Abstract Scalable approaches for enhancing therapeutic small extracellular vesicles (sEVs) production can facilitate the transition of sEVs from bench to bedside and beyond. Here, we present a user-friendly method to manipulate the extracellular mechanical microenvironment of umbilical cord-derived mesenchymal stem cells (MSCs), a promising cell type for generating sEVs with therapeutic benefits, to boost sEV secretion and regenerative bioactivity. The bioreactor system, called the programmable controlled rotating platform (PRP), is designed to apply normal stress on cell surface through centrifugal rotation culture. Experimental analyses suggested that the PRP can promote a 4-fold sEV secretion increase without affecting cell viability and sEV size when compared to the traditional static culture condition. More importantly, PRP-induced MSC-sEVs can significantly promote epithelial cell migration in vitro and accelerate corneal wound healing in a murine model, with suppressed inflammatory responses in wound bed tissue. Further mechanistic investigations revealed that this process involves the activation of cellular transcriptional signals implicated in sEV biogenesis. Concurrently, sEV cargo undergoes remodeling to enrich regenerative and immunoregulatory functions. These findings demonstrate the efficacy of our established platform in advancing sEV production and improving clinical performance, providing a novel sEV-based mechanism for ocular treatments, including corneal epithelialization and even retinal neural regeneration. Supplementary Information The online version contains supplementary material available at 10.1186/s12951-025-03556-y. Keywords: Mesenchymal stem cells, Extracellular vesicles, Mechanical stimulation, Corneal healing, Tissue regeneration Graphical Abstract [44]graphic file with name 12951_2025_3556_Figa_HTML.jpg Supplementary Information The online version contains supplementary material available at 10.1186/s12951-025-03556-y. Introduction Small Extracellular vesicles (sEVs) are heterogeneous nanoparticles (usually < 200 nm in diameter) that are either naturally or inducibly released by living cells [[45]1]. Interest in sEVs, especially those derived from mesenchymal stem cells (MSCs), as an emerging alternative to whole-cell therapeutics is increasing, driven by their innate advantages of safety and storage [[46]2], as well as the abilities to traverse biological barriers and transport bioactive molecules. Recent studies have emphasized the MSC-sEVs’ beneficial effects in various pre-clinical therapeutic trials (e.g., neurodegeneration, cancer, and vulnus) [[47]3–[48]5]. However, sEVs released by cells in their undisturbed state tend to be modest in quantity and generally contain ordinary inclusions, which poses a significant obstacle to their bench to bedside [[49]6]. Thus, efforts should be made to prioritize research on increasing sEV production and remodeling their functional cargo, as well as understanding the underlying mechanisms by which microenvironments influence their properties. Methods to enhance sEV production include chemical or physical triggering of parental cells. For instance, promoted cellular exocytosis and sEV release could be implemented via calcium phosphate treating and subsequently elevating cellular Ca^2+ levels [[50]7, [51]8]. Yet the introduction of exogenous inducers might modify the cell culture environment, potentially affecting both sEV purification and compositional assessment. Harnessing cells’ sensitivity to physical stimuli offers a promising alternative. Non-invasively manipulating the cellular microenvironment by applying mechanical stress like fluid shear [[52]9, [53]10], stretching [[54]11], hydrostatic pressure [[55]12], acoustic wave [[56]13], and stiff extracellular matrix [[57]14] can alter gene expression and cell function, thereby regulating sEV generation and bioactivity. These approaches avoid introducing exogenous substances into the extracellular environment, leading to a significant increase in sEV yield and enhanced functions such as wound healing and axonal growth. Despite their individual strengths, scalability challenges remain to simplify the operation processes and reduce the reliance on sophisticated equipment. In this study, we propose a simple and practical bioreactor, which we termed the programmable rotating platform (PRP), to apply controllable normal stress on the cell surface and enhance their sEV secretion (Fig. [58]1A). We chose human umbilical cord mesenchymal stem cells (MSCs) as an application paradigm, taking into account their low tumorigenicity, immunogenicity, easier acquisition, fewer ethical concerns, and established precedents in therapeutic [[59]15, [60]16]. Our modeling and experimental evaluations identified optimal parameters that maximize both cell viability and sEV yield. We validated the practicality of this approach by significantly enhancing the sEV production and demonstrating improved therapeutic bioactivity in wound healing models of cultured cells and in vivo murine corneal tissue (Fig. [61]1B and C). Additionally, we explored the underlying mechanisms by which PRP drives secretion of sEVs with regenerative bioactivity through transcriptional and protein analyses. Since PRP is programable, and easy to extend, this approach streamlines experiments for mass production of sEVs and can be adapted to diverse potential therapeutic applications. Fig. 1. [62]Fig. 1 [63]Open in a new tab Cell culture and stress analysis on the programmable rotating platform. (A) PRP-based cell culture and surface mechanical stress on MSCs. (B) Schematic plot depicting changes in cell behavior and EV functions under rotational culture. (C) Experimental diagram of corneal repair promotion using PRP-MSC-sEVs. (D) The optimal orientation of flasks for rotational cell culture. The bottom of the cell culture flask, facing the centrifugal side, was defined as “outside surface, OS”, while the bottom side facing the centripetal direction was defined as “inside surface, IS”. The medium volume used for OS centrifugation experiment was 50 mL (to ensure cells in the vertically placed flask fully contact the culture medium) and filled for IS (to reduce nutrient deprivation caused by liquid surface imbalance). Statistical significance was calculated by one-way ANOVA with post hoc Tukey’s pairwise multiple comparisons test (n = 3, **p < 0.01, ***p < 0.001) and the error bar represents the standard error of the mean (SEM). (E) Computational model of physical parameters of culture medium on PRP Results Culture and dynamic stress analysis of MSC on a programmable rotating platform The rotating culture platform is composed of a motor control station and a workbench featuring with a disk and eight concave cubic columns (Figure [64]S1). The programmable stepper motor drives the perpendicularly fixed culture flasks on the disk, executing a circular motion to enable rotational centrifugal cell culture. We first investigated the effect of surface orientation on cells. Results showed that cells on the outside surface (centrifugal side) maintained high viabilities (> 95%), yet those on the inside surface (centripetal side) experienced significant decreases in viability (< 90%) even at lower centrifugal force (Fig. [65]1D). We suggest that the rigid culture flask provides superior cell support during rotational centrifugal force compared to the flexible medium. Moreover, full medium loading is necessary under inside surface conditions to mitigate the effect of liquid surface imbalance on cell nutrient supply during centrifugation; this also raises the potential for liquid leakage. Consequently, subsequent experiments will employ the configuration with cells positioned on the outside surface, except where noted. We then conducted fluid simulation experiments to assess the forces experienced by cells under rotation (Fig. [66]1E). Apart from the support of the culture flask, the fluid exerted the primary external forces. Our results showed that after reaching a constant rotational speed (ω), the fluid and cells remained relatively stationary (i.e., shear stress from fluid flow was negligible), but cells were subjected to a constant surface pressure (i.e., normal stress) exerted by the fluid. Nonetheless, under variable speed conditions (ω’), both shear stress and pressure exerted by the fluid were minimal and short-lived. These results demonstrate that our PRP system successfully provided mechanical stimulation to cultured cells, facilitating the exploration of the impact of physical stress on intracellular and extracellular signals. Preparation of MSC-derived sEVs based on PRP Capitalizing on the cell’s ability to adapt to microenvironmental changes through the regulation of secretion, we optimized the centrifugation parameters to assess the impact of PRP-based culture on cells and sEVs. When examining the influence of centrifugal force, we found that more than 90% of the nanoparticle population falls within 40–200 nm even under different centrifugal force conditions (Fig. [67]2A). As expected, sEV yield increased with increasing centrifugal force, yet the yield decreased significantly when the constant centrifugal force exceeded 15×g (e.g., 20×g) (Fig. [68]2B). Simultaneously, excessive centrifugation force (more than 20×g) may pose potential risks such as significant thermal effect and culture flask detachment. Subsequent assays further suggested that overloads may impair cell viability relative to static culture conditions (L4 and 0×g) (Fig. [69]2C). Besides, while sEV yields were comparable under both constant force (F, corresponding to constant speed ω) and variable force (F’, corresponding to ω’) conditions, cell viability was reduced when experienced steep transitions between high and low centrifugal forces (Fig. [70]2D and Figure [71]S2). Meanwhile, equipment fatigue under variable loading conditions is another factor to consider, despite acute mechanical perturbation being thought to activate cells [[72]17]. Collectively, constant centrifugal force of 15×g was adopted for further experiments. Fig. 2. [73]Fig. 2 [74]Open in a new tab sEV production and cell viability under various culture conditions. (A) The mean size distribution of particles under different centrifugal force conditions. Culture medium within 12-hour was collected for sEV isolation. n = 12. (B) The total number of extracellular particles under different centrifugal forces. n = 4. (C) Survival rates of cells with different centrifugal forces. n = 3. (D) Particle yields in conditions of constant force (F) or variable forces (F’). n = 4. (E) The size distribution of particles under 15×g centrifugal force condition for different culture durations. n = 12. (F) Particle amount at different culture durations. n = 4. (G) Cell survival rates at different culture times. n = 3. (H) Incremental yield relative to the conventional culture (L4). n = 4. (I) Particle amount under different cell culture systems. n = 4. (J) Protein immunoblotting of sEV-positive and -negative markers under different conditions (n = 3). L4, horizontal flask with 4 mL medium and without centrifugation. L50, horizontal flask with 50 mL medium and without centrifugation. 0×g, vertical flask with 50 mL medium and without centrifugation. 15×g, vertical flask with 50 mL medium and under 15×g centrifugal force. Taking 15 − 11×g as an example, the variable force experiment is defined by switching between 15×g and 11×g conditions, with each condition maintained for 10 s. The switching is program-controlled and completed within 3 s. One-way ANOVA was used for multiple comparisons (B-C and F-J), and unpaired t-test was used for generating p values (D). Data was represented as mean ± SEM. *p or #p < 0.05, **p < 0.01, ***p or ###p < 0.001, ****p or ####p < 0.0001. For B-C, “*” indicates the statistical significance when compared to 15×g-12 h, “#” indicates the statistical significance when compared to 0×g-12 h. For F-H, “*” indicates vs 15×g-12 h, “#” indicates vs 15×g-24 h The influence of temporal conditions was also assessed in our study. Analogously, the working time having limited impact on the particle size distribution (Fig. [75]2E). We then observed that prolonging the rotation culture time is beneficial for harvesting more extracellular particles (Fig. [76]2F), yet long-term (24-hour) stimulation led to a marked decrease in cell viability (Fig. [77]2G). Furthermore, a 12-hour rotational culture demonstrated the highest enhancement efficiency in yield (4-fold) compared to the conventional static culture (Fig. [78]2H). Consequently, we subsequently collected cell-derived sEVs within 12 h only. Since rotational culture involves filling enough culture flasks with medium and vertically positioning them on PRP, we investigated the influence of both medium volume and flask orientation to ascertain that the observed increase in sEV yield was attributable to centrifugal force. Results showed that increasing medium volume (L50) or static vertical placement (0×g) could not significantly improve particle yield (Fig. [79]2I-J). Despite the potential for increased hydrostatic pressure and decreased oxygen tension under these conditions, only the group with additional centrifugal force showed significantly enhanced yield, confirming the efficacy of PRP culture for promoting sEV production. Additionally, the nanoparticles isolated under these conditions were found to present a cup-shaped characteristic (Figure [80]S3) and undetectable contaminant signals (Fig. [81]2J). PRP-based MSC-sEVs showing enhanced biological activity in corneal wound healing Building on previous findings that mechanical stimulation can endow sEVs with unique therapeutic potential including tissue repair and inflammation regulation [[82]18, [83]19], we thus subsequently explored the regenerative potential of PRP-MSC-sEVs using in vitro and in vivo models. According to the scratch assay using sEVs with equivalent amount (Fig. [84]3A), 15×g PRP products (PRP-sEVs) showed the most prominent effect on the migration of human corneal epithelial cells (Fig. [85]3B). Despite a minimal initial repair effect at 12 h, a dramatic improvement was observed at 24 h (Figure [86]S3A), signifying the most rapid healing rate induced by 15×g-sEV therapy. In addition, increasing the dosage of PRP-sEVs 4-fold (4×) accelerated wound healing of the epithelial cells, while a 4-fold dilution (1/4×) significantly slowed it down (Fig. [87]3C and Figure [88]S3B-C). The pronounced increase in cell proliferation further suggests the tissue repair potential of PRP-MSC-EVs (Fig. [89]3D). Since PRP-MSCs could secret more sEVs than others, these findings together support the idea that PRP-culture can drive both sEV production and their regenerative potential. Fig. 3. [90]Fig. 3 [91]Open in a new tab PRP-sEVs promote corneal epithelial wound closure in both in vitro and in vivo models. (A) Scratch assay images of human corneal epithelial cells (HCECs) cultivated in the medium supplementary with PBS (set as control), L4-sEVs, 0×g-sEVs, 15 − 11×g-sEVs and 15×g-sEVs. (B) Quantification of HCECs migration upon various sEV samples. PBS was set as control. n = 3, ****p < 0.0001 vs. 15×g group. (C) Quantitation of dose-dependent scratch assay. 2,500 15×g-sEVs per cell was set as 1× dose. n = 3, *p < 0.05 and ****p < 0.0001. (D) Proliferation assay of HCECs (n = 3, *p < 0.05 and **p < 0.01). (E) Representative fluorescein staining images representing the healing process of murine corneal epithelium. (F) The wound rate of corneal epithelial areas at the indicated times. n = 3, *p < 0.05, ***p < 0.001 and ****p < 0.0001. Two-way ANOVA was performed to investigate the effects of time and sample type (B, F) or dose (C) for scratch wound healing, while one-way ANOVA was performed for (D). The error bar represents the SEM To validate the preclinical efficacy and therapeutic advantages of PRP-MSC-sEVs, we adopted a murine corneal epithelial wound model and monitored the epithelial closure via fluorescein uptake (Fig. [92]3E). Remarkably, corneal epithelial wounds treated with 15×g-sEVs exhibited almost complete re-epithelialization within 36 h, significantly outperforming the PBS and static-cultured sEV (L4 and 0×g) treatment groups which displayed 13–35% residual wound area (Fig. [93]3F). Subsequent H&E staining revealed corneal edema, epithelial thinning, and disorganized stromal layer in the injured corneas, meanwhile the administration of 15×g PRP-sEVs led to the best restoration of cellular organization (Fig. [94]4A). We then observed that the 15×g PRP-sEVs significantly mitigate the corneal inflammatory response with reduced expression of TNF-α and IL-1β (Fig. [95]4B-C). Immunofluorescence staining further showed a significant recovery in the expression of cellular CK-3 and ZO-1 following PRP-sEVs treatment, suggesting a reversal of the compromised epithelial barrier function (Fig. [96]4D-E). Besides, it also effectively reduced the TUNEL signals representing cell apoptosis (Fig. [97]4F-G). Overall, based on our in vitro and in vivo therapeutic models, PRP-sEVs showed dual potentials toward both anti-inflammation and regeneration. Fig. 4. [98]Fig. 4 [99]Open in a new tab PRP-sEVs showing dual therapeutic potentials towards both anti-inflammation and regeneration in the murine model. (A) H&E staining the eyeballs at 36 h after treatments. (B) Representative immunohistochemical images of TNF-α and IL-1β after 36 h-treatments. (C) Relative mRNA expression of corneal TNF-α and IL-1β. (D) Immunofluorescence staining of corneal with CK-3 (green) and ZO-1 (red). The cell nuclei were stained blue with DAPI. (E) Statistics of relative fluorescence intensity of CK-3 and ZO-1. (F) Representative immunofluorescence histochemistry images of TUNEL (green). DAPI was used for staining the nuclei. (G) Statistics of relative fluorescence intensity of TUNEL. One-way ANOVA was performed for generating statistical p-value (n = 3, ***p < 0.001 and ****p < 0.0001 versus the normal group). All data was presented as mean ± standard deviation PRP culture modulates cellular transcription and behavior We further employed transcription analysis to explore the underlying molecular mechanisms of PRP culture strategies in regulating cell behavior. When considering all transcripts, GSEA showed that KEGG pathways related to protein synthesis (aminoacyl-tRNA biosynthesis) and protecting cells from environmental stress (base excision repair) were significantly enriched in the 15×g group, whereas the inflammatory response (e.g., MAPK, PI3K-Akt, and IL-17 signaling pathway) and cell death (necroptosis) related pathways were enriched in the 0×g group (Fig. [100]5A). The differential gene (p < 0.05, fold change > 1.5) analysis also demonstrated the distinct pathway enrichment profiles of two groups. While most of genes linked to inflammatory actives were markedly downregulated in the 15×g group, those associated with cell behaviors (cytoskeleton and cell adhesion) and vesicular transport (SNARE interactions in vesicular transport and endocytosis) were significantly upregulated (Fig. [101]5B-C and Figure [102]S4). Fig. 5. [103]Fig. 5 [104]Open in a new tab PRP culture induced alternations of cellular transcription and behavior. (A) The ridge plot of GSEA with transcriptions from 15×g and 0×g cells. (B) KEGG and KOG categories based on differentially expressed genes. p < 0.05 was the significant enrichment screening parameters for the GSEA, KEGG and KOG pathways. (C) Heat map showing the mRNAs significantly upregulated in the 15×g MSCs compared to the 0×g group (except CTNNB1). Molecules, such as those related to the ESCRT complex and adhesion family, were highlighted. (D) Western blot of cellular β-actin, connexin 43, and β-catenin under different conditions (n = 3, **p < 0.01, mean ± SEM) To explore the tangible effects of external stress on cellular behavior, we evaluated alterations in cell growth direction under a range of culture conditions (Figure [105]S5). The inherent spindle morphology of mesenchymal stem cells presents a unique opportunity for studying cellular behavior dynamics. After calculating the individual intersection angles, cellular behavior dynamics were assessed by comparing the correlation coefficient of distribution between the initial and final states (Figure [106]S5A-C). The weak correlation coefficient under conventional static conditions (horizontal placement without centrifugation) indicates that cells undergo migration and deviation in growth direction during culture (Figure [107]S5C). Vertical placement or rotation of the culture flask could reduce the deviation in cell orientation, with the vertical rotation group (i.e., experimental group) showing the most significantly stable results (correlation coefficient > 0.4, p < 0.05). Besides, although individual cell tracking was limited by the monitoring equipment, the consistent bulk data, coupled with observed slower enzymatic digestion and increased expression of β-catenin (Fig. [108]5D), strongly suggest that rotational culture alters cell adhesion and growth. Remodeling of sEV cargo via PRP stimulation We concurrently depicted the proteomic profiles for 15×g sEVs and 0×g sEVs to explore the potential molecular mechanisms involved in the enhanced therapeutic efficacy of PRP-MSC-sEVs. We identified 73 proteins upregulated and 83 proteins downregulated in the 15×g group (Table [109]S1). Interestingly, these differentially expressed proteins (DEPs) exhibited distinct subcellular localizations. Compared to the control group, PRP treatment enhanced the extracellular components of sEVs, while the untreated group enriched more proteins associated with cytoplasmic/nuclear localization (Figure [110]S6). The enrichment analysis using DEPs revealed that the sEVs derived from 15×g-stimulated MSCs accumulate more protein cargos associated with cellular proliferation, extracellular matrix, and congenital immunity (Fig. [111]6A-B). Among these, the complement and coagulation cascades were the most significantly enriched biological functions in 15×g sEVs. The complement system, a component of innate immunity, interacts with various components of the hemostatic pathway to maintain physiological homeostasis [[112]20]. Proteolytic components involved in coagulation and their associated regulatory factors contribute to the balance of hemostasis and thrombosis [[113]21]. These components are also key players in various important biological processes such as promoting cell proliferation and wound healing. Fig. 6. [114]Fig. 6 [115]Open in a new tab Comparative analysis of sEV proteins following PRP treatment. (A) Enriched KEGG terms for DEPs (15×g versus 0×g). (B) Line graph illustrating log2 fold change in protein abundance for proteins associated with cell cycle, cellular senescence, and the complement and coagulation cascades, comparing 15×g and 0×g sEVs (left). A rightward shift indicates an increased frequency of proteins upregulated in 15×g sEVs. Heatmap of significantly upregulated and downregulated proteins in 15×g sEVs related to the biofunctions of cell cycle, cellular senescence, and the complement and coagulation cascades (right). (C) 15×g-induced fold changes of cell mRNA and sEV protein abundance. (D) A volcano plot of the abundance correlation between mRNA and protein pairs. Pairs with highly positive (r > 0.8, p < 0.05) or negative (r < -0.8, p < 0.05) correlations were considered significant and highlighted in red and blue, respectively. Subsequent enrichment analysis was performed to uncover the biological significance of these significantly correlated genes Since cellular mRNA regulate protein biogenesis to a certain extent (r from 0.3 to 0.8) [[116]22, [117]23], we were prompted to whether cellular transcriptional levels also molding the sEV protein composition. Globally, our analysis of cell mRNA and corresponding sEV protein abundance under control (0×g) and PRP (15×g) conditions demonstrated a lack of correlation between their fold changes (Fig. [118]6C). Nevertheless, some of individual mRNA-protein pairs presented highly positive or negative correlation (Fig. [119]6D), suggesting the presence of both cooperative and antagonistic interactions in cellular transcription and the sorting of proteins into sEVs. Enrichment analysis further revealed a coordinated change in mRNA/protein levels associated with immune regulation and vesicle formation between cells and sEVs, while components related to the extracellular matrix exhibited a more variable correlation. Discussion Owing to their superior properties, including diverse components, stable membrane structure, and low immunogenicity, sEVs are increasingly investigated as therapeutic targets, drug delivery vehicles, and standalone therapies [[120]24, [121]25]. However, the limited production efficiency and endogenous cargo capacity hinder large-scale clinical applications. Appropriate mechanical stimulation, can significantly enhance sEV yield by modulating secretion mechanisms under specific loading conditions, without introducing external substances [[122]26]. In this study, we proposed a novel approach based on programmable rotating platform to stimulate sEV secretion. Our PRP platform stands out for its simplicity, cost-effectiveness, and compatibility with standard cell culture incubators, eliminating the need for specialized equipment. The well-established therapeutic applications of mesenchymal stem cells make them an ideal case study for this platform. Experimental data revealed that rotary culture is an efficient and safe method for MSC-sEV production, with normal stress from the rotating liquid medium identified as the dominant force mechanism confirmed by computational approach. Under condition of 15×g PRP culture for 12 h, 4-fold promotion in sEV yield could be readily achieved compared to static culture, without compromising cell viability (> 95%). Although prolonged stimulation could yield more extracellular particles, the concomitant decrease in cell viability (cell mortality rate > 10%) suggested the potential introduction of more apoptotic bodies. A combined in vitro (HCEC scratch assay and proliferation assay) and in vivo (murine corneal wound model) approach demonstrated that PRP-induced sEVs significantly outperformed sEVs from other conditions in promoting wound healing, inhibiting inflammation, and preserving corneal tissue integrity. These findings suggest that PRP-sEVs carry more bioactive molecules beneficial for tissue repair, potentially leading to better therapeutic outcomes. Mechanical stimulation triggers the activation of multiple intracellular signaling pathways, including MAPK and PI3K/Akt [[123]27, [124]28], which subsequently influences the generation, loading, and release of sEVs, as well as modifies the composition and abundance of their bioactive components, thereby influencing the biological outcomes and therapeutic effects. Among the signaling pathways identified, MAPK and PI3K are well-known to play critical roles in regulating cellular processes such as growth, differentiation, and apoptosis. However, our transcriptomic analysis revealed that these pathways were dysregulated and preferentially enriched in cells without centrifugating stress. This could be partially attributed to negative feedback mechanisms and the crosstalk between multiple pathways triggered by mechanical stress [[125]29]. More importantly, key genes in the MAPK and PI3K pathways known to be activated by mechanical stress, such as JNK2 (MAPK9), p53 (MAPK12), and PI3K subunits (PIK3CB, PIK3R3), were actually significantly upregulated in 15×g cells, consistent with previous studies. In addition, it’s worth mentioning that while β-actin and β-catenin expression (both transcriptional and protein levels) either increased significantly or subtly, connexin expression (associated with intercellular gap junctions) remained relatively stable (Fig. [126]5D and Figure [127]S4). These results are in line with the pathway analysis, suggesting that PRP stimulation primarily induced changes in cytoskeletal organization and cell adhesion. Our proteomic analysis of sEVs revealed a significant enrichment of pathways and proteins involved in inflammatory wound healing in 15×g sEVs compared to 0×g sEVs. Of note, among the DEPs of these categories, complement factor I (CFI) and complement C3 are upregulated and downregulated in 15×g sEVs, respectively (Fig. [128]6B). C3 serves as a central component in the complement cascade, initiating the formation of the membrane attack complex and inflammatory mediator C3α. CFI, on the other hand, acts as a regulatory factor, preventing excessive complement activation and tissue damage [[129]30]. Beyond corneal reepithelization, we also identified a retinal homeostasis associated DEP, that is, SERPINF1 (serpin family F member 1). Also known as pigment epithelium-derived factor, SERPINF1 is a secreted multifunctional protein that protects retinal neurons from light damage, oxidative stress, and glutamate excitotoxicity [[130]31]. Our results show a significant upregulation of SERPINF1 in the 15×g group and this finding was validated in an independent set of pooled samples (Table [131]S1), suggesting the neuroprotective potential of PRP-MSC-sEVs. Further, as expected, no significant correlation was found between the cellular transcriptome and the sEV proteome. This suggests that alterations in cellular mRNA levels following PRP stimulation do not directly dictate protein composition within sEVs, but are rather affected by multiple biological processes, such as post-transcriptional regulation [[132]32] and specific sorting mechanisms that govern protein incorporation into sEVs [[133]33]. Finally, when focused on the mRNA-protein pairs with significant correlations, we observed a coordinated involvement of cellular and sEV components in innate immunity and vesicular trafficking. These findings, coupled with the cellular pathway enrichment analysis, offer an underlying mechanistic explanation for the PRP system’s role in promoting MSC-sEV production. This study has several limitations. Firstly, the investigation primarily focused on MSCs, and the applicability of this method to other cell types remains to be determined. Different cell types may exhibit varying responses to mechanical stimulation, necessitating optimization of parameters such as rotational speed, pressure range, and duration for each cell type. Therefore, future studies are crucial to optimize parameters for diverse cell types and validate the methodological applicability. Furthermore, while a correlation between the endocytosis and increased sEV number was observed, the specific molecular signaling pathways and regulatory mechanisms involved in this process require further elucidation. For instance, we should delve deeper into how mechanical stimuli influence intracellular signaling pathways to promote sEV generation. After identifying core regulatory factors and their downstream networks, specific inhibitors should be used to treat key downstream nodes to observe cell and EV changes [[134]14]. Another limitation is that the increase of sEV yield did not reach the levels achieved by some other methods. However, the stimulation platform proposed in this study is convenient, scalable, and can enhance the biological activity of sEVs. Conclusion In this study, we developed an easy-to-use approach for enhancement of sEV production by subjecting MSCs to the continuous normal stress induced by well-controlled centrifugating force. This method resulted in a nearly 4-fold increase in the quantity of sEVs and meanwhile significantly enhanced their ability to promote corneal wound healing. Mechanical stimulation activated the cytoskeleton, cell adhesion, shear stress response, and endocytic pathway, thereby influencing cellular behavior and sEV biogenesis. The therapeutic potential of PRP-sEVs can be attributed in part to the selective sorting of bioactive molecules that are poised to reduce inflammation and promote tissue repair. Consequently, PRP-based mechanical stimulation presents a promising and effective strategy for promoting the sEV production, offering invaluable support for accelerating translational research and advancing the clinical application of MSC-sEVs. Experimental section Cell culture Human umbilical cord mesenchymal stem cells from 3 donors were purchased from Yingrun Biotechnologies Co., Ltd. in China. The cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycins (P/S). DMEM and FBS were purchased from Gibco (Waltham, Massachusetts, USA), P/S was purchased from Thermo Fisher Scientific, Inc. (Shanghai, China). For all experiments, HU-MSCs at passage 6 or earlier were used. The cells were cultured as a monolayer attached to the bottom of the culture flask. When reaching 80–90% confluence, the medium was replaced with serum-free medium for subsequent experiments. Additionally, immortalized human corneal epithelial cells (HCECs) were gifted by Prof. Kaihui Nan (Wenzhou Medical University). Both cell types were cultured in an incubator maintained at 5% CO[2] and 95% humidity. After treatment, the culture medium was supplemented with Calcein-AM and Propidium Iodide (PI) (C2015M, beyotime) for staining the live and dead cells, respectively. Composition and operation of the working platform The work platform consists of eight concave barriers and a central disk. Before placing it inside the incubator, the work platform must be thoroughly disinfected using UV irradiation and a 75% (v/v) ethanol solution to effectively prevent the intrusion and contamination of microorganisms. T25 culture flasks containing seeded cells were securely positioned within the eight grooves, ensuring firm contact with the inner walls to maintain stability during the experiment. Subsequently, the stepper motor controller parameters were adjusted. The controller then generated pulse signals and directional commands. Upon receiving these control signals, the stepper motor driver, equipped with an integrated logic circuit, precisely energized the stepper motor windings in a predefined sequence, facilitating the intended motor operation. To investigate the impact of PRP on cells, a multi-physics simulation software (COMSOL 6.2a) was employed for modeling elasticity mechanics and theoretical mechanics within the PRP. In this model, the culture flasks were assumed to be linear elastic materials, and the equations of solid mechanics were adopted. The framework for simulating turbulent flow is constructed using the Navier-Stokes equations (momentum equations), the continuity equation, turbulence model equations, dissipation rate equations, turbulence annual models, and turbulence generation terms. sEV Isolation and Characterization: For sEV Isolation, cell culture supernatant was collected and then processed following sequential centrifugation steps at 500 ×g for 10 min and 3,000 ×g for 30 min. After concentrated using an ultrafiltration tube (30 kDa, Millipore, USA), the sEV pellets were obtained after ultracentrifugation at 120,000 ×g for 120 min and finally resuspended in 200 µL of PBS for further analysis. The number and size distribution of nanoparticles were measured using Nanoparticle Tracking Analysis technology (NanoSight NS300, Malvern, USA). Each sample was measured three times, with a collection time of 30 s for each measurement. Samples for transmission electron microscopy (TEM) imaging was prepared according to our previous work [[135]34] and the imaging was performed on TEM (Tecnai 12, Philips company, Holland). Based on the analysis of Qubit TM protein assay kit (2216924, Invitrogen, USA), 40 µL of sEV samples and 20 µg of cell lysates were used for immunoblotting. The presence of various sEV-associated proteins was assessed using primary antibodies against Alix (Santa Cruz, sc53540), CD9 (Santa Cruz, sc13118), Calnexin (Santa Cruz, sc23954), Annexin V (Santa Crue, sc74438) and CD63 (Santa Cruz, sc55275), while the cell proteins was accessed using antibodies for GAPDH (Santa Cruz, sc365062), connexin 43 (Santa Cruz, sc-271837), β-catenin (Santa Cruz, sc7963), and β-actin (Santa Cruz, sc47778). All primary antibodies are diluted at 1:1000, and all secondary antibodies are diluted at 1:3000. In vitro scratch closure assay After seeding and reaching 90% confluence in 96-well culture plates, HCECs were serum-starved overnight. The cell monolayers were scratched along the centerline using a pipette tip, and subsequently treated by PBS or sEV (2,500 particles per cell) [[136]6, [137]13, [138]35] samples. Images of cell migration were observed and captured using an inverted fluorescence microscope (IX73, OLYMPUS) and analyzed by using Image J (v1.53t). Migration area = initial area - area at the current time point. In vitro cell proliferation assay 5-ethynyl-20-deoxyuridine (EdU) assay was performed to investigate PBS and sEVs on proliferation of HCECs. After PBS or sEV treated, cells were labeled with EdU solution (Beyotime), fixed with 4% (w/v) PFA solution, permeated by 0.3% (v/v) Triton X-100 solution (Sigma-Aldrich), labeled with click reaction buffer with Azide Alexa Fluor 488, and stained by Hoechst, subsequently. Fluorescence images were counted by Image J to analyze the rates of EdU positive cells. Animals Female C57BL/6 mice (7 to 8 weeks of age; body weight 20–24 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. The mice were housed in a specific-pathogen-free environment at the Shandong Eye Institute animal facility. All animal procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and approved by the Ethics Committee of Shandong Eye Institute (SDSYKYJS No.20240316). Corneal epithelial wound healing A corneal epithelium debridement model was established in line with an earlier publication [[139]36]. Subsequently, different sEVs (1 × 10^10 EVs per mL) [[140]37] are added. All procedures adhered to guidelines approved by the Ethics Board at the Shandong Eye Institute. Following sedation with an intraperitoneal pentobarbital sodium solution at 0.6% (dose of 50 mg/kg) and application of 0.5% proparacaine, the mid-cornea (approximately 2.5 mm in diameter) was removed from each mouse using an Algerbrush II burr under a 1FR Pro operating microscope from Zeiss (Jena, Germany). Wound recovery was monitored with 0.25% fluorescent dye to highlight epithelial disruptions, and images were captured using an SL-D701 system from TOPCON in Japan. The proportion of area with damage was measured via Fiji (ImageJ 1.53c) according to earlier accounts [[141]38]. Hematoxylin and Eosin (H&E) staining Thirty-six hours after corneal epithelium disruption, murine eyes were obtained and immersed in fixative for an overnight period. They then underwent dehydration, followed by paraffin infiltration, and were trimmed into 5 μm slices. Subsequently, each section was stained with H&E in combination with a diaminobenzidine-based reagent. Immunohistochemistry The eyeballs were fixed with tissue fixative, dehydrated, and sliced, and proceeded with dewaxing, rehydration, and antigen repair. Then, an endogenous catalase blocker blocks endogenous enzymes, and 5% BSA blocks them. The corneal slices were incubated with a primary antibody overnight at 4 °C. Primary antibodies used targeted TNF-α (Proteintech, 60291-1) and IL-1β (Abcam, ab9722). On the second day, the samples were incubated with HRP-conjugated secondary antibodies (Gene Tech, GK500510A) for 1 h at room temperature. The DAB chromogenic solution was used to catalyze the formation of brownish-yellow products. Finally, hematoxylin staining, dehydration, sealing, and Upright Microscope imaging (DM4B, Leica, Germany) were performed respectively. Quantitative real-time PCR An RNA isolation kit (TransZol Up Plus, Transgene Biotech, S30314) was employed to obtain total corneal RNA, and its concentration was measured on a NanoDrop One instrument (ND-one, Thermo, USA). Next, cDNA was generated with the cDNA Synthesis SuperMix (Vazyme, R323-01) following the manufacturer’s protocol. Subsequently, qPCR took place employing ChamQ SYBR qPCR Master Mix from Vazyme (Q711-02) on a QuantStudio5 device (Thermo, USA). Cycling started with a 30-second hold at 95 °C, then proceeded through 40 rounds featuring 5-second intervals at 95 °C and 30-second intervals at 60 °C. β-actin served as the internal control. All assays were carried out in triplicate. Primer sets are shown in Table [142]1. Table 1. List of primers used in this study Gene Forward Primers (5’--3’) Reverse Primers (5’--3’) m-actin CATGTTCGTCATGGGTGTGAA GGCATGGACTGTGGTCATGAG m-TNFα ACAAGGCTGCCCCGACTAC TGGGCTCATACCAGGGTTTG m-IL1β CTTTCCCGTGGACCTTCCA CTCGGAGCCTGTAGTGCAGTT [143]Open in a new tab Immunofluorescence For immunofluorescence staining, the mouse eyeballs were embedded by OCT tissue freezing medium at − 80℃ for 30 min, and sliced at a 7-µm thickness. Subsequently, the frozen sections were washed three times with PBS, fixed in 4% paraformaldehyde (PFA) for 15 min at room temperature, and permeabilized by 0.1% TritonX-100.Next, the sections were blocked for 1 h using 5% BSA, followed by primary antibody incubation with the ZO-1 (Thermo, 40-2200) and CK3 (BIOSS, 3646R) overnight at 4 °C. The next day, after rewarming and thorough cleaning of PBS-Tween, the samples were incubated with a secondary antibody (Abcam, ab150076) for 2 h in darkness at room temperature, and DAPI (Solarbio, C0065) was used for staining for 3 min. Then, samples were washed in PBS-Tween for 3 × 5 min and covered with coverslips. Immediately after covering, samples were viewed with an immunofluorescence microscope (LSM880, ECHO, USA). Tunel staining To detect the cell death levels in mouse cornea, we used a TUNEL kit (Abbkine, KTA2010), and performed as the manufacturer’s instructions. Samples were viewed with an immunofluorescence microscope (LSM880, ECHO, USA). RNA sequencing Total RNA was isolated using the Total RNA Extractor (Trizol) kit, followed by mRNA purification with oligo(dT) magnetic beads. First-strand cDNA was synthesized using random primers, and the second strand was generated with DNA polymerase, substituting dUTP for dTTP. RNA concentration was determined using the Qubit 2.0 RNA detection kit to assess the input amount for library construction. Then, adapter and index sequences were ligated to both ends of the cDNA fragments, followed by PCR amplification to a target size of 300–400 bp. Library size was verified with the Agilent 2100 Bioanalyzer, and concentration was measured using the Qubit. Libraries were pooled based on sequencing depth requirements and sequenced using paired-end 150-bp reads (PE150) to generate over 10 Gb of clean data. Post-sequencing, quality control measures were implemented to remove low-quality reads, and the high-quality data was aligned to the reference genome for subsequent gene expression and differential expression analyses. Proteomics Protein samples were prepared in 200 µL of 8 M urea buffer. Dithiothreitol was added to a final concentration of 5 mM, and samples were incubated at 37 °C for 45 min for reduction. Subsequently, iodoacetamide was added to a final concentration of 11 mM for alkylation in the dark at room temperature for 15 min. Following alkylation, 800 µL of 25 mM ammonium bicarbonate and 2 µg of trypsin (Promega, V5280) were added for overnight digestion at 37 °C. After digestion, the pH was adjusted to 2–3 with 20% trifluoroacetic acid and the peptides were desalted using C18 resin (Millipore, Billerica, MA). Peptide concentration was determined using the Pierce™ Quantitative Peptide Assay Kit (Thermo Fisher). For liquid chromatography and mass spectrometry (LC-MS) analysis, peptides were separated on a Vanquish Neo UHPLC nano-scale system (Thermo Fisher). Mobile phase A consisted of 0.1% formic acid in water, while mobile phase B consisted of 0.1% formic acid in 100% acetonitrile. Data-independent acquisition (DIA) analysis was performed using the Vanquish Neo system coupled to an Orbitrap Astral high-resolution mass spectrometer (Thermo Scientific). In positive ion mode, the parent ion scan range was 380–980 m/z with a resolution of 240,000 at 200 m/z, a normalized AGC target of 500%, and a maximum injection time (IT) of 5 ms. For MS2 analysis, DIA acquisition was performed with 299 scan windows, an isolation window of 2 Th, HCD collision energy of 25%, a normalized AGC target of 500%, and a maximum IT of 3 ms. Statistical analysis The statistical analysis was performed using SPSS 25.0 statistical software and GraphPad Prism 9.1 software. All data are expressed as means ± standard deviation (SD) or standard error of mean (SEM). One-way analysis of variance (ANOVA) and t-tests was used to determine the statistical differences for multiple and two-groups comparisons, respectively. A two-way ANOVA multiple comparison was employed to analyze the statistical differences for migration dynamics upon various sEV samples or sEV doses. Spearman correlation analysis was used for generating the correlation coefficient. Electronic supplementary material Below is the link to the electronic supplementary material. [144]Supplementary Material 1^ (14.3MB, docx) Acknowledgements