A Whole‐Course‐Repair System Based on Neurogenesis‐Angiogenesis Crosstalk and Macrophage Reprogramming Promotes Diabetic Wound Healing

Diabetic wound (DW) therapy is currently a big challenge in medicine and strategies to enhance neurogenesis and angiogenesis have appeared to be a promising direction. However, the current treatments have failed to coordinate neurogenesis and angiogenesis simultaneously, leading to an increased disability rate caused by DWs. Herein, a whole‐course‐repair system is introduced by a hydrogel to concurrently achieve a mutually supportive cycle of neurogenesis‐angiogenesis under a favorable immune‐microenvironment. This hydrogel can first be one‐step packaged in a syringe for later in situ local injections to cover wounds long‐termly for accelerated wound healing via the synergistic effect of magnesium ions (Mg2+) and engineered small extracellular vesicles (sEVs). The self‐healing and bio‐adhesive properties of the hydrogel make it an ideal physical barrier for DWs. At the inflammation stage, the formulation can recruit bone marrow‐derived mesenchymal stem cells to the wound sites and stimulate them toward neurogenic differentiation, while providing a favorable immune microenvironment via macrophage reprogramming. At the proliferation stage of wound repair, robust angiogenesis occurs by the synergistic effect of the newly differentiated neural cells and the released Mg2+, allowing a regenerative neurogenesis‐angiogenesis cycle to take place at the wound site. This whole‐course‐repair system provides a novel platform for combined DW therapy.

points through a combinatory function represent a promising direction for DW treatment.
Neuropathy has been demonstrated to be closely associated with impaired wound healing in diabetes. [6] Diabetic neuropathy was found in almost 90% of DWs, and sensory neuropathy can result in neuropathic pain and/or loss of sensation, significantly increasing the risk of lower extremity amputation. [6] Emerging evidence suggests that therapeutic strategies for the promotion of neurogenesis in DWs can beneficially accelerate wound healing. [7] However, neurogenesis is not an independent event, and it is tightly mediated by both complicated molecular mechanisms from surrounding neural and non-neural cells. [8] The cellular crosstalk between neural cells and vascular endothelial cells (ECs) plays a key role in neurogenesis. ECs can promote neurogenesis by supporting the high metabolic demands of neural cells, and the bioactive molecules released by the neuroepithelium can also drive the growth and maturation of blood vessels. [9] Both neural cells and ECs constitute a "neurovascular niche" that enhances cell function and accelerates DWs healing. [10] Recently, a variety of therapeutic strategies were developed to facilitate DW healing through enhancement of neurogenesis and/or angiogenesis. [11] However, neurogenesis and angiogenesis were regarded as two independent events, and there is a lack of knowledge on the crucial role of neurogenesisangiogenesis crosstalk in DW healing. Hence, novel strategies that coordinate the crosstalk between neurogenesis and angiogenesis under pathological immune microenvironments in DWs should be investigated to facilitate rapid wound healing.
Furthermore, it has been well-documented that timely shifting of the macrophage phenotype towards the M2 phenotype can convert the inflammation phase into the proliferation phase, leading to accelerated wound healing. [12] Recently, multiple agents have been selected as candidates for macrophage reprogramming and small extracellular vesicles (sEVs) are most favorable. [13] Kamerkar et al. [14] described engineered sEVs carrying an antisense oligonucleotide to regulate the reprogramming of tumor-associated macrophages. In contrast to sEVs derived from mammalian cells, medicinal plants-derived sEVs have many more sources, low immunological risks, and costefficient production. [15] Therefore, medicinal plant-derived sEVs may have considerable application potential in DW therapies.
As a kind of medicinal plant-derived sEV, ginseng-derived sEVs (G-sEVs) were successfully extracted and demonstrated to facilitate DW healing. [16] More important, G-sEVs were demonstrated to have a direct effect on neural differentiation of bone marrow-derived mesenchymal stem cells (BMSCs), which has been widely explored in the context of treating neural injuries. [17] However, a limited number of BMSCs in wound sites seriously restricts the efficacy of G-sEVs in promoting neurogenesis in DWs. Fortunately, it was reported that macrophages could capture extracellular magnesium ions (Mg 2+ ), thereby secreting cytokines such as interleukin-8 (IL-8) and C-C motif chemokine ligand 5 (CCL5), which in turn recruit BMSCs into injured tissues. [18] Thus, it holds that combining the strong efficacy of G-sEVs in promoting neurogenesis and the role of extracellular Mg 2+ in recruiting BMSCs will provide a therapeutic advantage in DW therapy.
Currently, a variety of versatile hydrogel dressings were designed for DW therapies, owing to the evident stability, high efficiency of drug encapsulation, good biocompatibility, and facilitation of local use in wound management. [4b,19] Of note, multifunctional dressings with antibacterial, pro-angiogenic, anti-inflammation, and reactive oxygen species (ROS)-scavenging functions appear to be an emerging trend of hydrogelbased therapy for DW healing. [19f,20] However, DW healing is a dynamic and complicated process including four continuous and overlapping phases, all of which, particularly the inflammation and proliferation phase, play prominent roles in tissue repair and regeneration. Unfortunately, the emerging multifunctional hydrogel dressings solely focus on the mediation of a single phase (commonly on inflammation or proliferation phase) of the healing process. Although certain beneficial outcomes were obtained in the utilization of these dressings, their universal efficacy and applicability remained limited due to the ignorance of the multi-factorial characteristic of DW.
Inspired by this biological knowledge, we have designed an in situ injectable hydrogel to deliver engineered G-sEVs as a wholecourse-repair dressing system. This hydrogel is conveniently fabricated by one-step mixing of the hyaluronic acid (HA)-adipic acid dihydrazide (ADH)/Mg solution with oxidized sodium alginate (OSA) solution and engineered G-sEVs suspension to form HA-ADH/OSA@Mg@sEVs based on the mild Schiff base reaction between HA-ADH and OSA and the chelation between Mg 2+ and OSA. In this way, our hydrogel dressing can easily achieve "Package firstly, injection when needed", which can meet the requirements for commercialization. During the healing process, the chelated Mg 2+ ions by OSA inside hydrogel slowly released for almost 7 days to promote the recruitment of BMSCs to the wound site, which could be induced toward neurogenic differentiation by the engineered G-sEVs. Concurrently, a pro-healing immune microenvironment could be achieved by G-sEVs-mediated macrophage reprogramming. At the proliferation stage, enhanced angiogenesis was observed due to the synergistic effect of the differentiated neural cells and the released Mg 2+ . This design puts forward the concept of supportive neurogenesis-angiogenesis crosstalk throughout the whole healing process to understand how it can promote the limited efficacy of current DW therapies using a whole-course-repair strategy.

Construction and Characterization of G-sEVs DM
In the local immune microenvironment of DW, excessive inflammatory response, caused by continuous infiltration of macrophages, is the main reason for the prolonged and unhealed wound. [19a] At the inflammation stage, the exudation of M1 macrophages is critical to establish resistance against the invasion of pathogenic bacteria. However, excessive infiltration of M1 macrophages can lead to a prolonged inflammatory phase, hindering the progression of wound healing. Thus, timely shifting of the macrophage phenotype towards M2 can convert the inflammation phase into the proliferation phase, leading to accelerated wound healing. [12] It was recently reported that didymin (DM) is capable of switching the macrophage phenotype from M1 toward M2 in ulcerative colitis. [21] As an ideal delivery vehicle for small molecules, sEVs have attracted accumulative interest in biomedical applications owing to their eminent advantages of good biocompatibility, facilitating intercellular communication, improving drug efficacy, and reducing side effects. [22] Therefore, we constructed engineering sEVs by the encapsulation of DM into G-sEVs to form G-sEVs DM for stable and sustained therapy of DW. The DM loading efficiencies for three different loading methods, including electroporation, freeze-thaw cycles, and direct incubation were measured using an indirect method by quantifying the free DM in the solution. Direct incubation of DM with G-sEVs for 1 h led to 8.10 ± 0.87% of the DM entrance to the G-sEVs, and repeated freeze-thaw procedure increased the encapsulation efficiency to 9.14 ± 2.35%, and electroporation led to an encapsulation rate of 40.70 ± 1.60%, indicating electroporation is a preferred method for encapsulation of DM into G-sEVs ( Figure S1A, (Supporting Information). Next, the morphology of G-sEVs and G-sEVs DM was studied by photographing them with transmission electron microscopy (TEM), and both exhibited a saucer-like morphology ( Figure S1B, Supporting Information). The distribution of particle size was measured by dynamic light scattering (DLS). The average diameters of G-sEVs and G-sEVs DM were 86.70 ± 6.61 and 77.20 ± 6.06 nm, respectively ( Figure S1C, Supporting Information). Flow cytometry (FCM) results further confirmed the surface markers of sEVs, such as CD9, CD63, and CD81 ( Figure S1D, Supporting Information).
The physiological stability of nanocarriers is of great importance for their application as a carrier of small molecules. To determine the stability of G-sEVs DM under physiological conditions, we monitored the changes in their particle diameter and zeta potential in 50% fetal bovine serum (FBS) for 7 consecutive days using DLS. The hydrodynamic diameter of G-sEVs DM slightly changed from 114.67 ± 3.09 nm to 122.67 ± 1.25 nm, and the surface zeta potential fluctuated from −21.33 ± 1.70 mV to −18.00 ± 2.16 mV ( Figure S1E, Supporting Information). In addition, after 24 h of incubation in phosphate buffer saline (PBS) and in FBS, the cumulative release of the DM from G-sEVs reached ≈96% and 99%, respectively ( Figure S2, Supporting Information). These results indicated that G-sEVs DM are stable under physiological conditions and can release DM in a sustained manner to induce effective regenerative responses.

G-sEVs DM Induces Macrophage to Switch from M1 to M2 Phenotype
To evaluate the effect of G-sEVs DM on macrophage reprogramming, we collected bone marrow-derived macrophages (BMDMs) from C57BL/6 mice and checked the regulatory role of G-sEVs DM on macrophage polarization in vitro. Interestingly, G-sEVs DM did not appear to affect the ratio of M2 macrophages in any of the groups pre-treated with IL-4 (20 ng mL −1 ) or IL-13 (20 ng mL −1 ) ( Figure S3A,B, Supporting Information). In contrast, a marked decrease in the ratio of F4/80 + CD86 + macrophages in the lipopolysaccharide-(LPS; 100 ng mL −1 ) and interferon-γ (IFN-γ; 20 ng mL −1 )-treated group was observed after co-incubation with G-sEVs DM for 48 h ( Figure S3C,D, Supporting Information). Next, the BMDMs were directed towards M1 polarization for 24 h and then treated with either PBS, G-sEVs, or G-sEVs DM for another 24 h. The results revealed that compared with the PBS-treated group, G-sEVs DM significantly induced the macrophages towards M2 phenotype polarization, while no obvious change in the M2 macrophage proportion was detected in the cells treated with PBS and G-sEVs ( Figure 3E,F). These results confirm that G-sEVs DM could induce reprogramming of macrophages from M1 to M2 phenotype due to the presence of DM in their structure.

Characterization of HA-ADH/OSA@Mg@sEVs
To satisfy the simultaneous delivery of Mg 2+ and G-sEVs DM , we developed an injectable and self-healing hydrogel using HA-ADH and OSA (Scheme 1). The successful synthesis of HA-ADH and OSA was proved by the 1 H nuclear magnetic resonance spectroscopy ( 1 H NMR) as shown in Figure S4 (Supporting Information). The preparation of the designed HA-ADH/OSA@Mg hydrogel required only one step by mixing the solution of HA-ADH/Mg and the solution of OSA in equal volume. This one-step preparation introduced two kinds of crosslinking to form the final HA-ADH/OSA@Mg hydrogel. One was the dynamic chemical crosslinking based on the Schiff base reaction between the hydrazide groups of HA-ADH and the aldehyde groups of OSA, and the other one was the dynamic physical crosslinking based on the chelation between the G blocks of OSA and Mg 2+ . The encapsulation of G-sEVs DM could also be successfully conducted by a convenient one-step process and the mild gelation condition ensured that there was no negative effect on the loaded G-sEVs DM . In addition, we studied the influence of sEVs' encapsulation on the hydrogel formation. As shown in Figure S5 (Supporting Information), the introduction of sEVs showed no significant effect on the gelation time ( Figure S5A,B, Supporting Information) and microstructure ( Figure S5C, Supporting Information) of the hydrogel, indicating that the HA-ADH/OSA@Mg hydrogel is a stable carrier for sEVs. At the same time, the Mg 2+ loading could be guaranteed by the chelation with OSA.
The HA-ADH/OSA hydrogel was prepared by mixing HA-ADH and OSA solution with the same concentration in equal volume. Firstly, we explored the effect of polymer concentration on the hydrogel gelation. We selected three groups of HA-ADH/OSA hydrogel with different final concentrations of HA-ADH and OSA as 1%, 2%, and 3%. As shown in Figure 1A,B, the increase of the HA-ADH and OSA concentration significantly decreased the gelation time. When the final concentration of HA-ADH and OSA increased from 1% to 2%, the gelation time decreased dramatically from 165 ± 15 min www.advmat.de www.advancedsciencenews.com to 25 ± 5 min. When the concentration of HA-ADH and OSA reached 3%, the average gelation time reached 7.5 ± 2.5 min. This should be contributed to the fact that the high HA-ADH and OSA concentrations may increase the reaction probability of hydrazide and aldehyde groups. Based on the 3% group, we further tested the effect of Mg 2+ addition on the hydrogel preparation. We fixed the final concentration of Mg 2+ at 75 mM. Our experiments ( Figure 1A,B) demonstrated the negligible influence of Mg 2+ on hydrogel gelation time, which indicated that the Schiff base reaction between HA-ADH and OSA played the dominant role in the crosslinking formation of the hydrogels. The internal microporous structure of HA-ADH/OSA hydrogel was visualized by Scanning Electron Microscope (SEM) ( Figure 1C). It exhibited that the higher concentration of HA-ADH and OSA could lead to a denser hydrogel network and the existence of Mg 2+ had no obvious influence on the hydrogel microstructure. The elemental mapping by Energy Dispersive Spectroscopy (EDS) ( Figure 1D) confirmed the homogeneous Mg 2+ distribution inside the HA-ADH/OSA@Mg hydrogel.
Since the microstructure changes caused by the alteration of HA-ADH and OSA concentration may affect the mechanical properties of the hydrogel, we tested the rheological time sweep for these three HA-ADH/OSA hydrogels (1%, 2%, and 3%) and HA-ADH/OSA@Mg hydrogel (3%+Mg). Figure 2A showed that the storage modulus (G′) greatly increased when the concentration of HA-ADH and OSA increased from 1% to 3%. This can be contributed to the increase of the crosslinking density inside hydrogels due to the increment of solid content. The addition of Mg 2+ could also lead to the further strengthening of hydrogel ( Figure 2B). This phenomenon strongly confirmed that the chelation between OSA and Mg 2+ could act as an efficient crosslinking to improve the mechanical property of hydrogel. This result also indicated the good incorporation of Mg 2+ inside hydrogel, which is fundamental for the long-term release of Mg 2+ from the hydrogel to attract stem cells.
Adv. Mater. 2023, 35, 2212300 Scheme 1. Schematic illustration of the beneficial role of HA-ADH/OSA@Mg@sEVs hydrogel. The preparation of the designed HA-ADH/OSA@Mg hydrogel required only one step by mixing the solution of HA-ADH/Mg and the solution of OSA in equal volume. The encapsulation of G-sEVs DM also could be included in this convenient one-step process and the mild gelation condition ensured that there was no negative effect on the loaded G-sEVs DM . At the same time, the Mg 2+ loading also could be guaranteed by the chelation with OSA. During the inflammation stage, the hydrogel promotes the recruitment of bone marrow derived-mesenchymal stem cells to the wound site through the released Mg 2+ , which will be induced toward neurogenic differentiation by the simultaneously released G-sEVs DM . Concurrently, a pro-healing immune microenvironment can be created by G-sEVs DM -induced macrophage reprogramming. At the proliferation stage, robust angiogenesis is further enhanced by the synergistic effect of the newly differentiated neural cells and the released Mg 2+ . Ultimately, a mutually supportive cycle of neurogenesis-angiogenesis will be formed at the wound.

www.advmat.de www.advancedsciencenews.com
We further compared the compressive mechanical properties between HA-ADH/OSA hydrogel and HA-ADH/OSA@Mg hydrogel ( Figure 2C,D; Figure S6, Supporting Information). The compressive results were consistent with the rheological results. Both the increase of the HA-ADH and OSA concentration and the addition of Mg 2+ were in favor of promoting Young's modulus and the breaking stress. Because these two kinds of crosslinking inside the HA-ADH/OSA@Mg hydrogel were dynamically reversible, this hydrogel had a good shearthinning property ( Figure 2E). This means that under a high shear force, the HA-ADH/OSA@Mg hydrogel can convert from "gel" to "sol" and when the shear force is removed, this system is able to reform into the "gel" state. This was the basic principle of the injectable property of our hydrogel ( Figure 2F; Video S1). At the same time, the reversible crosslinking also guaranteed the self-healing property of our hydrogel ( Figure 2G; Video S2). Finally, we characterized the sustained release of Mg 2+ of the HA-ADH/OSA@Mg hydrogel. As shown in Figure 2H, the release of Mg 2+ from the hydrogel was at a sustained rate, which makes it a desirable platform to recruit BMSCs to the wound sites for the long term through Mg 2+ release.

HA-ADH/OSA@Mg@sEVs Recruits BMSCs to the Wound Sites
To test the ability of HA-ADH/OSA@Mg@sEVs to recruit BMSCs, a transwell migration assay was utilized. The results suggest that both HA-ADH/OSA@Mg@sEVs and HA-ADH/ OSA@Mg can recruit BMSCs. However, no obvious cell migration was detected in the PBS-, HA-ADH/OSA-, and HA-ADH/ OSA@sEVs-treated groups (Figure 3A,B). Additionally, the protein expressions of chemokines, IL-8 and CCL5, in different groups were measured by quantitative real-time polymerase chain reaction (qRT-PCR). A significant increase in the above-mentioned chemokines was found in the groups that the formulation had Mg 2+ ( Figure 3C,D), suggesting that the HA-ADH/OSA@Mg@sEVs may be able to attract BMSCs to  www.advmat.de www.advancedsciencenews.com the wound sites through the release of Mg 2+ . Previous strategies that induce the differentiation of BMSCs into functional neural cells have proven to be useful for neurogenesis in DWs. Given this, we further investigated the effect of HA-ADH/OSA@ Mg@sEVs in the neural differentiation of BMSCs. It is beneficial in tissue engineering to develop controlled-release formulations that deliver multiple bioactive factors to construct an appropriate microenvironment. [23] Therefore, the simultaneous loading of diverse bioactive factors within a hydrogel is one of the feasible methods to achieve enhanced cellular responses for effective extracellular matrix (ECM) biosynthesis. However, due to the limitation of permeability and network interactions with bioactive molecules in hydrogels, ideal loading and simultaneous delivery of multiple drugs with different properties using hydrogels scaffold remain a great challenge. [24] In our study, it was not known whether Mg 2+ will affect the release behavior of G-sEVs DM when encapsulated in the hydrogel. Hence, the release kinetics of G-sEVs DM from the hydrogels was examined. As displayed in Figure 3E, ≈60% of G-sEVs DM remained in HA-ADH/OSA after 20 days of incubation in PBS and the cumulative release of G-sEVs DM from HA-ADH/OSA@sEVs was higher than HA-ADH/OSA. These results indicate that HA-ADH/OSA@Mg@sEVs could be a suitable delivery vehicle for G-sEVs DM in a sustained releasing manner. Next, the proneurogenesis effect of HA-ADH/OSA@Mg@sEVs was tested in vitro. As shown in Figure 3F,G, HA-ADH/OSA@Mg@sEVs effectively enhanced the neural differentiation of BMSCs, with significant nestin and β3-tubulin marker expressions observed on day 7 post-treatment. In addition, some neuron-like extensions, which are the morphological characteristics of neuronlike cells, were captured under phase-contrast microscopy ( Figure 3H; Figure S7

HA-ADH/OSA@Mg@sEVs Induces Macrophage Reprogramming
An anti-inflammatory microenvironment can shorten the inflammation phase and shift it into the proliferation phase in DWs. We thus investigated the effects of HA-ADH/OSA@Mg@ sEVs on reprogramming macrophages by examining the regulatory role of HA-ADH/OSA@Mg@sEVs in macrophage polarization in vitro. The results indicate that G-sEVs DM released from HA-ADH/OSA@Mg@sEVs did not affect the ratio of M2 macrophages in the BMDMs pre-treated with IL-4 (20 ng mL −1 ) and IL-13 (20 ng mL −1 ) (Figure 4A,B)   subsequently treated with either PBS, HA-ADH/OSA@Mg, or HA-ADH/OSA@Mg@sEVs for another 24 h. The results showed that the hydrogel loaded with G-sEVs DM could increase the ratio of M2 macrophages while no significant increase in M2 polarization was found in non-G-sEVs DM loaded groups ( Figure 4E,F). These results reveal that in a relatively short treating period (i.e., 24 h), although the markers of the M2 phenotype gradually appear, the intracellular markers of the M1 phenotype do not completely diminish in the HA-ADH/OSA@ Mg@sEVs group. However, longer treatment (48 h) induced a remarkable decrease of M1 markers, indicating macrophage reprogramming by HA-ADH/OSA@Mg@sEVs through the release of G-sEVs DM .

BMSCs-derived Neural Cells and Mg 2+ Synergistically Promote Angiogenesis
After the inflammation phase, the wound will enter the proliferation phase. Local blood supply to the wound during the proliferation phase is a factor that determines successful healing. [25] Thus, we next evaluated the pro-angiogenic effects of the wholecourse-repair system in vitro and in vivo. Human umbilical vein endothelial cells (HUVECs) were divided into five groups according to the following treatments: PBS, HA-ADH/OSA, HA-ADH/OSA@Mg, HA-ADH/OSA@sEVs, and HA-ADH/ OSA@Mg@sEVs. As shown in Figure 5A,B, the Edu staining assay indicated that HUVECs in the Mg 2+ -treated groups had a higher proliferation rate than the non-Mg 2+ -treated groups. Similarly, the cell scratch assay and tube formation assay revealed enhanced cell migration and angiogenic effects in Mg 2+ treated groups ( Figure S10A-E, Supporting Information). Next, a murine DW model was created in vivo, and the small animal doppler examination of the wound was performed after 10 days. The results suggest that there was better blood perfusion in the Mg 2+ -treated groups compared with the non-Mg 2+ -treated groups ( Figure 5C,D). Regenerative wound tissues were then collected from different groups and CD31 was stained by immunohistochemistry (IHC). Histological evaluation of the stained tissues revealed that there were a greater number of newly formed CD31 + blood vessels in the regenerative areas in the Mg 2+ -treated groups compared with the non-Mg 2+ -treated groups ( Figure 5E,F). We next assessed the effect of the newly differentiated neural cells on HUVECs. A cell coculture system was established using a trans-well device, in which the BMSCs-induced neural cells were put in the upper chamber and HUVECs in the lower chamber ( Figure 5G). As shown in Figure S10F,G (Supporting Information), in the presence of newly differentiated neural cells, HUVECs migration and tube formation were significantly enhanced compared with the PBS-treated group. The cells in the lower chamber and proteins were extracted, and the Enzyme-Linked Immunosorbent Assay (ELISA) results showed that the expressions of CGRP, NGF, and CRH were significantly higher when co-cultured with the newly differentiated neural cells ( Figure 5H). Moreover, to test whether angiogenesis can promote neurogenesis, we constructed a co-culture system using a trans-well device, in which the HUVECs were put in the upper chamber and BMSCs in the lower chamber. After incubation for 7 days, significant expres-sions of nestin and β3-tubulin were observed in the co-culture group compared with the control group (cultured with PBS). This result indicates angiogenesis can effectively direct neurogenesis in vitro ( Figure S11, Supporting Information). All these data indicate that the BMSCs-derived neural cells and Mg 2+ released from the hydrogel system have the potential to synergistically promote angiogenesis, which can in turn beneficially drive the BMSCs toward neural differentiation.

HA-ADH/OSA@Mg@sEVs Accelerates DW Healing In Vivo
To determine whether the whole-course-repair system could accelerate DW healing, wounds were created in a diabetic mouse model and the effects of the hydrogel system were evaluated. The wound closure rate in the HA-ADH/OSA@Mg@ sEVs group was significantly higher than in the other groups (Figure 6A-C). Next, the wound tissues were collected from the different treatment groups on day 14 post-wounding, and the elevation in the level of Mg 2+ in the tissues collected from the HA-ADH/OSA@Mg@sEVs group was verified by inductively coupled plasma mass spectrometry (ICP-MS) ( Figure 6D). Furthermore, the levels of the neuropeptide, including corticotropin-releasing hormone (CRH), nerve growth factor (NGF), and calcitonin gene-related peptide (CGRP) of the skin tissues were assessed by immunofluorescence staining, indicating that the neuropeptides were markedly enriched in HA-ADH/OSA@sEVs and HA-ADH/OSA@Mg@sEVs groups ( Figure 6E,F). Moreover, the formation of new epithelial and granulation tissue was further evaluated by histological examination. Wounds in hydrogel-treated groups showed a markedly greater thickness of granulation tissue than the untreated control group, with the HA-ADH/OSA@Mg@sEVs-treated group showing the highest degree of granulation tissue formation ( Figure 6G; Figures S12A,B and S13A, Supporting Information). Masson's trichrome staining showed both angiogenesis and collagen deposition in the hydrogel-treated wounds, with the HA-ADH/OSA@Mg@sEVs group showing the highest degree of both ( Figure 6H; Figures S12C and S13B, Supporting Information). In addition, the immunofluorescence results indicated that HA-ADH/OSA@Mg@sEVs can effectively recruit BMDMs to the wounds and enhance neurogenesis in vivo ( Figure S14, Supporting Information). These findings in the mouse model of DWs demonstrate the robust pro-healing effects of HA-ADH/OSA@Mg@sEVs on DWs.

HA-ADH/OSA@Mg@sEVs has a Good Biocompatibility
The above results suggest that our whole-course-repair system is highly efficacious in the treatment of DWs. However, systemic toxicity and the release of Mg 2+ in plasma would potentially impede their clinical applications. Accordingly, during the entire treatment period, the alterations in mouse body weight and the Mg 2+ levels in the plasma were regularly tested. In addition, at the end of treatment, the blood and vital organs were collected from the mice for routine blood analysis and H&E examination. As shown in Figure S15A (Supporting Information), there were no significant alterations in the body weight www.advmat.de www.advancedsciencenews.com Adv. Mater. 2023, 35, 2212300 of mice treated with HA-ADH/OSA, HA-ADH/OSA@Mg, HA-ADH/OSA@sEVs, and HA-ADH/OSA@Mg@sEVs, except for that of the mice treated with PBS that showed a significant decrease, presumably owing to the existence of long-term Figure 5. BMSCs-derived neural cells and Mg 2+ synergistically enhance HUVECs differentiation. A,B) The proliferation of HUVECs in different groups was tested by Edu staining. Scale bar: 20 µm. C,D) Wounds were created in a diabetic mouse model and divided into five groups according to the different treatments. At day 10 post-wounding, the blood perfusions among the five groups were detected by small animal doppler examination. E,F) Skin tissue was collected from the diabetic mice in the five groups, and CD31 IHC assay was conducted to detect the newborn blood vessel. Scale bar: 100 µm. G) The schematic illustration of the co-culture system. The BMSCs-derived neural cells were in the upper chamber, and the HUVECs were in the lower chamber. H) The ELISA results of CGRP, NGF, and CRH in the different groups. **p < 0.01, ***p < 0.001, ****p < 0.0001.

www.advmat.de www.advancedsciencenews.com
Adv. Mater. 2023, 35, 2212300 infections and hyperglycemia in the wound area. The Mg 2+ level in the plasma was tested by ICP-MS analysis. No significant alterations were found in the hydrogel-treated groups either on day 7 or at end of the treatment ( Figure S15B, Supporting Information), indicating that this hydrogel system has minimal risk of Mg 2+ poisoning. This conclusion was further verified by the blood analysis ( Figure S15C-E, Supporting Information). The values of three representative hematology markers (red blood cell count (RBC), hemoglobin (HGB), and white blood cell count (WBC) were all within the normal ranges. Moreover, as shown in Figure S16 (Supporting Information), the H&E staining of the major organs from the HA-ADH/OSA@Mg@ sEVs group further demonstrated no obvious tissue damage found in the HA-ADH/OSA@Mg@sEVs-treated mice compared with healthy mice, indicating good biocompatibility of HA-ADH/OSA@Mg@sEVs. Collectively, the above results all suggest the good biocompatibility of HA-ADH/OSA@Mg@ sEVs, thus giving them a great potential for serving as a wholecourse-repair system to be administrated on DWs.

The Underlying Mechanisms of DW Repair Enhanced by HA-ADH/OSA@Mg@sEVs
To uncover the underlying mechanisms of the accelerated wound healing process, we performed mRNA sequencing analysis of DWs treated with or without HA-ADH/OSA@ Mg@sEVs at day 3 and day 7 post-wounding. As shown in Figure S17A,B (Supporting Information), the expression profiles of the HA-ADH/OSA@Mg@sEVs group were markedly different from the profiles of the PBS-treated group. We found 235 and 95 genes with significantly increased transcript levels in HA-ADH/OSA@Mg@sEVs-treated group compared with the PBS-treated group on days 3 and 7, respectively. There were 272 genes that exhibited significantly decreased transcript levels in HA-ADH/OSA@Mg@sEVs-treated group on day 3 and 287 on day 7 ( Figure S17C, Supporting Information). Through Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEGs), we found that the downregulated DEGs in HA-ADH/OSA@Mg@sEVs-treated group at day 3 were distinctly enriched in the "cellular response to interferon-gamma", "response to interferon-gamma", "cell activation involved in immune response", and "inflammatory response" (Figure S17D, Supporting Information), and the upregulated DEGs at day 7 were enriched in "regulation of blood vessel size", "positive regulation of neuron differentiation", and "positive regulation of angiogenesis" (Figure S17E, Supporting Information). The statistics of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of DEGs indicated that HA-ADH/ OSA@Mg@sEVs downregulated hypoxia-inducible factor-1 (HIF-1), nuclear factor kappa-B (NF-κB), and peroxisome proliferator-activated receptor (PPAR) signaling pathways at day 3 ( Figure S17F, Supporting Information) and upregulated phosphoinositide 3-kinase (PI3K) signaling pathway and vascular endothelial growth factor (VEGF) signaling pathway, which were previously demonstrated in neurogenesis and angiogenesis at day 7 ( Figure S17G, Supporting Information). Together, these results indicate that down-regulated DEGs predominated in HA-ADH/OSA@Mg@sEVs-treated-group at day 3 probably owing to the anti-inflammatory effect of the released G-sEVs DM and up-regulated DEGs predominated at day 7 probably due to the activation of neurogenesis and angiogenesis in HA-ADH/ OSA@Mg@sEVs-treated-wounds.

Discussion
Currently, sEVs-based hydrogel dressings are widely used for promoting DW healing owing to their excellent anti-inflammatory, antimicrobial, and pro-angiogenic properties. [26] Emerging evidence now suggests that hydrogel dressings could facilitate the rapid healing process by functioning as an ideal carrier for the sustained release of sEVs, which could be taken by the target cells where they exert regulatory effects. [27] However, most of the previous studies focused solely or primarily on the antibacterial properties and proangiogenic functions of the dressings, [19f ] leaving neural restoration in wound healing largely unaddressed, particularly the crosstalks between neural cells and vascular endothelial cells. It also should be considered that wound healing is a complicated process that involves four continuous and overlapping phases, and the most successful outcome may rely on addressing all the phases.
Our group has developed a novel hydrogel dressing based on the whole-course-repair strategy that targets neurogenesisangiogenesis whilst providing a beneficial immune microenvironment at the wound site to promote neurogenesis and angiogenesis during the four wound healing processes. During the hemostasis and inflammation phase, the dressing is designed to recruit BMSCs to the wound sites through the sustained release of Mg 2+ , and the BMSCs could subsequently be directed towards neurogenic differentiation by G-sEVs DM , which can simultaneously provide an anti-inflammation microenvironment by reprogramming macrophages from the M1 phenotype to M2 phenotype. During the proliferation phase, the newly differentiated neural cells and the released Mg 2+ can synergistically promote angiogenesis to restore the blood perfusion of the wound site, which can in turn enhance neurogenesis activity. Ultimately, a synergistic cycle of neurogenesis-angiogenesis can be established at the wound site and facilitate rapid wound healing. Recent studies have reported the promising potential of G-sEVs as an efficient agent for promoting neurogenesis, and DM's strong reprogramming effects on macrophages. [16,21] In the present study, DM was loaded into G-sEVs using an engineering sEVs strategy, and we verified the potent neurogenesis and macrophage reprogramming effects of the engineered sEVs. We then incorporated G-sEVs DM into HA-ADH/OSA/Mg hydrogel and demonstrated robust pro-healing effects of this sEVs-loaded hydrogel system. This system likely acts as a chemical catalyst to activate the cellular communication between neural cells and vascular endothelial cells throughout the whole healing process of DW. Indeed, local injection of HA-ADH/ OSA@Mg@sEVs to the wound site of the diabetic mouse model resulted in markedly reduced inflammation, enhanced neurogenesis and angiogenesis, and accelerated wound closure of the wound. HA-ADH/OSA@Mg@sEVs at the pro-healing dose was well tolerated in wound healing-related functional cells (including endothelial cells, fibroblasts, etc.), and no signs of systemic toxicity were observed in the mice after local application of HA-ADH/OSA@Mg@sEVs. Our study demonstrated the promising prospect of HA-ADH/OSA@Mg@sEVs as a highly effective and safe pro-healing dressing for the treatment of DWs.
Wound repair and regeneration after injury require a dynamic shift in the immune microenvironment from a pro-inflammatory microenvironment to neutralize injury and remove dead or injured tissue to an anti-inflammatory microenvironment to facilitate migration and proliferation of reparative cell types and increase vascularization and nutrient supply. [28] Therefore, both pro-and anti-inflammatory reactions are critical to successful tissue repair and regeneration following injury. Macrophages participate in all phases of wound healing. [29] Classically activated M1 macrophages are induced by bacterial endotoxin, peptidoglycan, or a variety of intracellular stimuli, and are proinflammatory and contribute to host defense and clearance of damaged tissue. [30] Alternatively activated M2 macrophages are anti-inflammatory and can secrete various molecules to promote wound repair and tissue remodeling. [31] Unlike in acute wounds, the switch from M1 to M2 macrophages is dysregulated in chronic wounds. Around 80% of cells at the edges of chronic wounds are M1 macrophages, and this is considered to cause persistent inflammation in chronic wounds, leading to insufficient healing or scar formation. [32] Disruption in the switch from M1 macrophages to M2 macrophages is a hallmark of DWs. [25] Here, we found that treatment with HA-ADH/ OSA@Mg@sEVs led to a switch of the macrophage phenotype from pro-inflammatory to anti-inflammatory in DWs. Two possible mechanisms may contribute to this reprogramming function of the dressing: Firstly, the G-sEVs DM released from the hydrogel system may be taken by the M1 macrophages, and the active DM from the sEVs may then induce the phenotype switch. Secondly, we hypothesized that the restoration of neural functions may support the macrophage-reprogramming effects of HA-ADH/OSA@Mg@sEVs given the important role of the impaired sensory function in exacerbating inflammation in the wound sites. However, it should be noted that incubation with HA-ADH/OSA@Mg@sEVs resulted in induction of macrophage phenotype reprogramming, indicating that HA-ADH/OSA@Mg@sEVs can exert a direct reprogramming effect independent of its pro-neurogenesis effects.
The pathological microenvironment of DW can induce a non-enzymatic glycosylation reaction between the aldehyde group of sugar and amino group of protein, and eventually, lead to the accumulation of advanced glycation end products (AGE) which are prevalent in the diabetic vasculature. [33] The continuous accumulation of AGE at the wound sites can establish a long-term hyperglycemic microenvironment around the wound, and impair cell proliferation, migration, and differentiation. [34] Whether the released G-sEVs DM can be efficiently taken by the BMSCs and macrophages in the hyperglycemic microenvironment of the wounds needs future evaluation. In addition, our study focused on local neural restoration of DW; however, restoration of peripheral neural function may deliver signal transduction to the brain center and exert emotional mediation. Considering the evidence that supports the beneficial role of positive mood in promoting cell proliferation and differentiation, [35] further studies exploring the crosstalk between the peripheral and central nervous systems during the wound healing process may be useful in developing more systematic repair strategies.

Conclusion
In summary, we present a whole-course-repair system (HA-ADH/OSA@Mg@sEVs) based on engineered G-sEVs, delivered by an in situ injectable, self-healing, and bio-adhesive hydrogel. Unlike most of the currently available hydrogel dressings that focus on only one phase of DW healing, the designed HA-ADH/OSA@Mg@sEVs have shown to exert pro-healing effects throughout the inflammation and proliferation stages of DW healing. In this study, we verified the therapeutic efficacy of HA-ADH/OSA@Mg@sEVs in promoting wound closure in STZ-induced diabetic mice. Furthermore, we elucidated the mechanisms underlying the observed neurogenesis-angiogenesis crosstalk and macrophage reprogramming promoted by G-sEVs DM and Mg 2+ released from this system in the inflammation and proliferation phases of DW healing. Moreover, we demonstrated good biocompatibility of HA-ADH/OSA@Mg@sEVs after local injection in vivo. Our findings suggest that the HA-ADH/OSA@Mg@sEV hydrogel is simpler to prepare and easy to use, showing its great commercial potential as a diabetic wound dressing in the future.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.