Engineering Piezo1‐Mediated Calcium Signalling for Enhanced Extracellular Vesicle Biogenesis and Cargo Loading

ABSTRACT

Cell‐derived extracellular vesicles (EVs) are promising nanocarriers for therapeutic delivery platforms owing to their biocompatibility and capacity to protect and efficiently transport bioactive molecules. However, EV‐based therapeutics remain constrained by inefficient cargo loading and low production yields, which limit scalable biomanufacturing. To overcome these limitations, we exploited the use of the mechanosensitive ion channel Piezo1 as a robust regulator of EV biogenesis using HEK293FT cells co‐transfected with Piezo1 and the bioluminescent EV reporter PalmReNL. Activation of Piezo1 with Yoda1 (30 µM) increased PalmReNL‐EV release by 3‐fold, while GsMTx4 inhibited EV release by 80.7%. This effect was unaffected by removal of extracellular Ca²⁺ but was suppressed by intracellular Ca²⁺ chelation with BAPTA‐AM, indicating a reliance on intracellular Ca²⁺ mobilisation. Small EVs (sEVs) from Piezo1‐activated cells were purified by anion exchange chromatography and analysed by proteomics, identifying 48 proteins exclusively in Piezo1‐induced sEVs preparations among 148 total detected, including cytoskeletal and stress‐related factors, while preserving enrichment of extracellular matrix (ECM) structural components prominent in both conditions. Yoda1 treatment increased the release of both large EVs (lEVs) and sEVs, with a particularly pronounced increase in sEV production. As a proof of concept for therapeutic cargo delivery, Yoda1 stimulation increased the incorporation of exogenously expressed interleukin‐10 (IL‐10) into sEVs by up to 4‐fold, and the bioactivity of sEV‐associated IL‐10 was validated using IL‐10‐CyCLoPs reporter cells. However, exposing Piezo1‐overexpressing cells to 30 µM Yoda1 markedly delayed cell adhesion and spreading, indicating that excessive Piezo1 activation may constrain sustained production of therapeutic sEVs. Collectively, these results identify mechanotransduction as a key regulator of sEV biogenesis and underscore the need for precise temporal control, potentially achievable through ultrasound‐based modulation, for the rational engineering of next‐generation sEV therapeutics.

1. Introduction

Extracellular vesicles (EVs) are lipid bilayer‐enclosed nano‐sized particles secreted by virtually all cell types and play essential roles in intercellular communication by transporting proteins, lipids, and nucleic acids to recipient cells (Kalluri 2016; El Andaloussi et al. 2013). Because of their intrinsic biocompatibility, low immunogenicity and ability to deliver functional biomolecules to target cells and tissues, EVs have attracted significant attention as platforms for therapeutic delivery and diagnostics (Elsharkasy et al. 2020; Kumar et al. 2024). However, large‐scale biomanufacturing and precise and efficient cargo loading remain significant challenges to their clinical translation (Xu et al. 2025).

Mechanical signalling is a fundamental regulator of cell physiology, particularly in mechanically dynamic tissues such as the cardiovascular system, lungs and tumour microenvironments (Di et al. 2023; Zhu et al. 2025). While the effects of chemical and genetic stimuli on EV biogenesis are well‐documented, the influence of mechanical stimuli and mechanotransduction pathways on EV biogenesis and cargo sorting is only beginning to be elucidated (Thompson and Papoutsakis 2023). Emerging evidence indicates that cells exposed to shear stress (Kang et al. 2022), mechanical stretch (Xie et al. 2020) or changes in extracellular matrix stiffness (Wu et al. 2023) exhibit altered EV production efficiency and cargo composition. Despite these findings, the molecular mediators that govern this mechanosensitive regulation remain largely undefined (Thompson and Papoutsakis 2023).

Piezo1 is a mechanically gated ion channel that responds to changes in membrane tension, allowing cation influx into the cells, particularly Ca^{2 +}, thereby linking external mechanical forces to intracellular signalling cascades (Fu et al. 2025; Liu et al. 2022). Originally characterised in vascular endothelial and epithelial cells (Ranade et al. 2014), Piezo1 has since been implicated in a range of pathophysiological processes, including development (Qin et al. 2021), inflammation (Liu et al. 2022), epithelial barrier dysfunction (Liu et al. 2025) and cell volume regulation (Michelucci and Catacuzzeno 2024). Because Piezo1 robustly activates calcium‐dependent signalling pathways, it has emerged as a promising regulator of EV biogenesis and cargo sorting. Supporting this notion, endogenous Piezo1 activation has been shown to enhance EV release from red blood cells (Sangha et al. 2023). However, it remains unclear whether exogenous Piezo1 overexpression can similarly boost EV production. Moreover, the effects of Piezo1 overexpression on EV composition and cargo loading have not yet been characterised.

Here, we investigate how activation of exogenously overexpressed Piezo1 influences EV release, proteomic composition and engineered cell‐mediated loading of the anti‐inflammatory cytokine interleukin‐10 (IL‐10) as a proof‐of‐concept therapeutic cargo. Using HEK293FT cells stably overexpressing Piezo1 and selectively activated with the small‐molecule agonist Yoda1, we show that Piezo1 activation preferentially enhances the secretion of protein‐rich small EVs (sEVs), shifts their proteomic profile toward stress‐adaptive pathways while preserving core EV protein components, and increases IL‐10 incorporation. Together, these findings provide mechanistic insight into Piezo1‐driven EV biogenesis and suggest that controlled Piezo1 activation may enhance both EV yield and therapeutic cargo loading, addressing key challenges in the development of EV‐based therapeutics.

2. Materials and Methods

2.1. Plasmid DNA Constructs

All plasmids were constructed using standard PCR‐based cloning protocols. Plasmid sequences were confirmed by GENEWIZ (South Plainfield, NJ) prior to use in experiments. For stable expression of the EV reporter PalmReNL (Perez et al. 2022; Kanada et al. 2022; Babel et al. 2025), we previously developed a Sleeping Beauty transposon vector, pKT2/CAGXSP/PalmReNL (Perez et al. 2022) (available from Addgene, plasmid #182970). mPiezo1‐IRES‐eGFP (Coste et al. 2010) was a gift from Ardem Patapoutian (Addgene plasmid #80925). Due to the large size of the mouse Piezo1 gene (7.6 kb), a piggyBac transposon vector system was used for stable expression. pCMV‐hyPBase and pPB‐UbC were gifts from the Wellcome Trust Sanger Institute (Yusa et al. 2011). For vector construction, pPB‐UbC was cut with BamHI for cloning. The mPiezo1 gene was PCR‐amplified and fused with a FLAG tag using the following primer sets: (forward) 5′‐TTTGGCTTTTTTGTTAGACGgccgccaccatggagccgcacgtgctg and (reverse) 5′‐ Gcccaactgctcgtctgtgaag, (forward) 5′‐cttcacagacgagcagttgggc and (reverse) 5′‐Ggaagttgatgctcttcaggttggc, (forward) 5′‐Gccaacctgaagagcatcaacttcc and (reverse) 5′‐gtgacgcggcctctttctctct, and (forward) 5′‐Agagagaaagaggccgcgtcac and (reverse) 5′‐TCGTCATCGTCTTTGTAATCctccctctcacgtgtccacttaatc. The T2A‐puromycin resistance gene was amplified from pSpCas9(BB)‐2A‐Puro (PX459) V2.0 (Ran et al. 2013) (a gift from Feng Zhang, Addgene plasmid #62988) using the following primers: (forward) 5′‐GATTACAAAGACGATGACGATAAGGGCAGTGGAGAGGGCAGAG and (reverse) 5′‐GCACAGTCGGAATTCATGTCAGGCACCGGGCTTGC. All PCR fragments were assembled using the NEBuilder HiFi DNA Assembly Cloning kit (E5520S, NEB) and transformed into NEB Stable Competent E. coli (C3040I, NEB). For IL‐10‐encoding minicircle (MC) production (Saleh et al. 2026), E. coli strain, ZYCY10P3S2T (Kay et al. 2010), was transformed with the parental plasmid and pre‐cultured overnight in LB medium containing kanamycin (50 µg/mL). The culture was then expanded in TB medium with kanamycin (50 µg/mL) overnight. MC production was induced by adding equal volumes of LB containing 0.01% L‐arabinose and 0.04 mol/L NaOH, followed by 6 h incubation at 32°C.

2.2. Cell Culture

HEK293FT cells (R700‐07, Invitrogen) and HEK293 cells (ATCC) were cultured in Dulbecco's Modified Eagle Medium (DMEM; Gibco, 11995065) supplemented with 10% (vol/vol) foetal bovine serum (FBS; Cytiva, SH30071.03) and 1% penicillin/streptomycin (Gibco, 15070063) at 37°C in a 5% CO₂ atmosphere. For spontaneous integration of transgene, HEK293 cells were stably transfected with the mPiezo1‐IRES‐eGFP vector and selected with G418 (500 µg/mL) to establish constitutive Piezo1 expression. GFP‐positive cells were sorted using a FACSAria II cell sorter (BD Bioscience) equipped with a 488 nm blue laser for GFP detection. To generate HEK293FT cells stably expressing Piezo1 without GFP, cells were co‐transfected with pCMV‐hyPBase and pPB‐UbC‐mPiezo1‐2A‐Puro, followed by puromycin selection (2 µg/mL). Individual colonies were picked and expanded for analysis of Piezo1 expression by Western blotting. To assess IL‐10 receptor dimerisation induced by IL‐10 binding, an IL‐10‐CyCLoPs NIH‐3T3 reporter cell line (Lu et al. 2026) was cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin.

2.3. Yoda1 Treatment and Calcium Modulation

To investigate the effects of Piezo1 activation on EV production, we used HEK293 cells constitutively expressing PalmReNL and Piezo1/GFP. Cells were plated in a 96‐well black plate (30,000 cells/well) and cultured for 24 h. Yoda1 (SML1558, MilliporeSigma, Burlington, MA) was dissolved in DMSO at 10 mM to prepare a stock solution. To prevent Yoda1 precipitation, the stock was diluted in complete DMEM immediately prior to cell treatment. GsMTx4 (STG‐100, Alomone Labs) was dissolved in water at 250 µM. Culture medium was removed, followed by treatment with 30 µM Yoda1 (with or without GsMTx4 at 2.5 µM) for 6 h. Supernatants were collected and centrifuged at 16,000 g for 10 min to eliminate cells and debris before analysing bioluminescence signals in the supernatant after adding furimazine (25 µM). To assess the role of extracellular Ca²⁺ in EV production, cells were treated with Yoda1 under three conditions: complete DMEM, Ca²⁺‐free buffer (HBSS, 10 mM HEPES, 0.1% BSA, 0.8 mM MgCl₂) and Ca²⁺‐containing buffer (HBSS, 10 mM HEPES, 0.1% BSA, 1.8 mM CaCl₂). To evaluate the role of intracellular Ca²⁺ in EV production, cells were treated with BAPTA‐AM (21001, AAT Bioquest, Pleasanton, CA). A 20 mM BAPTA‐AM stock solution was prepared in DMSO and diluted with complete DMEM immediately before use. Cells (30,000 cells/well) were treated with BAPTA‐AM for 10 h, and the medium was replaced with complete DMEM containing both Yoda1 and BAPTA‐AM. The cells were then cultured for an additional 6 h before collecting supernatants for bioluminescence assays. Cell viability of EV producing cells stably overexpressing Piezo1 after 6‐h Yoda1 treatment was evaluated using the trypan blue exclusion assay with a TC20 Automated Cell Counter (Bio‐Rad). To ensure accurate quantification, Piezo1‐overexpressing cells that detached during Yoda1 treatment were collected and pooled with trypsinised adherent cells prior to analysis. To evaluate cell recovery following Piezo1 activation, the collected cells were replated, and their reattachment and adhesion were monitored over the subsequent 24 h.

2.4. EV Isolation

EVs were isolated from the conditioned medium of transfected and control HEK293FT cells. To prepare EV‐depleted medium, FBS was ultracentrifuged at 100,000 × g for 18 h at 4°C and filtered to remove residual vesicles (Shelke et al. 2014). Once cells reached ∼50% confluence, the culture medium was replaced with EV‐depleted medium, and cells were incubated for an additional 6 or 48 h to isolate EVs (Perez et al. 2022). Conditioned media (50 mL, from 4 × 10⁷ cells) were sequentially processed by centrifugation at 600 × g for 5 min to remove intact cells and large debris. Subsequently, to remove large vesicles, the supernatants were filtered through 0.2 µm PES membrane filters (Nalgene, 725–2520). Finally, small EVs (sEVs) were collected by a size‐based EV isolation method using 50 nm pore‐size track‐etched membranes (Whatman, WHA110603) mounted on vacuum filtration holders (Millipore, SX0002500) (Perez et al. 2023). Vesicles retained by the membranes were washed with PBS and collected. For proteomic analysis, to further enrich sEV subpopulations with high protein content based on their surface charge properties, an additional anion exchange chromatography step was performed, as previously performed (Saleh et al. 2026; Seo et al. 2022). Pre‐isolated sEVs were diluted five‐fold in 10 mM Tris‐HCl buffer (pH 7.4) to reduce the salt concentration prior to column loading. Samples (1 mL) were loaded onto a HiTrap DEAE Sepharose Fast Flow column (Cytiva, 17505501) equilibrated with 10 mM Tris‐HCl containing 30 mM NaCl (pH 7.4). After sample application, the column was washed with equilibration buffer and eluted using a linear NaCl gradient (30–500 mM), collecting 20 fractions of 500 µL each. Eluted fractions were analysed using an ÄKTA Pure HPLC system (GE Life Sciences) for UV absorbance and conductivity. The isolated protein‐enriched sEV fractions were pooled and concentrated to 1 mL using Amicon Ultra Centrifugal Filters, 100 kDa MWCO (Millipore, UFC8100). To preserve EVs during freezing, a 10% cryoprotectant solution (250 mM trehalose, 250 mM HEPES) was used. To separately enrich large EVs (lEVs) and sEVs, conditioned medium (50 mL, from 4 × 10⁷ cells) was sequentially processed by centrifugation at 600 × g for 5 min to remove intact cells and debris, followed by a second centrifugation at 16,000 × g to collect lEVs. The resulting supernatant was filtered through 0.2 µm PES membranes, and sEVs were subsequently isolated using 50 nm porous membranes.

2.5. Nanoparticle Tracking Analysis (NTA)

EVs were analysed for size and concentration using a ZetaView instrument (Particle Metrix). Prior to measurement, particle suspensions were diluted in PBS to fall within the optimal detection range of 1 × 10⁹ to 1 × 10¹⁰ particles/mL. Data acquisition was performed in automatic mode, incorporating real‐time quality control to ensure consistent video capture and particle tracking. Particle size (peak and mean diameters, nm) and concentration (particles/mL) were determined using software algorithms based on analysis of Brownian motion and light scattering properties.

2.6. Transmission Electron Microscopy (TEM)

To visualise the morphology of isolated EVs, samples were first fixed in 1% paraformaldehyde. A formvar‐coated gold grid was conditioned in a humidified chamber for 24 h before use, then placed on a 50 µL droplet of the sEV suspension and incubated for 20 min under cover. Following incubation, grids were sequentially transferred face‐down onto droplets of water for washing. Negative staining was performed using 0.4% uranyl acetate and 1.8% methylcellulose embedding solution (Cocozza et al. 2023). The excess stain was removed with filter paper. Grids were air‐dried prior to imaging on a JEOL JEM‐1400 Flash transmission electron microscope operated at 100 kV, equipped with a Matataki Flash sCMOS camera for high‐resolution image acquisition.

2.7. Protein Quantification

Total protein content of EVs and respective cell samples was measured using the Micro BCA Protein Assay Kit (Thermo Fisher Scientific, 23235) following the manufacturer's instructions. In brief, 1 µL of the sample was diluted in 99 µL of PBS, then mixed with the working reagent and incubated at 37°C for 2 h. Absorbance was recorded at 562 nm using a microplate reader. Protein concentrations were calculated by comparison to a standard curve generated from BSA standards ranging from 0.5 to 40 µg/mL, using linear regression analysis.

2.8. Fluorescence Microscopy and Bioluminescence Measurement

Cells were cultured in black or clear 96‐well cell culture plates. Fluorescence protein expression was imaged using a BZ‐X700 fluorescence microscope (KEYENCE, Osaka, Japan) or an Olympus CKX53 microscope equipped with an EP50 camera (Olympus Corporation, Tokyo, Japan). Images were analysed using ImageJ software (NIH, Bethesda, MD). For bioluminescence assays, purified or unpurified PalmReNL‐EV samples were mixed with furimazine (25 µM; Promega, N1110), and luminescence was measured using a Spark Multimode Microplate Reader (Tecan, Männedorf, Switzerland).

2.9. Western Blot Analysis

EVs were prepared for protein analysis using 15 µL of volume‐equivalent input. Samples were lysed in 4× Laemmli sample buffer (Bio‐Rad, 1610747) with or without β‐mercaptoethanol, depending on the target. Primary antibodies used were anti‐Piezo1 (1:1,000; Proteintech, 28511), anti‐GAPDH (1:1,000; Santa Cruz, sc‐365062), anti‐IL‐10 (1:5,000; Proteintech, 60269‐1‐lg), anti‐CD9 (1:1,000; Proteintech, 60232), anti‐CD63 (1:1,000; Thermo Fisher, 10628D; clone Ts63), anti‐CD81(1:3,000; Proteintech, 66866), anti‐FLAG (1:1,000; MilliporeSigma, F1804), anti‐TSG101 (1:1,000; Proteintech, 28283‐1‐AP) and anti‐Calnexin (1:1,000; Cell Signalling, 2679T). Proteins were resolved by SDS‐PAGE using 4%–20% Mini‐PROTEAN TGX Stain‐Free Gels (Bio‐Rad, 4568096) and transferred onto PVDF membranes (Millipore, IPFL00010). For intracellular protein detection, parental HEK293FT cells were lysed in RIPA buffer containing protease inhibitors, and the amount of 44 µg of total protein (from Micro BCA) was used. All membranes were blocked in PBS supplemented with 5% non‐fat dry milk and 0.05% Tween‐20 (PBST) for 30 min at room temperature and incubated overnight at 4°C with the following primary antibodies. After three washes in PBST, membranes were incubated with HRP‐conjugated secondary antibodies, either anti‐mouse (1:10,000; Cell Signalling, 7076) or anti‐rabbit (1:10,000; Cell Signalling, 7074), for 1 h at room temperature. Signal detection was performed using ECL Select Western Blotting Detection Reagent (GE Healthcare, RPN2235) and imaged using a ChemiDoc MP Imaging System (Bio‐Rad) or LI‐COR Odyssey Fc (LICORbio).

3. Proteomics and Mass Spectrometry Analysis

3.1. Proteolytic Digestion

The fractions of sEVs purified by anion exchange chromatography (25 µg) were re‐suspended in 100 mM Tris‐HCl (pH 8.5) containing 4% (w/v) sodium deoxycholate (SDC). Reduction and alkylation were performed by adding tris(2‐carboxyethyl)phosphine (TCEP) and chloroacetamide to final concentrations of 10 and 40 mM, respectively, followed by incubation at 45°C for 5 min with shaking at 2000 rpm using an Eppendorf ThermoMixer C. Trypsin, prepared in 50 mM ammonium bicarbonate, was added at a 1:100 ratio (wt/wt) and the mixture was incubated at 37°C overnight with shaking at 1500 rpm in the ThermoMixer. The final volume of each digest was 300 µL. After digestion, SDC was removed by phase extraction. Samples were then acidified to 1% trifluoroacetic acid (TFA) and subjected to C18 solid‐phase extraction using StageTips (Rappsilber et al. 2007) to remove salts.

3.2. LC/MS/MS Analysis

A 5 µL aliquot of each sample was automatically injected using a Thermo EASYnLC 1200 system onto a Thermo Acclaim PepMap RSLC C18 trapping column (0.1 mm × 20 mm). Peptides were trapped and washed for 5 min with Buffer A. Subsequently, bound peptides were eluted over 35 min onto a Thermo Acclaim PepMap RSLC C18 analytical column (0.075 mm × 250 mm) using the following linear gradient: 8%–19% Buffer B from 0 to 15 min; 19%–40% Buffer B from 15 to 24 min. After the gradient, the column was washed with 90% Buffer B for 10 min. All separations were carried out at a constant flow rate of 300 nL/min. The column temperature was maintained at 50°C using an integrated column oven (PRSO‐V2, Sonation GmbH, Biberach, Germany). Buffer compositions: Buffer A: 99.9% Water, 0.1% Formic acid; Buffer B: 80% acetonitrile, 0.1% formic acid, 19.9% water. Eluted peptides were introduced into a Thermo Scientific Q Exactive HF‐X mass spectrometer via a FlexSpray ion source. Full MS survey scans were acquired in the Orbitrap at a resolution of 60,000 (at m/z 200). The top 10 most intense precursor ions from each survey scan were selected for fragmentation by higher‐energy collisional dissociation (HCD), and MS/MS spectra were at a resolution of 15,000.

3.3. DDA Data Analysis

The resulting MS/MS spectra were converted to peak lists using Mascot Distiller v2.8.5 (www.matrixscience.com) and searched against a custom database containing all human protein sequences available from Uniprot (downloaded from Uniprot, www.uniprot.org) for background and appended with common laboratory contaminants (downloaded from www.thegpm.org, cRAP project) using the Mascot (Perkins et al. 1999) searching algorithm v 3.0.0. The Mascot output was then analysed using Scaffold v5.3.3 (www.proteomesoftware.com) to probabilistically validate protein identifications. Protein assignments that passed the Scaffold 1% false discovery rate (FDR) confidence filter were considered true identifications. Mascot search parameters included allowing up to two missed tryptic cleavage sites, a fixed modification of carbamidomethylation on cysteine residues, and variable modification for methionine oxidation. The peptide mass tolerance was set to ±10 ppm, and the MS/MS fragment ion tolerance was set to 0.02 Da. FDR was estimated by searching against a randomized version of the database.  A minimum of one unique peptide was required for protein identification. Since the proteomic analysis was based on technical LC‐MS/MS replicates from a single enriched sEV preparation, these data were considered descriptive of detectable sEV protein cargo and were not used for statistical comparisons among independent biological replicates.

3.4. Bioinformatics and Network Analysis

The proteins identified through label‐free proteomics were subjected to hierarchical clustering and protein‐protein interaction network analysis. The hierarchical heatmap was generated using the Heatmapper web tool (http://www.heatmapper.ca) (Babicki et al. 2016). To assess functional relationships among Control and Piezo1‐induced sEV proteins, interaction networks were built using the ToppCluster tool (Kaimal et al. 2010) (FDR correction, p value < 0.05). Final network images were exported as high‐resolution vector figures for inclusion in the manuscript.

3.5. Statistical Analysis

Unless otherwise specified in the figure legends, data are presented as mean ± standard deviation (SD) from at least three biological replicates. For mass spectrometry proteomics, the descriptive data presented are from technical replicates. Statistical comparisons between groups were made using unpaired two‐tailed Student's t‐tests or one‐way ANOVA followed by Tukey's post hoc test, as appropriate. p‐values < 0.05 were considered statistically significant. GraphPad Prism 9 was used for all analyses.

4. Results

4.1. Piezo1 Activation Enhances EV Release via Intracellular Calcium Signalling

To investigate whether Piezo1 activation promotes EV secretion, we engineered HEK293 cells to stably express both Piezo1 and PalmReNL, a bioluminescent EV reporter (Perez et al. 2022; Kanada et al. 2022; Babel et al. 2025). Successful overexpression of Piezo1 was confirmed by Western blotting (Figure 1a). Fluorescence microscopy revealed that GFP signals indicate Piezo1‐expressing cells, whereas tdTomato signals indicate PalmReNL‐positive cells (Figure 1b). The engineered cells were treated with 30 µM Yoda1, a chemical agonist for Piezo1 (Botello‐Smith et al. 2019), for 6 h, followed by collection of cell culture supernatants. Direct assessment of bioluminescence signals from PalmReNL reporters in supernatants, without EV isolation, revealed a 3‐fold increase upon Piezo1 activation compared to Piezo1‐expressing cells without Yoda1 treatment, suggesting enhanced release of PalmReNL‐EVs mediated by Piezo1 activation (Figure 1c). Treatment with the peptide‐based Piezo1 inhibitor GsMTx4 reduced EV release by 80.7%, supporting a role for Piezo1‐mediated mechanosensitive signalling (Figure 1d). Prior studies have shown that Piezo1 activation induces Ca²⁺ influx from the extracellular space (Wang 2025). To determine whether this influx is required for Piezo1‐mediated EV release, we cultured these Piezo1‐overexpressing reporter cells in Ca²⁺‐free buffer during a 6 h Yoda1 treatment. PalmReNL bioluminescence in the supernatant did not differ between Ca²⁺‐containing and Ca²⁺‐free conditions, indicating that extracellular Ca²⁺ entry is not required for Piezo1‐induced EV secretion (Figure 1e). In contrast, treatment with BAPTA‐AM, a membrane‐permeable Ca²⁺ chelator, inhibited Piezo1‐induced EV release in a dose‐dependent manner, reducing it by up to 58.3% (Figure 1f). Together, these results suggest that mobilisation of intracellular Ca²⁺ stores is sufficient to drive Piezo1‐mediated EV production.

FIGURE 1.

Intracellular Ca²⁺ is required for Piezo1‐induced EV release in stably transfected HEK293 cells. (a) Piezo1 overexpression was confirmed by Western blotting. (b) Fluorescence microscopy showed co‐expression of PalmReNL (red) and Piezo1/GFP (green) in HEK293 cells. Scale bars, 50 µm. (c) Yoda1‐induced production of PalmReNL‐labelled EVs was quantified by measuring bioluminescence in culture supernatants. Reporter cells were treated with 30 µM Yoda1 for 6 h before supernatant collection. (d) Piezo1/PalmReNL‐expressing HEK293 cells were treated with 30 µM Yoda1 in the presence or absence of 2.5 µM GsMTx4. EV release was assessed by measuring PalmReNL‐associated bioluminescence in the culture supernatant after treatment. (e) Bioluminescence analysis of PalmReNL‐EVs released from Piezo1/PalmReNL co‐expressing cells under Ca²⁺‐containing or Ca²⁺‐free conditions following Piezo1 activation with 30 µM Yoda1 for 6 h. (f) Effect of intracellular Ca²⁺ chelation on Piezo1‐induced EV release. Reporter cells were pretreated with 5–20 µM BAPTA‐AM for 10 h, followed by co‐treatment with 30 µM Yoda1 and BAPTA‐AM for an additional 6 h. Error bars represent SD (n = 4).

4.2. Biophysical and Morphological Assessment of Piezo1‐Induced EVs

To further assess the impact of Piezo1 channel activation on cellular EV production, we engineered HEK293FT cells to stably express Piezo1 without fluorescent reporter expression, using the piggyBac transposon system (Yusa et al. 2011). Lentiviral vectors and the Sleeping Beauty transposon system have limited capacity for large transgenes (Li et al. 2013). In contrast, the piggyBac system enabled efficient genomic insertion of the 7.6 kb Piezo1 coding sequence together with the ubiquitin C promoter and terminator elements (Figure S1a). For comparative analysis, sEVs were isolated from non‐transfected HEK293FT cells treated with 10 µM Yoda1 for 48 h (control) and from Piezo1‐overexpressing cells treated for 6 h. Because non‐transfected cells did not show enhanced sEV production in response to 10 µM Yoda1, the control condition was extended to 48 h to ensure sufficient sEV yield for isolation using conventional protocols. The isolated sEVs were characterised using established EV‐associated markers, including CD63 and TSG101 as positive sEV markers, GAPDH as an EV‐associated protein marker, and Calnexin as a negative marker for cellular contamination (Figure S1b). In our experimental setting, control sEV preparations showed no detectable GAPDH signal, whereas Piezo1‐induced sEVs efficiently incorporated GAPDH. Notably, overexpressed Piezo1 protein was not detected in sEVs regardless of Piezo1 activation (Figure S1b).

To improve purity for proteomic analysis, isolated sEVs were further purified by anion‐exchange chromatography based on their high protein content and negatively charged surfaces, as previously described (Saleh et al. 2026; Seo et al. 2022). Piezo1‐induced sEVs eluted from the anion‐exchange column at a lower ionic strength than control sEVs, indicating weaker interactions with the positively charged DEAE resin and suggesting alterations in vesicle surface charge‐related composition. Because Piezo1 was not detected in sEVs (Figure S1b), the altered elution profile likely reflects changes in the surface properties of the enriched sEV subpopulation rather than incorporation of Piezo1 into the EV membrane. These differences may arise from the distinct biogenesis or cargo‐loading processes associated with accelerated Piezo1‐driven sEV production compared with constitutive sEV release under standard culture conditions (Figure 2a,d). Nanoparticle tracking analysis (NTA) showed comparable sEV concentrations and size distributions between the two groups; however, sEVs from Piezo1‐overexpressing cells exhibited 3‐fold higher protein yield, indicating enhanced cargo protein loading in the sEVs induced by Piezo1 activation (Figure 2b,e). Transmission electron microscopy (TEM) showed no apparent morphological differences between the two groups (Figure 2c,f). We therefore analysed the sEV proteome to identify potential differences in protein composition.

FIGURE 2.

Characterisation of sEVs derived from control and Piezo1‐overexpressing cells following Yoda1 treatment. (a, d) Anion‐exchange chromatography profiles showing the elution patterns of protein‐enriched sEV fractions. (b, e) Nanoparticle tracking analysis (NTA) of particle size distribution (nm) and concentration (EVs/mL), with corresponding protein yields (µg/mL) quantified by protein assay. (c, f) Transmission electron microscopy (TEM) images of purified sEVs (arrowheads, higher magnification 60,000×), confirming their characteristic morphology and size.

4.3. Proteomic Landscape Detects Distinct Cargo Features in Piezo1‐induced sEVs

To explore the protein cargo associated with accelerated Piezo1‐induced sEV production, we performed label‐free proteomic analysis on enriched sEV fractions isolated from control and Piezo1‐overexpressing cells. Three technical LC‐MS/MS replicates were acquired from a single enriched sEV preparation per condition. Label‐free proteomic analysis of sEVs isolated from control and Piezo1‐overexpressing cells identified differences in protein cargo profiles (Supporting Information Data 1). A total of 148 proteins were identified (Figure 3a), including 79 shared (53%) between the two groups. In addition, 21 proteins were detected only in control‐derived sEV preparations, whereas 48 proteins were exclusive to Piezo1‐induced sEV preparations (Figure 3b). The modest depth of protein identification likely reflects the combination of a relatively short LC‐MS/MS acquisition time and the stringent sEV enrichment strategy employed in this study. Therefore, the proteomic dataset is not interpreted as a complete catalogue of possible sEV proteins. This proteomic analysis was designed as a proof‐of‐concept investigation to characterise detectable differences in protein cargo between enriched sEV populations rather than to achieve exhaustive proteome coverage. Moreover, the enrichment approach was designed to reduce the broader EV heterogeneity present in the original culture supernatant, thereby facilitating a more focused comparison of sEV‐associated protein cargo. The shared (Control/Piezo1) proteome was dominated by extracellular and particle‐associated proteins, with strong enrichment for extracellular matrix (ECM) structural components, such as FN1, THBS1, THBS4, COL6A1, COL6A3, COMP, POSTN and VTN (Supporting Information Data 1). In contrast, proteins uniquely detected in control sEVs were enriched for a basement membrane/cartilage ECM module, including pathways related to basement membrane organisation, extracellular matrix assembly, proteoglycan and collagen binding, calcium ion binding. Representative proteins in this group included COL2A1, HSPG2, LAMB1, NID1, TNC and SPARC. Over‐representation analysis of proteins uniquely observed in the Piezo1‐induced preparations indicated enrichment of intracellular proteostasis and biogenesis processes. Enriched categories included ATP‐dependent protein folding chaperone, unfolded protein binding, rRNA binding/structural constituent of ribosome, and cytoskeletal compartments such as the chaperonin‐containing T‐complex, ribosomal subunits, and polymeric cytoskeletal fibres. These enrichments were primarily driven by TRiC/CCT chaperonins (CCT2–CCT8 and TCP1), along with heat‐shock and trafficking/translation‐associated proteins, including HSPA6, HSP90B1, FKBP4, NPM1, EEF2, RPL/RPS proteins, TUBB, ACTN1, CLTC (Figure 3c and Supporting Information Data 1).

FIGURE 3.

Proteomic profiling of sEVs derived from Piezo1‐overexpressing and control HEK293FT cells under Yoda1 treatment. (a) Heatmap with hierarchical clustering showing protein abundance in sEVs from untransfected control cells and Piezo1‐overexpressing cells treated with 10 µM Yoda1 for 6 h (Piezo1⁺) or 48 h (control). The colour scale indicates relative protein abundance (log₂) (blue = minimum; orange = maximum) based on normalised protein intensity. For each protein, the representative value per condition was based on the median across technical replicates. (b) Venn diagram showing the overlap of proteins identified in sEVs from untransfected control cells and Piezo1‐overexpressing cells. (c) Molecular function enrichment analysis of protein signature associated with Control and Piezo1‐derived sEVs. Blank circle nodes = sample; red hexagon: genes; blue squares = molecular function.

Because sEV yield was matched by comparing accelerated Piezo1‐driven production (6 h) with prolonged baseline production (48 h), the pathway enrichment analyses likely reflect features associated with rapid EV biogenesis rather than a strictly genotype‐controlled comparison. Therefore, these explorative findings are interpreted as pathway‐level associations linked to enhanced sEV production kinetics rather than direct causal effects attributable solely to exogenous Piezo1 overexpression.

4.4. Yoda1‐Mediated Activation of Overexpressed Piezo1 Selectively Increases sEV Production

To investigate the impact of Piezo1 activation on the biogenesis of distinct EV subtypes, we next compared large EVs (lEVs) and sEVs, as defined by the MISEV guidelines (Welsh et al. 2024), isolated from the supernatants of HEK293FT cells stably overexpressing Piezo1 and treated with 30 µM Yoda1 for 6 h. Notably, Yoda1 treatment markedly increased the levels of tetraspanin markers in sEVs derived from Piezo1‐overexpressing cells, whereas lEV levels remained largely unchanged (Figure 4a). Specifically, CD9 and CD81 levels increased by 80‐fold and 27‐fold, respectively, in the sEV fraction of Yoda1‐treated Piezo1‐overexpressing cells compared with controls. To further assess the purity of the isolated EV fractions, GAPDH and calnexin (a negative EV marker) were analysed (Figure S2). Consistent with the marked increase in sEV production, Yoda1‐mediated Piezo1 activation substantially increased GAPDH levels in sEVs, whereas only a modest increase was observed in lEVs. In contrast, calnexin was undetectable in all EV fractions, supporting the purity of the isolated EV preparations. This subtype‐specific enhancement was further validated using bioluminescence assays in which Piezo1‐overexpressing HEK293FT cells were transfected with PalmReNL. These assays revealed a pronounced 9.6‐fold increase in sEV‐associated signals from Piezo1‐overexpressing cells compared with Yoda1‐treated control cells. In contrast, Yoda1‐treated Piezo1‐overexpressing cells exhibited no significant increase in lEV‐associated signals relative to Yoda1‐treated control cells, suggesting that Piezo1 activation preferentially promotes PalmReNL‐sEV release (Figure 4b). NTA indicated increases in both EV subtypes, with a more prominent rise in lEVs from Piezo1‐overexpressing cells upon 30 µM Yoda1 treatment (Figure 4c and Figure S3). This apparent disparity may reflect the enhanced sensitivity of PalmReNL reporters for detecting protein‐rich sEV populations, consistent with our previous observations comparing PalmReNL‐based bioluminescence measurements with NTA particle counts (Perez et al. 2022). Supporting this interpretation, treatment of Piezo1‐overexpressing cells with 30 µM Yoda1 resulted in a 19‐fold increase in total protein content within the sEV fraction, whereas no appreciable alterations were observed in the lEV fraction (Figure S3). Together, these findings indicate that Yoda1‐mediated Piezo1 activation promotes overall EV release, while exerting a particularly strong effect on the production of protein‐rich sEVs in Piezo1‐overexpressing cells.

FIGURE 4.

Yoda1‐mediated activation of Piezo1 enhances sEV release and marker expression. (a) Western blot analysis of CD9 and CD81 in lEVs and sEVs isolated from control (C), Yoda1‐treated (30 µM, Y1), Piezo1‐overexpressing (P1), and Piezo1‐overexpressing plus Yoda1‐treated (P1+Y1) HEK293FT cells under non‐reducing conditions. Quantification of CD9 and CD81 protein levels relative to control is shown on the right. (b) Bioluminescence measurements of PalmReNL‐labelled lEVs (left) and sEVs (right) derived from the same experimental groups. EVs were isolated from cell supernatants after 6 h of treatment with 30 µM Yoda1. Error bars represent SD (n = 4). (c) NTA showing particle concentrations of lEVs and sEVs under the indicated conditions.

4.5. Yoda1‐Mediated Piezo1 Activation Enhances Dose‐Dependent Release of Engineered sEVs, Providing a Proof of Concept for Regulated EV‐Based Therapeutic Delivery

To assess whether Piezo1 activation influences the incorporation of exogenously expressed therapeutic proteins into sEVs, both control and Piezo1‐overexpressing cells were transfected with a minicircle DNA vector encoding IL‐10 previously developed by our group (Saleh et al. 2026). At 48 h post‐transfection, cells were treated with 10 or 30 µM Yoda1 for 6 h prior to sEV isolation. Notably, transfected IL‐10 was effectively incorporated into sEVs, with Piezo1‐overexpressing cells showing 3‐fold and 4‐fold increases following treatment with 10 and 30 µM Yoda1, respectively, compared to control cells without Piezo1 overexpression (Figure 5a). Western blot analysis showed elevated levels of tetraspanin markers (CD9, CD63 and CD81) in sEVs derived from Piezo1‐overexpressing cells, with the strongest upregulation observed following high‐dose (30 µM) Yoda1 treatment (Figure 5a ). The purity of the isolated sEV fractions was further assessed by analysing GAPDH and calnexin (Figure S4). GAPDH was readily detected in Piezo1/Yoda1‐induced sEVs, consistent with enrichment of EV‐associated proteins. In contrast, calnexin, a negative cellular contamination marker, was detected in producer‐cell lysates but was undetectable in the corresponding sEV preparations, supporting the purity of the isolated vesicles. These results indicate that Piezo1 activation not only enhances sEV release but also facilitates therapeutic cargo loading, potentially through shared regulatory mechanisms that coordinate vesicle biogenesis and protein incorporation into sEVs. Morphological examination of EV producing cells revealed pronounced changes at the higher Yoda1 concentration (30 µM), including cell rounding and partial detachment, particularly in Piezo1‐overexpressing cells (Figure 5b). These observations likely reflect rapid cytoskeletal and membrane remodelling associated with increased sEV secretion following Piezo1 activation.

FIGURE 5.

Dose‐dependent enhancement of exogenous cargo loading into sEVs via Piezo1 activation. (a) Western blot analysis and quantification of IL‐10 under reducing conditions, and of CD63, CD9 and CD81 under non‐reducing conditions, in sEVs isolated from control (C) and Piezo1‐overexpressing (P) HEK293FT cells following transfection with IL‐10‐encoding minicircle DNA. Quantification of relative protein abundance is shown below each corresponding blot. An equal volume (15 µL) of sEV preparation was loaded for Western blot analysis. (b) Representative phase‐contrast images of control and Piezo1‐overexpressing cells following 6 h treatment with 10 or 30 µM Yoda1. Scale bars, 50 µm.

To determine whether the observed morphological changes reflected cytotoxicity, we assessed the viability of Piezo1‐overexpressing cells following Yoda1 treatment. No evidence of acute cytotoxicity was observed under any treatment condition during the 6‐h treatment period (Figure 6a). We next examined the long‐term consequences of Piezo1 activation by replating cells after exposure to low and high concentrations of Yoda1 to assess the recovery of these morphological alterations. Piezo1‐overexpressing cells treated with 10 µM Yoda1 appeared normal upon replating and readily re‐established adhesion, whereas cells treated with 30 µM Yoda1 exhibited markedly delayed reattachment and remained rounded even 24 h after Yoda1 removal (Figure 6b). Together, these results suggest that sustained or excessive Piezo1 activation can lead to persistent cytoskeletal alterations that may impact long‐term EV production.

FIGURE 6.

Assessment of cell viability and reversibility of Piezo1 activation following Yoda1 treatment in control and Piezo1‐overexpressing cells. (a) Cell viability was assessed by trypan blue exclusion after 6 h of treatment with increasing concentrations of Yoda1. Both adherent and detached cells were collected and combined following trypsinisation to ensure accurate viability measurements. Error bars represent SD (n = 3). (b) To evaluate the reversibility of Yoda1‐induced morphological changes, cells from each treatment group were collected, replated after Yoda1 removal (top images), and cultured for an additional 24 h (bottom images). Scale bars, 250 µm.

4.6. Enhanced sEV Production by Piezo1 Activation Retains Functional IL‐10 Cargo Activity

Finally, the bioactivity of Piezo1‐induced, IL‐10‐loaded sEVs was evaluated using the cytokine cellular locating platforms (CyCLoPs) reporter system (Lu et al. 2026), which is specific for IL‐10 signalling. The reporter cells co‐express IL‐10RA‐tTA and IL‐10RB‐TEV fusion proteins together with a tetracycline response element (TRE)‐H2B‐EGFP reporter, enabling expression of nuclear‐localised H2B‐EGFP upon CyCLoPs‐dependent activation of IL‐10 signalling (Figure 7a). Control HEK293FT and Piezo1‐overexpressing cells were transfected with IL‐10‐encoding minicircle DNA and treated with 30 µM Yoda1 for 6 h. To ensure equivalent starting conditions, equal numbers of control and piezo1‐overexpressing cells (1 × 10⁶ cells per well in a 6‐well plate) were seeded and transfected, and Yoda1 treatment was initiated 48 h post‐transfection. Isolated IL‐10‐loaded sEVs were subsequently incubated with IL‐10‐CyCLoPs reporter cells (20,000 cells per well in 96‐well plates) for 24 h, after which IL‐10 signalling activity was evaluated by measuring nuclear EGFP expression. Notably, only IL‐10‐loaded sEVs derived from Piezo1‐overexpressing cells following Yoda1 treatment induced a marked increase in nuclear EGFP signals in the reporter cells (Figure 7b). Quantification of EGFP‐positive nuclei revealed a 9.7‐fold increase in reporter induction compared with untreated reporter cells (Figure 7c). These findings demonstrate that Yoda1‐mediated Piezo1 activation enhances the production of bioactive IL‐10‐loaded sEVs capable of eliciting functional IL‐10 signalling in recipient cells.

FIGURE 7.

Bioactivity assessment of Piezo1‐induced IL‐10‐loaded sEVs. (a) Schematic illustration of the IL‐10‐CyCLoPs reporter system. IL‐10 associated with sEVs binds to IL‐10 receptors, triggering TEV protease‐mediated cleavage and release of C‐terminally tethered tTA, which activates reporter gene expression. Created with BioRender.com. (b) Fluorescence microscopy of nuclear H2B‐EGFP expression in IL‐10‐CyCLoPs cells following 24 h incubation with IL‐10‐loaded sEVs. Corresponding phase‐contrast images show the reporter cells in the same field of view. All fluorescence images were acquired using identical exposure time, gain, and illumination settings to enable quantitative comparison. Scale bars, 100 µm. (c) Quantification of EGFP reporter induction, expressed as the fold change in EGFP‐positive nuclei relative to untreated reporter cells. Error bars represent SD (n = 3).

5. Discussion

This study identifies Piezo1 as a potent regulator of sEV biogenesis and provides both mechanistic insight and translational potential for sEV‐based therapeutic engineering. Through controlled activation of Piezo1 with the small molecule agonist Yoda1, we achieved a selective and robust enhancement of sEV release, accompanied by efficient incorporation of the exogenously overexpressed therapeutic cytokine IL‐10 within the cells. These findings suggest that Piezo1 may serve as a biologically tunable platform for scalable and targeted sEV production, representing a significant step toward engineering next‐generation sEV‐based therapeutics.

Given Piezo1's established role as a mechanosensitive, Ca²⁺ permeable ion channel that transduces mechanical cues into intracellular biochemical signals (Di et al. 2023), it is surprising that our results revealed that Piezo1‐induced sEV release occurs independently of extracellular Ca²⁺ influx but critically depends on the mobilisation of intracellular Ca²⁺ stores. Our findings suggest that Piezo1 promotes sEV biogenesis not only through direct Ca²⁺ entry across the plasma membrane but also by mobilising Ca²⁺ from intracellular reservoirs. Although the mechanism by which Yoda1‐mediated activation of Piezo1 triggers Ca²⁺ release from the ER remains unclear, our observations are at least consistent with previous reports identifying intracellular Ca²⁺ as a key regulator of vesicle budding and trafficking (Kim et al. 2022; Williams et al. 2023). Previous studies indicate that Piezo1 can also regulate Ca^{2 +}‐independent signalling pathways, including PP2A–Akt (Shindo et al. 2025), RhoA/ROCK/PIP₂ signalling and actin remodelling (Deng et al. 2025), focal adhesion dynamics (Yao et al. 2022) and membrane mechanics (Zeng et al. 2022). Consistent with this, we show that the membrane‐impermeable peptide inhibitor GsMTx4 suppresses EV release, suggesting a contribution of downstream Piezo1‐associated signalling. These pathways may influence vesicle biogenesis by modulating cytoskeletal organisation, membrane tension, endosomal trafficking, multivesicular body (MVB) positioning and MVB–plasma membrane fusion. However, these mechanisms remain speculative and will require future mechanistic validation.

This calcium‐dependent signalling cascade appears to selectively enhance the release of protein‐rich sEVs, as evidenced by the marked increase in total protein content and the sEV markers CD9, CD81 and GAPDH in the isolated sEV fraction (Figure 4a; Figures S2 and S3). In contrast, these markers exhibited little or no change in the lEV fraction. This EV subtype‐specific response suggests that Piezo1 signalling differentially regulates vesicle release according to vesicle size and cellular origin, consistent with the distinct biogenesis of lEVs, which are generated predominantly through outward budding of the plasma membrane (Colombo et al. 2014). Given that sEVs are among the best‐characterised next‐generation therapeutic delivery platforms, with unique pharmacokinetic and biodistribution profiles (Murphy et al. 2019; Kang et al. 2021; Kamerkar et al. 2017), selective stimulation of sEV biogenesis through Piezo1 activation represents a valuable strategy for scalable and efficient production of therapeutic sEVs (Wan et al. 2024; Guo et al. 2021). Our findings extend previous studies demonstrating that mechanical stimulation or modulation of the mechanical microenvironment can enhance EV production and therapeutic potential (Kang et al. 2025; Ragni et al. 2025). A key distinction of the present study is the use of exogenous Piezo1 expression as a molecularly defined regulator of sEV biogenesis, allowing transient pharmacological activation to drive the production of therapeutic cargo‐loaded sEVs.

The enrichment of tetraspanin markers may have functional significance beyond indicating increased sEV abundance. Tetraspanins regulate membrane organisation and contribute to EV biogenesis, cargo sorting, cellular uptake and biodistribution (Duke et al. 2025; Rubinstein et al. 2025; Toribio and Yanez‐Mo 2022; Perez‐Hernandez et al. 2013). Therefore, their increased abundance in Piezo1‐induced sEVs may affect vesicle stability, target‐cell interactions, and cargo delivery. While our findings demonstrate enhanced sEV production and IL‐10 incorporation following Piezo1 activation, further studies are needed to determine whether these changes translate into improved therapeutic efficacy.

To further explore how Piezo1 activation influences sEV composition, we conducted proteomic profiling of a highly purified sEV subpopulation isolated by anion exchange chromatography and enriched for vesicles with high protein content and a negatively charged surface. This analysis revealed increased abundance of proteins associated with cytoskeletal remodelling, chaperone‐mediated protein folding and cellular stress responses, including heat shock proteins (Figure 3). Although Yoda1 functions as a chemical activator of Piezo1 channels, the observed proteomic alterations are consistent with Piezo1's established role in orchestrating cellular responses to mechanical stress (Yaganoglu et al. 2023). Our findings indicate that Piezo1 activation not only enhances sEV production but also selectively enriches stress‐protective and structural proteins within these vesicles, potentially through calcium‐dependent activation thresholds, post‐translational modifications and/or engagement of vesicle‐sorting machinery (Thompson and Papoutsakis 2023; Mir and Goettsch 2020). Notably, although we focused our proteomic analysis on a protein‐rich sEV subpopulation, the substantial overlap in proteomic profiles between Piezo1‐enhanced sEVs (6 h) and sEVs collected after prolonged Yoda1 exposure (48 h) supports the idea that Piezo1‐induced sEV biogenesis is a selective, regulated process rather than a nonspecific cellular stress response. Building on this observed selectivity, we found that Piezo1 activation also enhances the loading of engineered therapeutic cargo into sEVs. We previously demonstrated, using a proteinase protection assay, that IL‐10 is localised on the EV surface as either a monomer or an oligomer (Saleh et al. 2026). However, the biological activity of this surface‐associated IL‐10 remained unclear. Following transfection with a minicircle vector encoding IL‐10, Piezo1 activation promoted a dose‐dependent increase in the delivery of bioactive IL‐10 by sEVs, as demonstrated using the IL‐10‐CyCLoPs reporter system (Figure 7). These findings further support a role for Piezo1, and likely downstream calcium signalling, in regulating sEV biogenesis and cargo incorporation. This approach offers a promising strategy to improve the loading efficiency of therapeutic payloads, a major challenge in engineered vesicle development (Kim et al. 2024). Notably, although a 6‐h stimulation accelerated sEV production, continuous exposure of Piezo1‐overexpressing cells to high‐dose Yoda1 (30 µM) significantly delayed recovery of physiological integrity (Figure 6b), highlighting the need for precise temporal control of Piezo1 activation, potentially through approaches such as ultrasound‐based mechanical stimulation (Li et al. 2024).

Several key questions remain to be addressed. First, it will be crucial to determine whether Piezo1‐mediated enhancement of sEV production extends to clinically relevant cell types, such as mesenchymal stem cells and various immune cell types (Kumar et al. 2024; Wiklander et al. 2019), under both pharmacological and mechanical stimulation. Second, a limitation of this proteomic analysis is that Piezo1‐induced and control sEVs were collected over different timeframes, reflecting the accelerated sEV production following Piezo1 activation and the limited duration of Yoda1 treatment. Consequently, the observed differences may be influenced not only by Piezo1/Yoda1‐mediated regulation but also by EV production kinetics and collection periods. Time‐matched biological replicates with deeper proteomic coverage will be required to more rigorously define Piezo1‐dependent cargo‐loading mechanisms. Furthermore, only protein cargo was analysed, whereas EV‐associated RNAs are also selectively sorted (O'Brien et al. 2020). Therefore, accelerated sEV biogenesis may alter RNA cargo composition. Future studies incorporating RNA profiling and functional analyses will be needed to determine how Piezo1 activation influences RNA loading and RNA‐mediated signalling by sEVs. Third, future studies should evaluate the therapeutic effects of Piezo1‐enhanced sEVs using in vivo disease models to advance the development of sEV‐based therapeutic delivery platforms (Lu et al. 2024; Cheng and Hill 2022).

In summary, our findings establish Piezo1 as a novel and versatile regulator of sEV biogenesis. Activation of Piezo1 by the small‐molecule agonist Yoda1 triggers intracellular calcium signalling that markedly accelerates sEV production while maintaining the characteristic extracellular and particle‐associated protein signatures of sEVs, including enrichment of ECM‐associated structural components. Moreover, Piezo1 activation enhances the incorporation of engineered protein cargo into sEVs for functional therapeutic delivery. Together, these findings position Piezo1 as a promising molecular tool for advancing sEV biomanufacturing by enabling scalable and tunable control of vesicle production and cargo loading, two key challenges limiting the clinical translation of engineered sEV‐based therapeutics.

Author Contributions

Najla A. Saleh: methodology, software, data curation, formal analysis, validation, investigation, Writing – original draft, Writing – review and editing, supervision, visualization, project administration. Rozita Shafiq: data curation, formal analysis, investigation. Amanda P. Cicerone: data curation, formal analysis, software, investigation, methodology. Sadhana Kilangodi: investigation, data curation, formal analysis, writing – review and editing. Hayden Hartwig: writing – review and editing, investigation, formal analysis, data curation. George I. Mias: writing – review and editing, validation. Guangqing Lu: methodology, resources, writing – review and editing. Jun R. Huh: methodology, resources, writing – review and editing. Weiping Li: methodology, investigation. Cheri X. Deng: conceptualization, writing – review and editing, funding acquisition, methodology, supervision, validation. Masamitsu Kanada: conceptualization, funding acquisition, investigation, writing – review and editing, software, formal analysis, methodology, supervision, data curation, resources, project administration, visualization, validation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting information: jex270182‐sup‐0001‐FigureS1‐S4.docx

Supporting information: jex270182‐sup‐0002‐SuppData1.xlsx

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.