A scalable platform for EPSC-Induced MSC extracellular vesicles with therapeutic potential

Abstract

Background

Extracellular Vesicles (EVs) derived from mesenchymal stem cells (MSCs) have gained recognition as promising therapeutic and drug delivery agents in regenerative medicine. However, their clinical application is limited by donor variability, low scalability, and inconsistent therapeutic quality. To overcome these challenges, a robust and standardized production platform is urgently needed.

Methods

We developed a scalable biomanufacturing strategy by generating and expanding MSCs from extended pluripotent stem cells (EPSC) using a suspension bioreactor culture system. A fixed-bed bioreactor was integrated for automated, continuous expansion of iMSCs and downstream EV harvesting. EVs were isolated through a streamlined protocol and characterized for size, morphology, surface markers, and bioactivity. Therapeutic efficacy was assessed in a bleomycin-induced pulmonary fibrosis mouse model.

Results

iMSC-derived EVs (iMSC-EVs) exhibited comparable characteristics to primary MSC-EVs, including a size distribution of 70–80 nm, cup-shaped morphology, and expression of canonical EV markers (CD63, CD81, TSG101). iMSCs were expanded for up to 20 days in 3D culture, yielding > 5 × 10⁸ cells per batch using a suspension bioreactor culture system and producing ~ 1.2 × 10¹³ EV particles/day in a fixed-bed bioreactor. In vivo, iMSC-EVs significantly reduced Ashcroft fibrosis scores and bronchoalveolar lavage fluid protein levels in bleomycin-injured lungs, with therapeutic efficacy comparable to primary MSC-EVs.

Conclusions

This study establishes a scalable and standardized platform for producing high-quality iMSC-EVs using bioreactor-based systems. Our approach addresses key limitations in traditional EV production and sets the stage for AI-integrated, fully automated, GMP-compliant manufacturing of therapeutic EVs suitable for clinical translation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-025-04507-y.

Introduction

MSC-derived EVs have emerged as a promising cell-free therapeutic modality for regenerative medicine, due to their immunomodulatory, anti-inflammatory, and tissue-repair properties [1]. EVs are nanoscale lipid bilayer-enclosed vesicles that mediate intercellular communication by transferring functional cargoes such as proteins, lipids, and nucleic acids [2, 3]. In preclinical and clinical settings, MSC-EVs have shown therapeutic benefits in a wide spectrum of diseases, including pulmonary fibrosis, cardiovascular injury, and autoimmune disorders [4–7].

Compared to cell-based transplantation, EV therapies offer advantages in safety, biodistribution, and storage, while avoiding issues such as immune rejection and embolism [8]. Proof-of-concept studies in lung, heart, and other organs show that MSC-EVs suppress inflammation, limit fibrosis, and promote functional recovery [9–11]. In cardiovascular applications, porcine cardiac adipose tissue MSC-EVs have been shown to modulate inflammatory and reverse remodeling after myocardial injury in rodent and porcine models [12, 13]. However, despite growing enthusiasm, the translational application of EVs remains hampered by several critical bottlenecks, including donor variability, limited scalability, and lack of standardized production protocols [11].

Most current MSC-EV manufacturing relies on primary MSCs derived from bone marrow, adipose tissue, or umbilical cord [14]. These sources suffer from finite expansion capacity, phenotypic drift during in vitro passaging, and batch-to-batch heterogeneity [15]. Moreover, large-scale EV production requires robust cell sources and bioprocesses that can meet GMP standards and clinical demand. To overcome these limitations, induced pluripotent stem cell-derived MSCs (iMSCs) have been explored as renewable and consistent EV-producing sources [16]. iMSCs offer the potential for unlimited expansion and clonal stability, and can be gene-edited at the pluripotent stage, enabling the production of customized EVs. However, the challenge remains to translate these advantages into a clinically applicable, scalable EV production platform.

In parallel, recent advances in bioreactor technology have enabled higher-density adherent cell cultures with improved EV yield and process control [17, 18]. Among these, fixed-bed bioreactors support GMP-compatible production by providing uniform nutrient perfusion and minimal shear stress [19]. Yet, few studies have integrated pluripotent-derived iMSCs with fixed-bed bioreactor systems for EV production, and fewer still have benchmarked their functional performance in disease models.

This study aims to develop a fixed-bed bioreactor system for large-scale iMSC expansion and EV production, addressing key challenges in scalability and consistency. The specific objectives are: (1) to establish a robust and scalable platform for generating iMSCs using a microcarrier suspension system. iMSCs were directly differentiated from EPSCs, providing a stable, renewable, and phenotypically consistent cell source; (2) to develop a standardized biomanufacturing process for iMSC-EVs using a fixed-bed bioreactor; (3) to comprehensively characterize iMSC-derived EVs and compare them to primary MSC-derived EVs in terms of size, morphology, and protein markers; and (4) to evaluate the therapeutic efficacy of iMSC- EVs in a bleomycin-induced pulmonary fibrosis mouse model. By achieving these objectives, this study aims to demonstrate the potential of iMSC-EVs as a scalable and effective alternative to primary MSC-EVs, supporting their application in regenerative therapies.

Materials and methods

Induction of EPSC from iPSC

The protocol for EPSC induction was adapted from previously reported methodologies [20, 21]. iPSCs were cultured to 70–80% confluency, dissociated using TrypLE-DPBS (1:1), and seeded at 1 × 10⁵ cells per well onto Matrigel-coated plates in EPSC medium. After attachment (Day − 1), the medium was replaced with EPSC induction medium on Day 0 and refreshed every 2 days. On Day 6, EPSCs were evaluated for morphology and marker expression. The EPSC induction medium consisted of a 1:1 mixture of knockout DMEM/F-12 (Thermo Fisher Scientific, 12660012) and Neurobasal medium (Thermo Fisher Scientific, 21103049), supplemented with 0.5×B27 (Thermo Fisher Scientific, 12587010), 0.5×N2 (Thermo Fisher Scientific, 17502048), 5% knockout serum replacement (Thermo Fisher Scientific, 10828028), 1% GlutaMAX (Thermo Fisher Scientific, 35050061), 1% non-essential amino acids (Thermo Fisher Scientific, 11140050), 1% penicillin-streptomycin (Thermo Fisher Scientific, 15140122), 0.1 mM β-mercaptoethanol (Thermo Fisher Scientific, 31350010), recombinant human LIF (10 ng/mL, PeproTech, 300-05), CHIR99021 (1µM, Stemcell Technologies, 72054), (S)-(+)-dimethindene maleate (2µM, ApexBio, B6734), minocycline hydrochloride (2µM, MEC, HY-17412), Y-27,632 (2µM, Abcam, ab120129),XAV939 (2µM, Abcam, ab120897) and Human Recombinant Activin A (40 ng/mL, Stemcell Technologies, 78001).

Induction of trophoblast-like stem cells (TLCs) from EPSCs

EPSCs were cultured until reaching 70–80% confluency. In a sterile environment, the culture medium was removed, and cells were washed twice with pre-warmed DPBS. For cell dissociation, 2 mL of TrypLE solution (Thermo Fisher Scientific, 12563029) was added to T25 flasks, incubated at 37 °C for 2 min, followed by an additional 2 min after removal of the initial dissociation solution. Detached cells were resuspended in DMEM/F-12 (Thermo Fisher Scientific, 11320033), collected, centrifuged at 1000 rpm for 5 min, and counted using trypan blue. Cells were seeded onto Matrigel-coated plates at a density of 1 × 10^5 cells/well in EPSC complete medium, designated as Day − 1. After 24 h, EPSCs were transitioned to TLC differentiation medium, with full medium replacement every 2 days until Day 6. The TCLs induction medium consisted of a 1:1 mixture of DMEM/F-12 and IMDM medium(Thermo Fisher Scientific, 31980030), supplemented with 20% knockout serum replacement, 0.1 mM β-mercaptoethanol, Recombinant Human EGF (50 ng/mL, PeproTech, AF-100-15), CHIR99021 (2µM, Stemcell Technologies, 72054), A83-01 (0.5µM, Santa Cruz Biotechnology, sc-203791 ), SB431542 (1µM, Abcam, ab120163 ), Y-27,632 (5µM, Abcam, ab120129) and Valproic acid (0.8mM, Selleck, S3944 ). On Day 6, cells were analyzed by flow cytometry for MSC markers (CD44, CD73, CD90, CD105, CD166) and negative markers (HLA-DR, CD34, CD45).

Directed differentiation of TLCs into MSCs

Matrigel-coated T25 flasks were prepared and incubated at 37 °C for at least 30 min before seeding. Following TLC differentiation, cells were dissociated with TrypLE-DPBS (1:1), centrifuged at 200 g for 5 min, and resuspended in MSC complete medium consisted of α-MEM and 5% HPL (Helios Bioscience, HPCFDCRL50). Cells were seeded at a density of 8 × 10⁴ cells/cm², designated as passage 1 (P1), and cultured at 37 °C in 5% CO₂ with medium changes every 3 days. Upon reaching confluence (~ 100%), cells were sequentially passaged. Subsequent passages were performed at densities of 4 × 10⁴/cm² (P2), 2 × 10⁴/cm² (P3), and 1 × 10⁴/cm² (P4) using 0.1% gelatin-coated plates. Cells were analyzed for MSC markers (CD44, CD73, CD90, CD105, CD166) and negative markers (HLA-DR, CD34, CD45) upon reaching passage 4. Expansion continued to P12 for experimental use.

RNA extraction and quantitative PCR (qPCR) analysis

Total RNA was extracted from Trizol-preserved samples (Thermo Fisher Scientific, 15596018) thawed at room temperature. Chloroform (200µL/mL Trizol, Sigma-Aldrich, C2432) was added, mixed for 15 s, incubated for 2 min, and centrifuged (12,000 rpm, 20 min, 4 °C). The aqueous phase was transferred, mixed with ice-cold isopropanol, incubated at -20 °C for 20 min, and centrifuged (12,000 rpm, 10 min, 4 °C). The RNA pellet was washed with 70% ethanol, air-dried, and resuspended in 40 µL RNase-free water. cDNA was synthesized using a commercial kit (Beijing Zoman Biotechnology, ZR102-3). qPCR was performed in a 20 µL reaction with 2× qPCR mix (TransGen Biotech, AQ131-04), 2 µL cDNA, 0.5 µL primers, and nuclease-free water. Cycling conditions: 95 °C for 5 min; 40 cycles of 95 °C for 10 s, 60 °C for 20 s, 72 °C for 10 s; final melt curve at 37 °C for 2 min. Primer information is provided in table S1.

Live/dead cell staining and imaging

Cells or cell-embedded scaffolds were gently rinsed with PBS and transferred to 1 mL of fresh culture medium containing 0.5 µL of Live stain (green) and 2 µL of Dead stain (red), following the manufacturer`s protocol (Thermo Fisher Scientific, L3224). After incubation at 37 °C under 5% CO₂ for 15 min, scaffolds were washed four times with PBS, placed in fresh medium, and imaged using a spinning disk confocal microscope.

MSC tri-lineage differentiation

To assess MSC tri-lineage differentiation potential, cells were seeded in 6-well plates at 70–80% confluence. For adipogenesis, we used MesenCult™ Adipogenic Kit (Stemcell Technologies, 05412) and performed Oil Red O staining after 14 days. For osteogenesis, MesenCult™ Osteogenic Kit (Stemcell Technologies, 05465) was applied for 21 days, followed by Alizarin Red S staining. For chondrogenesis, MSC pellets were cultured in MesenCult™-ACF Chondrogenic medium (Stemcell Technologies, 05455) for 21–28 days, then stained with Alcian Blue. Differentiation was confirmed by microscopic visualization of lipid droplets, mineralized nodules, and glycosaminoglycans, respectively.

Single-cell RNA sequencing and data processing

Single-cell RNA sequencing (scRNA-seq) libraries were prepared and sequenced using the Illumina HiSeq 4000 platform with paired-end high-throughput sequencing. Raw sequencing data were processed using the BD Genomics platform, where low-quality reads were filtered, and the remaining reads were aligned to the reference genome to generate RDS files for downstream analysis. To ensure high-quality single-cell data, we applied stringent quality control measures using the Seurat package in R. Cells with mitochondrial gene expression < 10%, ribosomal gene expression < 8%, and a minimum of 500 detected genes were retained for further analysis. Batch effects across different single-cell samples were corrected before integration.

For dimensionality reduction, principal component analysis (PCA) was performed using the RunPCA function, followed by Uniform Manifold Approximation and Projection (UMAP) clustering with the RunUMAP function for visualization. Differential gene expression analysis was conducted using the FindAllMarkers function to identify marker genes for distinct cell populations. Gene expression visualization was performed using the ggplot2 package in R.

Expansion of iMSC in bioreactor

iMSCs were expanded via microcarrier-based culture in bioreactors to achieve scalable production. Firstly, 5.5 × 10⁷ iMSCs and 2.7 g of porous microcarriers 3D TableTrix^® (CytoNiche Biotech) were added to a sterile vivaSPIN bioreactor (CytoNiche Biotech) with 1.5 L MSC medium and with controlled parameters: temperature at 37◦C, pH at 7.2 and dissolved oxygen at 40%. Initial cell seeding (Day 0) involved intermittent stirring cycles: 35 rpm for 5 min, followed by 0 rpm for 55 min, repeated for 24 cycles to facilitate cell attachment to microcarriers. On Day 1, medium was added to reach a total volume of 2 L, and continuous stirring was set at 35 rpm. As cells entered stable proliferation, stirring speeds were increased to 55 rpm on Day 4.

Medium exchange

Medium exchanges were performed when glucose levels fell below 2 mmol/L, occurring on Days 3 and 5. The procedure involved halting stirring to allow microcarrier settling, then removing 1.0 L of spent medium and replacing it with fresh medium.

Cell harvesting

To harvest cells, stirring and temperature were stopped to allow microcarrier settling, and the culture supernatant was removed. Microcarrier suspension was transferred to a 1 L collection bottle. In a biosafety cabinet, a microcarrier dissociation solution (3D FloTrix Digest solution, 1 mg/mL, CytoNiche, R001-500) was prepared from a 10 mg/mL stock solution and filtered for sterility. After incubation in a 37 °C water bath for 30–40 min with gentle shaking, cells were collected by centrifugation at 1700 rpm for 5 min, washed twice with DPBS, and counted.

Cryopreservation

Cells were resuspended in iStock-120 cryopreservation solution, with 3.0 × 10⁸ cells aliquoted per 20 mL CryoMACS cryopreservation bag. A total of 5.75 × 10⁸ cells were harvested and divided across two freezing bags. Bags were labeled and cryopreserved in a programmable freezer, then stored in liquid nitrogen. Additionally, 4.0 × 10⁶ cells were reseeded in 2D culture to assess morphology, adherence, and marker expression by flow cytometry.

Extracellular vesicles production process

Fixed-Bed Pre-Culture

A total of 750 mL of 5% HPL-MEM-α complete medium is pumped into the fixed bed bioreactor (Tongteng Innovation Co., Ltd., CEL-G^® Culture Ad60 Series Fixed Bed Reactor System). The system is initiated in batch mode for pre-culture. After 24 h, samples are collected for mycoplasma and endotoxin testing. If the results meet quality control standards, cell seeding is initiated.

Cell seeding

Cryopreserved MSCs (300 million cells) are thawed. To minimize potential negative effects of high cell density, the cells are aspirated from the cryovial and resuspended in an equal volume of 5% HPL-MEM-α complete medium. The suspension is centrifuged at 1070 rpm for 5 min. The supernatant is discarded, and the cell pellet is resuspended in fresh 5% HPL-MEM-α complete medium, followed by cell counting for subsequent use.

Inoculation into Fixed-Bed Bioreactor: The batch culture mode is halted. The bioreactor inlet port is opened, and the prepared cells are introduced into the fixed-bed chamber. The inlet port is sealed, and batch mode is reactivated under the following conditions: temperature at 37 °C, pH 7.3, and dissolved oxygen (DO) at 50%.

Extracellular vesicles collection

After 48–72 h of culture, 500 mL of 5% HPL-MEM-α complete medium is replaced daily. The process involves activating the emptying mode to remove 500 mL of spent medium. Subsequently, the filling mode is activated to pump in 500 mL of fresh medium, followed by resumption of the circulation mode. The discarded medium is sampled, while the remainder is stored at 4 °C for backup. The collected samples undergo biochemical analysis, including pH measurement. Once stability is confirmed, the culture medium is switched to 3% HPL-MEM-α complete medium, with the same medium exchange method.

Extracellular vesicles isolation and purification

Prior to use, the entire circuit was sanitised by recirculating ~ 2 L of 0.5 mol L⁻¹ NaOH for ≥ 15 min and left static overnight. Within 24 h of operation, the system was flushed with sterile water for ≥ 30 min and equilibrated with 0.9% saline until pH 6.5–7.2. Processing parameters were set to an inlet pressure of 0.8-1.0 bar, a recirculation flow of ~ 300 mL min⁻¹, and a trans-membrane flux of 200–400 LMH. Daily conditioned medium (600–700 mL) was loaded through a 0.22 μm depth filter (Pall HP PDK11) at 5 mL min⁻¹, maintaining < 1 bar across the filter. The clarified feed was recirculated through the hollow-fiber module until the retentate volume reached 50–60 mL.

Saline was then added in 6–10 exchange volumes while maintaining the same pressures and flow; diafiltration continued until retentate conductivity and pH matched physiological values. The final EV suspension was sterile-filtered (0.22 μm, Pall KA02ECVP2G), sampled for size, particle count, protein markers, and sterility testing, and immediately stored at -80 °C, avoiding multiple freeze-thaw cycles. Freshly isolated EVs were used for all in vitro quality control experiments, while frozen EVs stored at − 80 °C were used for in vivo studies. All cartridges and filters were single-use replaced between batches to prevent cross-batch contamination.

EV size and distribution analysis

The size distribution of freshly isolated EVs was measured by the NanoCoulter counter (Resun Technology, co., LTD, Shenzhen) through a particle-by-particle manner, and the nanopore chip with the measuring range of 60–200 nm was chosen in the experiment.

Transmission electron microscopy (TEM)

Freshly isolated EV samples were prepared by diluting the purified EV suspension to a suitable concentration in PBS. A 10 µL aliquot of each EV sample was placed onto a Formvar-carbon-coated copper grid and allowed to adsorb for 1 min. Excess liquid was carefully removed using filter paper. Following adsorption, the EVs were negatively stained to enhance contrast. 2% uranyl acetate solution was applied to the grid for 1 min, after which excess stain was removed, and the grid was allowed to air dry for 5–10 min. Grids were then examined under a transmission electron microscope.

Western blot analysis of EV marker proteins

Freshly isolated EV samples were lysed in RIPA buffer supplemented with protease inhibitors and quantified using a BCA protein assay (Thermo Fisher Scientific, 23227). Equal amounts of protein (20ug) were loaded onto 10% SDS-PAGE gels and separated under denaturing conditions. Proteins were transferred to PVDF membranes, which were subsequently blocked with 5% non-fat milk in Tris-buffered saline with 0.1% Tween-20 (TBST) for 1 h at room temperature. Membranes were then incubated overnight at 4 °C with primary antibodies against CD81, CD63, CD9, TSG101, ALIX and Calnexin. Following primary incubation, membranes were washed in TBST and incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence detection system, with marker expression confirmed by the presence of corresponding bands. Antibody information is provided in table S2.

Flow cytometry

Cells were harvested at appropriate passages, washed twice with PBS, and resuspended in PBS supplemented with 2% fetal bovine serum (FBS, Thermo Fisher Scientific, A5256701). Aliquots containing 1 × 10⁵ cells per sample were incubated with fluorophore-conjugated monoclonal antibodies. Samples were incubated with antibodies for 30 min at 4 °C in the dark. After staining, cells were washed twice with flow cytometry buffer and resuspended in 500 µL of the same buffer. Unstained controls and isotype-matched controls were included to set gating parameters. Data acquisition was conducted on a flow cytometer, and at least 10,000 events were recorded for each sample. The presence of high levels of CD44, CD73, CD90, CD166, and CD105, along with the absence of CD34, HLA-DR, and CD45, was used to confirm the MSC identity and purity of the cell populations. Antibody information is provided in table S2.

CCK-8 proliferation assay

Human dermal fibroblasts (HDFs, IMMOCELL, IMP-H081) were seeded into 96-well plates at a density of 4,000 cells per well and allowed to adhere overnight in standard growth medium. Cells were then treated with 50 µg/mL (total protein) of freshly isolated MSC-EVs, iMSC-EVs, or acellular HPL-EVs. Serum-free medium served as a negative control. After 48 h of incubation, 10 µL of CCK-8 reagent (Beyotime, C0038) was added to each well and incubated for an additional 2 h at 37 °C. Absorbance was measured at 450 nm using a microplate reader. Results were normalized to the control group and analyzed using one-way ANOVA followed by Tukey’s post hoc test.

Scratch wound healing assay

Human dermal fibroblasts were seeded into 6-well plates at a density of 3 × 10⁵ cells per well and cultured in complete medium until reaching ~ 90% confluency. A uniform scratch was created using a sterile 200 µL pipette tip across the center of each well. Detached cells were gently removed by washing with PBS, and wells were subsequently incubated in serum-free medium containing 50 µg/mL total protein of either MSC-EVs, iMSC-EVs, or acellular HPL-EVs. Serum-free medium without EVs served as a negative control. Bright-field images were captured immediately after scratching (0 h) and after 48 h of treatment using a phase-contrast microscope. Wound closure was quantified by measuring the width of the scratch at three randomly selected regions per image using ImageJ software.

In vitro safety assessment of EVs using bone marrow-derived macrophages (BMDMs).

Bone marrow cells were isolated from the femurs and tibias of C57BL/6 mice (Charles River) and differentiated into macrophages by culturing in IMEM supplemented with 10% fetal bovine serum and 50 ng/mL GM-CSF (Thermo Fisher Scientific, PMC2011) for 7 days. To assess the pro-inflammatory effects of EVs, BMDMs were treated with iMSC-EV (5E8 particles/mL), MSC-EV (5E8 particles/mL), or lipopolysaccharide (LPS, 100 ng/mL, Sigma, L5293) for 4 h. Total RNA was extracted. The relative mRNA expression of TNFα, IL1β, and IL6 was quantified using qRT-PCR with GAPDH as the internal control. To evaluate the anti-inflammatory effects of EVs, BMDMs were pretreated with iMSC-EV or MSC-EV (5E8 particles/mL) for 2 days, followed by stimulation with LPS (100 ng/mL) for 4 h. The expression of TNFα, IL1β, and IL6 was similarly analyzed by qRT-PCR.

Pulmonary fibrosis mouse model

Animal housing

All mice were housed in a specific pathogen-free (SPF) facility under controlled environmental conditions (temperature 22 ± 2 °C, relative humidity 50 ± 10%, 12-hour light/dark cycle). Mice were housed in individually ventilated cages with a maximum of five animals per cage, provided with autoclaved bedding, nesting material, and environmental enrichment. The mice were acclimated for one week before the experiment and health status was monitored daily by trained personnel, and all procedures were conducted in accordance with institutional animal welfare guidelines.

Fibrosis induction

Ten-week-old male C57BL/6 mice were used to establish a bleomycin-induced pulmonary fibrosis model. Mice were randomly divided into 4 groups: Sham, DPBS (control), iMSC-EVs, MSC-EVs, with 6 mice in each group. Under 1.5% isoflurane (Lunan Better Phamacertical, 110302) anesthesia, mice received a single intratracheal instillation of bleomycin sulfate (2 mg/kg; MedChemExpress, HY-17565). Mice were randomly allocated to experimental groups, with the individual animal considered the experimental unit. Humane endpoints were established for animals exhibiting > 20% weight loss or severe respiratory distress. Post-procedure, mice were monitored daily for health status, body weight, and respiratory symptoms.

EV administration

To evaluate the therapeutic effects of EVs, iMSC- and MSC-derived EVs were administered via aerosol inhalation starting on Day 7 after bleomycin instillation. EVs were dosed at 20 µg/100 µL PBS and delivered every other day until Day 21 using a mesh nebulizer attached to a custom-built aerosol exposure chamber. Control animals received nebulized PBS under identical conditions. The delivery was standardized across groups to ensure equal exposure time. Investigators administering treatments were blinded to group allocation. Endpoint assessments included weight monitoring, histological analysis, and molecular evaluation of fibrosis and inflammation markers (n = 6).

Histological and biochemical analyses

On Day 21 post-bleomycin induction, Mice were sacrificed by cervical dislocation. lung tissues were harvested, fixed in 10% formalin, and embedded in paraffin (Leagene, DF0135). Five-micron-thick sections were cut and stained with hematoxylin and eosin (H&E) for general morphology, Masson’s Trichrome for collagen detection, and Sirius Red for enhanced visualization of collagen fibers. Fibrosis severity was quantified using the Ashcroft scoring system, which assesses fibrotic changes on a scale of 0 (normal) to 8 (severe fibrosis). Scoring was performed by two blinded, independent observers on five randomly selected fields per section, providing an unbiased evaluation of fibrosis extent across treatment groups. Lung injury and inflammation were evaluated by measuring total protein levels in bronchoalveolar lavage fluid (BALF), collected by instilling and aspirating 1 mL of sterile PBS into the trachea. BALF samples were centrifuged at 1,000 x g for 10 min at 4 °C, and the supernatant was analyzed for total protein concentration using a BCA assay. Elevated protein levels in BALF were considered indicative of increased alveolar-capillary permeability and lung injury.

In vivo safety assessment of EVs

Eight- to ten-week-old male C57BL/6 mice were used in an initial, gross-toxicity screen. Animals were randomly allocated to three groups: PBS (vehicle, n = 5), MSC-EV (n = 4), and iMSC-EV (n = 4). Investigators responsible for dosing and end-point analyses were blinded to group allocation. EVs (1 × 10¹¹ particles in 100 µL PBS) were delivered by tail-vein injection on Days 0, 2, 4, 6, 8, 10; controls received PBS only. Body weight and clinical appearance were recorded daily. On Day 13, mice were euthanised under isoflurane, and blood was collected for serum biochemistry (ALT, AST; Beckman Coulter AU480).

Major organs, including the heart, liver, spleen, lung, kidney, and testis, were harvested and fixed in 10% neutral buffered formalin (Sigma, R04586). Fixed tissues were embedded in paraffin, and stained with hematoxylin and eosin (H&E) for histological examination. Tissue sections were examined under a light microscope for evidence of inflammation, necrosis, or other pathological changes. Representative images were captured for comparison between control and EV-treated groups.

Statistical analysis

All statistical analyses were performed using GraphPad Prism (GraphPad Software). Data are presented as mean ± standard deviation (SD) u. Comparisons between two groups were conducted using an unpaired two-tailed Student`s t-test. For multiple group comparisons, one-way analysis of variance (ANOVA) followed by Tukey`s post hoc test was applied. A p-value of < 0.05 was considered statistically significant, with significance levels denoted as p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). All experiments were performed with at least three independent biological replicates (n ≥ 3) unless otherwise stated.

The work has been reported in line with the ARRIVE guidelines 2.0.

Results

Differentiation and phenotypic characterization of iMSCs

MSCs can be derived from trophoblast-like stem cells (TLCs) in a relatively short period, yielding cells that meet MSC criteria and demonstrate tri-lineage differentiation potential [22, 23]. Building on this approach, we established an innovative method for generating MSCs by first inducing extended pluripotent stem cells (EPSCs) from induced pluripotent stem cells (iPSCs). Following previously reported protocols [20, 21]we successfully derived EPSCs from iPSCs.

To characterize EPSCs, we first examined their morphology using phase-contrast microscopy, revealing distinct structural differences among iPSCs, H9 cells, and EPSCs. While iPSCs and H9 cells maintained tightly packed colony structures typical of pluripotent cells, EPSCs exhibited a more elevated, dome-like morphology (Figure S1A). To assess pluripotency characteristics, we performed immunofluorescence staining for the pluripotency marker OCT4 (Figure S1B). Both iPSCs and H9 cells displayed strong OCT4 expression, whereas EPSCs exhibited higher OCT4 levels. To further validate the differentiation potential of iPSCs, H9 cells, and EPSCs toward TLCs, we quantified the expression of key trophoblast markers, GATA3, KRT7, and HCGβ [24]via quantitative PCR (Figure S1C). Compared to iPSC- and H9-derived TLCs, EPSC-derived TLCs exhibited significantly higher expression levels, with fold increases exceeding 20-fold for GATA3, 40-fold for KRT7, and 100-fold for HCGβ (p < 0.001 for all markers). These findings indicate that EPSCs possess a strong predisposition for trophoblast differentiation, acquiring TLC characteristics more readily than iPSCs and H9 cells.

Having characterized EPSCs, we next induced their differentiation into MSCs. Phase-contrast microscopy tracked the morphological changes during this process, revealing a gradual transition from the elevated, dome-like EPSC colonies to a spindle-shaped, fibroblast-like morphology characteristic of MSCs (Fig. 1A). This transformation confirms the successful induction of an MSC phenotype. To verify MSC identity, we performed flow cytometry on differentiated iMSCs and compared them to primary MSC sources, including umbilical cord-derived MSCs (UC-MSCs) and adipose-derived MSCs (ADSCs) (Fig. 1B). All MSC populations exhibited high expression of canonical MSC surface markers, CD44, CD73, CD90, CD166, and CD105, while lacking the hematopoietic markers CD34 and CD45, confirming their mesenchymal origin. To assess the multipotency of MSCs from different sources, we induced iMSCs, UC-MSCs, and ADSCs to differentiate along adipogenic, osteogenic, and chondrogenic lineages, followed by lineage-specific staining (Fig. 1C). Oil Red O staining confirmed adipogenic differentiation through lipid droplet accumulation, Alcian Blue staining validated chondrogenic differentiation via glycosaminoglycan deposition, and Alizarin Red staining demonstrated osteogenic differentiation through calcium mineralization. All MSC types exhibited robust tri-lineage differentiation potential. While slight variations in staining intensity were noted, iMSCs displayed a differentiation capacity comparable to that of UC-MSCs and ADSCs, underscoring their viability as an alternative MSC source.

Fig. 1.

Generation, characterization, and differentiation potential of iMSC

(A) Schematic representation of the differentiation process from iPSCs to EPSCs, trophoblast-like stem cells (TLCs), and ultimately iMSCs. Representative phase-contrast images show distinct morphological transitions at each stage. Scale bar: 100 μm. (B) Flow cytometric analysis of MSC surface markers (CD44, CD73, CD90, CD166, and CD105) and negative markers (CD34, CD45, and HLA-DR) in iMSCs, UC-MSCs, and ADSCs. All MSC types exhibited high expression of positive markers and lacked hematopoietic markers, confirming their mesenchymal identity. (C) Tri-lineage differentiation potential of iMSCs, UC-MSCs, and ADSCs. All MSC types demonstrated robust differentiation potential, with comparable staining intensities. Scale bar: 50 μm. (D) UMAP-based clustering of ADSCs, iMSCs, and UC-MSCs, illustrating transcriptional heterogeneity across different MSC populations (up panel). Merged UMAP plot of ADSCs, iMSCs, and UC-MSCs, depicting the overall similarity in gene expression profiles (down panel)

To further confirm the functional and phenotypic consistency of iMSCs with primary MSCs, we conducted single-cell RNA sequencing analysis (Fig. 1D). This high-resolution approach allowed for a comprehensive transcriptional comparison between iMSCs, UC-MSCs, and ADSCs, enabling the evaluation of cellular heterogeneity, lineage-specific markers, and functional gene signatures across different MSC populations. Uniform Manifold Approximation and Projection (UMAP) clustering demonstrated that clustering of ADSCs, iMSCs, and UC-MSCs, with distinct subclusters labeled for different cell populations. Notably, iMSCs exhibit a transcriptional profile that is more closely aligned with UCMSCs, as evidenced by the overlapping cluster distributions (Fig. 1D, top). The bottom panel provides a merged UMAP plot, highlighting the degree of similarity between the three MSC populations (Fig. 1D, bottom). Expression analysis of key MSC markers (Figure S2), including CD44, THY1, and S100A4 [25, 26], confirmed that iMSCs retained canonical MSC identity. Additionally, genes associated with extracellular matrix organization and cell adhesion, such as fibronectin (FN1), collagen type I alpha 1 (COL1A1), and biglycan (BGN) [27]were consistently expressed across all three MSC types, indicating that iMSCs preserve the structural and functional characteristics of primary MSCs.

These findings further support the notion that iMSCs maintain a stable mesenchymal phenotype and functional gene expression profile, reinforcing their potential as a scalable and reproducible alternative to primary MSCs for regenerative medicine and EV-based therapeutics.

Comparative analysis of proliferation capacity and stability of iMSC, UC-MSC, and ADSC across passages

To further evaluate the phenotypic stability, genomic integrity, and proliferative capacity of iMSCs in comparison to UC-MSCs and ADSCs, we performed a series of analyses. Phase-contrast microscopy at passages P6, P8, and P12 revealed that all three MSC types retained their characteristic fibroblast-like, spindle-shaped morphology, with no visible signs of senescence or abnormal morphology at these stages (Fig. 2A). To assess phenotypic stability, we conducted flow cytometry analysis for key MSC surface markers at passages P6, P8, P12, and P16 (Fig. 2B and Figure S3). iMSCs consistently expressed high levels of MSC markers while maintaining minimal CD34 and CD45 expression across all passages up to P16, indicating sustained MSC identity. In contrast, UC-MSCs and ADSCs showed increasing CD45 expression at passage P16, with CD45 + cells exceeding 60% of the population. This suggests a loss of MSC purity and potential cellular aberrations in UC-MSC and ADSC cultures at later passages, highlighting the superior phenotypic stability of iMSCs during extended culture.

Fig. 2.

Characterization of iMSC, UC-MSC, and ADSC across multiple passages

(A) Phase-contrast microscopy images of iMSCs, UC-MSCs, and ADSCs at passages P6, P8, and P12. All three MSC types exhibit typical fibroblast-like, spindle-shaped morphology across passages, with no visible signs of senescence or morphological abnormalities. Scale bar: 50 μm. (B) Flow cytometry analysis of MSC surface markers (CD44, CD73, CD90, CD105, and CD166) and hematopoietic markers (CD34, CD45, and HLA-DR) at passages P6, P8, P12, and P16. (C) Representative karyotypes of iMSCs, UC-MSCs, and ADSCs were analyzed at passages P6, P8, P12, and P16 to assess genomic stability. (D) Cell proliferation was assessed using the CCK-8 assay at passages P4, P8, P12, and P16. All MSC types exhibited progressive increases in proliferation rates with passage number. Data are presented as mean ± SD, n = 3

To assess genomic stability during long-term expansion, we performed karyotype analysis on iMSCs, UC-MSCs, and ADSCs at passages P6, P8, P12, and P16 (Fig. 2C). iMSCs maintained a normal karyotype up to P12, suggesting a relatively high threshold for chromosomal stability. However, by P16, some karyotypic abnormalities emerged, indicating a potential limit to prolonged in vitro expansion. ADSCs exhibited chromosomal abnormalities as early as P6, suggesting a higher susceptibility to genomic instability. In contrast, UC-MSCs maintained a stable karyotype across all analyzed passages, reinforcing their genomic resilience during extended culture. To evaluate cell proliferation dynamics, a CCK-8 assay was conducted at P4, P8, P12, and P16 (Fig. 2D). All MSC types exhibited progressive increases in proliferation rates with passage number, confirming their continued proliferative capacity during in vitro expansion. While iMSCs, UC-MSCs, and ADSCs displayed comparable growth trends, subtle differences in proliferation rates emerged at later passages, suggesting potential variations in long-term expansion efficiency. These findings indicate that iMSCs maintain chromosomal stability over extended passages more effectively than ADSCs but may develop abnormalities at later passages, whereas UC-MSCs exhibit superior genomic integrity throughout expansion.

Together, these findings suggest that iMSCs not only preserve a stable MSC phenotype and karyotype at higher passages but also maintain proliferation capacity while exhibiting greater genomic and phenotypic stability.

Optimized scalable expansion of iMSCs in bioreactors using degradable microcarriers

To harness the high proliferative capacity of iMSCs for large-scale applications, we optimized iMSC expansion using degradable microcarriers in three different bioreactor setups: a 125 mL bioreactor, a 500 mL bioreactor, and a 5 L bioreactor, progressing through increasing culture volumes (Fig. 3A). Calcein AM and propidium iodide (PI) staining were used to assess cell viability and proliferation at different time points (Fig. 3B). Calcein AM staining (green) indicates high cell viability across all bioreactor conditions, while minimal PI staining (red) suggests low cell death. From Day 1 to Day 7 (Day 6 for the 500 mL bioreactor), the increasing density of green-fluorescent cells confirms active proliferation, with bright-field images showing the formation of dense 3D cell aggregates.

Fig. 3.

Scalable expansion and characterization of iMSCs in bioreactors

(A) Schematic representation of the stepwise expansion of iMSCs using degradable microcarriers in scalable bioreactor systems. iMSCs were first expanded in a 125 mL bioreactor, followed by sequential scale-up in 500 mL and 5 L bioreactors, with final cell harvesting and cryopreservation. (B) Cell viability assessment using Calcein AM and propidium iodide staining in 125 mL, 500 mL, and 5 L bioreactors at different time points (Day 1 and Day 7 for 125 mL and 5 L; Day 1 and Day 6 for 500 mL). Live cells fluoresce green (Calcein AM), while dead cells fluoresce red (PI). Bright-field (BF) images show the formation of dense 3D cell aggregates over time. Scale bar: 100 μm. (C) Growth curve tracking total iMSC numbers over time in 125 mL, 500 mL, and 5 L bioreactors. iMSCs exhibit an initial adaptation phase, followed by exponential expansion onward, with the 5 L bioreactor achieving the highest cell yield by day 20. Data are presented as mean ± SD, n = 3. (D) Flow cytometry analysis of MSC-specific markers and hematopoietic markers in iMSCs expanded in 125 mL, 500 mL, and 5 L bioreactors. Cells maintain high MSC marker expression and low hematopoietic marker expression across all conditions, confirming the retention of MSC identity after bioreactor expansion

To further characterize iMSC growth kinetics, we monitored total cell number over time (Fig. 3C). Minimal growth was observed during the initial three days, likely due to an adaptation phase as cells attached to the microcarriers and acclimated to the bioreactor environment. From day 4 onward, rapid cell proliferation was observed, with high confluency reached by day 7 in the 125 mL and 5 L bioreactors and by day 6 in the 500 mL bioreactor, at which point cells were harvested. These results demonstrate that iMSCs are well-suited for scalable expansion in bioreactor systems, with degradable microcarriers providing effective support for cell attachment, proliferation, and viability. To assess metabolic activity during iMSC expansion, we monitored lactate accumulation and glucose consumption across different bioreactor conditions (Figure S4). Lactate concentration (LAC) increased steadily from Day 1 to Day 7, correlating with the cell proliferation pattern (Figure S4A). Glucose concentration (GLU) progressively declined in all bioreactor conditions, with the most pronounced depletion occurring in the 500 mL bioreactor, indicating higher cellular metabolic demand (Figure S4B). Glucose consumption rates increased significantly from Day 3 onward, correlating with the onset of rapid cell proliferation. Notably, the 5 L bioreactor exhibited a more gradual decrease in glucose levels and a progressive increase in glucose consumption, likely due to its larger medium volume buffering metabolic fluctuations.

To verify that iMSCs retained their MSC identity after bioreactor expansion, we conducted flow cytometry analysis to assess the expression of key MSC markers and hematopoietic markers at the end of expansion in each bioreactor system (Fig. 3D). iMSCs maintained high expression of MSC markers and low expression of hematopoietic markers across all bioreactor conditions, confirming that bioreactor-expanded iMSCs retained their mesenchymal phenotype. These findings demonstrate that iMSCs can be efficiently expanded in scalable bioreactor systems using degradable microcarriers while preserving their MSC characteristics. The observed growth kinetics and phenotypic stability suggest that iMSCs are highly suitable for large-scale manufacturing, making them a promising candidate for clinical and regenerative medicine applications.

Optimized large-scale production of MSC- and iMSC-derived EVs using a fixed-bed bioreactor

A fixed-bed bioreactor system with a total culture surface area of 2 m² was established to enable the efficient production of EVs derived from MSCs and iMSCs. The workflow integrates cryopreserved working cell banks, cell seeding onto a fixed-bed bioreactor, and continuous collection of EV-rich supernatant, which is subsequently purified using a tangential flow filtration system (Fig. 4A). To assess cell proliferation within the fixed-bed bioreactor, samples were collected from the bioreactor membrane at days 7 and 14 post-seeding, and cell growth was analyzed via crystal violet staining (Fig. 4B). No significant differences in cell attachment or proliferation were observed between the two groups, indicating similar adaptation and expansion within the fixed-bed environment. Cell viability was assessed at days 7 and 14 using Calcein AM and propidium iodide (PI) staining (Fig. 4C). Strong Calcein AM fluorescence (green) and minimal PI staining (red) confirmed that both MSCs and iMSCs maintained high viability throughout the culture period. Besides, no excessive cell aggregation was detected during fixed-bed culture. crystal violet staining and live/dead imaging (Fig. 4B and C) showed an even cellular distribution across the matrix and uniformly high viability in both MSC and iMSC runs. These results highlight the suitability of the fixed-bed bioreactor for maintaining a homogeneous, viable cell population, which is critical for GMP-grade, reproducible EV manufacture.

Fig. 4.

Large-scale production and characterization of MSC- and iMSC-derived EVs using a fixed-bed bioreactor system

(A) Schematic representation of the fixed-bed bioreactor workflow for MSC and iMSC expansion and EV production. The process involves cell thawing and seeding, expansion in a fixed-bed bioreactor, continuous EV collection, and purification to obtain high-purity EVs. (B) Crystal violet staining of MSCs and iMSCs cultured within the fixed-bed bioreactor at Day 7 and Day 14, showing progressive cell proliferation and matrix colonization over time. (C) Live/dead staining of MSCs and iMSCs at Day 7 and Day 14 using Calcein AM (green, live cells) and propidium iodide (PI, red, dead cells). Both cell types exhibit high viability with minimal cell death, indicating a stable culture environment in the fixed-bed bioreactor. Scale bar: 100 μm. (D) Flow cytometry analysis of MSC-specific markers (CD44, CD73, CD90, CD105, CD166) and hematopoietic markers (CD34, CD45) in MSCs and iMSCs at Day 7 and Day 14. (E) Daily harvest (pre-CONC) and post concentrate (post-CONC) volumes from the fixed-bed bioreactor for MSCs and iMSCs. Data are mean ± SD. (F) Total EV output per day from MSCs and iMSCs, quantified by nanoparticle tracking analysis. Data are presented as mean ± SD

Given the extended culture period, flow cytometry was performed to evaluate MSC surface marker expression at days 7 and 14 post-seeding (Fig. 4D). Both MSCs and iMSCs consistently expressed high levels of MSC-specific markers while showing low expression of hematopoietic markers, confirming the retention of a stable MSC phenotype throughout expansion. This ensures a homogeneous and reliable cell population for EV production. To assess EV production capacity, we measured the daily volume of collected supernatant and purified EV yield from the bioreactor system (Fig. 4E). Approximately 600–700 mL of conditioned medium was harvested daily from fixed-bed cultures of both MSCs and iMSCs, with no significant difference between groups. After tangential-flow filtration, this volume was concentrated to 50–60 mL, yielding comparable EV suspensions. Nanoparticle tracking showed daily outputs of about 1.5 × 10¹³ particles for MSCs and 1.2 × 10¹³ particles for iMSCs (Fig. 4F), confirming that the two cell sources possess equivalent EV-production capacity.

To ensure the quality and consistency of EVs produced from MSC and iMSC using a fixed-bed bioreactor, we performed comprehensive characterization using multiple analytical techniques. NanoCoulter counter equipped with nanopore chips was conducted to determine the size distribution of MSC-EVs and iMSC-EVs (Figure S5A). Both EV types exhibited a similar size range, with an average diameter of approximately 70–80 nm, consistent with the typical EV profile. To further examine EV morphology, we employed transmission electron microscopy (TEM) (Figure S5B). TEM images revealed that both MSC-EVs and iMSC-EVs displayed the characteristic cup-shaped morphology, confirming the structural integrity of the isolated EVs. To verify EV identity and purity, we conducted Western blot analysis for key EV protein markers (Figure S5C). Both MSC-EVs and iMSC-EVs consistently expressed CD81, CD63, CD9, TSG101, and ALIX, hallmark proteins associated with EVs [28, 29]. Calnexin, an endoplasmic reticulum marker and negative control for EVs, was absent, confirming that the EV preparations were free from cellular contamination.

To assess the short-term storage stability of EVs produced from MSCs and iMSCs, we measured particle concentrations at baseline (T0), 1 month, and 2 months post-storage at -80 °C using a NTA counter (Figure S5D). To ensure comparability across experimental groups, EV preparations from MSCs and iMSCs, were first quantified by NTA, and their particle concentrations were adjusted to similar levels at baseline (T0) prior to downstream analysis. The results showed that both MSC-EVs and iMSC-EVs maintained stable particle concentrations over the 2-month period, with only minimal variation. To further validate the specificity of the observed functional effects, we examined whether residual vesicles present in the human platelet lysate (HPL)-supplemented medium could account for EV activity. Acellular HPL-derived EVs (designated “Medium”), isolated from unconditioned medium subjected to the same isolation procedure, were compared side-by-side with MSC-EVs and iMSC-EVs. In functional assays, acellular HPL-EVs failed to promote wound closure in scratch assays (Figure S5E and S5F) or enhance fibroblast proliferation (Figure S5G), in contrast to the pro-repair effects and pro-proliferative of MSC-EVs and iMSC-EVs. As shown in Figure S5H, the Medium group also contained significantly fewer particles, less than 10% of the concentration measured in either MSC- or iMSC-conditioned media, despite undergoing the same isolation procedure. These results confirm that the therapeutic effects observed in our study are attributable to actively secreted EVs from MSCs and iMSCs, rather than residual vesicles originating from the culture supplement.

Safety profile of iMSC-EVs and MSC-EVs

To evaluate the safety profile and immunomodulatory effects of iMSC-EVs and MSC-EVs, a series of in vitro and in vivo experiments were conducted. The pro-inflammatory potential of iMSC-EVs and MSC-EVs was assessed by measuring the expression of pro-inflammatory cytokines in bone marrow-derived macrophages (BMDMs). Macrophages were treated with EVs or lipopolysaccharide (LPS, positive control) for 4 h, followed by mRNA extraction and qRT-PCR analysis. As shown in Fig. 5A, LPS treatment significantly upregulated TNFα, IL1β, and IL6 expression, confirming its strong pro-inflammatory effects. In contrast, neither iMSC-EVs nor MSC-EVs induced significant cytokine expression, suggesting that both EV types lack pro-inflammatory activity.

Fig. 5.

Evaluation of the immunomodulatory effects and systemic safety of iMSC-EVs and MSC-EVs

(A) Pro-inflammatory potential of iMSC-EVs and MSC-EVs in bone marrow-derived macrophages (BMDMs). Macrophages were treated with iMSC-EVs, MSC-EVs, or LPS (positive control) for 4 h, followed by qRT-PCR analysis of TNFα, IL1β, and IL6 mRNA expression. Data are presented as mean ± SD (p < 0.001 vs. control). (B) Anti-inflammatory effects of iMSC-EVs and MSC-EVs in BMDMs. Macrophages were pretreated with iMSC-EVs or MSC-EVs for 2 days before LPS stimulation, followed by qRT-PCR analysis of TNFα, IL1β, and IL6 mRNA expression. Data are presented as mean ± SD. (C) Schematic representation of the in vivo EV safety study. Mice received intravenous injections of iMSC-EVs or MSC-EVs (10¹¹ particles/dose) on Days 0, 2, 4, 6, 8, and 10, and were sacrificed on Day 13 for safety assessments. (D) Body weight monitoring of EV-treated mice. No significant differences were observed between control and EV-treated groups throughout the study, indicating no systemic toxicity. Data are presented as mean ± SD. (E, F) Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels. Liver function markers remained within normal ranges across all groups, suggesting no hepatotoxic effects from EV administration. Data are presented as mean ± SD. (G) Histopathological analysis of major organs. Hematoxylin and eosin (H&E) staining of the heart, liver, spleen, lung, kidney, and testis revealed no significant abnormalities, structural damage, or inflammatory infiltration in EV-treated mice compared to the control group. Scale bar: 500 μm

To evaluate the anti-inflammatory potential of EVs, BMDMs were pretreated with iMSC-EVs or MSC-EVs for 2 days, followed by LPS challenge for 4 h. mRNA expression levels of TNFα, IL1β, and IL6 were quantified using qRT-PCR (Fig. 5B). MSC-EVs significantly reduced LPS-induced TNFα, IL1β, and IL6 expression compared to the LPS-only group. iMSC-EVs also mitigated LPS-induced inflammation, but their effects differed slightly across cytokines. Notably, iMSC-EVs exhibited a slightly stronger effect on IL6 suppression. Despite these differences, both EV types demonstrated anti-inflammatory properties, supporting their therapeutic potential for inflammatory and immune-related disorders.

The systemic safety of iMSC-EVs and MSC-EVs was further assessed in vivo. Mice received intravenous injections of EVs (10¹¹ particles/dose) on Days 0, 2, 4, 6, 8, and 10, and were sacrificed on Day 13 (Fig. 5C). Body weight was monitored throughout the study, with no significant differences observed between EV-treated and control groups, indicating no systemic toxicity (Fig. 5D). Liver function was assessed using serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels, which remained within normal ranges in all groups, suggesting no hepatotoxicity (Fig. 5E, F). To further assess tissue integrity, histopathological analysis of major organs (heart, liver, spleen, lung, kidney, and testis) was performed (Fig. 5G). No significant abnormalities, structural damage, or inflammatory infiltration were observed in EV-treated mice compared to controls, confirming the absence of EV-induced toxicity.

Collectively, these results demonstrate that iMSC-EVs and MSC-EVs are safe, as they do not trigger inflammatory responses in vitro or cause adverse effects in vivo. Furthermore, both EV types exhibit anti-inflammatory properties. These findings highlight the potential of EVs as therapeutic agents for immune-related and inflammatory diseases.

Therapeutic efficacy of iMSC-EVs and MSC-EVs in pulmonary fibrosis model

To assess the therapeutic potential of iMSC-EVs and MSC- EVs in treating pulmonary fibrosis, we administered EVs via intratracheal instillation to a bleomycin (BLM) induced lung fibrosis mouse model (Fig. 6A). Mice received treatments every two days from Day 7 to Day 21 post-bleomycin induction. Body weight monitoring revealed that mice treated with iMSC-EVs or MSC-EVs exhibited a significant recovery in weight compared to the BLM + PBS group, indicating an overall improvement in health and reduced disease severity (Fig. 6B). Both EV-treated groups maintained significantly higher body weights than the untreated fibrosis group, suggesting that iMSC-EVs and MSC- EVs effectively mitigate fibrosis-induced weight loss.

Fig. 6.

Therapeutic effects of iMSC-EVs and MSC-EVs in a bleomycin-induced pulmonary fibrosis mouse model

(A) Schematic representation of the experimental design. Pulmonary fibrosis was induced in mice by intratracheal instillation of bleomycin (5 mg/kg) on Day 0. Mice were treated every two days from Day 7 to Day 21 with either iMSC-EVs (20 µg/100 µL), MSC-EVs (20 µg/100 µL), or DPBS (control) before being sacrificed for analysis. (B) Body weight changes over time. Mice treated with iMSC-EVs or MSC-EVs exhibited significantly improved weight recovery compared to the BLM + PBS group, indicating a reduction in disease severity. Data are presented as mean ± SD. (C) Hematoxylin and eosin (H&E) staining of lung tissue. Representative images show severe alveolar structure disruption and fibrosis in the BLM + PBS group, whereas both iMSC-EVs and MSC-EVs treatments preserved lung architecture and reduced fibrotic lesions. Scale bar: 2.5 mm. (D) Masson’s trichrome staining for collagen deposition. The BLM + PBS group exhibited extensive collagen accumulation (blue staining), whereas both EV-treated groups showed reduced collagen deposition, suggesting attenuation of fibrosis. Scale bar: 2.5 mm. (F) Quantification of fibrosis severity using the Ashcroft score. Both iMSC-EVs and U iMSC-EVs treatments significantly decreased fibrosis scores compared to the BLM + PBS group, with no significant difference between the two EV-treated groups, indicating comparable therapeutic efficacy. Data are presented as mean ± SD. (G) Total protein levels in bronchoalveolar lavage fluid (BALF). EV-treated mice exhibited significantly lower BALF protein levels compared to the BLM + PBS group, suggesting reduced alveolar-capillary barrier damage and inflammation. Data are presented as mean ± SD

Histological analysis of lung tissue using hematoxylin and eosin (H&E) staining demonstrated that both iMSC-EVs and MSC-EVs treated mice exhibited reduced fibrotic lesions and improved alveolar structure compared to the BLM + PBS group (Fig. 6C). Masson’s trichrome staining further confirmed a reduction in collagen deposition in the lungs of EV-treated mice, as indicated by less intense blue staining, suggesting attenuation of excessive extracellular matrix accumulation (Fig. 6D). Sirius Red staining, which selectively marks collagen fibers, also showed a marked reduction in fibrosis and tissue remodeling in the EV-treated groups compared to the PBS-treated fibrosis group (Fig. 6E). These histopathological findings indicate that both iMSC-EVs and MSC-EVs effectively attenuate fibrosis and restore lung architecture.

To quantify fibrosis severity, the Ashcroft score was calculated from lung tissue sections. Both iMSC-EVs and MSC-EVs treatments significantly reduced Ashcroft scores compared to the BLM + PBS group, with no significant difference between the two EV-treated groups, indicating comparable therapeutic efficacy (Fig. 6F). Additionally, total protein levels in bronchoalveolar lavage fluid (BALF), a marker of lung injury [30]were significantly lower in the EV-treated groups compared to the BLM + PBS group, suggesting reduced alveolar-capillary barrier disruption and inflammation (Fig. 6G).

These findings demonstrate that both iMSC-EVs and MSC-EVs exhibit comparable therapeutic efficacy in alleviating bleomycin-induced pulmonary fibrosis in mice. The consistent reductions in fibrosis markers, lung tissue remodeling, and BALF protein levels confirm that both iMSC-EVs and MSC-EVs are effective therapeutic agents for fibrotic lung diseases.

Discussion

In this study, we developed a scalable platform for iMSC expansion and EV production using a fixed-bed bioreactor system. iMSCs were generated from EPSCs, offering a stable, reproducible, and renewable cell source. A fixed-bed bioreactor was employed to enable high-density iMSC expansion and continuous EV collection, ensuring batch-to-batch consistency. EVs were isolated using tangential flow filtration and size exclusion chromatography to achieve high purity while preserving EV functionality. Comprehensive characterization of iMSC-EVs was performed, including size distribution, morphology, protein markers, and therapeutic efficacy. A bleomycin-induced pulmonary fibrosis mouse model was used to evaluate the anti-fibrotic effects of iMSC-EVs. Both iMSC-EVs and MSC-EVs significantly reduced fibrosis, improved lung architecture, and modulated key inflammatory and fibrotic markers, demonstrating equivalent therapeutic efficacy.

Due to the unlimited proliferative capacity of EPSCs and iPSCs [31]we can mass-produce standardized iMSC populations with high consistency and reproducibility, overcoming the donor variability and limited expansion potential associated with primary MSCs. This scalability is particularly advantageous for clinical applications, where large-scale, quality-controlled cell sources are required for EV-based therapeutics. Furthermore, the genetic plasticity of EPSCs and iPSCs allows for precise gene editing using technologies such as CRISPR-Cas9, enabling the development of engineered cell lines with enhanced therapeutic properties [32, 33]. By introducing specific genetic modifications, we can generate engineered iMSCs that produce customized EVs with tailored cargo compositions, such as enriched anti-inflammatory miRNAs, optimized regenerative factors, or surface-modified targeting molecules [34].

The intrinsic scalability and genetic uniformity of iMSC cultures translate into clear operational advantages for therapeutic EV manufacturing. Whereas primary MSC sources, most commonly bone-marrow- or adipose-derived, exhibit donor-to-donor variation and phenotypic drift during extended expansion [35, 36]. This heterogeneity limits the consistency and therapeutic efficacy of MSC-EVs. In contrast, iMSCs originate from a single iPSC clone and retain a stable surface-marker profile. This homogeneity supports reproducible, large-batch EV production and is particularly valuable for chronic indications that demand frequent or high-dose EV administration. Using enzyme-dissolvable microcarriers, we scale iMSCs from 125mL to 5 L in 6–7 days with 95% viability, stable metabolism and phenotype. Harvest is single-step and expected to exceed 90% recovery, markedly higher than the 65% achieved by López-Fernández et al. with serum-based, polystyrene microcarriers that require detachment plus filtration [37]. Although iPSC products demand extra pluripotency testing, they remove donor variability, enable clonal banking and streamline downstream processing. Thus, our iMSC platform promises higher yields for industrial EV manufacture.

This study further demonstrates that both MSCs and iMSCs can be efficiently expanded in a fixed-bed bioreactor system, maintaining high viability, a stable MSC phenotype, and robust EVs production. The bioreactor system supports extended culture durations and continuous EVs collection, highlighting its potential for large-scale, clinical-grade EVs manufacturing. Nevertheless, industrial deployment of fixed beds remains limited, in part because the packed matrix imposes slower heat transfer, longer fluid-residence times and, at large scale, the risk of axial temperature or nutrient gradients; it also requires comparatively high medium volumes, which can erode cost-efficiency in continuous campaigns [38]. To mitigate these drawbacks, we incorporated an external recirculation loop with inline heat-exchange, used low-shear perfusion rather than true batch operation, and equipped the cartridge with distributed temperature probes to verify uniformity. Thus, while the inherent constraints of the platform must be acknowledged, targeted engineering controls can offset many of them, allowing fixed-bed bioreactors to offer a robust and scalable alternative for clinical-grade EV manufacturing.

For clinical applications, EV doses typically range from 1 × 10¹¹ to 1 × 10¹² particles per treatment, depending on the disease indication and delivery method [39–41]. Given the daily production capacity of 1.2 × 10¹³ particles, a single 2m² fixed-bed bioreactor can generate sufficient EVs to meet the needs of 12–120 clinical doses per day. Over a standard 14-day production cycle, this output corresponds to 280–2,800 clinical doses, demonstrating the feasibility of fixed-bed bioreactors in supporting large-scale EV manufacturing. The scalability, reproducibility, and automation of this system make it highly suitable for addressing the growing demand for EV-based therapies in clinical trials and beyond. By leveraging the robust proliferative capacity and consistent phenotype of iMSCs, this platform ensures a stable and homogenous cell source for EV production. Moreover, its ability to generate large quantities of EVs in a controlled and automated manner lays the foundation for the widespread clinical translation of EV-based therapies. These findings highlight the promise of fixed-bed bioreactor systems in supporting the large-scale, standardized production of therapeutic EVs.

While this study provides promising insights, several limitations should be acknowledged. First, further validation in humanized models or clinical settings is required to confirm the safety, efficacy, and optimal dosing of iMSC-EVs in patients. Additionally, our study focused on a single murine model of pulmonary fibrosis. Expanding this research to other fibrotic or inflammatory conditions could provide a broader understanding of iMSC-EV efficacy across multiple diseases. Moreover, our systemic safety assessment used a limited sample size, and all animals were male. This restricts the statistical power and generalizability of our findings. In future preclinical investigations, we will implement adequately powered, sex-balanced study designs to ensure robust evaluation of efficacy and safety. Addressing these questions will be critical for advancing iMSC-EVs toward clinical application and ensuring their long-term success as a regenerative therapy.

Conclusions

This study establishes iMSC-EVs as a scalable, consistent, and effective therapeutic tool for pulmonary fibrosis and potentially other inflammatory and fibrotic diseases. By addressing key challenges in EV production, including scalability and functional consistency, iMSC-EVs represent a viable alternative to primary MSC-EVs. The integration of fixed-bed bioreactors for large-scale manufacturing further enhances their clinical feasibility, paving the way for the next generation of cell-free regenerative therapies.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

iPSC

Induced pluripotent stem cell

EPSC

Extended pluripotent stem cell

TLC

Trophoblast-like stem cells

MSC

Mesenchymal stem cell

iMSC

Induced mesenchymal stem cell

UC-MSC

Umbilical cord-derived mesenchymal stem cell

ADSC

Adipose-derived mesenchymal stem cell

EVs

Extracellular vesicles

iMSC-EV

Extracellular vesicle derived from iMSC

MSC-EV

Extracellular vesicle derived from MSC

LPS

Lipopolysaccharide

ALT

Alanine aminotransferase

AST

Aspartate aminotransferase

BALF

Bronchoalveolar lavage fluid

Author contributions

L.W., Y.G. and S.G. supervised the research. L.W. and S.G. conceived of the experimental design. S.G., N.L., and Q.P. performed the differentiation experiments. S.G., N.L., and Q.P. performed all the in vitro molecular experiment and data analysis. S.G., F.W, and L.H. performed all the large-scale donor cell culture and EVs related experiment and data analysis. R.D., B.Z., and Y.Z. performed all the animal experiment and data analysis. S.G., J.W., and Z.N. performed all the single-cell transcriptome sequencing and data analysis. L.W. wrote the manuscript. S.T. and Y.G. edited and reviewed the manuscript. All authors read and approved the final manuscript.

Funding

Y.G. is supported by the National Natural Science Foundation of China (Grant No. 81170236 and 31127001). L.W. is supported by the Beijing Nova Program (20230484348).

Data availability

Single-cell RNA-seq raw data has been deposited at the Genome Sequence Archive of the National Genomics Data Center. Accession number is GSA for Human: HRA010880.

Declarations

Ethics approval and consent to participate

All experimental procedures involving iPSCs were approved by the Ethics Committee of Beijing Luhe Hospital. The research project titled “Peripheral blood collection from healthy people for iPSC cell bank construction and clinical use” was approved on July 24, 2023 (2021-LHYW-043-04). Donors provided written informed consent for their participation in the study and the use of their tissue for research purposes. Human dermal fibroblasts were purchased from a commercial vendor, IMMOCELL. The company has confirmed that there was initial ethical approval for the collection of human materials and the derivation of cell lines and that the donors had signed informed consent. See for details: http://www.immocell.com/?rydxb-13/7508.html.

All experimental procedures involving MSC cells approved by the Ethics Committee of Tsinghua Changgung Hospital. The research project titled “Clinical study on the safety, tolerability and preliminary efficacy of mesenchymal stem cell derived exosomes in the treatment of idiopathic pulmonary fibrosis” was approved on July 31, 2023 (No.23346-4-02). Donors provided written informed consent for their participation in the study and the use of their tissue for research purposes.

All animal procedures were performed in accordance with relevant laws and regulations governing the use and care of experimental animals and approved by the Institutional Animal Care and Use Committee (IACUC) of Allife Medicine (No. IAUCC-2024-160, Title: Safety and efficacy of mesenchymal stem cell-derived exosomes in the treatment of pulmonary fibrosis in mice; Date of approval: Feb 1, 2024).

Consent for publication

All authors confirm their consent for publication.

Competing interests

This study was partially supported by Allife Medicne, Inc. Other authors declare no conflict of interest.