the release of EVs from human pulmonary artery smooth muscle cells that are implicated in arterial hypertension [10]. Nevertheless, whether and how TGF-β can modulate tumor-derived EVs in order to assist tumorigenesis, remains poorly understood.

The term EVs collectively refers to a variety of secreted vesicles [11], such as exosomes, of endosomal origin, and microvesicles, originating from the plasma membrane [12]. Membrane budding followed by a fission process generates EVs secreted inside the lumen of multivesicular bodies (MVBs) or in the extracellular milieu [12]. EV biogenesis and sorting of their cargo is regulated by extracellular stimuli, which activate intracellular signaling, e.g. MEK/ERK1/2 kinases [13], and such signals act downstream via the endosomal sorting complexes required for transport (ESCRT) machinery, tetraspanins or synthesis of lipid/cholesterol constituents of EV membranes [12]. Hence, the present study aimed at examining whether TGF-β can regulate the secretion of tumor-derived EV subpopulations and their cargo. We elucidated mechanisms of TGF-β signaling contributing to EV release from breast cancer (BRCA) and lung adenocarcinoma (LUAD) cells, providing also insight into the protein composition of these EVs and their functional relevance as facilitators of oncogenic transformation.

Methods

Cells, reagents and treatments

Cell lines of human basal-TNBC (MDA-MB-231: HTB-26; Hs578T: HTB-126), luminal BRCA (ZR-75-1: CRL-1500; MCF7: HTB-22), LUAD (A549: CCL-185), non-small cell lung carcinoma (H1299: CRL-5803) and embryonic kidney epithelial-like (HEK293T: CRL-3216) were obtained from the American type culture collection (https://www.atcc.org/). Primary cultures of human breast fibroblasts (BNF2, BNF3) were previously described [14]. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Sigma-Aldrich AB, Stockholm, Sweden) or Roswell Park Memorial Institute-1640 (ThermoFisher Scientific, Stockholm, Sweden), and supplemented with 10% fetal bovine serum (FBS; Biowest, Almeco A/S, Esbjerg, Denmark), 100 U/mL penicillin and 100 µg/mL streptomycin (Merck/Millipore, Stockholm, Sweden). The non-tumorigenic breast epithelial MCF10A cells were obtained from Dr Fred Miller (Barbara Ann Karmanos Cancer Institute, Detroit, USA) and cultured in DMEM/F12 (ThermoFisher Scientific, Stockholm, Sweden), supplemented with 5% horse serum (Biowest, Almeco A/S, Esbjerg, Denmark), 20 ng/mL epidermal growth factor (PeproTech EC Ltd/ThermoFisher Scientific, Stockholm, Sweden), 100 ng/mL cholera toxin, 0.5 µg/mL hydrocortisone, 10 µg/mL insulin (the last three from Sigma-Aldrich AB, Stockholm, Sweden). All cells were kept in a humidified incubator at 37 °C and 5% CO₂. The cell lines were checked for mycoplasma presence and authenticated through STR profiling (Eurofins, Uppsala, Sweden). The cells were starved for 24 h in serum-free media prior to treatment for different time periods with 0.2-5.0 ng/mL recombinant human TGF-β1 (PreproTech EC Ltd/ThermoFisher Scientific, Stockholm, Sweden).

Cells were treated with inhibitors (i) LY2157299 (TβRi; #15409, Cayman Chemical Co, Ann Arbor, MI, USA), U0126 (MEKi; #19–147 Merck/Millipore, Stockholm, Sweden), AY9944 (DHCR7i; #14611, Cayman Chemical Co, Ann Arbor, MI, USA), simvastatin (HMGCRi; #S6196, Sigma-Aldrich AB, Stockholm, Sweden), heparin (#7980, STEMCELL Technologies UK Ltd, Cambridge, UK), GM6001 (MMPi; #CC1010, Sigma-Aldrich AB, Stockholm, Sweden), monoclonal anti-TGF-β1/2/3 antibody (TGF-β Ab; MAB1835, R&D Systems Inc./Biotechne, Abingdon, UK), gefitinib (EGFRi; #Y0001813, Sigma-Aldrich AB, Stockholm, Sweden), lapatinib (EGFRi# CDS022971, Sigma-Aldrich AB, Stockholm, Sweden), doxorubicin hydrochloride (Dox; #D1515, Sigma-Aldrich AB, Stockholm, Sweden) or paclitaxel (Taxol; #T7402, Sigma-Aldrich AB, Stockholm, Sweden). Drug toxicity was assessed after 48 h treatment by PrestoBlue HS Cell Viability Reagent (#P50200, ThermoFisher Scientific, Stockholm, Sweden), as previously described [15]. The fluorescence units from treated cells were normalized relative to those of vehicle-treated (0.1% dimethyl-sulfoxide (DMSO)) control cells. The 10 or 50% inhibitory concentration (IC₁₀ or IC₅₀) for the drugs was estimated based on dose-response curves. Experiments were performed in triplicates and data are expressed as mean ± SD.

EV isolation and characterization

EV isolation and characterization was performed as previously described [15], following the MISEV2018 guidelines [11]. The cells were incubated for 48 h in DMEM supplemented with 10 µg/mL insulin (Sigma-Aldrich AB, Stockholm, Sweden), 1× MEM non-essential amino acids (#11140-050), 10 µg/mL fibroblast growth factor-basic (#13256-029), 1 mM sodium pyruvate (#11360-039) and 0.1 mM β-mercaptoethanol (#21985-023; all from ThermoFisher Scientific, Stockholm, Sweden), in the presence or absence of TGF-β1 or the respective drug treatment. Additionally, MCF10A cells were incubated in the culture media without horse serum. The serum-free medium was centrifuged at 1,200 × g for 5 min to clear cell debris while measuring the corresponding number of cells. The medium supernatant was filtered through a 0.2 μm filter, was further concentrated 20× by tangential flow filtration on a 50 kDa Ultra-15 Centrifugal Filter (Amicon; Merck/Millipore, Stockholm, Sweden) by centrifugation at 1,200 × g for 30 min, and defined as the vesicular secretome fraction (VSF). EVs smaller than 200 nm were sub-fractionated based on the presence of the tetraspanin CD81 protein [16] and on GM1-gangliosides, the latter having high binding affinity for cholerae toxin chain-B (CTB) [15]. VSF (150 µl) was incubated with 40 µL of anti-CD81-magnetic Dynabeads™ (#10616D, ThermoFisher Scientific, Stockholm, Sweden) or 0.5 µg of biotinylated CTB (#C34779; ThermoFisher Scientific, Stockholm, Sweden) in 100 µl of 0.2 μm filtered PBS, respectively, for 60 min at 37 °C. Pre-washed Dynabeads MyOne Streptavidin T1 (50 µl, #65602; ThermoFisher Scientific, Stockholm, Sweden) were added to the CTB-EV mixture and incubated for 30 min at 25 °C. The magnetic beads containing CD81-EVs or CTB-EVs were immobilized and the corresponding supernatant was collected as the respective specific EV-depleted fraction. The beads were washed with PBS and the isolated EVs or corresponding EV-depleted fractions were stored at -20 °C for up to 30 days.

Nanoparticle number and size distribution were measured through nanoparticle tracking analysis (NTA), (NanoSight system; Malvern Panalytical, Malvern, UK, equipped with a 532 nm laser and the NTA 3.4 analytic software). Modal diameter and per cell ratio of nanoparticles were analyzed statistically. To visualize EVs through transmission electron microscopy (TEM), VSF was mixed with an equal volume of 4% w/v formaldehyde and further incubated with uranyl oxalate, pH 7.0, for 5 min, then stained in a drop containing 4% w/v uranyl acetate and 2% w/v methylcellulose on ice for 10 min, as described [17]. Images were acquired on a Tecnai™ G2 Spirit BioTwin TEM (ThermoFisher Scientific/FEI, Stockholm, Sweden) at 80 kV with an ORIUS SC200 CCD camera and Gatan Digital Micrograph software (both from Gatan Inc./Blue Scientific, Pleasanton, CA, USA). In addition, cryogenic-transmission electron microscopy (cryo-TEM) imaging was performed to better characterize EVs. Briefly, the EVs (4.0 µL; 1 × 10¹⁰ nanoparticles/mL) were incubated twice for 30 s before blotting on a carbon QuantiFoil R 2/2 + 2 nm C (Jena Bioscience, Jena, Germany), followed by vitrification in liquid ethane using a Vitrobot Mark IV (ThermoFisher Scientific, Waltham, USA). The grids were loaded into a TFS Glacios (ThermoFisher Scientific, Waltham, USA) and the images were acquired at an acceleration voltage of 200 kV, total dose of 196 eÅ2 and defocus of -3 μm. The images were recorded by a Falcon 3EC direct-imaging camera. Moreover, CD81-EVs were suspended in 50 µl of PBS and fixed with 2% w/v formaldehyde, 2.5% w/v glutaraldehyde, 0.1 M sodium cacodylate, pH 7.2, before scanning on a FEI Quanta 250 FEG scanning electron microscope (SEM; ThermoFisher Scientific, Waltham, USA). For uptake assays, EVs present in the VSF were labelled using PKH26 red (#MIDI26-1KT, Sigma-Aldrich, Stockholm, Sweden). For functional assays on recipient cells, EVs present in the VSF or the corresponding EV-depleted fractions were used after particle (VSF) or volume equivalent (depleted fractions) directly, or after treatment with 25 µM MMPi or 40 µg/ml proteinase K for 30 min at 37 °C, the latter followed by incubation with complete protease inhibitors (#11697498001, Roche Diagnostics Scandinavia AB, Bromma, Sweden) for additional 30 min prior to their application into recipient cells.

TEM of cell monolayers

Cells were fixed in 2.5% w/v glutaraldehyde, 1% w/v formaldehyde, 0.1 M PIPES, pH 7.4, and further incubated with 1% w/v osmium tetroxide (Analytical Standards AB, Landvetter, Sweden). The cells were placed in a 1:1 mixture of Epon Resin (Ted Pella Inc., Redding, CA, USA) and 99.9% v/v ethanol for 1 h, followed by two changes of 100% resin, the first for 2–4 h and the last overnight, until polymerization was achieved at 60 °C for 48 h. Ultrathin Sects. (50–70 nm) were cut in a Leica UC7 ultramicrotome (Leica Geosystems AB, Kista, Sweden) and placed on grids. The grids were contrasted in 5% w/v uranyl acetate and Reynold’s lead citrate for 10 and 2 min, respectively. Dried grids were examined in a Tecnai™ G2 Spirit BioTwin TEM (ThermoFisher Scientific/FEI, Stockholm, Sweden) at 80 kV with an ORIUS SC200 CCD camera and Gatan Digital Micrograph software (both from Gatan Inc./Blue Scientific, Pleasanton, CA, USA).

Transferrin uptake

To assess endocytic uptake of transferrin (TRF), MDA-MB-231 and ZR-75-1 cells were placed on ice for 10 min and incubated for 1 h with 50 µg/mL Alexa Fluor-488-conjugated TRF (#T13342, ThermoFisher Scientific, Stockholm, Sweden). For flow cytometric analysis, 1 × 10⁴ live cells were counted as events according to the equipment’s setting (Acuri, BD Bioscences, Stockholm, Sweden). The samples were displayed in histograms and plotted according to the median of positive cells incorporating TRF. In addition, TRF uptake was evaluated by direct fluorescence in an inverted microscope (Nikon-Eclipse Ti-U, Nikon Europe B.V., Amstelveen, The Netherlands), equipped with a CCD camera (Andor multi pixel sCMOS camera, Oxford Instruments, Abingdon, UK). For this purpose, 50 × 10⁴ MDA-MB-231 cells were fixed with 4% w/v formaldehyde for 15 min and stained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich AB, Stockholm, Sweden).

Correlation analysis of gene expression

Comparisons of gene expression in tumor (BRCA, LUAD) versus normal samples were retrieved from the GEPIA2 server [18]. Overall survival in breast and lung cancer patients was predicted using the KM plotter database (KMplot) [19]. Samples were stratified and compared using auto-select cut-off with p-value calculated by log-Rank test. Using GEPIA2 RNA-seq cancer patient data, co-expression of TGFB1 mRNA with GSEA hallmark signatures for TGF-β signaling, EMT, protein secretion and cholesterol homeostasis in LUAD and BRCA was analyzed (Supplementary Table S1) based on the Pearson correlation statistic (R > 0.2, p-value < 0.05). Correlation of gene expression and chemotherapy response in BRCA patients was conducted using the database from ROC Plotter (http://www.rocplot.org) [20]. Patients were categorized as non- or responders to treatment, according to the relapse-free survival status at 5 years post-surgery.

RNA extraction and RT-qPCR

RNA extraction and real-time RT-qPCR were performed as previously described [21]. Gene expression levels were determined after normalization to GAPDH levels. Primer sequences of the genes are described in Supplementary Table S2.

Immunoblotting

Immunoblotting was performed as previously described [21]. The lysis buffer contained complete protease inhibitors (#11697498001, Roche Diagnostics Scandinavia AB, Bromma, Sweden), phosphatase inhibitors (PhosSTOP; #4906837001, Sigma-Aldrich AB, Stockholm, Sweden), and protein content was determined using Bradford Reagent (#5000006, Bio-Rad Laboratories Inc., Sundbyberg, Sweden). In immunoblot analysis of EVs, equal volume of VSF or CD81- and CTB-enriched populations was analyzed in order to capture the difference in EV numbers contained in these preparations. The antibodies used in this study are described in Supplementary Table S3. ImageJ bundled with Java 1.8.0_172 (National Institutes of Health, Bethesda, MD, USA) was used to normalize intensity levels according to the expression of loading control proteins GAPDH, β-ACTIN or β-TUBULIN.

siRNA transfection analysis

Once the cells reached 70% of confluence, transfections were performed with siLentFect (#170–3360, Bio-Rad Laboratories Inc., Sundbyberg, Sweden) and the respective siRNA (20 nM), according to the manufacturer. After 48 h, cells were collected for knockdown validation and seeded for further experiments. The human-specific siRNA oligonucleotides (Supplementary Table S4) were several distinct siRNAs against the target mRNA.

Total cholesterol quantification

A cholesterol quantification Kit (#MAK043, Sigma-Aldrich AB, Stockholm, Sweden) was used to measure the amount of total cholesterol. Briefly, EVs or cells treated with TGF-β1 or inhibitors were extracted with 200 µL of chloroform: isopropanol: IGEPAL (7:11:0.1) in a micro-homogenizer for 5 min. The samples were centrifuged at 13,000 × g for 10 min and the organic phase was transferred to a new tube and dried at 50 °C to remove chloroform. The dried lipids were dissolved with 200 µL of the Cholesterol Assay Buffer and vortexed until the mixture was homogeneous. In a 96-well flat-bottom plate, 50 µL of the Reaction Mix was added to the respective wells with samples or blank. The cholesterol concentration was determined by colorimetry. The data were expressed as percentage of total cholesterol normalized to protein content, determined using Bradford Reagent (#5000006, Bio-Rad Laboratories Inc., Sundbyberg, Sweden).

Liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS)

The protein profile of MDA-MB-231 CD81- and CTB-EVs was determined by LC-ESI-MS/MS and analyzed by label-free quantification, as previously described [22]. To approximately 70 µl of EV sample, 70 µl of lysis buffer was added (4% w/v SDS, 50 mM HEPES, pH 7.6, 1 mM DTT). Proteins were digested with trypsin (sequencing grade modified, Pierce/ThermoFisher Scientific, Stockholm, Sweden), using a modified protocol for SP3 protein clean-up [22], followed by SP3 peptide cleanup. Each sample was separated using a Thermo Scientific Dionex nano LC-system in a 3 h 5–40% ACN gradient coupled to Thermo Scientific High Field QExactive. The MS raw files were searched using Sequest-Percolator or Target Decoy PSM Validator under the software platform Proteome Discoverer 1.4 (ThermoFisher Scientific, Stockholm, Sweden) against the Uniprot Homo sapiens database and filtered to a 1% false discovery rate (FDR) cut-off. A precursor ion mass tolerance of 10 ppm was applied and product ion mass tolerances of 0.02 Da for HCD-FTMS and 0.8 Da for CID-ITMS. The algorithm considered tryptic peptides with maximum 2 missed cleavages, carbamidomethylation (C) as fixed modifications and oxidation (M) as variable modifications. All proteomics data have been deposited to PRIDE (https://www.ebi.ac.uk/pride/) under accession number PXD039591 and also presented in Supplementary Table S5.

Comparative label-free analysis of the proteomic profiles was performed using Perseus (version 1.6.14.0) by uploading the results obtained by the Proteome Discoverer, and following a previously established pipeline [23] with the following modifications. Prior to the analysis, the data were manually filtered in order to remove obsolete protein IDs from UniProt, duplicates and protein IDs corresponding to only one replicate. The data were normalized and transformed in logarithmic scale using log₂ transformation and all the missing values were replaced by 0 using imputation in Perseus. Principal component analysis (PCA) in Perseus was used using as a cut-off method the Benjamini-Hochberg FDR with a cut-off value of 0.05. The volcano plot was made using -log(p-value) derived by t-test and the log₂(Fold Change) of each comparison with an FDR value 0.05 and visualization with GraphPad. Protein-protein interaction networks for the unique or common proteins on CD81- or CTB-EVs (Ctrl versus TGF-β1-treatment) were constructed using STRING (version 11.5; http://string-db.org/), with the required high confidence score (> 0.7). Subsequent GO-biological process (BP), molecular function (MF), cellular component (CC) and Reactome enrichment analyses were performed.

ELISA

Mature TGF-β1 (referred also as active TGF-β1) was measured using the human TGF-β1-Duoset ELISA sandwich kit according to the manufacturer’s instructions (#DB100C; R&D Systems, Oxon, UK). Acidification with 1 N HCl, followed by neutralization with 1.2 N NaOH/0.5 M HEPES at room temperature, was used to release the active TGF-β1 from its latent complex and measure the total TGF-β1 in the sample. Recombinant human latent TGF-β1 (L-TGF-β1: #SRP0300; Merck, Stockholm, Sweden) was used to monitor TGF-β1 activation induced by EVs.

Gelatinase and collagenase proteolytic activity

The EnzChek™ Gelatinase/Collagenase Assay kit (ThermoFisher Scientific, Stockholm, Sweden) was used to evaluate gelatinase or collagenase activity associated to EVs. Briefly, 100 µg/mL of DQTM fluorescein-conjugate gelatin solution, was incubated for 1 h at room temperature with VSF, VSF depleted from CD81-positive EVs or in the presence of MMPi. Proteolytic activity, which is proportional to fluorescence intensity, was measured with a microplate reader (495 nm excitation/515 nm emission). For each condition, background fluorescence was corrected by subtracting the blank value. EV-free media was used as negative control and the relative normalized gelatin/collagen proteolytic activity is expressed as averages from biological triplicates.

Luciferase assays

Luciferase assays in MDA-MB-231 and MCF10A cells transiently transfected with the CAGA₁₂-luciferase promoter-reporter and treated with TGF-β1, EVs, EV-depleted fractions or inhibitors were performed using the firefly and renilla dual-luciferase Assay kit (Biotium, Fremont, CA, USA), as described [24]. Relative normalized luciferase activity is expressed as averages from triplicate determinations. Each experiment was repeated at least twice.

Cell culture wound-healing assay

To evaluate migration, cells treated with TGF-β1 or EVs for 48 h were seeded (3 × 10⁴ cells/well) in complete medium into Culture-Insert 2 well 24 (#80242; Ibidi GmBh, Gräfelfing, Germany), as previously described [24]. The silicone insert was removed once the cells reached confluence and detached cells were removed by washing twice with PBS, followed by addition of fresh culture media. Wound closure was observed at different time points using a Zeiss Axioplan microscope (objective 10×) with MRC digital camera (Zeiss, Jena, Germany). Wound surface area was quantified by ImageJ as the percentage of open wound area per condition.

Matrigel invasion assay

Transwell inserts (#351152) for 24-well plates (6.5 mm diameter, 8 μm pore) were coated with 300 µg/mL Matrigel matrix (#734 − 0269; both from Corning, New York City, NY, USA) diluted in coating buffer (10 mM Tris, pH 8.0, 0.7% w/v NaCl) and incubated at 37 °C for 1 h. Cells (5 × 10⁴) treated with VSF, VSF depleted from CD81-positive EVs or MMPi for 48 h were seeded in serum‐free DMEM in the upper chamber, and DMEM/10% FBS was placed in the lower chamber and incubated at 37 °C to allow the invasion through the Matrigel barrier for 18 h. After incubation, the inserts were fixed in methanol and stained with DAPI (Sigma-Aldrich AB, Stockholm, Sweden). Nuclei were counted in 10 pictures per insert, taken with the 20× objective, using ImageJ. Data were expressed as the percentage of invasion based on the ratio of the mean number of cells invading through Matrigel matrix per mean number of cells in the uncoated support.

Immuno- and direct fluorescence microscopy

The cells were seeded in a 12-well plate followed by treatments with TGF-β1, VSF or VSF^+ PKH26 for the indicated time periods. Immunofluorescence was performed as previously described [25]. Fixed cells were incubated with primary antibodies (Supplementary Table S3) in 1% BSA/PBS overnight at 4 °C, followed by incubation with Alexa-Fluor-488 or Alexa-Fluor-546 secondary antibodies (Supplementary Table S3) at a dilution of 1:500 in PBS for 1 h at room temperature. Cell nuclei were stained with DAPI (Sigma-Aldrich AB, Stockholm, Sweden) at a dilution 1:1,000 in PBS for 10 min at room temperature. The cells were examined by an inverted microscope (Nikon-Eclipse Ti-U, Nikon Europe B.V., Amstelveen, The Netherlands) equipped with a CCD camera (Andor multi pixel sCMOS camera, Oxford Instruments, Abingdon, UK). Ten to 15 random pictures were taken with 10× or 20× objectives at the same exposure time. When appropriate, the scores were given in a blind manner.

Zebrafish extravasation assay

Staging and embryo production of Tg(Fli1:EGFP) zebrafish (Danio rerio), whose vasculature is marked in green, were conducted as described [25]. Empirical determination of sample size that provided power for discrimination between conditions, was used. For this reason, 200 embryos were injected per condition in order to reach a final number of viable embryos of more than 100. No randomization method was applied and the microinjector was blinded to the groups of injected cancer cells. MDA-MB-231 cells treated or not with VSF^Ctrl or VSF^{+ TGF−β1} were stained with 4 ng/µl CM-Dil Dye (ThermoFisher Scientific, Stockholm, Sweden) for 30 min at 37 °C. At 48 h post-fertilization, approximately 400 CM-Dil Dye-labelled cells in the presence or absence of doxorubicin (5 µM) were micro-injected as described [25]. Based on pre-established criteria, only microscopy-verified, correctly injected and viable zebrafish were retained at 34 °C and imaged automatically (ImageXpress Nano, Molecular Devices) 24 and 48 h post-implantation; cells extravasated from the circulation at the posterior part of the fish were counted (150 zebrafish per cell line and biological condition).

Statistical analyses

All data (except when otherwise indicated) are presented as the mean ± SEM. Comparisons were performed using one-way or two-way ANOVA, followed by multiple paired comparisons conducted by means of the Tukey’s or Bonferroni’s post-test method when applicable. The Mann–Whitney test was used to evaluate associations between gene expression and chemotherapy treatment response. Additional statistical methods are explained in individual figure legends. The data were analyzed with GraphPad Prism 10.1 (GraphPad Software, San Diego, CA). A p-value < 0.05 was necessary to determine statistically significant differences.

Results

TGF-β induces EV release

TGF-β preserves integrity of tissues and has different functions during cancer development depending on cancer type and stage [8]. KMplot analysis showed a correlation between high TGF-β1 protein levels and poor overall survival of BRCA patients (Supplementary Fig. S1A), consistent with the pro-tumorigenic role of TGF-β in BRCA. BRCAs are classified into several subtypes, including luminal-like BRCAs expressing estrogen receptor-α (ER) and/or progesterone receptor (PgR), and lacking expression of the human epidermal growth factor receptor 2 (HER2), HER2⁺ BRCAs and basal-like, HER2/ER/PgR-negative, known as triple-negative (TNBC) BRCAs [26]. We investigated the role of TGF-β signaling in the modulation of tumor-derived EVs using as model systems BRCA cell lines from distinct subtypes, the TNBC MDA-MB-231 and Hs578T, and the luminal-B ZR-75-1. Upon stimulation with TGF-β1 for 48 and 120 h, secreted EVs smaller than 200 nm were evaluated initially for their size and number through NTA (Supplementary Fig. S1B shows representative nanoparticle traces). Although no significant difference in the EV modal size was noted (Fig. 1A), TGF-β1 stimulation led to a significant increase in the number of EVs (EV^{+ TGF−β1}) secreted by the three BRCA cells; concomitant treatment with a TGF-β receptor type I kinase inhibitor (TβRi, 2.0 µM) blocked the release of EV^{+ TGF−β1} (Fig. 1B). Similarly, TGF-β1 stimulation did not impact the modal size, but increased significantly the numbers of EV^{+ TGF−β1} released by A549 (LUAD), H1299 (non-small cell lung carcinoma), MCF7 (luminal-A BRCA), MCF10A (non-tumorigenic breast epithelial) and BNF3 (normal breast fibroblast) cells (Supplementary Fig. S1C), suggesting a general stimulatory effect of TGF-β on EV secretion by different tumor and non-tumorigenic cells. In line with this finding, there was a significant and positive correlation (R = 0.6060, p = 0.0392) between the mean expression of cellular TGFB1 mRNA measured in each individual cell model (Supplementary Fig. S1D) and the mean number of nanoparticles released by each of the seven cell types examined (Fig. 1C), suggesting that autocrine TGF-β1 increases the titer of secreted EVs. It should be noted that the MCF7 cells were not included in this correlation analysis and the absolute number of nanoparticles monitored in each one and the same cell model could vary over experiments (Fig. 1, Supplementary Fig. S1). In contrast, the impact of TGF-β stimulation, biological or chemical perturbations (see below) always resulted in reproducible relative differences on nanoparticle numbers.

The EVs were isolated as three complementary pools, i.e. VSF covering many EV populations, CD81-EVs, enriched in the tetraspanin CD81, and cholerae toxin chain-B-selected EVs (CTB-EVs), enriched in GM1-gangliosides. Similar volume of collected MDA-MB-231 (MDA EVs), Hs578T, A549 and H1299 EV populations were characterized based on the presence of the EV markers TSG101, CD81, CD9, ALIX and β-ACTIN, whose increased levels upon TGF-β stimulation reflect the higher EV numbers contained in the same volume of EV preparation (Fig. 1D, Supplementary Fig. S1E-G). Additionally, MDA EVs were visualized by cryo-TEM demonstrating lipid bilayer continuity, particulate/filamentous luminal content, a rich surface decoration reminiscent of a glycocalyx and an apparent diameter of 60–140 nm, without obvious architectural or structural differences caused by TGF-β stimulation (Fig. 1E, Supplementary Fig. S1H). MDA CD81-EVs on immunobeads were characterized by SEM (Fig. 1F), whereas Hs579T, ZR-75-1, MCF7, MCF10A, A549 and H1299 EVs were visualized by TEM (Supplementary Fig. S1I), confirming relatively homogeneous populations of membrane-surrounded nanoparticles.

To identify endosomal compartments, MDA-MB-231 cells were incubated with bovine serum albumin (BSA)-Au particles, revealing that TGF-β stimulation did not alter the morphology of MVBs (Supplementary Fig. S2A). However, in the absence of 3D analysis of the ultrathin sectioned cells we were limited to accurately analyze MVB size or number of intraluminal vesicles. Moreover, we performed experiments of transferrin (TRF) uptake by MDA-MB-231 and ZR-75-1 cells, as measured by flow cytometry and direct fluorescence microscopy (Supplementary Fig. S2B-D). Since TGF-β1 stimulation did not affect TRF uptake it is unlikely that the higher release of EV^{+ TGF−β1} is due to higher rates of endosomal trafficking towards MVBs.

MEK/ERK1/2 signaling mediates EV secretion in response to TGF-β

To better understand the molecular signaling driving EV^{+ TGF−β1} secretion, a time-course experiment was performed including 6, 24, 48 and 120 h of stimulation of MDA-MB-231 cells with TGF-β1. TGF-β1 stimulation for 24 h or longer was required to enhance release of EVs (Supplementary Fig. S2E), which had similar morphologies at all time points tested (Supplementary Fig. S2F). We scrutinized the kinetics of TGF-β-induced EV secretion by stimulating cells with TGF-β1 followed by addition of TβRi at different time points. TβRi addition 8 h post-TGF-β successfully abrogated EV^{+ TGF−β1} release, while blocking signaling 24 h after its onset was ineffective, as noted by NTA and ALIX, TSG101 and CD81 expression (Supplementary Fig. S2G, H). We conclude that enhanced secretion of EVs by TGF-β follows slow kinetics and is dependent on factors activated during the first 8 h of TGF-β signaling.

Silencing of SMAD2, SMAD3 or both in MDA-MB-231 cells (Fig. 1G, Supplementary Fig. S3A, B) followed by TGF-β1 stimulation for 48 h, did not block the number of EV^{+ TGF−β1} (Fig. 1H), suggesting that TGF-β-induced SMAD signaling may not be required to enhance EV release. The effectiveness of the SMAD2 and SMAD3 silencing was verified by analyzing the expression of established gene targets of TGF-β signaling, MMP9 (matrix metalloprotease 9), whose induction by TGF-β1 was reduced after SMAD3 (but not SMAD2) silencing, and whose biological action will be scrutinized deeper later, and TGFBI (TGF-β-induced), whose induction by TGF-β was reduced after SMAD2 (but not SMAD3) silencing (Supplementary Fig. S3C, D). Thus, the absence of a clear SMAD2/3 impact on EV^{+ TGF−β1} secretion indicates that alternative non-SMAD signaling could be induced by TGF-β, such as the MEK/ERK1/2 pathway (see introduction). A MEK inhibitor (MEKi; U0126) that exhibited mild 0–20% toxicity in MDA-MB-231, ZR-75-1, A549 or embryonic HEK293T cells at 5 µM (Supplementary Fig. S4A), blocked basal and TGF-β1-stimulated ERK1/2 phosphorylation in MDA-MB-231 cells, without perturbing SMAD2 phosphorylation (Fig. 1I). MEKi decreased the number of MDA EVs maximally by 50%, whereas, TGF-β1 stimulation for 48 h was unable to induce any measurable EV secretion in the presence of MEKi (Fig. 1J). Similar findings were observed in ZR-75-1 and A549 cells treated with MEKi for 48 h (Supplementary Fig. S4B-E). It is worth noting that while the MEKi reduced basal levels of EV release in BRCA cells, it did not affect significantly the release of EVs by A549 cells (LUAD), whereas the negative impact of the MEKi on EV^{+ TGF−β1} secretion applied to all cell models tested. In order to monitor specificity of MEK/ERK1/2 activation downstream of TGF-β receptors, we asked whether EGF receptors that are well-established alternative activators of MEK/ERK1/2 signaling, could mediate the TGF-β effects on EV secretion. By employing two independent EGFR tyrosine kinase inhibitors, gefitinib or lapatinib, at concentrations that were not toxic to A549 and MDA-MB-231 cells (Supplementary Fig. S4F, G), we found that neither of the two chemicals inhibited ERK1/2 phosphorylation by TGF-β1 or TGF-β1-stimulated EV secretion (Fig. 1K, L, Supplementary Fig. S4H, I). On the other hand, fibronectin, whose protein expression in response to TGF-β is known to involve crosstalk between TGF-β and EGF signaling [27], was reduced partially when the two cell models were treated with the two different EGFR inhibitors (Fig. 1K, Supplementary Fig. S4H). Therefore, our data support the notion that TGF-β stimulation enhances the release of tumor-derived EVs, via sustained MEK/ERK1/2 and most likely additional late signaling mediators.

TGF-β-induced EV release involves increased cholesterol accumulation

The human tumor datasets of TCGA were used to identify possible links between TGFB1 mRNA expression and the signature of genes related to different biological pathways, including pathways that could orchestrate EV biogenesis, such as cholesterol homeostasis [12]. As expected, the gene set enrichment analysis (GSEA) of BRCA and LUAD datasets revealed positive and significant correlations between TGFB1 mRNA expression and the signature of genes classified as hallmarks for TGF-β signaling and EMT (Fig. 2A, Supplementary Fig. S5A, Supplementary Table S1; note that the correlations were stronger in LUAD and reached levels of up to 0.54 but not higher). Using TGF-β signaling and EMT terms as positive controls helped us point our attention to the significant correlation between TGFB1 levels and the gene signature of cholesterol homeostasis in both BRCA and LUAD datasets (Fig. 2B, Supplementary Table S1; note however the lower R values). Cholesterol is enriched in the membrane of microvesicles and exosomes and is one of the central contributors to EV biogenesis and release [12, 28]. This established knowledge stimulated further analysis in the direction of cholesterol biosynthesis. Furthermore, the significant correlation between TGFB1 mRNA expression and the cholesterol homeostatic gene signature (Fig. 2B), is consistent with previous findings, outside the context of EV biogenesis, on regulation of expression of genes modulating cholesterol biosynthesis and distribution by TGF-β signaling [29, 30]. In line with these reports from non-cancerous cell types, we measured a significant increase of total cholesterol levels in MDA-MB-231, ZR-75-1 and A549 cells after 48 h of TGF-β1 stimulation (Fig. 2C, D; Supplementary Fig. S5B). Additionally, MEKi decreased the intracellular levels of cholesterol in MDA-MB-231 and A549 cells by 25–50% relative to the control, and abrogated completely the TGF-β1-induced increase in cholesterol level (Fig. 2C, D), indicating that the MEK/ERK1/2 pathway mediates both increase of cellular cholesterol levels and EV secretion by TGF-β1. Interestingly, TGF-β1 stimulation also enhanced significantly the total cholesterol content of EVs secreted by MDA-MB-231 and A549 cells (Fig. 2E; Supplementary Fig. S5C).

To uncover the mechanism of total cholesterol induction by TGF-β1, we assessed the mRNA expression of 9 out of the 73 genes from the GSEA cholesterol homeostasis gene signature, i.e. DHCR7, FABP5, HMGCR, LDLR, MVD, PMVK, SC5D, SQLE and STARD4 (Supplementary Table S1, Supplementary Fig. S5D) in MDA-MB-231, ZR-75-1, MCF10A and A549 cells treated with TGF-β1 for 48 and 120 h (Fig. 2F; Supplementary Fig. S6A-C). DHCR7 and SQLE were significantly induced by TGF-β1 in all breast epithelial cells, while FABP5, HMGCR, LDLR, MVD, PMVK, SC5D and STARD4 were also increased after TGF-β1 stimulation, but in a cell type-dependent manner. Critically, TGF-β signaling required MEK/ERK1/2 activation in order to induce the expression of DHCR7, FABP5 and SQLE in MDA-MB-231 and ZR-75-1 (Fig. 2G; Supplementary Fig. S6D) cells. DHCR7 protein analysis in MDA-MB-231 (Fig. 2H), ZR-75-1 and A549 (Supplementary Fig. S6E, F) cells validated the mRNA expression analysis.

The expression of genes involved in cholesterol synthesis, including DHCR7 and SQLE, is regulated by the sterol regulatory element-binding protein-2 (SREBP2) [31], whose phosphorylation is induced by ERK1/2 [32]. Analysis of the time course of SREBP2 phosphorylation in response to TGF-β1 in MDA-MB-231 (Fig. 2I) and A549 (Supplementary Fig. S6F) cells, revealed accumulation of phospho-SREBP2 after TGF-β1 stimulation at 24 and 48 h, but not at 6 h. Importantly, the presence of MEKi successfully blocked the phosphorylation of SREBP2 in MDA-MB-231 (Fig. 2H) and A549 (Supplementary Fig. S6F) cells, suggesting that TGF-β induces the activation of SREBP2 via MEK/ERK1/2-mediated phosphorylation. Consistently, the time course of DHCR7 mRNA expression in response to TGF-β stimulation showed late regulation (after 6 h; Fig. 2J), which matched well the induction of phosphorylation of SREBP2 and the kinetics of EV induction by TGF-β. Notably, SREBP2 silencing in MDA-MB-231 cells (Fig. 2K), decreased the total cholesterol levels and EV secretion (Fig. 2L, M), while in the absence of SREBP2, TGF-β1 stimulation failed to increase either cholesterol or EV secretion and to induce the expression of DHCR7 and SQLE (Fig. 2L-N), without affecting SERPINE1 induction (Supplementary Fig. S6G). Note that SERPINE1, also known as plasminogen activator inhibitor 1 (PAI1; Fig. 2F-I; Supplementary Fig. S6A-D, F-H), whose transcription is induced after formation of a p53/TFE3/Sp1/SMAD complex bound to the SERPINE1 promoter in response to TGF-β [33,34,35], was used as a positive control for TGF-β signaling activation. Moreover, we silenced SMAD2, SMAD3 or both (efficiency of protein silencing shown in Supplementary Fig. S6I). SMAD2 or SMAD2/3 silencing did not perturb the TGF-β-induced expression of DHCR7 and SQLE in MDA-MB-231 cells, whereas SMAD3 silencing led to a strong decrease in their basal expression (Supplementary Fig. S6J). It should not go unnoticed that TGF-β modulated LDLR mRNA expression, which appeared to be context-dependent, as its levels were enhanced only in MCF10A but not in the tumor cells (Fig. 2F, G; Supplementary Fig. S6A-D). SMAD-specific interference also failed to provide convincing evidence for a transcriptional input to LDLR gene regulation (Supplementary Fig. S6J). We conclude that a TGF-β receptor to MEK/ERK1/2 to SREBP2 signaling cascade leads to induction of cholesterol biosynthesis and secreted EV accumulation.

DHCR7 contributes to EV release in response to TGF-β

The expression of the TGF-β-regulated genes (DHCR7, FABP5, HMGCR, MVD, SQLE and STARD4) was further correlated with clinical features using cancer patient datasets (Supplementary Fig. S7-S9). Among these genes, only DHCR7 was highly expressed in both BRCA and LUAD patients in comparison to healthy controls (Supplementary Fig. S7), and its higher expression significantly correlated with (i) the poor overall survival of BRCA (Supplementary Fig. S8A) and lung cancer (Supplementary Fig. S9A) patients, and (ii) with the lack of response to chemotherapies in BRCA patients (Supplementary Fig. S8B). Therefore, the cumulative data pointed to the clinical relevance of DHCR7, the rate-limiting enzyme that catalyzes the conversion of 7-dehydrodesmosterol to desmosterol in the Bloch pathway and the conversion of 7-dehydrocholesterol to cholesterol in the Kandutsch-Russell pathway (Supplementary Fig. S5D) [36].

We investigated whether DHCR7 regulation could be a central mechanism by which TGF-β signaling increases cholesterol levels and EV^{+ TGF−β1} secretion. Silencing of DHCR7 in MDA-MB-231 (Supplementary Fig. S10A, B; siDHCR7_C being the only siRNA decreasing DHCR7 expression effectively) and ZR-75-1 (Supplementary Fig. S10C) cells, led to a 50% decrease of total cholesterol levels in MDA-MB-231 in the absence of TGF-β (Fig. 3A), whereas no significant reduction in basal total cholesterol level was seen in ZR-75-1 cells (Fig. 3B). Upon stimulation with TGF-β1, total cholesterol levels remained low and suppressed in both cell models (Fig. 3A, B). Importantly, DHCR7 silencing decreased the number of EVs secreted from MDA-MB-231 and ZR-75-1 cells by 20% and 50% respectively (Fig. 3C, D), confirming the established role of cholesterol on EV biogenesis. Notably, after DHCR7 silencing, TGF-β1 stimulation did not increase the EV levels (Fig. 3C, D).

In order to validate independently the importance of DHCR7 silencing in BRCA cells, a DHCR7 chemical inhibitor (DHCR7i; AY9944) was employed at a non-toxic IC₁₀ concentration (20 µM) (Supplementary Fig. S10D). In contrast to DHCR7 knockdown, DHCR7i did not reduce the total cholesterol levels, although it efficiently blocked the cholesterol induced by TGF-β in MDA-MB-231 cells (Fig. 3E). Furthermore, treatment with DHCR7i decreased the basal numbers of EVs released by 50% and TGF-β1 stimulation failed to enhance EV release in either of the BRCA cells (Fig. 3F, Supplementary Fig. S10E). Finally, we administered simvastatin, an inhibitor of HMG-CoA reductase (HMGCR), another rate-limiting enzyme that initiates cholesterol biosynthesis (Supplementary Fig. S5D; cytotoxic response shown in Supplementary Fig. S10F). Simvastatin at 2.5 µM effectively decreased the total amount of cholesterol (Supplementary Fig. S10G), and blocked completely the induction of total cholesterol by TGF-β1 in MDA-MB-231 cells (Fig. 3E). Moreover, simvastatin decreased the number of EVs secreted by MDA-MB-231 cells by 60% and prevented any TGF-β-mediated induction of EV^{+ TGF−β1} release, phenocopying the DHCR7i (Fig. 3F). Simvastatin had a similar effect on EV secretion in the LUAD A549 cells but did not neutralize the effect of TGF-β1 (Supplementary Fig. S10H).

Although the primary effect of TGF-β signaling characterized in this study is the enhanced production of EVs, we also sought to explore possible effects of the compounds that blocked the release of EV^{+ TGF−β1} from tumor cells in the context of doxorubicin (Dox) or paclitaxel (taxol) cytotoxicity. Of note, resistance to anticancer drugs is facilitated by complex mechanisms, including TGF-β, EMT and tumor-derived EVs [6, 8]. Thus, upon examining cytotoxicity to Dox and taxol via dose-response analysis (Supplementary Fig. S10I, J), MDA-MB-231, ZR-75-1 and A549 cells were treated for 48 h in the absence or presence of MEKi, DHCR7i or simvastatin (Fig. 3G, Supplementary Fig. S10K, L). The combination of inhibitors with the cytotoxic drugs led to a significant decrease in tumor cell viability when compared to the chemotherapeutic treatment in the absence of the MEK and DHCR7 inhibitors (Fig. 3G, Supplementary Fig. S10K, L). Cytotoxicity correlated to higher levels of cleaved PARP and caspase3 in MDA-MB-231 cells treated with Dox, suggesting that cell death was mediated by apoptosis (Fig. 3H). These data show that inhibitors of TGF-β-induced EV secretion potentiate the cytotoxic effect of chemotherapeutic drugs. Altogether, our data suggest that, mechanistically, TGF-β1 activates MEK/ERK1/2 signaling, which in turn phosphorylates SREBP2 and enhances expression of enzymes involved in cholesterol synthesis, including DHCR7, which positively regulates tumor-derived EV release (Fig. 3I). Thus, regulation of DHCR7 expression and cholesterol elevation provide the intermediate link for the long-term effect that TGF-β signaling exhibits in regulating EV release.

TGF-β signaling modulates a subset of the EV cargo proteome

We investigated whether TGF-β could also change constituents of the EV protein cargo. To this end, the total protein content of isolated CD81-EVs and CTB-EVs from MDA-MB-231 cells was subjected to LC-ESI-MS/MS. PCA revealed that both CD81-EV^{+ TGF−β1} and CTB-EV^{+ TGF−β1} acquired different profiles of protein content after 120 h of TGF-β1 stimulation, compared to EV^Ctrl (Fig. 4A, B). Moreover, the protein content differed among CD81-EV and CTB-EV subpopulations secreted by MDA-MB-231 cells under the same condition (Supplementary Fig. S11A, B). From the 985 proteins detected in the subpopulation of CD81-EVs, 303 were unique in EV^Ctrl, while 143 proteins were specific in EV^{+ TGF−β1} (Fig. 4C). Additionally, 2,392 proteins were detected in the CTB-EV fraction, of which 319 proteins were exclusive to EV^Ctrl and 275 were unique to EV^{+ TGF−β1} (Fig. 4D). Lists of unique proteins in CD81-EVs and CTB-EVs are shown in Supplementary Tables S6 and S7, respectively. The common proteins carried by CD81-EV (n = 539, 54.72%) and CTB-EV (n = 1,797, 75.12%) fractions are shown as volcano plots (Fig. 4E, F). Besides the expected enrichment of “extracellular vesicle/exosome”-related terms, these proteins were significantly clustered into key pathways using the STRING algorithm, based on Gene Ontology (GO) of Cellular Component (CC), Biological Process (BP), Molecular Function (MF) and Reactome (REAC) terms. Associations to “cellular process”, “protein/RNA binding”, “cell adhesion molecule”, “cadherin binding”, “infectious disease” and “immune system” were found in both EV subpopulations (Supplementary Fig. S11C, D), indicating a possible redundant function transferred by different tumor-derived EVs to recipient cells.

The unique proteins identified in the CD81- and CTB-EV cargoes of the EV^Ctrl and EV^{+ TGF−β1} groups shared analogous CC terms relevant for EV networks (Fig. 4G, H, Supplementary Fig. S11E, F). However, despite the fact that identical MF terms were found in CD81-EVs, such as cadherin and cell adhesion molecule binding, it was evident that CD81-EV^{+ TGF−β1} were enriched in MF terms associated with TGF-β signaling (e.g., “growth factor binding”, “signaling receptor binding” and “TGF-β binding”), including the BP terms, “response to wounding and cellular response to growth factor stimulus”, when compared to CD81-EV^Ctrl (Fig. 4G, H). Similarly, CTB-EV^{+ TGF−β1} were enriched in BP terms associated with TGF-β signaling, such as “TGF-β receptor signaling pathway”, “cellular response to TGF-β stimulus”, “response to growth factor”, “cellular response to growth factor stimulus” and “transmembrane receptor serine/threonine kinase signaling” (Supplementary Fig. S11E, F). Both CD81-EV^{+ TGF−β1} and CTB-EV^{+ TGF−β1} protein profiles were enriched in REAC terms of “extracellular matrix organization”, “elastic fiber formation” and “collagen degradation” (Fig. 4H, Supplementary Fig. S11F). Furthermore, key proteins of the unique groups belonged to GO terms also shared by the common groups. The latter terms of “extracellular exosome”, “extracellular vesicle” and “extracellular region” included one of the highest top-hit proteins, MMP9 (Fig. 4F and see further analysis below).

Of note, TGF-β1 and TGF-β2 were detected in both CD81- and CTB-EV subpopulations, but without significant differences among the groups and below the cut-off thresholds of the volcano plots (Fig. 4E, F). Thus, we validated our MS analysis, confirming by ELISA that TGF-β1 protein is carried by EVs without significant differences among the total TGF-β1 carried by Ctrl or TGF-β-induced EVs (mean average of 243,74 pg/mL, equivalent to 1 × 10⁹ nanoparticles) (Fig. 4I). Nevertheless, it is important to mention that the levels of active TGF-β1 carried by EV^{+ TGF−β1} were significantly higher compared to EV^Ctrl, even though the majority of TGF-β1 associated to EVs is present in the latent complex (Fig. 4I). Altogether, the cumulative proteomic data confirm that activation of TGF-β signaling can modify the sorting of protein cargo of tumor-derived EVs and that TGF-β signaling and extracellular matrix regulators, such as MMPs, are important constituents of such tumor-derived EVs.

Next, we sought to understand better the loading of protein cargo found uniquely or enriched into different EV fractions by comparing the presence of specific EV-associated proteins with the respective cellular mRNA expression upon TGF-β stimulation for 48 and 120 h in the MDA-MB-231 and ZR-75-1 cells (Fig. 4J, Supplementary Fig. S12A). Four genes CDK1, BMP4, NOG and TGFBR3, whose proteins were more abundantly enriched in EV^Ctrl, and five genes, ITGB5, MMP2, MMP9, MMP10 and TGFBI, whose proteins were more abundantly enriched in EV^{+ TGF−β1}, were selected. Among these are genes encoding the TGF-β co-receptor TGFBR3/β-glycan, the latent TGF-β-activating enzymes MMP2 and MMP9 and TGFBI, the collagen-associated protein of the extracellular matrix that is a highly responsive gene to TGF-β signaling, as described above. MDA-MB-231 and ZR-75-1 cells stimulated with TGF-β1 presented a significant reduction of CDK1, BMP4 and TGFBR3 expression, and at the same time, MMP2, MMP9 and TGFBI were significantly enhanced. Of importance, not all the EV protein cargo reflected the transcriptional effects induced by TGF-β, as noted for NOG, ITGB5 and MMP10, suggesting that sorting of different cargo proteins into tumor-derived EVs does not strictly follow the abundance of the respective mRNA.

One of the common proteins between EV^Ctrl and EV^{+ TGF−β1}, which was also highly enriched in EV^{+ TGF−β1} was MMP9 (Fig. 4C, F, Supplementary Table S6), a matrix metallopeptidase highly expressed in BRCA and LUAD compared to non-tumorigenic samples (Supplementary Fig. S12B). Moreover, MMP9 mRNA expression correlated with poor overall survival of TNBC and LUAD patients (Supplementary Fig. S12C, D), but did not correlate with chemotherapy response in TNBC patients (Supplementary Fig. S12E). The relevance of MMP9 mRNA expression in predicting TNBC and LUAD patient outcome, suggests that EV-associated MMP9 could possibly identify circulating EVs in cancer patients. In addition, the co-presence of MMP9 and latent TGF-β1 (Fig. 4I) as cargo of EVs may be linked to the mechanism of activation of latent TGF-β carried by EVs [7]. In general, the proteomic profile revealed that MDA-MB-231-derived EVs carry a complexity of signaling molecules which could deliver simultaneously actionable pathways in recipient cells.

EV internalization is not required to activate TGF-β signaling

To examine the potential for intercellular communication mediated by MDA EVs, we first characterized the routes of EV interaction with recipient cells by labelling the EV membranes with the lipid dye PKH26 [37]. No significant difference was found in the modal size of fluorescent MDA EVs, suggesting lack of aggregation after labelling (Supplementary Fig. S13A). We then monitored the internalization of fluorescent MDA EVs to recipient MDA-MB-231 cells (homotypic transfer) or to MCF7 and MCF10A (heterotypic transfer) immunostained for EEA1, a scaffold protein of early endosomes. At the earlier time point (1–3 h), only a fraction of fluorescent MDA EV^Ctrl and EV^{+ TGF−β1} co-localized with EEA1, suggesting endocytic uptake (Fig. 5A; Supplementary Fig. S13B, C), while additional cellular compartments appeared to interact with fluorescent MDA EVs, which suggests either membrane fusion, ligand-receptor interaction or endosomal escape. At the later time point of 18 h, a dramatic decrease of fluorescent MDA EVs was observed in the cells, suggesting that EVs were taken up at the early time points and were then processed intracellularly, leading to the loss of the lipophilic PKH26 label (Fig. 5A; Supplementary Fig. S13B, C). Additionally, the use of heparin (50 µg/mL), shown previously to prevent EV uptake by recipient cells [38], successfully decreased the uptake of fluorescent MDA EVs not only in recipient MDA-MB-231 cells, but also in ZR-75-1 and MCF10A cells (Supplementary Fig. S13D-F).

We monitored quantitatively whether similar numbers of MDA EV^Ctrl or EV^{+ TGF−β1} (1 × 10⁹ nanoparticles) could activate TGF-β signaling. This was determined by the CAGA₁₂-luciferase reporter, which is highly specific and monitors only SMAD transcriptional activity, in other words, TGF-β family-specific signaling [24]. The EVs induced TGF-β signaling in recipient MDA-MB-231 cells after 8 h incubation, while depletion of the CD81-positive EV population reduced the effect of EVs on TGF-β signaling by approximately 30%, suggesting that approximately one third of the measured response depends on CD81-specific EV subpopulations (Fig. 5B, Supplementary Fig. S14A). Of note, although no significant differences were observed between EV^Ctrl and EV^{+ TGF−β1} when 10⁹ nanoparticles per mL were used, dose-dependent experiments revealed a higher potency of the EV^{+ TGF−β1} to activate the reporter under non-saturating conditions (0.05–0.5 × 10⁹ nanoparticles per mL; Fig. 5C). Furthermore, when EV uptake was blocked by heparin, MDA EVs still activated the CAGA₁₂-luciferase reporter (Fig. 5B). On the other hand, TGF-β signaling induced by MDA EVs was significantly decreased when the EVs were pre-incubated with a neutralizing monoclonal antibody against TGF-β1/2/3 (Fig. 5B); the anti-TGF-β Ab concentration used (5 µg/mL) was sufficient to block effectively up to 1 ng/mL of recombinant mature TGF-β1 (Supplementary Fig. S14B) and almost completely 5 ng/mL (Fig. 5B). As expected, the presence of TβRi completely blocked CAGA₁₂-luciferase reporter activity under all conditions (Fig. 5B). Moreover, upon 8 h treatment, MDA EVs also induced TGF-β signaling in MCF10A cells (Supplementary Fig. S14C). Additionally, the phosphorylation levels of SMAD2, SMAD3 and ERK1/2 were increased after 1 and 6 h of EV interaction with recipient cells (Fig. 5D). Monitoring at a later time point, 48 h, increased levels of FN1, PAI1, pERK1/2 and pAKT were determined by immunofluorescence microscopy (Supplementary Fig. S14D, E; FN1, pERK1/2) and immunoblotting (Fig. 5E, all markers, Supplementary Fig. S14F; FN1, PAI1, pERK1/2) in recipient MDA-MB-231 and ZR-75-1 cells treated with MDA EVs. Additionally, MDA EVs significantly enhanced TGFBI and MMP9 mRNA expression in MDA-MB-231 (Fig. 5F). Depletion of the CD81-positive EV population reduced significantly the TGF-β signaling response of these two genes, demonstrating via this independent assay that CD81-specific EV subpopulations contributed to these gene responses (Supplementary Fig. S14G). Similarly, in MCF7, BNF3 and BNF2 cells, MDA EVs induced MMP2, MMP9, MMP10 and MMP14 expression (Supplementary Fig. S14H-J). As a word of caution, we emphasize that the original TGF-β stimulus provided to cancer cells in all experiments, which led to increased EV secretion, cannot be directly extrapolated to phenocopy the target cell signaling initiated by latent TGF-β carried as cargo by these EVs. As exemplified by out signaling analyses, comparison of pSMAD2 and pSMAD3 versus pERK1/2 and pAKT levels measured in recipient MDA-MB-231 and ZR-75-1 cells, indicated that pERK1/2 and pAKT were induced by MDA EVs to a much stronger degree than by TGF-β (Fig. 5D, E, Supplementary Fig. S14E, F). This suggests that other signals carried by MDA EVs promoted the activation of these pathways in parallel to the EV cargo of latent TGF-β.

The above data uncover that MDA EVs do activate TGF-β signaling and this signal does not require EV internalization by the recipient cells. Instead, EV interactions with cell surface TGF-β receptors might be sufficient to drive signaling, being accompanied by parallel pathways not described here in depth that are simultaneously activated upon interaction of EVs with the recipient cells.

EV MMPs and surface proteins increase TGF-β signaling and invasive phenotype

We investigated further the requirement for EV cargo TGF-β, MMP and other proteins to mechanistically induce TGF-β signaling responses in recipient cells. As pointed out by our MS analysis (Fig. 4), EV^{+ TGF−β1} carried higher levels of MMP2, MMP9 and integrins, which are known molecules promoting the release of the active TGF-β dimer from its latent complex [39, 40]. This could explain partially the significantly higher levels of active TGF-β1 monitored in EV^{+ TGF−β1} (Fig. 4I). In order to demonstrate the endogenous capacity of EVs and associated proteins to activate the latent form of TGF-β, a constant amount (2.5 ng) of recombinant latent-TGF-β1 was either activated by acidification with 1 N HCl for 10 min or was treated with similar numbers of MDA EVs (1 × 10⁹ nanoparticles) for 30 min at room temperature. Notably, both VSF^Ctrl and VSF^{+ TGF−β1} induced a significant release of TGF-β1 from their respective recombinant latent complex pools, and EV^{+ TGF−β1} exerted such activation in a more efficient manner (2.23-fold increase in activation) than EV^Ctrl (1.47-fold increase in activation), and as efficiently as acidification (2.4-fold increase in activation) (Fig. 5G), presumably due to cargo or associated proteins enriched in EV^{+ TGF−β1}.

To assess the functionality of MMPs associated with EVs, we measured the gelatin/collagen proteolytic activity of EVs before or after depletion of the CD81-positive EV populations, serving as specificity control. As shown in Fig. 5H, similar numbers of MDA EVs (1 × 10⁹ nanoparticles) enhanced matrix lysis in comparison to EV-free media (used as reference); depletion of the CD81-positive EV population decreased significantly the proteolytic activity of these EVs. In agreement with all previous observations, we also found significantly higher lytic activity associated with EV^{+ TGF−β1} compared to EV^Ctrl, since these EVs carry higher levels of MMP2 and MMP9 (Fig. 5H). Furthermore, dose-dependent experiments more convincingly revealed the higher potency of the EV^{+ TGF−β1} to promote such proteolytic activity (Supplementary Fig. S15A).

We employed an MMP inhibitor (MMPi; GM600), targeting MMP1 (Ki: 0.4 nM), MMP2 (Ki: 0.5 nM), MMP3 (Ki: 27 nM), MMP8 (Ki: 0.1 nM) and MMP9 (Ki: 0.2 nM), used at 25 µM, a dose that was not toxic to cells (Supplementary Fig. S15B). Pre-treatment of EVs with MMPi for 30 min at 37 °C efficiently blocked the gelatin/collagen proteolytic activity promoted by MDA EVs (Fig. 5H). Finally, treatment of the EVs with MMPi or proteinase K, the latter stripping the vesicular surface from attached or embedded membrane proteins, significantly inhibited the MDA EV-induced CAGA₁₂-luciferase activity in (homotypic) MDA-MB-231 and (heterotypic) BNF3 cells (Fig. 5I, Supplementary Fig. S15C). Since MMPi was equally effective as proteinase K in blocking EV-mediated TGF-β signaling, MMP9 (or another MMP) must play an important role in facilitating TGF-β signaling by tumor cell-derived EVs, most logically, by activating the predominant latent TGF-β form (Fig. 4I). Of note, similar concentration of MMPi was more efficient blocking VSF^Ctrl activity than VSF^{+ TGF−β1}, which also corroborates the higher levels of MMPs on EVs stimulated with TGF-β1 (Fig. 5I, Supplementary Fig. S15C). As control for these experiments, bioactive, recombinant TGF-β1 was pre-treated with MMPi or proteinase K, and as expected, MMPi had minor (MDA-MB-231) or no (BNF3) effect (Fig. 5I, Supplementary Fig. S15C). Furthermore, proteinase K significantly blocked the signaling by recombinant TGF-β1 in both MDA-MB-231 and BNF3 cells, which attests to the efficiency of protein digestion under the employed conditions (Fig. 5I, Supplementary Fig. S15C). Hence, our data indicate that MMP activity carried by EVs (such as MMP9 and other MMPs) activates TGF-β ligand from its latent form carried by the EVs, thus inducing signaling in recipient cells.

Since EVs can activate TGF-β signaling in recipient cells, it was natural to assume that such intercellular crosstalk mediated by tumor-derived EVs could provide invasive potential. Thus, MDA-MB-231 and ZR-75-1 cells pre-stimulated for 48 h with TGF-β1 or with similar numbers of MDA EV^Ctrl and EV^{+ TGF−β1}, demonstrated higher motility compared to untreated controls, without significant differences between EV^Ctrl and EV^{+ TGF−β1} (Supplementary Fig. S15D, E). In addition, MDA-MB-231 and BNF3 cells seeded across Matrigel, acquired a more invasive phenotype after the cells had been pre-treated with either MDA EV^Ctrl or EV^{+ TGF−β1} for 48 h (Fig. 5J, Supplementary Fig. S15F, G). Recipient cell pre-incubation with the MMPi for 30 min before EV delivery to them, showed that MMPi alone did not affect the motility of MDA-MB-231 and BNF3 cells, but abrogated the invasion promoted by MDA EVs (Fig. 5J, Supplementary Fig. S15G), indicating that MMPi did not act on cellular MMPs that support invasion. Depletion of the CD81-positive MDA EV population reduced the invasive response of the recipient cells significantly, demonstrating again a role for CD81-specific EV subpopulations to this biological response (Supplementary Fig. S15F).

Related to the migration and invasion experiments, the availability of breast fibroblasts allowed us to test the established actions of TGF-β/MMPs on fibrogenic responses. BNF2 and BNF3 fibroblasts treated with either MDA EV^Ctrl or EV^{+ TGF−β1} for 48 h, acquired an elongated morphology and showed increased FN1 protein levels intracellularly and in their extracellular matrix, an established fibrogenic response (Supplementary Fig. S16). Immunostaining for α-Smooth Muscle Actin (α-SMA), the hallmark protein of activated and contractile fibroblasts, revealed significant α-SMA expression by the fibroblasts, without further increase of α-SMA after MDA EV treatment (Supplementary Fig. S16). These fibroblast experiments confirm that the MDA EVs can transfer TGF-β and MMPs, explaining the spindle-like morphological transformation of these cells. Furthermore, invasiveness in vivo was assessed. CM-Dil Dye-labeled MDA-MB-231 cells treated for 48 h with either MDA EV^Ctrl or EV^{+ TGF−β1} were injected in the duct of Cuvier of transgenic Tg(Fli1:EGFP) zebrafish embryos, an established in vivo system for EV functionality [41]. After 24 h, there was a significant increase in the number of labeled cells that extravasated from the blood vessels and invaded the collagenous matrix of the tail in comparison to the non-treated cells (Fig. 5K). Hence, our data strengthen the relevance of EV-associated MMPs activating TGF-β signaling and enhancing invasive phenotype in the recipient cells.

EVs transfer resistance to cytotoxic drugs

Resistance to anticancer drugs is facilitated by complex mechanisms, including TGF-β, EMT and tumor-derived EVs [6, 8]. To determine the effect of MDA EVs on cell viability and drug-resistance, equal numbers of MDA EVs were added to recipient MDA-MB-231 cultures (homotypic transfer) in the absence or presence of Dox or taxol co-treatment. In the absence of cytotoxic drugs, TGF-β1 stimulation suppressed the proliferation of MDA-MB-231 cells, while treatment with either MDA EV^Ctrl or EV^{+ TGF−β1} for 48 h significantly increased the percentage of Ki67-positive cells (Supplementary Fig. S17A). The proliferation response to EV treatment reflected the respective numbers of viable cells (Supplementary Fig. S17B), which was further validated in MCF10A, BNF3, MCF7 and ZR-75-1 cells (Supplementary Fig. S17C-F, respectively). Furthermore, EV-treated MDA-MB-231 and ZR-75-1 cells became more resistant to the cytotoxic drugs (Supplementary Fig. S17B). To confirm that the resistance to cytotoxicity was EV-mediated, MDA-MB-231 cells were treated with MDA EVs depleted of their CD81⁺ vesicular fraction. Drug response assays confirmed that CD81-EV depletion abrogated the proliferation and resistance-transfer phenotypes compared to non-depleted MDA EVs (Supplementary Fig. S17B). Notably, EVs collected from the non-tumorigenic MCF10A cells did not induce proliferation or drug resistance in MDA-MB-231 or ZR-75-1 cells (Supplementary Fig. S17G, H).

To validate the latter data in vivo, we treated MDA-MB-231 cells with or without EVs and injected the cells in zebrafish embryos that lived in water containing vehicle or 5 µM Dox (Fig. 5L). The number of extravasated cell clusters 48 h after injection in the presence of Dox was significantly lower relative to untreated cells (compare the quantification -scale difference- in Fig. 5K and L). Under the latter condition, a significant increase in the number of extravasated cell clusters from MDA-MB-231 cells treated with EV^Ctrl or EV^{+ TGF−β1} compared to the non-EV treated cells was observed (Fig. 5L), indicating that MDA EVs increased the proportion of drug-resistant MDA-MB-231 cells circulating in the vasculature of the fish.

In order to connect the transfer of chemoresistance to TGF-β signaling, the recipient cells were pre-incubated with the same as previously used TβRi. In contrast to recombinant TGF-β1 that reduced recipient cell viability, TβRi preserved the cell viability (Supplementary Fig. S17I). In the presence of taxol or Dox, TβRi significantly reduced the enhanced resistance (measured as viability) caused by either recombinant TGF-β1, VSF^Ctrl or VSF^{+ TGF−β1} (Supplementary Fig. S17J, K). Hence, MDA EVs enhance cell viability in tumor and non-tumorigenic cells and consequently increase the chemoresistance of cancer cells, at least in part due to the activation of TGF-β signaling. Taken together, our data show that an equal number of EV^Ctrl or EV^{+ TGF−β1} induced similar responses to cell motility and drug resistance in recipient cells, emphasizing that the primary effect of TGF-β signaling is the enhanced production of EVs.

Discussion

We investigated EV biology by studying normal human breast epithelial cells and fibroblasts in addition to breast and lung cancer cells. The signaling mechanism initiated by TGF-β receptors, is followed by MEK/ERK1/2 kinase activation and SREBP2 phosphorylation, which regulates DHCR7 expression and cholesterol biosynthesis, thus providing inputs to one of the major routes of EV biogenesis (Fig. 3I). Furthermore, the analysis of the protein-based cargo of EVs led to the demonstration that TGF-β (primarily in its latent form) and MMP9 carried by EVs, the latter specially enriched in EVs generated after TGF-β stimulation, have biological activity. TGF-β carried as EV cargo signals on recipient cells through plasma membrane receptors and does not require EV uptake to elicit function. MMP9 (or other MMPs) in EVs seem to be involved in activation of latent TGF-β carried by the EVs and also in EV-mediated cell invasiveness (Fig. 5M).

Our finding that MEK/ERK1/2 primarily mediates TGF-β-induced EV secretion, agrees with a report, not related to TGF-β, that implicated MEK/ERK1/2 signaling in EV production by MVBs in colorectal and renal cancer cells [13]. One of the factors that control EV biogenesis is lipid/cholesterol synthesis, since the membranes of EVs are enriched in cholesterol, and microvesicles are specifically secreted from lipid rafts [28]. Accordingly, simvastatin, which inhibits HMGCR, limits the cholesterol pool, and inhibits EV secretion [42], as we also demonstrated in this study. Conversely, cholesterol can also modulate the rate of secretion and the cargo of EVs in BRCA cells [43]. We demonstrate that DHCR7 expression [36] is induced by TGF-β via activation of MEK/ERK1/2 leading to phosphorylation of SREBP2. Our observations are in line with other findings of SREBP2 activation by ERK1/2 [32], and of induction of DHCR7 expression by SREBP2 [31]. Interestingly, the late response of SREBP2 phosphorylation and DHCR7 expression to TGF-β stimulation, beyond being novel, it is compatible with the slow signaling process observed regarding the secretion of EVs induced by TGF-β and the roles of cholesterol in EV biogenesis. Of note, TGF-β not only regulates DHCR7 and SQLE expression, as in our models, but can also induce HMGCR expression in keratinocytes, leading to increased cholesterol synthesis [30]. Thus, the current study expands a critical aspect of EV biology related to the contribution of cholesterol synthesis. Additionally, we demonstrate that combinations of FDA-approved inhibitors of cholesterol biosynthesis with established cytotoxic drugs [44], sensitize BRCA and LUAD cells to undergo cell death. The interesting possibility that the effect of statins involves suppression of EV biogenesis, deserves to be examined.

The unbiased proteomic analysis of CD81- and GM1-enriched EVs showed that TGF-β stimulation modulated the cargo of small EVs (< 200 nm), for instance enriching EVs with MMP9 among a few hundred other proteins, but without altering dramatically their morphology or functionality. Furthermore, TGF-β led to the regulation of several EV-associated proteins. A study where the MDA-MB-231 EV proteome was compared to the EV proteome of non-malignant MCF10A cells (outside the context of growth factor signaling like TGF-β), concluded that specific protein cargo reflected the secretome of each cell type [45]. A similar conclusion was reached from BRCA EV proteomic studies that confirmed that the EV proteome can indicate the cell of origin and thus discriminate between epithelial and mesenchymal BRCA cells [46]. Proteomic analysis of EVs secreted from colorectal cancer cells that carry mutant, kinase deficient TGF-β type II receptors, revealed an EV cargo enriched in extracellular matrix and nucleosomal proteins [47], but it is not known whether such enrichment was due to altered TGF-β signaling or whether it reflected cell adaptation to the mutation. Our model proposes that TGF-β primarily influences the rate of EV secretion by elevating cholesterol levels, and regulates the sorting of a select number (a few hundred) of EV cargo content, including modulation of the levels of pre-existing EV cargo proteins. We studied deeper two cargo proteins present in both EV^Ctrl and EV^{+ TGF−β1} since they are functionally interlinked. TGF-β, whose levels were not further modulated in EV^{+ TGF−β1} relative to EV^Ctrl, and MMP9 whose levels were selectively increased in EV^{+ TGF−β1}. Signaling and invasiveness studies also served as functional correlates to the presence of latent TGF-β and MMP9 on EVs. Such signaling analyses should be interpreted with caution as several of the signaling proteins studied (e.g. pERK1/2, pAKT) are not unique to TGF-β signaling and can provide signaling readouts of several alternative EV cargo proteins. Since SMAD C-terminal phosphorylation is a unique signaling readout of TGF-β family ligands throughout animal evolution, we safely conclude that EVs activate downstream signaling that represents partially cargo TGF-β (or another member of its family) and partially cooperating cargo molecules.

TGF-β in association with its proteoglycan co-receptor β-glycan/TGFBR3, both carried by EVs, induces the transition of fibroblasts to myofibroblasts [5]. Similarly, mast cell-secreted EVs carry both latent and mature TGF-β on their surface, and upon interaction with recipient mesenchymal stem cells, the TGF-β associated with EVs can signal from endosomes after EV internalization [48]. We present a complementary mechanism, since heparin, an established inhibitor of EV uptake [38] did not block signaling in the cancer cell studied. Thus, TGF-β signaling mediated by EVs can also be exerted by ligand tethered to the EV surface. EVs also carry the α_vβ₆ integrin that associates with and activates latent TGF-β [49], or the tetraspanins CD9 and CD81, which can form complexes with the signaling TGF-β receptors [50], possibly providing a supra-molecular machinery for TGF-β activation. Overexpression of MMP9 has been reported to proteolytically cleave latent TGF-β and activate this signaling pathway in BRCA cells, propagating malignant cell invasiveness [40, 51]. We propose that TGF-β signaling can lead to specific enrichment of MMP9 on EVs, which can activate latent TGF-β in the EV cargo, a mechanism compatible with the well-established auto-induction principle of TGF-β [7]. It will also be interesting to analyze the co-segregation of different TGF-β family ligands with MMP family members in EV populations generated by BRCA or LUAD tumors. However, it is worth presenting a critical view on the above model and more general on the validity of EV cargo proteins [2]. In the absence of detailed topological studies such as cryo-electron tomography, the EV cargo, and especially secreted proteins that are identified tethered to the EV membrane (e.g., latent TGF-β1, MMPs, etc.) may represent genuine proteins properly packaged during EV biogenesis; alternatively, such proteins can be coating the EV surface due to a protein corona effect or they may even represent fortuitously associated proteins due to the method of EV preparation. The previous rigorous demonstration of latent TGF-β1 on the surface of exosomes [48], generates optimism that the molecular mechanism described here may well take place on the EV surface under physiological conditions. Lastly, our evidence on transfer of chemoresistance by EVs, is consistent with previous reports whereby resistance to doxorubicin or paclitaxel was transferred via TNBC EVs to non-malignant MCF10A cells [52]. Furthermore, TGF-β as EV cargo has been linked to transfer of resistance to anti-cancer drugs in head-and-neck squamous cell carcinoma and BRCA [15, 53], a phenotype that here was partially inhibitable by a TGF-β type I receptor inhibitor.

Conclusions

In summary, this study provides a fresh view on signaling mechanisms that regulate the secretion of EVs and highlights the role of cholesterol biosynthesis, TGF-β signaling and matrix metalloproteases as mechanisms of regulation of latent TGF-β cargo carried by EVs, leading to diverse pro-tumorigenic biological effects on recipient cancer cells.

Data and materials availability

The proteomics data are deposited in to PRIDE (https://www.ebi.ac.uk/pride/) under Accession Number PXD039591. For reviewer use, the details are: https://www.ebi.ac.uk/pride/; Username: reviewer_pxd039591@ebi.ac.uk Password: gBpMY8J0. The additional experimental resources used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations

BP:: GO-biological process
BRCA:: Breast cancer
CC:: GO-cellular component
CTB:: Cholerae toxin chain-B
Dox:: Doxorubicin
EMT:: Epithelial-mesenchymal transition
ER:: Estrogen receptor-α
ESCRT:: Endosomal sorting complexes required for transport
EV:: Extracellular vesicle
FDR:: False discovery rate
GO:: Gene ontology
GSEA:: Gene set enrichment analysis
HER2:: Human epidermal growth factor receptor 2
HMGCR:: HMG-CoA reductase
LC-ESI-MS/MS:: Liquid Chromatography Electrospray Ionization Tandem Mass Spectrometry
L-TGF-β1:: Latent TGF-β1
LUAD:: Lung adenocarcinoma
MF:: GO-molecular function
MMP:: Matrix metalloprotease
MVB:: Multi-vesicular body
NTA:: Nanoparticle tracking analysis
PCA:: Principal component analysis
PgR:: Progesterone receptor
REAC:: Reactome
SEM:: Scanning electron microscopy
α-SMA:: α-Smooth muscle actin
SREBP2:: Sterol regulatory element-binding protein-2
TEM:: Transmission electron microscopy
TGF-β:: Transforming growth factorβ
TGFβRI:: TGF-βreceptor type I
TGFβRII:: TGF-βreceptor type II
TGFBI:: TGF-β-induced
TNBC:: Triple-negative breast cancer
TRF:: Transferrin
VSF:: Vesicular secretome fraction
Xi:: Inhibitor of molecule X

References

Raposo G, Stoorvogel W. Extracellular vesicles: exosomes, microvesicles, and friends. J Cell Biol. 2013;200(4):373–83.
Article CAS PubMed PubMed Central Google Scholar
Lucotti S, Kenific CM, Zhang H, Lyden D. Extracellular vesicles and particles impact the systemic landscape of cancer. EMBO J. 2022;41(18):e109288.
Article CAS PubMed PubMed Central Google Scholar
Witwer KW, Van Balkom BWM, Bruno S, Choo A, Dominici M, Gimona M, et al. Defining mesenchymal stromal cell (MSC)-derived small extracellular vesicles for therapeutic applications. J Extracell Vesicles. 2019;8(1):1609206.
Article CAS PubMed PubMed Central Google Scholar
Hoshino A, Kim HS, Bojmar L, Gyan KE, Cioffi M, Hernandez J, et al. Extracellular vesicle and particle biomarkers define multiple human cancers. Cell. 2020;182(4):1044–e6118.
Article CAS PubMed PubMed Central Google Scholar
Webber J, Steadman R, Mason MD, Tabi Z, Clayton A. Cancer exosomes trigger fibroblast to myofibroblast differentiation. Cancer Res. 2010;70(23):9621–30.
Article CAS PubMed Google Scholar
Rodrigues-Junior DM, Tsirigoti C, Lim SK, Heldin CH, Moustakas A. Extracellular vesicles and transforming growth factor b signaling in Cancer. Front Cell Dev Biol. 2022;10:849938.
Article PubMed PubMed Central Google Scholar
Tzavlaki K, Moustakas A, TGF-b Signaling. Biomolecules. 2020;10(3):487.
Article CAS PubMed PubMed Central Google Scholar
Massagué J, Sheppard D. TGF-b signaling in health and disease. Cell. 2023;186(19):4007–37.
Article PubMed PubMed Central Google Scholar
Chanda D, Otoupalova E, Hough KP, Locy ML, Bernard K, Deshane JS, et al. Fibronectin on the Surface of Extracellular vesicles mediates Fibroblast Invasion. Am J Respir Cell Mol Biol. 2019;60(3):279–88.
Article CAS PubMed PubMed Central Google Scholar
de la Cuesta F, Passalacqua I, Rodor J, Bhushan R, Denby L, Baker AH. Extracellular vesicle cross-talk between pulmonary artery smooth muscle cells and endothelium during excessive TGF-b signalling: implications for PAH vascular remodelling. Cell Commun Signal. 2019;17(1):143.
Article PubMed PubMed Central Google Scholar
Théry C, Witwer KW, Aikawa E, Alcaraz MJ, Anderson JD, Andriantsitohaina R, et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J Extracell Vesicles. 2018;7(1):1535750.
Article PubMed PubMed Central Google Scholar
van Niel G, D’Angelo G, Raposo G. Shedding light on the cell biology of extracellular vesicles. Nat Rev Mol Cell Biol. 2018;19(4):213–28.
Article PubMed Google Scholar
Hikita T, Uehara R, Itoh RE, Mitani F, Miyata M, Yoshida T, et al. MEK/ERK-mediated oncogenic signals promote secretion of extracellular vesicles by controlling lysosome function. Cancer Sci. 2022;113(4):1264–76.
Article CAS PubMed PubMed Central Google Scholar
Gioni V, Karampinas T, Voutsinas G, Roussidis AE, Papadopoulos S, Karamanos NK, et al. Imatinib mesylate inhibits proliferation and exerts an antifibrotic effect in human breast stroma fibroblasts. Mol Cancer Res. 2008;6(5):706–14.
Article CAS PubMed Google Scholar
Rodrigues-Junior DM, Tan SS, Lim SK, Leong HS, Melendez ME, Ramos CRN, et al. Circulating extracellular vesicle-associated TGFb3 modulates response to cytotoxic therapy in head and neck squamous cell carcinoma. Carcinogenesis. 2019;40(12):1452–61.
CAS PubMed Google Scholar
Szvicsek Z, Oszvald A, Szabo L, Sandor GO, Kelemen A, Soos AA, et al. Extracellular vesicle release from intestinal organoids is modulated by apc mutation and other colorectal cancer progression factors. Cell Mol Life Sci. 2019;76(12):2463–76.
Article CAS PubMed PubMed Central Google Scholar
Thery C, Amigorena S, Raposo G, Clayton A. Isolation and characterization of exosomes from cell culture supernatants and biological fluids. Curr Protoc Cell Biol. 2006;(3):3 22.
Tang Z, Kang B, Li C, Chen T, Zhang Z. GEPIA2: an enhanced web server for large-scale expression profiling and interactive analysis. Nucl Acids Res. 2019;47(W1):W556–60.
Article CAS PubMed PubMed Central Google Scholar
Lanczky A, Gyorffy B. Web-based Survival Analysis Tool tailored for Medical Research (KMplot): development and implementation. J Med Internet Res. 2021;23(7):e27633.
Article PubMed PubMed Central Google Scholar
Fekete JT, Györffy B. ROCplot.org: validating predictive biomarkers of chemotherapy/hormonal therapy/anti-HER2 therapy using transcriptomic data of 3,104 breast cancer patients. Int J Cancer. 2019;145(11):3140–51.
Article CAS PubMed Google Scholar
Caja L, Tzavlaki K, Dadras MS, Tan E-J, Hatem G, Maturi NP, et al. Snail regulates BMP and TGFb pathways to control the differentiation status of glioma-initiating cells. Oncogene. 2018;37(19):2515–31.
Article CAS PubMed PubMed Central Google Scholar
Moggridge S, Sorensen PH, Morin GB, Hughes CS. Extending the compatibility of the SP3 paramagnetic bead Processing Approach for Proteomics. J Proteome Res. 2018;17(4):1730–40.
Article CAS PubMed Google Scholar
Psatha K, Kollipara L, Voutyraki C, Divanach P, Sickmann A, Rassidakis GZ, et al. Deciphering lymphoma pathogenesis via state-of-the-art mass spectrometry-based quantitative proteomics. J Chromatogr B Analyt Technol Biomed Life Sci. 2017;1047:2–14.
Article CAS PubMed Google Scholar
Papoutsoglou P, Rodrigues-Junior DM, Morén A, Bergman A, Pontén F, Coulouarn C, et al. The noncoding MIR100HG RNA enhances the autocrine function of transforming growth factor b signaling. Oncogene. 2021;40(21):3748–65.
Article CAS PubMed PubMed Central Google Scholar
Tsirigoti C, Ali MM, Maturi V, Heldin C-H, Moustakas A. Loss of SNAI1 induces cellular plasticity in invasive triple-negative breast cancer cells. Cell Death Dis. 2022;13(9):832.
Article CAS PubMed PubMed Central Google Scholar
Harbeck N, Gnant M. Breast cancer. Lancet. 2017;389(10074):1134–50.
Article PubMed Google Scholar
Sundqvist A, Vasilaki E, Voytyuk O, Bai Y, Morikawa M, Moustakas A, et al. TGFb and EGF signaling orchestrates the AP-1- and p63 transcriptional regulation of breast cancer invasiveness. Oncogene. 2020;39(22):4436–49.
Article CAS PubMed PubMed Central Google Scholar
Pfrieger FW, Vitale N. Cholesterol and the journey of extracellular vesicles. J Lipid Res. 2018;59(12):2255–61.
Article CAS PubMed PubMed Central Google Scholar
Ramji DP, Singh NN, Foka P, Irvine SA, Arnaoutakis K. Transforming growth factor-b-regulated expression of genes in macrophages implicated in the control of cholesterol homoeostasis. Biochem Soc Trans. 2006;34(Pt 6):1141–4.
Article CAS PubMed Google Scholar
Yamane T, Muramatsu A, Shimura M, Kobayashi-Hattori K, Oishi Y. Transforming growth factor-b1 induces cholesterol synthesis by increasing HMG-CoA reductase mRNA expression in keratinocytes. Biosci Biotechnol Biochem. 2016;80(7):1379–81.
Article CAS PubMed Google Scholar
Horton JD, Goldstein JL, Brown MS. SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest. 2002;109(9):1125–31.
Article CAS PubMed PubMed Central Google Scholar
Kotzka J, Lehr S, Roth G, Avci H, Knebel B, Muller-Wieland D. Insulin-activated Erk-mitogen-activated protein kinases phosphorylate sterol regulatory element-binding Protein-2 at serine residues 432 and 455 in vivo. J Biol Chem. 2004;279(21):22404–11.
Article CAS PubMed Google Scholar
Datta PK, Blake MC, Moses HL. Regulation of plasminogen activator inhibitor-1 expression by transforming growth factor-b-induced physical and functional interactions between smads and Sp1. J Biol Chem. 2000;275(51):40014–9.
Article CAS PubMed Google Scholar
Hua X, Liu X, Ansari DO, Lodish HF. Synergistic cooperation of TFE3 and smad proteins in TGF-b-induced transcription of the plasminogen activator inhibitor-1 gene. Genes Dev. 1998;12(19):3084–95.
Article CAS PubMed PubMed Central Google Scholar
Kawarada Y, Inoue Y, Kawasaki F, Fukuura K, Sato K, Tanaka T, et al. TGF-b induces p53/Smads complex formation in the PAI-1 promoter to activate transcription. Sci Rep. 2016;6:35483.
Article CAS PubMed PubMed Central Google Scholar
Prabhu AV, Luu W, Li D, Sharpe LJ, Brown AJ. DHCR7: a vital enzyme switch between cholesterol and vitamin D production. Prog Lipid Res. 2016;64:138–51.
Article CAS PubMed Google Scholar
Puzar Dominkus P, Stenovec M, Sitar S, Lasic E, Zorec R, Plemenitas A, et al. PKH26 labeling of extracellular vesicles: characterization and cellular internalization of contaminating PKH26 nanoparticles. Biochim Biophys Acta Biomembr. 2018;1860(6):1350–61.
Article CAS PubMed Google Scholar
Atai NA, Balaj L, van Veen H, Breakefield XO, Jarzyna PA, Van Noorden CJ, et al. Heparin blocks transfer of extracellular vesicles between donor and recipient cells. J Neurooncol. 2013;115(3):343–51.
Article CAS PubMed Google Scholar
Shi M, Zhu J, Wang R, Chen X, Mi L, Walz T, et al. Latent TGF-b structure and activation. Nature. 2011;474(7351):343–9.
Article CAS PubMed PubMed Central Google Scholar
Yu Q, Stamenkovic I. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-b and promotes tumor invasion and angiogenesis. Genes Dev. 2000;14(2):163–76.
Article PubMed PubMed Central Google Scholar
Androuin A, Verweij FJ, van Niel G. Zebrafish as a preclinical model for Extracellular Vesicle-based therapeutic development. Adv Drug Deliv Rev. 2021;176:113815.
Article CAS PubMed Google Scholar
Kulshreshtha A, Singh S, Ahmad M, Khanna K, Ahmad T, Agrawal A, et al. Simvastatin mediates inhibition of exosome synthesis, localization and secretion via multicomponent interventions. Sci Rep. 2019;9(1):16373.
Article PubMed PubMed Central Google Scholar
Baek AE, Krawczynska N, Das Gupta A, Dvoretskiy SV, You S, Park J et al. The cholesterol metabolite 27HC increases secretion of Extracellular vesicles which promote breast Cancer progression. Endocrinology. 2021;162(7).
Beckwitt CH, Brufsky A, Oltvai ZN, Wells A. Statin drugs to reduce breast cancer recurrence and mortality. Breast Cancer Res. 2018;20(1):144.
Article CAS PubMed PubMed Central Google Scholar
Risha Y, Minic Z, Ghobadloo SM, Berezovski MV. The proteomic analysis of breast cell line exosomes reveals disease patterns and potential biomarkers. Sci Rep. 2020;10(1):13572.
Article CAS PubMed PubMed Central Google Scholar
Wen SW, Lima LG, Lobb RJ, Norris EL, Hastie ML, Krumeich S, et al. Breast Cancer-derived exosomes reflect the cell-of-origin phenotype. Proteomics. 2019;19(8):e1800180.
Article PubMed Google Scholar
Fricke F, Michalak M, Warnken U, Hausser I, Schnolzer M, Kopitz J, et al. SILAC-Based quantification of TGFBR2-Regulated protein expression in Extracellular vesicles of microsatellite unstable colorectal cancers. Int J Mol Sci. 2019;20(17):4162.
Article PubMed PubMed Central Google Scholar
Shelke GV, Yin Y, Jang SC, Lasser C, Wennmalm S, Hoffmann HJ, et al. Endosomal signalling via exosome surface TGFb-1. J Extracell Vesicles. 2019;8(1):1650458.
Article CAS PubMed PubMed Central Google Scholar
Fedele C, Singh A, Zerlanko BJ, Iozzo RV, Languino LR. The avb6 integrin is transferred intercellularly via exosomes. J Biol Chem. 2015;290(8):4545–51.
Article CAS PubMed PubMed Central Google Scholar
Wang HX, Hemler ME. Novel impact of EWI-2, CD9, and CD81 on TGF-b signaling in melanoma. Mol Cell Oncol. 2015;2(1).
Dong H, Diao H, Zhao Y, Xu H, Pei S, Gao J, et al. Overexpression of matrix metalloproteinase-9 in breast cancer cell lines remarkably increases the cell malignancy largely via activation of transforming growth factor b/SMAD signalling. Cell Prolif. 2019;52(5):e12633.
Article PubMed PubMed Central Google Scholar
Ozawa PMM, Alkhilaiwi F, Cavalli IJ, Malheiros D, de Souza Fonseca Ribeiro EM, Cavalli LR. Extracellular vesicles from triple-negative breast cancer cells promote proliferation and drug resistance in non-tumorigenic breast cells. Breast Cancer Res Treat. 2018;172(3):713–23.
Article CAS PubMed PubMed Central Google Scholar
Tan Y, Luo X, Lv W, Hu W, Zhao C, Xiong M, et al. Tumor-derived exosomal components: the multifaceted roles and mechanisms in breast cancer metastasis. Cell Death Dis. 2021;12(6):547.
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank the Clinical Proteomics Mass Spectrometry facility at Science for Life Laboratory, Stockholm, Sweden for the proteome analysis. We thank Kenta Okamoto (Uppsala University) for assistance with SEM, Daniel Larsson and Anna Sundborger-Lunna (Uppsala University) for assistance with cryo-TEM, and Oleksandr Voytyuk (Uppsala University) for providing an optimized AcylRAC protocol and reagents. TEM was performed at BioVis – Biological Visualization, Rudbeck laboratory, Uppsala University.

Funding

Open access funding provided by Uppsala University.

This research was supported by the Swedish Cancer Society, grants CAN2018/469 and CAN2021/1506Pj01H (AM); the Swedish Research Council, grant 2018–02757 and 2023–02865 (AM) and 2020 − 01291 (CHH); the European Research Council, grant 787472 (CHH); Lars Hierta Memorial Foundation, grant FO2023-0501 (DMRJr); O.E. och Edla Johanssons Vetenskapliga Stiftelsen (DMRJr); Stiftelsen Längmanska kulturfonden, grant BA24.0451 (DMRJr); The Sigurs and Elsa Golje Memorial Foundation (LA2024-0099) (DMRJr).

Author information

Authors and Affiliations

Department of Medical Biochemistry and Microbiology, Science for Life Laboratory, Biomedical Center, Uppsala University, Box 582, Uppsala, SE-751 23, Sweden
Dorival Mendes Rodrigues-Junior, Chrysoula Tsirigoti, Carl-Henrik Heldin & Aristidis Moustakas
Astra Zeneca, Pepparedsleden 1, Mölndal, SE-431 83, Sweden
Chrysoula Tsirigoti
Laboratory of Biochemistry, School of Medicine, Faculty of Health Sciences, Aristotle University of Thessaloniki, Thessaloniki, GR-541 24, Greece
Konstantina Psatha & Michalis Aivaliotis
Laboratory of Cell Proliferation & Ageing, Institute of Biosciences and Applications, National Centre for Scientific Research ‘Demokritos’, Athens, GR-153 10, Greece
Dimitris Kletsas

Authors

Dorival Mendes Rodrigues-Junior
View author publications
You can also search for this author inPubMed Google Scholar
Chrysoula Tsirigoti
View author publications
You can also search for this author inPubMed Google Scholar
Konstantina Psatha
View author publications
You can also search for this author inPubMed Google Scholar
Dimitris Kletsas
View author publications
You can also search for this author inPubMed Google Scholar
Michalis Aivaliotis
View author publications
You can also search for this author inPubMed Google Scholar
Carl-Henrik Heldin
View author publications
You can also search for this author inPubMed Google Scholar
Aristidis Moustakas
View author publications
You can also search for this author inPubMed Google Scholar

Contributions

D.M.R.J.: Conceptualization; resources; experimental design; funding acquisition; methodology; data acquisition; validation; visualization; data analysis; data interpretation; writing-original draft; writing-revision, review and editing. C.T.: Methodology; data acquisition; visualization; writing-review and editing. K.P.: Methodology; data analysis; writing-review and editing. D.K.: Resources; data interpretation; writing-review and editing. M.A.: Methodology; data analysis; writing-review and editing. C.H.H.: Resources; data interpretation; funding acquisition; writing-revision, review and editing. A.M.: Conceptualization; resources; data interpretation; funding acquisition; supervision; project administration; writing-original draft; writing-revision, review and editing.

Corresponding author

Correspondence to Aristidis Moustakas.

Ethics declarations

Ethics approval and consent to participate

The experiments do not involve human subjects. The zebrafish experiments were performed at the core facility at Karolinska Institutet, Stockholm, Sweden, which holds ethical permits from the Stockholm Prefecture Ethics Committee.

Consent for publication

All authors have read the final version of the manuscript and agree with its publication.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article

Rodrigues-Junior, D.M., Tsirigoti, C., Psatha, K. et al. TGF-β induces cholesterol accumulation to regulate the secretion of tumor-derived extracellular vesicles. J Exp Clin Cancer Res 44, 42 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13046-025-03291-0

Download citation

Received: 25 October 2024
Accepted: 14 January 2025
Published: 06 February 2025
DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13046-025-03291-0

TGF-β induces cholesterol accumulation to regulate the secretion of tumor-derived extracellular vesicles

Abstract

Background

Method

Results

Conclusion

Graphical Abstract

Blurb

Background

Methods

Cells, reagents and treatments

EV isolation and characterization

TEM of cell monolayers

Transferrin uptake

Correlation analysis of gene expression

RNA extraction and RT-qPCR

Immunoblotting

siRNA transfection analysis

Total cholesterol quantification

Liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS)

ELISA

Gelatinase and collagenase proteolytic activity

Luciferase assays

Cell culture wound-healing assay

Matrigel invasion assay

Immuno- and direct fluorescence microscopy

Zebrafish extravasation assay

Statistical analyses

Results

TGF-β induces EV release

MEK/ERK1/2 signaling mediates EV secretion in response to TGF-β

TGF-β-induced EV release involves increased cholesterol accumulation

DHCR7 contributes to EV release in response to TGF-β

TGF-β signaling modulates a subset of the EV cargo proteome

EV internalization is not required to activate TGF-β signaling

EV MMPs and surface proteins increase TGF-β signaling and invasive phenotype

EVs transfer resistance to cytotoxic drugs

Discussion

Conclusions

Data and materials availability

Abbreviations

References

Acknowledgements

Funding

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Additional information

Publisher’s note

Electronic supplementary material

Supplementary Material 1

Supplementary Material 2

Rights and permissions

About this article

Cite this article

Share this article

Keywords

Journal of Experimental & Clinical Cancer Research

Contact us