This study was approved by the Ethics Committee of the First Hospital of Hunan University of Chinese Medicine (NO.HN-LL-KY-2020-013-01). All experiments were performed strictly in accordance with the Declaration of Helsinki, and informed consent was obtained from all the participants.

Umbilical-cord blood from healthy, full-term newborns was obtained from the Obstetrics and Gynecology Department of the First Hospital, Hunan University of Chinese Medicine, Changsha, China. Heparin (20 U/mL) was added to the cord blood for anticoagulation. The isolated cord blood mononuclear cells were resuspended in DMEM (D5796, Sigma, Saint Louis, MO, USA) with 10% FCS (04-001-1ACS, Gibco, Carlsbad, CA, USA). The cell suspension (3 $\times$ 10${}^6$/mL) was inoculated into a plate with slides and cultured in an incubator (DH-160I, Shanghai Santeng, Shanghai, China) at 37 °C with 5% CO${}_2$ and saturated humidity. The cells were identified using flow cytometry and immunofluorescence (IF) staining. Afterward, the EPCs were treated with the corresponding concentrations of ASIV (0, 50, 100, and 200 mg/L) for 72 h. As previously described [24], the exosomes were harvested by centrifugation at 200,000 $\times$g for 60 min, and treated and examined by transmission electron microscope (TEM, JEM1400, Jeol, Beijing, China).

Male, 8-week-old Sprague-Dawley rats (n = 96) weighing 250–300 g were purchased from Hunan SJA Laboratory Animal Co., Ltd. (Changsha, Hunan, China). The experimental protocol was approved by the Animal Experimentation Ethics Committee of the First Hospital of Hunan University of Chinese Medicine (NO. ZYFY20201018-2). Rats were housed alone and exposed to a 12/12-h light/dark cycle at 22–24 °C and with ad libitum access to food and water. One week later, they were divided into control (n = 12) and diabetic (n = 84) groups. After fasting for 12–16 h, rats in the diabetic group were given a single intraperitoneal injection of streptozotocin (STZ) solution (65 mg/kg body weight, in 0.1 M citrate buffer, pH = 4.5, S0130, Sigma-Aldrich, Saint Louis, State of Missouri, USA) to induce a type I diabetes model [25, 26]. Blood glucose levels were randomly monitored daily after the first 72 h. When three consecutive random blood-glucose concentrations were more than 16.7 mmol/L, a successful model of diabetes was considered to have been established. After 14 d, rats in the control and diabetic groups were anesthetized and shaved. The dorsal skin of the rats was disinfected, and four pieces of 1.5 $\times$ 1.5-cm${}^2$ full-layer skin were removed from the deep fascia on both sides of the rat spine. All the rats were injected intraperitoneally with 100 mg/kg bromodeoxyuridine (BrdU, Sigma, St. Louis, MO, USA). STZ, rapamycin, and wortmannin were purchased from Sigma. Rats in the diabetic group were randomly divided into the following groups: an EPCexo group, a low dose of ASIV-EPCexos group (low-ASIV-EPCexo), a medium dose of ASIV-EPCexos group (mid-ASIV-EPCexo), a high dose of ASIV-EPCexos group (hi-ASIV-EPCexo), a hi-ASIV-EPCexo + rapamycin group, and a hi-ASIV-EPCexo + wortmannin group. Rats in the EPCexo group were injected with 500 mg/L EPCexos. The low-, mid-, and hi-ASIV-EPCexo groups were treated with ASIV-EPCexos at different concentrations (250, 500, and 1000 mg/L). Rats in the hi-ASIV-EPCexo + rapamycin and hi-ASIV-EPCexo + wortmannin groups were treated with rapamycin (37 nM, mTOR blocker, ab120224, AbcamCam-bridge, Cambridgeshire, UK) or wortmannin (15 µg/kg, PI3K blocker, ab120148, AbcamCam-bridge, Cambridgeshire, UK), respectively, following treatment with 1000 mg/L ASIV-EPCexos. EPCexos were locally injected around the wounds [27, 28]. Wound photographs were recorded at 0, 3, 7, 10, and 14 d, and the percentage of wound healing area was calculated using the Image-Pro Plus image analysis system following skin resection. Rats were euthanized at 14 d, and the skin was excised and preserved for further analysis.

The expression of the surface markers CD31 (+), CD34 (+), CD45 (–), and CD133 (+) of EPCs was evaluated using flow cytometry. After approximately 12 d of subculture, cells were obtained. The resuspended cells were precipitated with 100 µL of 0.5% BSA-PBS. Subsequently, 2 µL CD34-FITC (bs-0646R-FITC, Bioss, Beijing, China), 1 µL CD133-FITC (bs-0209R-FITC, Bioss, Beijing, China), 3 µL CD31-FITC (bs-0195R-FITC, Bioss, Beijing, China), and 1 µL CD45-FITC (bs-0522R-FITC, Bioss, Beijing, China) antibodies were incubated with the cells at 37 °C in the dark for 30 min. After washing and centrifugation, the cells were analyzed by flow cytometry (A00-1-1102, Beckman, Brea, California, USA).

The expression of CD34 and VEGF receptor 2 (VEGFR2) of EPCs was detected using IF staining. IF staining was used to evaluate the expression of CD31, $\alpha$SMA, CD45, BrdU, and VEGF in the skin tissue. Paraffin-embedded skin tissue blocks were cut into 4-µm sections. These were dewaxed and rehydrated using xylene and ethanol and subsequently immersed in citrate buffer (pH 6.0) and heated to retrieve the antigens. The sections were stained with sodium borohydride solution and Sudan black and rinsed with distilled water. The slides and sections were placed in 5% BSA and incubated for 60 min. Slides and sections were incubated overnight at 4 °C with diluted primary antibodies against CD34 (ab81289, 1:50, Abcam, Cambridge, Cambridgeshire, UK), VEGFR2 (ab39378, 1:50, Abcam, Cambridge, Cambridgeshire, UK), CD31 (11265-1-AP, 1:100, PTG, Wuhan, Hubei, China), $\alpha$SMA (55135-1-AP, 1:100, PTG, Wuhan, Hubei, China), CD45 (ab33923, 1:100, Abcam, Cambridge, Cambridgeshire, UK), BrdU (#5292, 1:100, CST, Danvers, Massachusetts, USA), and VEGF (ab46154, 1:100, Abcam, Cambridge, Cambridgeshire, UK). The next day, the slides and sections were incubated at 37 °C for 90 min with anti-rabbit or mouse IgG (SA00013-4 or SA00013-1, 1:200, Proteintech, Chicago, Illinois, USA). Slides and sections were incubated with 4${}^{\prime }$,6-diamidino-2-phenylindole and observed using a fluorescence microscope.

Following treatment of EPCs with ASIV for 24 h, the medium was removed, discarded, and replaced with 100 µL medium containing 10% CCK8 (NU679, Dojindo, Kumamoto, Japan). The absorbance was measured at 450 nm using a microplate reader (MB-530, HEALES, Shenzhen, Guangdong, China).

Intervention EPCs were hydrolyzed into single cells. A cell suspension, prepared in a serum-free basal medium (1 $\times$ 10${}^5$/mL), was added to the upper Transwell chamber (33318035, Corning, Corning, NY, USA). A medium with 10% FBS was then added to the lower chamber. After 48 h, the medium was discarded, and the cells in the upper compartment were cleaned. The cells in the lower compartment were fixed with a 1:1 acetone–methanol solution for 20 min. The cells were stained and subsequently photographed using an inverted microscope (DSZ2000X, Cnmicro, Beijing, China).

Corning Matrigel Basement Membrane Matrix (356234, BD Biosciences, Franklin Lakes, NJ, USA) was added to each well of a 48-well plate until the wells were evenly covered without bubbles. The 48-well plates were then incubated at 37 °C for 1 h. Cells were digested with 0.25% trypsin and suspended before the addition of 7.5 $\times$ 10${}^4$ cells/well to the 48-well plate. Photographs were taken (100$\times$ magnification) after incubation at 37 °C for 6 h.

An ExoQuantTM Overall Exosome Capture and Quantification Assay Kit (#K1201-100, Biovision, San Francisco, CA, USA) was used for quantitative analysis of exosomes. All procedures were carried out following the guidelines. The absorbance was obtained at 450 nm. The concentration of the exosomes was calculated after plotting a standard curve.

A 30 µg sample of denatured protein was added to each well of the gel. Samples were electrophoresed for 130 min at a constant voltage of 75 V. Target proteins were transferred from the gel to a nitrocellulose (NC) membrane at a constant current of 300 mA. Following the transfer process, the NC membrane was subjected to blocking. The primary antibodies against CD63 (25682-1-AP), CD9 (20597-1-AP), CD81 (66866-1-Ig), mTOR (66888-1-Ig), ras homolog enriched in brain (Rheb, 15924-1-AP), PI3K (67071-1-Ig), AKT (10176-2-AP), PIP3 (17552-1-AP), PDK1 (10026-1-AP), and $\beta$-actin (60008-1-Ig) were purchased from Proteintech (Chicago, Illinois, USA), whereas those against VEGFa (ab46154), VEGFb (ab185696), VEGFc (ab9546), fibroblast growth factor (FGF, ab208687), and angiotensin-1 (Ang-1, ab183701), p-mTOR (ab109268), eNOS (ab76198), and tuberous sclerosis 2 (TSC2, ab32554) were purchased from Abcam (Cambridge, Cambridgeshire, UK). The primary antibody was incubated overnight at 4 °C. The NC membrane was incubated with anti-mouse/rabbit IgG (SA00001-1 or SA00001-2, Proteintech, Chicago, Illinois, USA) for 90 min. Protein bands were visualized by SuperECL Plus hypersensitive luminescent solution (K-12045-D50, Advansta, Beijing, China). Photographs were taken using a ChemiScope 6100 chemiluminescence imaging system (CLINX, Shanghai, China).

HE staining was used to observe the morphological changes in rat skin. Paraffin-embedded rat-skin tissue samples were cut into 4-µm-thick sections using a slicer. The sections were dewaxed and rehydrated using xylene and ethanol and stained with HE. The sections were examined under a microscope (BA210T, Motic, Xiamen, Fujian, China).

Paraffin-embedded rat skin tissues were cut, dewaxed, and rehydrated. The sections were incubated with hematoxylin solution, rinsed with distilled water, and stained with acid fuchsin solution. Next, the sections were then incubated with 1% phosphomolybdic acid and blue aniline solutions for 5 min. The samples were examined under a microscope.

Total RNA was extracted, and the cDNA was obtained by reverse transcription (CW2569, Beijing CWBIO Co., Ltd., Beijing, China). The primer sequences for the target genes are presented in Table 1. Primers were synthesized by Shanghai Sangon Biotech. A PCR system containing the fluorescent dye UltraSYBR Mixture (CW2601, Beijing CWBIO Co., Ltd., Beijing, China) was prepared, and the reaction was performed with QuantStudio1 (Thermo, Waltham, Massachusetts, USA). The fluorescence signal was monitored in real time, and the internal reference was $\beta$-actin. The gene levels were calculated (2${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{Ct}}$ method).

Statistical analyses were conducted using GraphPad Prism 8.0.1 (GraphPad Software, Inc., San Diego, CA, USA), and the data were expressed as mean $\pm$ standard deviation. The differences between the two groups were analyzed utilizing the Student’s t-test. For comparing differences among multiple groups, a one-way analysis of variance (ANOVA) was performed, followed by a Tukey post-hoc test. Statistical significance was set at p 0.05.

EPCs were identified using flow cytometry and IF staining. The expression of CD31, CD34, and CD133 was positive, but the expression of CD45 was negative, indicating that the endothelial cells were successfully extracted (Supplementary Fig. 1). As depicted in Fig. 1A, the viability of EPCs in the 50, 100, and 200 mg/L groups exhibited a substantial enhancement in comparison to the control group. The cell migration and tube-formation abilities of EPCs in the 50, 100, and 200 mg/L groups were appreciably improved (Fig. 1B,C). EPCs in the 100 mg/L group showed optimal ability. Therefore, ASIV promotes the progress of EPCs, and a concentration of 100 mg/L exhibited the optimal effect.

Fig. 1.

ASIV promotes EPC progress. (A) Representative graphs of CCK-8 showing cell viability. (B) Representative light micrographs and graphs showing cell migration by Transwell assay. (C) Representative light micrographs and graphs of tube formation assay revealing the tube formation ability. *p 0.05 vs the control group. Magnification = 100$\times$, scale bar = 100 µm. ASIV, astragaloside IV; EPC, endothelial progenitor cell; CCK-8, cell counting kit-8.

We next examined the effect of ASIV on EPCexo secretion. The EPCexo characteristic markers CD9, CD63, and CD81 levels were detected [17]. The expression of those markers was positive in all the groups (Fig. 2A), indicating the successful extraction of EPCexos. In TEM, EPCexos showed a “cup holder” structure in the four groups (Fig. 2B). EPCexos were quantified using ELISA. As shown in Fig. 2C, the concentration of EPCexos was appreciably higher in the 50, 100, and 200 mg/L groups than in the control group, with the highest concentration observed in the 100 mg/L group. These findings indicate that ASIV facilitates EPCexo secretion, with the optimal effect observed at a concentration of 100 mg/L.

Fig. 2.

ASIV promotes EPCexo secretion. (A) The WB analysis of shows the expression of EPCexo characteristic markers CD9, CD63, and CD81. (B) Representative micrographs of EPCexo morphology were obtained via TEM. Scale bar = 60 nm. (C) Comparison of EPCexo content via ELISA in the different groups. *p 0.05 vs the control group. EPCexo, EPC-derived exosomes; WB, western blotting; ELISA, enzyme-linked immunosorbent assay; TEM, transmission electron microscope.

The chronic-wound model in STZ-stimulated diabetic rats was used to investigate whether ASIV-EPCexos promote type I diabetic-wound healing. The EPCexo group exhibited a higher rate of wound healing compared to the model group (Fig. 3A,B). After ASIV-EPCexo treatment, wound healing rates in the low-, mid-, and hi-ASIV-EPCexo groups were substantially higher than in the EPCexo group, and the hi-ASIV-EPCexo group exhibited optimal healing. As opposed to the model group, the positive staining rates of BrdU and VEGF were substantially raised in the EPCexo group (Fig. 3C–E). Moreover, the low-, mid- and hi-ASIV-EPCexo groups exhibited stronger positive staining for BrdU and VEGF compared to the EPCexo group, and the hi-ASIV-EPCexo group demonstrated the highest staining intensity. Granulation tissue formation, epithelial reformation, collagen deposition, fibroblast proliferation, and angiogenesis are hallmarks of wound healing [29]. In contrast to the model group, the epidermal thickness in the EPCexo and low-, mid-, and hi-ASIV-EPCexo groups gradually decreased. The collagen area was higher in the low-, mid-, and hi-ASIV-EPCexo groups than in the model group (Fig. 4A–C). IF staining results revealed that the number (No.) of CD31 and $\alpha$SMA positive cells were raised in the EPCexo and low-, mid-, and hi-ASIV-EPCexo groups (Fig. 4D–G). Rapamycin and wortmannin reversed the above changes in the hi-ASIV-EPCexo group. In conclusion, ASIV-EPCexo accelerated skin-wound healing in diabetic rats, with 1000 mg/L ASIV-EPCexo showing an optimal effect.

Fig. 3.

ASIV-EPCexos may promote type I diabetic-wound healing via facilitating BrdU and VEGF expression. (A) Representative images of the diabetic wounds of rat skin at different time points. (B) Comparison of wound healing rate of rat skin at different time points. (C) Representative micrographs of IF staining showing the expression of BrdU and VEGF. (D,E) Statistical analysis of the expression of BrdU and VEGF. *p 0.05 vs the model group. #p 0.05 vs the EPCexo group. &p 0.05 vs the hi-ASIV-EPCexo group. Magnification = 400$\times$, scale bar = 25 µm. n = 12 rats/group. IF, immunofluorescence; BrdU, bromodeoxyuridine; hi-ASIV-EPCexo, a high dose of ASIV-EPCexos.

Fig. 4.

ASIV-EPCexos may promote type I diabetic-wound healing via modulating CD31 and $\alpha$SMA expression. (A) Representative micrographs of HE staining. (B) Representative micrographs of MT staining. (C) Statistical analysis of (A,B). (D,E) Representative micrographs of IF staining showing CD31 and $\alpha$SMA expression. (F,G) Statistical analysis of (D,E). *p 0.05 vs the model group. #p 0.05 vs the EPCexo group. &p 0.05 vs the hi-ASIV-EPCexo group. Magnification = 100$\times$ or 400$\times$, scale bar = 100 µm or 25 µm. n = 12 rats/group. HE, hematoxylin and eosin; MT, masson’s trichrome.

To study the mechanism by which ASIV-EPCexos accelerate type I diabetic wound healing, we analyzed the changes in skin tissue at the molecular level. The expression levels related to vascular growth (VEGFa, VEGFb, VEGFc, FGF, and Ang-1) in the EPCexo group were higher than in the model group. These expression levels were further increased following ASIV-EPCexo treatment. Compared to the hi-ASIV-EPCexo group, these expression levels in the hi-ASIV-EPCexo + rapamycin and hi-ASIV-EPCexo + wortmannin groups were substantially decreased (Fig. 5A,B). We then examined the expression levels of the pathway proteins and genes in the samples. In terms of the model group, the expression levels of p-mTOR, mTOR, p-mTOR/mTOR, Rheb, eNOS, PI3K, AKT, PIP3, PDK1, and TSC2 in the EPCexo group were raised (Fig. 5C–F), although several of them were not substantially different. In addition, the expression levels of genes and proteins related to these pathways were higher in the low-, mid-, and hi-ASIV-EPCexo groups than in the EPCexo group. However, the expression of these genes and proteins decreased in the hi-ASIV-EPCexo + rapamycin and hi-ASIV-EPCexo + wortmannin groups. The results suggest that ASIV-EPCexos may activate the PI3K/AKT/mTOR pathway, thereby promoting angiogenesis.

Fig. 5.

ASIV-EPCexos promote angiogenesis via the PI3K/AKT/mTOR pathway in diabetic wounds. (A) Representative graphs of RT-qPCR showing the relative mRNA expression of VEGFa, VEGFb, VEGFc, FGF, and Ang-1. (B) The WB analysis shows the protein expression of VEGFa, VEGFb, VEGFc, FGF, and Ang-1. (C,D) Representative graphs of RT-qPCR showing the expression of mTOR, Rheb, eNOS, PI3K, AKT, PIP3, PDK1, and TSC2. (E) The WB analysis shows the protein expression of mTOR and p-mTOR. (F) The WB analysis shows the protein expression of Rheb, eNOS, PI3K, AKT, PIP3, PDK1, and TSC2. *p 0.05 vs the model group. #p 0.05 vs the EPCexo group. &p 0.05 vs the hi-ASIV-EPCexo group. n = 12 rats/group. FGF, fibroblast growth factor; Ang-1, angiotensin-1; Rheb, ras homolog enriched in brain; PIP3, phosphatidylinositol triphosphate; PDK1, PH domain of 3-phosphoinositide-dependent protein kinase 1; TSC2, tuberous sclerosis 2.

Interleukin (IL)-1$\beta$-dependent inflammation and immunity have important healing effects in diabetic wounds [30, 31]. Therefore, we next investigated the expression of inflammatory factors and immune cells in the skin of diabetic wounds. As shown in Fig. 6A, the expression levels of the IL-6 and IL-1$\beta$ genes were lower in the EPCexo group than in the model group. Furthermore, their expression in the low-, mid-, and hi-ASIV-EPCexo groups was even lower than in the EPCexo group. However, their expression was higher in the hi-ASIV-EPCexo + rapamycin and hi-ASIV-EPCexo + wortmannin groups than in the hi-ASIV-EPCexo group. IF staining analysis revealed that the number of CD45 immune cells was lower in the EPCexo group than in the model group (Fig. 6B,C). The CD45 immune cell content was lower in the low-, mid-, and hi-ASIV-EPCexo groups compared to the EPCexo group, whereas the rapamycin and wortmannin inhibitors reversed this effect. These results suggest that ASIV-EPCexos inhibit inflammation in skin wounds through the PI3K/AKT/mTOR pathway.

Fig. 6.

ASIV-EPCexos inhibit inflammation via the PI3K/AKT/mTOR pathway in type I diabetic-wound healing. (A) Representative graphs of RT-qPCR showing the mRNA expression of IL-6 and IL-1$\beta$. (B,C) Representative micrographs and statistical analysis of CD45 immune cells. *p 0.05 vs the model group. #p 0.05 vs the EPCexo group. &p 0.05 vs the hi-ASIV-EPCexo group. Magnification = 400$\times$, scale bar = 25 µm. n = 12 rats/group.