Skin-mimetic bilayer hydrogel enhances spatiotemporal coordination of neuro-immune-vascular interactions to accelerate diabetic wound healing

Abstract

Diabetic wound healing is a complex spatiotemporal process that requires stage-specific interventions to address disrupted neuro-immune interactions and impaired angiogenesis. However, achieving such precise coordination with a single regenerative dressing remains a considerable challenge. Drawing inspiration from native skin physiology, we have developed an intelligent, conductive, skin-mimetic bilayer hydrogel that spatially segregates functions and temporally orchestrates the repair process. This design features a robust, anisotropic upper layer that provides protection and serves as an efficient conduit for electrical stimulation, combined with a responsive lower layer that adheres to the wound and enables on-demand drug delivery. Specifically, the lower hydrogel releases calcitonin gene-related peptide in response to the early inflammatory microenvironment, effectively suppressing the pro-inflammatory M1 macrophage phenotype and promoting its transition to the pro-repair M2 phenotype. Subsequently, the conductive upper layer sustains the release of magnesium ions and synergizes with electrical stimulation to significantly enhance endothelial cell migration and tube formation via activation of the VEGF signaling pathway. Transcriptomic analysis reveals that this combination fosters a pro-regenerative microenvironment by enriching pathways related to extracellular matrix organization and angiogenesis. This skin-mimetic structure-to-function design offers a practical strategy for staged, precise wound repair in diabetes and provides a generalizable framework for chronic tissue regeneration.

Graphical abstract

This study presents a skin-mimetic bilayer hydrogel inspired by natural skin, comprising an anisotropic tough conductive upper layer and a responsive adhesive lower layer. It sequentially releases CGRP and Mg²⁺ upon electrical stimulation to coordinate neuro-immune interactions and neurovascularization spatially and temporally. This neuro-immune-vascularization approach closely replicates the biological cascade of diabetic wound healing.

Highlights

• •A skin-mimetic bilayer hydrogel enables spatially segregated, time-sequenced therapy for diabetic wound healing.
• •Combined adhesive and conductive layers modulate inflammation and regeneration without functional interference.
• •A CGRP-mediated neuro-immune-vascular cascade coordinates healing and angiogenesis in diabetic wounds.

1. Introduction

Diabetes mellitus (DM) is a globally prevalent disease marked by challenging wound management [1]. These wounds are prone to infection and delayed healing due to chronic hyperglycemia, which increases oxidative stress and promotes microbial colonization, creating a microenvironment that hinders tissue repair [2]. DM-related peripheral neuropathy disrupts neuro-immune regulation, prolonging inflammation and causing persistent immune dysfunction [3]. During the proliferative phase, inadequate angiogenesis restricts oxygen and nutrient delivery, impairing neural regeneration and slowing tissue repair. Therefore, stage-specific modulation and remodeling are essential for effective wound healing in DM patients [4].

During the initial inflammatory phase, timely and precise immunomodulation is essential. Such regulation mitigates the deleterious oxidative microenvironment and promotes macrophage polarization, thereby establishing favorable conditions for the initiation of tissue repair. Common strategies include the application of antioxidants and reactive oxygen species (ROS) scavenging systems [[5], [6], [7]], anti-inflammatory small molecules or growth factors [[8], [9], [10]], metal catalysts and nanozymes [[11], [12], [13]], as well as exosomes to attenuate persistent inflammation and restore immune homeostasis [[14], [15], [16]]. However, many of these approaches neglect the interactions within the neuro-immune system. Calcitonin gene-related peptide (CGRP), a key neuropeptide, has the capacity to reprogram macrophage polarization and enhance local microcirculation [17]; nevertheless, it undergoes rapid degradation in vivo [18]. Therefore, its effective utilization necessitates encapsulation with controlled, on-demand release. In this context, stimuli-responsive hydrogels have garnered considerable attention due to their ability to detect pathological cues such as glucose [19], ROS [20], or matrix metalloproteinases [21], provide tissue adhesiveness and conformability [22], and simultaneously exhibit antibacterial activity alongside drug delivery functions [23]. Despite these advances, the healing of DM wounds represents a spatiotemporally complex process, wherein therapeutic demands evolve temporally from early anti-inflammatory modulation to late-stage pro-regenerative support. Spatially, the superficial barrier and the deeper wound bed require distinct structural and functional interventions. Consequently, a single regenerative dressing incorporating multiple functional components often encounters challenges in precisely addressing the multi-stage requirements of DM wound healing in a temporally controlled manner.

During the proliferative phase, angiogenesis plays a critical role in facilitating subsequent neural repair [24]. The formation of new blood vessels restores oxygen and nutrient delivery, removes metabolic waste products, and provides paracrine signals that support regeneration and tissue remodeling, thereby enabling coordinated neurovascular recovery [25]. To enhance vascularization, various strategies have been investigated, including electrical stimulation (ES) [26], administration of pro-angiogenic growth factors [27], responsive release systems [28], provision of structural or orientational guidance to promote endothelial angiogenesis [29], and the use of metal ions to stimulate endothelial migration while improving the regenerative microenvironment [30]. Among these approaches, ES is recognized as a highly efficient, convenient, and precisely controllable therapeutic modality that has been extensively applied in tissue repair [[31], [32], [33]]. ES activates key signaling pathways, such as the vascular endothelial growth factor (VEGF) pathway, thereby promoting endothelial cell proliferation, migration, and tube formation, which significantly accelerates the repair process [34]. However, the efficacy of ES depends on stable electrical coupling to the wound site, necessitating a compatible conductive medium [35]. Conductive hydrogels, characterized by their high water content and tissue compatibility, represent an ideal conductive medium [36]. Nevertheless, in the moist and dynamic wound environment, efficient charge transfer cannot be achieved solely by increasing the concentration of conductive fillers within an isotropic network. Instead, microstructural engineering of conduction pathways is essential. The construction of anisotropic conductive networks, such as aligned channels or fibers, reduces the percolation threshold, shortens charge-transfer pathways, and decreases interfacial impedance, thereby enhancing electron and ion conduction [37]. Simultaneously, this oriented architecture facilitates stress dissipation and mechanical reinforcement, providing biomimetic structural guidance that supports directed cellular growth and functional recovery [38].

An optimal wound dressing functions as a temporary skin substitute; thus, it is essential to accurately replicate the native composition, structure, and function of the skin. The skin's bilayer architecture comprises a dense, robust epidermis that provides protection and a porous, bioactive dermis that facilitates cellular infiltration and exchange [39]. Guided by this principle, we developed a skin-mimetic bilayer hydrogel engineered to correspond with the distinct phases of wound healing (Fig. 1). The upper layer, designed to emulate the protective and supportive functions of the epidermis, offers a durable barrier and a conductive medium. In contrast, the lower layer, which mimics the adhesive and bioactive properties of the dermis, adheres to the wound to ensure conformal contact and enables on-demand drug delivery. Specifically, the upper layer consists of an anisotropic poly (vinyl alcohol) (PVA) hydrogel incorporating poly (3,4-ethylenedioxythiophene):poly (styrene sulfonate) (PEDOT:PSS) as the conductive component and nano magnesium oxide (MgO) to impart antibacterial activity and sustained release of magnesium ions (Mg²⁺). Consequently, this layer provides protection, electrical conductivity, antibacterial effects, and proangiogenic functionality. The anisotropic network is achieved through a process of structural remodeling: initially, weak salting out combined with mechanical training rearranges polymer chains and initiates orientation; subsequently, strong salting out accompanied by stretching enhances the degree of orientation and mechanical performance; finally, annealing increases crystallinity and stabilizes the microstructure, followed by water bath desalting to remove residual salts and improve biocompatibility. This methodology yields a dense, robust, and oriented structure functionally analogous to the stratified epidermis. The lower layer comprises a hydrogel containing quaternary ammonium cations and dynamic boronic ester bonds, loaded with CGRP. The quaternary ammonium cations confer broad-spectrum antibacterial activity and interfacial adhesion, while the dynamic boronic esters enable reversible adhesion and responsiveness to the local microenvironment. As a result, the lower layer exhibits antibacterial activity, responsiveness to ROS and glucose, wet adhesion, and immunomodulatory functions. Its composition and porous network emulate the extracellular matrix-rich and cell-instructive dermal layer. Both layers share antibacterial properties and drug delivery capabilities; however, the upper layer emphasizes electrical conduction and protection, whereas the lower layer prioritizes responsiveness and adhesion. This skin-mimetic bilayer hydrogel design orchestrates therapeutic factors temporally and spatially, establishing a novel paradigm for staged treatment and precise repair of diabetic wounds.

Fig. 1.

The skin-mimetic bilayer hydrogel with anisotropic conductive network and responsive adhesion synergistically modulates macrophage polarization and promotes angiogenesis.

2. Results and discussion

2.1. Structural characterization of the anisotropic upper layer hydrogel

To evaluate whether the upper PVA hydrogel establishes a stable and oriented structure following reconstruction while incorporating functional components, we performed comprehensive multiscale and stepwise characterizations across five distinct sample groups. Initially, scanning electron microscopy (SEM) was utilized to investigate the microstructural morphology. The pristine PVA exhibited an irregular and isotropic network, whereas the processed AP sample demonstrated strip-like features aligned unidirectionally. The incorporation of PEDOT:PSS in APC, nano MgO in M@AP, or both in M@APC preserved the characteristic texture and maintained pronounced orientation (Fig. 2a). Moreover, SEM analysis of the unannealed intermediate samples revealed an oriented structure, confirming the successful reconstruction of polymer chains (Fig. S1). To further elucidate the internal architecture, various sections of M@APC were examined. The transverse section displayed a uniform and interconnected porous network, while the longitudinal section and fracture surface exhibited an oriented structure parallel to the processing direction (Fig. S2–S4). Subsequently, energy-dispersive X-ray spectroscopy (EDS) mapping was employed to verify the spatial distribution of functional phases. Sulfur and magnesium elements were homogeneously distributed within the APC and M@AP hydrogels, respectively, indicating effective and uniform incorporation of conductive and inorganic components within the network (Fig. S5). As shown in Fig. 2b, the reconstructed samples exhibited sharper diffraction peaks relative to PVA, indicative of enhanced crystallinity. Diffraction peaks observed at approximately 38.0°, 50.8°, and 58.6° in M@AP and M@APC correspond to the (101), (102), and (110) planes of magnesium hydroxide, attributed to the hydration of nano MgO [40]. Given the critical role of orientation in conduction pathways, small-angle X-ray scattering (SAXS) analysis was conducted (Fig. 2c–e). M@APC displayed two sharp azimuthal peaks at 90° and 270°, a higher Hermans’ orientation parameter compared to PVA, and a transformation of the two-dimensional scattering pattern from circular to spindle-like. Comparison between 0% and 100% strain further demonstrated that applied stretching promotes the formation of fibrous crystals and facilitates the development of oriented channels [41]. Atomic force microscopy (AFM) was employed to assess surface structure and roughness; PVA exhibited low and irregular surface roughness, whereas M@APC presented periodic grooves forming continuous channels aligned with the orientation direction (Fig. 2f). Additionally, three-dimensional laser confocal imaging of fluorescently labeled PVA skeletons revealed parallel fiber bundles consistent with the processing direction, corroborating SEM observations (Fig. 2g). Collectively, these multidimensional analyses demonstrate that the reconstruction process ensures uniform incorporation of PEDOT:PSS and nano MgO, while establishing stable, oriented crystalline regions and continuous transport channels. This structural configuration provides a robust foundation for subsequent charge transfer under hydrated conditions and mechanical stability.

Fig. 2.

Preparation and microstructures of upper-layer anisotropic hydrogels. a) SEM images (Scale bars: 40 μm, 10 μm) and b) XRD patterns of various hydrogel groups. c) The corresponding scattering intensity (I) vs. azimuthal angle (θ) curve, d) Herman's orientation parameter, and e) SAXS patterns of hydrogels with random (PVA) and aligned (M@APC) nanofibrils after 0% and 100% strain, respectively. f) AFM phase images (Scale bars: 4 μm) and g) confocal images (Scale bars: 50 μm) of hydrogels with random and aligned nanofibrils.

2.2. Functional performance of the upper layer hydrogel

To systematically assess the functionality of the upper hydrogel, five distinct sample groups were characterized. Thermodynamic analyses were conducted to examine alterations in crystalline domains. As shown in Fig. 3a–c, the exothermic peaks of M@AP and M@APC in the differential scanning calorimetry (DSC) profiles shifted toward higher temperatures, indicating the formation of increased crystalline regions. Thermogravimetric (TG) and derivative thermogravimetric (DTG) analyses demonstrated that the decomposition temperatures of AP and APC exceeded those of PVA, suggesting a more stable network structure. Furthermore, M@AP and M@APC exhibited a rapid mass loss phase, which, together with XRD results, is likely attributable to the dehydration of magnesium hydroxide, inducing early melting of certain crystalline domains [42]. Consistent results were obtained from measurements of water content and crystallinity; oriented samples showed a significant reduction in water content accompanied by an increase in crystallinity (Fig. 3d and S6), indicative of simultaneous densification and crystallization. Subsequently, the anti-swelling behavior was evaluated, with all groups approaching near saturation within 36 h. The swelling ratio of M@APC was approximately 3%, whereas even the blank PVA exhibited a swelling ratio below 20%. Photographic evidence revealed no appreciable volume change, collectively demonstrating excellent swelling resistance (Fig. 3e and S7). Mechanical properties were then investigated, revealing that the M@APC hydrogel possessed superior toughness and strength at the macroscopic scale (Fig. 3f). Tensile stress-strain curves (Fig. 3g–i) indicated that the construction of the oriented structure enhanced both modulus and fracture strength, and the incorporation of nano-sized MgO significantly improved toughness. This enhancement is likely attributable to the inorganic phase acting as physical crosslinking points and heterogeneous nucleation sites, which restrict polymer chain mobility and facilitate stress transfer and energy dissipation [43]. Cyclic loading and unloading tests exhibited similar trends, with M@APC demonstrating the greatest dissipated energy (Fig. S8). In incremental strain cyclic loading-unloading tests, the hysteresis area increased moderately with strain, attributable to sacrificial hydrogen bonds providing effective energy dissipation. The modulus decreased slightly after each cycle, suggesting that a minor fraction of semicrystalline domains unfolded as strain increased (Fig. 3j) [44]. Further continuous cyclic loading-unloading tests confirmed the material's stability; as shown in Fig. 3k, the stress-strain curves of M@APC consistently overlapped over 100 cycles and remained close to the initial level after 1200 s of cycling (Fig. S9), indicating robust anti-fatigue behavior under repeated stress. Additionally, M@APC was capable of lifting an object approximately 500 times its own weight, demonstrating reliable macroscopic load-bearing capacity (Fig. 3l). Regarding electrical properties, the formation of oriented channels and the introduction of conductive components enhanced conductivity. Although the incorporation of nonconductive MgO caused a slight decrease in conductivity, it did not compromise overall electrical connectivity (Fig. 3m). As depicted in Fig. 3n, M@APC consistently illuminated an LED when used as a connector, confirming effective electronic transport. Finally, inductively coupled plasma (ICP) measurements indicated a relatively steady release rate of Mg²⁺ ions from M@APC, providing a sustained supply for subsequent applications (Fig. 3o). Collectively, these convergent findings demonstrate that the upper hydrogel integrates an anisotropic framework with functional constituents, thereby conferring robust mechanical properties, efficient electronic transport, and drug delivery capabilities.

Fig. 3.

Functional characterization of the upper layer hydrogel. a) DSC curves, b) TG analysis, and c) DTG curves of PVA, AP, APC, and M@AP, M@APC. d) Crystallinity under dry and wet conditions, and e) swelling ratios in phosphate-buffered saline (PBS) solution of the different hydrogel groups. f) Macroscopic photo of M@APC hydrogel. g) Tensile stress−strain curves of PVA, AP, APC, M@AP, and M@APC, h) accompanied by their corresponding tensile stress and strain, i) as well as Young's modulus and toughness values. j) Cyclic loading-unloading tests of M@APC hydrogel with an incremental strain. k) Multiple cyclic loading-unloading curves of M@APC hydrogel at 50% strain. l) High strength M@APC hydrogel (0.6 g) demonstration by loading a 3 kg bucket. m) Conductivity of the different hydrogel groups. n) Macroscopic photo of the M@APC hydrogel lighting up a light bulb. o) The drug release curves of the upper layer hydrogel at different times.

2.3. Preparation and functional evaluation of the lower layer hydrogel

To elucidate the interrelationships among chemical structure, microstructure, mechanical properties, and biological function of the lower hydrogel, a comprehensive series of characterizations was conducted, encompassing synthesis verification and functional evaluation. As illustrated in Fig. 4a, the preparation of the lower hydrogel involved four sequential steps. Initially, a polyester backbone was synthesized via epoxide ring-opening polymerization. Subsequently, ethylenediamine side chains were introduced through a Michael addition reaction, followed by grafting of phenylboronic acid groups via a quaternization reaction to yield the target polymer, PEEP. Finally, blending PEEP with a PVA solution produced a network crosslinked by dynamic boronate ester bonds, which are responsive to glucose and ROS. The lower hydrogel is designated as PEEP-PVA (PP) with composition x and y, where x and y represent the weight fractions of PEEP and PVA, respectively, in the total polymer content. Fourier-transform infrared (FTIR) spectroscopy revealed characteristic bands at 1715 and 1660 cm⁻¹, corresponding to the C=O stretching vibrations of ester and amide groups, respectively. Additionally, absorption features at 1612 and 1322 cm⁻¹ were attributed to benzene ring C=C stretching and C–B vibrations of boronate ester groups, collectively confirming the successful synthesis of PEEP (Fig. 4b). The successful preparation of PEEP was further corroborated by the presence of characteristic proton signals in the ¹H NMR spectrum (Fig. S10). Given that wet adhesion is a critical function of the lower hydrogel, the PP hydrogel effectively anchors the upper hydrogel layer onto the skin, facilitating intimate conformity and stable attachment during wrist movement, thereby demonstrating excellent macroscopic adhesiveness (Fig. 4c). Microstructural analysis (Fig. 4d) indicated that increasing the PEEP content resulted in smaller pore sizes and a denser network structure, attributable to a higher degree of crosslinking induced by the increased density of dynamic boronate ester bonds. This structural refinement provides a foundation for enhanced subsequent performance. Consistently, the ROS scavenging capacity increased with PEEP content, with the 15-5 PP formulation exhibiting the most pronounced antioxidant activity, suggesting that a higher density of dynamic boronate ester bonds promotes antioxidant behavior (Fig. 4e and S11). Adhesion mechanics testing further demonstrated that the reversible dynamic bonds, in conjunction with electrostatic interactions provided by quaternary ammonium cations, synergistically enhanced adhesion strength as the PEEP content increased (Fig. 4f and g) [45]. Rheological measurements revealed viscous-dominated behavior at low frequencies, with the sample maintaining a viscous flow state throughout the strain sweep, supporting favorable spreading and conformal adhesion on the wound surface (Fig. 4h and i). Based on these combined properties, the 15-5 PP formulation was selected for subsequent investigations. Stimuli responsiveness constitutes another critical functional attribute of the lower hydrogel, enabling intelligent regulation of degradation and drug release in response to microenvironmental cues. Degradation assays (Fig. 4j) demonstrated that 15-5 PP retained over 70% of its mass in PBS over five days. In contrast, under diabetic microenvironment conditions, the hydrogel exhibited accelerated degradation, culminating in approximately 90% cumulative drug release by day five (Fig. 4k and l). In summary, through the synergistic action of dynamic boronate ester bonds and cationic moieties, a lower hydrogel was developed that is both adhesive and stimuli-responsive. Its dynamic crosslinking network facilitates precise conformal contact with the wound and enables sensing of diabetic microenvironmental cues, thereby achieving on-demand controlled drug delivery.

Fig. 4.

Preparation and characterization of the lower layer hydrogel. a) Synthetic route of the PEEP and b) corresponding FTIR spectral analysis. c) Macroscopic adhesion photos of the lower layer hydrogel. d) SEM images of PP hydrogels with different formulations (5–5 PP, 10–5 PP, 15–5 PP). Scale bar: 200 μm. e) The quantification of DPPH scavenging efficiency of PP hydrogels with different formulations. f) Wet lap-shear adhesion strength and g) adhesion force of PP hydrogels with different formulations. h) Frequency-sweep and i) Strain-sweep rheology of PP hydrogels showing storage (G′) and loss (G″) moduli. j) Degradation of PP hydrogel in PBS, 25 mM glucose, 1 mM H₂O₂, and DM-mimicking degradation curve, with k) simulated drug daily release profile and l) cumulative release curve in different media. m) Agar-plate images and n) antibacterial rate showing the antibacterial performance of different samples against E. coli and S. aureus. Data were analyzed by one-way ANOVA followed by Tukey's post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS, not significant.

2.4. Antibacterial activity and biocompatibility of the bilayer hydrogel

Given the high prevalence of early infections in DM wounds, effective antibacterial activity is a critical attribute of wound dressings. However, it is essential to balance antibacterial efficacy with the minimization of cytotoxic effects. Accordingly, we initially focused on optimizing the concentration of nano MgO through cytotoxicity screening. The CCK-8 assay results indicated that a concentration of 20 mg/mL induced mild cytotoxicity, leading to the selection of 10 mg/mL as the optimal dosage (Fig. S12). Using this optimized concentration, antibacterial assays demonstrated that both the upper and lower hydrogel layers inhibited bacterial growth. The lower layer exhibited superior bactericidal activity, attributable to quaternary ammonium cations that disrupt bacterial membranes via electrostatic interactions. In contrast, the upper layer exerted its antibacterial effect through the sustained release of Mg²⁺ ions and the alkalinity generated by MgO hydrolysis, which compromised bacterial membrane potential and integrity. Consequently, the bilayer configuration achieved inhibition rates approaching 100%, indicating a synergistic antibacterial effect (Fig. 4m and n). Furthermore, a comprehensive evaluation of the biocompatibility of the skin-mimetic bilayer hydrogel was performed. Combined live/dead staining and CCK-8 assays revealed high cell viability and typical spreading morphology, confirming the absence of significant cytotoxicity induced by the materials (Fig. S13 and 14). Additionally, the hemolysis assay demonstrated a low hemolysis ratio, confirming excellent hemocompatibility (Fig. S15). Meanwhile, histological analysis following subcutaneous implantation in mice corroborated the excellent in vivo biocompatibility of the hydrogel. The main organ structures in the material-treated group were comparable to those in the blank control group, with no evident pathological alterations observed (Fig. S16). Collectively, these findings demonstrate the favorable biocompatibility of the skin-mimetic bilayer hydrogel and substantiate its biosafety for application as a wound dressing. For clarity and conciseness in subsequent biological experiments, abbreviated nomenclature for the skin-mimetic bilayer hydrogel was introduced, as summarized in Table 1.

Table 1.

Abbreviation details table for all groups involved.

	PVA	Anisotropy	Conductivity	MgO	PEEP	CGRP	ES	DM
AP	+	+	-	-	-	-	-	-
APC	+	+	+	-	-	-	-	-
M@AP	+	+	-	+	-	-	-	-
M@APC	+	+	+	+	-	-	-	-
PP	+	-	-	-	+	-	-	-
C@PP	+	-	-	-	+	+	-	-
DM-C@PP/APC	+	+	+	-	+	+	-	+
DM-C@PP/M@APC	+	+	+	+	+	+	-	+
DM-C@PP/APC-E	+	+	+	-	+	+	+	+
DM-C@PP/M@APC-E	+	+	+	+	+	+	+	+

2.5. The polarization regulation of RAW264.7 in vitro

To evaluate the immunoregulatory potential of C@PP, we first examined its effects on macrophage polarization. RAW 264.7 cells cultured under DM conditions were analyzed via immunofluorescence, flow cytometry, and RT-qPCR to assess the expression of key polarization markers across various treatment groups. Immunofluorescence analysis revealed that relative to DM, DM-C@PP significantly decreased the fluorescence intensity of the M1 marker CD86 while concurrently increasing CD206 expression (Fig. 5a and S17). Furthermore, flow cytometric analysis was conducted to assess the modulatory effects of C@PP on macrophage phenotypes (Fig. 5b–d). The findings revealed a significant shift in macrophage polarization within the DM-C@PP group, characterized by a marked reduction in M1 macrophages alongside a substantial increase in M2 macrophages. Consequently, the M2/M1 polarization ratio in the DM-C@PP group was 6.9-fold higher than that observed in the DM group. Consistent with these findings, RT-qPCR results are shown in Fig. 5e and f, both DM-C@PP and DM-PP treatments significantly suppressed the M1 proinflammatory markers iNOS and TLR4, with DM-C@PP demonstrating the most pronounced inhibition, indicative of a strong negative regulation of the M1 phenotype. Notably, DM-PP also reduced M1 marker expression, an effect likely attributable to the intrinsic ROS scavenging properties of PP; the addition of CGRP further enhanced this inhibitory effect. In contrast, regarding M2 markers, DM-PP induced only a modest increase in CD206 expression compared to DM alone, whereas DM-C@PP markedly upregulated both CD206 and CD200R1, with the latter exhibiting a particularly substantial increase. These results suggest that DM-C@PP more effectively promotes macrophage polarization toward a pro-repair M2 phenotype. Collectively, these results indicate that C@PP combines the ROS scavenging function of PP with the regulatory activity of CGRP to exert potent immunomodulatory effects.

Fig. 5.

Immunomodulatory and pro-migration properties of the skin-mimetic bilayer hydrogel in vitro. a) Immunofluorescence staining of M2/M1 RAW264.7 macrophage polarization markers CD206 and CD86 in different treatment groups. Scale bar: 50 μm. b-d) Typical flow cytometry expression profiles of CD86 and CD206 in M1-polarized RAW264.7 macrophages cultured in different groups, and quantitative analysis. e-f) Effects of different treatments on the relative mRNA expression levels of M1/M2 RAW264.7 macrophage polarization marker genes (iNOS, TLR-4, CD206, CD200R1). g) The transwell assay demonstrated the effect of different treatments on cell migration ability. Scale bar: 200 μm. Data were analyzed by one-way ANOVA followed by Tukey's post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS, not significant.

2.6. The migration regulation of human umbilical vein endothelial cells (HUVECs) in vitro

Angiogenesis is primarily driven by the directed migration and proliferation of endothelial cells and is intricately regulated by the immune microenvironment. Inflammatory mediators associated with M1 macrophages and elevated levels of ROS inhibit endothelial cell migration, whereas factors secreted by M2 macrophages facilitate extracellular matrix remodeling and promote migration [46]. To investigate this mechanism, HUVECs were treated with conditioned media derived from macrophages stimulated by C@PP, in combination with Mg²⁺ supplementation and exogenous electrical stimulation. Endothelial migration was subsequently evaluated using transwell assays. As depicted in Fig. 5g and S18, the DM group exhibited the lowest number of migrating cells, whereas the DM-C@PP/APC group showed a significant increase, indicating that the immunomodulated microenvironment effectively alleviates migration inhibition under diabetic conditions. The addition of Mg²⁺ alone (DM-C@PP/M@APC) or ES alone (DM-C@PP/APC-E) further enhanced migration. Notably, the combined treatment (DM-C@PP/M@APC-E) resulted in the highest migration density, approximately double that of the DM group, and was significantly superior to all other groups. This finding was corroborated by scratch assays, which demonstrated a consistent pattern, with the DM group exhibiting the least migration and the DM-C@PP/M@APC-E group displaying the greatest migratory capacity (Fig. S19). Furthermore, the effect of oriented surface structures on cell crawling behavior was examined. SEM images (Fig. S20) revealed that cells on non-oriented surfaces predominantly exhibited a rounded morphology, whereas cells on oriented surfaces were well-spread, spindle-shaped, and aligned with the surface texture. Fluorescence staining supported these observations, showing stripe-like cellular alignment on oriented surfaces (Fig. S21). Three-dimensional confocal imaging further elucidated the alignment of cell morphology and cytoskeletal organization along the imposed orientation (Fig. S22). Collectively, these results demonstrate a synergistic effect of immunomodulation, Mg²⁺, and ES on endothelial cell migration and suggest that contact guidance provided by biomimetic oriented structures facilitates directional migration.

2.7. Synergistic therapeutic approaches promote angiogenesis in vitro

Endothelial cell migration is essential for the successful formation of nascent vascular networks. To further elucidate whether the observed synergistic effect facilitates network formation, we systematically assessed the angiogenic capacity of HUVECs at the genetic, protein, and morphological levels. As depicted in Fig. 6a–c, compared to the DM group, the DM-C@PP/APC group exhibited significant upregulation of CD31, VEGF, and HIF-1α mRNA at days 3 and 7, indicating that the immunomodulatory environment established by C@PP/APC provides a critical proangiogenic foundation. Building upon this, both the DM-C@PP/M@APC and DM-C@PP/APC-E groups demonstrated further increases in the expression of these genes. Notably, the DM-C@PP/M@APC-E group consistently exhibited the most pronounced upregulation at all time points, significantly surpassing all other groups. VEGF protein expression was subsequently evaluated via immunofluorescence. The DM group showed weak red fluorescence, indicative of severely impaired VEGF synthesis under diabetic conditions, whereas the DM-C@PP/APC group displayed a marked increase, confirming that the immunomodulated milieu promotes VEGF production. The DM-C@PP/M@APC-E group exhibited the strongest and most extensive VEGF signal, approximately 1.8-fold higher than that of the DM group (Fig. 6d and f). Importantly, Matrigel tube formation assays provided functional validation of these findings. Relative to the DM group, the DM-C@PP/APC group enhanced tubular network formation, with further improvements observed in the DM-C@PP/M@APC and DM-C@PP/APC-E groups. The DM-C@PP/M@APC-E group generated the most complete and dense capillary-like networks (Fig. 6e). Quantitative analyses of node number, total tube length, and mesh number consistently demonstrated superior performance of this group compared to all others (Fig. 6g–i). To elucidate the underlying mechanism at the protein level, Western blot analysis (Fig. 6j and k) confirmed that the DM-C@PP/M@APC-E group exhibited the highest expression levels of angiogenesis-related proteins. Collectively, these multidimensional experiments demonstrate that the DM-C@PP/M@APC-E group achieves synergistic enhancement by integrating immune microenvironment remodeling, magnesium ion release, and exogenous electrical stimulation, thereby markedly improving the angiogenic capacity of HUVECs at the genetic, protein, and functional morphological levels.

Fig. 6.

Angiogenic regulation effects of skin-mimetic bilayer hydrogel in vitro. a-c) The mRNA expression levels of CD31, VEGF, and HIF-1α in HUVECs cultured under different treatments for 3 and 7 days were determined by RT-qPCR. d) Representative immunofluorescence images of VEGF expression (red) in HUVECs cultured under different treatments for 3 days. Cell nuclei were counterstained with DAPI (blue). Scale bar: 100 μm. e) Fluorescence micrographs and AngioTool topological network analysis of tubular structures formed by HUVEC cells treated with different treatments. Scale bar: 100 μm. f) Quantitative analysis of fluorescence intensity of VEGF by immunofluorescence staining. g-i) Quantitative analysis of tube formation-related parameters, including the number of meshes, the number of main branches, and the number of nodes. j) Western blot analysis images and k) relative protein expression levels of angiogenesis-related markers (CD31, HIF-1α, VEGF) in HUVECs from different groups after 7 days. Data were analyzed by one-way ANOVA followed by Tukey's post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS, not significant.

2.8. The biological mechanisms of angiogenesis by transcriptomic analysis

To further investigate the mechanisms by which Mg²⁺ and ES promote angiogenesis, we conducted cell transcriptome assays. We first assessed data quality and sample reproducibility. Pearson correlation heatmaps in Fig. 7a showed coefficients greater than 0.98 across all samples, indicating good within-group consistency. Results of differential expression analysis showed that, compared with the DM-C@PP group, the DM-C@PP/M@APC treatment group only induced 32 differentially expressed genes. In contrast, the combined treatment group DM-C@PP/M@APC-E detected a total of 1210 differentially expressed genes. Furthermore, we also compared the DM-C@PP/M@APC-E group with the DM-C@PP/M@APC group and detected 996 differentially expressed genes (Fig. 7b–c and S23). Gene Ontology (GO) functional enrichment analysis was performed to reveal the molecular mechanisms of electrical stimulation combined with magnesium ion therapy. Results indicated that comparing DM-C@PP/M@APC-E to both DM-C@PP and DM-C@PP/M@APC groups revealed a significant enrichment in biological processes associated with the extracellular matrix and cell junctions. Specifically covering key pathways such as integrin binding, cell junction assembly, collagen-containing extracellular matrix, and conferring tensile strength (Fig. 7d and e). This finding strongly confirms that electrical stimulation plays a core leading role in regulating cell adhesion, driving matrix remodeling, and enhancing the mechanical integrity of nascent tissues. Furthermore, the pathway of metal ion transmembrane transporter activity was also significantly activated in the comparison between DM-C@PP and DM-C@PP/M@APC (Fig. S24). We believe this reflects the specific cellular response to magnesium ion release, embodying its regulatory biological effects in enhancing cation transmembrane transport and maintaining ionic homeostasis. Kyoto Encyclopedia of Genes and Genomes (KEGG) results showed that, in the comparisons of DM-C@PP and DM-C@PP/M@APC-E as well as DM-C@PP/M@APC and DM-C@PP/M@APC-E, shared significantly enriched pathways such as TNF signaling pathway, IL-17 signaling pathway, ECM-receptor interaction, and Inflammatory mediator regulation of TRP channels were detected, which strongly confirms the core role of electrical stimulation in regulating inflammatory responses, matrix interactions, and inflammatory mediator regulation (Fig. 7f–g and S25). Most importantly, the comparison between DM-C@PP and DM-C@PP/M@APC-E also identified the focal adhesion and cell adhesion molecule pathways as being significantly enriched. The appearance of these key pro-angiogenic signaling pathways indicates that the combination of magnesium ions and electrical stimulation is not a simple functional overlay but significantly enhances the angiogenic potential at the molecular level through synergistic effects, thereby providing stronger biological support for tissue regeneration.

Fig. 7.

Multi-omics characterization of different treatment conditions. a) Pearson correlation heatmap illustrating the global transcriptomic similarity among all samples. High intra-group correlations indicate strong biological reproducibility, whereas the distinct inter-group separation highlights the divergent transcriptional responses induced by different treatments. b-c) Transcriptional variations among treatment groups identified by volcano plot analysis. Enrichment results for the three treatment comparisons are displayed. Red and blue dots highlight significantly upregulated and downregulated genes. d-e). Functional disparities among treatment groups identified by GO enrichment analysis. The enrichment of key biological pathways was compared across three different treatments. The number of genes, statistical significance, and enrichment factor are represented by the size of the dots, their colors, and their positions on the x-axis, respectively. Core functional terms with significant differences are highlighted with red boxes. f-g) KEGG pathway enrichment analysis of differential metabolites. Enrichment results for the three treatment comparisons are displayed. Bubble size corresponds to the count, and color corresponds to the -log₁₀ (P value). Pathways with critical differences are highlighted in red.

2.9. Skin-mimetic bilayer hydrogel promotes DM wound healing

To evaluate whether the skin-mimetic bilayer hydrogel, incorporating CGRP-mediated immunoregulation alongside Mg²⁺ release and electrical stimulation, can enhance angiogenesis in vitro and facilitate integrated healing within a complex pathological environment, we further assessed its pro-healing efficacy using a diabetic mellitus wound model (Fig. 8a). Chronic diabetes mellitus was induced in mice via streptozotocin administration, followed by the creation of 8 mm full-thickness skin defects on the dorsal surface. Macroscopic images obtained at successive time points (Fig. 8b and c) demonstrated progressive wound area reduction across all groups. Notably, the DM-C@PP/APC and DM-C@PP/M@APC groups exhibited significantly accelerated wound closure compared to the untreated DM group, indicating that CGRP delivery and Mg²⁺ release contributed positively to the repair process. Furthermore, the DM-C@PP/M@APC-E group, which combined ES with hydrogel treatment, achieved the most rapid healing, with a closure rate of 98.57% ± 0.85% by day 12, approaching complete wound closure and suggesting that ES further enhances healing in diabetic wounds. Histological analyses were performed on wound specimens collected at each time point using hematoxylin and eosin (H&E) and Masson's trichrome staining. As shown in Fig. 8d, H&E sections at day 12 revealed discontinuous reepithelialization and persistent inflammatory cell infiltration in the DM group, whereas the DM-C@PP/APC and DM-C@PP/M@APC groups exhibited increased reepithelialization and more mature granulation tissue. The DM-C@PP/M@APC-E group demonstrated the least inflammatory infiltration and a more complete epidermal layer. Masson's trichrome staining (Fig. 8e) revealed more uniform and organized collagen deposition in the treatment groups relative to the DM group, with the DM-C@PP/M@APC-E group displaying the most regular collagen fiber orientation and highest packing density. These findings suggest that the combined application of ES, Mg²⁺, and CGRP promotes collagen remodeling and dermal reconstruction.

Fig. 8.

DM wound healing effects of the skin-mimetic bilayer hydrogel in vivo. a) Schematic timeline of the in vivo wound healing experiment in a DM mouse model. b) Representative photographic images of wounds from different treatment groups on days 0, 3, 5, 7, 9, and 12. Scale bar: 4 mm. c) Quantitative analysis of wound closure rate over time. d-e) Histopathological analysis of wound tissues by H&E staining and Masson's trichrome staining. Representative H&E and Masson's stained sections of wound areas from different treatment groups on post-operative day 6 and day 12. The blue dashed lines demarcate the wound area. Scale bars: 50 μm, 500 μm.

2.10. Skin-mimetic bilayer hydrogel regulates the neuro-immune interactions

In DM wounds, prolonged early inflammation substantially hinders tissue regeneration, making a timely transition from the inflammatory to the proliferative phase critical for effective repair. Early inflammation in DM wounds may persist for 3 to 7 days post-surgery; therefore, day 6 was selected as a pivotal time point for analysis, aligning with the CGRP release profile of the C@PP hydrogel. Immunofluorescence staining was utilized to assess inflammatory modulation. As depicted in Fig. 9a and d-e, infiltration of F4/80-positive macrophages was evident across all groups. Relative to the DM group, the CD86-positive area was significantly reduced in the other three groups, indicating suppression of the pro-inflammatory macrophage phenotype. Conversely, the CD206-positive area was significantly increased in the DM-C@PP/APC and DM-C@PP/M@APC groups, suggesting that CGRP functions as a principal mediator of immunoregulation. The addition of Mg²⁺ ions and electrical stimulation conferred further synergistic effects, thereby abbreviating the inflammatory phase. Subsequent to immune modulation, successful progression into stable proliferation and remodeling largely depends on effective reconstruction of vascular and neural networks. To evaluate neural regeneration, CGRP expression in wound tissue was examined via immunofluorescence. As shown in Fig. 9b, qualitative observations at day 6 revealed that the DM-C@PP/M@APC-E group exhibited the densest CGRP distribution and clear colocalization with nuclei. Quantitative analysis in Fig. 9f further confirmed these findings, indicating that while all treatment groups exhibited elevated CGRP expression, the DM-C@PP/M@APC-E group demonstrated a significantly greater CGRP-positive area of 1.25% ± 0.05% than the other groups, providing compelling evidence of the efficacy of the synergistic strategy. By day 12, the CGRP signal was markedly diminished relative to day 6, and its localization shifted from the perinuclear region toward the cell membrane, indicating dynamic downregulation during repair and suggesting that early CGRP delivery and microenvironmental improvement predominantly occur during the initial phase [47]. Immunohistochemical analysis of GAP43 corroborated these findings. As shown in Fig. 9c, the DM-C@PP/M@APC-E group displayed more continuous and orderly nerve fiber-like structures at the wound edge and base. Quantitative analysis further demonstrated that at days 6 and 12, all treatment groups exhibited higher GAP43 expression than the DM group, with the greatest increase observed in the DM-C@PP/M@APC-E group (Fig. 9g). Collectively, these results indicate that the on-demand release of CGRP from the lower hydrogel, combined with the magnesium-containing oriented conductive network in the upper hydrogel and exogenous electrical stimulation, synergistically promotes axonal sprouting and elongation, thereby accelerating neural reconstruction within the wound.

Fig. 9.

Skin-mimetic bilayer hydrogel in DM wounds in vivo immune regulation and promoting neural repair effect. a) Immunofluorescence staining for the pan-macrophage marker F4/80 (green) and the M1 macrophage marker CD86 (red) or the M2 macrophage marker CD206 (red) in wound sections from DM mice treated with different treatment groups. Nuclei are counterstained with DAPI (blue). Scale bar: 50 μm. b) Representative images of immunofluorescence staining for CGRP (red) and nuclei (blue, DAPI) in wound sections from different treatment groups on day 6 and day 12. Scale bar: 50 μm. c) Representative immunohistochemical images of wound sections from different treatment groups at 6 days and 12 days post-injury show GAP43 protein expression (brown staining). Increased brown intensity indicates increased GAP43 expression. Scale bar: 50 μm. d-e) Quantitative analysis of the positive area of CD86 and CD206 by immunofluorescence staining. f) Quantitative analysis of the positive area of CGRP by immunofluorescence staining. g) Quantitative analysis of the GAP43-positive area by immunohistochemistry. Data were analyzed by one-way ANOVA followed by Tukey's post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS, not significant.

2.11. Skin-mimetic bilayer hydrogel promotes angiogenesis

Nerves and blood vessels engage in a reciprocal interaction throughout the wound healing process. A mature and continuous vascular network supplies oxygen and nutrients while facilitating the removal of metabolic waste, thereby establishing a critical foundation for neural regeneration and tissue repair. Building upon this concept, we performed dual immunofluorescence staining for CD31 and α-SMA to systematically evaluate angiogenesis and vascular maturation in diabetic mellitus (DM) wounds. Image analysis revealed that on day 6, the DM group exhibited sparse and discontinuous CD31 signals with minimal α-SMA coverage; in contrast, the DM-C@PP/APC group demonstrated more continuous CD31-positive structures with nascent lumens. Furthermore, vessel density was increased in both the DM-C@PP/M@APC and DM-C@PP/M@APC-E groups. Correspondingly, the α-SMA-positive area followed a similar trend; notably, the fluorescence area in the DM-C@PP/M@APC-E group measured 1.6% ± 0.05%, representing a 1.9-fold increase relative to the DM group and significantly exceeding that of the other two groups (Fig. 10a). By day 12, these differences became more pronounced: CD31 staining in the DM-C@PP/M@APC-E group revealed mature vessel-like structures, the α-SMA-positive area further increased, and colocalization with CD31 was markedly enhanced, collectively indicating progressive vascular maturation and stabilization (Fig. 10b). Quantitative analyses presented in Fig. 10c and d corroborated these findings, demonstrating stepwise increases in CD31 and α-SMA-positive areas at both day 6 and day 12, with the DM group exhibiting the lowest levels, the DM-C@PP/APC and DM-C@PP/M@APC groups showing significantly higher levels, and the DM-C@PP/M@APC-E group displaying the highest levels. In summary, CGRP-mediated immunoregulation combined with magnesium ion supplementation and exogenous electrical stimulation synergistically promotes angiogenesis and vascular maturation in DM wounds, thereby providing the metabolic support and regenerative microenvironment essential for neural regeneration and reparative cellular processes.

Fig. 10.

Skin-mimetic bilayer hydrogel in DM wounds in vivo angiogenic effect. a-b) Representative immunofluorescence images of wound sections stained for the CD31 (green) and the α-SMA (red) at day 6 and day 12 post-injury. Nuclei were counterstained with DAPI (blue). Scale bar: 50 μm. c-d) Quantitative analysis of α-SMA and CD31-positive areas in wound tissues. Data were analyzed by one-way ANOVA followed by Tukey's post hoc test. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS, not significant.

3. Conclusion

To the best of our knowledge, the skin-mimetic bilayer hydrogel developed in this study represents a novel strategy that integrates anisotropic conductive networks with a ROS- and glucose-responsive adhesive layer to address the sequential healing requirements of diabetic wounds. This dual-layer configuration enables time-sequenced therapy, advancing from early-stage immune modulation to late-stage angiogenic and neurogenic regeneration. Mechanistically, the simultaneous release of Mg²⁺ ions and electrical stimuli synergistically activates endothelial signaling pathways and extracellular matrix remodeling, while CGRP-mediated macrophage polarization contributes to immune niche remodeling. Transcriptomic analyses further demonstrate that electrical and ionic stimulation cooperatively regulate gene expression associated with angiogenesis, neurogenesis, and cytokine–receptor interactions. In vivo experiments reveal that the bilayer hydrogel significantly accelerates wound closure, reduces inflammation, and promotes vascular maturation and nerve regeneration in DM mouse models. In summary, these findings establish a neuro-immune-vascularization strategy for chronic wound healing, and the biomimetic, stage-adaptive design presented herein may guide the future development of intelligent materials for complex tissue regeneration.

CRediT authorship contribution statement

Haomin Wang: Writing – review & editing, Writing – original draft, Validation, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. Lei Yang: Writing – review & editing, Writing – original draft, Validation, Methodology, Formal analysis, Data curation, Conceptualization. Hao Jiang: Validation, Investigation, Formal analysis. Zheng Li: Supervision, Investigation. Xiaojun Zhou: Writing – review & editing, Methodology, Investigation, Data curation. Baokun Wang: Validation, Methodology. Zheyu Zhang: Supervision. Guoqing Wang: Writing – review & editing, Visualization, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization. Shuo Chen: Writing – review & editing, Supervision, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization. Chuanglong He: Writing – review & editing, Visualization, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization.

Ethics approval and consent to participate

This study does not involve human participants, human data, or human tissues. All the experimental animal procedures were performed by following local animal welfare laws and guidelines and approved by the Animal Ethics Committee of Donghua University (No: DHUEC-NSFC-2022-27).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix. ASupplementary data

The following is the Supplementary data to this article:

Data availability

The data supporting the findings of this study can be made available by the corresponding author upon reasonable request.