Skin-mimetic bilayer hydrogel normalizes diabetic wound healing by orchestrating inflammatory cell dynamics: An early intervention strategy

Abstract

Diabetic ulcers, characterized by persistent chronic inflammation and impaired healing, pose a significant clinical challenge. Inspired by the physiological healing process observed in healthy wounds, we develop an early intervention strategy by the spatiotemporal modulation of inflammatory cell dynamics to restore inflammation homeostasis in diabetic wounds. To achieve this, a skin-mimetic bilayer hydrogel that mimics the layered structure, composition, and pore size distribution of human skin is developed to allow for the sequential release of neuropeptide substance P and interleukin-10. Our findings demonstrate that this bilayer hydrogel achieves layer-specific pore size distributions and seamless interlayer integration through spontaneous dynamic crosslinking, along with clinically relevant multifunctionality. Notably, it enables rapid release of substance P from its loose bottom layer to enhance neutrophil recruitment within the first day and promote pro-inflammatory macrophage infiltration during the early inflammatory phase. Subsequently, the dense top layer enables delayed and sustained release of interleukin-10, which induces M2c macrophage polarization to facilitate inflammation resolution and support scarless wound closure. Overall, this study provides an effective early intervention strategy to reinstate the physiological healing process in diabetic wounds, thereby preventing chronic ulcer progression and non-healing outcomes.

Graphical abstract

Highlights

• •Skin-mimetic bilayer hydrogel is designed to mimic the layered structure and composition of human skin.
• •Spatiotemporal modulation of inflammatory cell dynamics restores inflammation homeostasis in diabetic wounds.
• •Spontaneous dynamic crosslinking endows facile synthesis and clinically relevant multifunctionality.
• •An effective early intervention strategy is established for preventing chronic ulcer progression.

1. Introduction

The high and rising prevalence of diabetes and diabetic foot ulcers (DFUs) poses major challenges to global healthcare system. According to the 11th Global Diabetes Atlas by the International Diabetes Federation (IDF), an estimated 589 million adults have diabetes in 2025, and 19-34% of patients will be affected by DFUs [1]. Clinical evidence shows that, 65% of healed ulcers recur within five years, and 20% of recurrences lead to amputation [2]. The high prevalence, treatment resistance, and severe complications associated with DFUs impose substantial clinical and socioeconomic burdens on patients and healthcare systems. Therefore, early intervention to halt the progression of diabetic wounds into chronic, refractory inflammation is regarded as an essential clinical strategy for improving healing efficiency.

Cutaneous wound healing is a pathophysiological process involving four temporally ordered phases: hemostasis, inflammation, proliferation, and remodeling [3,4]. After injury in healthy wounds, neutrophils, monocytes and macrophages are major immune cells that cooperate during the onset, progression and resolution of inflammation [5,6]. However, systemic pathological factors in diabetics often cause cutaneous abnormalities even before wound formation. For instance, resident mast cells in the skin of both diabetic humans and mice demonstrate excessive degranulation under baseline conditions, yet fail to mount an amplified response after injury, resulting in diminished secretion of the neuropeptide substance P (SP) [[7], [8], [9]]. This impairment leads to reduced infiltration of neutrophils and monocytes during the early healing phase, typically within 24 h post-injury [10]. Studies have indicated that attenuated acute inflammatory response in early diabetic wounds is one of the key pathological mechanisms causing the persistent infiltration of pro-inflammatory macrophages and subsequent delay in healing [[11], [12], [13]].

Macrophage dysfunction further exacerbates the pathophysiology of diabetic wounds through multiple interconnected mechanisms [14,15]. Among them, defective efferocytosis of apoptotic neutrophils contributes to a self-amplifying inflammatory milieu that aggravates tissue damage and promotes chronic wound formation. Macrophages exhibit remarkable plasticity, enabling their functional phenotypes to dynamically adapt to microenvironmental signals [16]. During the early inflammatory phase, M1-like pro-inflammatory macrophages are predominantly activated, whereas M2-like anti-inflammatory macrophages emerge later to resolve inflammation and facilitate tissue repair [17,18]. In particular, the M2c macrophage subtype, induced by interleukin-10 (IL-10) or transforming growth factor-β (TGF-β), plays a critical role in the later stages of inflammation. This subtype contributes to inflammation resolution through potent efferocytic activity and sustained secretion of endogenous IL-10 [[19], [20], [21]]. These findings highlighted two distinct yet interrelated pathological features in diabetic ulcers: insufficient acute inflammation due to impaired neutrophil recruitment in the early inflammatory phase, and unresolved chronic inflammation stemming from defective efferocytosis and impaired macrophage phenotypic regulation in the later inflammatory phase. Therefore, sequentially and precisely modulating inflammatory cell behaviors to restore spatiotemporal coordination in the healing process represents significant therapeutic potential for achieving high-quality diabetic wound repair.

In this study, inspired by the physiological healing cascade in healthy wounds, we designed a skin-mimetic bilayer hydrogel for sequential delivery of SP and IL-10 to spatiotemporally modulate inflammatory cell dynamics, particularly neutrophils and macrophages (Fig. 1). The bilayer hydrogel recapitulates human's skin structure and composition through: (i) a dense top layer formed via dynamic thiol-metal ion coordination using keratin (Ker), a predominant structural protein of epidermis, and copper ions (Cu²⁺) as the building blocks, providing a skin barrier against external bacterial infection [[22], [23], [24], [25]]; and (ii) a loose bottom layer constructed via the formation of dynamic Schiff base bonds, using gelatin (Gel) and oxidized hyaluronic acid (OHA) as the material matrix, emulating the ECM microenvironment for tissue regeneration. It is noteworthy that, spontaneous thiol-aldehyde dynamic crosslinking between the two layers ensures seamless interfacial integration (Fig. 1a). Owing to the dynamic cross-linking gelation mechanism and the formulation design, this bilayer hydrogel exhibits clinically relevant multifunctionality including desirable mechanical resilience, shape adaptability, and skin adhesion. More importantly, this bilayer hydrogel maintains excellent biocompatibility, owing to its natural composition and a simple, mild synthesis process that requires no external cross-linkers, additives, or material modifications. Functionally, it also serves as a sequential drug delivery system, in which the bottom layer enables rapid release of SP within the first day to enhance early neutrophil recruitment and acute inflammatory responses, while the top layer provides delayed and sustained release of IL-10 during the later phase to promote M2c macrophage polarization, driving efferocytosis of apoptotic neutrophils and inflammation resolution (Fig. 1b and c). Using a diabetic murine model of full-thickness skin defects, we demonstrate that this bilayer hydrogel normalizes diabetic wounds to a healthy healing trajectory by reprogramming inflammatory cell dynamics across the onset, progression and resolution phases of inflammation. This coordinated modulation reestablishes tissue homeostasis and facilitates scarless regeneration in diabetic wounds.

Fig. 1.

Design and application of a skin-mimetic bilayer hydrogel for orchestrating inflammatory cell dynamics to normalize healing in diabetic wounds. a Fabrication mechanism of skin-mimetic bilayer hydrogel via dynamic crosslinking. b Skin-mimetic bilayer hydrogel-mediated sequential release of SP and IL-10 for normalization of inflammatory cell dynamics in diabetic wounds. c Skin-mimetic bilayer hydrogel-mediated spatiotemporal modulation of inflammatory cell dynamics facilitates scarless closure in diabetic wounds.

2. Results

2.1. Fabrication and evaluation of skin-mimetic bilayer hydrogel

To mimic both the composition and microstructure of human skin, we fabricated a bilayer hydrogel using natural polymers that are abundant in human skin as the matrix materials. The top layer was designed to replicate the epidermis by employing keratin as the matrix material. Keratin is a cysteine-rich fibrous structural protein produced by keratinocytes, primarily found in the epidermis of the skin [[26], [27], [28], [29]]. A dense, dynamic hydrogel (denoted as Ker/Cu hydrogel) was prepared through a thiol-Cu²⁺ coordination reaction between thiol groups in keratin and Cu²⁺ ions (Fig. 2a). This gelation process was achieved by simple mixing at room temperature, eliminating the need for chemical modification of keratin or additional crosslinkers. The hydrogel's microstructure could be tuned by varying the mass ratios of keratin to copper sulfate (Ker/Cu = 1/0.04, 1/0.06, and 1/0.08), as shown by scanning electron microscopy (SEM). The resulting hydrogels exhibited interconnected, uniform porous structures with pore sizes ranging from 35 to 65 μm (Fig. 2b). Higher Cu²⁺ concentrations led to larger pores, attributed to the dominant crosslinking mechanism shifting from disulfide (S-S) bonds between keratin's thiol groups to S-Cu-S bonds formed with Cu²⁺ (Fig. 2c). A higher swelling ratio was achieved at a higher level of Cu²⁺ concentration, further confirming the presence of larger pores within the hydrogel (Fig. S2a). Due to the dynamic thiol-Cu²⁺ coordination, two separate Ker/Cu 0.06 hydrogels fused completely within 30 min of contact, demonstrating excellent self-healing properties (Fig. 2d). The rheological properties of the Ker/Cu 0.06 hydrogel were first characterized by amplitude sweep. As shown in Fig. 2e, both the storage (G′) and loss (G″) moduli remained constant at strains below 100%, with G′ consistently surpassing G″, confirming a stable gel network. Structural yielding commenced as the strain exceeded 100%, marked by a sharp decline in G’. The G′ and G″ curves intersected at a strain of approximately 700%. To verify the dynamic nature of this network, an alternating step strain test was conducted. Under high shear strain (γ = 1000%), the hydrogel liquefied (G’’ > G′), but it rapidly recovered its gel state (G’ > G″) upon returning to low strain (γ = 1%). This process could be repeated several cycles (Fig. 2f). The incorporation of Cu²⁺ also endowed the hydrogel with potent antibacterial activity. Inhibition zone and colony counting assays against S. aureus revealed clear antibacterial efficacy across all tested Cu²⁺ concentrations, and higher Cu²⁺ levels correlated with larger inhibition zones and significantly reduced bacterial viability (Fig. 2g–i). Owing to its dense microstructure and potent antibacterial activity, the top layer hydrogel provided robust resistance to external bacterial infections, a crucial characteristic for wound protection.

Fig. 2.

Fabrication and evaluation of top layer hydrogel and bottom layer hydrogel. a Photograph of Ker/Cu hydrogel and its dynamic cross-linking gelation mechanism via thiol-Cu²⁺ coordination reaction between thiol groups in keratin and Cu²⁺. b Representative SEM images of Ker/Cu hydrogels at three concentrations of Cu²⁺, and their average pore sizes. c Elemental mapping images of Ker/Cu hydrogel. d Macroscopic self-healing test of Ker/Cu 0.06 hydrogel. e Rheological test of Ker/Cu 0.06 hydrogel by amplitude sweep. f Rheological behaviors of Ker/Cu 0.06 hydrogel with alternate strain cycles from 1% to 1000%. g Antibacterial ability of Ker/Cu hydrogels at three Cu ratios against S. aureus tested by inhibition zone. h Photographs of bacterial colonies on the agar plates after contact with Ker/Cu hydrogels at three Cu²⁺ ratios for 12 h. PBS was set for control. i Bacterial viability of Ker/Cu hydrogels. j Photograph of Gel/OHA hydrogel and its dynamic cross-linking gelation mechanism via Schiff base reaction between amino groups of gelatin and aldehyde groups of oxidized hyaluronic acid. k Representative SEM images of Gel/OHA hydrogels at three ratios of Gel/OHA, and their average pore sizes. l Macroscopic self-healing test of Gel 3/OHA hydrogel. m Rheological test of Gel 3/OHA hydrogel by amplitude sweep. n Rheological behaviors of Gel 3/OHA hydrogel with alternate strain cycles from 1% to 700%. o Photographs of Gel 3/OHA hydrogel adhered to a finger joint and adapted well to finger movements. All data were generated from at least three independent experiments and presented as the means ± standard deviation. Statistical analysis was performed by one-way ANOVA. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001.

For the bottom layer, we selected gelatin and oxidized hyaluronic acid (OHA), natural polymers abundant in the dermal layer of human skin, to construct a loose, dynamic hydrogel via Schiff base formation between amino groups in gelatin and aldehyde groups in OHA (denoted as Gel/OHA hydrogel) (Fig. 2j). OHA was synthesized by sodium periodate oxidation of HA, with an aldehyde substitution degree of 28.6% confirmed by hydroxylamine hydrochloride titration (Fig. S1a). The Gel/OHA hydrogel could also be fabricated by a simple mixing step at room temperature. FTIR spectroscopy confirmed the formation of imine bonds in the hydrogel, as evidenced by the disappearance of the characteristic aldehyde peak at 1730 cm⁻¹ and the concomitant appearance of a new peak at 1630 cm⁻¹ corresponding to C=N stretching vibrations (Fig. S1b). The Gel/OHA hydrogel displayed a porous architecture with pore sizes ranging from 120 to 220 μm across different mass ratios (2/2, 3/2, and 4/2), which were significantly larger than those in the top layer (Fig. 2k). A positive correlation was observed between the swelling ratio and the pore size of the Gel/OHA hydrogels (Fig. S2b).The macroscopic self-healing test and rheological tests including amplitude sweep and alternating step strain sweep confirmed the dynamic crosslinking mechanism of the Gel/OHA hydrogel (Fig. 2l–n). Additionally, this hydrogel could adhere to skin, withstanding finger joint movements without detachment (Fig. 2o), demonstrating its potential for wound dressing applications.

A bilayer hydrogel was constructed by integrating a top-layer Ker/Cu 0.06 hydrogel with a bottom-layer Gel 3/OHA hydrogel through spontaneous interfacial crosslinking. The process was achieved by bringing the two layers into contact at room temperature for about 1 h, which induced a dynamic thiol-aldehyde addition reaction between thiol groups present in the keratin (Ker, top layer) and aldehyde groups in the oxidized hyaluronic acid (OHA, bottom layer). Notably, this assembly process required no external adhesives, chemical crosslinkers, or material pretreatment, yet yielded a seamlessly integrated structure with no interfacial gaps (Fig. 3a). Cross-sectional SEM imaging confirmed tight interfacial bonding and layer-specific pore size distributions. The top layer exhibited a dense porous microstructure with an average pore size of approximately 48 μm, while the bottom layer displayed larger pores averaging around 167 μm (Fig. 3b). This structure closely mimics the epidermal-dermal architecture of human skin [30,31]. The interfacial bonding strength was confirmed to be sufficient to resist mechanical delamination (Fig. 3c) and remained stable even after swelling of the bilayer hydrogel in water (Fig. 3d). To further validate the dynamic crosslinking mechanism between two layers, two Ker/Cu 0.06 and Gel 3/OHA hydrogels were cut and the separate pieces formed an integrated hydrogel after contact capable of withstanding a tensile force without breakage (Fig. 3e). Mechanical compression tests demonstrated that the bilayer hydrogel kept structural integrity at strains >80% (Fig. 3f). The bilayer hydrogel conformally adhered to dynamic surfaces, such as finger joints, without detachment or interfacial failure, owing to the suitable tissue adhesion of the bottom layer, robust interfacial bonding between layers, and excellent mechanical compliance (Fig. 3g). Moreover, it enabled painless, residue-free removal, highlighting its potential for clinically convenient wound dressing applications. The in vitro biocompatibility of the hydrogels was evaluated through hemolysis and cytotoxicity assays. The hemolysis test demonstrated excellent blood compatibility for all hydrogel formulations including Gel 3/OHA, Ker/Cu 0.06, and bilayer hydrogel, with hemolysis rates below the 5% threshold for biomaterial safety (Fig. 3h). Cytocompatibility was assessed using L929 mouse fibroblasts through CCK-8 proliferation assays and Live/Dead staining. Quantitative analysis showed no statistically significant differences in the cell viability between hydrogel-treated groups and PBS control using a hydrogel leaching solution collected on day 1 (Fig. 3i). This result was corroborated by fluorescence microscopy (Fig. S3a), which showed an increase in live (green) cells over prolonged incubation, with few dead (red) cells observed. To further assess the long-term cytotoxicity of the bilayer hydrogel, leaching solutions collected on days 6 and 12 were co-cultured with L929 cells. CCK-8 assays confirmed that the bilayer hydrogel exhibited no apparent cytotoxic effects (Fig. S3b). These results demonstrated the excellent cytocompatibility of all hydrogels, showing promise for biomedical applications. The in vivo biocompatibility and biodegradation of the bilayer hydrogel were evaluated via subcutaneous implantation in mice. The degradation process was monitored at 6, 9, and 14 days post-implantation. A gradual reduction in the hydrogel volume was observed over the 14-day period. Histopathological analysis of the local skin tissue on day 6 showed no significant inflammatory response induced by the hydrogel (Fig. S4a). Furthermore, H&E staining of main organs (liver, spleen, and kidney) on day 6 confirmed the absence of notable pathological alterations, collectively demonstrating a good in vivo biocompatibility of the bilayer hydrogel (Fig. S4b).

Fig. 3.

Construction and evaluation of skin-mimetic bilayer hydrogel. a Photograph of the bilayer hydrogel and schematic illustration of interfacial bonding mechanism. b Representative SEM images of bilayer hydrogel. c Interfacial bonding strength test of bilayer hydrogel. d Photographs of bilayer hydrogel after swelling in water. e Macroscopic self-healing test of Gel 3/OHA and Ker/Cu 0.06 hydrogels. f Mechanical compressive curves of bilayer hydrogel. g Skin adhesion and peeling off test. h Hemolysis assay of the hydrogels. Inset: photographs of red blood cells after incubation with the hydrogels. Deionized water and PBS were used as positive and negative controls, respectively. i Viability of L929 mouse fibroblasts after incubation in the leaching solution of hydrogels for 1, 3 and 5 days. j Schematic of drug release from bilayer hydrogel. kIn vitro release profile of RhB. lIn vitro release profile of BSA-FITC. All data were generated from at least three independent experiments and presented as the means ± standard deviation. Statistical analysis was performed by one-way ANOVA. ns, not significant.

The sequential drug release kinetics of the bilayer hydrogel were investigated using Rhodamine B (RhB) and Fluorescein-labeled Bovine Serum Albumin (BSA-FITC) as model drugs to simulate SP and IL-10, respectively (Fig. 3j). The cumulative release profile of RhB demonstrated a rapid release of approximately 75% from the bottom layer within the first 24 h (Fig. 3k). This rapid release can be attributed to the loose, porous microstructure of the bottom hydrogel layer. In contrast, a delayed and sustained release profile of BSA-FITC was obtained after the first day, owing to the dense structural architecture of the top hydrogel layer and the overall bilayer design (Fig. 3l). A cumulative release of about 65.79% was achieved within eleven days. These results demonstrated that the bilayer hydrogel facilitated a sequential drug release profile, characterized by an initial rapid release followed by a delayed and sustained release after the first day.

2.2. In vitro assay for neutrophil migration, macrophage polarization and macrophage efferocytosis

The chemotactic effect of SP on neutrophil migration was evaluated using a transwell migration assay (Fig. 4a). Prior to experimentation, HL-60 cells were differentiated into neutrophil-like cells (dHL-60) via dimethyl sulfoxide (DMSO) induction [32]. Wright-Giemsa staining revealed distinct morphological changes in the cells. Undifferentiated HL-60 cells exhibited rounded nuclei, while dHL-60 cells developed multi-segmented nuclei, a characteristic feature of neutrophil maturation, after 7 days of DMSO induction (Fig. 4b). To investigate SP-mediated chemotaxis, dHL-60 cells were seeded in the upper chamber of a transwell system, while the leaching solution of SP-loaded Gel 3/OHA (SP@Gel/OHA) was added in the lower chamber. The leaching solution of Gel 3/OHA without SP loading was set for comparison. Following a 12-h incubation, SP@Gel/OHA significantly enhanced dHL-60 cell transmigration, while Gel 3/OHA without SP elicited negligible migration (Fig. 4c and d). This result indicated a potent chemotactic activity of SP@Gel/OHA.

Fig. 4.

In vitro assay of inflammation cell modulation under stimulation of SP-loaded Gel/HA and IL-10-loaded Ker/Cu. a Schematic of neutrophil migration test using a transwell system after treatment with the leaching solution of SP@Gel/HA. Gel/HA and blank culture medium were set for comparison. b Wright-Giemsa staining of HL-60 cells before and after differentiation. c Photograph of dHL-60 cells migrating to the lower chamber. d Quantitative analysis of neutrophil migration after treatments with SP@Gel/HA and Gel/HA. Untreated group serves as a control. e Schematic of macrophage polarization after treatment with LPS, IL-10, or IL-10/LPS. f Representative fluorescence images of macrophages after different treatment. Red: iNOS (M1 marker); Green: CD163 (M2c marker); Blue: DAPI (nuclear staining). g Schematic of macrophage efferocytosis test toward apoptotic dHL-60 cells under different treatments. h Flow cytometry plots of dHL-60 cells before and after apoptosis induction. i Representative fluorescent images of macrophage efferocytosis toward apoptotic dHL-60 cells under different treatments. Macrophages and apoptotic cells were stained green and red, respectively. All data were generated from at least three independent experiments and presented as the means ± standard deviation. Statistical analysis was performed by one-way ANOVA. ns, not significant; ∗∗∗∗p < 0.0001.

The phenotypic switch in macrophages plays a pivotal role in inflammation resolution and high-quality wound repair. Exogenous IL-10 is known to promote macrophage polarization toward the M2c subtype (also referred to as “acquired deactivation macrophages”) through activation of the IL-10 receptor and subsequent downstream signaling via signal transducer and activator of transcription 3 (STAT3) [33]. To evaluate the polarization effects of IL-10-loaded Ker/Cu hydrogel, M0 macrophages were co-cultured with LPS, IL-10, or a combination of IL-10 and LPS (IL-10/LPS) (Fig. 4e). M0 macrophages were differentiated from THP-1 cells, a human monocytic cell line, using phorbol 12-myristate 13-acetate (PMA). IL-10 was obtained by incubating the hydrogel in the culture medium to obtain the leaching solution. Fluorescence microscopy analysis demonstrated that LPS stimulation induced high expression of inducible nitric oxide synthase (iNOS) in macrophages, consistent with a pro-inflammatory phenotype. In contrast, macrophages exposed to either IL-10 or IL-10/LPS exhibited markedly elevated CD163 expression, a defining marker of the M2c macrophage subtype, with a higher upregulation observed in the IL-10 treatment group (Fig. 4f).

M2c macrophages are characterized by their efferocytic capacity, which allows them to efficiently clear apoptotic cells such as neutrophils following acute inflammation. This process further contributes to the resolution of inflammation [10,34]. We next studied the efferocytosis ability of M0 macrophages toward apoptotic dHL-60 after treatment with LPS, IL-10, or IL-10/LPS (Fig. 4g). Apoptosis in dHL-60 cells was induced via UV irradiation [35,36]. Flow cytometry analysis confirmed successful induction of apoptosis, as indicated by a significant increase in Annexin V+/PI + staining in irradiated dHL-60 cells compared to non-irradiated controls (Fig. 4h). Subsequently, apoptotic dHL-60 cells were co-cultured with M0 macrophages that had been pretreated with LPS, IL-10, or IL-10/LPS. To enable clear visualization, macrophages were labeled with CellTracker Green, while apoptotic dHL-60 cells were stained with pHrodo Red to distinguish engulfed cells. Fluorescence microscopy analysis showed that macrophages treated with IL-10 and IL-10/LPS, particularly the former, exhibited significantly higher efferocytic activity compared to the LPS-treated group (Fig. 4i). These results demonstrated that SP effectively recruited neutrophils, and IL-10 promoted macrophage polarization toward the M2c subtype, which displayed enhanced capacity for apoptotic cell clearance.

2.3. Bilayer hydrogel normalizes diabetic wound healing

We further evaluated the sequential inflammation modulating and wound healing efficacy of the SP/IL-10-loaded bilayer hydrogel using a full-thickness cutaneous wound model in diabetic mice (Fig. 5a). Prior to the main experiment, the SP loading content was optimized to achieve neutrophil infiltration levels in diabetic wounds comparable to those in normal wounds, thereby avoiding excessive recruitment that could promote neutrophil extracellular trap (NET) formation and exacerbate inflammation. At an SP loading of 5 μg/mL, neutrophil infiltration in SP-treated diabetic wounds increased to a level similar to normal wounds by 24 h and subsequently declined comparably in both groups by 72 h, with no obvious tissue irritation observed (Fig. S5). Subsequently, the diabetic wounds were divided into four treatment groups: bilayer hydrogel loaded with SP and IL-10 (SP/IL-10@Bilayer), bilayer hydrogel loaded with SP in the bottom layer (SP@Bilayer), bilayer hydrogel loaded with IL-10 in the top layer (IL-10@Bilayer), and saline solution (Model). A full-thickness cutaneous wound model in healthy mice was also used as a control, which was treated with saline solution (Normal). Wound closure was monitored on days 0, 3, 6, 9, and 12 (Fig. 5b and c). Among all the groups, the model group showed a lowest healing rate at each time point. SP@Bilayer had a comparable healing rate to normal by day 3 but lagged thereafter, suggesting a fast release of SP in wounds, which enhanced acute inflammation responses compared to model group. IL-10@Bilayer surpassed normal healing rates by day 3, however, its healing progression slowed subsequently. Of note, the SP@Bilayer group showed faster wound closure than the IL-10@Bilayer one between days 9 and 12, although the difference did not reach statistical significance. These findings indicate that upregulating acute inflammation through the rapid release of SP did not impede healing progression of diabetic wounds. Conversely, anti-inflammatory intervention alone delayed wound healing at later stages when administered in the setting of an inadequate acute inflammatory response. In contrast, SP/IL-10@Bilayer exhibited a similar healing rate to normal at each time point, with both achieving complete wound closure by day 12. This indicates that bilayer hydrogel-mediated sequential release of SP and IL-10 not only enhanced acute inflammation but promoted its resolution, ultimately enabling diabetic wounds to follow a physiological healing trajectory comparable to that of healthy wounds.

Fig. 5.

In vivo evaluation of bilayer hydrogel for diabetic wound healing. a Experimental timeline of the wound healing test in a diabetic model. b Representative photographs and schematic diagrams of wounds on days 0, 3, 6, 9 and 12 post-injury. Diabetic wounds were treated with SP/IL-10@Bilayer, SP@Bilayer, IL-10@Bilayer, and saline solution (Model), respectively. Healthy mice treated with saline solution were set as Normal. c Quantitative analysis of relative wound areas at each time point. d Representative H&E staining images of wounds on days 6 and 12. Black dash lines and black triangles indicated the length of wound area and epidermal thickness, respectively. e Quantitative analysis of epidermal thickness in wounds on day 12. f Representative Masson's trichrome staining images of wounds on days 6 and 12. g Quantitative analysis of collagen deposition on day 12. h Representative Sirius Red staining images of wounds on day 12. i Ratios of collagen I to III of wounds. All data were generated from at least three independent experiments and presented as the means ± standard deviation. Statistical analysis was performed by one-way ANOVA. # means significant difference compared to the normal group. #p < 0.05, ##p < 0.01, ###p < 0.001 and ####p < 0.0001; ∗ means significant difference compared to the model group. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001; & means significant difference compared to SP/IL-10@Bilayer. & p < 0.05, && p < 0.01 and &&& p < 0.001.

H&E staining of wound tissues revealed that the SP/IL-10@Bilayer group exhibited comparable wound dimensions to the normal group on days 6 and 12, whereas other diabetic groups had significantly larger wounds on day 12 (Fig. 5d and S6). Moreover, epidermal thickness in the SP/IL-10@Bilayer group approximated normal levels, significantly lower than those in model and SP@Bilayer groups (Fig. 5e). Masson's trichrome staining of wound tissues demonstrated that the SP/IL-10@Bilayer group exhibited well-aligned collagen fibers, organized in wavy bundles within the dermal layer (Fig. 5f). Quantitative analysis further revealed that the collagen deposition in SP/IL-10@Bilayer closely resembled that of healthy tissue, while being significantly higher than those in model and SP@Bilayer (Fig. 5g). To further characterize collagen organization, Sirius Red staining was performed and analyzed under polarized light microscopy, where collagen I (mature, thick fibers) and collagen III (immature, thin fibers) appeared yellow-orange and green, respectively (Fig. 5h). Compared to the normal group, both Model and SP@Bilayer showed significantly elevated collagen I/III ratios. In contrast, SP/IL-10@Bilayer maintained a ratio similar to that of the normal group, suggesting its potential to mitigate fibrotic scar formation (Fig. 5i). The scar score of the wounds in the SP/IL-10@Bilayer group was also lower than that in the other groups (Fig. S7). It should be noted that although the IL-10@Bilayer group exhibited comparable collagen deposition and a collagen I/III ratio to those of the normal group, it showed disordered collagen fiber arrangement in the dermal layer, as revealed by Masson's trichrome staining.

2.4. Bilayer hydrogel orchestrates inflammatory cell dynamics

Studies have shown that the dysregulation of inflammatory cell behaviors is a major pathological factor leading to the persistent chronic inflammation and consequent delayed wound healing [37,38]. We further investigated the sequential inflammatory modulation of SP/IL-10@Bilayer on early inflammatory cell recruitment and later macrophage polarization, with SP@Bilayer and IL-10@Bilayer for comparison. Saline-treated healthy wounds (Normal) and diabetic wounds (Model) were also utilized as controls. The early neutrophil infiltration in each group was assessed by immunohistochemical staining of Ly-6G⁺ cells at 8 h, 1 d and 3 d after injury (Fig. 6a). Fig. 6b and c demonstrated that Ly-6G expression in the normal group significantly increased by day 1 while decreased by day 3 to a level comparable to that at 8 h. In contrast, the model group exhibited a slight increase in Ly-6G expression throughout the test period, which remained significantly lower than that in the normal group on day 1. This result verified an inadequate neutrophil recruitment in diabetic wounds during the early inflammatory phase [39]. In wounds treated with SP-loaded bilayer hydrogels including SP@Bilayer and SP/IL-10@Bilayer, Ly-6G expression was upregulated by 8 h and reached levels comparable to healthy wounds by day 1, confirming a fast delivery of SP within the first day and an effective neutrophil recruitment by day 1 in diabetic wounds. By day 3, neutrophil infiltration in SP/IL-10@Bilayer decreased to a level comparable to that of the normal group. In contrast, neutrophil numbers in SP@Bilayer, while reduced, remained higher than the normal level. The expression level of Ly-6G in the IL-10@Bilayer group remained similar to that of the model group during the test period. These findings suggested that the sequential delivery of SP and IL-10 in the SP/IL-10@Bilayer group not only enhanced early neutrophil recruitment but also suppressed prolonged neutrophil infiltration. Moreover, CXCL-1 levels were measured on day 1 post-injury as a key indicator of the initial neutrophil influx. CXCL-1 functions as a major neutrophil-specific chemoattractant, driving their rapid migration to the wound site within hours post-injury [40,41]. qRT-PCR analysis revealed that CXCL-1 expression in wounds treated with SP-loaded bilayer hydrogels was comparable to that in the normal group, and significantly elevated relative to those in the model and IL-10@Bilayer groups, further confirming effective neutrophil infiltration during the early inflammatory phase in diabetic wounds (Fig. 6d).

Fig. 6.

Bilayer hydrogel orchestrates inflammatory cell dynamics during the early inflammation phase of diabetic wound healing. a Experimental timeline for assay of early neutrophil recruitment. b Immunohistochemical staining for Ly-6G in wounds at 8 h, 1 d and 3 d after injury. Diabetic wounds were treated with SP/IL-10@Bilayer, SP@Bilayer, IL-10@Bilayer, and saline solution (Model), respectively. Healthy mice treated with saline solution were set as Normal. c Quantitative analysis of Ly-6G⁺ cells in each group. d Relative expression of CXCL-1 on day 1. e Relative expression of MCP-1 on day 1. f Experimental timeline for assay of M1 macrophage infiltration. g Immunofluorescence staining for iNOS in wounds on days 1, 3 and 6 after injury. h Quantitative analysis of iNOS ⁺ cells in each group. i-k Relative expressions of macrophage-associated pro-inflammatory cytokines including TNF-α, IL-1β and IL-6 on day 3. l Schematic illustrating the dynamic modulation of inflammatory cells during the early inflammation phase of diabetic wounds by SP/IL-10@Bilayer. All data were generated from at least three independent experiments and presented as the means ± standard deviation. Statistical analysis was performed by one-way ANOVA. # means significant difference compared to the normal group. #p < 0.05, ##p < 0.01 and ###p < 0.001; ∗ means significant difference compared to the model group. ∗p < 0.05; & means significant difference compared to SP/IL-10@Bilayer. & p < 0.05 and && p < 0.01.

Studies show that during the acute inflammation phase in healthy wounds, neutrophils release proteases like MMP-9 to activate MCP-1 (also called CCL2), a key monocyte chemoattractant, by liberating its latent form from extracellular matrix (ECM)-bound reservoirs, thereby generating a chemotactic gradient that recruits circulating monocytes to the wound site [42]. The recruited monocytes subsequently differentiate into macrophages, orchestrating the progression from neutrophil-dominated acute inflammation to macrophages-dominated tissue repair. Therefore, we further determined the MCP-1 level in each group on day 1 post-injury by qRT-PCR to assess the priming of macrophage infiltration (Fig. 6e). Compared to healthy wounds, the model group showed a significant reduction in MCP-1 expression on day 1. Treatment with the SP-loaded bilayer hydrogels restored MCP-1 to a level comparable to that in healthy wounds, whereas the IL-10@Bilayer group exhibited no significant difference from the model group. The dynamic modulation of macrophage polarization was also assessed throughout the inflammatory progression and resolution phases in each group. The distribution of M1-like pro-inflammatory macrophages in wound tissues was evaluated using immunofluorescence staining on days 1, 3, and 6 post-injury (Fig. 6f). As shown in Fig. 6g and h, wounds in the healthy wounds exhibited a marked increase in iNOS⁺ cells by day 3, which subsequently declined noticeably by day 6. In contrast, the model group showed only a slight rise in M1 macrophage infiltration throughout this period, with significantly fewer iNOS⁺ cells compared to healthy wounds on day 3. A similar kinetic pattern was observed in wounds treated with SP-loaded bilayer hydrogels during 1-3 days. However, by day 6, SP/IL-10@Bilayer showed iNOS⁺ cell levels comparable to healthy wounds, whereas SP@Bilayer still remained higher levels. The IL-10@Bilayer group showed a consistently lower distribution of iNOS⁺ cells compared to the model group, although this difference did not reach statistical significance. qRT-PCR analysis revealed consistent expression patterns of pro-inflammatory cytokine expression such as TNF-α, IL-1β, and IL-6 on day 3 across these groups (Fig. 6i–k). Our findings indicated that the SP/IL-10@Bilayer treatment facilitated the restoration of a physiological inflammatory progression during the early healing phase of diabetic wounds, as evidenced by enhanced neutrophil recruitment and M1 macrophage infiltration that closely mirrored those in healthy wounds (Fig. 6l).

We further investigated the effect of SP/IL-10@Bilayer on macrophage polarization during the later phase of diabetic wound healing, with particular emphasis on its role in promoting M2c macrophage polarization (Fig. 7a). SP@Bilayer and IL-10@Bilayer were set for comparison. Saline-treated diabetic wounds and healthy wounds were used as controls. Immunofluorescence staining for CD163 was performed to track the temporal dynamics of M2c macrophages from days 3 to 12 post-injury. (Fig. 7b and c). All groups exhibited a similar trend that CD163⁺ cells increased by day 6 before gradually declining. However, the model group and SP@Bilayer had consistently lower CD163⁺ levels, with no significant difference between them. IL-10 loaded bilayer hydrogels including SP/IL-10@Bilayer and IL-10@Bilayer achieved comparable CD163⁺ populations to healthy wounds on day 6. To further validate M2c macrophage polarization, we assessed the expression levels of Mer receptor tyrosine kinase (MerTK), IL-10 and TGF-β1 on day 6. MerTK expression is a characteristic marker of M2c macrophage subtype and plays an essential role in the efferocytosis of apoptotic cells [43]. As shown in Fig. 7d, both the model and SP@Bilayer groups exhibited significantly lower MerTK expression compared to healthy wounds. The IL-10@Bilayer group showed a partial recovery in MerTK expression, while the SP/IL-10@Bilayer group reached levels similar to those in healthy wounds. The secretion of anti-inflammatory cytokines such as IL-10 and TGF-β1 in the IL-10-loaded bilayer hydrogel groups were elevated to near-normal levels on day 6, significantly surpassing those in the model and SP@Bilayer groups (Fig. 7e and f). These findings indicate that the delayed release of IL-10 mediated by SP/IL-10@Bilayer effectively promoted M2c macrophage polarization to a physiologically relevant level during the later healing phase of diabetic wounds, facilitating the resolution of inflammation via efferocytosis of apoptotic cells and secretion of anti-inflammatory cytokines (Fig. 7g).

Fig. 7.

Bilayer hydrogel modulates M2c macrophage polarization in the later healing phase of diabetic wound. a Experimental timeline for assay of M2c macrophage polarization. b Immunofluorescence staining for CD163 in wounds on days 3, 6, 9 and 12 after injury. c Quantitative analysis of CD163⁺ cells in each group. d-f Relative expression of MerTK, IL-10, and TGF-β1 on day 6. g Schematic illustration of M2c macrophage polarization regulated by SP/IL-10@Bilayer and its contribution to inflammation resolution. All data were generated from three independent experiments and presented as the means ± standard deviation. Statistical analysis was performed by one-way ANOVA. # means significant difference compared to the normal group. #p < 0.05 and ##p < 0.01; ∗ means significant difference compared to model group. ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001; & means significant difference compared to SP/IL-10@Bilayer group. & p < 0.05.

3. Discussion

The present study demonstrated that diabetic wounds exhibited an attenuated early inflammatory response, marked by reduced neutrophil infiltration within the first day post-injury and impaired pro-inflammatory macrophage polarization by day 3 compared to healthy wounds. This leads to persistent macrophage recruitment and M1 macrophage accumulation, thereby maintaining a chronic state of low-grade inflammation that is marked by an elevated ratio of pro-inflammatory M1 to anti-inflammatory M2 macrophages [44]. Rapid delivery of SP within the first day effectively augmented neutrophil recruitment by day 1, which subsequently declined by day 3. The neutrophils further stimulated immune cell influx, evidenced by elevated pro-inflammatory macrophage infiltration by day 3, reaching levels comparable to those in healthy wounds. This intervention accelerated the healing rate at each timepoint compared to untreated diabetic wounds, underscoring the critical role of an adequate early inflammatory response for healing initiation.

Most existing research predominantly focused on the necessity of anti-inflammatory actions in diabetic wound healing. Our work demonstrates that treatment that suppressed inflammation without permitting an adequate early inflammatory response (IL-10@Bilayer group) failed to accelerate wound closure compared to that allowing adequate early inflammation (SP@Bilayer group). As supported by established literature, neutrophils fulfill multifaceted functions in wound healing, encompassing pathogen defense, debris clearance, inflammatory modulation, and support for angiogenesis through the provision of a provisional matrix scaffold [43,45]. In the context of inflammation regulation, they not only amplify inflammation by recruiting monocytes/macrophages and lymphocytes but also contribute to its resolution. Following apoptosis, the clearance of neutrophils via macrophage-mediated efferocytosis facilitates the release of anti-inflammatory signals, thereby promoting the transition from the inflammatory phase to the proliferative stage of healing [46,47]. Therefore, moderate modulation of the acute inflammatory response, particularly recruitment of neutrophils, represents a promising early intervention strategy to reset the healing trajectory in diabetic wounds.

By comparison with healthy wounds, the resolution of inflammation in the SP treatment group remained delayed during later stages, revealing the limitations of inflammation modulation exclusively in the early phase. We identified the M2c macrophage subset as a critical mediator of inflammation resolution and scarless wound repair through enhanced efferocytosis of apoptotic neutrophils, elevated secretion of IL-10, and modulation of the late-phase inflammatory microenvironment via suppression of pro-inflammatory mediators and upregulation of anti-inflammatory factors. It is noteworthy that, IL-10 treatment has been reported as an anti-inflammatory and proregenerative treatment approach in diabetic wound repair, which harvested accelerated healing outcomes than untreated diabetic wounds [19,48]. However, our findings demonstrated that IL-10 treatment alone by IL-10@Bilayer showed accelerated wound healing by day 3 but delayed in later phases compared to normal wound healing processes. Meanwhile, IL-10 treatment alone caused disordered collagen fiber deposition during tissue regeneration. This observation further underscores the critical need to enhance the early inflammatory response in diabetic wound healing. We propose that the premature or excessive suppression of neutrophil activity and the pro-inflammatory response may compromise the initial debridement of the wound matrix. This, in turn, could provide a disorganized template for subsequent fibroblast migration and collagen deposition, ultimately leading to aberrant tissue repair [49]. Current therapeutic strategies for diabetic wounds predominantly target anti-inflammatory processes, especially by promoting macrophage polarization toward proregenerative phenotype [50]. This study suggests that management of diabetic wounds should be tailored according to the specific wound pattern. Specifically, for early-wound pattern, targeting early inflammatory response might be an effective yet underutilized strategy.

Precise temporal regulation of inflammatory cell dynamics is a fundamental prerequisite for the normalization of diabetic wound healing. This study developed a skin-mimetic bilayer hydrogel that enables the sequential delivery of SP and IL-10. This bilayer hydrogel replicates the structural hierarchy of human skin, consisting of a dense top layer that mimics the epidermal barrier for mechanical protection, and a loose bottom layer that emulates the ECM architecture to facilitate cellular infiltration and tissue regeneration. By systematically modulating pore size, the loose bottom layer facilitated rapid release of SP within the first day, while the dense top layer provided delayed and sustained release of IL-10 starting after day one. In contrast to existing bilayer hydrogels, our design offers several unique advantages: (1) It is constructed entirely from natural skin-derived polymers or their simple derivatives [51,52]; (2) Both layers are formed through dynamic crosslinking mediated by the polymers’ inherent functional groups, which imparts essential wound dressing characteristics to the hydrogel, including layer-specific pore size distributions, mechanical resilience, shape adaptability, and tissue adhesion; (3) The bilayer structure can be readily assembled via spontaneous interlayer bonding upon contact between individually pre-formed layers, resulting in seamless interfacial integration [53,54]; (4) The entire fabrication process of the hydrogel eliminates the need for exogenous crosslinkers or complex manufacturing steps, thereby endowing the hydrogel with excellent biocompatibility and satisfying the critical prerequisites for its clinical translation.

4. Conclusions

In summary, this study develops a skin-mimetic bilayer hydrogel for spatiotemporal modulation of inflammatory cell dynamics in diabetic wounds. This hydrogel recapitulates the architecture and composition of human skin, exhibiting spontaneous dynamic integration between layers and clinically applicable multifunctionality. Specifically, the bilayer system facilitates the restoration of the physiological healing trajectory in diabetic wounds by enabling a rapid release of SP from the bottom layer to enhance early neutrophil recruitment and acute inflammation post-injury, as well as a delayed and sustained release of IL-10 from the top layer to promote M2c macrophage polarization and scarless repair at later stages. Overall, this work provides a robust platform for spatiotemporally coordinating the inflammatory cascade. We believe our findings will establish a physiology-inspired early intervention strategy for diabetic wound management, with the potential to prevent chronic ulcer progression and fibrotic scar formation.

5. Experimental section

5.1. Materials

Keratin was extracted from duck feathers using our previously established method [55], with detailed information provided in the Supporting Information. All chemical reagents were used as received from commercial sources without further purification. Specifically, copper sulfate (CuSO₄) was acquired from Shanghai Aladdin Biochemical Technology Co., Ltd. Hyaluronic acid (HA, sodium salt, Mw = 200-400 kDa) and porcine skin gelatin were obtained from Bloomage Freda Biopharm Co., Ltd. and Beijing InnoChem Science & Technology Co., Ltd., respectively. Fluorescent markers including rhodamine B (RhB), FITC-labeled bovine serum albumin (BSA-FITC), Carboxyfluorescein diacetate, succinimidyl ester (CFDA SE), and DiD were purchased from Shanghai Titan Scientific Co., Ltd. and Dalian Meilun Biotechnology Co., Ltd. Substance P (SP) was sourced from MedChemExpress (USA), and both recombinant human IL-10 and murine IL-10 were obtained from PeproTech (USA).

5.2. Preparation of Ker/Cu and Gel/OHA hydrogels

The keratin/copper (Ker/Cu) hydrogels were synthesized by mixing keratin solution at a final concentration of 17 w/v% with CuSO₄ solution at varying final concentration of 0.04, 0.06 or 0.08 w/v%, respectively, at room temperature. After stirring for several seconds, the mixture was placed statically at room temperature to form a hydrogel. The resulting hydrogels were designated as Ker/Cu 0.04, Ker/Cu 0.06, and Ker/Cu 0.08, respectively. The Gel/OHA hydrogels were also prepared through a facile mixing method at ambient temperature. OHA was synthesized by the sodium periodate oxidation method, according to the literature [56,57]. For hydrogel fabrication, 4% w/v of OHA solution was mixed with gelatin solution at varying concentrations of 4%, 6%, or 8% w/v, respectively, in a volume ratio of 1:1 at 37 °C. After stirring for several seconds, the mixture was placed statically at room temperature to form a hydrogel. The resulting hydrogels were designated as as Gel 2/OHA, Gel 3/OHA, and Gel 4/OHA based on their gelatin content.

5.3. Preparation of bilayer hydrogel

The bilayer hydrogel was fabricated using a layer-by-layer gelation approach. First, the bottom layer was prepared by casting 500 μL of a Gel and OHA mixed solution into a circular mold (about 18 mm in diameter), followed by static gelation at room temperature for 10 min. Subsequently, the top layer was constructed by depositing 500 μL of a keratin and CuSO₄ mixed solution onto the pre-formed bottom layer, which was then allowed to gel at room temperature for 30 min. The total thickness of the resulting bilayer hydrogel was approximately 3 mm. For the SP and/or IL-10 loaded bilayer hydrogel, SP was incorporated into the bottom layer by mixing with the Gel/OHA solution to achieve a final concentration of 5 μg/mL, and IL-10 was mixed with the keratin/CuSO₄ solution at a final concentration of 250 ng/mL [48,58,59]. The resulting bilayer hydrogels were used for the in vitro experiments. To more clearly demonstrate the morphological characteristics of the bilayer hydrogel in the height direction, the preparation conditions of the bilayer hydrogel shown in Fig. 3a were slightly adjusted by using 500 μL of a Gel and OHA mixed solution and 500 μL of a keratin and CuSO₄ mixed solution, coupled with a circular mold of about 12 mm in diameter.

5.4. In vitro drug release test

The drug release profiles of bilayer hydrogel were determined using RhB and BSA-FITC as model drugs. RhB was loaded in the bottom layer, and BSA-FITC was loaded in the top layer. One piece of bilayer hydrogel was placed in 2 mL PBS (pH 7.4) with the height of the liquid below 1 mm and incubated at 37 °C under gentle agitation to simulate physiological conditions. At predetermined time intervals, 1 mL of release medium was collected and replaced with fresh PBS. The release amounts of RhB and BSA-FITC were measured by fluorescence spectroscopy.

5.5. Transwell migration assay

HL-60 promyelocytic leukemia cells (ZQ 0433, Shanghai Zhong Qiao Xin Zhou Biotechnology Co., Ltd.) were cultured in RPMI-1640 medium (Biosharp) supplemented with 20% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37 °C in a humidified 5% CO₂ atmosphere. For neutrophil-like differentiation, cells were seeded at a density of 5 × 10⁵ cells/mL and treated with 1.25% dimethyl sulfoxide (DMSO) in complete medium for 7 days, with medium replacement every 48 h. Wright-Giemsa staining was performed to observe the morphology of differentiated HL-60 (dHL-60) cells according to the reference. A 24-well transwell system was used to examine the chemotactic potential of SP@Gel/OHA hydrogel, with Gel/OHA hydrogel for comparison. In brief, 5 × 10⁴ cells were inoculated in the upper chamber, and 700 μL of the hydrogel leaching solution was added to the lower chamber. The leaching solution was obtained by incubating 30 mg hydrogel in 15 mL culture medium for 24 h. The culture medium without additional inducers was used as a control. After 12 h incubation at 37 °C, migrated cells were quantified using a hemocytometer.

5.6. Polarization of macrophages

The human monocytic leukemia cell line THP-1 (CL-0233, Pricella Life Science & Technology Co., Ltd.) was maintained in complete growth medium consisting of RPMI-1640 supplemented with 10% FBS, 0.05 mM β-mercaptoethanol, and 1% penicillin/streptomycin at 37 °C in a 5% CO₂ humidified incubator. To induce differentiation into M0 macrophages, THP-1 cells were seeded at 5 × 10⁵ cells/mL in RPMI-1640 complete medium containing 10 ng/mL phorbol 12-myristate 13-acetate (PMA) and incubated for 48 h at 37 °C under 5% CO₂. Subsequently, differentiated M0 macrophages were stimulated with Lipopolysaccharide (LPS, 100 ng/mL) for M1 polarization, IL-10 leaching solution (IL-10, 20.94 ng/mL) for M2c polarization, and combination of LPS and IL-10 working solution (IL-10/LPS) for co-stimulation. After another 48 h, macrophage cells were harvested and stained with antibodies against human iNOS (Servicebio, GB11119) and CD163 (Abcam, ab182422) for 30 min. Fluorescence images were acquired using fluorescence microscope.

5.7. Assays for neutrophil apoptosis and macrophage efferocytosis

To induce apoptotic neutrophils, DiD-labeled differentiated HL-60 cells were exposed to UV irradiation (254 nm, 6W, 0.02J/s/cm²) for 15 min, followed by Annexin V-FITC/PI flow cytometry analysis to confirm the occurrence of apoptosis. Concurrently, macrophages treated with LPS (M1 polarization), IL-10 (M2c polarization), or IL-10/LPS combination were fluorescently labeled with CFDA-SE, while untreated macrophages served as control. Following three PBS washes to remove excess dye, counted macrophages were co-cultured with apoptotic neutrophils at a 1:5 ratio in fresh medium for 8 h at 37 °C under 5% CO₂. After incubation, non-phagocytosed neutrophils were eliminated through three gentle PBS washes, and efferocytosis was quantified via fluorescence microscopy by enumerating DiD⁺ apoptotic neutrophils internalized within CFDA-SE⁺ macrophages across multiple fields of view.

5.8. In vivo diabetic wound healing assessment

6-7 weeks old male C57BL/6J mice were purchased from Shanghai Jiesijie Laboratory Animal Co., Ltd. Forty-five mice were kept under specific pathogen-free conditions with a 12 h light/12 h dark cycle. All mice had free access to food and water. Diabetic mice were induced using male C57BL/6J mice through intraperitoneal injection of freshly prepared streptozotocin (STZ, 50 mg/kg in citrate buffer, pH 5.5) for 7 consecutive days following daily 12-h fasting periods. Diabetes was confirmed when fasting blood glucose levels exceeded 16.7 mmol/L for 3 consecutive days. Both diabetic and healthy mice were anesthetized and subjected to aseptic creation of 6-mm circular full-thickness dorsal wounds. Diabetic mice were randomly allocated into four groups (n = 18) and treated with saline solution (model), SP-loaded bilayer hydrogel (SP@Bilayer), IL-10-loaded bilayer hydrogel (IL-10@Bilayer), and SP/IL-10 loaded bilayer hydrogel (SP/IL-10@Bilayer). Healthy mice received saline treatment were set as controls. To match the wound dimensions, SP- and/or IL-10-loaded bilayer hydrogels were fabricated using a 6-mm diameter mold. Specifically, 50 μL of a Gel/OHA mixture containing 5 μg/mL SP was layered with 50 μL of a keratin/CuSO₄ mixture containing 250 ng/mL IL-10. The resulting hydrogels measured approximately 3 mm in height and 6 mm in diameter. A single patch of the bilayer hydrogel was applied to the wound on day 0 after wounding. Wound healing progression was quantitatively monitored every 3 days. The scar size and scar score of the wound tissues were analyzed based on the Vancouver scar score system [60].

5.9. Histological, immunohistochemistry and immunofluorescence assay

H&E and Masson staining were performed to assess the healing process of the wounds. The expression levels of collagen I and collagen III were assessed by Sirius Red staining and imaged under polarized light microscope. Neutrophil recruitment was evaluated by immunohistochemical staining for Ly-6G (Servicebio, GB12229). The infiltration of pro-inflammatory (M1) macrophages and polarization of M2c macrophages were analyzed by immunofluorescence staining using antibodies against iNOS (Servicebio, GB11119) and CD163 (Servicebio, GB14027), respectively.

5.10. Assay of inflammation-related factors

Real-time PCR experiments were used to assess relative mRNA expressions of inflammatory factors at specific time points: neutrophil-related chemokines (CXCL1, MCP-1) on day 1; pro-inflammatory cytokines (IL-6, IL-1β, TNF-α) on day 3; and anti-inflammatory factors (IL-10, TGF-β1) along with the efferocytosis-related macrophage marker MerTK on day 6. All primer sequences are listed in Table S1. The relative mRNA expression level of the target gene was normalized to the control group and calculated using the 2^−ΔΔCt method.

5.11. Statistical analysis

All quantitative data are expressed as mean ± standard deviation from at least three independent experiments. Statistical analyses were conducted using GraphPad Prism 8.0 (GraphPad Software, USA). Comparisons among multiple groups were performed using one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. Statistical significance was defined as p < 0.05 for all comparisons.

CRediT authorship contribution statement

Mingrui Cui: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Hongtao He: Writing – review & editing, Resources. Hanzhi Lu: Methodology, Investigation, Formal analysis, Data curation. Xuexia Yang: Resources, Methodology. Shuo Chen: Writing – review & editing. Xiaojun Zhou: Writing – review & editing. Fulun Li: Validation, Supervision, Resources. Chuanglong He: Validation, Supervision, Resources, Project administration, Funding acquisition. Guang Yang: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

All animal experiments were conducted with approval from the Animal Care Committee of Donghua University, Shanghai, China (approval # DHUEC-NSFC-2023-27).

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 A. Supplementary data

The following is the Supplementary data to this article.