Bioelectric Reawakening by a Self‐Powered Thermoelectric Hydrogel Accelerates Diabetic Ulcer Repair

ABSTRACT

Diabetic foot ulcer (DFU) remains difficult to heal due to disrupted endogenous bioelectricity together with persistent infection, inflammation, and oxidative stress. Reinstating wound bioelectricity therefore represents an attractive therapeutic strategy, and thermoelectric materials are particularly suited to this purpose by harvesting the natural skin–air temperature gradient without external power input. Here, a self‐powered ionic thermoelectric dual‐network hydrogel is developed to simultaneously reconstruct wound bioelectric cues and remodel the hostile DFU microenvironment. The hydrogel generates wound‐relevant microcurrents under physiological temperature gradients, while luteolin and Zn^{2 +} are incorporated as complementary bioactive modules to suppress bacterial burden and excessive inflammation, thereby establishing a pro‐regenerative niche. Meanwhile, the catechol‐containing dual‐network architecture imparts strong wet adhesion and robust mechanical stability for conformal wound coverage. Mechanistically, this study provides, to our knowledge, the first evidence that thermoelectric stimulation reprograms fibroblast repair behavior through bioelectric transduction into a Ca^{2 +}/calmodulin‐dependent phosphoinositide 3‐kinase/protein kinase B (PI3K/Akt) and extracellular signal‐regulated kinase (Erk) signaling network. The hydrogel exhibits broad‐spectrum antibacterial activity, immunomodulatory effects, and pro‐angiogenic capacity in vitro, and accelerates wound healing by 66.84% in diabetic rats. This work establishes a self‐powered strategy that integrates bioelectric restoration with microenvironment remodeling for DFU repair.

A self‐powered thermoelectric hydrogel is shown to reawaken wound‐site bioelectric cues while simultaneously remodeling the diabetic ulcer microenvironment. Integrated Zn^{2 +} and luteolin delivery suppresses infection and inflammation, whereas thermoelectric stimulation promotes cell migration, angiogenesis, and tissue regeneration. This synergistic strategy enables accelerated repair of infected diabetic wounds and highlights a bioelectric route for advanced wound therapy.

1. Introduction

Diabetic foot ulcer (DFU) is notoriously recalcitrant and progressive, with amputation rates approaching 20% [1]. Beyond multidrug‐resistant infection and the self‐perpetuating cycle of chronic inflammation and oxidative stress, accumulating evidence indicates that attenuation or loss of the wound's endogenous electric field (EEF) is a key contributor to nonhealing [2, 3, 4, 5]. In intact epithelia, asymmetric distributions of ion channels and transporters establish a transepithelial potential; upon injury this potential is short‐circuited, generating a lateral EEF oriented negatively at the wound center [6, 7]. This physiological EEF guides periwound cell migration, proliferation, and lineage commitment, thereby orchestrating efficient tissue repair [8, 9, 10, 11]. In diabetes, reduced expression and function of ion channels/transporters, coupled with electrolyte imbalance, compromise EEF integrity and further hinder healing cascades [12, 13].

Current electroactive dressings, such as triboelectric and piezoelectric systems, have shown promise in promoting angiogenesis. However, their reliance on exogenous ultrasound or mechanical actuation, combined with the generation of spatially disordered electric fields, limits clinical translatability and deviates from physiological EEF characteristics [13, 14, 15, 16, 17, 18]. In contrast, thermoelectric (TE) materials offer an intrinsically powered strategy by converting low‐grade thermal energy into bioelectric signals [19, 20]. Sodium 2‐acrylamido‐2‐methyl‐1‐propanesulfonate (AMPS‐Na), a prototypical ionic thermoelectric (i‐TE) monomer, transduces skin‐air temperature differentials into electrical energy via the ionic Seebeck effect [21, 22, 23]. This design confers two key advantages: self‐sustainability, as energy is derived directly from physiological thermal gradients, and biological alignment, as the generated electric field mimics native wound healing currents. Nevertheless, AMPS‐Na‐based hydrogels alone lack the mechanical robustness and wet adhesion required for on‐wound application, and electrical stimulation (ES) alone cannot simultaneously address infection, inflammation, and oxidative stress—the triad of pathological barriers in DFU [23].

To address these limitations, we propose a tri‐modal synergistic strategy integrating TE, antibacterial, and anti‐inflammatory functionalities, via a dual‐network hydrogel constructed through multi‐component interactions between AMPS‐Na, dopamine methacrylamide (DPMA), luteolin, and ZnSO₄ (Scheme 1). The AMPS‐Na backbone enables i‐TE conversion, while catechol moieties from DPMA confer strong wet adhesion [24]. Zn^{2 +} coordinates with catechol groups to form a metal–organic network (MON), and luteolin intercalates within the polymer matrix via π–π stacking and hydrogen bonding—together forming a dual‐network architecture (AMPS‐Na main chain – (DPMA–Zn^{2 +}) coordination network – luteolin interfill) that markedly enhances mechanical robustness and cohesion.

SCHEME 1.

Design and multifunctionality of the ADZ@Lut hydrogel. ADZ@Lut integrates an AMPS‐Na–based i‐TE network with DPMA, luteolin, and Zn^{2 +} to yield a dual‐network hydrogel. The skin–air temperature gradient drives Soret‐type ionic thermodiffusion, restoring a wound‐directed electric field and enabling sustained release of Zn^{2 +} and luteolin to suppress bacterial colonization and modulate inflammation. TE stimulation further promotes angiogenesis and activates Ca^{2 +}/calmodulin‐dependent phosphoinositide 3‐kinase/protein kinase B (PI3K/Akt) and extracellular signal‐regulated kinase (Erk) signaling, thereby enhancing fibroblast proliferation and migration.

Beyond structural reinforcement, this dual network delivers complementary bioactivities. Luteolin exerts chemical antibacterial effects by inhibiting bacterial ATP production and downregulating resistance gene expression, while Zn^{2 +} mediates physical envelope disruption—providing a two‐pronged attack against multidrug‐resistant bacteria that mitigates the limitations of single‐agent therapy [25, 26]. In parallel, luteolin activates nuclear factor erythroid 2–related factor 2 signaling to boost antioxidant enzyme activities and attenuates reactive oxygen species (ROS)‐driven inflammatory cascades, while Zn^{2 +} suppresses NLR family pyrin domain‐containing 3 (NLRP3) inflammasome activation to facilitate resolution of inflammation—jointly remodeling the immune–oxidative microenvironment through multiple pathways [27, 28]. The AMPS‐Na–mediated, sustained and oriented microcurrent further enhances transmembrane transport of Zn^{2 +}/luteolin via electropermeative effects, balancing early antibacterial action with long‐term tissue repair [29, 30, 31, 32, 33, 34].

In this study, we systematically characterize the structural, mechanical, TE, and bioactive properties of the hydrogel. We validate its antibacterial/antibiofilm efficacy, immunomodulatory activity, and pro‐regenerative effects in vitro, and demonstrate accelerated wound healing in a diabetic rat model. By integrating metal ions, natural products, and TE materials into a single platform, this work presents a self‐powered, multi‐target, and highly adaptable smart dressing for chronic refractory wounds, including DFU, postoperative infected wounds, and pressure ulcers.

2. Results

2.1. Characterization of Hydrogel Dressings

To decouple matrix‐intrinsic properties from cargo‐mediated effects, four comparator hydrogels were engineered for physicochemical characterization and functional evaluation: (i) AD (AMPS‐Na/DPMA; no cargo), (ii) AD@Lut (matrix loaded with luteolin), (iii) ADZ (matrix loaded with ZnSO₄), and (iv) ADZ@Lut (dual‐loaded with luteolin and Zn^{2 +}). Scanning electron microscopy (SEM) revealed that all formulations formed interconnected 3D porous networks (Figure 1a). The AD matrix exhibited comparatively large pores with smooth walls and uniform thickness. Incorporation of luteolin preserved the overall architecture while introducing dispersed, cluster‐like features adherent to pore surfaces. In ADZ, a subset of pores became narrower with increased surface roughness, whereas ADZ@Lut integrated both signatures—polyphenolic clusters and micron‐scale particulates—within a noticeably denser network. Elemental mapping by energy‐dispersive x‐ray spectroscopy (EDS) confirmed homogeneous distributions of constituent elements across all hydrogels (Figure 1b). To probe chemical structure and intermolecular interactions, Fourier transform infrared (FTIR) spectroscopy was performed (Figure 1d). The AD hydrogel displayed characteristic bands corresponding to −OH/N─H stretching (3321 cm^{− 1}), amide I (C═O stretching, 1651 cm^{− 1}), amide II (N─H bending, 1543 cm^{− 1}), and −SO₃ ⁻ S═O stretching (1039 cm^{− 1}), confirming successful UV‐initiated copolymerization of AMPS‐Na and DPMA. Upon luteolin incorporation, a new aromatic skeletal vibration emerged at 1606 cm^{− 1}, accompanied by marked broadening and attenuation of the −OH band, indicative of extensive hydrogen bonding between DPMA catechol groups and the luteolin aromatic framework. In Zn‐containing hydrogels, further red‐shifting and broadening of the −OH band were observed, consistent with catechol–Zn^{2 +} coordination; these spectral features were most pronounced in the presence of both luteolin and Zn^{2 +}. Collectively, these results support the coexistence of hydrogen bonding and metal–polyphenol coordination, establishing a secondary, dynamically crosslinked network that underlies the enhanced mechanical integrity and tissue adhesion of the hydrogel.

FIGURE 1.

Structural, mechanical, swelling/ release and TE characterization of AD, AD@Lut, ADZ and ADZ@Lut hydrogels. (a) Representative SEM images of the four formulations. (b) EDS elemental maps. (c) Photographic demonstration of the conformal adhesion and mechanical adaptability of ADZ@Lut on skin and at jointed sites, as well as under bending, twisting, compression, and stretching. (d) FTIR spectra of the four hydrogels. (e) Frequency‐sweep rheology showing solid‐like viscoelastic behavior (G′ ≫ G″) across formulations. (f) Tensile stress–strain curves demonstrating enhanced strength and toughness after Zn^{2 +} incorporation and dual‐network formation. (g) Swelling curves of four groups of hydrogels. (h) Cumulative release profiles of Zn^{2 +} and luteolin over 21 days. (i–l) TE parameters: (i) power factor, (j) Seebeck coefficient, (k) thermal conductivity, and (l) ionic conductivity of four groups of hydrogels. (m) Open‐circuit voltage (Voc) output of ADZ@Lut under stepwise temperature gradients (ΔT = 3, 5, and 10 K). (n) On‐skin demonstration and infrared thermography (left, photographic image; right, thermal image) showing a typical skin–air temperature gradient (∼10 K) at the dressing site. (o) In situ voltage measurement on skin after 10 min contact, demonstrating ∼90 mV output from ADZ@Lut under realistic conditions. Sample size n = 3 for all quantitative experiments. Data are presented as mean ± SD. Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test. P < 0.05 was considered statistically significant; ns, not significant.

Given the intended cutaneous application, adhesion to skin and jointed regions as well as mechanical compliance under bending, twisting, compression, and tension were assessed (Figure 1c). The hydrogel conformed readily to dynamic surface deformations, indicating suitability for on‐skin use during routine body movements. Robust adhesion across diverse tissue substrates was further confirmed (Figure S1). Under oscillatory shear (0.1–10 rad s^{− 1}), all formulations exhibited viscoelastic solid–like behavior with storage moduli (G′) exceeding loss moduli (G″) by a wide margin (Figure 1e). In uniaxial tensile testing, incorporation of ZnSO₄ markedly increased both ultimate strain and ultimate stress (Figure 1f), consistent with reinforcement by the metal–phenolic secondary network. Young's modulus was likewise elevated in ADZ and ADZ@Lut (Figure S2a).

Beyond adequate mechanical integrity, hydrogels intended for wound management must exhibit well‐regulated swelling behavior and sustained moisture retention. The AD control displays pronounced swelling yet poor water retention, attributable to its looser polymer network and inferior mechanical robustness (Figure 1g; Figure S2c,d). In contrast, the ADZ@Lut formulation, reinforced by multiple luteolin–Zn^{2 +}–mediated dynamic crosslinks, achieves moderate and well‐controlled swelling together with excellent water‐retention capacity. This favorable fluid‐management profile enables efficient sequestration of wound exudate while maintaining a moist healing microenvironment, which is essential for tissue repair and for the long‐term stability of TE sensing and signaling during on‐skin operation [35, 36].

To evaluate sustained delivery, cumulative release was monitored over three weeks (Figure 1h). A phase‐complementary pattern was observed: Zn^{2 +} shows an initial burst followed by a plateau, meeting early‐phase antibacterial/antibiofilm needs; luteolin releases more gradually and continuously, providing prolonged antioxidant cues. The observed phase‐complementary release is passively governed by the distinct molecular states, interaction strengths, and diffusivities of the two cargos within the dual‐network hydrogel. Zn^{2 +}, as a small hydrated ion, is only partially involved in catechol–metal coordination, while a weaker bound fraction remains mobile within the hydrated network and can therefore diffuse out more rapidly at the early stage [37, 38]. In contrast, luteolin has lower aqueous solubility and is likely further retained in the matrix by additional intermolecular interactions, including hydrogen bonding and aromatic interactions with the catechol‐containing network, thereby resulting in a slower and more sustained release profile [39].

The ADZ@Lut hydrogel functions as an i‐TE material. When a temperature gradient is established across the hydrogel—for instance, between the skin (∼32°C–34°C) and the ambient air (∼22°C–25°C)—cations and anions undergo selective thermodiffusion due to their distinct Soret coefficients [40, 41]. This process establishes an ionic concentration gradient and an intrinsic electric field, yielding a measurable thermovoltage in an open‐circuit configuration. The voltage output approximately follows the relationship ΔV = αiΔT, where αi represents the ionic Seebeck coefficient (mV·K^{− 1}).

Notably, despite pronounced differences in microstructure and macroscopic mechanical properties among the four hydrogel formulations, their thermal conductivity and ionic conductivity remained within the same order of magnitude, with Seebeck coefficients consistently in the range of ∼8–10 mV K^{− 1} (Figure 1i–l; Figure S2b). This convergence is consistent with the design principles of an AMPS‐Na–based i‐TE network: thermal transport is dominated by the aqueous phase, whereas ion migration is primarily dictated by the fixed charge density arising from the constant AMPS‐Na content. Modulation of secondary crosslinks through hydrogen bonding and metal–polyphenol coordination may locally alter network tortuosity and microviscosity, but these effects do not override the matrix‐governed transport processes. Collectively, these findings indicate that ADZ@Lut preserves the fundamental TE framework required for stable electric‐field generation while, by design, endowing the hydrogel with enhanced toughness, strong wet adhesion, and bioactive antibacterial and anti‐inflammatory functionality.

EEF at epithelial wounds typically range from tens to hundreds of mV mm^{− 1} and function as essential guidance cues for directed cell migration [4, 8, 42]. To assess practical performance under variable thermal conditions relevant to real‐world use, we measured the Voc of ADZ@Lut in response to prescribed temperature differentials (ΔT = 3, 5, and 10 K) using a temperature‐controlled stage coupled with real‐time voltage acquisition (Figure S3). The device generated voltages comparable in magnitude to physiological EEF across commonly encountered temperature gradients (Figure 1m). The cycling stability of the thermoelectric output was further examined by recording the voltage response for 40 min under repeated on/off thermal stimulation (2 min on an isothermal stage and 2 min at room temperature, ΔT = 10 K). The voltage response remained highly reproducible over multiple cycles, with peak values fluctuating within a stable range and showing no obvious decay (Figure S4). Moreover, when applied directly to human skin, an ADZ@Lut patch produced an output of 90.3 mV under a representative skin–air temperature difference of approximately 10 K, substantiating the translatability of the bench‐top measurements to an on‐body operating context (Figure 1n, o).

Together, these physicochemical characterizations establish ADZ@Lut as a robust, conformal, and functionally integrated hydrogel that combines mechanical resilience, controlled hydration dynamics, bioactive cargo release, and stable i‐TE output. This unified materials platform provides the structural and electrical prerequisites for subsequent antibacterial, immunomodulatory, and pro‐regenerative functions demonstrated in vitro and in vivo.

2.2. Broad‐Spectrum Antibacterial and Antibiofilm Performance of ADZ@Lut

Effective infection control is a prerequisite for the successful treatment of DFU [43]. We assessed the bactericidal activity of four hydrogel formulations against two clinically prevalent pathogens—methicillin‐resistant Staphylococcus aureus (MRSA) and Escherichia coli (E. coli)—using agar plate–based inhibition assays (Figure 2a–c). As expected, the blank control and matrix‐only hydrogel (AD) exhibited negligible bactericidal activity. In contrast, the AD@Lut and the ADZ each showed substantial antibacterial effects, whereas the dual‐loaded formulation (ADZ@Lut) achieved the most pronounced inhibition against both Gram‐positive and Gram‐negative bacteria. Quantitative analysis revealed that inhibition of MRSA reached 87.87% for AD@Lut, 82.58% for ADZ, and 97.57% for ADZ@Lut (Figure 2b). Against E. coli, the corresponding inhibition rates were 89.11%, 78.43%, and 96.82%, respectively (Figure 2c). The superior efficacy of the dual‐loaded hydrogel relative to the single‐agent formulations is consistent with emerging antimicrobial strategies in which orthogonal bactericidal mechanisms act synergistically to enhance bacterial clearance, suppress regrowth, and reduce the risk of dose‐dependent cytotoxicity [44].

FIGURE 2.

Antibacterial and antibiofilm activities of AD, AD@Lut, ADZ, and ADZ@Lut against MRSA and E. coli. (a) Representative colony‐forming unit (CFU) agar plates after 24 h of co‐incubation with each formulation. (b,c) Quantification of bacterial survival for MRSA (b) and E. coli (c). (d) Biofilm formation and disruption evaluated by crystal violet staining (top) and confocal laser scanning microscopy (CLSM) live/dead imaging (bottom); numbers adjacent to the crystal violet wells indicate residual biofilm coverage relative to the total well area. (e,f) Depth‐resolved fluorescence intensity profiles of MRSA (e) and E. coli (f) biofilms obtained by CLSM, quantifying the distribution of viable bacteria across biofilm thickness. (g) Representative transmission electron microscopy (TEM) images of bacteria after treatment. (h) Schematic illustration of the antibacterial and antibiofilm mechanisms of ADZ@Lut. For all quantitative measurements, n = 6 independent experiments; data are presented as mean ± SD. Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test. *p < 0.05, ***p < 0.001, and ****p < 0.0001.

In view of the fact that biofilms constitute a major barrier to infection control in chronic diabetic wounds—by shielding bacteria from both systemic antibiotics and topical therapies—we next examined biofilm inhibition and eradication (Figure 2d–f) [45]. In crystal violet assays, the cargo‐containing groups (AD@Lut, ADZ, and ADZ@Lut) markedly suppressed biofilm formation, with the dual‐loaded formulation exhibiting the greatest reduction in biomass (Figure 2d). Consistently, live/dead confocal imaging of established biofilms revealed extensive bacterial killing after treatment with these hydrogels. Depth‐resolved z‐stacks further demonstrated substantially diminished viable signals across the entire biofilm thickness in the ADZ@Lut group for both MRSA and E. coli (Figure 2e,f), corroborating the crystal violet results. Supplementary experiments further showed that ADZ@Lut retained strong antibacterial and antibiofilm activity against drug‐resistant E. coli (Figure S6).

To obtain morphological evidence of antibacterial action, we conducted SEM and TEM analyses (Figure S5; Figure 2g). SEM micrographs showed intact cocci/rods with smooth surfaces in the blank and AD groups, whereas AD@Lut, ADZ, and particularly ADZ@Lut induced characteristic envelope damage, including surface corrugation, invagination, and partial collapse, accompanied by markedly reduced bacterial surface coverage. TEM further corroborated disruptions of the cell wall/membrane, submembranous vacuolization, and cytoplasmic leakage, with the most severe ultrastructural deterioration observed in the ADZ@Lut group. These morphological signatures are consistent with those reported for effective antibiofilm materials and antimicrobial interfaces, in which membrane disruption leads to osmotic imbalance and eventual lysis [46].

To clarify whether electrical stimulation contributed directly to the antibacterial and antibiofilm effects, supplementary experiments were performed under external thermoelectric stimulation (Figure S7). In both the colony‐counting assay and the mature‐biofilm live/dead assay, Blank + ES did not show an obvious difference from Blank, and ADZ@Lut + ES did not exhibit a significant additional improvement compared with ADZ@Lut alone (Figure S8). Therefore, under the present experimental conditions, the antibacterial and antibiofilm activities were mainly attributed to the released Zn^{2 +} and luteolin rather than to a directly measurable antibacterial effect of electrical stimulation.

A schematic overview (Figure 2h) summarizes the individual and synergistic antibacterial actions of luteolin and Zn^{2 +} against planktonic bacteria and biofilms. Prior untargeted metabolomics has implicated luteolin in perturbing flavin‐dependent energy metabolism, bacterial chemotaxis, and glycerophospholipid pathways—effects consistent with membrane destabilization, bioenergetic collapse, and progressive structural disintegration [47]. By contrast, Zn^{2 +} destabilizes bacterial envelopes and engages specific regulatory targets, thereby amplifying envelope stress responses and ultimately compromising cellular viability [48]. ADZ@Lut integrates complementary chemical and ionic actions, including polyphenol driven antibiofilm activity, virulence attenuation, Zn^{2 +} induced envelope damage, and signaling stress, thereby providing a mechanistic basis for the enhanced bactericidal efficacy and deeper biofilm eradication observed in this study.

2.3. Immunomodulation of Macrophage Polarization by ADZ@Lut

Beyond persistent infection, a defining pathological feature of diabetic wounds is chronic, unresolved inflammation, which prevents the transition from the inflammatory to the reparative phase of healing [49, 50]. Macrophages play a central role in this process, as the balance between proinflammatory M1 and proresolving M2 phenotypes critically governs inflammation resolution and tissue remodeling [51, 52]. We therefore investigated the capacity of ADZ@Lut to modulate macrophage polarization using immunofluorescence analyses. Lipopolysaccharide (LPS) stimulation induced preferential polarization toward an M1‐like phenotype [53]. Experimental groups included an untreated control, an LPS‐treated positive control, and co‐culture groups in which LPS‐treated cells were exposed to extracts from hydrogels differing in composition. Live/dead staining further confirmed that the hydrogel extracts did not induce obvious cytotoxicity in RAW264.7 macrophages under the same treatment conditions used for the polarization experiments (Figure S9a).

The expression of the M1 marker inducible nitric oxide synthase (iNOS) and the M2 marker arginase‐1 (Arg‐1) was quantified (Figure 3a, c). Hydrogels containing either luteolin or Zn^{2 +} partially attenuated M1 polarization while promoting M2 differentiation, whereas the dual‐loaded ADZ@Lut produced the most pronounced skewing toward an M2‐like phenotype. Relative to the LPS control, ADZ@Lut reduced the M1 population by 54.19% and increased the M2 population by 1.74‐fold (Figure 3e, f). Flow cytometric analysis was consistent with the immunofluorescence results. Representative gating and distribution plots are shown in Figure 3b,d. Quantitative analysis revealed marked reductions in the CD86⁺ (M1‐like) subset accompanied by concomitant increases in the CD206⁺ (M2‐like) subset following ADZ@Lut treatment across independent replicates (Figure 3g,h). The gating strategy for identification of M1 and M2 macrophages is illustrated in Figure S10a,b.

FIGURE 3.

ADZ@Lut regulates macrophage polarization and attenuates intracellular oxidative stress. (a) Representative immunofluorescence images of iNOS in RAW264.7 macrophages under the indicated treatments: untreated control (Control −), LPS‐stimulated positive control (Control +), AD, AD@Lut, ADZ, and ADZ@Lut. F‐actin is stained in red and nuclei are counterstained with 4′,6‐diamidino‐2‐phenylindole (DAPI, blue). Bottom, 3D reconstructed fluorescence images. (b) Flow cytometric analysis of M1 macrophages (CD86⁺F4/80⁺). (c) Immunofluorescence images of Arg‐1 and corresponding 3D fluorescence intensity reconstructions. (d) Flow cytometric analysis of M2 macrophages (CD206⁺F4/80⁺). (e,f) Quantification of iNOS (e) and Arg‐1 (f) fluorescence intensities. (g,h) Statistical analysis of the proportions of M1 (CD86⁺, g) and M2 (CD206⁺, h) macrophages among total F4/80⁺ cells. (i) Representative images of intracellular ROS levels in RAW264.7 cells assessed by dihydroethidium (DHE) staining. (j) Schematic illustration of the immunomodulatory and antioxidative mechanisms of ADZ@Lut via the concerted release of luteolin and Zn^{2 +}. Data are presented as mean ± SD (n = 7). Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test. ns, not significant; **p < 0.01, ***p < 0.001, ****p < 0.0001.

To further verify the immunomodulatory effect of ADZ@Lut at the transcriptional level, quantitative real‐time polymerase chain reaction (qRT‐PCR) was performed in RAW264.7 macrophages (Figure S11). ADZ@Lut significantly downregulated the pro‐inflammatory genes tumor necrosis factor (Tnf) and iNOS, as well as the inflammasome‐related gene NLRP3. In contrast, ADZ@Lut markedly increased the expression of the anti‐inflammatory genes Arg‐1 and interleukin‐10 (Il‐10). Notably, tumor necrosis factor alpha‐induced protein 3 (Tnfaip3), which encodes the nuclear factor kappa B (NF‐κB) negative‐feedback regulator A20, was further upregulated after ADZ@Lut treatment, suggesting enhanced termination of excessive inflammatory signaling. Mechanistically, luteolin functions as a polyphenolic immunomodulator that suppresses NF‐κB, mitogen‐activated protein kinase (MAPK), and activator protein 1 signaling, attenuates oxidative stress and NLRP3 inflammasome activation, and consequently biases macrophages toward an M2‐like phenotype [54, 55, 56]. In contrast, Zn^{2 +} acts as an ionic cue that engages A20/Tnfaip3‐mediated negative feedback on NF‐κB signaling, modulates lysosomal zinc homeostasis and inflammasome activity, and directly promotes M2 polarization at the material–tissue interface [57, 58]. Together, these two components operate in a complementary manner—chemical restraint of proinflammatory transcription (luteolin) coupled with ionic signal–driven immunoreprogramming (Zn^{2 +})—providing a mechanistic explanation for the superior anti‐inflammatory and pro‐regenerative outcomes observed with the dual‐loaded hydrogel.

To further verify that the observed immunomodulatory effect was not restricted to the immortalized RAW264.7 cell line, we performed supplementary validation in primary bone marrow‐derived macrophages (BMDMs). Importantly, under LPS stimulation, both immunofluorescence staining and flow cytometry analysis in BMDMs reproduced the same overall trend observed in RAW264.7 cells, with reduced pro‐inflammatory related signals and increased anti‐inflammatory related signals after ADZ@Lut treatment (Figure S9b–i). These findings further support that the anti‐inflammatory and immunomodulatory effect of ADZ@Lut is not limited to RAW264.7 cells and can also be reproduced in primary macrophages.

To substantiate the antioxidant contribution of luteolin, we performed 2,2‐diphenyl‐1‐picrylhydrazyl (DPPH) and 2,2′‐azino‐bis(3‐ethylbenzothiazoline‐6‐sulfonic acid) (ABTS) radical‐scavenging assays (Figure S12) in parallel with intracellular ROS measurements using DHE staining (Figure 3i; Figure S13). In both RAW264.7 macrophages and L929 fibroblasts, hydrogels containing luteolin significantly reduced DHE fluorescence intensity relative to controls, indicating effective ROS neutralization and supporting the shift toward an M2‐dominant, pro‐healing immune milieu.

2.4. Proliferation‐Phase Support via Self‐Powered Bioelectric Cues

Upon effective infection control and resolution of acute inflammation, diabetic wounds transition into the proliferative phase [59]. In physiological settings, the endogenous transepithelial electric field promotes cell proliferation and directed migration [13]. Fibroblasts serve as the primary effector cells at this stage; ES induces front–rear polarization, cytoskeletal remodeling, and focal adhesion reorganization, thereby improving collagen fiber alignment and facilitating wound‐edge contraction [60]. Endothelial cells orchestrate angiogenesis through chemokine‐ and ES‐guided migration, forming perfusable microvascular networks that re‐establish tissue perfusion and oxygen/nutrient delivery [61]. Notably, accumulating evidence indicates that endogenous bioelectric signaling is compromised in diabetic wounds as a consequence of epithelial ion‐transport dysfunction and impaired Na⁺/K⁺‐ATPase–driven transepithelial potentials [3, 42, 62, 63].

Leveraging the intrinsic temperature gradient at the wound–air interface, the TE hydrogel (ADZ@Lut) converts low‐grade, noninvasive thermal energy into a stable microcurrent, thereby reconstructing a wound‐relevant electric field without the need for an external power supply. In this set of experiments, four groups were included: blank control, AD, ADZ@Lut, and ADZ@Lut (+). The ADZ@Lut (+) group denotes cells cultured with ADZ@Lut extracts and additionally subjected to TE stimulation using the setup shown in Figure S7 (ΔT = 10 K). Cytocompatibility and pro‐proliferative effects were evaluated by live/dead staining and Cell Counting Kit‐8 (CCK‐8) assays (Figure 4a–c). Neither the material itself nor the mild ES induced detectable cytotoxicity; after 24 to 72 h, green fluorescence from viable cells in the ADZ@Lut (+) group was significantly greater than that in the blank and hydrogel‐only controls (Figure 4a; Figure S12a–c). Consistent with the imaging results, CCK‐8 analysis showed that on days 3 and 5, cell numbers of L929 fibroblasts and human umbilical vein endothelial cells (HUVECs) in the ES group reached approximately 1.5‐fold those of the blank and hydrogel‐only groups (Figure 4b, c).

FIGURE 4.

TE stimulation promotes cell viability, migration, and angiogenesis in vitro. (a) Representative live/dead staining images of L929 fibroblasts and HUVECs after exposure to the indicated treatments. Live cells are shown in green and dead cells in red. (b,c) CCK‐8 assays showing the proliferation of L929 cells (b) and HUVECs (c) at indicated time points. (d,e) Transwell migration assay of L929 fibroblasts: (d) Representative images of migrated cells after 12 h and (e) Quantification of migrated cells per field. (f,g) Transwell migration assay of HUVECs: (f) Representative images after 12 h and (g) corresponding quantification. (h–k) Scratch wound‐healing assays of L929 fibroblasts (h,j) and HUVECs (i,k). Representative images at 0 and 24 h are shown in (h) and (i); quantitative analyses of migration rates are shown in (j) and (k). (i–n) In vitro angiogenesis assessed by Matrigel tube formation assay: (i) representative images of capillary‐like network formation, (m) total tube length, and (n) number of junctions. (o) Schematic illustrating that the TE field generated by ADZ@Lut under a physiological temperature gradient promotes cell proliferation, migration, and angiogenesis. Data are presented as mean ± SD (n = 5). Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test. ns, not significant; **p < 0.01, ***p < 0.001, ****p < 0.0001.

We next evaluated ES‐enhanced cell migration. In Transwell migration assays (12 h), the numbers of migrated L929 fibroblasts and HUVECs in the ADZ@Lut (+) group increased to 5.3‐fold and 2.4‐fold of the control group, respectively (Figure 4d–g). In scratch assays (24 h), wound‐closure rates for L929 cells and HUVECs in the ADZ@Lut (+) group were accelerated by 3.1‐fold and 1.5‐fold versus control (Figure 4h–k), indicating that the natural temperature difference at the skin surface is sufficient for ADZ@Lut to generate an electric field that promotes directional migration. In addition, Matrigel tube‐formation assays showed significant increases in total tube length and junction number in the ES group, supporting an ES‐enhanced angiogenic capacity in vitro (Figure 4i–n).

To further verify whether this pro‐regenerative effect could also be reproduced in rat fibroblasts, we performed additional scratch and CCK‐8 assays. Consistent with the results observed in L929 cells, the ADZ@Lut (+) group markedly promoted rat fibroblast migration and proliferation compared with the Blank, AD, and ADZ@Lut groups (Figure S15).

Collectively, these results demonstrate that ADZ@Lut restores a wound‐relevant bioelectric cue and promotes key processes of the proliferative phase. These include fibroblast‐mediated wound contraction and endothelial cell–mediated formation of functional vascular networks that support nutrient and oxygen delivery. These processes are commonly impaired in diabetic wounds (Figure 4o).

2.5. RNA Sequencing Analysis

To elucidate the transcriptional programs underlying the pro‐migratory and pro‐proliferative effects of TE stimulation, we performed mRNA sequencing (RNA‐seq) on fibroblasts harvested on day 3 from the ADZ@Lut (+) and ADZ@Lut groups. Principal component analysis (PCA) revealed a pronounced separation between the two conditions, indicating substantial global transcriptional remodeling (Figure 5a). Consistently, volcano plots and heat maps highlighted widespread differential gene expression, with 1,640 genes up‐regulated and 764 genes down‐regulated in the ADZ@Lut (+) group relative to ADZ@Lut (Figure 5b,d).

FIGURE 5.

Transcriptomic profiling implicates Ca^{2 +}‐dependent PI3K–Akt signaling in TE stimulation–enhanced fibroblast migration and proliferation. (a) 3D PCA plot showing global transcriptional separation between the ADZ@Lut (+) and ADZ@Lut groups. (b) Volcano plot of differentially expressed genes (DEGs) identified by RNA sequencing, with genes up‐regulated (red) and down‐regulated (blue) in ADZ@Lut (+) relative to ADZ@Lut. (c) Chord diagram illustrating representative up‐regulated genes and their associated Gene Ontology (GO) terms. (d) Heat map of DEGs comparing ADZ@Lut (+) and ADZ@Lut. (e) GO molecular function enrichment highlighting overrepresentation of extracellular matrix (ECM) binding, structural constituent of ECM, cell adhesion molecule binding, and calmodulin binding. (f) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showing activation of PI3K–Akt signaling, focal adhesion, ECM–receptor interaction, and cytoskeleton‐associated pathways. (g‐i) Gene set enrichment analysis (GSEA) of the calcium signaling pathway (g), PI3K–Akt signaling pathway (h), and Ras signaling pathway (i), demonstrating significant enrichment in ADZ@Lut (+). RNA‐seq was performed using biological triplicates (n = 3 per group). Statistical significance was defined by Benjamini–Hochberg adjusted p values with |log₂ fold change| ≥ 1 and padj < 0.05.

A chord diagram (Figure 5c) summarizes representative up‐regulated genes and their associated GO terms. GO enrichment analysis revealed strong overrepresentation of ECM binding and cell adhesion molecule binding (Molecular Function), wound healing, cell–substrate adhesion, and ECM organization (Biological Process), as well as cell–substrate junction and collagen‐containing ECM (Cellular Component) (Figure 5e; Figure S16). Collectively, these signatures indicate reinforced fibroblast–matrix interactions together with active ECM synthesis and remodeling, thereby providing the structural scaffold and contractile forces required for effective wound closure. These findings are consistent with and strongly support the observed enhancement of fibroblast migration and proliferation under TE stimulation. Notably, enrichment of the Molecular Function term calmodulin binding was observed in the electrically stimulated group, accompanied by prominent changes in transcripts associated with Ca^{2 +} influx and Ca^{2 +}‐dependent signaling, including STIM2, ORAI family members, and CAMK2A/B. These patterns suggest that Ca^{2 +} entry represents a critical upstream signal through which TE stimulation promotes fibroblast proliferation and migration.

KEGG pathway analysis demonstrated significant enrichment of the PI3K–Akt signaling pathway, together with downstream effector modules commonly regulated by PI3K/Erk signaling—including focal adhesion, ECM–receptor interaction, and regulation of the actin cytoskeleton (Figure 5f) in agreement with the observed migration‐ and proliferation‐associated phenotypes.

To capture pathway‐level trends beyond gene‐wise differential expression, we further performed GSEA. In the ADZ@Lut (+) group, the calcium signaling, PI3K–Akt, Ras, and phosphatidylinositol signaling pathways were significantly enriched, accompanied by coordinated activation of adhesion‐ and cytoskeleton‐related programs (focal adhesion, ECM–receptor interaction, regulation of the actin cytoskeleton, and adherens junction) (Figure 5g–i; Figure S17). Collectively, these orthogonal analyses converge on a unified mechanistic model in which TE stimulation induces Ca^{2 +} entry, initiates calmodulin‐dependent signaling, and activates the PI3K–Akt axis, thereby reinforcing adhesion, extracellular matrix organization, and cytoskeletal remodeling to drive fibroblast migration and proliferation and ultimately accelerate wound repair.

2.6. Validation of the Intracellular Downstream Mechanism

To experimentally interrogate the pathways implicated by RNA sequencing, we performed a series of targeted perturbation experiments (Figure 6). Live‐cell imaging with Fluo‐4 AM revealed that ES markedly elevated intracellular Ca^{2 +} levels in fibroblasts (Figure 6a,b). This rapid Ca^{2 +} influx is concordant with the RNA‐seq–identified enrichment of calcium signaling and calmodulin‐binding gene sets, supporting a model in which Ca^{2 +} functions as a primary upstream trigger of ES‐evoked responses. Given that phosphorylated Akt and Erk are key downstream signaling nodes controlling cell proliferation and migration, we next assessed their activation by Western blotting [64]. ES robustly increased phosphorylation of Akt and Erk (Figure 6c–e). Together with the KEGG and GSEA results, which showed that the PI3K–Akt pathway and downstream adhesion‐ and cytoskeleton‐related programs were among the most significantly enriched pathways, these data indicate that ES activates pro‐survival and pro‐motility signaling at the protein level.

FIGURE 6.

TE stimulation activates Ca^{2 +}–calmodulin–PI3K–Akt/Erk signaling to promote fibroblast proliferation and migration. (a) Representative fluorescence images of intracellular Ca^{2 +} in fibroblasts stained with Fluo‐4 under blank, ADZ@Lut, and ADZ@Lut (+) conditions; nuclei are counterstained with Hoechst. (b) Quantification of Fluo‐4 fluorescence intensity. (c) Western blot analysis of phosphorylated and total Akt and Erk under the indicated treatments. (d,e) Quantification of phosphorylation of Akt (p‐Akt)/Akt and phosphorylation of Erk (p‐Erk)/Erk ratios. (f,g) Representative fluorescence images of intracellular Ca^{2 +} (f) and corresponding quantification (g) following treatment with the Ca^{2 +} chelator BAPTA‐AM. (h) Western blot showing inhibition of Akt and Erk phosphorylation upon Ca^{2 +} chelation. (i,j) Quantification of p‐Akt/Akt and p‐Erk/Erk ratios following Ca^{2 +} chelation. (k,l) Scratch‐wound assays and quantitative analysis of fibroblast migration after Ca^{2 +} blockade. (m–o) Western blot analysis and quantification of p‐Akt/Akt and p‐Erk/Erk ratios in the presence of the calmodulin inhibitor W‐13. (p–r) Western blot analysis and quantification of p‐Akt/Akt and p‐Erk/Erk ratios after treatment with the PI3K inhibitor LY294002. (s) Schematic illustration of the proposed Ca^{2 +}‐dependent signaling cascade triggered by TE stimulation, in which Ca^{2 +} influx activates calmodulin, leading to PI3K activation and subsequent phosphorylation of Akt and Erk, thereby promoting proliferation and migration. Data are presented as means ± SD (n = 4). Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test. ns, not significant; **p < 0.01, ***p < 0.001, ****p < 0.0001.

To determine whether Ca^{2 +} entry is required for downstream kinase activation, intracellular Ca^{2 +} was chelated with BAPTA‐AM [65]. BAPTA‐AM markedly suppressed the ES‐induced rise in cytosolic Ca^{2 +} (Figure 6f,g) and, critically, eliminated ES‐triggered phosphorylation of Akt and Erk (Figure 6h‐j). Functionally, Ca^{2 +} chelation abrogated the ES‐mediated acceleration of wound closure in scratch assays (Figure 6k, l), indicating that Ca^{2 +} influx is indispensable for both kinase activation and the associated pro‐migratory phenotype in fibroblasts.

Given the pronounced enrichment of calmodulin‐associated terms in GO Molecular Function analysis, we next asked whether Ca^{2 +} signals are transduced through calmodulin to activate downstream kinases. Pharmacological inhibition of calmodulin with W‐13 substantially attenuated ES‐induced phosphorylation of Akt and Erk (Figure 6m–o) [66], consistent with a Ca^{2 +}‐dependent, calmodulin‐mediated relay linking ion influx to kinase activation. These findings provide direct molecular evidence supporting the transcriptomics‐based prediction that calmodulin couples Ca^{2 +} signaling to PI3K–Akt pathway activation.

To further resolve pathway hierarchy, we pharmacologically inhibited PI3K with LY294002 [67]. Despite ongoing electrical stimulation, PI3K blockade markedly reduced phosphorylation of both Akt and Erk (Figure 6p–r). Given the pronounced enrichment of the PI3K–Akt pathway in our KEGG and GSEA analyses, the decrease in p‐Akt upon PI3K inhibition was anticipated. Notably, p‐Erk was also substantially diminished, indicating that, under these conditions, PI3K activity is required for Erk phosphorylation, rather than Erk being engaged primarily through the canonical MAPK cascade—consistent with the absence of MAPK–Erk enrichment in our transcriptomic profiles. This interpretation accords with recent reports demonstrating strong PI3K dependence of Erk phosphorylation, including RAF‐independent Erk activation in lymphocytes, PI3K–Ras crosstalk that potentiates Erk signaling during tissue morphogenesis, and PI3Kα‐dependent GAB1–Erk activation in carcinoma cells [68, 69, 70].

Integrating these multi‐level readouts, we propose the mechanistic schema depicted in Figure 6s: TE stimulation → Ca^{2 +} influx → calmodulin‐dependent PI3K activation → Akt and Erk phosphorylation → enhanced fibroblast migration and wound‐relevant reparative behaviors.

2.7. In Vivo Wound Repair in Diabetic Rats Under TE Hydrogel Treatment

To determine whether ADZ@Lut recapitulates in vivo the dual functionalities observed in vitro, including bioactive cargo delivery and on‐wound thermoelectric stimulation, we established an infected full‐thickness excisional wound model on the dorsum of diabetic rats (Figure 7a). Control groups included an untreated cohort and a group dressed with a commercial Tegaderm film. The experimental arms comprised (i) the matrix‐only hydrogel (AD), (ii) the hydrogel extract derived from ADZ@Lut (Lut–Zn^{2 +}), and (iii) the intact ADZ@Lut hydrogel. To ensure positional stability during unrestricted animal movement, all hydrogel dressings were secured with a gauze wrap despite their intrinsic wet adhesion (placement photographs are shown in Figure S18). A schematic illustration in Figure 7b depicts the temperature gradient established across the upper and lower surfaces of ADZ@Lut, which drives a continuous TE microcurrent at the wound interface.

FIGURE 7.

ADZ@Lut accelerates healing of infected full‐thickness wounds in diabetic rats. (a) Experimental scheme illustrating the establishment of a streptozotocin‐induced diabetic rat model with full‐thickness infected wounds and the treatment timeline, including hydrogel application, serial wound imaging, and histological analyses. (b) Schematic illustration of the temperature gradient across ADZ@Lut on the rat dorsum, created by the skin–air interface, which drives ionic thermodiffusion and generates an in situ bioelectric field. (c) Representative gross images of wound closure at days 0, 3, 6, 10, and 14 across treatment groups (Blank, Tegaderm, AD, Lut–Zn^{2 +} extract, and ADZ@Lut), with corresponding wound area maps. (d) Heat map summarizing the normalized wound areas over time. (e) Quantification of wound closure rates during the 14‐day treatment period. (f) Hematoxylin and eosin (H&E)–stained sections at day 3 showing inflammatory infiltration, with green arrows indicating dense neutrophil accumulation. (g) Representative bacterial colony plates from wound tissues at day 3. (h) H&E staining at day 14 assessing re‐epithelialization and tissue regeneration, with yellow arrows denoting regenerated epidermis. (i) Masson's trichrome staining at day 14 evaluating collagen deposition and extracellular matrix remodeling, with red arrows indicating regenerated skin appendages. Data are presented as means ± SD (n = 6). Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test.

Serial planimetric analysis at days 0, 3, 6, 10, and 14 revealed a clear hierarchy of therapeutic efficacy (Figure 7c–e): ADZ@Lut achieved the most rapid reduction in wound area, followed by AD and Lut–Zn^{2 +}, whereas Tegaderm offered only marginal benefit over the untreated control. Heat map visualization (Figure 7d) and closure‐rate plots (Figure 7e) corroborated the accelerated healing trajectory in the ADZ@Lut group. By day 6, wound length in the ADZ@Lut cohort had decreased to 41% of that in the untreated group (Figure S19a). These results indicate a dual contribution from material composition and the dressing‐generated microcurrent: AD provides a baseline structural benefit, the Lut–Zn^{2 +} extract confers a soluble‐factor effect, and the intact ADZ@Lut integrates antibacterial and anti‐inflammatory chemistry with in situ bioelectric guidance to maximize wound closure.

At day 3, H&E staining revealed pronounced neutrophil infiltration in the untreated, Tegaderm, and AD groups (green arrows, Figure 7f), whereas the Lut–Zn^{2 +} and ADZ@Lut groups exhibited visibly attenuated inflammatory cell density. Consistently, quantitative plating assays demonstrated a substantial reduction in recoverable bacteria in the Lut–Zn^{2 +} and ADZ@Lut groups relative to the other treatments (Figure 7g; colony counts are summarized in Figure S19c). These findings indicate that the chemical payloads (Zn^{2 +} and luteolin) effectively suppress early bioburden and dampen acute inflammation, thereby restoring a microenvironment conducive to proliferative‐phase repair. By day 14, H&E staining revealed a more continuous and thinner neoepidermis in the ADZ@Lut group (yellow arrows, Figure 7h; thickness quantification in Figure S19b), in contrast to the thicker and irregular epithelial layers observed in the untreated and Tegaderm groups. Masson's trichrome staining (Figure 7i) further demonstrated denser and more highly organized collagen deposition in ADZ@Lut, with a significantly greater collagen fraction than in the other treatments (Figure S19d). Notably, cutaneous appendages (red arrows, Figure 7i) were more frequently observed and more completely regenerated in the ADZ@Lut group, indicating a shift toward functional tissue restoration rather than scar‐dominated repair.

To further validate the in vivo efficacy of ADZ@Lut against resistant Gram‐positive infection, we established an independent MRSA‐infected diabetic wound model. Compared with the Blank and AD groups, ADZ@Lut accelerated wound closure, attenuated histological inflammatory response, and significantly reduced bacterial burden in the wound tissue (Figure S20).

2.8. ADZ@Lut Coordinates Inflammation–Proliferation–Remodeling to Speed Diabetic Wound Repair

To delineate the therapeutic efficacy and mechanistic basis of ADZ@Lut in vivo, we performed immunofluorescence and immunohistochemical analyses of full‐thickness diabetic wound tissues at defined stages of healing. During the early inflammatory phase (Figure 8a), immunofluorescence imaging revealed pronounced immunomodulatory effects in the Lut–Zn^{2 +} and ADZ@Lut groups. As shown in Figures 8b, e, wounds treated with Lut–Zn^{2 +} or ADZ@Lut exhibited markedly reduced expression of iNOS and myeloperoxidase (MPO; a marker of neutrophil infiltration) relative to the control, Tegaderm, and AD groups. Concurrently, expression of Arg‐1 was significantly elevated (Figure 8c). Collectively, these findings underscore a central role for Lut–Zn^{2 +} release in mitigating the initial inflammatory burst and establishing an immunological milieu conducive to subsequent tissue regeneration. This shift from a pro‐inflammatory to a pro‐regenerative phenotype is essential to resolving the chronic inflammatory burden characteristic of diabetic wounds [71].

FIGURE 8.

ADZ@Lut orchestrates inflammation resolution, angiogenesis, and matrix remodeling in diabetic wounds. (a) Representative immunofluorescence images at day 3 (inflammatory phase) showing iNOS (red)/CD68 (green)/DAPI (blue) and Arg‐1 (red)/CD68 (green)/DAPI (blue), together with immunohistochemical staining of MPO in wound tissues from the indicated groups. (b,c) Quantification of mean fluorescence intensity (MFI) of iNOS (b) and Arg‐1 (c). (e) Quantification of MPO expression as mean optical density (MOD). (d) Representative immunofluorescence images at day 14 (proliferative phase) showing platelet endothelial cell adhesion molecule‐1 (CD31, red)/ alpha‐smooth muscle actin (α‐SMA, green)/DAPI (blue), Ki67 (red)/DAPI (blue), and vascular endothelial growth factor (VEGF, red)/DAPI (blue). (f,h) Quantification of MFI of CD31 (f) and α‐SMA (h). (i,j) Quantification of MFI of Ki67 (i) and VEGF (j). (g) Representative immunofluorescence images at day 14 during the remodeling phase showing collagen I (Col I, red)/DAPI (blue) and collagen III (Col III, red)/DAPI (blue). (k,l) Quantification of MFI of Col I (k) and Col III (l). Data are presented as mean ± SD (n = 6). Statistical significance was determined by one‐way ANOVA with Tukey's multiple comparisons test. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

As wounds entered the proliferative phase (Figure 8d), we evaluated key indicators of angiogenesis and cellular proliferation. Immunostaining for CD31 and α‐SMA revealed significantly enhanced formation of mature vasculature, as evidenced by increased CD31⁺/α‐SMA⁺ structures, in the AD and ADZ@Lut groups (Figure 8d,f,h). In addition, expression of Ki67 (a marker of proliferation) and vascular endothelial growth factor (VEGF; a central proangiogenic factor) was most pronounced in wounds treated with ADZ@Lut (Figure 8d,i,j). Collectively, these findings indicate that the ADZ@Lut hydrogel not only provides a robust scaffold for endothelial migration but also actively stimulates vigorous vascularization and cellular activity, both of which are critical for supplying nutrients and oxygen to regenerating tissue.

By day 14, we further evaluated deposition and composition of Col I and Col III. Because mature repair tissue is characterized by enrichment of mechanically robust, highly cross‐linked Col I relative to the more compliant Col III [72], we examined group‐wise shifts in collagen composition. Wounds treated with AD and ADZ@Lut displayed significantly increased Col I levels and diminished Col III deposition (Figure 8g,k,l), indicative of accelerated and higher‐quality extracellular matrix remodeling. This collagen profile predicts enhanced mechanical strength and a reduced risk of wound recurrence [73].

To validate the signaling pathways engaged by ADZ@Lut in vivo, we examined phosphorylation of Akt and Erk in wound tissues. Immunofluorescence analysis (Figure S21a‐c) revealed significantly elevated levels of p‐Akt and p‐Erk in the AD and ADZ@Lut groups. This activation is consistent with the Ca^{2 +} influx/PI3K/Akt and Erk signaling axes delineated in our in vitro experiments, substantiating mechanistic concordance across experimental scales.

Biosafety represents a primary consideration for clinical translation. Hemolysis testing (Figure S22) confirmed that erythrocyte lysis induced by all hydrogel extracts was negligible and well below the 5% threshold for biomaterial compatibility, indicating excellent hemocompatibility [74]. In addition, histological examination of major organs (heart, liver, spleen, lung, and kidney) at the end of the study showed no pathological abnormalities or tissue injury in any group (Figure S23). Overall, ADZ@Lut coordinates a stage‐coupled regenerative cascade involving suppression of inflammation and neutrophil infiltration, macrophage reprogramming, enhanced angiogenesis and proliferation, and ordered collagen remodeling, thereby overcoming major pathological barriers to diabetic wound repair.

3. Discussion and Conclusion

In this study, we present a tri‐synergistic thermoelectric–antimicrobial–anti‐inflammatory strategy to overcome three entrenched barriers to diabetic wound repair—infectious burden, chronic inflammation/oxidative stress, and attenuation of endogenous bioelectric cues. We engineered a tissue‐conformal dual‐network hydrogel with strong wet adhesion, high toughness under physiological hydration, and favorable cyto‐ and hemocompatibility, enabling effective deployment across diverse on‐skin wound environments. Functionally, this dressing delivers broad‐spectrum antibacterial and antibiofilm activity together with immunomodulatory and antioxidant effects, while simultaneously promoting fibroblast and endothelial proliferation and migration, thereby accelerating closure in an infected diabetic wound model. Mechanistically, by integrating RNA sequencing with protein‐level validation, we delineate a bioelectric transduction axis whereby i‐TE microcurrents—harvested from the passive skin–air temperature gradient—trigger a Ca^{2 +} influx → calmodulin → PI3K–Akt cascade, with PI3K‐dependent Erk activation, ultimately converging on adhesion‐ and cytoskeleton‐associated programs that govern motility and survival. Collectively, these findings establish a self‐powered, multi‐target hydrogel paradigm that couples sustained on‐wound microcurrent delivery with microenvironment remodeling, offering a practical and translational route to address the long‐standing challenge of chronically infected diabetic wounds.

A key advantage of TE dressings over piezoelectric or triboelectric systems is their ability to generate directional electric fields that emulate physiological wound fields—an essential cue for guiding cell migration [75, 76, 77]. Tan et al. reported a wireless TE hydrogel and, through combined in vitro/in vivo assays and single‐cell transcriptomics, demonstrated proangiogenic activity and accelerated healing in diabetic wounds [77]. Although conceptually aligned with our approach, their system relies on a photothermal trigger to impose a temperature gradient; consequently, electric‐field generation still depends on exogenous physical activation. Such dependence necessitates scheduled interventions and specialized hardware, increasing cost and potentially undermining patient adherence. By contrast, ADZ@Lut is genuinely self‐powered, harvesting the natural skin–air temperature differential to generate a stable field without external input, thereby markedly improving practicality and portability.

Qin et al. likewise report a “self‐powered TE hydrogel” that accelerates wound repair [78]; however, their platform is an electronic thermoelectric (e‐TE) system based on Ag₂Se in which the Seebeck potential arises from electron transport. Our platform, in contrast, is i‐TE. As endogenous bioelectric fields in tissues are ion‐transport–driven, i‐TE more faithfully recapitulates the native physiological modality [41]. A further practical distinction lies in materials economics and efficiency: conventional e‐TE materials typically exhibit Seebeck coefficients in the µV·K^{− 1} range and entail higher cost, whereas i‐TE hydrogels deliver mV·K^{− 1}‐level thermopower using low‐cost ionomers, providing larger voltages per unit ΔT at clinically relevant temperature biases [41]. Consistent with this, the Ag₂Se‐based e‐TE system displays a Seebeck coefficient approximately an order of magnitude lower than that of our i‐TE ADZ@Lut hydrogel [78]. What's more, neither of the above TE hydrogel designs explicitly addresses infection control or chronic inflammation, limiting applicability in the microbially burdened, oxidative, and immune‐dysregulated milieu of DFU [79]. ADZ@Lut was explicitly engineered for this clinical reality: luteolin and Zn^{2 +} confer antibacterial/antibiofilm, anti‐inflammatory, and antioxidant functions, while the i‐TE backbone restores a wound‐relevant electric field.

Rather than a simple additive mixture, our integration of an i‐TE matrix with a polyphenolic small molecule (luteolin) and Zn^{2 +} was conceived as a stage‐aware, complementary system tailored to the multifactorial pathology of diabetic wounds. Luteolin functions as a flavonoid immunomodulator and antibiofilm agent by attenuating NF‐κB/MAPK‐driven inflammatory programs, constraining toxin production and biofilm development, and reinforcing antioxidant defenses, features that make it particularly attractive for incorporation into advanced wound‐care hydrogels and composite dressings [80, 81]. These attributes have been extensively documented in the biomaterials literature, including double‐network or smart‐release hydrogel systems where luteolin enhances infection control and dampens inflammatory damage while supporting tissue repair [82, 83, 84]. In parallel, Zn^{2 +} provides orthogonal mechanisms: rapid bactericidal activity via membrane destabilization and metabolic interference; immunomodulation that biases macrophages toward pro‐resolving phenotypes; and anti‐inflammasome actions through metallothionein signaling and suppression of NLRP3 activation [37, 85, 86, 87, 88]. Beyond biological effects, Zn^{2 +} coordination reinforces the polymer network through catechol–metal interactions, thereby improving toughness and wet adhesion without compromising i‐TE performance [89]. The two cargos are thus synergistic yet mechanistically distinct: luteolin confers chemical suppression of microbial virulence and inflammatory transcription, whereas Zn^{2 +} introduces ionic membrane stress and inflammasome control. When deployed within an i‐TE scaffold that restores a directional bioelectric cue, they jointly establish an antibacterial and anti‐inflammatory microenvironment permissive to subsequent cellular proliferation and migration. This complementarity reduces dependence on any single agent and mitigates the risk of dose‐related toxicity.

One limitation of the present study is that our transcriptomic analysis was confined to fibroblasts; comprehensive single‐cell multi‐omics and phosphoproteomic profiling at the whole wound level would enable more precise delineation of cell type specific responses to i‐TE stimulation and resolution of pathway dynamics in vivo. In conclusion, ADZ@Lut integrates an AMPS‐Na i‐TE backbone with Zn^{2 +}/luteolin chemistry to reconstitute wound‐relevant microcurrents while concurrently controlling infection and inflammation. This integrated design accelerates repair in an infected diabetic wound model and establishes a new paradigm for deploying TE materials in the treatment of refractory chronic ulcers.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma73728‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.