Simultaneous broad-spectrum antibacterial and regenerative therapy for diabetic wounds using sonosensitive peptide composite hydrogel

Abstract

Diabetic wound infections remain highly refractory to treatment due to persistent bacterial colonization, oxidative stress, dysregulated inflammation, vascular insufficiency, and impaired extracellular-matrix remodeling. Current therapeutic strategies remain limited by poor tissue penetration, inadequate infection control, and ineffective inflammatory-oxidative microenvironment modulation. Here, we engineered a sonosensitive antimicrobial peptide composite hydrogel, FFRK8@ZnO₂@fHAMA, by integrating human host-defense-peptide-derived FFRK8, microenvironment-responsive ZnO₂ microspheres, and fish-collagen-modified hyaluronic acid methacrylate (fHAMA) into an injectable bioactive matrix. Upon low-intensity pulsed ultrasound (LIPUS) irradiation, this hydrogel acts as a multifunctional regenerative dressing that couple broad-spectrum bacterial eradication with oxidative-stress attenuation, macrophage repolarization, angiogenic activation and extracellular-matrix reconstruction. Distinct from conventional passive dressings or antibiotic-dependent therapies, FFRK8@ZnO₂@fHAMA enables non-invasive and spatiotemporally precise activation, sustained local therapeutic retention and coordinated immune-redox-metabolic microenvironment remodeling while maintaining favorable biosafety. This sonosensitive peptide hydrogel offers a powerful bioactive strategy for repairing infected diabetic wound and may inspire next-generation therapeutic modality for chronic non-healing tissue regeneration.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-05014-9.

Introduction

Infected diabetic wounds, particularly diabetic foot ulcers (DFUs), represent one of the most severe complications of diabetes, carrying substantial risks of non-healing, amputation, and mortality with considerable healthcare burden [1, 2]. Globally, DFUs affect approximately 15–25% of patients with diabetes during their lifetime and account for a major proportion of diabetes-related hospitalizations and healthcare expenditures.[3, 4] Despite considerable advances in debridement, off-loading, vascular intervention, and infection management, clinical healing outcomes remain unsatisfactory, especially following the onset of infection [5, 6]. This therapeutic failure largely arises from a profoundly dysregulated wound microenvironment in which chronic hyperglycemia, neuropathy, and microvascular dysfunction collectively sustain persistent inflammation, excessive oxidative stress, endothelial and fibroblast dysfunction, impaired extracellular matrix remodeling, and inadequate tissue perfusion [7, 8]. Infection further exacerbates this pathological milieu through biofilm formation, sustained cytokine activation, and redox disequilibrium, thereby suppressing angiogenesis and disrupting the transition from the inflammatory phase to the proliferative and remodeling phases of wound healing [9, 10]. Consequently, management of infected diabetic wounds requires not only efficient bacterial eradication but also simultaneous alleviation of inflammation and oxidative stress, restoration of pro-regenerative cellular functions, and promotion of functional tissue regeneration [8, 11].

The therapeutic efficacy of conventional antibiotics in chronic wound infections is increasingly undermined by biofilm-associated tolerance and the escalating prevalence of multidrug-resistant pathogens [5, 6, 12]. Antimicrobial peptides (AMPs) have emerged as promising alternatives owing to their membrane-targeting mechanisms, broad-spectrum antimicrobial activity, low propensity for resistance development, and immunomodulatory properties [13–15]. However, their clinical translation remains constrained by the limited antimicrobial efficacy of short peptide sequences. Moreover, highly potent AMPs often rely on long or cyclic sequences and complex chemical modifications, which may increase the risks of hemolysis and off-target cytotoxicity [16]. In this context, FFRKSKEK (FFRK8) represents a particularly promising candidate. Derived from the human host-defence peptide LL-37, FFRK8 is a short, low-haemolytic peptide that addresses several critical limitations of AMP development by integrating synthetic simplicity and favourable biosafety with ultrasound-activated, rapid, and broad-spectrum antibacterial activity [17]. Upon ultrasound stimulation, FFRK8 achieved > 99% bactericidal activity against clinically isolated multidrug-resistant pathogens [17], including Staphylococcus aureus, methicillin-resistant Staphylococcus aureus (MRSA), Escherichia coli (E. coli), Staphylococcus epidermidis, Enterobacter cancerogenus, and Pseudomonas aeruginosa, whilst exhibiting negligible cytotoxicity and minimal intrinsic antibacterial activity in the absence of sonication. Mechanistically, the diphenylalanine motif confers piezoelectric responsiveness upon FFRK8, thereby enabling ultrasound-amplified membrane penetration, transient extracellular bactericidal reactive oxygen species (ROS) generation, and disruption of bacterial electron transport processes [17]. Collectively, these properties establish FFRK8 as a compelling non-antibiotic antibacterial module for the development of ultrasound-responsive biomaterials targeting infected diabetic wounds.

Ultrasound has emerged as an attractive external stimulus for diabetic wound therapy owing to its non-invasive nature, clinical accessibility, deep tissue penetrability, and spatio-temporal controllability [18]. Unlike conventional passive topical delivery, sonosensitizers enable on-demand therapeutic activation and enhanced tissue penetration within poorly vascularized wound environments [19, 20]. Although recent advances have demonstrated capabilities such as oxygen release, nanozyme activation, and controlled drug delivery [21, 22], many systems still depend on structurally complex gas carriers or multifunctional nanotherapeutics, such as perfluorocarbon emulsions and microbubble-based platforms, or multifunctional nano-therapeutics incorporating catalytic nanozymes, photothermal nanoparticles, or multicomponent drug-delivery nanocarriers. Therefore, a substantial need persists for a structurally simple yet mechanistically integrated sonosensitizer for the treatment of infected diabetic wounds.

Among various candidate inorganic components, zinc peroxide (ZnO₂) has attracted considerable attention because it decomposes under pathological acidic conditions to generate hydrogen peroxide and oxygen, thereby simultaneously modulating antibacterial activity and hypoxic microenvironments [23]. This property is particularly relevant to infected diabetic wounds, where bacterial burden, tissue hypoxia, oxidative imbalance, and impaired vascular regeneration are closely interconnected [10, 11, 23]. Meanwhile, hyaluronic acid methacrylate (HAMA) hydrogels have been widely employed as wound dressings owing to their excellent biocompatibility, high water-retention capacity, extracellular-matrix-mimetic characteristics, and ability to prolong the local retention of bioactive agents within irregular wound beds [24, 25]. Further functional modification with fish collagen may enhance their bioactivity by introducing additional matrix-mimetic and cell-interactive properties. Consequently, fish-collagen-modified HAMA (fHAMA) hydrogels may provide a more favorable scaffold for incorporating responsive antimicrobial and microenvironment-modulating components for repairing infected diabetic wound [26, 27].

Based on these considerations, we developed a multi-functional FFRK8@ZnO₂@fHAMA hydrogel for the treatment of infected diabetic wounds (Fig. 1A). This platform integrates sonosensitive broad-spectrum antibacterial activity with local microenvironment modulation, thereby promoting inflammation resolution, oxidative stress attenuation, angiogenesis, and tissue regeneration. The therapeutic efficacy of this hydrogel was systematically evaluated both in vitro and in vivo using infected diabetic wound models, while integrated transcriptomic and metabolomic analyses were performed to elucidate the underlying mechanisms (Fig. 1B). By integrating ultrasound-triggered infection control with microenvironmental remodeling and multi-omics mechanistic interrogation, this study proposes a translationally relevant therapeutic strategy for the management of complicated infected diabetic wounds.

Fig. 1.

Schematic illustration of FFRK8@ZnO₂@fHAMA as a sonosensitive nanoparticle formulation to accelerate infected diabetic wound healing. (A) Scheme shows the fabrication process of FFRK8@ZnO₂@fHAMA hydrogel. (B) Acting mechanisms of FFRK8@ZnO₂@fHAMA in infected diabetic wound treatment

Results

Physicochemical characterization of FFRK8@ZnO₂@fHAMA

We first evaluated the as-constructed FFRK8@ZnO₂@fHAMA system by systematically characterizing its particle morphology, crystalline structure, component integration, hydrogel microarchitecture, and release behavior. As shown in Fig. 2A, the synthesized ZnO₂ particles exhibited a relatively uniform spherical morphology with a rough, nanopitted surface under scanning electron microscopy (SEM). Such surface features are likely generated by the self-assembly of primary ZnO₂ nanocrystals during peroxide precipitation, followed by solvent evaporation and slight interparticle shrinkage during drying. This surface morphology is favorable for subsequent peptide adsorption and homogeneous incorporation into the hydrogel matrix, as it provides a stable interfacial framework while minimizing the risk of local aggregation. After peptide loading, transmission electron microscopy (TEM) revealed that FFRK8@ZnO₂ retained its structural integrity without obvious collapse or fragmentation (Fig. 2B), indicating that peptide assembly on the particle surface did not disrupt the ZnO₂ core structure. Energy-dispersive spectroscopy (EDS) elemental mapping further demonstrated the spatial co-distribution of Zn and O signals from the inorganic phase together with N and C signals attributable to the peptide component (Fig. 2C and S1), supporting the successful incorporation of FFRK8 onto ZnO₂ rather than mere physical mixing. This finding is particularly important because the therapeutic performance of the system depends upon the structurally integrated cooperation between the sonosensitive peptide and the responsive inorganic carrier.

Fig. 2.

Characterization of FFRK8@ZnO₂@fHAMA. (A) Scanning electron microscope (SEM) image of ZnO₂ nanospheres. Scale bar, 50 nm. (B, C) Transmission electron microscope (TEM) image (B) and energy dispersive spectrometer (EDS) elemental mapping images (C) of FFRK8@ZnO₂ nanospheres. Scale bar, 50 nm and 20 nm, respectively. (D, E) Crystal lattice structure (D) and X-ray diffraction (XRD) pattern (E) of ZnO₂ nanospheres. (F) Hydrogel solid-liquid transformation process. (G) SEM images of the fHAMA hydrogel and FFRK8@ZnO₂@fHAMA. Scale bar, 100 μm and 50 μm, respectively. (H) EDS elemental mapping images of FFRK8@ZnO₂@fHAMA. Scale bar, 100 μm. (I) Fourier transform infrared (FT-IR) spectra of FFRK8, ZnO₂, fHAMA, ZnO₂@fHAMA, and FFRK8@ZnO₂@fHAMA. (J-K) Release curves of FFRK8 peptides (J) and Zn²⁺ ions (K) from FFRK8@ZnO₂@fHAMA with or without LIPUS-activation under different pH conditions (pH = 7.5 and 5.5)

The crystalline structure of the inorganic phase was subsequently confirmed by high-resolution lattice imaging and X-ray diffraction (XRD) analysis (Fig. 2D and E). The presence of well-defined lattice fringes indicated that the synthesized ZnO₂ possessed a highly ordered crystalline structure rather than an amorphous peroxide precipitate (Fig. 2D). Consistently, the XRD pattern verified the successful formation of crystalline ZnO₂ (Fig. 2E and S2). The diffraction peaks corresponded well with the reference pattern, confirming the formation of crystalline ZnO₂. And no impurity peaks from ZnO were observed, indicating the high phase purity of the synthesized nanoparticles. Importantly, this structural confirmation is not merely descriptive but also functionally significant, as the peroxide-containing crystalline core underpins the microenvironment-responsive reactivity of ZnO₂, thereby providing the physicochemical foundation for its antimicrobial and wound-modulating activities.

At the hydrogel level, the fish-collagen spectrum exhibited characteristic broad proton resonance peaks. In contrast, the modified HAMA system exhibited signals corresponding to the vinyl and methyl protons of methacryloyl groups (Figure S3), confirming the successful synthesis of the fHAMA hydrogel precursor. Following modification, the hydrogel precursor underwent rapid UV-triggered gelation to form a stable composite hydrogel (Fig. 2F). The resulting FFRK8@ZnO₂@fHAMA hydrogel also exhibited favorable formability and plasticity (Figure S4 and S5). SEM revealed that the fHAMA scaffold possessed a porous three-dimensional architecture, and this interconnected network was well preserved after incorporation of FFRK8@ZnO₂ (Fig. 2G). EDS mapping of the composite hydrogel further demonstrated the broad distribution of Zn- and O-related signals throughout the matrix (Fig. 2H and S6), suggesting the relatively homogeneous incorporation of peptide-loaded particles without obvious phase separation. And X-ray photoelectron spectroscopy (XPS) further validated these elemental constituents and supported the successful incorporation of individual components (Figure S7). Such a homogeneous distribution and the construction are desirable because diabetic wound dressings require both prolonged local retention and uniform therapeutic exposure throughout the wound bed.

Fourier-transform infrared (FT-IR) spectroscopy further confirmed the successful integration of the multiple components within the composite system (Fig. 2I). Compared with the spectra of the individual constituents, FFRK8@ZnO₂@fHAMA exhibited the coexistence of characteristic absorption bands corresponding to FFRK8, ZnO₂, and fHAMA, together with peak shifts and overlaps indicative of interfacial interactions among the different components. These spectral characteristics support the successful incorporation of the peptide-loaded inorganic component into the hydrogel network, rather than the simple physical coexistence of three isolated materials. Subsequently, zeta potential evaluation showed the presence of negative charge in composite hydrogel due to abundant carboxyl groups in fHAMA (Figure S8). Finally, taking into account the dynamic changes in the pH value of the infected diabetic wound, the release of the FFRK8 peptide and Zn²⁺ ions was detected under different pH conditions. The release profile showed that FFRK8 and Zn²⁺ was gradually released from the fHAMA matrix over time (Fig. 2J and K), indicating that fHAMA functioned as a sustained local reservoir rather than a bolus-release carrier. Moreover, negligible H₂O₂ generation was observed at physiological pH (7.5), whereas acidic conditions markedly promoted peroxide release from FFRK8@ZnO₂@fHAMA + LIPUS (Figure S9A). Dissolved oxygen measurements further demonstrated continuous oxygen evolution (Figure S9B). This pH-responsive behavior is consistent with the known decomposition characteristics of ZnO₂ and supports its suitability for infected diabetic wounds, where the locally acidic microenvironment facilitates the controlled peroxide generation. From a therapeutic perspective, the release of FFRK8 gradually approached a plateau after approximately 2 days, providing a rational basis for determining treatment intervals and optimizing therapeutic protocols in animal models. This trend was also confirmed in rats wound of FFRK8 release simulations. Through quantitative ELISA analysis, FFRK8@ZnO₂@fHAMA + LIPUS group demonstrated substantially higher FFRK8 concentrations within wound area and a prolonged retention profile compared with free FFRK8 (Figure S10). Furthermore, owing to the relatively alkaline microenvironment of infected diabetic wounds during the early stage of treatment [28–30], the release of the FFRK8 peptide predominates the initial therapeutic phase. As the antibacterial effect progressively reduces the infectious burden, the dysregulated wound microenvironment gradually transitions toward a mildly acidic state, which further facilitates peptide release while simultaneously promoting Zn²⁺ ions release to exert anti-inflammatory and tissue-reparative effects (Fig. 2J and K).

In vitro antibacterial, antioxidative, and anti-inflammatory performance under LIPUS activation

Infection represents a major barrier to diabetic wound healing. Wound dressings with effective antimicrobial activity are therefore essential for preventing pathogen invasion and promoting repair [31]. We evaluated the antibacterial efficacy of LIPUS-activated FFRK8@ZnO₂@fHAMA, as effective infection control is essential for alleviating excessive inflammatory responses and restoring impaired tissue regeneration during the early stage of diabetic wound healing. The bactericidal activity of FFRK8@ZnO₂@fHAMA under LIPUS stimulation was evaluated against Gram-positive MRSA and Gram-negative E. coli. SEM images showed that MRSA and E. coli in the PBS group retained intact morphology with smooth cell membranes, whereas incomplete formulations or LIPUS alone induced only limited structural perturbation (Fig. 3A and S11). In contrast, bacteria treated with FFRK8@ZnO₂@fHAMA, particularly under LIPUS activation, exhibited pronounced membrane deformation, rupture, and cytoplasmic leakage (Fig. 3A). These findings were corroborated by live/dead fluorescence staining and biofilms fluorescence. The FFRK8@ZnO₂@fHAMA + LIPUS group predominantly exhibited red fluorescence, whereas control, incomplete-formulation, and LIPUS-alone groups showed strong green fluorescence or mixed green/red fluorescence (Fig. 3B, S12, S13). These results indicate that the material induced direct bactericidal damage rather than merely bacteriostatic effects. Mechanistically, LIPUS amplifies the antibacterial activity of FFRK8 through enhanced membrane penetration, piezoelectric polarization, transient extracellular bactericidal ROS generation, and disruption of bacterial electron transport [17].

Fig. 3.

Antibacterial, anti-inflammatory, and antioxidant functions of FFRK8@ZnO₂@fHAMA+LIPUS. (A) SEM images of MRSA and E. coli with different treatment for 4 h at 37℃. Representative images were from 3 independent experiments. Scale bar, 500 nm. (B) Representative MRSA and E. coli live/dead staining images and relative quantitative pie chart (upper right of each image) after different treatments. Live bacteria stained with SYTO 9 (green) and dead bacteria stained with PI (red), respectively. Scale bar, 20 μm. (C, D) Confocal images (C) and relative quantification (D) of intracellular oxidative stress-associated ROS in RAW 264.7 macrophages under different treatment conditions (n = 3). Scale bar, 50 μm. (E, F, G) JC-1 immunofluorescence (E) and relative quantification analysis (F, G) of mitochondrial membrane potential in RAW 264.7 cells under different treatment conditions (n = 3). Scale bar, 50 μm. (H) Schematic illustrating promotion of macrophage polarization under FFRK8@ZnO₂@fHAMA+LIPUS. (I, J, K) Immunofluorescence staining (I) and quantification of CD86 (J) and CD206 (K) in LPS-incubated RAW264.7 cells (n = 3). Scale bar, 50 μm. (L) Western blot analysis of CD86 and CD206 protein isolated from LPS-incubated 264.7 macrophages treated with different materials, respectively. The data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 indicate significant differences compared with the Control + PBS group, while ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001 indicate significant differences compared with the LPS/H₂O₂ + PBS group

Given this antibacterial mechanism, we next established oxidative stress and inflammatory models in RAW264.7 macrophages through 24 h lipopolysaccharide (LPS) stimulation or 4 h H₂O₂ exposure, respectively, to evaluate antioxidative capacity. Fluorescence imaging with the ROS-sensitive probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) showed that all material-treated groups reduced intracellular fluorescence intensity compared with the LPS + PBS group, with the most pronounced reduction observed in the FFRK8@ZnO₂@fHAMA + LIPUS group (Fig. 3C and D). However, LIPUS alone appears to have minimal effects (Figure S14). These findings suggest that the LIPUS-activated composite hydrogel effectively alleviated LPS-induced intracellular ROS accumulation. Notably, the ROS contributing to antibacterial activity are transient extracellular reactive oxygen species generated during LIPUS-mediated activation of FFRK8, which primarily act on the bacterial membranes and electron transport processes. By contrast, that measured in RAW264.7 macrophages represent intracellular oxidative stress induced by bacterial infection and inflammatory stimulation. As bacterial burden is reduced and inflammatory signaling is alleviated, intracellular oxidative stress correspondingly decreases. Therefore, these results demonstrate successful restoration of redox homeostasis within host cells. In addition, mitochondrial membrane potential (ΔΨm) was assessed using the JC-1 fluorescent probe [32]. Cells in the control group predominantly exhibited red fluorescence, reflecting JC-1 aggregation under normal ΔΨm, with minimal green fluorescence corresponding to the monomeric form. In contrast, H₂O₂ exposure induced a marked fluorescence shift from red to green (Fig. 3E and S15A), indicative of mitochondrial depolarization. Treatment with FFRK8@ZnO₂@fHAMA, particularly under LIPUS activation, markedly reversed this alteration by effectively restoring red fluorescence to levels comparable to the control group while substantially reducing green fluorescence (Fig. 3F-G and S15B).

To further evaluate the therapeutic potential of LIPUS-activated FFRK8@ZnO₂@fHAMA during the inflammatory phase of wound healing, in which macrophages predominantly adopt an M1-like pro-inflammatory phenotype, we investigated its immunomodulatory capacity to promote M2-like anti-inflammatory macrophage polarization using an LPS-stimulated macrophage inflammatory model (Fig. 3H) [33–35]. Firstly, flow cytometry analysis of bone marrow-derived macrophages (BMDMs) demonstrated that LPS stimulation markedly increased the proportion of CD86-positive macrophages, whereas treatment with FFRK8@ZnO₂@fHAMA + LIPUS significantly reduced CD86 expression and increased CD206 expression, indicating a shift toward an anti-inflammatory M2-like phenotype (Figure S16 and S17). Consistently, immunofluorescence staining of RAW 264.7 macrophage further confirmed these findings, showing that LPS-induced M1-like polarization was characterized by elevated CD86 expression and reduced CD206 expression. In contrast, FFRK8@ZnO₂@fHAMA + LIPUS treatment effectively reversed this polarization pattern, as evidenced by decreased CD86 and enhanced CD206 signals (Fig. 3I-K). Notably, LIPUS treatment alone exhibited minimal effects on macrophage polarization (Figure S18), suggesting that the immunomodulatory activity was mainly attributed to the FFRK8@ZnO₂@fHAMA system. Furthermore, western blot (WB) and real-time quantitative polymerase chain reaction (RT-qPCR) analyses were highly consistent with these immunofluorescence findings (Fig. 3L, S19, S20). Compared with the LPS + PBS group, macrophages treated with FFRK8@ZnO₂@fHAMA + LIPUS exhibited significantly lower expression of pro-inflammatory mediators, including CD86, interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α), together with elevated expression of anti-inflammatory markers such as CD206, IL-10, and CD163 [22]. These findings collectively indicate the potent anti-inflammatory and macrophage-reprogramming capacity of LIPUS-activated FFRK8@ZnO₂@fHAMA.

In vitro promoted healing by FFRK8@ZnO₂@fHAMA under LIPUS activation

Impaired angiogenesis is a hallmark of chronic diabetic wounds, and endothelial dysfunction substantially contributes to deficient tissue oxygenation, impaired granulation tissue formation, and delayed wound closure [1, 36]. To evaluate the pro-angiogenic potential of these sonosensitive composite hydrogels, we co-cultured human umbilical vein endothelial cells (HUVECs) with the hydrogels and subsequently subjected to LIPUS stimulation. Immunofluorescence staining for the angiogenic markers CD31 and vascular endothelial growth factor (VEGF) revealed that LIPUS-activated FFRK8@ZnO₂@fHAMA markedly upregulated both CD31 and VEGF expression compared with control, incomplete-formulation, and LIPUS-alone groups, indicating potent pro-angiogenic capacity (Fig. 4A-C and S21). To further elucidate the molecular mechanism, angiogenesis-related markers were evaluated in HUVECs. WB and RT-qPCR analyses are also consistent with these immunofluorescence findings (Figure S22, S23), further supporting the enhanced angiogenic phenotype observed in immunofluorescence staining.

Fig. 4.

Pro-angiogenic and tissue-healing functions of FFRK8@ZnO₂@fHAMA+LIPUS. (A, B, C) Immunofluorescence staining (A) and relative quantification of CD31 (B) and VEGF (C) in HUVEC under different treatment conditions (n = 3). Scale bar, 50 μm. (D, E, F) Schematic diagram for cellular experiments of scratch assay (D), tube-formation assay (E), and Transwell migration assay (F). (G, H) Representative migration images (G) and relative quantification (H) of HUVECs after different treatments for 0 h and 24 h (n = 3). Scale bar, 500 μm. (I, J, K) Representative images of HUVECs tube formation assay (I) and quantitative analysis of total number of junctions (J) and vessel area percentage (K) (n = 3). Scale bar, 200 μm. (L, M) Representative Transwell assays images (L) and quantification (M) of HUVECs after different treatments (n = 3). Scale bar, 100 μm. The data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 indicate significant differences compared with the PBS (control) group

To further elucidate the mechanisms underlying the healing-promoting activity of FFRK8@ZnO₂@fHAMA, we performed a series of in vitro angiogenesis-related assays using HUVECs, including scratch wound-healing, tube-formation, and Transwell migration assays (Fig. 4D-F). The scratch assay demonstrated that the combination of LIPUS and FFRK8@ZnO₂@fHAMA significantly enhanced HUVECs migration compared with incomplete-formulation groups. Notably, the wound-closure reached 80.98% at 24 h following FFRK8@ZnO₂@fHAMA + LIPUS treatment (about 2.7-fold higher than the control group) (Fig. 4G and H). In addition, tube-formation analysis showed that the FFRK8@ZnO₂@fHAMA + LIPUS group exhibited significantly more junctions and larger vessel than the other groups, suggesting that the combination effectively promoted angiogenic activity of HUVECs in vitro (Fig. 4I-K). Similarly, the Transwell migration assay confirmed that the FFRK8@ZnO₂@fHAMA + LIPUS group exhibited the highest number of transmigrated cells, significantly exceeding that of the control group (Fig. 4L and M). Additionally, the angiogenic effect of LIPUS alone was limited (Figure S24), indicating that ultrasound primarily serves as an external activation to trigger the therapeutic functions of the FFRK8@ZnO₂@fHAMA system rather than acting as an independent therapeutic agent. Collectively, these results indicate that the material not only upregulated angiogenesis-related markers but also promoted the essential cellular behaviors required for vascular regeneration, including endothelial activation, lateral motility, directional migration, and vascular network assembly [37, 38].

FFRK8@ZnO₂@fHAMA under LIPUS activation accelerates healing of MRSA-infected diabetic full-thickness dorsal skin wound

Given its inflammation-resolving, antibacterial, pro-angiogenic, and reparative matrix-remodeling properties, LIPUS-activated FFRK8@ZnO₂@fHAMA presents a promising bioactive platform for diabetic wound healing. To evaluate its therapeutic efficacy, we established a streptozotocin (STZ)-induced diabetic rat model maintained on a high-fat/high-sugar diet, combined with MRSA infection (150 µL, 10⁸ CFU/mL), and created full-thickness dorsal skin defects (10 mm in diameter) [22, 39]. Successful diabetes induction was confirmed by random blood glucose concentrations exceeding 16.7 mmol/L, measured from tail-vein blood samples one week after STZ administration (Figure S25). The excisional wounds were treated every other day for four applications of fHAMA, ZnO₂@fHAMA, FFRK8@ZnO₂@fHAMA, or FFRK8@ZnO₂@fHAMA + LIPUS, with PBS-treated wounds serving as controls (Fig. 5A). Macroscopic observation revealed that wounds in the PBS group remained largely unhealed throughout the observation period, whereas progressive closure was observed in all material-treated groups, with the most rapid healing occurring in the FFRK8@ZnO₂@fHAMA + LIPUS group (Fig. 5B-D).

Fig. 5.

FFRK8@ZnO₂@fHAMA + LIPUS accelerates MRSA-infected diabetic full-thickness dorsal skin wound healing in rats. (A) Experimental timeline for in vivo dorsal skin wound (10 mm in diameter) healing studies. (B) Representative photographs of the dorsal wound healing process after different treatments, and trace of wound closure from day 0 to day 14. (C) Heatmap represents quantification of the wound area. (D) Quantitative analysis of wound closure percentage (n = 6). (E, F) Bacterial colonies (E) and quantification (F) of MRSA-infected wounds on days 3 in each group (n = 6). (G, H) H&E staining (G) and scar index (H) of dorsal wound tissues on day 14 (n = 6). Scale bar, 500 μm and 100 μm. (I, J) Masson staining (I) and collagen volume fraction (J) of dorsal wound tissues on day 14 (n = 6). Scale bar, 500 μm and 100 μm, respectively. The data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 indicate significant differences compared with the PBS (control) group

Wound exudates were collected from the wound margins on day 3. Notably, the FFRK8@ZnO₂@fHAMA + LIPUS group exhibited only trace bacterial growth with markedly lower colony counts than the other groups (Fig. 5E and F), attributable to the enhanced bactericidal activity of FFRK8 under LIPUS activation. This finding is consistent with the in vitro antibacterial results and indicates that LIPUS activation effectively preserved and amplified the anti-MRSA activity of FFRK8 within the wound microenvironment. Given that persistent bacterial burden sustains inflammatory cytokine activation and impedes tissue reconstruction [1, 8], early bacterial clearance likely represented a critical initiating event enabling the wound to transition from an infection-dominated inflammatory state toward productive healing.

Consistent with the accelerated macroscopic wound closure, hematoxylin-eosin (H&E) staining revealed significant histological improvement in FFRK8@ZnO₂@fHAMA + LIPUS-treated wounds, evidenced by a lower scar index and reduced inflammatory cell infiltration compared with untreated diabetic controls (Fig. 5G and H). Moreover, greater collagen deposition was observed in the FFRK8@ZnO₂@fHAMA + LIPUS group than in other groups, indicating enhanced extracellular matrix (ECM) remodeling (Fig. 5I and J). As collagen deposition is essential for cell proliferation, differentiation, and wound healing,³¹ these findings further support the pro-regenerative effect of the treatment.

Immunohistochemical staining demonstrated that IL-6 and TNF-α remained highly expressed in PBS-treated wounds, whereas all material-treated groups reduced expression of these pro-inflammatory cytokines to varying degrees, with the most pronounced suppression observed in the FFRK8@ZnO₂@fHAMA + LIPUS group (Fig. 6A-C). Enzyme-linked immunosorbent assay (ELISA) analysis of tissue lysates further confirmed the gradual restoration of inflammatory homeostasis, evidenced by decreased levels of IL-6 and TNF-α (Figure S26).

Fig. 6.

Effects of FFRK8@ZnO₂@fHAMA+LIPUS on inflammation resolution, vascular growth promotion, and fibrosis acceleration in MRSA-infected diabetic full-thickness dorsal skin wound. (A, B, C) Representative images (A) and quantitative statistics (B, C) of IL-6 and TNF-α immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. (D, E, F) Representative images (D) and quantitative statistics (E, F) of iNOS/CD68 and Arg-1/CD68 immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. (G, H, I) Representative images (G) and quantitative statistics (H, I) of VEGF/CD31 immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. (J, K, L) Representative images (G) and quantitative statistics (H, I) of Collagen Ⅰ/α-SMA immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. The data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 indicate significant differences compared with the PBS (control) group

Subsequently, immunofluorescence staining was performed to investigate immune modulation and angiogenic activity during the wound healing. Red fluorescence labeled the M1 macrophage marker inducible nitric oxide synthase (iNOS), whereas green fluorescence labeled the M2 macrophage marker arginase 1 (Arg-1). Relative fluorescence intensity was quantitatively analyzed to assess the immune status of the wound microenvironment. Wounds treated with PBS or incomplete formulations exhibited abundant pro-inflammatory M1 macrophage infiltration. In contrast, the FFRK8@ZnO₂@fHAMA group exhibited a marked shift in macrophage polarization within granulation tissue, predominantly showing the anti-inflammatory M2 phenotype. Notably, compared with the non-LIPUS group, the sonication group exhibited a higher proportion of M2 macrophages (Fig. 6D-F), indicating that LIPUS stimulation enhanced the biological activity of the composite hydrogel.

The restoration of immune homeostasis was accompanied by recovery of vascularization-related signals. Immunofluorescence staining for the vascular markers VEGF and CD31 was performed to evaluate the effect of LIPUS-activated FFRK8@ZnO₂@fHAMA on wound angiogenesis. VEGF/CD31 co-staining revealed the weakest vascular-associated signals in PBS-treated wounds, whereas the FFRK8@ZnO₂@fHAMA + LIPUS group exhibited the strongest fluorescence intensity (Fig. 6G-I), providing in vivo evidence that the composite hydrogel enhanced neovascularization and thereby contributed to reversing the hypovascular microenvironment associated with diabetes. Finally, collagen I/α-smooth muscle actin (α-SMA) staining demonstrated that the ultrasound-activated group exhibited the strongest matrix-remodeling-associated signals (Fig. 6J-L). On day 14, elevated expression of α-SMA and collagen I primarily reflects productive myofibroblast activation and reparative extracellular matrix reconstruction rather than pathological fibrosis [40]. Thus, these results demonstrate that FFRK8@ZnO₂@fHAMA + LIPUS suppresses inflammatory cytokines, rebalances macrophage phenotype, restores vascular regeneration, and advances reparative matrix remodeling in MRSA-infected diabetic wounds.

FFRK8@ZnO₂@fHAMA + LIPUS accelerates healing of E. coli-infected diabetic distal limb ulcer wounds

Encouraged by the favorable healing outcomes in the full-thickness dorsal wound model, we then evaluated the therapeutic efficacy of the platform in diabetic distal limb ulcers, which represent a more clinically challenging scenario owing to their high susceptibility to infection, ischemia, repetitive mechanical stress, and delayed healing [1, 41]. Diabetic wound infections frequently involve Gram-negative pathogens, among which E. coli is a well-recognized contributor to poor clinical outcomes [1, 8]. The model was established by creating a circular excisional wound in a diameter of 10 mm on the thigh of diabetic rats, followed by topical inoculation with an E. coli suspension (150 µL, 10⁸ CFU/mL) at the wound site (Fig. 7A). Throughout the treatment period, the FFRK8@ZnO₂@fHAMA + LIPUS group exhibited the most rapid wound closure among all groups and displayed skin appearance closely resembling normal tissue, without obvious hypertrophic or keloid-like scar formation by day 14 (Fig. 7B-D). In contrast, wounds in the PBS and incomplete-formulation groups exhibited delayed closure and unsatisfactory healing outcomes.

Fig. 7.

FFRK8@ZnO₂@fHAMA+LIPUS accelerates E. coli-infected diabetic distal limb wound healing in rats. (A) Experimental timeline for in vivo distal limb ulcer wound (10 mm in diameter) healing studies. (B) Representative photographs of the distal limb ulcer wound healing process after different treatments, and trace of wound closure from day 0 to day 14. (C) Heatmap represents quantification of the wound area. (D) Quantitative analysis of wound closure rate (n = 6). (E, F) Bacterial colonies (E) and quantification (F) of E. coli-infected wounds on days 3 in each group (n = 6). (G, H) H&E staining (G) and scar index (H) of distal limb ulcer wound tissues on day 14 (n = 6). Scale bar, 500 μm and 100 μm. (I, J) Masson staining (I) and collagen volume fraction (J) of distal limb ulcer wound tissues on day 14 (n = 6). Scale bar, 500 μm and 100 μm. The data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 indicate significant differences compared with the PBS (control) group

Bacterial burden analysis on day 3 demonstrated that viable E. coli counts remained highest in the PBS group and were most effectively reduced by FFRK8@ZnO₂@fHAMA + LIPUS treatment (Fig. 7E and F), confirming that the composite hydrogel retained potent broad-spectrum antibacterial activity in vivo. Histological analysis further revealed more advanced tissue continuity and a lower scar index in the ultrasound-activated group compared with the control group, while Masson’s trichrome staining demonstrated the highest collagen volume fraction in the same group (Fig. 7G-J). These findings indicate that the material not only suppressed Gram-negative bacterial infection but also promoted progression of the wound toward structural repair and extracellular matrix reconstruction. The sustained therapeutic efficacy of the platform in this more challenging wound environment further highlights its clinical translation potential.

Furthermore, in the E. coli-infected diabetic distal limb wound model, PBS-treated tissues displayed the strongest IL-6 and TNF-α staining signals (Fig. 8A-C), indicative of persistent inflammatory activation. In contrast, the FFRK8@ZnO₂@fHAMA + LIPUS group showed the greatest reduction in both cytokines, consistent with ELISA results (Figure S27). These findings suggest that the platform effectively alleviated the chronic inflammatory burden associated with Gram-negative bacterial infection rather than functioning solely as a local antimicrobial agent. Macrophage polarization analysis further revealed abundant iNOS-positive inflammatory macrophages and weak Arg-1 signals in the PBS group, whereas FFRK8@ZnO₂@fHAMA + LIPUS markedly reversed this pattern (Fig. 8D-F). In parallel, VEGF/CD31 co-staining demonstrated that LIPUS-activated FFRK8@ZnO₂@fHAMA most effectively enhanced angiogenesis-related signals (Fig. 8G-I), while collagen I/α-SMA staining demonstrated the most advanced reparative matrix remodeling (Fig. 8J-L). Therefore, the therapeutic mechanisms observed in the dorsal wound model were consistently reproduced in the diabetic distal limb wound model, as evidenced by alleviated inflammation, macrophage reprogramming, enhanced angiogenesis, and progression of the wound bed toward a mature remodeling phase. The biological activity and therapeutic mechanisms of FFRK8@ZnO₂@fHAMA + LIPUS were therefore highly consistent across distinct infected diabetic wound models. This cross-model reproducibility substantially strengthens the translational significance of the platform, indicating that it functions as a broad-spectrum, ultrasound-activated microenvironment-remodeling biomaterial rather than a site- or strain-specific wound dressing.

Fig. 8.

Effects of FFRK8@ZnO₂@fHAMA+LIPUS on inflammation resolution, vascular growth promotion, and fibrosis acceleration in E. coli-infected diabetic distal limb wound. (A, B, C) Representative images (A) and quantitative statistics (B, C) of IL-6 and TNF-α immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. (D, E, F) Representative images (D) and quantitative statistics (E, F) of iNOS/CD68 and Arg-1/CD68 immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. (G, H, I) Representative images (G) and quantitative statistics (H, I) of VEGF/CD31 immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. (J, K, L) Representative images (G) and quantitative statistics (H, I) of Collagen Ⅰ/α-SMA immunohistochemistry staining of infected wound tissues (n = 6). Scale bar, 100 μm. The data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 indicate significant differences compared with the PBS (control) group

Transcriptomic and metabolomic profiling reveal a shared reparative program induced by FFRK8@ZnO₂@fHAMA+LIPUSTr

To comprehensively investigate the transcriptional dynamics induced by LIPUS-activated FFRK8@ZnO₂@fHAMA treatment, total RNA was extracted from MRSA-infected dorsal wounds and E. coli-infected distal limb wounds in well-established diabetic rat models for transcriptomic analysis. Differential expression analysis of MRSA-infected diabetic wounds revealed 2,020 upregulated and 1,928 downregulated genes in the FFRK8@ZnO₂@fHAMA + LIPUS group compared with the untreated group (Fig. 9A). Similarly, in the E. coli-infected distal limb wound model, 3,132 upregulated and 2,836 downregulated genes were identified in the FFRK8@ZnO₂@fHAMA + LIPUS group (Fig. 9B). Venn diagram analysis revealed that 3,065 differentially expressed genes (DEGs) were shared between the two models (Fig. 9C), strongly suggesting that the platform induced a conserved biological response across anatomically and microbiologically distinct infected diabetic wound environments.

Fig. 9.

Transcriptomic analysis of infected diabetic wound tissues. (A, B) The differentially expressed genes (DEGs) analyzed by volcano plots of MRSA-infected diabetic full-thickness dorsal wound (A) and E. coli-infected diabetic distal limb wound (B) between untreated and FFRK8@ZnO₂@fHAMA+LIPUS group. (C) Venn diagram demonstrating the identification of 3065 common DEGs of two models. (D) KEGG enrichment analysis of common DEGs of two models between untreated and FFRK8@ZnO₂@fHAMA+LIPUS group. (E) Gene Ontology (GO) enrichment analysis of biological process for DEGs. (F, G) Gene set enrichment analysis (GSEA) in KEGG database (F) and GO databases (G) of the DEGs in MRSA-infected diabetic full-thickness dorsal wound models

Moreover, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis identified that these shared DEGs were significantly enriched in several inflammation- and healing-related pathways, including the tumor necrosis factor (TNF) signaling pathway, NF-κB signaling pathway, and ECM-receptor interaction pathway (Fig. 9D). TNF and NF-κB signaling are widely recognized as central regulators of inflammatory progression, and previous studies have demonstrated that inhibition of Tnf or Nfkb1 effectively alleviates inflammatory responses [42–45]. Gene Ontology (GO) analysis further implicated that dysregulated genes in diabetic wounds were significantly enriched in biological processes associated with inflammation and oxidative stress, collagen remodeling, cell differentiation, cell migration, and responses to bacterial infection (Fig. 9E, S28, S29). This expression profile reflects a molecular landscape characterized by persistent inflammation, impaired tissue regeneration, and bacterial infection, all hallmarks of non-healing diabetic wounds. Gene set enrichment analysis (GSEA) demonstrated that ECM-receptor interaction and collagen fibril organization were positively enriched following treatment, whereas TNF signaling, NF-κB signaling, defense responses to bacterial infection, and inflammatory responses to wounding were negatively enriched (Fig. 9F and G, S30, S31). Collectively, these findings indicate that FFRK8@ZnO₂@fHAMA + LIPUS shifted the wound transcriptomic landscape away from an infection-dominated inflammatory state toward a reparative program characterized by extracellular matrix reconstruction and angiogenic activation.

Given that metabolites directly reflect infection-associated alterations in the wound microenvironment, we performed metabolomic profiling of tissues from diabetic MRSA-infected dorsal wounds and E. coli-infected distal limb wounds. Principal component analysis (PCA) demonstrated high reproducibility and strong clustering consistency among samples (Fig. 10A). Notably, despite substantial metabolic differences between the distinct infection models, the overall metabolic profiles became more convergent following LIPUS-activated FFRK8@ZnO₂@fHAMA treatment (Fig. 10A). Volcano plot analysis demonstrated extensive metabolic reprogramming in both models, with 715 differentially expressed metabolites (DEMs) identified in the full-thickness dorsal wound model (Fig. 10B) and 456 DEMs in the distal limb ulcer model (Fig. 10C), including 69 metabolites shared between the two models (Fig. 10D, S32). These observations suggest that the therapeutic effects of FFRK8@ZnO₂@fHAMA + LIPUS were accompanied not only by transcriptional reprogramming but also by extensive remodeling of wound-tissue metabolism. Hub genes and key metabolites were identified through STRING network analysis and MetaboAnalyst-based KEGG pathway analysis, respectively (Figure S33, S34). Correlation analysis between hub-gene expression and key-metabolite abundance across all samples highlighted significant associations between wound-healing-related pathways and multiple differential metabolites (Fig. 10E and F). Although these correlations do not establish direct causality, they strongly suggest that inflammatory resolution, ECM remodeling, and metabolic redox adaptation were tightly coordinated during treatment. Furthermore, WB was performed to evaluate key components of the NF-κB signaling pathway in tissues from the distal limb ulcer model. Compared with the PBS group, incomplete-formulation groups moderately reduced the protein levels of phosphorylated IκBα, phosphorylated p65, and TNF-α, whereas LIPUS-activated FFRK8@ZnO₂@fHAMA produced the most pronounced inhibitory effect (Fig. 10G-J), indicating that the treatment effectively blocked NF-κB activation. Overall, FFRK8@ZnO₂@fHAMA + LIPUS suppresses infection-sustaining inflammatory signaling while restoring extracellular matrix communication and metabolic homeostasis, thereby enabling infected diabetic wounds to transition from chronic inflammatory arrest toward productive tissue regeneration.

Fig. 10.

Metabolomics analysis of infected diabetic wound tissues. (A) Principal component analysis (PCA) of the metabolomics data. (B, C) The differentially expressed metabolites (DEMs) analyzed by volcano plots of MRSA-infected diabetic full-thickness dorsal wound (B) and E. coli-infected diabetic distal limb wound (C) between untreated and FFRK8@ZnO₂@fHAMA+LIPUS group. (D) Venn diagram demonstrating 69 common DEMs of two models. (E) Heatmap showing Pearson correlation coefficient analysis between hub-genes and key-metabolite. The cell color represents the correlation R value, with red representing a positive correlation and blue representing a negative correlation (Pearson correlation: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). (F) A Sankey map of DEGs (left) in transcriptomics and DEMs (right) in metabolomics, according to the correlation of expression level. (G) Western blot analysis of pivotal protein of NF-kappa B signaling pathway isolated from the distal limb ulcer tissue lysates treated with different materials, respectively. (H, I, J) Determination of gray value of target protein p-Iκbα (H), p-p65 (I), and TNF-α (J) (n = 3)

Biocompatibility evaluation of FFRK8@ZnO₂@fHAMA+LIPUS

Adequate biosafety is essential for the clinical translation of wound dressings. LIPUS-activated FFRK8@ZnO₂@fHAMA was therefore systematically evaluated both in vitro and in vivo. Hemolysis assays demonstrated that neither the incomplete formulations, the composite hydrogel, nor LIPUS treatment induced significant hemolysis (Figure S35). Cytocompatibility was further evaluated using RAW264.7 macrophages and HUVECs. Fluorescence live/dead staining revealed no significant increase in dead cells following any treatment (Figure S37) and CCK-8 assays confirmed negligible effects on macrophage viability (Figure S38). Structural disruption of the cell membrane leads to the release of enzymes, including lactate dehydrogenase (LDH), which has relatively stable enzyme activity. We therefore assessed the LDH concentrations by OD₄₉₀ in culture medium. The LDH assay displayed similar results observed in the fluorescence staining and CCK-8 assay, indicated satisfactory cytocompatibility of incomplete-formulation and composite hydrogel system (Figure S39). In vivo biosafety was evaluated in healthy rats. Histopathological examination of the heart, liver, spleen, lungs, and kidneys revealed no observable tissue damage, inflammatory infiltration, or morphological abnormalities, indicating the absence of significant organ toxicity associated with LIPUS-activated FFRK8@ZnO₂@fHAMA or its individual components (Figure S39). Furthermore, comprehensive haematological and serum biochemical analyses performed at the experimental endpoint revealed no statistically significant differences in white blood cell (WBC) or red blood cell (RBC) counts, nor in hepatic function indicators (ALT and AST) or renal function indicators (CREA and BUN), compared with non-wounded diabetic rats (Figure S40).

Discussion

This study demonstrates that the therapeutic efficacy of LIPUS-activated FFRK8@ZnO₂@fHAMA in infected diabetic wounds arises from coordinated microenvironmental reprogramming rather than antibacterial activity alone. Infected diabetic wounds are sustained by a tightly interconnected pathological network in which bacterial persistence, chronic inflammatory signaling, oxidative stress, endothelial dysfunction, and impaired extracellular matrix remodeling mutually reinforce one another [46–48]. Within this framework, the sonosensitive antimicrobial peptide-based composite hydrogel moves beyond the conventional concept of passive wound coverage and instead functions as an active regulator of the wound microenvironment.

Unlike conventional hydrogel dressings that mainly function through passive drug release, the present platform integrates ultrasound-responsive peptide activation, ZnO₂-mediated microenvironment modulation, and extracellular matrix-mimicking hydrogel retention into a unified therapeutic system, enabling coordinated antibacterial activity and wound microenvironment reprogramming. A central strength of this platform lies in the use of FFRK8 as its antibacterial core. AMPs are attractive alternatives to conventional antibiotics owing to their broad-spectrum activity and lower propensity for resistance development. However, their clinical translation remains constrained by the limited efficacy of short peptides and the structural complexity often required for highly potent peptide candidates [13, 49, 50]. FFRK8, a short human-derived peptide, addresses these limitations through its synthetic simplicity and favorable biosafety profile. Loading FFRK8 onto ZnO₂ and embedding it within the fHAMA hydrogel further improved its suitability for wound therapy by enhancing local retention at the wound site. This feature is essential because even highly effective peptides are unlikely to exert sustained therapeutic effects in chronically infected wounds if rapidly diluted, degraded, or cleared from the wound bed. And the experimentally confirmed peroxide/oxygen generation further supports the rationale for selecting ZnO₂ as the inorganic component. LIPUS is equally critical to the therapeutic mechanism of the system. In this platform, ultrasound functions not merely as a physical adjunct but as the activating trigger that transforms the retained formulation into an active therapeutic depot. Compared with more structurally complex sonosensitizers based on gas carriers, nanozyme cascades, or multifunctional nanodrugs, LIPUS-activated FFRK8@ZnO₂@fHAMA achieves broad therapeutic functionality through a comparatively streamlined material design [19–21]. From a biomaterial perspective, this relative simplicity enhances mechanistic clarity and supports the reproducibility and translational potential of the platform.

Mechanistically, the data support a sequential yet interdependent model of wound repair. The marked membrane disruption observed in both MRSA and E. coli is consistent with the established sonosensitive behavior of FFRK8, including ultrasound-enhanced bacterial membrane interaction and ROS-associated injury [17]. However, the significance of the rapid antibacterial effect lies in the downstream reparative processes it enables. In diabetic wounds, persistent bacterial burden sustains cytokine production, oxidative stress amplification, and progressive tissue injury. FFRK8@ZnO₂@fHAMA + LIPUS effectively alleviated this burden, thereby enabling the wound microenvironment to transition out of the infection-dominated inflammatory state. The macrophage polarization results further strengthen this interpretation. The shift from M1 to M2 polarization indicates restoration of immune plasticity rather than mere suppression of inflammation, a process particularly critical for diabetic wound healing. The vascularization- and matrix-remodeling findings fit coherently within this mechanistic framework. Once inflammatory and oxidative constraints were alleviated, endothelial activity recovered, as evidenced by increased VEGF/CD31 expression, enhanced cellular migration, and improved capillary-like network formation. This observation is biologically consistent with the established dependence of diabetic wound angiogenesis on a permissive immune-redox microenvironment rather than isolated pro-angiogenic stimulation [51, 52]. The subsequent increase in collagen deposition and α-SMA-associated remodeling further indicates that improvements in immune and vascular conditions were successfully translated into productive structural repair. Collectively, these findings suggest that FFRK8@ZnO₂@fHAMA + LIPUS restored a biologically coordinated transition from inflammatory arrest to reparative remodeling.

A vital strength of the study is the reproducibility of the therapeutic mechanism across two distinct infected diabetic wound models. Although the dorsal MRSA-infected wound model and the distal limb E. coli-infected ulcer model differed substantially in anatomical context and microbial challenge, the treatment induced a highly consistent therapeutic trajectory characterized by reduced bacterial burden, attenuated inflammatory signaling, macrophage repolarization, enhanced angiogenesis, and improved matrix reconstruction. This consistency substantially strengthens the translational relevance of the platform, particularly given the heterogeneity and polymicrobial nature of diabetic wound infections. Integrated transcriptomic and metabolomic analysis further supported this interpretation by demonstrating that the observed phenotypic repair response was accompanied by restoration of metabolic homeostasis. Notably, methionine serves as a methyl donor for N⁶-methyladenosine modification [53], which in turn regulates macrophage polarization [54, 55]. These findings highlight the close interplay between metabolic regulation and immune reprogramming, linking the observed tissue repair process to a defined mechanistic signaling axis rather than to a purely descriptive phenotype.

Several limitations and challenges should be addressed in future investigations. First, although protein experiments strongly support NF-κB-related inflammatory signaling as a potential mechanism underlying this therapeutic response, we acknowledge that additional secondary mechanisms may contribute given the complexity of wound microenvironment modulation and may not have been fully captured in the present study. Alternative pathways warranting further investigation include TNF signaling, ECM-receptor interaction, collagen fibril organization, and defense responses to bacterial infection. However, it should be noted that these omics analyses primarily reveal pathway associations rather than direct causal relationships. Future studies combining pathway-specific functional assays, such as protein phosphorylation analysis or pharmacological inhibition, will be valuable for further signaling mechanistic validation. Second, although the M1/M2 macrophage classification was adopted for clarity and consistency with existing literature, macrophage polarization in vivo should be regarded as a dynamic continuum comprising multiple intermediate activation states beyond this simplified dichotomous framework [56, 57]. Third, the optimal LIPUS parameters may require further evaluation in large-animal models and future human studies. Compared with rats, large animals and humans possess distinct anatomical and physiological characteristics, including differences in skin, muscle, and adipose tissue thickness. Therefore, LIPUS parameters may require additional optimization to maximize therapeutic efficacy and tissue penetration. Encouragingly, such optimization appears feasible given the rapid development of LIPUS-based therapies, several of which have already received U.S. Food and Drug Administration (FDA) approval for diverse clinical applications, including fresh fractures and nonunion [58–60]. Finally, LIPUS-activated FFRK8@ZnO₂@fHAMA still faces several challenges with respect to clinical translation. Owing to the current insufficiency of preclinical data and the lack of standardized guidelines for biosafety evaluation, further systematic validation and long-term safety assessment will be required before clinical application. Nevertheless, these limitations do not diminish the central contribution of the work.

In conclusion, this study presents a sonosensitive FFRK8@ZnO₂@fHAMA hydrogel as a multifunctional platform for infected diabetic wound repair. By integrating broad-spectrum antibacterial activity with microenvironment modulation, this system enabled bacterial clearance, oxidative-stress attenuation, inflammation resolution, angiogenesis promotion, and tissue remodeling. In both MRSA-infected dorsal wounds and E. coli-infected distal limb wounds, LIPUS-activated FFRK8@ZnO₂@fHAMA markedly accelerated healing and restored a more regenerative wound microenvironment. These findings support LIPUS-activated biomaterials as a promising strategy for complicated infected diabetic wounds and other chronic non-healing tissues.

Methods and materials

Animals, cell lines, bacteria strains, and materials

All animal studies complied with policies of the National Ministry of Health and were approved by the Institutional Review Board of Chongqing Medical University (IACUC-CQMU-2025–11162). Male Sprague–Dawley (SD) rats (170–200 g, SPF), and high-fat fodder (> 45% fat) were obtained from Enswell Biotechnology Co., Ltd. (Chongqing, China). All animal experiments were conducted exclusively with male animals to eliminate the potential confounding effect of estrogen on wound healing and glycemic control in the diabetic model. STZ (≥ 98%) was purchased from Maokang Biotech Co., Ltd. (Shanghai, China).

Murine macrophages (RAW264.7, TIB-71), HUVECs (CRL-4053), MRSA (43300), and E. coli (25922) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cell culture reagents, including Dulbecco’s modified Eagle’s medium (DMEM), endothelial cell medium (EC-medium), fetal bovine serum (FBS), and penicillin–streptomycin, were purchased from Gibco (Grand Island, NY, USA). For bacterial culture, LB medium was purchased from Solarbio Life Sciences (Beijing, China), while TSB broth (CM0129) and bacterial culture agar were purchased from OXOID (Hampshire, UK).

LPS was purchased from Solarbio. Triton X-100 (P0096), ROS detection kits (S0033S), Transwell inserts (FTW004-6Ins), crystal violet staining solution (C0121, 100 mL), LDH cytotoxicity assay kit, and SYTO 9/PI live/dead bacterial double stain kit (D1245) were gained from Beyotime Biotechnology (Shanghai, China). The mitochondrial membrane potential assay kit with JC-1 (KGA1904-10), serum TNF-α and IL-6 ELISA kits were purchased from KeyGEN Biotechnology Co., Ltd. (Nanjing, China). Cell Counting Kit-8 (CCK-8) and Live/Dead staining kits were purchased from Dojindo (Kumamoto, Japan). DAPI was gained from Thermo Fisher Scientific (D1306). Matrigel was purchased from Corning (Corning, NY, USA). Primary antibodies against CD86, CD206, CD31, VEGF and Alexa Fluor-conjugated secondary antibodies were acquired from Abmart Inc. (Shanghai, China). Animal histological staining reagents were sourced from Saivs Biotechnology Co., Ltd. (Shanghai, China). RNA-later for transcriptomics samples collection was provided by Shanghai Biotechnology Corporation.

FFRK8 peptide (sequence: FFRKSKEK; purity > 95%), hyaluronic acid methacryloyl (HAMA; EFL-HAMA-150 K), and fish-collagen were synthesized and obtained from Qiangyao Biotechnology Co. (Shanghai, China). Zinc acetate dihydrate, hydrogen peroxide (30 wt%), sodium citrate, sodium hydroxide, Polyvinylpyrrolidone (PVP) and glucose water (5%) were purchased from Aladdin (Shanghai, China).

Preparation of ZnO₂ microspheres

ZnO₂ microspheres were synthesized via a modified one-pot peroxide precipitation method [23]. Anhydrous zinc acetate (1.0 g) and polyvinylpyrrolidone (PVP, 0.3 g) were dissolved in anhydrous ethanol (50 mL) under magnetic stirring to form a homogeneous precursor solution. Ammonia solution was then introduced to regulate the local coordination environment of Zn²⁺ and promote the formation of spherical peroxide intermediates. Subsequently, hydrogen peroxide (5 mL) was added dropwise under continuous stirring, and the reaction mixture was maintained at 37 ℃ for 24 h for particle growth. After completion of the reaction, the resulting ZnO₂ microspheres were collected by centrifugation (6000 rpm, 10 min), washed three times with deionized water and twice with ethanol, and vacuum-dried at 40℃ for 24 h. All post-synthesis procedures were conducted under reduced light exposure and mild-temperature drying conditions. The dried ZnO₂ microspheres were stored in the dark at 4℃ before further use.

Peptide loading onto ZnO₂ microspheres

The FFRK8 crude product was purified by preparative high-performance liquid chromatography and confirmed by electrospray ionization mass spectrometry. The peptide powder was lyophilized and stored at −20℃ until use [17]. FFRK8 was loaded onto ZnO₂ microspheres through electrostatically driven adsorption. Briefly, ZnO₂ microspheres were dispersed in PBS (pH 7.4) at 2 mg/mL by gentle bath sonication for 3 min to obtain a uniform suspension. In parallel, FFRK8 was dissolved in sterile deionized water at 2 mg/mL. The peptide solution was then added dropwise to the ZnO₂ suspension at a mass ratio of peptide: particle = 1:4 under gentle stirring. The mixture was incubated on a rotator at 25℃ for 12 h to allow adsorption equilibrium. After incubation, the FFRK8@ZnO₂ complexes were collected by centrifugation at 6,000 rpm for 5 min at 4℃ and washed twice with PBS to remove unbound peptide. The pellet was redispersed in sterile PBS and stored at 4℃ for immediate use.

FFRK8 loading efficiency was quantified by the bicinchoninic acid (BCA) assay. Briefly, ZnO₂ particles (20 mg) were dispersed in PBS (10 mL, 2 mg/mL), followed by dropwise addition of FFRK8 solution (20 mg in 10 mL sterile deionized water) under gentle stirring to allow peptide adsorption. After incubation, the suspension was centrifuged, and the supernatant containing unbound peptide was collected for quantification at 540 nm (OD₅₄₀) according to the FFRK8 standard calibration curve.

The loading efficiency (%) of FFRK8 peptide on ZnO₂ nanoparticles was calculated according to the following formula:

where W₀ and W_f represent the initial and unbound peptide amounts, respectively.

The OD₅₄₀ values of the supernatants from three independent experiments were 0.1862, 0.1981, and 0.1717, corresponding to residual peptide masses of 2.51, 2.79, and 2.16 mg, respectively. The calculated loading efficiencies were 87.45%, 86.05%, and 89.20%, indicating that the peptide-loading process exhibited good reproducibility and stability.

Fabrication of FFRK8@ZnO₂@fHAMA hydrogel

The HAMA precursor solution was prepared by dissolving HAMA in sterile PBS at 2.0 wt% at 37℃ under light protection. LAP was then added at 0.25 wt% and the solution was gently vortexed until fully dissolved [61]. The freshly prepared FFRK8@ZnO₂ suspension was then slowly introduced into the HAMA precursor under gentle stirring to obtain a homogeneous composite pre-gel. Then, the composite pre-gel was transferred into cylindrical polytetrafluoroethylene molds and irradiated using a 365 nm UV lamp at an intensity of 10 mW/cm² for 60 s to form FFRK8@ZnO₂@fHAMA hydrogel in situ. All precursor components were prepared under aseptic conditions and passed through a 0.22 μm filter before the addition of ZnO₂ microspheres. The final formulation contained 1.5 wt% HAMA, 0.2 wt% LAP, and 1.0 wt% FFRK8@ZnO₂.

Characterization of FFRK8@ZnO₂ and FFRK8@ZnO₂@fHAMA

For nanoparticle characterization, the morphology of ZnO₂ microspheres and FFRK8@ZnO₂ was observed by SEM. The microstructure, dispersion state, and peptide-loading-associated surface changes were further examined by TEM. Elemental mapping coupled with TEM was used to analyze the spatial distribution of Zn, O, Cl, and N elements, thereby verifying peptide incorporation onto the ZnO₂ particles. High-resolution TEM was further used to observe the lattice fringes of ZnO₂ and evaluate its crystalline structure, while powder XRD was performed to confirm the crystal phase of the ZnO₂ component, standard diffraction data (PDF-2, v2004) were from International Centre for Diffraction Data (ICDD) database. The surface chemical composition of the composite system was analyzed by XPS. FT-IR was used to identify the characteristic functional groups of FFRK8, ZnO₂, fHAMA, FFRK8@ZnO₂, and FFRK8@ZnO₂@fHAMA, and to verify successful assembly of the peptide-loaded inorganic particles within the hydrogel network through the coexistence or shift of characteristic absorption bands.

The in vitro release behavior of FFRK8 from FFRK8@ZnO₂@fHAMA was evaluated in PBS at 37℃. Briefly, pre-weighed hydrogel samples were immersed in release medium under gentle shaking. At predetermined time points, aliquots of the supernatant were collected and replaced with an equal volume of fresh PBS. The amount of released FFRK8 was quantified, and the cumulative release percentage was calculated accordingly.

Cell culture

RAW264.7 cells were cultured in high-glucose DMEM medium, supplemented with 10% FBS and 1% penicillin/streptomycin (P/S). HUVECs were maintained in EC-medium, according to the manufacturer’s instructions. Cells were incubated at 37℃ in a humidified atmosphere containing 5% CO₂ (ESCO, CLM-170B-8-TC). All cell lines used in this study were routinely tested and confirmed to be free of mycoplasma contamination.

Bacterial strains and growth conditions

MRSA and E. coli were used as representative Gram-positive and Gram-negative bacterial strains, respectively. Frozen MRSA bacterial stocks were first streaked onto tryptic soy broth agar (TSB agar, 1.5%) plates and incubated at 37℃ for 24 h, while E. coli bacterial stocks were streaked onto Luria–Bertani (LB) agar plates. A single colony was then picked and inoculated into liquid TSB medium (MRSA) and liquid LB medium (E. coli), followed by incubation at 37℃ in a shaking incubator to obtain an actively growing bacterial suspension. Before each experiment, the bacterial suspension was collected and adjusted to 1 × 10⁸ CFU/mL. All bacterial handling procedures were conducted under sterile conditions.

LIPUS treatment

The ultrasound probes were manufactured by Chongqing Ronghai Engineering Research Center of Ultrasound Medicine Co., Ltd. LIPUS treatment was performed at frequency of 1.0 MHz, acoustic intensity of 1.2 W/cm², duty cycle of 50%, pulse duration of 20 s, and total treatment of 15 min (all within physiotherapy level [18]) both in vitro and in vivo [17].

Bacterial live/dead staining

Bacteria in the logarithmic growth phase were collected and incubated with the ultrasound-responsive biomaterial, with or without LIPUS stimulation. After treatment, bacterial samples were stained with live/dead bacterial viability kit containing SYTO 9 and propidium iodide (PI). Following incubation in the dark, fluorescence images were acquired using a confocal laser scanning microscope (Zeiss LSM 880, Carl Zeiss, Oberkochen, Germany) under identical imaging settings for all groups. Live bacteria with intact membranes were visualized in green, whereas membrane-compromised or dead bacteria were visualized in red. Bactericidal efficacy was qualitatively assessed based on the relative distribution of green and red fluorescence.

Electron microscopy observation

The morphology of MRSA and E. coli was examined by electron microscopy to further investigate bacterial membrane damage after treatment. Bacteria subjected to different treatments were collected by centrifugation, washed gently with phosphate-buffered saline, and fixed with glutaraldehyde solution at 4℃. The fixed bacterial pellets were sequentially dehydrated in graded ethanol, dried, mounted on conductive stubs, sputter-coated with gold, and observed using a field-emission scanning electron microscope.

Macrophage assay in vitro

RAW264.7 macrophages (2 × 10⁶ cells/well) were seeded in 6-well plates and cultured at 37℃ in a humidified atmosphere containing 5% CO₂ for 24 h. After stimulated with LPS (100 ng/mL, 24 h), the culture medium was replaced with ultrasound-responsive biomaterial, and incubated for another 24 h. Cells without LPS considered as the control group. All groups were pretreated with 100 mM glucose for 10 min.

1. Fluorescence imaging of intracellular ROS levels

Following treatment, the cells were incubated with DCFH-DA working solution (10 µM) at 37℃ for 30 min in the dark. The cells were then washed twice with PBS to remove excess probe, and fluorescence images were acquired using the confocal laser scanning microscope at an excitation wavelength of 488 nm under identical imaging parameters for all groups. Intracellular ROS levels were qualitatively evaluated based on the intensity of green fluorescence.

• (2)Macrophage phenotype modulation assay

The cells were treated with PBS, fixed with 4% paraformaldehyde (PFA) for 10 min, permeabilized with 0.2% Triton X-100, and blocked with 4% bovine serum albumin (BSA). The samples were then incubated with primary antibodies against CD86 and CD206 (1:400 dilution) overnight. And, Alexa Fluor-conjugated secondary antibodies (1:400 dilution) were applied for 2 h at room temperature in the dark. Cell nuclei were counterstained with DAPI. Fluorescence images were acquired using the confocal laser scanning microscope under identical imaging settings for all groups. The fluorescence intensity (FI) of CD86 and CD206 was quantified using ImageJ software (National Institute of Health, Bethesda, MD, USA). Regions of interest (ROIs) were used background-corrected integrated density measurements and presented as CD86 (red), CD206 (green), and DAPI (blue).

• (3)Mitochondrial membrane potential assay kit with JC-1

RAW264.7 cells were stimulated with H₂O₂ (1 mM) for 4 h instead of LPS. Cells without H₂O₂ considered as the control group. After treatment, the culture medium was removed, and the cells were incubated with JC-1 staining working solution at 37℃ for 20 min in the dark. The cells were then washed twice with JC-1 staining buffer and immediately observed using the confocal laser scanning microscope using identical imaging parameters for all groups. Quantitative analysis of fluorescence intensity was performed using ImageJ software.

• (4)Flow Cytometry

After standard culture, BMDMs were incubated with different treatments for 24 h. Cells were then blocked with Fc receptor blocking buffer and stained with fluorochrome-conjugated antibodies for phenotypic characterization. Fluorescence signals were acquired using a flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA), and the data were analyzed using FlowJo software (BD Biosciences, v10.8.1).

Endothelial cells assay in vitro

HUVECs (2 × 10⁶ cells/well) were seeded in 6-well plates and cultured at 37℃ in a humidified atmosphere containing 5% CO₂ until the cells reached appropriate confluence. The cells were then incubated with the corresponding material-conditioned medium prepared in advance, with or without LIPUS stimulation, following the grouping strategy used in the immunofluorescence, migration, tube-formation or scratch assays. Cells with PBS treated considered as the control group.

1. Detection of CD31 and VEGF in HUVECs

After treatment, the cells were washed twice with PBS, fixed with 4% PFA for 30 min at room temperature, permeabilized with 0.1% Triton X-100 in PBS for 10 min, and blocked with 1% BSA for 1 h. The samples were then incubated at 4 ℃ with primary antibodies against CD31 and VEGF (1:400 dilution) overnight. After PBS washing, the cells were incubated with the corresponding fluorophore-conjugated secondary antibodies for 1 h at room temperature in the dark. Cell nuclei were counterstained with DAPI for 8 min. Fluorescence images were acquired using the confocal laser scanning microscope, and the fluorescence intensity of CD31 and VEGF was quantified using ImageJ software. Relative fluorescence intensity was normalized to those of the control group.

• (2)Transwell Migration Assay

HUVECs were harvested and resuspended in serum-free EC-medium, and 5 × 10⁴ cells were seeded into the upper chamber of each Transwell insert. The lower chamber was filled with FBS (10%) as a chemoattractant. The ultrasound-responsive biomaterials were prepared in advance and added to the lower chamber, with or without LIPUS stimulation. Cells were then incubated at 37℃ and 5% CO₂ for 24 h. After incubation, gently removed non-migrated cells, the migrated cells on the lower surface were fixed with 4% PFA and stained with crystal violet solution. The stained cells were imaged under inverted microscope (Eclipse Ti, Nikon, Tokyo, Japan), and the number of migrated cells was quantified by counting three randomly selected microscopic fields for each insert. Migration data were presented as the mean number of migrated cells.

• (3)Tube-formation assay

For tube-formation assay, 6-well plates were precoated with Matrigel and incubated at 37 ℃ for 30 min to gelation. HUVECs (2 × 10⁶ cells/well) were then seeded onto the Matrigel-coated wells in EC-medium. The ultrasound-responsive biomaterials were prepared in advance and added to the plates, with or without US stimulation. The plates were then incubated at 37℃ in a humidified atmosphere containing 5% CO₂. After 24 h, capillary-like network formation was observed under the inverted microscope. Quantitative analysis was performed using ImageJ software with the Angiogenesis Analyzer plugin, including number of junctions, and meshes area.

• (4)Scratch assay

When HUVECs cultured and reached a confluent monolayer, a linear scratch was then created in each well using a sterile 200 µL pipette tip. The wells were gently washed with PBS to remove detached cells and cell debris, and the medium was replaced and added different ultrasound-responsive biomaterials. Images of the scratched areas were acquired at 0 h and 24 h under the inverted microscope. The wound closure area was quantified by ImageJ software, and the migration percentage was calculated according to the following formula:

Diabetic models establishment of rats

All rats housed under specific pathogen-free (SPF) conditions at 20 ± 3℃, 40–70% relative humidity, with 12 h light/dark cycles. After one week of acclimatization, the rats were fed with high-fat diets (> 45% fat) and glucose water (5%) for 4 weeks. STZ was then freshly dissolved in citrate buffer (pH = 4.5) and administered by intraperitoneal injection (30 mg/kg) once weekly for two consecutive weeks. Rats’ blood glucose measurements > 16.7 mmol/L were considered successfully diabetic models and were included in the subsequent wound-healing experiments.

Diabetic infected wound model establishment

For the MRSA and E. coli-infected wound model, 60 diabetic SD rats were anesthetized with inhaled isoflurane. Then, the dorsal or right-hind limb hair was shaved and the skin was disinfected. A full-thickness excisional wound with an area of 1 cm² was created on the dorsal region of 30 rats, and on the right-hind limb of another 30 rats. The MRSA (150 µL, 1 × 10⁸ CFU/mL) or E. coli suspension (150 µL, 1 × 10⁸ CFU/mL) was spread onto the wounds. The successfully modeled rats were than randomly allocated into groups (n = 6). Wound healing was monitored by photographing and measuring wound areas at predetermined time points, while bacterial infection was assessed by swab culture and spread plate methods on day 3.

Wound photographing and closure analysis

Digital photographs of the wounds were acquired at predetermined time points (0, 3, 7, 10, 14 days) under identical camera settings with a scale marker placed adjacent to each wound. The wound margins were manually traced, and the wound area was quantified using ImageJ software. The relative wound closure rate was calculated according to the following formula:

where A₀ represents the initial wound area and A_t represents the residual wound area at the indicated time point.

Bacterial burden assessment

To evaluate the early antibacterial effect in vivo, tissue exudate from the infected area were aseptically harvested on 3-day after treatment. The collected tissue exudate was subjected to quantitative culture on strain-appropriate agar plates. After incubation under standard culture conditions, visible colonies were counted, and the bacterial burden was expressed as colony-forming units (CFU).

Histological analysis of diabetic rats

For histological evaluation, wound tissues were harvested at experimental end point, fixed in 4% paraformaldehyde, dehydrated, embedded in paraffin, and sectioned into 5 μm slices (RM2255 microtome, Leica). H&E staining was performed to assess tissue morphology and inflammatory cell infiltration. Masson’s trichrome staining was used to evaluate collagen deposition and overall extracellular matrix remodeling. Quantitative analysis of histological parameters was performed using ImageJ software.

Immunofluorescence of diabetic rats

For immunofluorescence staining, paraffin sections were deparaffinized, rehydrated, subjected to antigen retrieval, and blocked with bovine serum albumin. The sections were then incubated with primary antibodies against the indicated markers overnight at 4℃, followed by incubation with species-matched fluorophore-conjugated secondary antibodies at room temperature in the dark. Nuclei were counterstained with DAPI, and fluorescence images were acquired by the confocal laser scanning microscope. Depending on the experimental purpose, immunofluorescence staining was performed for macrophage phenotype markers (CD68 (1:4000 dilution), iNOS (1:4000 dilution) and Arg-1 (1:4000 dilution)), angiogenesis-related markers (CD31 (1:5000 dilution) and VEGF (1:5000 dilution)), and tissue fibrosis-related markers (α-SMA (1:3000 dilution) and Col-1 (1:3000 dilution)). Fluorescence intensity was quantified using ImageJ software.

Immunohistochemistry of diabetic rats

For immunohistochemistry, paraffin sections were deparaffinized, rehydrated, and subjected to antigen retrieval, followed by quenching of endogenous peroxidase activity and blocking with serum or bovine serum albumin. The sections were incubated with primary antibodies against the indicated target proteins at 4℃ overnight, followed by HRP-conjugated secondary antibodies. Signal development was performed using 3,3-diaminobenzidine, and counterstained with hematoxylin. Representative inflammatory markers included TNF-α (1:4000 dilution) and IL-6 (1:4000 dilution). Positive staining was quantified as the ratio of integrated optical density and positive area using Image Pro plus software (v6.0, Media Cybernetics, Bethesda, MD, USA).

Enzyme-linked immunosorbent assays

To evaluate the systemic inflammatory response, the wound tissue lysates were collected from rats on days 7 and 14 after treatment. The supernatant was subsequently separated by centrifugation (3000 rpm, 10 min). The supernatant samples and standards were added to 96-well ELISA plates pre-coated with TNF-α or IL-6 antibodies and incubated by the kit protocol. Then, enzyme-linked detection antibodies were added, followed by substrate development. The absorbance was measured using a microplate reader (SpectraMax iD5, Molecular Devices, San Jose, CA, USA) at 450 nm, and cytokine concentrations were calculated from the corresponding standard curves. All samples were analyzed in duplicate, and the results were expressed as pg/mL.

Real-time quantitative polymerase chain reaction (RT-qPCR)

RAW 264.7 macrophages were harvested and total RNA was extracted with TRIzol^™ reagent (Invitrogen, Carlsbad, CA, USA). Complementary DNA (cDNA) was synthesized from total RNA using the PrimeScript^™ RT-PCR Kit (Takara, Tokyo, Japan). Quantitative real-time PCR (qPCR) was subse quently performed with SYBR^® Green Master Mix (Qiagen, Venlo, The Netherlands) on a real-time detection system. Primers were ordered from Liantai Biotech (Shanghai, China). The primer sequences were listed in Table S1 (Supporting Information).

Western blotting analysis of macrophage and skin tissue

Total proteins were extracted from macrophages and skin tissues of distal limb ulcer models using a commercial protein extraction kit and quantified by BCA assay. Equal amounts of protein were separated on 7.5% SDS-PAGE gels (PG211; Epizyme, Cambridge, MA, USA) and subsequently transferred onto PVDF membranes (Merck, Darmstadt, Germany). Membranes were incubated overnight at 4℃ with the respective primary antibodies. Band intensities were quantified by densitometric analysis and normalized to β-Actin.

Biosafety assessments

1. Hemolysis assay:

Fresh whole blood was collected from healthy rats and treated rats into anticoagulant-containing tubes and centrifuged at 1100 g for 4 min at 4℃. The erythrocyte pellet was washed three times with saline until the supernatant became clear. Purified RBCs were then resuspended in physiological saline to obtain a 4% (v/v) RBC suspension. RBC suspension (1 mL) was incubated with biomaterial samples at 37℃ for 12 h. Ultrapure water and physiological saline were used as the positive and negative controls, respectively. After incubation, the samples were centrifuged (3000 rpm, 10 min) at 4℃, and the absorbance of the supernatant was measured at 542 nm using the microplate reader. The hemolysis percentage (%) was calculated according to the equation:

• (2)Hematological and biochemical tests

To assess systemic biosafety in vivo, whole-blood samples were collected from rats at the experimental endpoint into EDTA-containing tubes. Routine hematological parameters, including WBC, RBC, hemoglobin (HGB), hematocrit (HCT), and platelet count (PLT), were measured using an automated hematology analyzer according to the manufacturer’s instructions. Serum biochemical indices were analyzed to assess hepatic (ALT, AST) and renal (BUN, CREA) function.

• (3)CCK-8 assay

RAW264.7 macrophages (2 × 10⁶ cells/well) were seeded in 6-well plates and cultured under standard conditions. The cells were then incubated with the indicated biomaterial formulations for 24 h. At the end of the incubation period, 10% CCK-8 working solution was added to each well, and the plates were further incubated for 70 min (37℃, 5% CO₂). The absorbance at 450 nm was measured using the microplate reader. Relative cell viability was calculated as the ratio of the absorbance of the treated to the control groups.

• (4)LDH assay

RAW264.7 macrophages (2 × 10⁶ cells/well) and HUVECs (2 × 10⁶ cells/well) were seeded in 6-well plates and cultured under standard conditions separately. After incubated 24 h with the indicated biomaterial formulations, LDH kits were employed to evaluate the LDH release of different groups. Triton X-100 (1%, 15 min) were used as the positive control for maximal LDH release. The absorbance at 450 nm was measured using the microplate reader.

Transcriptome sequencing and analysis

Fresh tissue was processed using Qiagen RNA Lipid Tissue Kit. Quality of the RNA was determined using TapeStation (Agilent Technologies, Santa Clara, CA, USA). The sequencing strategy was paired-end 150 bp for NovaSeq 6000. FastQC package (v0.11.9) were used to perform the sequencing quality control, and reads were a aligned to the rat reference genome using Hisat2 (v2.2.1) with default parameters [62]. Then, StringTie (v1.3.3b) was used for assembly and quantification of counts matrix and TPM (transcripts per kilobase of exon model per million mapped reads) of each annotated gene [63].

We used DESeq2 (v1.42.0) to detect significantly DEGs [64], with a threshold of fold change (FC) > 2 and a Benjamini-Hochberg corrected p < 0.05. Subsequently, we utilized DAVID (https://david.ncifcrf.gov/) for KEGG and GO enrichment analysis to identify significantly enriched pathways [65]. The Gene Set Enrichment Analysis (GSEA) were performed by KEGG database and GO enrichment pathways. Then, DEGs were used to construct PPI network through STRING web tools [66], and all parameters were set to default values. The PPI results were analyzed and visualized by Cytoscape (V3.8.2) [67, 68]. The hub genes were chosen based on degree connectivity via cytoHubba [69] in Cytoscape. After maximal clique centrality (MCC) algorithm [70], all genes were ranked according to their intramodular connectivity, and only the top ten genes were selected as hub genes.

Metabolomics sequencing and analysis

For tissue samples used in metabolomics testing, they were collected at the end of the experiment, flash-frozen in liquid nitrogen, and then stored at −80℃ for subsequent analysis. After homogenizing and incubation on ice, the samples were centrifuged and injected into the LC-MS/MS system for analysis [71]. The mobile phases consisted of eluent, with a gradient profile optimized for comprehensive metabolite coverage. The mass spectrometer was operated in both positive and negative polarity modes.

Raw data were processed using Compound Discoverer 3.3 (Thermo Fisher Scientific, Waltham, MA, USA) for peak alignment, picking, and quantitation. Data were normalized against total spectral intensity. Molecular formulas were predicted, and peaks were matched against databases like mzCloud and mzVault for qualitative and quantitative results. Then, these metabolites were annotated using the KEGG, HMDB, and LIPIDMaps databases. Principal Component Analysis and Partial Least Squares Discriminant Analysis (PLS-DA) were conducted using metaX63. Univariate analysis (t-test) was employed to calculate statistical significance (p-values). Metabolites with variable importance in prediction (VIP) ≥ 1, p < 0.05 and FC > 2 were considered differentially expressed metabolites [53]. Enrichment analysis for metabolites was performed by the MetaboAnalyst 6.0 online web [72].

Statistical analysis

Data are presented as mean ± SD. In vitro experiments were performed with three independent replicates (n = 3), and in vivo experiments were conducted with six animals per group (n = 6). Statistical analyses and data visualization were performed in R using the ggplot2 package (v3.5.2). Statistical significance was determined by a one-way ANOVA test followed by Tukey’s multiple comparison analysis. Differences were considered statistically significant at p < 0.05.

Author contributions

Y.L.: Methodology, Validation, Formal analysis, Investigation, Resources, Data Curation, Writing - Original Draft, Writing - Review & Editing, Visualization; B.C.X.: Methodology, Investigation, Writing - Original Draft, Writing - Review & Editing; X.Z.: Validation, Investigation, Writing - Review & Editing; G.H.L.: Resources, Data Curation, Writing - Review & Editing; H.Y.Z.: Formal analysis, Investigation, Resources, Writing - Review & Editing; Y.X.Z.: Investigation, Resources, Writing - Review & Editing; X.C.W.: Data Curation, Writing - Review & Editing; Z.Y.Z.: Writing - Review & Editing; Z.W.: Writing - Review & Editing; Y.Z.: Writing - Review & Editing; Y.G.: Writing - Review & Editing; C.R.Y.: Writing - Review & Editing; J.Y.Z.: Conceptualization, Formal analysis, Writing - Review & Editing, Supervision; W.Y.Q.: Conceptualization, Methodology, Validation, Formal analysis, Writing - Review & Editing, Supervision; Y.F.Z.: Conceptualization, Methodology, Validation, Writing - Review & Editing, Supervision, Funding acquisition.

Funding

This work was financially supported by the National Natural Science Foundation of China (12474450), Chongqing Talent Program (2024030029), and the Fundamental Research Funds for the Central Universities (YG2026QNB30).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All experimental and research procedures described in this manuscript have been conducted in strict compliance with the policy of the National Ministry of Health of China. All research procedures have obtained official approval from the Institutional Review Board of Chongqing Medical University (IACUC-CQMU-2025-11162). This study did not involve the use of human data or tissues. Thus, Consent to Participate: not applicable.

Competing interests

The authors declare no competing interests.