PTH Derivative Promotes Chronic Wound Healing via Synergistic Multicellular Stimulating and Exosomal Activities

Background: Chronic diabetic wounds are a disturbing and rapidly growing clinical problem. Parathyroid hormone related peptide (PTHrP-2) was assumed as multifunctional factor in angiogenesis, brogenesis and re-epithelization. This study aims to test PTHrP-2 eciency and mechanism in chronic wound healing. Methods: Through repair phenomenon in vivo some problems were detected, and further research on their mechanisms was made. In vivo therapeutic effects of PTHrP-2 was determined by HE, Masson, microl and immunohistochemical staining. In vitro direct effects of PTHrP-2 was determined by proliferation, migration, Vascular Endothelial Grown Factor and collagen I secretion of cells and Akt/ Erk1/2 pathway change. In vitro indirect effects of PTHrP-2 was study via exosomes. Exosomes from PTHrP-2 untreated and treated HUVECs and HFF-1 cells were insolated and identied. Exosomes were co-cultured with original cells, HUVECs or HFF-1 cells, and epithelial cells. Proliferation and migration and pathway change were observed. PTHrP-2-HUVEC-Exos was added into in vivo wound to testify its hub role in PTHrP-2 indirect effects in wound healing. Results: In vivo, PTHrP-2 exerted multifunctional pro-angiogenesis, pro-rbogenesis and re-epithelization effects. In vitro, PTHrP-2 promoted proliferation and migration of endothelial and broblast cells, but had no effect on epithelial cells. Therefore, we tested PTHrP-2 indirect effects via exosomes. PTHrP-2 intensied intercellular communication between endothelial cells and broblasts and initiated endothelial-epithelial intercellular communication. PTHrP-2-HUVEC-Exos played hub role in PTHrP-2 indirect effects in wound healing. Conclusion: The ndings of this study indicate that PTHrP-2, a multifunctional factor, can promote chronic wound healing via synergistic multicellular stimulating and exosomal activities. sulfate-polyacrylamide gel


Background
With the improvement of people's living standards and the acceleration of urbanization, people's diet and lifestyle have changed, leading to an increasing incidence of diabetes worldwide (1). Diabetic chronic wound is one of the most common complications of diabetes which costs billions of dollars, and causes a high socioeconomic burden (2). Traditional diabetic chronic wound therapies include dressing changes, repeated debridement and amputation (3). Regenerative approaches, such as approaches using cytokines and hydrogels for tissue engineering are being developed and hold promise for chronic wound healing (4,5). Currently, cytokines such as epithelial cell growth factor (EGF) and basic broblast growth factor (b-FGF) are widely used in clinic and promote healing(6, 7). However, these cytokines are often expensive and have strict limitations regarding use currently used factors exert their healing abilities in either the dermal layer or corium layer of chronic wounds(8).
Skin tissue contains two layers, the dermal layer and corium layer (9), and is composed of three major cell types: epithelial cells, brocytes and endothelial cells (10)(11)(12). Fibrocytes, bers, endothelial cells and capillaries construct the corium layer, whereas epithelial cells compose the dermal layer. Three cell types are closely related in anatomic and functional aspects. Collagens form the major scaffold for the corium layer are important in wound healing (11). Endothelial cells and epithelial cells attach within or on the scaffold and form capillaries and the dermal layer. The capillaries supply oxygen and nutrients to the brocytes and epithelial cells, whereas the dermal layer protects the corium layer form bacteria and radiation (13). The three cell types communicate and interact with one another via paracrine activity (14).
Therefore, an optimal approach might involve the targeting of all three cell types to promote wound healing via synergistic effects.
PTH, a hormone secreted by the parathyroid, is an important factor in bone turnover and calcium phosphorus metabolism (15,16). It is used as an anabolic agent in osteoporosis therapy. Previous studies by other scholars and our research group have shown that PTH can not only stimulate osteoblasts but also endothelial cells and promote angiogenesis (17). In addition, there is evidence that PTH promotes brogenesis and collagen deposition (18,19). PTH1R, a major receptor of PTH, has been found to be expressed in endothelial cells, broblasts and epithelial cells, which indicated the activating basis of PTH in skin tissue (20)(21)(22)(23). Therefore, we hypothesized that PTH might promote wound healing via multicellular stimulation and multilayer skin tissue repair.
In the present study, we tested this hypothesis. We developed a novel PTH derivative, PTHrP-2, and applied it to diabetic chronic wounds in rats. Macroscopic, histological and radiological analyses veri ed that PTHrP-2 can signi cantly promote wound healing via proangiogenesis, pro brogenesis and reepithelization. In vitro results con rmed these ndings. Endothelial cells and broblasts were activated by PTHrP-2 and showed enhanced proliferation, migration and angiogenic or brogenic activity. However, epithelial cells did not show any activation under PTHrP-2 treatment. The difference between in vivo results and in vitro results led use to hypothesize that PTHrP-2 promotes re-epithelization via indirect intercellular communications.
The exosome is a recently discovered extracellular vesicle that is important in intercellular communications (24). Recent studies have reported that drugs, hormones and environmental factors can alter exosome contents and intercellular communication (25). Therefore, we attribute in vivo reepithelization promotion by PTHrP-2 to exosome-mediated changes in intercellular communication.
Exosomes were isolated from PTHrP-2-stimulated endothelial cells. Self-activation and broblast-and epithelial cell-activation by exosomes from PTHrP-2-stimulated endothelial cells were observed in vitro and in vivo studies.

Materials And Methods
Synthesis and loading of PTHrP-2 To synthesize PTHrP-2, the FMOC/tBu solid-phase method was used(26). The whole sequence of PTHrP- KKLQD-VHNF-EEE. The three glutamic acid (Glu) residues at the C terminal and the phosphorylated serine (Ser) residue at the N-terminal are the major variations in PTHrP-2 relative to the 1-34 amino acids of PTH. Gel ltration was applied for initial puri cation of the crude peptide. In order to release PTHrP-2 sustainably in vivo, PTHrP-2 loading was applied in calcium alginate hydrogel. According to previous studies (27,28), 2% sodium alginate solution was mixed with PTHrP-2, and 1.5% CaCl2 solution was mixed into gel. The nal product was PTHrP-2@Ca-Alg.
In vivo study Induction of diabetes and excisional wound splinting model preparation Forty-ve male Sprague-Dawley rats aged 8 weeks were selected to construct a diabetic rat model according to a previously established method. All protocols obtained approval from the Animal Care and Experimental Committee of the Sixth People's Hospital a liated with Shanghai Jiao Tong University School of Medicine. Before operation, the rats fasted for one night to measure the baseline blood glucose levels. Streptozotocin (65 mg/kg b.w., i.p.) was intraperitoneally injected, and blood glucose levels were measured at three time points, on days 1, 3 and 7. After observation for two weeks, 24 rats with glucose levels of over 300 mg/dl were selected as experimental rats for follow-up operation. According to the results of the in vitro experiment, the diabetic rats were divided into three groups, control, Ca-Alg, and PTHrP-2@ Ca-Alg, to evaluate the ability of PTHrP-2@ Ca-Alg to repair diabetic skin wounds.
Animals and surgical procedure A rodent model of full-thickness skin wounds in diabetes was established. After the wound was successfully established, Ca-Alg and PTHrP-2@ Ca-Alg were placed on the wound surfaces of the animals. After surgery, sterile gauze was used to x the wound surface. The rats were observed every day to ensure that the dressings were intact. After the operations, the animals were placed in a controlled temperature environment and continued to be fed the diabetic diet, and their bedding was replaced every day.

Measurement of wound size reduction
Postoperative photos were taken with a camera (Canon, Japan) at the following four time points: day 0, day 3, day 7 and day 14. A model diagram of wound repair was constructed, and the changes in wound area and repair status were analyzed by ImageJ. The amount of wound closure was determined using the formula percent wound size reduction = ¼ [(A0-At)/A0] 100, where A0 is the initial wound area (t ¼ 0), and At is the wound area at each time point.

Micro l perfusion and microcomputed tomography
Micro l was used to evaluate neovascularization during wound healing in the diabetic rats. The experimental rats were euthanized 14 days after surgery. The hair was removed from the chest of each rat, and scissors were used to cut open the chest. After clamping the descending aorta and incising the inferior vena cava, the left ventricle was penetrated with an angiocatheter. Then, 100 ml of heparinized saline and 20 ml of Micro l (MV-122; Flow Tech, USA) were successively perfused at 2 ml/min. To ensure the polymerization and solidi cation of the contrast agent, the experimental samples were incubated at 4°C overnight. On the second day after the operation, the sample was pruned and scanned with microcomputed tomography (Micro CT) at a resolution of 9 mm to detect new blood vessels. Using 3D Creator software, 3D images were reconstructed. The blood vessel area and number of blood vessels in the wound were also determined using this software.
Histologic, immunohistochemical and immuno uorescence analysis For histology, the samples were dehydrated, embedded in para n and sliced into sections (~6μm thick). Neuroepithelial length and collagen deposition were observed via hematoxylin and eosin (H&E) and Masson's trichrome staining. Immunohistochemistry and immuno uorescence were applied to observe angiogenesis and broblasts in the wound eld. For immunohistochemistry, the sections were rehydrated and treated with antigen retrieval. After incubation with the primary antibody against CD31 (1:200, Abcam, Cambridge, UK) at 4°C overnight, the sections were incubated with a biotinylated secondary antibody and an ABC complex and stained with DAB substrate. All sections were counterstained with hematoxylin and observed under a light microscope. For immuno uorescence, the sections were rehydrated and blocked with 1.5% goat serum (Merck-Millipore). After incubation with primary antibodies against CD31 (1:200, Abcam, Cambridge, UK) and α-SAM (1:50, Abcam, Cambridge, UK) at 4°C overnight, the sections were incubated with Alexa Fluor 488-and Cy3-conjugated secondary antibodies and DAPI (Sigma-Aldrich) for visualization. The sections were observed via confocal laser scanning microscopy. Angiogenesis was determined in six sections from different samples. For each section, six high-power elds containing the entire portion of the wounds were randomly observed, and the newly formed blood vessels were evaluated. All counting procedures were conducted separately by two pathologists.

Cell proliferation and migration
The proliferation of HUVECs, HFF-1 cells, and HaCaTs was analyzed with the CCK-8 method. HUVECs were cultured in medium under control conditions or with 0.1 nM, 1 nM or 10 nM PTHrP-2 (n = 4). The cells were inoculated in 96-well culture plates (Corning, USA) at a density of 2×10 3 cells per well and cultured for 1, 3 and 7 days according to the different conditions of each group. HFF-1 cells and HaCaTs were cultured under the same conditions, but the initial number of cells was 1.5×10 3 . Then, 100 µl of culture medium containing 10% CCK-8 was added to each well of the 96-well plate, and the plates were incubated for 2 h. The absorbance value of each sample was immediately measured at 450 nm by a microtiter plate reader (BioTek, Winooski, USA).
The migration of HUVECs, HFF-1 cells and HaCaTs was tested with a transwell assay (3422, Corning, USA). HUVECs, HFF-1 cells and HaCaTs at a density of 2×10 4 cells were inoculated in the upper chamber and cultured in 200 µl of serum-starved medium, whereas the lower chamber contained 500 µl of complete medium. The cells were xed and stained for 10 min with 0.1% crystal violet after incubation for 24 h. The migrated cells were photographed by microscopy (Olympus IX 70, Tokyo, Japan) and counted by ImageJ.

Angiogenic characters
MatrigelTM (BD Bioscience) was thawed in advance in a 4°C refrigerator overnight. In precooled 24-well plates, 200 µl of Matrigel was added to each well, and the plates were then incubated at 37°C for 1 h.
HUVECs that had been precultured for 48 h in medium under different conditions were digested with trypsin and counted. A total of 1x10 5 pretreated HUVECs were added to the 24-well plates containing Matrigel, and the samples continued to be cultured in the treated medium. The tube-forming ability of HUVECs was observed by microscopy (Olympus IX 70, Tokyo, Japan) after culture in a humidi ed 37°C/5% CO2 incubator for 8 h. Statistical analysis of the number of tubes in the microscope (Olympus IX 70, Tokyo, Japan) photos was carried out with Image J.
After culturing in medium under the control condition or with 0.1 nM, 1 nM or 10 nM PTHrP-2 (n = 4) for 3 days, HUVECs and HFF-1 cells were xed with 4% paraformaldehyde for 15 min and then washed with PBS three times. Next, the cells were permeabilized with 0.25% Triton X-100 for 15 min and then blocked with 3% bovine serum albumin (BSA) for 1 h. After washing with Phosphate Buffered Saline (PBS), we added anti-VEGF (1:200, ABclonal, China) and Incubated with the cells in a 4°C refrigerator overnight; then, we added the secondary antibodies in darkness. One hour later, the cells were washed with PBS three times for 5 min each time. The cytoskeletons were then stained with 5 g/ml of phalloidin (1:200, Yeasen, China) at room temperature for 45 min. Then, the slides were washed with PBS three times for 5 min each time. After the nal wash, the samples were stained by adding 4′,6-diamidino-2-phenylindole (DAPI, 1:200, Solarbio) in PBS for 10 min, followed by imaging. The cells were visualized using a confocal microscope (Leica, Solms, Germany).
The VEGF secretion from HUVECs was detected by an enzyme-linked immunosorbent assay (ELISA). A total of 1x10 5 HUVECs were seeded in medium under the control condition or with 0.1 nM, 1 nM or 10 nM PTHrP-2 (n=4) in 6-well plates. After the cells were cultured for 3 days, the supernatants of the samples were collected, and the contents of VEGF released from the samples were detected in strict accordance with the manufacturer's instructions using an ELISA kit (NeoBioscience, China).

Fibrogenic characters
Immuno uorescence (Anti-Collagen I, Abcam, UK) and enzyme-linked immunosorbent assay (ELISA) (Human Pro-Collagen I alpha 1 DuoSet ELISA, R&D systems, USA) were used to determine the type I collagen in HFF-1 cells.

Western blotting to evaluate angiogenic and brogenic characteristics
For Western blotting, exosomes or cells were lysed. The lysates were diluted with 5 × loading buffer at a ratio of 1:5 and heated at 95°C for 5 min. The protein extracts were separated by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene uoride membranes (Immobilon P, Millipore, Billerica, USA). The membranes were blocked with 5% nonfat milk or BSA for 1 h and then incubated with primary antibodies overnight at 4°C and with HRP-linked secondary antibodies for 1 h at room temperature. The protein bands were then visualized using an enhanced chemiluminescence (ECL) substrate kit (Merck Millipore, USA).

Exosome isolation
The HUVEC and HFF-1 exosomes were isolated by ultracentrifugation. In brief, when HUVEC and HFF-1 cultures reached 80% con uence, the culture medium was removed, and the cells were washed three times with PBS. Then, serum-free medium was used for culturing. PTHrP-2 was added to the medium of the PTHrP-2-treated group at this time. After 48 h of culturing, conditioned media were collected and centrifuged at 300 × g for 10 min and 2000 × g for 15 min to remove dead cells and debris. The supernatants were then ltered via a 0.22-µm lter (Micropore) and centrifuged at 100000 × g for 1.5 h twice. Then, the pellets were resuspended in PBS.

Exosome characterization
The morphology of exosomes was identi ed by TEM (JEM-1400, JEOL, Japan). Western blotting analysis was used to verify the exosome markers Alix, TSG101 (Protein Tech, USA) and CD9 (Abcam, USA). DLS was applied to determine the exosome size distribution. Particle concentration, particle size and the video frame of exosomes were analyzed by a Flow NanoAnalyzer (FNA) (NanoFCM, China) and nanoparticle tracking analysis (NTA) (ZetaView PMX 110, Particle Metrix, Meerbusch, Germany). The protein concentration of exosomes was quantitatively detected by a BCA protein assay kit.

Exosome internalization
The puri ed exosomes were labeled with the red uorescent dye PKH26 (Sigma-Aldrich, Germany) according to the manufacturer's protocol. Subsequently, PKH26-labeled exosomes were added to the medium and incubated with HUVECs, HFF-1 cells and HaCaTs for 24 h. Afterwards, the cells were xed and washed with PBS 3 times and then blocked with QuickBlock™ Blocking Buffer for Immunol Staining (Beyotime, China). The cytoskeleton was then exposed to 5 g/ml of phalloidin (1:200, Yeasen) at room temperature for 45 min. Then, the slides were washed with PBS three times for 5 min each time. After the nal wash, the samples were stained by adding DAPI (1:200, Solarbio) in PBS for 10 min and then imaged. The cells were visualized using a confocal microscope (Leica, Solms, Germany).

Exosome-mediated intercellular communication
Exosomes were extracted from HUVECs and HFF-1 cells in the PTHrP-2-treated and untreated groups using the method described above. Exosomes from the treated groups (PTHrP-2-HUVEC-Exos and PTHrP-2-HFF-1-Exos) and untreated groups (HUVEC-Exos and HFF-1-Exos) were cultured together with HUVECs and HFF-1 cells, and the proliferation, migration and tube formation experiments were performed as described above.

Effects of PTHrP-2-treated-exosomes on HaCaTs
The exosomes extracted above were co-cultured with HaCaTs to evaluate the proliferation and migration ability of HaCaTs according to the methods described above. According to the characterization results of HaCaTs, the mechanisms were explored by western blot.

In vivo validation of PTHrP-2-HUVEC-Exos
The effect of PTHrP-2-HUVEC-Exos in vitro was veri ed by subcutaneous injection in diabetic rat wounds.
Histopathological methods were used to analyze the rats 7 days after operation. HE staining, Masson staining, CD31 immunohistochemical staining and CD31/α-SMA dual immuno uorescence staining experiments were carried out according to the methods described above.

Statistical analysis
All experiments, both in vitro and in vivo, were repeated at least three times. Data are representative of these experiments and are shown as the mean ± standard deviation (SD). The means of multiple groups were compared with one-way analysis of variance (ANOVA). The independent sample test was used to compare means between two groups. Statistical analysis was conducted using GraphPad Prism software, and P < 0.05 was considered statistically signi cant.

Results
Evaluation of PTHrP-2@Ca-Alg effects on wound healing in vivo Fig. 1A outlines the process of the animal experiment. Fig. 1B showed untreated wounds and wounds treated with Ca-Alg or PTHrP-2@Ca-Alg at 4 time points. Over time, the size of the wounds in all three groups decreased by various degrees, with the wounds in the PTHrP-2 group and the PTHrP-2@Ca-Alg group becoming smaller than those in the untreated group (Fig. 1C). The wounds treated with PTHrP-2@Ca-Alg were nearly healed by day 14. According to the quantitative data analysis of wound closure (equation (1)), the areas of wounds in the group treated with PTHrP-2@Ca-Alg were smaller than those in the other groups at three time points, and among the three groups, the PTHrP-2@Ca-Alg group exhibited the highest wound healing rate. The process of wound healing in rats is illustrated in Fig. 1D.
Fourteen days after surgery, Micro CT was used to evaluate the vascular formation of Ca-Alg-treated, PTHrP-2@Ca-Alg-treated and untreated wounds. The reconstructed three-dimensional images ( Fig. 2A) showed that the vascular density in the PTHrP-2@Ca-Alg-treated group was signi cantly higher than that in the other two groups. The quantitative data analysis of the number of newly formed blood vessels (Fig.  2B) and the area and number of blood vessels showed that these variables were signi cantly higher in the PTHrP-2 treatment group than in the Ca-Alg group and control group.
CD31 immunohistochemical staining and CD31/α-SMA dual immuno uorescence staining of the wound tissue at 7 and 14 days after surgery showed new blood vessel formation and mature blood vessels (Fig.  2C). The quantitative data analysis of the newly formed vascular density, i.e., the number of CD31positive cells per mm2, con rmed the increase in the number of wound vessels following treatment with PTHrP-2@Ca-Alg (Fig. 2D). At day 14, blood vessel density was much higher in the PTHrP-2@Ca-Alg group than in other three groups. The number of mature blood vessels increased in all four groups from the 7th day to the 14th day after the wound surface operation but was lower than the number of new blood vessels in each group. The number of mature blood vessels on the wound surface was signi cantly increased in the PTHrP-2@Ca-Alg group relative to the numbers in the other groups on the 14th day after surgery.
According to the histological analysis of the Masson's trichrome staining, there was a signi cant difference in treatment effect among the three groups (Fig. 3A). Compared with the control group, the PTHrP-2@Ca-Alg-treated group showed more extensive collagen deposition and greater collagen ber thickness. The processed images of the PTHrP-2@Ca-Alg-treated group revealed improved arrangement of collagen bers, similar to that of normal skin, which re ects the positive roles of PTHrP-2 in ECM deposition and collagen alignment. In general, the PTHrP-2@Ca-Alg group had greater numbers structures resembling hair follicles and sebaceous glands than did the other groups. According to the optical microscopy of H&E staining (Fig. 3B), new epithelial tissue was formed in the wounds of the three groups. The initial width of each wound was 2 cm; in the gure, the black arrow indicates the length of the new epithelium. As shown in the gure (Fig. 3C), the wound healing effect in the PTHrP-2@Ca-Alg treatment group was signi cantly better than that in the other three groups on days 7 and 14.
In vitro study of PTHrP-2 effects HUVECs (Fig. 4A), HFF-1 cells (Fig. 4B) and HaCaTs (Fig. 4C) were cultured in medium containing different concentrations for 1, 3 and 7 days, respectively, and the proliferation of these cell lines is shown in Fig. 2. In HUVECs and HFF-1 cells, the proliferation rates of cells in the treatment groups were higher than that of cells in the untreated groups, and the proliferation effect became more pronounced as the concentration increased beginning at 3 days. At 7 days, we observed that PTHrP-2 stimulated the proliferation of HUVECs and HFF-1 cells; 10 nM PTHrP-2 had the strongest effect on cell proliferation. However, in the CCK-8 assay, we found that PTHrP-2 had no signi cant effect on proliferation of HaCaTs, indicating that there was no cytotoxicity.
The transwell assays showed that the numbers of migrated cells in the treatment groups were signi cantly higher than those in the untreated groups. Among the four groups, the group treated with 10 nM PTHrP-2 exhibited the strongest migration capacity. By measuring the tube formation activity of HUVECs, we determined the angiogenic potential of PTHrP-2 at different concentrations. The HUVECs treated with PTHrP-2 formed elongated tube structures on the substrate gel substrate layer after 8 h of incubation, whereas without PTHrP-2 treatment, the HUVECs formed incomplete or sparse tubular networks. The HUVECs treated with 10 nM PTHrP-2 produced the most blood vessels among the treatment groups, and the blood vessels were largely complete (Fig. 4D, E).
In Fig. 4F, the cytoskeleton is stained red by phalloidin, the nucleus is stained blue by DAPI, and the VEGF secreted by the cells is green. At 3 days, the amounts of blue and green uorescence in the treatment groups were signi cantly greater than those in the untreated group. PTHrP-2 strongly stimulated VEGF secretion from HUVECs and HFF-1 cells. Among the four treatment conditions, 10 nM PTHrP-2 was the most effective. To further explore the mechanism of the PTHrP-2 effect on HUVEC and HFF-1 cells, the levels of VEGF secreted from cells cultured in different concentrations of PTHrP-2 were detected by an ELISA kit. In the PTHrP-2 treatment group, the VEGF content in the supernatant of cultured HUVECs was the highest after 3 days, and 10 nM PTHrP-2 provided the best conditions for HUVEC secretion of VEGF.
Similar results were obtained in HFF-1 cells. After treatment in the 10nM PTHrP-2 group, type I collagen of HFF-1 cells was the highest, followed by 1nmpthrp-2. However, in the HFF-1 cells, there was no difference in type I collagen levels between the 0.1nM and untreated treatment (Fig. 4G).
To explore the mechanism of PTHrP-2 action at the protein level, Western blotting was performed in the untreated groups and the groups treated with different concentrations of PTHrP-2. The results (Fig. 4H) suggest that PTHrP-2 may activate the PI3K/Akt and Erk1/2 signaling pathways. Compared with the untreated group, the treatment groups showed increased phosphorylation of Akt and Erk, which indicated the activation of the two signaling pathways in HUVECs and HFF-1 cells. Among the groups, the 10 nM treatment group exhibited the most pronounced effect.

Evaluation of the PTHrP-2-treated exosomes
Characterization of PTHrP-2-treated exosomes The nanoparticles puri ed from HUVECs and HFF-1 cells treated with PTHrP-2 were characterized by TEM, DLS and Western blotting. TEM (Fig. 5A and 6A) experiments with PTHrP-2-treated exosomes showed that most of the extracted nanoparticles were spherical or cup-shaped, which indicated the presence of exosomes. We observed the presence of exosome markers, such as Alix, TSG101 and CD9, by Western blotting (Fig. 5B and 6B). These markers con rmed that the particles were exosomes. The sizes of the PTHrP-2-treated exosomes were directly determined using a DLS system called the Nanosizer system, which ranges from 40 to 100 nm ( Fig. 5C and 6C) When the samples were further concentrated for data analysis, we found that the exosome concentration in the treated group was approximately 1.5 times greater than that in the untreated group through FNA ( Fig. 3 and 4D) and NTA ( Fig. 5E and 6E) detection. After exosome proteins were extracted, BCA protein assay (Beyotime, China) was performed, and the data were in accordance with the concentration ratio ( Fig. 5F and 6F). The cytoskeleton, PKH26labeled exosomes (PKH26-Exos) and cell nuclei are stained green, red and blue, respectively, in the images (Fig. 5G and 6G) collected with laser scanning confocal microscopy. PKH26-Exos were located in the area around the nucleus. A red area in the perinuclear region was observed in more than 90% of the HUVECs and HFF-1 cells. These data indicate that the exosomes can be successfully internalized by HUVECs and HFF-1 cells.
In vitro study of the PTHrP-2-treated exosomes To investigate the effect of PTHrP-2 on exosomes, we extracted exosomes from untreated and treated HUVECs and HFF-1 cells. The extracted exosomes were tested for their effects on the migration of HUVECs and HFF-1 cells as described above. Relative to control treatment and treatment with the untreated exosomes, treatment with the PTHrP-2-treated exosomes signi cantly increased the proliferation (Fig. 5H and 6H)  Based on the results of the in vitro experiments, HUVEC-Exos and PTHrP-2-HUVEC-Exos were used for the treatment of diabetic rat wounds by subcutaneous injection. In the HE staining, Masson staining, CD31 immunohistochemical staining and CD31/α-SMA dual immuno uorescence staining experiments, PTHrP-2-HUVEC-Exos was found to yield signi cantly better effects than the other two treatments. After 14 days of treatment with PTHrP-2-HUVEC-Exos, the rates of epithelialization, angiogenesis and collagen arrangement of the wound surface were signi cantly enhanced (Fig. 7E).

Discussion
Skin tissue is composed of a dermal layer and corium layer and three cell types: endothelial cells, broblasts and epithelial cells. Wound healing can be divided into four phases: a coagulation phase, an in ammation phase, a reconstruction phase and a maturation phase (29). All three cell types are involved in these phases. Diabetic conditions impair the activation of these three cell types, impairing wound healing and leading to chronic wounds (30). In healing chronic wounds, the three cell types are equally important. Hyperglycemia caused by diabetes can destroy the microvascular structure, thereby affecting the functions of blood vessels and the ability of cells to deliver oxygen (31,32). Collagen, a component of the corium layer, can also be impaired by hyperglycemia (33). Furthermore, some studies have reported that hyperglycemia can inhibit epithelium cell proliferation and migration. These phenomena may inhibit wound healing and lead to the formation of chronic wounds (34). The three types support and protect each other. Simultaneous activation of the three cell types does not have simple additive effects but rather a synergistic effect. Different from traditional bioactive factors, PTH has multiple roles in wound healing via multicellular activation. In vivo, PTHrP-2 induced rapid wound closure. H&E staining revealed increased re-epithelization of the wound by PTHrP-2. Masson trichrome staining revealed denser collagen deposition and a better collagen array, approaching that of normal skin, due to PTHrP-2. Immuno uorescence and immunohistochemistry for CD31 and a-SMA and micro l perfusion revealed a denser neovascularization and more mature capillary structure under PTHrP-2 treatment. These results suggested that PTHrP-2 is effective in promoting chronic wound healing and that its healing effects are multicellular. PTHrP-2 is a multifunctional factor that exerts synergistic effects on the three cell types and wound healing.
To clarify the mechanism of action of PTHrP-2 in wound healing, we conducted in vitro experiments to study the effects of PTHrP-2 on the three cell types. Endothelial cells and broblasts exhibited activation under PTHrP-2 conditions. Proliferation, migration, capillary-structure formation, VEGF and type I collagen secretion were promoted by PTHrP-2. In previous studies, PTH-treated cells were found to play major roles via PTH1R and the downstream Akt/Erk 1/2 pathway (35,36). Furthermore, the Akt/Erk 1/2 pathway is important in angiogenesis and brogenesis. Therefore, we examined Akt/Erk 1/2 pathway change under PTHrP-2 conditions in endothelial cell and broblasts. The results indicated that the Akt/Erk 1/2 pathway is a major pathway through which PTHrP-2 acts on endothelial cells and broblasts.
However, these promoting effects were not seen in HaCaTs under PTHrP-2 intervention. Proliferation and migration of HaCaTs were not enhanced by PTHrP-2. The in vitro results were not in line with the in vivo results. HaCaTs do have PTH receptors on their surfaces. Some studies reported that PTH promotes antimicrobial peptide expression on epithelial cells (37). Our experiment results and previous studies indicated that PTHrP-2 may have major antibacterial effects on epithelial cells but not proliferation or migration activity. We formulated a second hypothesis to explain the promotion of re-epithelization in the in vivo experiment. We hypothesized that PTHrP-2 promotes re-epithelization by altering the intercellular communication among the three cell types. The exosome, a newly discovered extracellular vesicle, was postulated as major mediator of intercellular communication in this study.
Exosomes from endothelial cells and broblasts without PTHrP-2 interference were carefully isolated, identi ed and added to the three cell types. Exosomes from endothelial cells and broblasts without PTHrP-2 interference do have some self-activation and interactivation abilities. In addition, endothelial cell exosomes have some ability to promote HaCaTs proliferation and migration. These results suggest exosomal-mediated cellular interactions in skin tissue. Endothelial cells, broblasts and epithelial cells undergo natural exosomal interactions to maintain skin tissue homeostasis.
Exosomes from endothelial cells and broblasts under PTHrP-2 interference showed stronger selfactivation and interactivation abilities than did those without such interference. Importantly, PTHrP-2stimulated endothelial cell exosomes exhibited signi cant promotion effects on HaCaTs proliferation and migration, which may explain PTHrP-2's ability to cause re-epithelization. Previous studies reported that certain drugs and environments, such as hypoxia, may alter exosome contents and thereby alter the intercellular messages communicated by exosomes(38-40). The exosome counts revealed that PTHrP-2 enhances exosome secretion. Through enhancing exosome production and strengthening exosome abilities, PTHrP-2 reinforces the intercellular interactions among endothelial cells, broblasts and epithelial cells and produces an exosomal intercellular network.
When investigating the mechanism of action of exosomes obtained following PTHrP-2 treatment, we selected the PI3K-Akt pathway for HaCaTs due to its canonical role in re-epithelization (41,42). Gsk3β is a typical regulation of the negative Wnt signaling pathway. Many researchers have described PI3K/AKT and Gsk3β/β-catenin signals as playing a key role in skin development and wound healing (43). Based on the ndings of previous studies, Gsk3β/β-catenin and PI3K/Akt signaling pathway was selected for veri cation. The results showed that canonical pathways were activated by PTHrP-2-interfered exosomes. This nding provides a good explanation for the enhanced proliferation and migration ability of HaCaTs due to PTHrP-2-HUVEC-Exos. We did not identify the factor that initiates these canonical pathways; this topic will be explored in our next study.
Tests of the effects of PTHrP-2-interfered exosomes on chronic wound healing in vivo were conducted.
Both natural endothelial cell exosomes and PTHrP-2-interfered exosomes exhibited wound healing properties. However, PTHrP-2-interfered exosomes exhibited much stronger healing abilities than natural endothelial cell exosomes in proangiogenesis, pro brogenesis and re-epithelization. The results con rmed our speculation.
In this study, we carried out in vivo and in vitro experiments from the perspective of drugs to promote tissue repair. Although the results of this study are satisfactory, there are still some defects in revealing the underlying mechanism of chronic wound in diabetes mellitus and the factors that led to lead PTHrP-2stimulated enhancement of exosome abilities.In the follow-up study, we consider to dig deeply into the pathogenesis of diabetes. Further studies will be conducted using in vivo and in vitro models of diabetes, such as high-sugar models or transgenic animals and the factors and related pathways will be investigated in our future work.

Conclusion
In this study, we hypothesized that a PTH derivative can be applied to chronic wound healing for its multicellular stimulatory effects. Our rst in vivo study validated our hypothesis, but the in vitro results did not provide evidence of the promotion of wound re-epithelization by PTHrP-2. Therefore, we further hypothesized that PTHrP-2 alters intercellular communication among endothelial cells, broblasts and epithelial cells. Exosomes were isolated from PTHrP-2 stimulated endothelial cells and showed a strong ability to activate epithelial cells. The ndings indicate that the mechanism of PTHrP-2 in promoting chronic wound healing can be attributed to direct stimulation and indirect exosomal activity. PTHrP-2 is an ideal multifunctional chronic wound healing agent that has multicellular synergistic effects.

Availability of data and materials
The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.
Ethics approval and consent to participate All protocols obtained approval from the Animal Care and Experimental Committee of the Sixth People's Hospital a liated with Shanghai Jiao Tong University School of Medicine.

Consent for publication
Not applicable

Competing interests
There are no con icts to declare.