miR-138 from Adipose-Derived Stem-Cell Exosomes Accelerates Wound Healing in Diabetic Rats Through Targeting SIRT1/PTEN Signaling to Promote Angiogenesis and Fibrosis


 Background: Exosome (Exo) secretion by adipose-derived mesenchymal stem cells (ADSCs) promotes cutaneous wound healing by transfer of bioactive molecules. miR-138 is a micro (mi)RNA that stimulates endothelial progenitor cells and promotes proliferation and locomotion of human scar fibroblasts (HSFs). However, the underlying molecular mechanism is unclear. Methods: miR-138 activity in Exo-mediated healing was investigated by subcutaneous injection of exosomes isolated with ultracentrifugation from control ADSCs or ADSCs expressing miR-138 into full-thickness skin wounds in a rat diabetes model. Wound healing was evaluated by wound closure rate, histological evaluation, and immunofluorescence staining. Cultures of HSFs and human mammary epithelial cells (HMECs) were treated with Exos from wild-type and miR-138-modified ADSCs under high glucose conditions (HG). Cell proliferation and apoptosis were measured with cell counting kit (CCK)-8 assay and flow cytometry. Cell migration was assayed in Transwell chambers, and the effect of miR-138 on Exo-mediated angiogenesis and protein expression were evaluated. The miR-138 target was identified with a luciferase reporter assay. Results: ADSC-Exos were incorporated by endothelial cells and HSFs, and miR-138 enhanced cell proliferation and migration and suppressed apoptosis under HG conditions. Exos promoted endothelial tubule formation by HSFs, and western blotting showed that miR-138 mediated the therapeutic effects by reversing SIRT1/PTEN-mediated PI3K/Akt and ERK1/2 signaling inhibition. The luciferase reporter assay confirmed that miR-138 interacted with the 3'-UTR of SIRT1 to suppress SIRT1 mRNA expression. Injection of miR-138-enriched Exos into skin wounds accelerated re-epithelialization, reduced scar width, and enhanced angiogenesis. Conclusions: In conclusion, miR-138-enriched Exos promoted wound healing in a rat diabetes model by targeting SIRT1, and their effects in promoting soft tissue wound healing warrant further study.


Introduction
Dermatological diseases occur frequently in patients with diabetes and can manifest as fungal and bacterial and infections, granuloma annulare, diabetic dermopathy, or necrobiosis lipoidica diabeticorum [1,2]. Clinical applications of regenerative treatments with growth factors and other cell therapies have been successful in treating diabetes-associated skin injuries, including diabetic foot ulcers [3].
Mesenchymal stem cells (MSCs) in normal skin take part in wound repair, and adipose-derived mesenchymal stem cells (ADSCs) have been veri ed to improve wound healing in diabetic rats [4][5][6]. Angiogenesis has crucial functions in wound healing, and chronic wounds in diabetes are characterized by decreased angiogenesis [7]. We have previously reported that ADSCs exosomes enhance wound healing and prevent injury to cells transplanted into a high glucose environment [8].
Exosomes are extracellular vesicles of 30-150 nm that are secreted into the extracellular milieu by a majority of cell types. They have been investigated in uids including blood, semen, urine, and the tissue extracellular matrix. They carry cargo, which participate in mediating interactions among organs and cells [9,10]. Exosomes have physiological activities relevant in both health and disease [11,12]. Exosomes have been found to be paracrine mediators that induce healing by transferring bioactive molecules, including micro (mi)RNAs, proteins and other compounds [13,14]. The ADSCs protective effects in wound healing have been suggested to be modulated by exosomes. ADSC exosomes can deliver miRNAs, which then ameliorate the high glucose environment in diabetes mellitus.
miRNAs are 19-25 nucleotide noncoding RNAs that suppress post-transcriptional gene expression through imperfect complementary sequence pairing at their target genes' 3′ untranslated regions (3′UTRs), thereby causing translational repression or mRNA degradation [15,16]. miRNAs have been demonstrated to participate in regulating 33.3% of all human genes, and thus, are involved in many biological functions [17,18]. Several studies have suggested that miRNAs regulate endothelial progenitor cell functions and may be novel therapeutic options for treating dermatological diseases.
SIRT1 is a silent information regulator 2 (Sir2) mammalian homolog that modulates wound healing. In diabetes [24,25], SIRT1 de ciency inhibits PTEN expression, which enhances the PI3K/Akt and ERK signaling activation [26]. ERK and PI3K/Akt signaling have been veri ed to promote cutaneous wound healing [27][28][29]. We established that the exosomes secreted by miR-138-modi ed ADSCs have therapeutic effects in angiogenesis and broblast function, in vitro, under high-glucose conditions and on cutaneous wound healing in a rat diabetes model. miR-138's activity in promoting exosome-mediated wound healing was investigated.

Animals and ethics statement
The Animal Care and Utilization Committee of the Shanghai Tenth People's Hospital approved all animal procedures. We obtained male Sprague-Dawley female rats at eight weeks of age from the SLAC Laboratory Animal Co. Ltd, Shanghai, China, and individually housed them in independent ventilated cages under 24 °C to 26 °C under constant humidity with a 12-hour light/dark cycle. All procedures were approved by the Ethics Committee of the the Shanghai Tenth People's Hospital, Shanghai, and were conducted by following the guidelines. We conducted surgical processes under anesthesia and made every effort to minimize suffering. We anesthetized rats by intraperitoneal injection of 30 mg/kg sodium pentobarbital before sacri ce and the rats were sacri ce by spinal dislocation.
ADSC isolation, culture, and identi cation Brie y, we harvested adipose tissue from normal rats or healthy humans. The Ethics Committee of the Shanghai Tenth People's Hospital, Shanghai approved all procedures (Human20170621 for humans, Animal20170513 for animals). We conducted procedures by following the guidelines. We conducted surgical processes under anesthesia and made every effort to minimize suffering. We washed tissues with phosphate-buffered saline (PBS) and minced them before digestion with 0.2% collagenase I (Sigma-Aldrich, St. Louis, USA) for 1 hour at 37 °C with intermittent shaking. We washed the digested tissue with Dulbecco's modi ed Eagle's medium (DMEM) (Sigma-Aldrich, St. Louis, USA) containing 15% FBS (Gibco BRL, Frederick, USA) and centrifuged the samples at 1000 rpm for 10 minutes to remove mature adipocytes. We resuspended the pellets in DMEM with 15% FBS, with penicillin (100 U/ml) and streptomycin (100 μg/ml), and cultured cells at 37 °C and 5% CO 2 . We detached ADSCs reaching 80%-90% con uence with 0.02% ethylenediaminetetraacetic acid/0.25% trypsin (Sigma-Aldrich, St. Louis, USA) for 5 minutes at room temperature and replated them. We utilized uorescein isothiocyanate (FITC)-or phycoerythrin (PE)conjugated antibodies against CD29, CD44, CD90, CD105, and von Willebrand factor (vWF) for phenotypic analysis. We used IgG-matched isotype as the internal control for each antibody. We grew normoxic ADSC cultures in 95% air (20% O 2 ) and 5% CO 2 .

Multilineage differentiation of ADSCs
To validate ADSC multilineage differentiation, we cultured third-passage mouse ADSCs in adipogenic differentiation medium (Sigma-Aldrich, St. Louis, USA) and stained them with oil red O after 2 weeks, or cultured them in osteogenic differentiation medium (Sigma-Aldrich, St. Louis, USA) and stained them with alizarin red after 3 weeks.

ADSC-derived exosome isolation and identi cation
After reaching 80%-90% con uence, ADSCs were rinsed with PBS and cultured in FBS-free endothelial cell growth medium-2MV, supplemented with 1× serum replacement solution for another 2 days. We removed conditioned culture medium and centrifuged the cells at 300 g for 10 minutes and at another 2000 g for 10 minutes to remove cellular debris and apoptotic cells. After centrifugation at 10,000 g for 0.5 hour, we ltered the supernatant (0.22 μm, Millipore, Billerica, USA), transferred 15 mL of supernatant to an Amicon Ultra-15 Centrifugal Filter Unit (100 kDa, Millipore) and centrifuged it at 4000 g to concentrate the sample to ~1 ml. We washed the ultra ltration unit twice with PBS and ltered the samples again at 4000 g to achieve a 1 ml volume. We added a 20% volume of Exoquick exosome precipitation solution (System Biosciences, USA) to the ultra ltered liquid and mixed the sample through inversion. After incubation for half a day, we centrifuged the mixture at 1500 g for 0.5 hour and aspirated the supernatant. We resuspended the exosome pellets in 500 μL PBS. We performed all procedures at 4°C. We determined the exosome protein content with a Pierce bicinchoninic acid Protein Assay Kit (Thermo Fisher Scienti c, USA). We stored exosomes at −80 °C until use. We utilized western blotting and transmission electron microscopy to analyze the exosomes. We determined size by dynamic light scattering with a Nanosizer. The size distribution was plotted with the particle radius (nm) on the X-axis and the percentage on the Y-axis.

Cell culture
We cultured HSFs and HEK293T (FuHeng Biology, Shanghai, China) in high glucose DMEM (Gibco BRL, Grand Island, USA) with 10% FBS. We cultured human mammary endothelial cells (HMECs, Cell Bank of the Chinese Academy of Sciences, Shanghai, China) in MCDB131 medium (Gibco BRL) containing 10% FBS, 2 mM L-glutamine (Sigma-Aldrich), epidermal growth factor (Sigma-Aldrich, 10 ng/ml) and hydrocortisone (1 μg/ml, Sigma-Aldrich). We maintained cells under 37 °C and 5% CO 2 in a humidi ed environment. HMECs and HSFs were stimulated with 5.5 mM or 30 mM glucose for 1 day, and 100 μg/ml of exosomes was added to the cultures to assess protection against high glucose injury.

Exosome labeling and uptake
We used a PKH67 uorescent linker kit (Sigma-Aldrich, St. Louis, USA) to label exosomes. We added PKH67 dye (400 μL) to exosome suspensions and incubated the samples for 5 minutes at room temperature. We added an identical volume of exosome-depleted bovine serum albumin to stop the reaction, and washed exosomes twice with PBS to remove any unbound dye. We incubated HSFs or HMECs with exosomes labeled for 3 hours, xed and stained samples with diamidinophenylindole (DAPI), and observed and photographed them under a confocal microscope.

RNA interference or overexpression
We purchased miR-138 inhibitors from RiboBio (Guangzhou, China), and conducted transfection according to standard procedures. In brief, we transferred cells to culture plates with six wells and transfected them through incubation in complete medium containing ADSC-exosomes (200 μg/well, at 100 μg/ml) and Lipofectamine 2000 (Thermo Scienti c), or an equivalent volume of PBS for 1 day. For miR-138 overexpression, ADSCs were transfected with an miR-138 mimic (5′-AGCGUGUGUUGUGAAUCAGGCCG-3′) synthesized by GenePharma (Shanghai, China) by Lipofectamine 2000, as described previously.

Flow cytometry
We used ow cytometry to assay HMECs apoptosis after FITC-conjugated annexin V and propidium iodide (PI) staining. We washed cells twice before adjusting them to 1×10 6 cells per ml in cold D-Hanks buffer. We added annexin V-FITC (10 μl) and PI (10 μl) to 100 μL cell suspension and incubated them for 15 minutes at room temperature in the dark. Before analysis, we added 400 µl binding buffer to each sample without washing. We performed each assay in at least triplicate.

Tubule formation assay
In vitro neovascularization was assayed in matrices of human brin. After treatment, we seeded serum-starved HMECs in endothelial basal medium onto plates coated with Matrigel (10 5 cells per well into six well plates) (BD Biosciences, Franklin Lakes, USA) and incubated them at 37°C for 12 hours. We observed tubular structures that formed in the Matrigel and photographed under phase-contrast microscopy, and the newly formed tube lengths in ten randomly selected elds per well were measured.

Transwell assay
We treated cells with exosomes, miR-138-exosomes, or miR-138 inhibitor under HG conditions. After 24 h, we starved the cells in serum-free medium for another 12 hours, performed trypsin digestion, and seeded 1×10 5 cells in the top chambers of 24 well Transwell culture inserts (Promega, Fitchburg, WI, USA).
Medium with 20% serum was used as a chemoattractant. We xed the cells for 10 min with formalin of 4% after 24 h culture.

Cell counting kit (CCK)-8 assay
We assayed HSF and HMEC proliferation with a CCK-8 kit (BD Biosciences, Franklin Lakes, USA) by using standard procedures. We cultured transfected cells in 96-well plates with exosomes under HG conditions for 24 h before adding 90 μL fresh culture medium and 10 μl CCK-8 reagent. We detected the absorbance at 450 nm with a microplate reader after incubation at 37°C for 2 hours.

Luciferase reporter assay
The 3′-UTR target sequence for miR-138 miRNA in the SIRT1 gene was predicted with the argetScan online tool. Wild-type and 3′-UTR mutant SIRT1 were performed and cloned into the pMIR re y luciferase-expressing vector. We cotransfected HEK293T cells at 70% con uence with 500 ng of pMIR-SIRT1wt/pMIR-SIRT1-Mut and 50 nM of miR-138 mimics with a Lipofectamine 2000 transfection kit (Thermo Scienti c) for the luciferase assays. We assayed luciferase activity with the Dual-Luciferase Reporter System (Promega). We independently performed ve assays.

Rat diabetes skin wound model
We induced diabetes in rats through a single intraperitoneal injection of 100 mg/kg in 0.01 M pH 4.3 sodium citrate. We measured the blood glucose daily and controlled it to between 16.7 and 33.3 mmol/l through administration of insulin at 6-18 U/day (Wan-Bang Biochemical Medicine Co. Ltd, Xuzhou, China). After 1 month, we established the subcutaneous wound model after iso urane inhalation anesthesia (2%). We made a single round full-thickness skin wound on the dorsal hind foot with a disposable 5 mm skin biopsy punch and Westcott scissors. Eighteen rats were allocated randomly to subcutaneously injection of 200 μg ADSC-exosomes in PBS of 100 μl or the same volume of PBS at four sites around the wound (25 μl per site). The rats were killed after 15 days, and skin specimens were harvested for histopathological evaluation. Each group have 6 rats. We anesthetized rats by intraperitoneal injection of 30 mg/kg sodium pentobarbital before sacri ce and the rats were sacri ce by spinal dislocation.

Measurement of wound contraction
On days 0, 7, and 15 before wound harvest, the sizes of the treated and control wounds were measured with a ruler and photographed (DMC-LX5GK, Panasonic, Japan). We calculated the ulcer area with Image-Pro Plus 4.5 (http://www.mediacy.com/imageproplus).

Immunohistochemistry and immuno uorescenceassays
We xed skin tissue samples in 10% formalin solution, embedded them in para n, and sectioned them at 5 μm. We stained tissue sections with Masson's trichrome for histological evaluation. We performed immuno uorescence staining of CD31 to measure histopathological alterations in angiogenesis and used terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) to detect apoptotic cells. We examined sections with a uorescence microscope (Nikon, Tokyo, Japan) or an Axiophot light microscope (Zeiss, Oberkochen, Germany) and photographed them with a digital camera.

Statistical analysis
We denoted continuous variables as means ± SD (standard deviation). We performed one-way analysis of variance for comparisons in GraphPad Prism (GraphPad, La Jolla, USA). A p-value ≤ 0.05 was considered to indicate a signi cant difference.

Characterization of ADSC-exosomes
We previously suggested that exosomes participate in mesenchymal stem cell-mediated wound healing in a diabetes foot ulcer model [7]. Exosome transfer of erythroid 2-related factor 2 (Nrf2) nuclear factor accelerated the cutaneous wound healing by promoting vascularization. miR-138 has been reported to enhance endothelial progenitor cell function in a mouse venous thrombosis model [20]. The involvement of miR-138 in exosome-mediated wound healing was con rmed in ADSCs isolated from human cell lines and rat adipose tissue. The isolated ADSCs had a typical cobblestone-like morphology (Fig. 1A). Immuno uorescence staining was positive for the expression of surface mesenchymal cell markers including CD29, CD90, CD44, and CD105, but negative for the endothelial markers CD34 and vWF (Fig. 1B-H). The oil red O and alizarin red staining data veri ed adipocyte differentiation and osteoblast differentiation (Fig. 1I, J).
Dynamic light scattering showed that the particle diameters ranged from 50 to 120 nm (Fig. 1K), in agreement with the distributions of exosome size reported in [30]. The ADSC-exosomes had cup-or sphere-shaped morphologies in transmission electron micrographs (Fig. 1L), similar to previously described exosomes [30]. The expression of exosome marker proteins such as CD81, CD63, CD31, and CD9 in ADSC-exosomes was con rmed with western blotting (Fig.   1M). The data demonstrated that both ADSC-exosomes and cellular components expressed CD63, CD81 CD31, and CD9, and that the nanoparticles were exosomes.

ADSC-exosomes transfer miR-138 into broblasts and endothelial cells
ADSCs overexpressing miR-138 were constructed to investigate the exosome-mediated promotion of wound healing. qPCR analysis con rmed miR-138 overexpression in the transformed ADSCs, as compared with wild-type ADSCs ( Fig. 2A). To determine whether ADSC-exosomes were internalized by endothelial cells and broblasts, a prerequisite for subsequent exosomal miRNA transfer, we incubated ADSC-exosomes labeled with the green uorescent dye PKH67 with HSFs and HMECs, and washed them to remove unbound exosomes, and xed them. We stained DNA with DAPI. Fluorescence microscopy demonstrated that PKH67-labeled exosomes entered into HSFs and HMECs (Fig. 2B), and qPCR con rmed the presence of miR-138 in the recipient cells after culture with ADSC-exosomes. miR-138 expression increased in both HSFs and HMECs after treatment with exosomes, and in ADSCs transfected with miR-138, as compared with wild-type ADSCs (Fig. 2C, D). qPCR con rmed that binding of miR-138 to the SIRT1 3′-UTR downregulated the mRNA and protein expression of SIRTI and its downstream gene PTEN in both HSFs and HMECs after exosome treatment, compared with controls ( Fig. 2E-H).
miR-138 is involved in ADSC-exosome-mediated angiogenesis, migration, and inhibition of apoptosis of HMECs in high-glucose cultures HMECs were cultured in medium with 30 mM glucose and PBS, exosomes, miR-138-exosomes, or miR-138-exosomes plus miR-138 inhibitor for 24 h. The Transwell assays indicated that high glucose inhibited HMEC migration, but exosome treatment reversed the effect. miR-138 overexpression in exosomes increased the recovery of HMEC migration, but miR-138 inhibitor treatment signi cantly inhibited the protective effect of exosomes (Fig. 3A, B). Flow cytometry analysis revealed that miR-138-exosome treatment signi cantly suppressed the apoptosis of HMECs under high glucose and that miR-138 inhibitor treatment reversed the protective effect of exosomes (Fig. 3C, D). Thus, miR-138 appears to have a protective effect against high glucose induced apoptosis in HMECs. The formation of tubules by HMECs on Matrigel-coated culture wells was used as an in vitro model of angiogenesis, as assessed by the number of branches that formed. miR-138-exosomes reversed the inhibition of angiogenesis in response to high glucose, an effect that was blocked by miR-138 inhibitor treatment (Fig. 3E, F). The co-culture assay between ADSCs and HMECs found that miR-138-modi ed ADSCs promotion angiogenesis of HMECs under HG condition, but after added the ADSCs exosome inhibitor, the promotion effect was reversed (Fig. S1). Suggestion that exosomes play a mediating role.
qPCR indicated that miR-138-exosome treatment suppressed the SIRT1 and PTEN mRNA expression due to high glucose and that cotreatment with the miR-138 inhibitor reversed the miR-138-exosome effect (Fig. 3H, I). The co-culture assay between ADSCs and HMECs found that miR-138-modi ed ADSCs promotion angiogenesis of HMECs under HG condition, but after added the ADSCs exosome inhibit the promotion effect was reversed (Fig. 3J), suggestion that the angiogenesis promotion effect of miR-138 was ADSC-exosome mediated. The in vitro functional assays in HMEC cultures indicated that miR-138 is involved in ADSC-exosome-mediated activation and the promotion of angiogenesis, cell migration, and activation, as well as the inhibition of apoptosis, possibly by targeting SIRT1/PTEN signaling.
The effects of ADSC-exosomes on HSFs were evaluated in cultures with 30 mM glucose, which were treated with PBS, exosomes, miR-138-exosomes, or miR-138-exosomes plus miR-138 inhibitor for 24 h. High glucose inhibited HFS migration in Transwell chambers, an effect reversed by exosome treatment. miR-138 overexpression increased the recovery of HSF migration, although the effect was largely reversed by miR-138 inhibitor treatment (Fig. 4A, B). Flow cytometry analysis indicated that miR-138-exosome treatment protected against high glucose-induced HSF apoptosis and that miR-138 inhibitor treatment reversed the protective effect (Fig. 4C, D). The CCK-8 assays indicated that high glucose conditions inhibited cell proliferation and that treatment with exosomes and 138-exosomes enhanced the proliferative activity of HSFs cultured in high glucose medium. Cotreatment with miR-138 inhibitor interfered with the effects of exosome treatment (Fig. 4E). The qPCR results showed that miR-138-exosome treatment suppressed the expression of SIRT1 and PTEN mRNA occurred due to high glucose, and that cotreatment with miR-138 inhibitor reversed the effect (Fig. 4F, G). In vitro functional assays of HSF cultures indicated that miR-138 was involved in ADSC-exosome-mediated activation, the promotion of cell migration, and apoptosis inhibition, possibly through targeting SIRT1/PTEN signaling. The RT-qPCR detection also found that miR-138 expression in ADSCs were decreased after transfected with inhibitor, but increased after transfected with mimic. The result also show that miR-138 overexpression inhibit SIRT1 and PTEN expression. Inhibit miR-138 expression promotion SIRT1 and PTEN expression (Fig. 4H-M).

Exosomes from miR-138-modi ed ADSCs accelerate cutaneous wound healing in diabetic rats
We investigated ADSC-exosomes' effects on wound healing in full-thickness cutaneous wounds in the feet of rats with streptozotocin-induced diabetes in response to subcutaneous injection of equivalent volumes of exosomes, miR-138-exosomes, or PBS exosome diluent. Wound closure was signi cantly accelerated by miR-138 exosomes, compared with the control PBS treatment. The wounds treated with miR-138-exosomes were nearly closed on day 14, but large scar areas were visible in control wounds (Fig. 7A, B). Masson-stained tissue from wounds treated with miR-138-exosomes included more collagen bers than tissue from controls after 14 days (Fig. 7C, D). CD31 staining indicated that microvascular development was signi cantly more extensive after treatment with exosomes and miR-138-exosomes than after the control treatment (Fig. 7E, F). TUNEL staining indicated that miR-138-exosome treatment signi cantly suppressed apoptosis of skin tissue, as compared with control treatment (Fig.7G, H). Western blotting con rmed that miR-138-exosome treatment strongly suppressed apoptosis protein expression, including that of caspase-3 and Bax (Fig. 7I-K), and SIRT1 and PTEN (Fig. 7L-N), as compared with control treatment.

Discussion
This study provides the rst demonstration of miR-138 activity in ADSC-exosome-mediated wound healing. The local transplantation of exosomes from ADSCs overexpressing miR-138 into skin wounds shortened the time of wound closure and increased the rates of collagen deposition, re-epithelialization, and new blood vessel formation. Scar formation was decreased. ADSC-exosomes were internalized, and miR-138 was transferred to both broblasts and endothelial cells. The upregulation of miR-138 expression activated and enhanced the migration of both broblasts and endothelial cells and suppressed apoptosis after exposure to high glucose culture conditions. Previous studies have found that up-regulation of miR-138 inhibits hypoxia-induced cardiomyocyte apoptosis via down-regulating lipocalin-2 expression [35]. miR-138 protects cardiomyocytes from hypoxia-induced apoptosis via MLK3/JNK/cjun pathway [36]. In this study, we found that transfer of miR-138 to endothelial cells by ADSC-exosomes enhanced cultured endothelial cell angiogenesis.
Bioinformatics and a luciferase reporter assay con rmed that miR-138 interacted with SIRT1 and decreased the expression of SIRT1 mRNA and protein, which were associated with decreased PTEN expression. The observed inhibitory effect of SIRT1/PTEN on ERK1/2 and PI3K/Akt signaling is consistent with ndings from previous reports [26].
Exosomes in human umbilical cord blood-derived epithelial progenitor cells (hucMSC-exo) have been found to promote wound healing in rats with streptozotocin-induced diabetes through enhancing angiogenesis via Erk1/2 signaling [27]. Akt pathway activation by hucMSC-Exo is also associated with a heat stress-induced decrease in apoptosis in a rat skin burn model [37]. PTEN is a PI3K/Akt pathway negative regulator that acts by dephosphorylating PIP3 (phosphatidylinositol 5, 4, 3 phosphate) to PIP2 (phosphatidylinositol 5, 4 phosphate) [38,39]. It suppresses the activation of ERK1/2 [40,41], and PTEN overexpression promotes apoptosis under stress conditions [42]. SIRT1 regulation of PTEN acetylation controls its localization and activity due to cellular damage and intracellular stress [26,33]. In our study, miR-138's effects on exosome-mediated wound healing were associated with inhibition of SIRT1 expression, and the downregulation expression of SIRT1 suppressed PTEN expression, but promotion phosphorylation of Akt and ERK1/2. The activation of PI3K/Akt and ERK1/2 signaling reversed HG induced cell apoptosis (Figure. 7O). Our study also found that miR-138 from ADSCs exosomes accelerates wound healing in diabetic rats by promote angiogenesis. miR-183 overexpression activated the PI3K/Akt pathway which led to upregulation of downstream target genes including VEGF and CD34. Therefore, growth and angiogenesis of endothelial cells were promoted [43].
Exosomes are paracrine mediators that transfer proteins and genetic material to target cells. Exogenous exosomal molecules might regulate target protein or gene expression as well as recipient cell function [44,45]. For instance, colorectal cancer exosomes are used to transfer mRNAs that enrich endothelial cell proliferation and facilitate angiogenesis [46]. Exosomes derived from human umbilical cord MSCs have been demonstrated to transfer miR-181c and to attenuate burn-induced in ammation [47]. Human umbilical cord blood exosomes have been reported to enhance cutaneous wound healing via miR-21-3pmediated promotion regarding broblast function and angiogenesis [30].

Conclusions
In this study, miR-138 transferred by ADSC-exosomes mediated wound healing primarily through preventing cell death. The wound healing promoted by exosomes from ADSCs overexpressing miR-138 suggests that exosome transplantation may have potential clinical applications in diabetes-induced skin injury treatment.

Declarations
Availability of data and materials Availability of data and materials can be assessed both in the "Methods" section, the "Results" section.

Consent for publication
Not applicable.
Author Contributions XL, WZ and MZ performed the experiments and analyzed the data, JH and YS conceived experiments and revised the draft, HY and ML conceived the studies and drafted the manuscript with feedback from all authors. All authors read and approved the nal version.