Abstract

   Non-healing skin wounds pose significant clinical challenges, with
   biologic products like exosomes showing promise for wound healing.
   Saliva and saliva-derived exosomes, known to accelerate wound repair,
   yet their extraction is difficult due to the complex environment of
   oral cavity. In this study, as a viable alternative, we established
   human minor salivary gland organoids (hMSG-ORG) to produce exosomes
   (MsOrg-Exo). In vitro, MsOrg-Exo significantly enhanced cell
   proliferation, migration, and angiogenesis. When incorporated into a
   GelMA-based controlled-release system, MsOrg-Exo demonstrated
   controlled release, effectively improving wound closure, collagen
   synthesis, angiogenesis, and cellular proliferation in a murine skin
   wound model. Further molecular analyses revealed that MsOrg-Exo
   promotes proliferation, angiogenesis and the secretion of growth
   factors in wound sites. Proteomic profiling showed that MsOrg-Exo’s
   protein composition is similar to human saliva and enriched in proteins
   essential for wound repair, immune modulation, and coagulation.
   Additionally, MsOrg-Exo was found to modulate macrophage polarization,
   inducing a shift towards M1 and M2 phenotypes in vitro within 48 h and
   predominantly towards the M2 phenotype in vivo after 15 days. In
   conclusion, our study successfully extracted MsOrg-Exo from hMSG-ORGs,
   confirmed the effectiveness of the controlled-release system combining
   MsOrg-Exo with GelMA in promoting skin wound healing, and explored the
   potential role of macrophages in this action.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-024-02811-y.

   Keywords: Wound Healing, Exosome, Organoid, Human minor salivary gland,
   Macrophage

Background

   Chronic non-healing skin wound remains a challenging condition for
   clinical management. The use of biologic products has been shown to
   promote wound healing and regeneration [[40]1–[41]5]. Previous studies
   have investigated the therapeutic potential of exosomes in wound
   repair, with the majority of studies focusing on stem cell-derived
   exosomes [[42]6–[43]11]. Saliva contains a variety of bioactive
   components, including growth factors, antimicrobial peptides and
   immunoglobulins, and has been demonstrated to accelerate oral and skin
   wound healing [[44]12–[45]15]. In addition, it has been revealed that
   saliva-derived exosome enhanced cutaneous wound healing by promoting
   angiogenesis [[46]16]. Therefore, saliva represents a promising source
   of exosomes for therapeutic use in skin wound treatment.

   3D cultured organoids are composed of a large number of mature,
   functional cells and can produce abundant exosomes, overcoming the
   limitation of low yield in traditional 2D cell culture [[47]17]. The
   therapeutic use of organoid-derived exosome has been studied in various
   diseases [[48]18–[49]21], while its application in the context of skin
   wound healing remains largely unexplored. Due to the complex nature of
   oral environment [[50]22, [51]23], the extraction of pure, sterile
   saliva-derived exosome is extremely challenging in clinical settings.
   As an alternative, establishing salivary gland organoids in vitro
   enables the production of saliva-derived exosomes that can be used for
   bioengineering techniques.

   Human minor salivary glands (hMSG) are universally present throughout
   the oral mucosa and can be easily accessed [[52]24], offering
   advantages over major salivary glands for tissue harvesting. Harvesting
   major salivary glands is more technically demanding and associated with
   complications, including fistula formation and scarring at the incision
   site. Thus, minor salivary glands are the ideal tissue source for
   organoid construction.

   Previous studies have successfully developed a preliminary model of
   hMSG organoid (hMSG-ORG) [[53]25, [54]26]. In this study, we aimed to
   (1) incorporate human minor salivary gland organoid-derived exosome
   (MsOrg-Exo) with a Gelatin Methacryloyl (GelMA) hydrogel to establish a
   sustained-release system, and to (2) investigate the therapeutic effect
   of MsOrg-Exo loaded GelMA in skin wound healing.

Methods

hMSG-ORG culture

   The minor salivary glands samples used in this study were obtained from
   patients undergoing lip reduction surgery. Participants ranged in age
   from 17 to 54 years (N = 13, mean = 35.69). Patients with a history of
   lip inflammation, lip trauma, or autoimmune diseases were excluded. To
   isolate cells from the tissues, the samples were dissected into small
   fragments. The enzymatic digestion process used 0.25% (w/v) collagenase
   type II (Worthington, US) for 1 h and TrypLE Express (Thermo Fisher,
   US) for 10 min. The enzymatic solution was then filtered through a
   70 μm filter. Isolated cells were embedded in Growth Factor-Reduced
   Matrigel (Corning, US) in 24-well plates, with the Matrigel solidifying
   for 20 min at 37 °C. A hMSG-ORG growth medium (HGM) containing the
   essential growth factors and small molecules was added to the wells
   (Supplementary Table [55]1), and the plates were incubated at 37 °C and
   5% carbon dioxide (CO[2]). Media were changed every 3 days. Organoids
   were subcultured every 7 days at a ratio of 1:2 to 1:4, and resuspended
   in cell freezing media (Corning, US) for storage at -80 °C. The
   brightfield images of organoids was visualized by Incucyte S3
   (Sartorius, USA) at fixed intervals.

Cell culture

   To isolate hMSG mesenchymal stem cells (hMSG-MSCs), the hMSG tissues
   were minced into pieces under 1 mm3 and placed at the bottom of 25 cm2
   culture flasks. The culture flask was placed vertically. Then, 3.5 ml
   of DMEM-High medium (HyClone, US) containing 10% FBS was added to the
   flask. After 6 h of incubation at 37 °C and 5% CO2, the culture flask
   was placed horizontally. Media were changed every 3 days. After
   observing the hMSG-MSCs migrating from the tissue under microscope,
   cells were passaged using 0.25% trypsin (HyClone, US) at 80% confluence
   in a 1:3 split ratio. Excessive tissue fragments were filtered out
   during the first passage.

   HUVEC, HaCaT, HSF and RAW 264.7 cell lines were all cultured with the
   aforementioned procedures in DMEM-High medium.

Tissue and organoid histology

   The organoids and tissue were fixed with 4% paraformaldehyde
   (Servicebio, China), dehydrated overnight at 4 °C in 30% sucrose, and
   embedded in paraffin. The embedded samples were then sliced into
   5-µm-thick sections. Hematoxylin and eosin (H&E), periodic acid-Schiff
   (PAS), Masson’s trichrome, and Alcian Blue staining were performed
   under standard protocols.

   For immunofluorescence staining, primary antibodies included E-Cad
   (1:200; Proteintech, US), AQP5 (1:500; Santa Cruz, US), CK5 (1:200;
   ABclonal, China), CK7 (1;200; Proteintech, US), AMY1 (1:800; CST, US).
   The fluorescent secondary antibodies included Alexa Fluor
   594/488-conjugated IgG (1:1000; CST, US). Nuclei were then stained with
   DAPI (CST, US) for 10 min. Images were acquired using a confocal
   microscope (Zeiss, Germany).

   Immunohistochemistry staining used anti-CD31 (1:300; Bioss, China) and
   anti-KI67 (1:200; Abcam, US) antibody. After deparaffinization,
   rehydration and antigen retrieval, the sections were incubated
   overnight at 4 °C with the primary antibodies. The sections were then
   incubated for 30 min at room temperature with an HRP-conjugated
   secondary antibody (ZSGB-bio, China). Diaminobenzidine (DAB) reagent
   (ZSGB-bio, China) was applied for 1 min. Nuclei were counterstained
   with hematoxylin for 30 s. ImageJ software was used to quantify the
   integrated optical density.

Functional swelling assay

   Compounds for the stimulant treatment were prepared in following
   concentration: carbachol 100 µM (Sigma, US), isoproterenol 1 µM (Sigma,
   US), vasoactive intestinal peptide (VIP) 200 nM (Tocris, UK). An
   equivalent volume of DMSO was added for control. Brightfield images of
   the organoids were captured using Incucyte at 5-minute interval over a
   2-hour period.

Isolation and identification of MsOrg-Exo

   The supernatant from the hMSG-ORGs culture was collected each time the
   medium was changed. After collecting a sufficient volume,
   ultracentrifugation was performed for exosome isolation. Specifically,
   the supernatant was centrifuged at 1000 g for 10 min, followed by
   6000 g for 5 min, then 12,000 g for 30 min, and finally,
   ultracentrifuged at 100,000 g for 90 min. The MsOrg-Exo was resuspended
   in PBS and stored at -80 °C.

   The structure of MsOrg-Exo was observed using transmission electron
   microscopy (TEM; JEOL, Japan). The protein amount of MsOrg-Exo was
   further quantified using a BCA protein assay kit (Applygen, China).
   Nanoparticle tracking analysis (NTA) was performed using NanoSight
   (Malvern Panalytical, UK) for particle size measurement. MsOrg-Exo and
   hMSG-ORG of equal protein amount were lysed for western blot analysis
   following standard protocol.

Exosome uptake experiment

   Before experiment, MsOrg-Exo were stained according to the instructions
   of the exosome membrane staining kit. Specifically, DiR dye (50:1) was
   added to an appropriate amount of exosome suspension, and the solution
   was mixed and incubated at 37 °C for 30 min. The solution was then
   transferred to a filter tube and centrifuged at 3000 g until the upper
   layer is free of liquid. The stained exosomes were collected from the
   filter membrane using PBS. For the exosome uptake experiment, HUVEC
   cells were treated with the stained exosomes. After 24 h of treatment,
   the cells were fixed, and immunofluorescence staining was performed
   following the steps described above.

Cell migration assay

   Experiments were conducted using fully confluent HUVEC, HaCaT and HSF
   cells grown in a 6-well plate. A straight scratch was made on the cell
   surface using a 200 µL pipette tip, and the wells were washed three
   times with PBS. Each well was then incubated with either PBS or 500
   ng/µL MsOrg-Exo at 37 °C. Images were captured at 12 and 24 h after
   treatment, and the area of wound was measured using ImageJ software.

Cell proliferation assay

   HUVEC, HaCaT and HSF cells were seeded in 96-well plates at a density
   of 3000 cells/well in DMEM-High basic medium, and then treated with 500
   ng/µL MsOrg-Exo or an equal volume of PBS. The cell area of each well
   was sequentially measured every hour over the following 36 h using the
   Incucyte system.

Tube formation assay

   HUVECs were seeded at a density of 3 × 10⁴ cells per well on
   Matrigel-coated 96-well plates, and then treated with 0, 100, 300, or
   500 µg/mL MsOrg-Exo in DMEM-High medium for 8 h. Brightfield images
   were captured at 4 and 8 h. The number of branches were counted using
   ImageJ software.

Preparation and characterization of GelMA

   Firstly, 0.25% (w/v) solution of lithium
   phenyl-2,4,6-trimethylbenzoylphosphinate (LAP; EFL, China) was prepared
   by dissolving 25 mg of LAP in 10 mL PBS. The LAP solution was mixed
   with dried GelMA (EFL, China) at various concentrations and heated to
   70 °C for complete dissolution. The GelMA solutions were then
   immediately filtered through a preheated 0.22 μm sterile injection
   filter. Next, GelMA hydrogel was formed by crosslinking with 405 nm
   ultraviolet light for 10 s. MsOrg-Exo was added to the GelMA solution
   before crosslinking.

   Scanning electron microscopy (SEM; JEOL, Japan) was used to examine the
   surface morphology of GelMA. After being frozen and lyophilized using a
   freeze dryer, the GelMA samples were taped to a stub and sputter-coated
   with platinum before being introduced into the SEM chamber for imaging.
   To detect the methacrylamide signals, crosslinked GelMA was cut into
   small pieces and dissolved in deuterated reagent. Proton nuclear
   magnetic resonance (H-NMR) was performed using spectrometer (Bruker,
   Germany) to test for the H spectrum. The storage modules and loss
   modules of GelMA before and after crosslinking were measured using
   rheometer (TA, US). GelMA solution pre-mixed with LAP was placed into
   the instrument. 60 s into the testing procedure, the samples were
   irradiated with 405 nm UV light for 20 s, after which the test went on
   for an additional 60 s. The compression test use electronic universal
   testing machine (MTS, China). The appropriate type of compression
   fixture was selected according to the sample size requirements. The
   initial force was set to 0.01 N, and the compression rate was set to
   0.2 mm/min. The test was started, and the equipment recorded the
   variation of force with displacement during the compression test. For
   the degradation experiment, cross-linked GelMA samples were immersed in
   PBS at 37 °C. Every 2 days, three samples were collected and dried
   under vacuum. The weight of the dried samples was measured, and the
   degradation ratios were calculated to construct the degradation curve.

Mouse skin wound model

   Mice (C57BL/6J, male, 7 weeks) were randomly divided into 4 groups
   (N = 5): (1) Blank; (2) 80 µL 10% GelMA; (3) 80 µL 10% GelMA with 500
   ng/µL MsOrg-Exo; (4) 80 µL PBS with 500 ng/µL MsOrg-Exo. After
   anesthesia, two full-thickness skin wounds with a diameter of 8 mm were
   created on the back. The prepared silicone ring was sutured to the skin
   around the wound so that the wound edges aligned with the inner circle
   of the silicone ring, according to Wang et al [[56]27]. For groups 2)
   and 3), the GelMA droplets were crosslinked by irradiating with 405 nm
   UV light for 20 s. For group 4), the MsOrg-Exo solution was injected
   into the dermis from the inner side of the wound. The wounds were
   covered with Tegaderm dressing and wrapped with an elastic bandage.
   Macroscopic assessment of the wounds was performed on days 0, 5, 10,
   and 15. The wound edges were manually segmented using ImageJ. On day
   15, all mice were anesthetized to harvest wound tissue for further
   analysis, and then euthanized using cervical dislocation maneuver.

Exosome controlled-release test

   In vitro, 25 µL GelMA hydrogel containing 500 ng/µL MsOrg-Exo were
   immersed in PBS in a 96-well plate, denoted as GelMA + MsOrg-Exo
   system. The supernatant was collected every day for 15 days and the
   protein content was measured using BCA Protein Assay Kit.

   In vivo, 80 µL of GelMA containing DiR-labeled MsOrg-Exo was applied to
   the wound on the left side of the mouse and immediately crosslinked
   using 405 nm UV light for 20 s. For right-sided wounds, 80 µL of
   MsOrg-Exo solution was injected into the dermis from four directions
   around the wound. Mice were anesthetized and subjected to IVIS imaging
   on day 0 and day 7.

RNA extraction and real-time qPCR

   Total RNA was extracted from tissue using TRIzol reagent (Invitrogen,
   US). The RNA was then reverse transcribed into cDNA using the FastKing
   RT Kit (TIANGEN, China) following the manufacturer’s instructions.
   Quantitative reverse transcription polymerase chain reaction (qRT-PCR)
   was carried out using QuantiNova SYBR Green PCR Master Mix (QIAGEN,
   Germany) on a Bio-Rad CFX system. Primer sequences are listed in
   Supplementary Table [57]2.

Proteomics analysis

   Saliva from healthy individuals (N = 3) and MsOrg-Exo samples (N = 3)
   were collected, and protein quantification was performed using the BCA
   method, followed by standard processing [[58]28]. Briefly, samples were
   mixed with urea buffer in Vivacon tubes and centrifuged. TCEP, IAA, and
   trypsin were then added sequentially, with incubation at the
   corresponding temperatures. After another centrifugation, the peptides
   were collected and reconstituted with ammonium solution. The samples
   were then subjected to elution, chromatographic separation, and mass
   spectrometry analysis using the proteomics analysis system (Thermo
   Scientific, US). Proteomic data were analyzed using MaxQuant software
   and compared with the UniProt database. The top 150 proteins with the
   highest abundance in MsOrg-Exo were selected for GO and KEGG pathway
   enrichment analysis.

Flow cytometry analysis

   The treated RAW 246.7 cells were collected from a 6-well plate using a
   cell scraper. After centrifugation, the cells were resuspended in
   TruStain FcX buffer (BioLegend, US) and incubated for 10 min. Following
   another centrifugation, the cell pellet was resuspended in a mixture of
   cell surface protein fluorescent antibodies diluted 1:100 and incubated
   at room temperature in the dark for 20 min. The cells were then washed
   with cold PBS, resuspended in fixation buffer (Invitrogen, US), and
   incubated at room temperature in the dark for 20 min. The pellet was
   centrifuged and washed three times with permeabilization buffer
   (Invitrogen, US). Subsequently, the intracellular protein fluorescent
   antibodies were diluted 1:50, and the cell pellet was resuspended in
   this mixture and incubated at room temperature in the dark for 60 min.
   Finally, the cells were washed three times with Permeabilization
   buffer, resuspended in an appropriate amount of cold PBS, and analyzed
   by flow cytometry. The primary antibodies used in this study were as
   follows: anti-F4/80-AF488, anti-CD80-BV421, anti-CD86-PE,
   anti-CD206-APC (BioLegend, US).

Opal multiplexed immunofluorescence

   PerkinElmer Opal 7-Color Manual IHC kit (Perkin-Elmer, US) was used to
   multiplexed immunofluorescence staining. According to the protocol,
   after deparaffinization, rehydration and antigen retrieval, the
   sections underwent the following five steps for each marker
   consecutively: (1) blockage with antibody diluent; (2) incubation with
   primary antibody for 1 h; (3) detection using Opal Polymer HRP Ms + Rb
   secondary antibody; (4) visualization using Opal TSA plus agent; (5)
   microwave treatment in citrate buffer. The primary antibodies used for
   each protein were as follows: anti-CD86 (1:500; CST, US), anti-CD206
   (1:500; CST, US) and anti-F4/80 (1:500; CST, US).

Statistical analysis

   The results are presented as mean ± standard error. The data were
   analyzed using GraphPad Prism (Version 9.0.0; US). Unpaired two-tailed
   student’s t-test was used for analysis between two groups. One-way
   ANOVA was used for analysis among multiple groups. P-values < 0.05 were
   considered significant (*P < 0.05, **P < 0.01, ***P < 0.001).

Results

Establishment of hMSG-ORG and exosome isolation

   Single-cell growth was tracked continuously. Formation of small,
   spherical, acinar-like organoids with a diameter of approximately
   100 μm was observed on the third day. Subsequently, the acini gradually
   increased in size, reaching a diameter of 1–2 mm after two weeks of
   culture (Fig. [59]1a).

Fig. 1.

   [60]Fig. 1
   [61]Open in a new tab

   Establishment of hMSG-ORG and exosome isolation. A. Bright-field images
   of primary culture of hMSG-ORG from day 0 to day 13 (scale bar
   = 500 μm) B. H&E / PAS / Alcian Blue staining of hMSG and hMSG-ORG
   showing salivary mucins in hMSG-ORG (scale bar = 100 μm) C.
   Immunofluorescence images of salivary gland markers in hMSG and
   hMSG-ORG (nuclei stained with DAPI; scale bar = 100 μm) D. Bright-field
   images showing acinar swelling in hMSG-ORG after stimulation with
   Carbachol / Isoprenaline / VIP (scale bar = 1 mm) E. TEM images of
   MsOrg-Exo (scale bar = 100 nm) F. Mean particle diameter and peak value
   of MsOrg-Exo measured by NTA G. Western blot analysis of exosomal
   markers CD81, TSG101, and the endoplasmic reticulum marker Calnexin in
   MsOrg-Exo and hMSG-ORG H. Exosomal protein content obtained from equal
   amounts of hMSG-MSC supernatant or hMSG-ORG supernatant measured by BCA
   assay (N = 3)

   Histological staining of hMSG-ORGs showed structural similarity with
   normal hMSG tissue. H&E staining of hMSG-ORGs revealed abundant,
   variably sized luminal structures. Salivary mucins were visible within
   larger lumens, as stained purple-red on PAS staining and blue on Alcian
   Blue staining (Fig. [62]1b). For immunofluorescence staining, the
   acinar markers AMY1 and AQP5 were both expressed in hMSG-ORGs, with
   AMY1 displaying a characteristic ring-like structure around the basal
   surface of the acini, similar to hMSG tissue. The epithelial cell
   marker E-Cadherin was expressed around the periphery of acini and
   lumens in both hMSG and hMSG-ORGs. Additionally, the basal cell marker
   CK5 and luminal cell marker CK7 were expressed. Moreover, the luminal
   cells were surrounded by basal cells in hMSG-ORGs, which is a hallmark
   histologic finding in hMSG tissue (Fig. [63]1c). These findings
   indicated that the hMSG-ORGs cultured in vitro largely mimicked
   salivary gland characteristics.

   To verify whether hMSG-ORGs can mimic salivary gland function in vitro,
   we performed functional swelling assays using a muscarinic receptor
   agonist carbachol, a β-receptor agonist isoproterenol, and a
   non-adrenergic non-cholinergic peptide neurotransmitter VIP. Real-time
   observation showed significant swelling of acinar-like structures in
   hMSG-ORGs within 2 h of drug administration, whereas no significant
   changes were observed in hMSG-ORGs treated with DMSO, indicating that
   hMSG-ORGs can respond to different neurotransmitter stimuli in vitro
   (Fig. [64]1d).

   After the establishment of hMSG-ORGs, the culture supernatant was
   collected for exosome extraction (MsOrg-Exo). TEM scanning of MsOrg-Exo
   exhibited a cup-shaped morphology with a diameter of approximately
   100 nm (Fig. [65]1e). NTA showed that the average particle size of
   MsOrg-Exo was 133.9 ± 67.0 nm (Fig. [66]1f). Exosomal surface proteins
   CD81 and TSG101 were shown to be expressed at higher levels in
   MsOrg-Exo (Fig. [67]1 g). Furthermore, compared to exosomes derived
   from hMSG-MSCs, a significantly increased number of exosomes were
   extracted from an equal volume of supernatant collected from the
   hMSG-ORG culture system (Fig. [68]1 h). These results indicated that
   MsOrg-Exo exhibited the fundamental characteristics of exosomes, and
   had advantages over exosomes extracted from the 2D culture system.

In Vitro effects of MsOrg-Exo on cell migration, proliferation, and
angiogenesis

   We treated HUVEC cells with DiR-labeled MsOrg-Exo, resulting in
   abundant red fluorescent particle accumulation around the cell nuclei,
   which indicated that MsOrg-Exo could be internalized (Fig. [69]2a).
   Subsequently, we treated HUVEC cells, HaCaT cells, and HSF cells with
   MsOrg-Exo. Scratch assay results showed that 12- and 24-hours
   post-treatment, the remaining scratch area in HUVEC, HaCaT, and HSF
   cells was significantly smaller compared to the control group
   (Fig. [70]2b-c). The area of cell growth in HUVEC, HaCaT, and HSF cells
   treated with MsOrg-Exo was larger compared to the control group, at any
   given time point within 36 h post-treatment (Fig. [71]2d). Based on
   these results, we concluded that MsOrg-Exo promoted the migration and
   proliferation of skin cells in vitro.

Fig. 2.

   [72]Fig. 2
   [73]Open in a new tab

   In vitro effects of MsOrg-Exo on cell migration, proliferation, and
   angiogenesis. A. Immunofluorescence images showing the internalization
   of MsOrg-Exo by HUVEC cells (exosomes labeled with DiR, nuclei stained
   with DAPI) B. Scratch assay conducted on HUVEC, HaCaT, and HSF cells
   treated with MsOrg-Exo, with bright-field images captured at 0 h and
   24 h post-treatment (scale bar = 100 μm) C. Statistical analysis of the
   remaining scratch area at 0 h, 12 h, and 24 h (N = 5) D. Proliferation
   assay of HUVEC, HaCaT, and HSF cells treated with MsOrg-Exo, with cell
   growth area continuously tracked by Incucyte over 36 h (N = 3) E.
   Bright-field images of HUVEC cells in the tube formation assay at 4 h
   and 8 h with MsOrg-Exo concentration of 500 µg/ml (scale bar = 100 μm)
   F. Statistical analysis of the number of branches formed at MsOrg-Exo
   concentrations of 0, 100, 300, and 500 µg/ml (N = 5)

   In vitro tube formation assays demonstrated that HUVEC cells began to
   form branching structures approximately 4 h post MsOrg-Exo treatment,
   and more complete tubular structures after 8 h. In contrast, HUVEC
   cells treated with PBS during the same period exhibited almost no
   significant branching or tubular structures (Fig. [74]2e). Statistical
   analysis of tube formation under different concentrations of MsOrg-Exo,
   using the number of branch structures formed after 8 h as the
   observation index, showed significant differences between the control
   group (0 µg/ml) and the 100, 300, and 500 µg/ml groups (Fig. [75]2f).
   These findings suggested that MsOrg-Exo promoted angiogenesis in vitro.

Development of an exosome controlled-release system using GelMA

   Before constructing the controlled release system, the mechanical
   properties of GelMA were first characterized. Upon exposure to 405 nm
   ultraviolet light for 20 s, GelMA transitioned from a highly flowable
   liquid to a solid gel (Fig. [76]3a). SEM images revealed that as the
   concentration of GelMA increased, its micropores slightly decreased in
   size (Fig. [77]3b). H-NMR analysis of the cured GelMA showed two
   characteristic peaks around 5.5 ppm, indicating the presence of two
   methacrylate vinyl groups (Fig. [78]3c). Rheological results
   demonstrated that cross-linked GelMA is an elastic solid with a storage
   modulus significantly higher than the loss modulus, consistent with the
   general properties of hydrogels (Fig. [79]3d). Compression test showed
   that GelMA is easily deformable and resistant to fracturing under
   stress (Fig. [80]3e). Additional degradation experiment proved that
   after 14 days of immersion, GelMA degraded by less than 30%
   (Fig. [81]3f). The mechanical properties of GelMA made it a suitable
   choice as a long-term wound dressing material in animal experiments.

Fig. 3.

   [82]Fig. 3
   [83]Open in a new tab

   Development of an exosome controlled-release system using GelMA. A.
   Appearance of 10% GelMA before and after crosslinking with 405 nm UV
   light for 20 s (left, before crosslinking; right, after crosslinking)
   B. SEM images showing pore sizes of 5% / 10% / 15% GelMA C. NMR
   spectrum of GelMA, with arrows indicating the characteristic peaks of
   methacrylate vinyl groups D. Changes in storage and loss module of 10%
   GelMA before and after photo-crosslinking E. Stress-strain curve of 10%
   GelMA F. Degradation ratio of 10% GelMA in 14 days G. In vitro release
   profiles of MsOrg-Exo loaded in GelMA at different concentrations
   (N = 3) H. 7-day release experiment in mice. The left wound was treated
   with 10% GelMA loaded with MsOrg-Exo, and the right wound was treated
   with an injection of MsOrg-Exo. MsOrg-Exo was pre-labeled with DiR, and
   images were captured using IVIS

   For in vitro exosome controlled-release experiments, MsOrg-Exo in the
   5% GelMA system was fully released in 12 days, while in the 15%
   systems, cumulative release rates were around 70% by day 14. Higher
   concentrations of GelMA prolonged the release duration of exosomes
   (Fig. [84]3e). We subsequently applied the GelMA + MsOrg-Exo system to
   mice for in vivo controlled-release experiments. We tracked the
   sustained release of exosomes using DiR-labeled MsOrg-Exo. After 7
   days, strong fluorescence was still present at the wound site with
   GelMA sustained release, whereas MsOrg-Exo injected subcutaneously on
   the other side had been completely absorbed (Fig. [85]3f). The
   GelMA + MsOrg-Exo system represented a controlled-release system for
   exosomes both in vitro and in vivo.

In vivo promotion of wound healing by MsOrg-Exo

   The normal wound healing in mice primarily relies on skin contraction,
   while human wound healing is mainly achieved through cell proliferation
   and active migration. To better simulate human wound conditions, we
   used silicone rings to fix the wound, thereby minimizing the skin
   contraction in mice.

   On day 5 of administration, there were no noticeable differences in
   healing among the four groups upon visual inspection. By day 10, the
   GelMA + MsOrg-Exo controlled-release system group (hereafter referred
   to as the “controlled-release group”) showed significant granulation
   tissue and fibrous tissue proliferation, with the wound edges migrating
   towards the center. By day 15, the blank group’s wound edges showed no
   substantial proliferation or migration, only slight skin contraction;
   the GelMA control group (hereafter referred to as the “GelMA group”)
   showed a small amount of white fibrous tissue at the wound edges; the
   controlled-release group wounds were nearly completely healed with
   almost no subcutaneous tissue visible and robust hair growth around the
   wound; the healing condition of the group with injected MsOrg-Exo
   (PBS + MsOrg-Exo group, hereafter referred to as the “injection group”)
   was intermediate between the above two groups (Fig. [86]4a).
   Quantitative analysis of wound area showed that on day 10 and day 15,
   the healing rate of the sustained release group was significantly
   higher than that of the other three groups (Fig. [87]4b-c).

Fig. 4.

   [88]Fig. 4
   [89]Open in a new tab

   In vivo promotion of wound healing by MsOrg-Exo. A. Photographs and
   simulated images of mouse wound healing at days 0, 5, 10, and 15
   (silicone ring inner diameter = 9 mm) B-C. Statistical analysis of
   mouse wound healing rates at days 0, 5, 10, and 15 (N = 5)

   On day 15, skin around the mouse wounds was sampled for analysis. H&E
   staining showed a larger area of granulation tissue with new blood
   vessels in the controlled-release and injection groups, whereas the
   blank group’s wound tissue was thinnest (Fig. [90]5a). Masson trichrome
   staining revealed richer blue-stained collagen fibers in the wound
   areas of the controlled-release and injection groups, with the most
   pronounced effect in the controlled-release group (Fig. [91]5b). The
   expression level of related genes in wound skin tissue was quantified
   using qRT-PCR. Compared with the blank group and the GelMA group, the
   controlled-release group showed significantly higher expression levels
   of VEGFα, HGF, bFGF, TGFβ, CK10, and CK19. Compared with the injection
   group, the controlled-release group also showed higher expression
   levels of these genes, though without statistical significance
   (Fig. [92]5c). Immunohistochemical staining showed significantly more
   and denser CD31-positive cells in the controlled-release group,
   indicating a notable promotion of angiogenesis (Fig. [93]5d).
   Additionally, the controlled-release group had more KI67-positive
   nuclei, suggesting a higher number of actively proliferating cells; the
   injection group also showed increased cell proliferation, though
   without statistical significance (Fig. [94]5e).

Fig. 5.

   [95]Fig. 5
   [96]Open in a new tab

   Proteomic analysis of MsOrg-Exo. A. H&E staining of mouse wound at
   postoperative day 15, with blue lines indicating wound width and arrows
   pointing to areas of epidermal discontinuity (scale bar = 500 μm) B.
   Masson staining of mouse wound at postoperative day 15, with
   blue-stained areas indicating collagen fibers (scale bar = 500 μm) C.
   qPCR analysis of expression levels of growth factors VEGFα, HGF, bFGF,
   TGFβ, epidermal differentiation marker CK10, and hair follicle stem
   cell marker CK19 in wound tissue at postoperative day 15 (N = 5) D-E.
   Immunohistochemical staining and semi-quantitative analysis of protein
   expression levels of CD31 and KI67 in wound tissue at postoperative day
   15 (scale bar = 100 μm)

   These results suggested that the GelMA + MsOrg-Exo controlled-release
   system promoted wound healing in mice by regulating growth factor
   expression, angiogenesis, and cell proliferation.

Proteomic analysis of MsOrg-Exo

   Next, we performed proteomic analysis of MsOrg-Exo and compared it with
   the human saliva proteome. The results showed a high overlap between
   the protein species in MsOrg-Exo and human saliva proteins
   (Fig. [97]6a), suggesting that MsOrg-Exo may exert functions similar to
   human saliva to a considerable extent.

Fig. 6.

   [98]Fig. 6
   [99]Open in a new tab

   Proteomic analysis of MsOrg-Exo. A. Venn diagram of proteins in
   MsOrg-Exo and human saliva; 790 proteins identified in MsOrg-Exo, 954
   proteins identified in human saliva samples, with 740 proteins common
   to both B. Differential clustering heatmap of proteins in MsOrg-Exo and
   human saliva (N = 3) C-D. GO and KEGG pathway enrichment analysis of
   the top 150 proteins expressed in MsOrg-Exo

   In differential protein clustering analysis, we identified several
   upregulated proteins in MsOrg-Exo that are related to wound healing,
   including VCL, which mediates the epidermal growth factor pathway;
   COL1A and COL2A, important components of skin fibers; ITGB1, which
   regulates keratinocyte migration; SPARC and THBS1, which promote
   fibroblast proliferation and migration; MAP2K2, related to inflammation
   induction; LAMB2, which promotes endothelial cell adhesion, migration,
   and angiogenesis; and RHOA, which promotes epidermal stem cell
   proliferation and migration (Fig. [100]6b). These upregulated proteins
   potentially mediated the healing process in skin wounds treated with
   MsOrg-Exo.

   In the Biological Process (BP) section of GO analysis, the pathway with
   the most enriched proteins was “wound healing” (16.55%). The subsequent
   categories were “negative regulation of proteolysis” (12.95%),
   “epidermis development” (12.95%), and “humoral immune response”
   (12.23%), all of which are involved in different aspects of wound
   healing regulation (Fig. [101]6c). In KEGG analysis, 16 pathways were
   statistically significant, with the most enriched pathway being “focal
   adhesion” (8.67%), which could assist in the adhesion connection of
   organoids with Matrigel during growth. The next enriched pathways were
   “complement and coagulation cascades” (8%) and the “PI3K-Akt pathway”
   (8%); the former could participate in the inflammatory immune response
   and coagulation during wound healing, while the latter could regulate
   skin cell proliferation and apoptosis, promoting angiogenesis by
   participating in key processes such as protein synthesis, metabolism,
   and the cell cycle (Fig. [102]6d).

Macrophages mediate the regulation of wound healing by MsOrg-Exo

   In the proteomic analysis, we observed that the proteins in MsOrg-Exo
   are highly related to immune responses. Therefore, we attempted to
   treat RAW 264.7 cells with MsOrg-Exo and conducted subsequent
   observations and analyses. Bright field images showed that after 48 h
   of culture, approximately 5% of RAW cells in the control group
   underwent natural differentiation. In contrast, about 60% of RAW cells
   treated with MsOrg-Exo exhibited varying degrees of polarization,
   becoming polygonal or developing pseudopodia (Fig. [103]7a). Further
   flow cytometry analysis revealed that after 24 and 48 h of MsOrg-Exo
   treatment, the expression of M1 and M2 markers in RAW cells was
   significantly increased. CD80 (M1 marker) showed the highest
   proportion, while CD206 (M2 marker) exhibited the greatest fold
   increase, with the degree of increase being positively correlated with
   the treatment duration (Fig. [104]7b). These results demonstrate that
   MsOrg-Exo can promote macrophage polarization in vitro in a short
   period.

Fig. 7.

   [105]Fig. 7
   [106]Open in a new tab

   Macrophages mediate the regulation of wound healing by MsOrg-Exo. A.
   Bright-field images of RAW 264.7 cells treated with MsOrg-Exo at 24 h
   and 48 h, with arrows indicating polygonal polarized RAW cells (scale
   bar = 100 μm) B. Flow cytometry analysis of macrophage polarization
   markers CD80, CD86, and CD206 in RAW cells treated with MsOrg-Exo
   (n = 3) C-D. qPCR analysis of M1 markers CD80, TNFα, IL1β, and M2
   markers IGF1, IL10, IL13 expression levels in wound tissue of mice at
   postoperative day 15 (N = 5) E-F. Opal staining (nuclei stained with
   DAPI) and semi-quantitative analysis (N = 5) of macrophage-related
   proteins CD86, CD206, and F4/80 in wound tissue of mice at
   postoperative day 15

   Next, we performed qRT-PCR analysis on the wound tissue of mice on the
   15th day to verify macrophage polarization in vivo. Among M1-related
   genes, the level of CD80 was significantly downregulated in the
   controlled-release group, while the expression level of the
   pro-inflammatory factor TNFα showed no significant difference between
   the groups. The expression level of IL1b, which promotes M1
   polarization, was relatively lower in the controlled-release group but
   without statistical significance (Fig. [107]7c). Among M2-related
   genes, the expression level of the M2 marker IGF1 and M2 mediator IL10
   were significantly upregulated in the controlled-release group. The
   expression level of IL13, which stimulates M2 polarization, was
   upregulated in both the controlled-release and injection groups but
   without statistical significance (Fig. [108]7d). We then performed Opal
   staining to detect the expression of CD86 and CD206 in the same tissue
   section. The results showed that on the 15th day, there was no
   significant difference in the expression level of CD86 between the
   groups (Fig. [109]7e), while the expression level of CD206 in the
   controlled-release group was significantly higher than in the other
   three groups (Fig. [110]7f). These results indicated that 15 days
   post-administration, the recruitment of M2 macrophages in the
   controlled-release group could be detected within the tissue.

   Therefore, we believe that MsOrg-Exo may assist macrophages in
   transitioning to both M1 and M2 types in the early stages of wound
   healing (within 48 h). In the later stages of wound healing, it assists
   macrophages in transitioning more towards the M2 type, which is in
   accordance with the commonly accepted role of macrophages throughout
   the wound healing process.

Discussion

   In this study, we developed hMSG-ORGs, which can be cultured and
   passaged long-term in vitro. We verified the presence of mucin and the
   localized expression of salivary gland marker proteins such as AMY1,
   AQP5, CK5, and CK7. The MsOrg-Exo isolated from the culture supernatant
   through ultracentrifugation are the first exosomes successfully
   extracted from human salivary gland organoids. Their morphology, size,
   and surface proteins were verified by various methods, meeting the
   general standards of exosomes. In vitro experiments showed that
   MsOrg-Exo could be taken up by cells and promoted the migration and
   proliferation of HUVEC, HaCaT, and HSF cells while enhancing the tube
   formation capability of HUVEC cells. In vivo, MsOrg-Exo incorporated
   into GelMA facilitated wound healing on the backs of mice within 15
   days. Proteomic analysis provided deeper insights into the protein
   composition of MsOrg-Exo, which overlaps significantly with that of
   human saliva, suggesting MsOrg-Exo as a potential alternative to
   saliva. Furthermore, the proteins enriched in MsOrg-Exo positively
   influenced wound healing through various pathways, including
   immune-related ones. Further analysis indicated that MsOrg-Exo might
   promote wound healing by inducing macrophage polarization during
   different stages of the healing process.

   The construction of salivary gland organoids began in 2008 when Clevers
   team used salivary gland stem cells for 3D culture in vitro and
   successfully performed orthotopic transplantation in mice [[111]29]. In
   2015, research group pioneered the isolation and culture of adult minor
   salivary gland epithelial stem/progenitor cells (hMSG-EpiPCs) and
   hMSG-MSCs, and then mixed them for 3D culture to construct functional
   hMSG-ORG [[112]26]. This approach is advantageous as it provides a
   clear cell composition for further research. However, when replicating
   this method, we found that hMSG-EpiPCs quickly aged during planar
   passage in vitro, with most cells failing to persist beyond five
   passages. Thus, we need to explore alternative methods. In 2016, the
   Coppes team reported separating salivary gland cells through enzymatic
   digestion and directly seeding the digested cells into hydrogels for
   organoid culture [[113]30]. Our study used a similar approach. We found
   that cells from digested human minor salivary glands could develop into
   acinar-like structures in Matrigel within 3 days, and they could be
   passaged after about 1 week of culture. In 2022, Yeo-Jun Yoon
   successfully established organoids from the three major human salivary
   glands that could be stably cultured and passaged for four months
   [[114]31]. In our research, the longest cultured hMSG-ORGs retained
   their acinar morphology and could be passaged for about two months,
   potentially due to the structural and functional differences between
   major and minor salivary glands. However, the digestion method allows
   all cell types to be incorporated into the hMSG-ORGs. Our study
   confirmed the differential localization of acinar and duct markers in
   hMSG-ORGs, but further analysis is needed to identify specific subtypes
   or other potential cell types.

   As mentioned earlier, salivary exosomes have been found to promote
   wound healing [[115]16], and they may also play a role in treating
   autoimmune diseases and controlling the progression of oral cancer
   [[116]32]. Furthermore, existing research on salivary gland-derived
   exosomes all involves cultivating salivary gland stem cells, including
   mesenchymal stem cells from the labial glands [[117]33] or parotid
   glands [[118]34], to extract exosomes for the treatment of salivary
   gland injuries or functional disorders. Previous studies have also
   attempted to culture salivary gland-like organoids using dental pulp
   stem cells to extract exosomes [[119]35, [120]36], but no further
   downstream research was conducted. Our study is the first to extract
   exosomes from genuine salivary gland organoids and demonstrate their
   significant role in wound healing.

   Wound healing is another theme of this study. Building appropriate
   animal models is crucial in wound healing research. Rats and mice,
   particularly mice, are widely used due to their easy maintenance,
   straightforward surgical procedures, and abundant research reagents
   [[121]37]. However, the natural wound healing pattern in rodents
   differs from humans. Mouse subcutaneous tissue is primarily composed of
   the panniculus carnosus muscle, which causes rapid wound contraction
   after injury [[122]38]. In contrast, human wound healing relies on
   re-epithelialization and granulation tissue formation without muscle
   involvement. Thus, specialized surgical procedures may be needed to
   reduce skin contraction when constructing skin wound models in mice. In
   some previous studies, researchers ignored this effect and directly
   used the difference between the experimental and control groups to
   demonstrate drug efficacy without additional treatment on mouse dorsal
   skin [[123]39, [124]40]. However, this method has drawbacks: the
   natural healing time for mouse dorsal skin is short, with significant
   wound reduction observed three days post-operation and complete healing
   within seven days, making it challenging to observe changes promptly.
   In this study, we primarily referred to the mouse wound model protocol
   proposed by Wang et al. in 2013 [[125]27], using silicone rings to
   restrict skin contraction. This method is widely used in studies
   investigating wound healing in mice [[126]41, [127]42].

   Different phenotypes of macrophages are recruited at different stages
   of wound healing. After tissue injury, once monocytes enter the wound,
   bacterial LPS and pro-inflammatory cytokines from tissue cells trigger
   their differentiation into M1-like macrophages in the early stage of
   wound healing, clearing pathogens and cellular debris [[128]43,
   [129]44]. In the later stage of healing, IL-4 and IL-13 from tissue
   cells stimulate macrophages to switch to the M2 type, releasing factors
   like VEGF and TGF-β, which promote cell growth, collagen production,
   and angiogenesis [[130]45, [131]46]. In chronic wounds,
   pro-inflammatory M1 macrophages persist and fail to transition to the
   anti-inflammatory M2 phenotype, leading to delayed healing [[132]47].
   In our study, MsOrg-Exo promoted the short-term polarization of RAW
   cells to M1 in vitro and helped accelerate the polarization of
   macrophages to M2 in vivo for faster wound healing in mice. However,
   due to limited mice availability, we conducted only a single round of
   sampling. At the time of sampling, all four wound groups had entered
   the proliferation phase. Thus, we could not verify the effect of
   MsOrg-Exo in vivo during the early stages of wound healing
   (inflammation stage), particularly on macrophages. Future studies
   should include more groups for comparison of histological and molecular
   biology changes at different healing stages.

   The human body contains a large number of minor salivary glands located
   superficially [[133]24], making them more suitable for research and
   clinical studies than the three major salivary glands. Our research
   utilized hMSG to construct engineered hMSG-ORGs and demonstrated the
   pro-wound healing effects of MsOrg-Exo both in vitro and in vivo. This
   breakthrough in the application of salivary gland organoids provides a
   strong theoretical basis for future clinical research on MsOrg-Exo in
   regenerative medicine. Further research is needed to explore the
   potential clinical use of autologous or allogeneic MsOrg-Exo for
   treatment or to combine MsOrg-Exo with new biocompatible materials for
   dressings, or to use MsOrg-Exo as an alternative to stem cell therapy
   for salivary gland injury.

Conclusions

   Our study established hMSG-ORG to produce MsOrg-Exo, which
   significantly enhanced cell proliferation, migration, and angiogenesis
   in vitro. When combined with GelMA, control-released MsOrg-Exo improved
   wound healing in vivo by promoting collagen synthesis, angiogenesis and
   growth factor secretion. Proteomic analysis revealed that MsOrg-Exo is
   enriched in proteins essential for wound repair. Additionally,
   MsOrg-Exo modulated macrophage polarization, shifting towards M2
   phenotype in long-term treatment.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [134]Supplementary Material 1^ (94.2KB, pdf)

Acknowledgements