Abstract

   Sepsis is a critical condition with high mortality, often leading to
   acute lung injury (ALI) due to uncontrolled inflammatory responses and
   alveolar epithelial damage. Extracellular vesicles (EVs), particularly
   mesenchymal stem cell‐derived EVs, have shown therapeutic potential in
   sepsis‐related organ dysfunction by transferring RNAs and proteins.
   However, their clinical use is limited by low efficacy and yield. To
   address this, 3D‐cultured MSCs (3D‐MSCs) are generated using
   MicroTissues 3D Petri Dish. These 3D‐MSCs demonstrate improved
   protection and proliferation of MLE‐12 cells in vitro. Mechanistic
   studies are conducted to explore the enhanced protective effects of
   3D‐MSCs derived EVs (3D‐EVs) in a septic‐ALI model. Proteomic and
   molecular analyses of 3D‐EVs revealed that they are enriched in
   hepatocyte growth factor (HGF). HGF helps maintain the barrier function
   of damaged alveolar epithelium through the PI3K‐AKT signaling pathway.
   Overall, 3D‐EVs effectively ameliorate sepsis‐induced ALI and enhance
   prognosis by enriching and delivering HGF, suggesting that their
   application represents a promising treatment strategy for septic ALI.

   Keywords: acute lung injury, extracellular vesicles, mesenchymal stem
   cell, sepsis, 3D culture
     __________________________________________________________________

   3D culture of MSCs upregulates gene expression of matrix
   metalloproteinases (MMPs), growth factors (GFs), and increases the
   production of extracellular vesicles (EVs). Proteomic analysis reveals
   642 upregulated proteins in 3D‐EVs, including HGF. These 3D‐EVs,
   enriched with HGF, protect lung epithelial cells from apoptosis and
   barrier damage in septic ALI via the PI3K/AKT pathway. Figure created
   with Biorender.com.

   graphic file with name ADVS-12-2500637-g008.jpg

1. Introduction

   Sepsis is a common and severe disease with high mortality in critically
   ill patients, imposing a heavy economic burden.^[ [42]^1 ^] Over the
   years, the incidence has increased gradually due to population aging.
   Among the multiple organ dysfunctions that sepsis can cause, the lung
   is often the first target of damage because of its susceptibility,
   developing into acute lung injury (ALI) and acute respiratory distress
   syndrome (ARDS). The pathogenesis of ALI/ARDS is characterized by
   diffuse alveolar damage, which encompasses injury to pulmonary
   capillary endothelial cells and alveolar epithelial cells, along with
   interstitial and alveolar edema.^[ [43]^2 ^] Despite the growing
   concern surrounding sepsis‐induced lung injury, no effective
   pharmacological therapy has been identified.

   Given the lack of effective pharmacological therapies for
   sepsis‐induced lung injury, mesenchymal stromal cells (MSCs) have
   emerged as a promising therapeutic option. Bone marrow MSCs (BMSCs)
   represent a population of multipotent stem cells endowed with
   self‐renewal capabilities.^[ [44]^3 ^] The administration of MSCs has
   been shown to mitigate alveolar‐capillary injury and reduce mortality
   in pre‐clinical models of ALI, owing to their multifaceted therapeutic
   properties, including anti‐inflammatory, immunomodulatory,
   regenerative, proangiogenic, and antimicrobial effects.^[ [45]^4 ^] A
   growing body of evidence suggests that the primary mode of MSC action
   is mediated through their secretome, particularly extracellular
   vesicles (EVs).^[ [46]^5 ^] EVs secreted by MSCs (MSC‐EVs) mirror the
   cells' inherent activity by delivering functional biomolecules, such as
   lipids, cytosolic proteins, and biologically active RNAs, to sites of
   tissue damage.^[ [47]^6 ^] Experimental evidence has demonstrated that
   MSCs can transfer EVs to target cells, effectively promoting the
   recovery of epithelial and endothelial tissues and facilitating
   microbial clearance and alveolar fluid reabsorption.^[ [48]^7 ^] Our
   previous research, along with that of others, has positioned MSC‐EVs as
   a promising avenue for cell‐free therapy in ALI.^[ [49]^8 ^] Studies
   have also shown that the paracrine properties of MSCs can be modulated
   through various preconditioning methods and techniques.^[ [50]^9 ^] For
   instance, the shift from a monolayer to a nonadhesive, 3D culture
   environment has been found to significantly alter the composition of
   the secretome, potentially augmenting its therapeutic efficacy.^[
   [51]^10 ^] Dynamic 3D culture of MSC‐EVs has been shown to rejuvenate
   senescent stem cells and enhance their immunomodulatory potential.^[
   [52]^11 ^] Moreover, exosomes produced from 3D‐cultured MSCs exhibit
   increased yield, albeit to varying degrees.^[ [53]^12 ^] Collectively,
   these findings underscore the potential of 3D cell cultures to overcome
   the limitations hindering the broad application of EVs.

   In this study, we harnessed a non‐adhesive 3D aggregation culture
   system to cultivate MSCs and evaluated the therapeutic efficacy of
   3D‐MSCs on injured lung epithelial cells in vitro. To elucidate the
   underlying mechanisms, we conducted RNA‐sequencing analysis, offering a
   comprehensive view of transcriptome alterations following 3D culture
   generation, with a particular emphasis on cellular remodeling and EVs
   generation pathways. Concurrently, it has been established that the
   paracrine factors derived from MSC spheroids can enhance tissue repair
   and immunomodulatory effects.^[ [54]^13 ^] Consequently, we purified
   and characterized the 3D‐cultured MSC‐derived EVs in their entirety.
   Subsequently, we evaluated the roles of 2D‐EVs and 3D‐EVs in preserving
   alveolar epithelial barrier integrity using a sepsis‐induced ALI mouse
   model. Our proteomic analyses revealed that the therapeutic mechanisms
   of these EVs involve, at least in part, the transfer of enriched
   hepatocyte growth factor (HGF). Therefore, our data provide evidence
   that 3D‐EVs may exert a significant impact on the reparative processes
   of lung epithelial tissue in sepsis.

2. Experimental Section

2.1. Culture of MSCs from Bone Marrow

   The BMSCs were obtained from Cyagen Biosciences (Guangzhou, China) and
   cultured in DMEM/F12 (Gibco) containing 10% fetal bovine serum (Gibco)
   and 1% penicillin/streptomycin (Gibco). The cells were placed in 100 mm
   cell culture dishes (Corning, NY, USA) and maintained at a temperature
   of 37 °C with 5% CO2 and 90% humidity. The growth medium was replaced
   every three days, and cell passage was conducted once the cells reached
   90% confluence. After removing the culture medium, the cells that
   adhered to the surface were rinsed two times using phosphate‐buffered
   saline (PBS). Subsequently, these cells were collected by exposing them
   to 1 mL of 0.25% trypsin and 1 mm ethylenediaminetetraacetic acid
   solution for a duration of 2 min at a temperature of 37 °C. This was
   followed by adding 5 mL of complete medium to counteract the effects of
   trypsin. Ultimately, the cells were reintroduced into a fresh medium at
   a ratio of 1:3 dilution. For all subsequent experiments, the
   utilization of BMSCs was limited to passages 3–5. The confirmation of
   BMSCs' identity involved flow cytometric detection of common MSC
   markers. This entailed the use of FITC‐conjugated monoclonal antibody
   against CD45, CD49e, and CD90, PE‐conjugated monoclonal antibody
   against CD29, along with their respective isotype controls
   (eBioscience, San Diego, CA, USA).

2.2. Preparation of Agarose Micro Multi‐Well Dishes and 3D MSCs Spheroids

   81‐wells rubber micro‐molds (Figure [55]1A) with a diameter of 400 µm
   were procured from Micro Tissues Inc (Providence, RI, USA). The molds
   were subjected to autoclave sterilization and subsequently dried in an
   ultraviolet irradiation setting for a duration of 1 h. Two percent
   (g/mL) agarose powder (Sigma) was mixed with DMEM/F12 and autoclave
   sterilized. While the mixture was still in liquid form, 500 µL of it
   was carefully pipetted into each mold. Following solidification, the
   micro‐well plates were extracted employing a sterilized shovel. The
   agarose dishes were placed in 6‐well culture Petri dishes with 2 mL
   DMEM/F12 containing 1% P/S and incubated at 37 °C with 5% CO2 for 24 h.
   Prior to adding MSCs to the micro‐well plates, the medium was removed
   from both the culture dishes and the micro‐well plates. Next, 200 µL of
   cell suspension, containing 2 × 10^5 cells mL^−1, was carefully added
   to each well of the micro‐well plates. Then, 2 mL of DMEM/F12 with 10%
   exosome‐depleted FBS and 1% P/S was added to the 12‐well culture Petri
   dishes (Corning, NY, USA). The culture dishes were incubated in a 37 °C
   incubator of 5% CO2 for 48 h and the medium was collected for EVs
   isolation.

Figure 1.

   Figure 1
   [56]Open in a new tab

   Characterization of 2D and 3D‐MSCs. A) Schematic diagrams of the
   establishment of 3D culture systems. B) Formation of MSC spheroids in
   3D culture dishes. C) Calcein‐AM staining analysis of 2D, 3D‐MSCs and
   readhered cultured 3D‐MSCs. D,E) Flow cytometry analysis of surface
   markers of 2D and 3D‐MSCs. F,G) FSC/SSC analysis of 2D and 3D‐MSCs. FSC
   (Scattered light) reflects the relative size of cells. Mean ± SE, n =
   4, ^** P < 0.01.

2.3. Treatment and detection of MLE‐12 Cells

   The murine lung epithelial cell line MLE‐12, procured from the American
   Type Culture Collection (Manassas, VA, USA), was cultured in Dulbecco's
   Modified Eagle Medium (DMEM; Gibco) supplemented with 10% fetal bovine
   serum (FBS; Gibco) and 1% penicillin/streptomycin (Gibco). The cells
   were seeded to a 6‐well culture plate at a density of 10^6 cells per
   well and cultured for 24 h. Subsequently, the cells were challenged
   with lipopolysaccharide (LPS; 1 µg mL^−1; Sigma, USA) and tumor
   necrosis factor‐alpha (TNF‐α; 10 ng mL^−1; Sigma, USA) for 24 h to
   induce a model of pulmonary injury in vitro. Concurrently, the cells
   were co‐cultured with 1 × 10^5 stem cells or treated with 10 µg mL^−1
   of EVs to assess their potential protective effects. After a 24‐h
   incubation, the cells were harvested for subsequent analysis.

   The quantification of early and late apoptotic events was performed
   using the Annexin V‐FITC/PI Apoptosis Detection kit (BD Biosciences).
   After a preliminary wash, cells were resuspended in annexin V‐binding
   buffer and subjected to staining with FITC‐conjugated Annexin V and
   propidium iodide (PI) for a duration of 15 min. The subsequent analysis
   was conducted employing a flow cytometer (BD FACSAria, Rockville, MD)
   to distinguish apoptotic populations.

   Reactive oxygen species (ROS) levels were determined using the DCFH‐DA
   assay kit (Solarbio, China), adhering to the manufacturer's protocol.
   Cells, pre‐treated as previously described, were incubated with DCFH‐DA
   in a dark chamber at 37 °C for 20 min, after which the fluorescence
   intensity was assessed using a flow cytometer to indicate ROS
   production.

   Mitochondrial membrane potential (MMP) was evaluated using the JC‐1
   assay kit, following the manufacturer's guidelines. Cells, pre‐treated
   as outlined, were incubated with JC‐1 (5 m m) for 30 min, and the
   resulting fluorescence emissions were analyzed by flow cytometry. The
   MMP was calculated based on the ratio of red to green fluorescence
   emissions.

   Cell proliferation was assessed using the EdU Cell Proliferation Kit
   and Cell Cycle Analysis Kit (Beyotime, China), in accordance with the
   manufacturer's instructions. For the scratch wound assay, MLE‐12 cells
   were cultured to confluence in 12‐well plates at a seeding density of 1
   × 10^5 cells per well. A sterilized 1 mL pipette tip was utilized to
   create a scratch in the cell monolayer, which was then treated with
   MSCs or EVs, as detailed in the preceding sections. After a 24 h
   incubation, microscopy was employed to observe cellular migration and
   to quantify the remaining open wound area.

2.4. Isolation and Identification of EVs

   After a 48‐hour incubation period, the culture medium was collected and
   processed through a series of centrifugation steps to remove floating
   cells, debris, and cellular remnants. Initial centrifugation was
   performed at 300 g for 10 min, followed by 2000 g for 10 min, and
   finally 10000 g for 30 min. The supernatant was then filtered through a
   0.22‐µm pore size filter to ensure sterility and purity.
   Ultracentrifugation at 100000 g for 90 min was subsequently employed to
   pellet EVs, which were once washed with PBS. The purified EVs were
   resuspended in PBS and stored at −80 °C for further analysis.

   Concurrently, the size distribution and concentration of the EVs were
   determined using nanoparticle tracking analysis (NTA) with the Flow
   NanoAnalyzer (NanoFCM, China). The morphology of the EVs was examined
   under transmission electron microscopy (TEM, Japan) to provide insights
   into their structural characteristics. To semi‐quantify the EV
   concentration, the protein content of EVs derived from the culture
   medium of both 2D‐MSCs and 3D‐MSCs was measured using the bicinchoninic
   acid (BCA) protein assay kit (Beyotime, China), as previously
   described. The presence of EV protein markers, including HSP70, Tsg101,
   and CD81, was confirmed by Western blot analysis, with specific
   antibody details provided in Table [57]S1 (Supporting Information).

2.5. Quantitative Real‐Time Polymerase Chain Reaction (qPCR)

   The Trizol reagent (Invitro gen, Calif) was used to extract total RNA
   from MSCs, MLE‐12 cells, and lung tissue samples. Following the
   manufacturer's instructions, the Prime Script RT Master Mix (Takara,
   China) was utilized to synthesize complementary DNAs (cDNAs). RT qPCR
   was conducted on a Light Cycler 480 detection system with the iTaq
   universal SYBR Green Super Mix. Employing GAPDH as the endogenous
   control, primers were designed using the Primer 5.0 software
   ([58]http://www.premierbiosoft.com/primerdesign/) and list as follow:
   F/Oct4: AGTTGGCGTGGAGACTTTGC, R/Oct4: CAGGGCTTTCATGTCCTGG; F/Nanog:
   GAACGCCTCATCAATGCCTGCA, R/Nanog: GAATCAGGGCTGCCTTGAAGAG; F/Sox2:
   AACGGCAGCTACAGCATGATGC, R/Sox2: CGAGCTGGTCATGGAGTTGTAC; F/Hgf:
   TATTGCCCTATTTCCCGTTGTGAAG, R/Hgf: CCTACTGTTGTTTGTGTTGGAATGC, F/Gadph:
   ACAACTTTGGTATCGTGGAAGG, R/Gadph: GCCATCACGCCACAGTTTC. The 2‐ΔΔCt method
   was utilized to calculate the mRNA expression levels.

2.6. Animal Procedures

   6–8 weeks male C57BL/6 mice (18–22 g body weight) were purchased from
   Beijing Shanghai JieSiJie Laboratory Animal Company (Shanghai, China).
   All mice were housed in a standard animal care room with a 12 h
   light/dark cycle and had free access to food and water. All experiments
   and surgical procedures were executed following the criteria of the NIH
   and authorized by the Medical Ethics Committee of Shanghai Pulmonary
   Hospital.

   The sepsis‐associated lung injury model mice were created by cecal
   ligation and puncture (CLP) according to previous studies.^[ [59]^14 ^]
   Sham group received the same procedure expect for CLP. The treatment
   group nebulized PBS corresponding EVs (10^7 particles diluted in 1 mL
   PBS) at 2, 6, and 12 h postoperatively, while one of the CLP group
   nebulized PBS as a negative control. 24 h after operation, the mice
   were sacrificed to obtain serum, and lung samples for molecular biology
   experiments and histological observation. Bronchoalveolar lavage was
   performed by injecting the lungs with 0.9% saline (0.5 mL) through the
   main bronchus, and this procedure was repeated three times. The
   bronchoalveolar lavag fluids (BALF) was collected for protein
   concentrations and cell count determine.

   In the survival experiments, animals were subjected to the various
   treatments as previously described, and their survival status was
   meticulously documented on a daily basis. The mice were euthanized on
   day 7 after survival data had been collected.

2.7. Lung Histopathology

   The lung tissues were preserved using a 4% solution of paraformaldehyde
   for a duration of 24 h. After that, the preserved lung tissues were
   embedded in paraffin using established techniques. Following this, the
   paraffin blocks were sliced into 5 µm thick sections and subjected to
   hematoxylin and eosin (H&E) staining. The resulting sections were
   examined using a microscope, and the assessment of lung tissue injury
   severity was evaluated using the following criteria,^[ [60]^15 ^]
   including intra‐alveolar congestion, intra‐alveolar hemorrhage,
   presence of neutrophils in air spaces or vessel walls, and the
   thickness of the alveolar wall/hyaline membrane. A numerical score
   ranging from 0 to 4 was assigned to each criterion, indicating the
   level of injury severity, with higher scores representing more severe
   damage. The total score for each criterion was then calculated as the
   lung injury score.

   After dewaxing and hydration, five µm paraffin sections were blocked
   with 1% BSA for 15 min in a humid box. The sections were then incubated
   with the primary antibody at 4 °C overnight to allow for optimal
   antigen‐antibody interaction. Subsequent to three washes with PBS, the
   sections were incubated with either FITC or Cy3‐conjugated secondary
   antibodies in the dark at room temperature for 1 hour to ensure
   specific signal detection. Nuclear counterstaining was performed with
   DAPI for 5 min to visualize cell nuclei. The slides were ultimately
   examined using a fluorescence microscope to capture the
   immunofluorescence signals.

   Terminal deoxynucleotidyl transferase‐mediated dUTP nick‐end‐labeling
   (TUNEL) assay staining was also performed on paraffin sections using
   the In Situ Cell Death Detection Kit. A commercial in situ cell death
   detection kit (Servicebio, China) was used to assay apoptosis in this
   study according to the manufacturer's protocol. The TUNEL‐positive
   cells, indicative of apoptotic activity, were enumerated under a
   fluorescence microscope, providing a quantitative measure of cell death
   within the tissue sections.

2.8. In Vivo Imaging of the EVs

   In vivo imaging of EVs was conducted using DiD Perchlorate‐labeled EVs,
   which were resuspended in PBS for nebulized administration. After an
   hour post‐administration, the mice were euthanized, and their vital
   organs were harvested. These organs were then subjected to imaging
   using the IVIS Lumina imaging system (Xenogen Corporation, Hopkinton,
   MA, USA) to detect the distribution and accumulation of the labeled
   EVs. DiD Perchlorate: λem = 646 nm, λex = 663 nm.

2.9. Pulmonary Function Test In Mice

   Pulmonary function measurements were performed on the FinePointe Series
   R&C system (EMMS, USA). Prior to the measurements, the animals were
   administered anesthesia. The mice were intubated and connected with the
   ventilator for mechanical ventilation to collect ventilation data.
   Pulmonary function data is analysis for minute volume (MV), dynamic
   compliance, lung resistance, changes in pleural pressure, and
   expiratory flow at 50% expired volume.

2.10. Western Blotting

   Lung tissue and cells were homogenized in ice‐cold PBS to obtain
   samples containing total proteins. The protein concentration was
   measured using a BCA Assay Kit (Takara, Japan). Subsequently, the
   protein samples were separated through electrophoresis and transferred
   to a polyvinylidene fluoride membrane. The membrane was then treated
   with a 5% fat‐free milk solution to block any nonspecific binding
   sites, following which it was incubated with primary antibodies.
   Following the washing step, a HRP‐conjugated secondary antibody against
   rabbit or goat IgG (dilution of 1:2000, Bioword, U.S.A.) was introduced
   and incubated. The immunoreactive bands were then detected using an ECL
   chemiluminescence kit (Amersham, Arlington Heights, IL). The protein
   intensity was quantified using Image Lab software and normalized to the
   internal control protein, β‐Actin or GAPDH. Previous studies have shown
   that GAPDH is stable during the cell three‐dimension culture.^[ [61]^16
   ^]

2.11. Enzyme‐Linked Immunosorbent Assay (ELISA)

   The serum concentration of IL‐6, and TNF‐α were assessed by commercial
   assay kit according to the manufacturer's protocol (Thermo Fisher
   Scientific; USA).

2.12. RNA Sequencing and Data Analysis

   Total RNA was extracted from various samples using Trizol reagent
   (ThermoFisher Scientific, 15 596 026). RNA sequencing and analysis were
   conducted by Beijing Gene+ (China), with differentially expressed genes
   identified at p ≤ 0.05 and |log[2]FC|≥1. For Kyoto Encyclopedia of
   Genes and Genomes (KEGG) pathway and Gene Enrichment analysis, a
   combination of DAVID Functional Annotation Bioinformatics
   ([62]https://david.ncifcrf.gov/) and CytoScape
   ([63]http://www.cytoscape.org) was used in our study.

2.13. Proteomics and Data Analysis

   EVs proteins from the 2D or 3D‐cultured MSCs were collected and further
   sent to Wuhan Metware Biotechnology for proteomics analysis.
   Differentially expressed proteins (DEPs) were screened based on the
   thresholds of p ≤ 0.05, and |log[2]FC|≥1. Afterward, the screened DEPs
   were subjected to multiple bioinformatic analyses, including Gene
   Ontology (GO) annotation and KEGG pathway enrichment analysis.

2.14. Small Interfering RNA (siRNA) Experiments

   Cells were seeded at a density of 5 × 10^5 per well in 12‐well plates
   and cultured to reach 50% confluence. Target gene knockdown was
   achieved by transfecting cells with siRNAs and EZ Trans reagents (LIFE
   iLAB BIO, China), following the manufacturer's protocol.
   Post‐transfection, cells were incubated at 37 °C with 5% CO[2] for 48
   h. Transfection efficiency was subsequently assessed using quantitative
   real‐time polymerase chain reaction (qRT‐PCR) analysis. The siRNAs used
   were synthesized by Sangong Biotech (China); with the sequences are
   listed as follow: GGAGAUACUACACCUACAATT (5′‐3′), UUGUAGGUGUAGUAUCUCCTT
   (3′‐5′).

3. Result

3.1. Scheme of 3D Spheroid Culture and Characterization of 3D‐MSCs

   After three serial passages, impurities were removed and the remaining
   cells were homogeneous, spindle shape, showing good growth activity
   (Figure [64]1C). The phenotypic profile of BMSCs was determined by flow
   cytometry. The results showed that the isolated cells positively
   expressed CD29, CD49e, and CD90 (over 95% of positive rate), but
   negatively expressed CD45. (Figure [65]1D,E). Figure [66]1A shows the
   protocol of agarose micro multi‐well dishes. Almost all BMSCs were
   generated into 3D spheroid after seeding in 3D culture dish for 12 h
   (Figure [67]1B). After the stem cell spheres were replanted in Petri
   dishes for 48 h, the BMSCs in the spheres radiated around the spheres
   and maintained good vitality (Figure [68]1C). Flow cytometric results
   revealed that over 95% of these cells expressed CD29, CD49e and CD90,
   and were negative for CD45, consistent with the current BMSCs criteria
   (Figure [69]1D,E). However, the relative size of cells was decreased to
   70%–80% of the original size by flow cytometric forward scatter (FSC)
   (Figure [70]1F,G). In fact, the cells in the core of sphere may
   maintain a smaller volume, which indicates that the cells have
   structure remodeling during the 3D culture process.

3.2. 3D‐MSCs Ameliorated Lung Epithelial Cell Injury Model In Vitro

   To investigate the therapeutic efficacy of 3D‐MSCs on lung epithelial
   injury in sepsis‐induced ALI in vitro, a non‐contact transwell
   co‐culture system was employed (Figure [71]2A). PKH26 (red) labeling
   was utilized to track the delivery of MSC‐derived EVs to recipient
   cells. Fluorescence imaging revealed the presence of PKH26‐positive
   spots on MLE‐12 cells, signifying a high uptake efficiency of EVs by
   the recipient cells (Figure [72]2B). The loss of mitochondrial membrane
   potential (MMP) was associated with early apoptosis and mitochondria
   function. JC‐1 probe was used to analyze the effects of MSCs on the
   loss of MMP in injured MLE‐12 cells. The results indicated that the
   3D‐MSCs played a more significant role in preserving mitochondrial
   fitness compared to 2D‐MSCs in MLE‐12 cells (Figure [73]2C,I).
   Additionally, 3D‐MSCs significantly reduced the levels of reactive
   oxygen species (ROS) in injured cells (Figure [74]2D,F). To evaluate
   the anti‐apoptotic effects of EVs on MLE‐12 cells, an Annexin V/PI
   assay was conducted. Results demonstrated that the 3D‐MSCs group
   exhibited superior anti‐apoptotic effects compared to the 2D‐MSCs group
   (Figure [75]2E,G). The scratch wound assay demonstrated a significant
   improvement in the migratory ability of MLE‐12 cells after a 24‐h
   co‐culture with 3D‐MSCs, compared to the cell model group (Figure
   [76]S1A, Supporting Information; Figure [77]2H). EdU staining was
   employed to assess the influence of MSCs on the proliferation of LPS
   and TNF‐α induced MLE‐12 cells (Figure [78]2J). Flow cytometry was used
   to quantify the EdU‐positive cell ratio, revealing a significant
   increase in the fluorescence density of EdU‐positive cells in the
   3D‐MSCs group relative to the injury model group (Figure [79]S1B,
   Supporting Information; Figure [80]2K). Cell cycle assays further
   substantiated the proliferative effects observed in the 3D‐MSCs group,
   with 35.5% of cells entering the S and G2 phases, a higher percentage
   compared to the LPS and TNF group and the 2D‐MSCs group
   (Figure [81]2L,M). All these data suggest that 3D‐MSCs exhibit superior
   therapeutic effects in epithelial injury in vitro, suggesting their
   potential as an effective treatment for sepsis‐induced ALI.

Figure 2.

   Figure 2
   [82]Open in a new tab

   3D‐MSCs alleviated lung epithelial damage in vitro. A) The protocol of
   2D or 3D MSC co‐culture with injury lung epithelial cells. B) Uptake of
   2D and 3D‐EVs in MLE‐12 cells detected by fluorescence microscope. C,I)
   MMP assay by fluorescence microscope and flow cytometry (Scale bar:
   50 µm). Increases in green fluorescence indicate perturbed MMP. D,F)
   intracellular ROS assay (count) and quantification of ROS. E,G) Annexin
   V/PI assay and quantification of apoptosis rate. H) Cell migration
   analysis. J,K) Cell proliferation was measured using EdU assays and
   quantification. (Scale bar: 100 µm). L,M)The cell cycle assay was
   conducted to analyze the proliferation viability. Mean ± SE, n = 4. * P
   < 0.05, ** P < 0.01.

3.3. Integral Analysis of Transcriptome of 3D‐MSCs

   To assess the influence of 3D culture on gene expression and to uncover
   the mechanisms behind the enhanced therapeutic potential of 3D‐MSCs, we
   performed a high‐throughput RNA‐sequencing analysis of the 3D‐MSCs
   transcriptome. First, Principal Component Analysis (PCA) was employed
   to delineate the genetic distinctions between 3D‐MSCs and 2D‐MSCs, as
   depicted in Figure [83]S2B (Supporting Information), revealing distinct
   gene expression clusters associated with each culture method. A total
   of 4795 Differentially Expressed Genes (DEGs) were identified from the
   expression profile datasets, with a log2 fold change > 1 and an
   adjusted p‐value < 0.05 (Figure [84]3A). To visualize the expression
   patterns of these DEGs, a heatmap was constructed to illustrate the
   expression matrix of the top 200 DEGs, which included 176 up‐regulated
   genes and 24 down‐regulated genes (Figure [85]S2C, Supporting
   Information). To identify hub genes, the top 200 DEGs were subjected to
   analysis using the STRING database, and a Protein‐Protein Interaction
   (PPI) network was generated using Cytoscape software (Figure [86]3B).
   Subsequently, the top 10 genes with high connectivity scores were
   identified using the CytoHubba/MCC plugin in Cytoscape, indicating
   their potential significance in the PPI network or in 3D cell spheroid
   formation (Figure [87]3C). The gene with the highest ranking among them
   is Acta2, which plays a role in cell motility, structure, and
   integrity. It also shows obvious downregulation at the protein level
   (Figure [88]3D). A detailed list of these top 10 genes, including their
   full names and functions, is presented in Table [89]1 . Additionally,
   the mRNA of growth factors in MSCs cultured in 3D exhibited substantial
   upregulation (Figure [90]3E). Genes that positively regulate cell
   proliferation were also extensively upregulated, potentially linking to
   their augmented therapeutic potency (Figure [91]S2E, Supporting
   Information). GO analysis annotated the DEGs and their gene products
   with respect to biological processes (BP), cellular components (CC),
   and molecular functions (MF), as shown in Figure [92]3G. The
   upregulation of genes associated with EV biogenesis (Figure [93]3F) and
   extracellular matrix remodeling (Figure [94]S2D, Supporting
   Information) implies a potential role for EVs in 3D‐MSC therapy. KEGG
   pathway analysis indicated that the DEGs were predominantly enriched in
   metabolic pathways, including “Metabolism of xenobiotics by cytochrome
   P450,” “Drug metabolism–other enzymes,” and “Ascorbate and aldarate
   metabolism” (Figure [95]3H). Furthermore, the expression of
   stemness‐associated genes Oct4, Sox2, Nanog, and Klf4 was upregulated
   in 3D spheroid‐cultured MSCs, suggesting the induction of a more
   primitive phenotype compared to 2D culture conditions (Figure [96]3I).

Figure 3.

   Figure 3
   [97]Open in a new tab

   RNA‐sequencing analysis of 2D and 3D‐MSCs. A) Volcano plots of
   significant DEGs (3D/2D). Red dots represent up‐regulated genes, blue
   dots represent down‐regulated genes. B) The PPI network of DEGs. C) The
   top 10 hub genes ranked using MCC plug‐in of Cytoscape. D) The protein
   expression level of the most significantly downregulated gene Acta2. E)
   Expression cluster of Growth factor family genes. F) Expression heatmap
   of genes related to EV biogenesis. G) GO enrichment analysis of DEGs.
   H) KEGG pathway analysis of DEGs. I) Stem cell gene expression. Mean ±
   SE, n = 4. ^* P < 0.05, ^** P < 0.01.

Table 1.

   Ten hub genes and their functions.
   Rank Gene symbol Description Function logFC adj P
   1 Acta2 actin alpha 2, smooth muscle Encodes one of six different
   highly conserved actin proteins. −4.99 1.00E‐4
   2 Hmgcs2 3‐hydroxy‐3‐methylglutaryl‐CoA synthase 2 Enables
   hydroxymethylglutaryl‐CoA synthase activity. 5.34 3.18E‐3
   3 Lcn2 lipocalin 2 Enables iron ion binding activity. Involved in
   innate immune response; positive regulation of cold‐induced
   thermogenesis; and siderophore transport. 6.89 1.78E‐4
   4 Mmp13 matrix metallopeptidase 13 A member of the matrix
   metalloproteinase family that plays a role in degradation of
   extracellular matrix proteins 5.45 1.42E‐4
   5 Mmp2 matrix metallopeptidase 2 A member of the matrix
   metalloproteinase family of extracellular matrix‐degrading enzymes 5.34
   7.71E‐4
   6 Cd68 CD68 antigen‐ Involves in several processes, including cellular
   response to lipopolysaccharide; cellular response to oxidised
   low‐density lipoprotein particle stimulus. 4.19 1.52E‐3
   7 Mmp3 matrix metallopeptidase 3 A member of the matrix
   metalloproteinase family of extracellular matrix‐degrading enzymes.
   8.92 5.78E‐4
   8 Mmp12 matrix metallopeptidase 12 A member of the matrix
   metalloproteinase family of extracellular matrix‐degrading enzymes.
   4.95 2.56E‐3
   9 Eln elastin Encodes elastin of the extracellular matrix protein 4.48
   4.58E‐4
   10 Dcn decorin Encodes preproprotein is processed to produce a mature
   protein that regulates collagen fibril assembly in the extracellular
   space. 5.15 2.96E‐4
   [98]Open in a new tab

3.4. Characterization of 3D‐MSCs Derived EVs

   Extensive research has highlighted the therapeutic potential of EVs
   derived from MSCs through their paracrine signaling capabilities for
   various conditions.^[ [99]^17 ^] Consequently, EVs were isolated via
   ultrafiltration and ultracentrifugation for further experimental
   analysis (Figure [100]4A). Transmission electron microscopy (TEM)
   images revealed that both 2D‐EVs and 3D‐EVs exhibited the
   characteristic cup‐base shape (Figure [101]4B). Western blotting
   confirmed the presence of specific EV protein markers, including HSP70,
   TSG101, and CD81, which were enriched in both 2D‐EVs and 3D‐EVs
   (Figure [102]4C). Interestingly, the mean and median diameter of the
   3D‐EVs (73.08  and 66.97 nm) were significantly smaller than 2D‐EVs
   (84.31  and 78.57 nm, respectively) (Figure [103]4D,E). It seems not to
   be reported in most previous related studies, except the study of Yuan
   et al.^[ [104]^11 ^] The supernatant of quantitative cell count was
   collected by ultracentrifugation, 3D‐EVs groups showed more particle
   precipitation (Figure [105]4F). The result of nanoFCM showed that the
   number of 3D cultured production was 2‐fold than 2D group
   (Figure [106]4G,H), while the EV‐based protein content was no
   difference between the two groups (Figure [107]4I). These results
   indicated that the harvested EVs from both 2D and 3D cultures conform
   to the typical criteria; however, there exist discrepancies in their
   properties.

Figure 4.

   Figure 4
   [108]Open in a new tab

   Characterization of EVs derived from 2D and 3D cultured MSCs. A)
   Protocol of EV isolation from MSCs supernatant. B) The morphology of
   EVs from 2D and 3D‐MSCs was detected by TEM (Scale bar: 200 nm). C)
   Protein expression of MSC‐EVs positive markers (Hsp70, TSG101, and
   CD81). D,E) The particle size distribution of MSC‐EVs was measured by
   nanoflow cytometer. F) Ultracentrifugation enrichment of 2D and 3D‐EVs.
   G,H) quantitative analysis of EV production from 2D and 3D culture. I)
   Total protein quantification in 2D and 3D‐EVs. Mean ± SE, n = 4. ^*
   P < 0.05, ^** P < 0.01.

3.5. 3D‐EVs Ameliorated Sepsis‐Induced ALI Model In Vivo

   We next investigated the protective effects of 3D‐EVs in the septic ALI
   model in vivo, and the animal procedure is shown in Figure [109]5A.
   First, to demonstrate that EVs could be taken in by lung epithelial
   tissue in vivo, EVs were labeled with a lipophilic far‐red fluorescent
   dye DID and administered to mice via inhalation. After an hour, the in
   vivo tracing showed that the 2D‐EVs and 3D‐EVs groups enriched red
   fluorescence in lung tissue (Figure [110]5B), and the red fluorescence
   was observed in mice lung alveolus by frozen fluorescent sections,
   indicating the effective uptake of EVs by the lung epithelial tissue
   (Figure [111]5C). Histological staining showed that CLP induced acute
   injury responses, such as interstitial edema, infiltration of
   inflammatory cells, destruction of alveoli (Figure [112]5D,E), and lung
   cells damage (Figure [113]5I,J). All these symptoms were significantly
   relieved after EV administration, with particularly noticeable
   improvements observed in the 3D‐EVs group. Meanwhile, BALF analysis
   results revealed that administration of 3D‐EVs significantly reduced
   the total protein concentration and inflammatory cell exudation in BALF
   (Figure [114]5G,H), accompanied by reduction of lung wet‐to‐dry ratio
   (Figure [115]5F). EVs also reduced the levels of inflammatory
   (cytokines IL‐6 and TNF‐α) in serum, although no significant difference
   was observed between the two intervention groups (Figure [116]5K,L).

Figure 5.

   Figure 5
   [117]Open in a new tab

   The effects of 3D‐EVs on ameliorating sepsis‐induced acute lung injury
   A) The schematic diagram of study design and procedure in vivo. B)
   Fluorescence imaging of tissues from different groups with or without
   nebulization. C) Detection of DID‐labeled EVs uptake by lung tissues in
   both fluorescence and bright field (Scale bar: 50 µm). D)
   Representative images of H&E‐stained lung tissue sections. (Top: Scale
   bar: 500 µm Bottom: Scale bar: 100 µm). E) The pathological lung injury
   score was calculated according to the H&E staining. F) Wet‐to‐dry ratio
   of lung tissues. G) Total protein concentration in BALF. H) Total cell
   counts in BALF. I), J) TUNEL staining of apoptotic cells and
   quantitative analysis of fluorescence intensity by ImageJ (Scale bar:
   200 µm). Levels of K) IL‐6 and L) TNF‐α in plasma measured by ELISA.
   Mean ± SE. ^* P < 0.05, ^** P < 0.01.

3.6. 3D‐EVs Maintained Lung Epithelial Barrier Integrity via PI3K/AKT Pathway

   The alveolar epithelium consists of gas‐exchange‐facilitating alveolar
   type 1 epithelial cells (AT1 cells, HOPX positive) and alveolar type 2
   epithelial cells (AT2, ProSPC positive), which are responsible for the
   secretion of pulmonary surfactants, antioxidants, and other crucial
   defensive molecules in the lung. Here, we examined the changes in AT1
   and AT2 distribution in response to EV treatment post‐CLP injury to
   assess epithelium damage and rescue (Figure [118]6A,C). Meanwhile, to
   demonstrate the effects of 3D‐EVs on the alveolar epithelial barrier,
   we assayed the expression of ZO‐1 and occludin (Figure [119]6A,C),
   which maintain the integrity of the alveolar epithelial barrier. We
   observed that the protein levels of ZO‐1, occludin, ProSPC, and HOPX
   declined in pulmonary tissues of CLP model mice (Figure [120]6B).
   However, these effects were reversed by inhalation of 3D‐EVs.
   Additionally, EVs did not significantly alter the relative number of
   ProSPC‐positive cells (Figure [121]S4, Supporting Information), which
   may be associated with the early activation of endogenous repair
   mechanisms in ALI.^[ [122]^18 ^] Collectively, 3D‐EVs exhibit superior
   restorative effects on compromised lung epithelium and epithelial
   barrier integrity compared to 2D‐EVs.

Figure 6.

   Figure 6
   [123]Open in a new tab

   3D‐EVs alleviated sepsis‐induced epithelial barrier damage via PI3K/AKT
   pathway in vivo. A), B) Expression of ZO‐1, Occludin, ProSPC, and HOPX
   in lung tissues across different groups, as detected by Western
   blotting, was quantified using ImageJ and normalized to β‐Actin. C)
   Immunofluorescence staining showing expression and distribution of ZO‐1
   (Red), occludin (Red), ProSPC (Green), and HOPX (Green) in different
   groups of lung tissues (Scale bar: 200 µm). Quantification of pulmonary
   function: D) minute volume (MV); E) dynamic compliance; F) lung
   resistance; G)changes in pleural pressure; H) expiratory flow at 50%
   expired volume. I) Survival curves of mice in different groups (n =
   12). J,K) The protein level of PI3K, AKT, PI3K phosphorylation, and AKT
   phosphorylation was examined by Western blotting. β‐Actin served as a
   loading control. The relative quantification of the detected signals
   was determined using ImageJ and normalized to β‐Actin. Mean ± SE. ^*
   P < 0.05.

   The integrity of the lung epithelial barrier is crucial for maintaining
   pulmonary function and preventing pathogenic invasion.^[ [124]^19 ^]
   Subsequent pulmonary function analyses revealed that 3D‐EVs
   counteracted the decrease in pulmonary compliance induced by ALI,
   underscoring their therapeutic potential (Figure [125]6E). Notably,
   airway resistance parameters in the CLP model were unaltered, and no
   emphysema was observed. Consequently, other associated indicators
   displayed negative outcomes (Figure [126]6D,F–H). The administration of
   3D‐EVs to mice with sepsis‐induced ALI also enhanced their survival
   rates (Figure [127]6I). The engagement of the PI3K/AKT signaling
   pathway has been demonstrated to mitigate cellular oxidative stress and
   to foster cell proliferation, among other beneficial effects.^[
   [128]^20 ^] By modulating the expression of key genes and proteins
   involved in these processes, the PI3K/AKT pathway serves as a critical
   regulator of cellular homeostasis and response to stress.^[ [129]^21 ^]
   Abnormal activation of the canonical PI3K/AKT signaling pathway has
   been strongly correlated with alveolar destruction, fibrosis, and loss
   of epithelial barrier integrity. To ascertain whether 3D‐EVs modulate
   epithelial injury and barrier stability through the PI3K/AKT pathway,
   we measured the phosphorylation levels of PI3K and AKT in lung protein
   lysates. Immunoblotting analysis demonstrated reduced expressions of
   p‐AKT and p‐PI3K in the model group, with a subsequent restoration of
   these expressions in the lungs of 3D‐EVs treated mice
   (Figure [130]6J,K).

3.7. 3D‐EVs Enriched and Delivered HGF, Maintaining the Integrity of the Lung
Epithelial Barrier

   Previous RNA‐sequencing studies have revealed a broad upregulation of
   growth factor gene clusters in 3D‐MSCs. To ascertain if the
   corresponding EVs encapsulate functional proteins that contribute to
   therapeutic outcomes, we conducted a proteomics analysis of these EVs.
   Data normalization via the Robust Multichip Average (RMA) method
   rendered the distributions clear and stable, with selected values in
   the violin plot used for subsequent analyses (Figure [131]S3A,
   Supporting Information). Principal component analysis (PCA) was then
   employed, revealing a distinct segregation between DEPs from 2D‐EVs and
   D‐EVs EVs in the 3D score plot (Figure [132]S3B, Supporting
   Information). The volcano plots of DEPs are depicted in Figure [133]7A.
   The heatmap of the top 200 significant DEPs is presented in
   Figure [134]7B, with hierarchical clustering analysis demonstrating the
   clear discrimination between DEPs from 3D‐EVs and 2D‐EVs. In terms of
   protein subcellular localization, the proteome was predominantly
   composed of nucleic (28.52%), cytoplasmic (23.48%), and extracellular
   (20.32%) components (Figure [135]7C). A total of 1073 DEPs were
   integrated into a protein‐protein interaction (PPI) network and
   visualized using Cytoscape (Figure [136]S3C, Supporting Information).
   KEGG pathway analysis, comparing 3D‐EVs to 2D‐EVs, identified the
   PI3K/AKT, MAPK, Ras, and Rap1 signaling pathways as being DEP‐enriched
   (Figure [137]S3D, Supporting Information). GO analysis revealed that
   3D‐EVs contained proteins involved in cell adhesion, proteolysis, and
   extracellular matrix organization (Figure [138]7D). The differential
   expression of proteins implicated in vesicle‐mediated transport and EV
   biogenesis is illustrated in Figure [139]7E,F, suggesting that 3D
   culture systematically modifies the EVs derived from MSCs. Furthermore,
   the Venn diagram displays genes and proteins that are concurrently
   upregulated or downregulated in the transcriptomics of 3D‐MSCs and
   proteomics of 3D‐EVs. Notably, there were 82 co‐upregulated and 35
   co‐downregulated genes or proteins, with HGF being the only upregulated
   growth factor in both expression profiles (Figure [140]7G).
   Subsequently, Western blotting analysis confirmed that 3D‐EVs contained
   higher levels of HGF protein compared to 2D‐EVs (Figure [141]S5,
   Supporting Information).

Figure 7.

   Figure 7
   [142]Open in a new tab

   Proteomics analysis of 2D and 3D‐EVs. A) Volcano plots of DEPs
   (3D‐EVs/2D‐EVs). Red dots represent up‐regulated proteins, blue dots
   represent down‐regulated proteins. B) Heatmap of top 200 DEPs from
   proteomics analysis. C) Protein subcellular location of DEPs. D) GO
   enrichment analysis of DEPs. E,F) Heatmap of EVs‐related proteins. G)
   Venn diagram of co‐expressed genes and proteins identified in DEGs and
   DEPs.

   Among other growth factors, HGF is a well‐recognized anti‐apoptotic and
   pro‐angiogenic factor. Generally, cells express Met and respond to HGF
   by activating PI3K/AKT and ERK1/2 signalling.^[ [143]^22 ^] To evaluate
   the role of HGF in the protective effects of 3D‐EVs against
   sepsis‐induced ALI and epithelial barrier damage, we performed HGF
   silencing using siRNA in vitro (Figure [144]8A). First, to confirm the
   cellular uptake in vitro, EVs were labeled with a red fluorescent lipid
   dye PKH26 and incubated with cells. After 24 h, all of the EVs were
   internalized by cells in the presence of LPS and TNF‐α
   (Figure [145]8B). Flow cytometry apoptosis assay demonstrated a
   significant reduction of the damaging effect of inflammation factors on
   MLE‐12 cells, an effect that was attenuated in the HGF siRNA group
   (Si‐HGF‐3D‐EVs) (Figure [146]8C,D). The scratch wound assay revealed
   that the migratory capacity of MLE‐12 cells was substantially enhanced
   after a 24‐hour treatment with 3D‐EVs, as compared to the TNF‐α group.
   However, the healing of the scratched surface was significantly delayed
   in the HGF siRNA group (Figure [147]8E,F). Next, Western blot results
   revealed that HGF siRNA administration partially blocked the
   therapeutic effects of 3D‐EVs on the expression of tight junction
   proteins, ZO‐1, and occludin (Figure [148]8G). As shown in
   Figure [149]8H, 3D‐EVs significantly alleviated tight junction protein
   damage or discontinuous distribution. While HGF‐Si‐3D‐EVs weakened this
   effect. To determine whether HGF derived from 3D‐EVs affected
   epithelial function recovery by activating the PI3K/AKT pathway,
   phosphorylation level of PI3K, AKT was measured in vitro. Immunoblot
   analysis indicated that 3D‐EVs reversed the decrease in the
   injury‐induced p‐AKT/AKT ratio and p‐PI3K/PI3K ratio, while
   HGF‐Si‐3D‐EVs reduced this effect (Figure [150]8G,I). These cumulative
   results indicate that 3D‐EVs can partly reverse sepsis‐induced
   down‐regulation of the PI3K/AKT pathway in mouse lung by delivering HGF
   and thus ameliorating lung injury (Figure [151]9 ).

Figure 8.

   Figure 8
   [152]Open in a new tab

   3D‐EVs delivered HGF to protect lung epithelial cells from apoptosis
   and epithelial barrier damage. A) The protocol of 2D, 3D‐EVs, or
   Si‐HGF‐3D‐EVs co‐culture with injured lung epithelial cells. B) Uptake
   of 2D, 3D‐EVs, or Si‐HGF‐3D‐EVs in MLE‐12 cells detected by
   fluorescence microscope. C), D) Annexin V/PI assay and quantification
   of apoptosis rate. E,F) A scratch experiment was used to assess cell
   migration (Scale bar: 200 µm). G,I)The protein expression level of
   ZO‐1, occludin, PI3K, AKT, PI3K phosphorylation, and AKT
   phosphorylation was examined by Western blotting. β‐Actin served as a
   loading control. The relative quantification of the detected signals
   was determined using ImageJ and normalized to β‐Actin. H)
   Immunofluorescence staining showing expression and distribution of ZO‐1
   (Red) in MLE‐12 cells (Scale bar: 100 µm). Mean ± SE, n = 4. ^*
   P < 0.05, ^** P < 0.01.

Figure 9.

   Figure 9
   [153]Open in a new tab

   MSCs cultured in 3D conditions up‐regulated the expression of matrix
   metalloproteinases (MMPs) and growth factors (GFs), and increased the
   production of extracellular vesicles (EVs). Proteomic analysis showed
   that 642 proteins were up‐regulated in 3D‐EVs (including HGF).
   Moreover, 3D‐EVs are enriched with HGF and deliver it to protect lung
   epithelial cells from apoptosis and barrier damage in septic ALI
   through the PI3K/AKT pathway. The figure is created by Biorender.com.

4. Discussion

   Over the past few decades, preclinical studies have highlighted the
   therapeutic potential of EVs, particularly those derived from MSCs, in
   tissue regeneration and repair.^[ [154]^23 ^] This has motivated
   researchers to harness EVs, particularly those derived from MSCs, as
   cell‐free therapeutics for a variety of diseases. Notably, EVs from
   MSCs have been shown to be as efficacious as their parent cells in
   promoting tissue regeneration, as evidenced by their therapeutic
   potential across diverse tissue injury models.^[ [155]^17a ^]

   Despite some progress, sepsis‐induced ALI remains a significant
   challenge in critical care, characterized by high mortality and limited
   treatment options. Our previous studies have shown that MSCs and their
   exosomes protect against sepsis‐induced ALI.^[ [156]^24 ^] However,
   challenges such as low production yields, variable bioactivity, and
   impurity issues hinder their broader application.^[ [157]^25 ^] Recent
   insights suggest that environmental shifts in the microenvironments of
   primary cells can alter EV content. In response, researchers are
   exploring various 3D culture techniques to enhance EV production and
   therapeutic efficacy. Preliminary studies have indicated that MSCs
   cultured in 3D spheroids can reduce collagen deposition, improve
   pulmonary function, and offer therapeutic benefits in animal models of
   lung disease.^[ [158]^26 ^] However, the impact of 3D‐EVs with specific
   cargo on septic ALI repair remains underexplored. In this study, we
   describe a novel, cost‐effective technique for rapid MSC spheroid
   formation and thorough characterization of the EVs they produce, with
   the aim of clarifying their therapeutic mechanisms in ALI.

   During spheroid formation, MSCs exhibited elevated expression of
   stemness genes without altering surface markers. Previous studies have
   suggested the consensus that maintenance of stem cell stemness is
   highly associated with ageing‐related diseases. Increasing MSCs'
   stemness has been extensively used to enhance their therapeutic
   effectiveness in treating various diseases.^[ [159]^27 ^] In vitro
   experiments involving non‐contact co‐culture of MSC spheroids and
   injured lung epithelial cells demonstrated that aggregated MSCs
   significantly enhance therapeutic efficacy by mitigating early injury
   and promoting cell proliferation. Additionally, 3D spheroid MSCs are
   approximately one‐fourth the size of traditionally cultured MSCs,^[
   [160]^27b ^] potentially enhancing their lung migration capabilities.
   Our research supports this, showing a reduction in size for MSCs
   undergoing spheroid formation. However, the risk of embolism from
   direct intravascular injection of cell spheres remains.^[ [161]^28 ^]
   Transcriptomic analysis of 3D‐MSCs revealed a notable enrichment in the
   matrix metalloproteinases family, crucial for extracellular matrix
   degradation.^[ [162]^29 ^] This, along with the downregulation of
   cytoskeleton genes and reduced cell size, suggests significant cellular
   restructuring during spheroid formation. Intriguingly, there is a
   positive correlation between the quantity of shed vesicles, their lytic
   enzyme content, and the invasive potential of various cell lines.^[
   [163]^29 ^] This also confirms that 3D culture reduces MSC size by
   increasing vesicle excretion, likely mediated by reduced cytoskeleton
   tension.^[ [164]^16 ^] Similar to previous studies, these processes may
   contribute to increased EVs yield, resulting in smaller and more
   uniform EVs.^[ [165]^11 ^] Furthermore, the internal core cells of the
   sphere exhibit altered metabolic pathways and increased
   pro‐regenerative paracrine signaling molecules and immunomodulatory
   factors, enhancing the therapeutic effects of MSCs.^[ [166]^30 ^]

   The nebulization of EVs has been established as an effective
   administration route for lung diseases.^[ [167]^26 , [168]^31 ^] EVs,
   when nebulized, are distributed in the lungs, primarily localizing in
   pulmonary macrophages and airway epithelial cells.^[ [169]^32 ^] Upon
   reaching the lungs, EVs transfer growth factors and other contents to
   injured alveoli for therapeutic purposes.^[ [170]^33 ^] In this study,
   the therapeutic effectiveness of 3D‐EVs in lung injury of CLP mice was
   verified through continuous inhalation therapy. Inhalation of EVs
   partially restored pulmonary structure and inhibited cell apoptosis,
   although slight lung wall thickening persisted. Interestingly, the
   injury score and other parameters were significantly higher in the
   2D‐EVs group compared to the 3D‐EVs group. Alveolar epithelial cells,
   critical for maintaining alveolar defense barrier integrity and
   regulating inflammation and external irritants,^[ [171]^34 ^] are
   compromised in lung injury, leading to increased alveolar permeability,
   pulmonary edema, and subsequent pathological alterations.^[ [172]^27a
   ^] The destruction of the pulmonary barrier often involves the
   disruption of tight junctions, evidenced by reduced expression of key
   tight junction markers such as occludin, Claudin‐1, and ZO‐1.^[
   [173]^35 ^] In this study, 3D‐EVs reversed the down‐regulation of
   occludin and ZO‐1 protein expression associated with ALI barrier
   function impairment. The infiltration of cells and total protein in
   BALF was also decreased. Pulmonary function test analysis demonstrated
   the efficacy of EVs in resolving pulmonary function, with 3D‐EVs
   exhibiting superior therapeutic effects compared to 2D‐EVs and being
   the sole treatment capable of reversing the reduction in dynamic lung
   compliance induced by ALI. This effect may be attributed to their
   ability to enhance type II pneumocyte synthesis^[ [174]^36 ^] and
   maintain alveolar structural integrity. The normal synthesis and
   secretion of surfactant protein C play a critical role in ensuring
   alveolar stability and proper pulmonary function by reducing alveolar
   surface tension and promoting alveolar re‐expansion.^[ [175]^37 ^]

   Proteomics has been instrumental in elucidating the molecular
   mechanisms underlying EV‐mediated lung repair. In this study, 51
   proteins were implicated in the PI3K−AKT signaling pathway, known to
   delay cellular senescence and extend organismal lifespan through
   diverse cellular processes.^[ [176]^38 ^] Several proteins were
   predicted to regulate the MAPK signaling pathway, central to cell
   proliferation and survival, suppression of the aging process, and
   promotion of BMMSC differentiation.^[ [177]^39 ^] Consistent with
   previous studies,^[ [178]^11 ^] there was minimal overlap between DEGs
   and secreted DEPs, suggesting specific sorting mechanisms governing EV
   cargo. Notably, HGF was the only growth factor significantly
   upregulated in both the parental transcriptome and EV proteomics data.
   Previously, HGF has been shown to exhibit anti‐apoptotic and
   anti‐fibrotic properties through the activation of the PI3K/AKT/mTOR
   pathway, contributing to lung barrier restoration in ALI.^[ [179]^35 ,
   [180]^40 ^] Okouchi et al. found that enhancing endogenous HGF levels
   in the lungs specifically inhibits alveolar epithelial cell apoptosis,
   mitigating inflammation and fibrosis following bleomycin‐induced
   injury.^[ [181]^41 ^] Thus, we posit that HGF derived from 3D‐EVs
   contributes to their therapeutic effects. Furthermore, we discovered
   that Si‐HGF‐3D‐EVs diminished the protective effects of 3D‐EVs on
   maintaining lung epithelium barrier integrity. Therefore, 3D‐EVs
   enriched with and delivering HGF present an effective approach for
   alveolus epithelial injury treatment.

   However, there are several potential limitations to the current study.
   First, we did not conduct parallel controlled trials between 3D‐MSCs
   and derived EVs to compare their therapeutic effects. Future research
   in this direction will aid in selecting cell therapy approaches with or
   without cells under various disease conditions. Second, only one dose
   of EVs was administered in this study, guided by previous experience. A
   dose‐response study to establish the optimal dosage would be beneficial
   for future pre‐clinical and clinical research. Third, while our data
   indicate that 3D‐EVs can effectively ameliorate sepsis‐ALI via HGF
   delivery, further investigation is needed to reveal the specific
   mechanisms and identify alternative substances playing therapeutic
   roles in EVs.

   In summary, our study demonstrates that 3D culture‐derived MSCs and
   their EVs enhance anti‐injury ability and maintain alveolar epithelial
   barrier function in experimental sepsis‐induced ALI. Compared to
   conventional 2D culture conditions, 3D spheroid culture has been shown
   to enhance EV yield. Furthermore, 3D‐EVs have the potential to enrich
   and deliver HGF to injured tissues, activating the PI3K/AKT signaling
   pathway to exert beneficial effects. Given the high morbidity and
   mortality rates associated with ALI and the current lack of effective
   therapies, these findings offer promising prospects for the development
   of ALI treatments.

Conflict of Interest

   The authors declare no conflict of interest.

Supporting information

   Supporting Information
   [182]ADVS-12-2500637-s001.docx^ (1.8MB, docx)

Acknowledgements