Cascaded, self-decomposing messenger in MXene@Cu-MOF heterostructures for on-demand infection control and pro-angiogenic wound repair

Abstract

Infected diabetic wounds are sustained by excessive reactive oxygen species (ROS), persistent bacteria, and poor angiogenesis. We present a hydrogel-encapsulated MXene@Cu-MOF platform made by in-situ growth of copper metal-organic framework (Cu-MOF) nanocrystals on Ti₃C₂T_x MXene sheets, forming a unified 2D/3D heterostructure. MXene provides ROS modulation and efficient photothermal conversion; the Cu-MOF acts as a degradable “functional backpack” that releases a self-decomposing messenger (Cu²⁺) in response to acidic microenvironments and near-infrared (NIR) irradiation. The interface suppresses MXene restacking, ensures uniform MOF dispersion, and converts photothermal input into gated ionic output. The platform scavenged multiple radicals, lowered intracellular ROS in macrophages, and showed robust, cyclable photothermal heating with NIR-amplified Cu²⁺ release in dispersion and GelMA hydrogels. Under NIR, it achieved broad antibacterial activity against Escherichia coli and Staphylococcus aureus and promoted endothelial proliferation, migration, tube formation, and up-regulation of VEGF, eNOS, HIF-1α, and FGF2. In a bacteria-infected diabetic wound model, MXene@Cu-MOF/GelMA accelerated closure, enhanced re-epithelialization, increased collagen deposition with maturation from collagen III to I, reduced tissue oxidative stress, shifted macrophages toward a CD206-positive reparative phenotype, and increased CD31-positive microvessels. Day-14 RNA-seq revealed enrichment of antibacterial defense, tightly regulated inflammatory programs, angiogenesis, and antioxidant pathways. This interface-engineered dressing delivers cascaded, on-demand therapy in which photothermal energy actively gates therapeutic ion generation, unifying infection control and vascularized regeneration in a single platform.

Graphical abstract

1. Introduction

Diabetic wounds remain a major clinical burden characterized by persistent inflammation, recurrent or polymicrobial infection, excessive reactive oxygen species (ROS), and impaired angiogenesis [1,2]. These intertwined pathologies create a hostile microenvironment that resists standard care and single-mode interventions, leading to delayed closure, high risk of amputation, and heavy socioeconomic costs [3]. Effective therapy therefore requires temporally coordinated actions that (i) quench acute oxidative stress to blunt inflammatory cascades, (ii) eradicate planktonic bacteria to reduce reinfection, and (iii) reprogram the wound bed toward regeneration with pro-angiogenic cues—all while minimizing systemic exposure and enabling on-demand intensification when infection surges [4].

Stimuli-responsive dressings and nanoplatforms have emerged to address parts of this challenge, exploiting pH, enzymes, or light to modulate local drug delivery [[5], [6], [7]]. Yet most existing systems combine multiple components via physical mixing or loose encapsulation [8]. The performance of such assemblies is frequently compromised by issues of (i) aggregation and poor dispersion, including the restacking of 2D nanosheets and the clustering of metal-organic framework (MOF) particles. This leads to a reduction in effective surface area and diminished mass/heat transfer efficiency [9] (ii) poorly coupled modalities where one treatment (e.g., photothermal heating); does not actively gate or amplify the other (e.g., antimicrobial ion release), yielding additive rather than truly synergistic effects [10]; and (iii) limited capacity to stage actions along the wound-healing timeline. Importantly, wound pH is heterogeneous in space and time [11]. While chronic wounds can be alkaline at the macroscale, localized infection microdomains may acidify due to bacterial metabolism, providing a rational trigger for pH-responsive carriers [12,13]. As a result, current platforms frequently underperform in bacteria-infected diabetic wounds, where therapy must be both microenvironment-sensitive and readily escalated [14,15].

MXene—exfoliated transition-metal carbides/nitrides such as Ti₃C₂T_x—are promising building blocks for infectious-wound care due to their inherent ROS-scavenging capacity, broadband near-infrared (NIR) absorption enabling efficient photothermal conversion, and surface terminations (–O, –OH) that afford chemical anchoring sites [16,17]. Copper-based MOF (Cu-MOFs; e.g., HKUST-1 or Cu-doped ZIF-8) complement MXenes by providing reservoirs of bioactive Cu²⁺ ions with broad-spectrum antibacterial activity and pro-angiogenic signaling potential [18]. Notably, many Cu-MOFs are labile under mildly acidic conditions typical of infected wounds, allowing pH-triggered release of Cu²⁺ “as needed” [19]. However, simply blending MXene with Cu-MOF in hydrogels rarely overcomes the fundamental problems of component segregation and uncontrolled burst or, conversely, insufficient release [20]. Moreover, physical mixtures generally cannot convert photothermal energy into an active, programmable gate for chemical delivery; the two modalities remain neighbors rather than collaborators [21]. Previous MXene–MOF systems have largely relied on ex situ mixing or loose decoration of MOF particles on MXene sheets. In these constructs, the two components coexist without a well-defined, engineered heterointerface and mainly provide only parallel photothermal and ion-release functions [22]. In such constructs, MXene is not explicitly exploited as a nucleation scaffold to direct MOF growth, and photothermal heating typically does not serve as an active gate for controlled ion generation. As a result, these MXene-MOF composites often behave as additive combinations rather than truly integrated platforms that convert light input into programmable chemical outputs.

Here we propose a unified 2D/3D heterostructure that couples these functionalities at the nanoscale interface: an MXene@Cu-MOF nanoplatform constructed by in-situ growth of Cu-MOF nanocrystals directly on ultrathin Ti₃C₂T_x sheets. In this architecture, the MXene acts as a conductive, ROS-quenching and photothermal “framework”, while the Cu-MOF forms a conformal, degradable “functional backpack” [23,24]. Surface oxygenated groups on MXene serve as metal-ion anchoring and heterogeneous nucleation sites, guiding selective MOF crystallization on the basal plane and edges [25]. This heterostructural integration is designed to (i) prevent MXene restacking by interposing MOF spacers, (ii) ensure uniform MOF dispersion and intimate interfacial contact, and (iii) establish short, efficient pathways for heat and mass transfer between the photothermal core and the labile MOF domains.

We further embed the MXene@Cu-MOF heterostructure within a biocompatible gelatin methacrylate (GelMA) hydrogel to realize a moist, conformal wound dressing that localizes therapy at the wound bed. The resulting system executes a cascaded, time-staged program aligned with wound pathophysiology. Immediately after application (post-operative days 0–3), exposed MXene surfaces scavenge ROS to dampen inflammatory signaling (“first barrier”), while the acidic milieu of infection triggers pH-responsive, baseline copper release from the MOF (“self-regulated messenger”). When clinical or microbiological signs indicate escalation, NIR irradiation activates the MXene photothermal effect to raise local temperature [22]. Crucially, in this heterostructure the delivered heat is not merely bactericidal: it accelerates MOF decomposition at the MXene–MOF interface, yielding a burst of Cu²⁺ that synergizes with hyperthermia to eradicate resilient bacteria—an active photothermal–chemical synergy that is difficult to achieve in physically mixed systems. As inflammation subsides (days 3–14), residual Cu²⁺ release transitions to a low-dose, longer-tail profile within the hydrogel matrix, supporting endothelial migration and tubulogenesis and thereby promoting neovascularization in the wound tissue. Compared with previously reported MXene- or MXene–MOF-based wound dressings, our interface-engineered MXene@Cu-MOF/GelMA system therefore acts as a unified, cascade-programmed platform that couples nanoscale photothermal–chemical conversion with staged antioxidative, antibacterial, and pro-angiogenic functions specifically tailored to infected diabetic wounds.

This work specifically addresses three unmet needs in the field. First, it replaces loosely associated composites with a single, interface-engineered heterostructure, transforming two co-loaded agents into a coupled system where photothermal energy gates and amplifies ion release. Second, it introduces the concept of a ‘self-decomposing messenger’: Cu²⁺ ions that are generated only when and where microenvironmental acidity (and, if needed, controlled hyperthermia) demand them, yielding negative feedback against infection while minimizing off-target exposure. Third, by leveraging MOF spacers to inhibit MXene stacking, it resolves a central materials bottleneck—dispersion and accessible surface area—thereby enhancing both ROS quenching and heat/mass transport.

We hypothesize that (i) in-situ growth of Cu-MOF on Ti₃C₂T_x yields a stable MXene@Cu-MOF heterostructure with suppressed aggregation and strengthened interfacial coupling; (ii) this coupling enables microenvironment-specific (pH-responsive) and externally programmable (NIR-amplified) Cu²⁺ release that converts photothermal input into an on-demand chemical output; and (iii) the resulting cascaded therapy—progressing from ROS mitigation to antibacterial eradication and then to pro-angiogenic support—will accelerate closure of bacterial-infected diabetic wounds relative to physical mixtures and single-modality controls. To test these hypotheses, we systematically characterized the MXene@Cu-MOF heterostructure, evaluated its antioxidant, antibacterial, and pro-angiogenic functions in vitro, and assessed its therapeutic performance in a bacteria-infected diabetic wound model (Scheme 1).

Scheme 1.

Conceptual illustration of the hydrogel-encapsulated MXene@Cu-MOF heterostructure and its cascaded, on-demand therapeutic program for infected diabetic wounds. Ultrathin Ti₃C₂T_x MXene nanosheets serve as a conductive, ROS-modulating and photothermal framework; Cu-MOF nanocrystals are in-situ grown on MXene basal planes and edges, forming an interface-engineered 2D/3D heterostructure that is uniformly distributed within a moist, conformal hydrogel dressing. Stage I (days 0–3): immediately after application, exposed MXene surfaces scavenge ROS and attenuate inflammatory signaling; the mildly acidic infection microenvironment induces pH-responsive Cu-MOF degradation with baseline Cu²⁺ release for antibacterial action, providing self-regulated dosing that scales with acidity. Stage II (on-demand escalation): NIR irradiation activates MXene photothermal conversion, raising local temperature; interfacial heating accelerates MOF decomposition and produces a transient burst of Cu²⁺, yielding photothermal–chemical synergy that kills planktonic bacteria, and reduces bacterial burden while the hydrogel matrix helps confine heat within the wound bed. Stage III (days 3–14): as infection subsides, residual slow Cu²⁺ release supports endothelial cell migration and tubulogenesis, promotes neovascularization and collagen remodeling, and favors macrophage polarization toward a reparative phenotype.

2. Materials and methods

2.1. Reagents and materials

All chemicals and reagents were of analytical grade, including titanium aluminum carbide (Ti₃AlC₂, MAX phase), copper(II) nitrate trihydrate, 1,3,5-benzenetricarboxylic acid (H₃BTC), 2-methylimidazole, zinc nitrate hexahydrate, tris(hydroxymethyl)aminomethane, hydrochloric acid, lithium fluoride, hydrogen peroxide (H₂O₂, 30 %), lipopolysaccharide (LPS), phorbol esters, phosphate-buffered saline (PBS), ethanol, sodium acetate, acetic acid, and nitric acid (trace-metal grade). GelMA with a degree of methacrylation of approximately 70 % was used for all experiments. A constant concentration of 5 % (w/v) GelMA and 0.25 % (w/v) photoinitiator lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) was used to prepare the precursor solution. Unless otherwise specified, they were purchased from Sigma-Aldrich (USA).

RAW264.7 macrophages and Human Umbilical Vein Endothelial Cells (HUVECs) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured as described in the following sections. Escherichia coli (E. coli) ATCC 25922 and Staphylococcus aureus (S. aureus) ATCC 25923 were employed for antibacterial assays. All experiments were conducted using ultrapure water (resistivity 18.2 MΩ cm). A total of 30 male C57BL/6 mice (8–10 weeks old, weighing 22–26 g) were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China).

2.2. Synthesis of Ti₃C₂T_x MXene nanosheets

Ti₃C₂T_x was produced by selective etching of Ti₃AlC₂ using a LiF/HCl mixture, followed by delamination. Briefly, LiF was dissolved in 9 M HCl, Ti₃AlC₂ powder was added slowly with stirring in an ice bath, and the etching proceeded at 35 °C for 24 h. The slurry was repeatedly washed by centrifugation and resuspension in water until supernatant pH was approximately 6. The sediment was sonicated (ice bath, 30–60 min) under nitrogen to delaminate layers, then centrifuged at low speed to remove unexfoliated particles. The supernatant containing few-layer Ti₃C₂T_x was stored under nitrogen at 4 °C and used within two weeks.

2.3. In-situ growth of Cu-MOF on MXene

Ti₃C₂T_x (MXene) dispersion (0.5–2.0 mg/mL) was mixed with Cu(NO₃)₂ solution in water–ethanol (1:1 v/v), followed by slow addition of H₃BTC under stirring. The mixture was maintained at 25–60 °C for 2–12 h at pH 3–4, allowing heterogeneous nucleation on MXene basal planes and edges. The product was collected by centrifugation, washed with ethanol and water, and re-dispersed in water.

MXene was pre-equilibrated with Zn²⁺/Cu²⁺ solution, after which 2-methylimidazole solution was added rapidly to induce ZIF-8 growth with partial Cu substitution. The composite was washed with methanol and water and stored at 4 °C.

Copper loading was quantified by inductively coupled plasma–optical emission spectrometry (ICP-OES) after acid digestion in 2 % HNO₃. All concentrations reported “with respect to nanosheets” refer to the mass of Ti₃C₂T_x in the composite.

The in-situ growth of Cu-MOF on Ti₃C₂T_x MXene was carried out under fixed precursor concentrations, MXene-to-Cu²⁺ ratio, temperature, reaction time, and stirring/dispersion conditions, which were optimized to afford uniformly distributed nanoscale Cu-MOF crystals on the MXene surfaces.

2.4. Preparation of MXene@Cu-MOF/GelMA hydrogels

5 % (w/v) GelMA was dissolved in PBS at 37 °C with 0.25 % (w/v) LAP. MXene@Cu-MOF was homogenized into the prepolymer solution at the indicated nanosheet concentrations. Hydrogels were formed by photopolymerization under 405 nm light (20 mW/cm²) for 30–60 s in sterile molds or directly on tissue for in vivo use. Hydrogels were equilibrated in PBS before assays.

2.5. Swelling behavior

Disk-shaped hydrogels (GelMA, MXene/GelMA, Cu-MOF/GelMA, MXene@Cu-MOF/GelMA; Ø ≈ 8 mm, thickness ≈ 2 mm) were photocrosslinked, lyophilized, and weighed to obtain dry weight W_d. Dried samples were immersed in PBS (pH 7.4, 37 °C). At 0.5, 1, 2, 4, 8 and 12 h, hydrogels were removed, gently blotted, and weighed to obtain W_s (t). The swelling ratio was calculated as Q(t)= (W_s (t) - W_d)/W_d.

2.6. Rheological properties

Cylindrical hydrogels were equilibrated in PBS at 37 °C and measured on a parallel-plate rheometer. A strain sweep (0.1–10 %) at 1 Hz was used to determine the linear viscoelastic region; frequency sweeps (0.1–10 Hz) were then performed at 37 °C and 1 % strain to obtain storage modulus (G′) and loss modulus (G″) for all formulations.

2.7. Adhesive performance

Porcine skin substrates (≈25 × 25 mm) were overlapped with an area of ∼1 cm². Prepolymer solution was placed in the overlap region, photocrosslinked, and equilibrated in PBS at 37 °C for 30 min. Lap-shear tests were carried out on a universal testing machine at 10 mm/min until failure. Adhesive strength was calculated from the maximum force divided by the overlap area. For MXene@Cu-MOF/GelMA, additional measurements were performed after NIR irradiation under the same conditions used in biological experiments.

2.8. Electron and probe microscopy

Dried powders or hydrogel cryo-fracture surfaces were sputter-coated with gold. Scanning Electron Microscopy (SEM; SU8010, Hitachi, Japan) imaging and energy-dispersive X-ray spectroscopy (EDS) elemental mapping were performed using X-MaxN 80 (voltage 5–15 kV; Oxford Instruments, UK). For bacterial morphology, fixed and dehydrated cells (see Antibacterial assays) were imaged on conductive stubs. Drops of dilute dispersions were cast on carbon-coated copper grids and dried. Transmission electron microscopy (TEM) and high-angle annular dark-field scanning TEM with EDS line/mapping were acquired (200–300 kV). Samples were deposited on freshly cleaved mica. Particle diameters of MXene@Cu-MOF heterostructures were measured based on TEM images. Tapping-mode atomic force microscopy (AFM) was performed on a Dimension Icon system (Bruker, USA); height maps and 3D renderings were processed in peak force tapping mode. Attenuated total reflectance-Fourier Transform Infrared Spectroscopy (FTIR, 4000-400 cm⁻¹) were collected on a Nicolet iS50 spectrometer (Thermo Fisher Scientific, USA). Spectra were baseline-corrected and vertically offset as indicated. X-ray photoelectron spectra (XPS) were acquired with Al Kα radiation (1486.6 eV). Charge correction was applied to C 1s at 284.8 eV. High-resolution regions for Ti 2p, C 1s, O 1s, F 1s, Cu 2p, and N 1s were deconvoluted with mixed Gaussian–Lorentzian functions. Powder X-ray diffraction (XRD) was recorded using Cu Kα radiation (λ = 1.5406 Å). Patterns for delaminated MXene emphasized the (002) reflection and its harmonics.

Dynamic light scattering (DLS) and electrophoretic light scattering were measured at 25 °C using a Malvern Zetasizer Nano ZS (Malvern Panalytical, UK). For DLS, intensity-weighted hydrodynamic sizes were obtained after brief sonication in water. ζ potentials were measured in 1 mM KCl at the indicated pH.

2.9. In vitro antioxidant assays

ABTS•⁺ was generated by reacting ABTS with potassium persulfate for 12–16 h in the dark and diluting to an absorbance of approximately 0.70 at 734 nm. Samples at 0–1.0 mg/mL were incubated with ABTS•⁺ for 6–10 min, and percent scavenging was calculated from absorbance decrease. For DPPH, ethanolic DPPH solution (A₅₁₇ ∼0.9) was mixed with samples for 30 min in the dark. Superoxide scavenging was assessed using a xanthine–xanthine oxidase–nitroblue tetrazolium system monitored at 560 nm. Hydroxyl radical (·OH) scavenging was evaluated using a Fenton reaction with salicylic acid probe, monitoring the chromophore at 510 nm. All assays included appropriate blanks and standards, and results were normalized to radical controls. To correct for particle-induced light scattering or intrinsic absorption, for each concentration, a sample-matched blank containing the same concentration of MXene, Cu-MOF, or MXene@Cu-MOF without ABTS/DPPH was measured, and its absorbance was subtracted from that of the corresponding reaction mixture prior to calculating scavenging efficiency; in addition, samples were mildly centrifuged and the supernatant was used for measurement when needed to reduce turbidity.

RAW264.7 macrophages were cultured in Dulbecco's Modified Eagle Medium supplemented with 10 % fetal bovine serum and 1 % penicillin–streptomycin at 37 °C and 5 % CO₂. For oxidative challenge, cells were exposed to 200–500 μM H₂O₂ for 1 h or 1 μg/mL LPS for 12–24 h. 2′,7′-Dichlorodihydrofluorescein diacetate (DCFH-DA, 10 μM) was loaded in serum-free medium for 30 min at 37 °C, followed by washes and treatment with PBS, GelMA, MXene/GelMA, or MXene@Cu-MOF/GelMA under the indicated NIR regimen. Fluorescence imaging was performed on a confocal microscope with identical exposure settings, and flow cytometry was conducted on a BD FACSCelesta flow cytometer (BD Biosciences, USA), gating viable single cells. Data were background-corrected and normalized to controls.

2.10. Macrophage polarization assays

RAW264.7 cells were primed with LPS (1 μg/mL, 12 h) for M1 or with IL-4 (20 ng/mL, 24 h) for M2 controls. After treatments, cells were fixed with 4 % paraformaldehyde, permeabilized, and blocked. Primary antibodies against CD86 and CD206 were incubated overnight at 4 °C followed by species-appropriate fluorophore-conjugated secondary antibodies. Nuclei were counterstained with 4′,6-Diamidino-2-Phenylindole (DAPI). Images were acquired under identical settings. For Western blotting, lysates were prepared in RIPA buffer, separated by SDS–PAGE, transferred to PVDF membranes, and probed for interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), and interleukin-1β (IL-1β) with β-actin as loading control. Bands were visualized by chemiluminescence and quantified densitometrically.

2.11. Photothermal measurements

An 808 nm continuous-wave laser with adjustable irradiance (0.5–1.5 W/cm²) was used. Samples (PBS, GelMA, MXene@Cu-MOF dispersion, and MXene@Cu-MOF/GelMA) were irradiated for 10 min unless stated otherwise. Infrared thermal imaging was captured with a FLIR A655sc infrared thermal camera (FLIR Systems, Inc., USA), and temperature–time traces were extracted from the same region of interest. Concentration and irradiance dependencies were evaluated at nanosheet concentrations of 0.5–1.5 mg/mL and irradiances of 0.5–1.5 W/cm². Stability was tested over ten heating-cooling cycles (10 min irradiation followed by 10 min natural cooling). Photothermal conversion efficiency can be calculated from the heating-cooling curves using the Roper method.

2.12. Copper release and hydrogel degradation

For release from MXene@Cu-MOF or MXene@Cu-MOF/GelMA, samples were placed in dialysis cassettes (molecular-weight cutoff 3.5–10 kDa) immersed in PBS (pH 7.4) at 37 °C under gentle shaking. At specified times, aliquots of the external medium were withdrawn and replaced with fresh buffer. For NIR-triggered release, samples were irradiated at 808 nm, 1.5 W/cm², for 10 min at each time point. Cu²⁺ concentrations were quantified by ICP-OES after acidification with 2 % HNO₃ against certified standards. Hydrogel mass loss in PBS was determined by recording wet mass at baseline and after incubation with or without intermittent NIR exposure; values were reported as normalized mass remaining.

2.13. Antibacterial testing in vitro

Both E. coli and S. aureus were grown to mid-log phase, washed, and diluted to approximately 10⁶-10⁷ colony-forming units (CFU)/mL. Suspensions were incubated with PBS, GelMA, MXene@Cu-MOF, or MXene@Cu-MOF/GelMA for 10–30 min at 37 °C and then irradiated with 808 nm NIR (1.5 W/cm², 10 min) unless otherwise stated. Aliquots were serially diluted and plated on Luria-Bertani agar for overnight incubation at 37 °C, followed by digital imaging and image-based colony area analysis. Live/Dead staining used SYTO9 and propidium iodide according to the manufacturer's instructions; fluorescence microscopy was performed under fixed exposure. For SEM, bacteria were fixed in 2.5 % glutaraldehyde, dehydrated through graded ethanol, dried, sputter-coated, and imaged as described above.

2.14. Endothelial assays in vitro

HUVECs were cultured in endothelial growth medium at 37 °C and 5 % CO₂. For EdU incorporation, cells were treated with the indicated samples, irradiated as appropriate, and incubated with EdU (10 μM) for 2 h, followed by fixation and click-chemistry detection. For scratch migration, confluent monolayers were scratched with a sterile pipette tip, rinsed, treated, and imaged at baseline and specified times; wound areas were quantified with ImageJ. For tube formation, growth-factor-reduced Matrigel was polymerized in 96-well plates, HUVECs were seeded at 2 × 10⁴ cells per well, treated, and imaged after 6 h; branch points and total tube length were quantified with ImageJ. For Reverse Transcription Quantitative Polymerase Chain Reaction (RT-qPCR), RNA was extracted with TRIzol, reverse-transcribed, and amplified with SYBR Green on a real-time PCR system. Primers targeted Vascular Endothelial Growth Factor (VEGF), Endothelial Nitric Oxide Synthase (eNOS), Hypoxia-Inducible Factor 1α (HIF-1α), and Fibroblast Growth Factor 2 (FGF2); expression was normalized to GAPDH using the 2^−ΔΔCt method. Primers used in the qPCR were listed in Table S1.

2.15. Animal studies

All procedures were approved by the Institutional Animal Care and Use Committee of Ruijin Hospital (Shanghai, China) and conformed to national guidelines. Male C57BL/6 mice were rendered diabetic by intraperitoneal streptozotocin (50 mg/kg for 5 consecutive days). Fasting blood glucose was measured after one week; mice with values above 300 mg/dL were enrolled. Under isoflurane anesthesia, dorsal hair was removed and two full-thickness excisional wounds were created with a 6 mm biopsy punch. A silicone splint was adhered around each wound to minimize contraction. Wounds were inoculated with 10⁷ CFU of S. aureus to establish infection. Treatments were applied immediately: PBS, GelMA, MXene/GelMA, Cu-MOF/GelMA, or MXene@Cu-MOF/GelMA. Where indicated, NIR irradiation (808 nm, 1.5 W/cm², 10 min) was delivered once daily on postoperative days 0–4 using a fixed distance and spot size. Wounds were photographed at days 0, 5, 9, and 14 with a scale bar; areas were quantified by planimetry in ImageJ.

At days 9 and 14, animals were euthanized and wound tissues were harvested with margins of healthy skin, fixed in 4 % paraformaldehyde, and paraffin-embedded. Sections (5 μm) were stained with hematoxylin and eosin (H&E) for general morphology and re-epithelialization, and with Masson's trichrome for collagen deposition. Epidermal thickness was measured at standardized distances from the wound center. For immunohistochemistry (IHC), antigen retrieval was performed in citrate buffer, followed by blocking, incubation with primary antibodies to collagen I or collagen III, and HRP-based detection. Positive area fractions were quantified from at least five random fields per section using blinded analysis.

Bacterial burden was assessed by fluorescence in-situ hybridization (FISH) using pan-bacterial probes, with DAPI counterstain. Tissue reactive oxygen species were evaluated by dihydroethidium (DHE) staining and imaging under fixed exposure settings. Macrophages were labeled with antibodies against CD68 (pan-macrophage) and CD206 (M2-associated), and blood vessels were visualized by CD31 immunofluorescence. For each marker, at least five fields per section and three sections per wound were analyzed. Quantifications included bacterial counts per field, DHE mean fluorescence intensity, CD68⁺ and CD206⁺ cells per field, and CD31⁺ microvessels per field.

Total RNA from day-14 wound tissues in MXene@Cu-MOF/GelMA treated (EXP) and GelMA treated (CON) were extracted with the RNeasy kit, and RNA integrity was confirmed (RIN ≥7). Poly(A) libraries were prepared using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (Catalog No. E7760S, New England Biolabs, USA) according to the manufacturer's instructions. and sequenced on an Illumina NovaSeq 6000 to produce 150 bp paired-end reads. Reads were quality-checked with FastQC, trimmed with Trimmomatic, aligned to the mouse reference genome (GRCm39) with STAR, and gene-level counts were generated with featureCounts. Differential expression genes (DEGs) used DESeq2 with default dispersion estimation and Benjamini–Hochberg correction. Significance was defined as adjusted P < 0.05 and absolute log₂ fold change ≥1 unless otherwise stated. Variance-stabilized data were used for principal component analysis (PCA). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) over-representation were performed with clusterProfiler; gene set enrichment analysis (GSEA) used the Molecular Signatures Database gene sets and the fgsea implementation. Heatmaps display gene-wise z-scores.

2.16. Statistics

Data are reported as mean ± standard deviation (SD) unless otherwise noted. The number of biological replicates is indicated in figure legends. Normality was assessed by Shapiro–Wilk when appropriate. Two-group comparisons used unpaired two-tailed Student's t-test. Multi-group comparisons used one-way ANOVA with Tukey's post hoc test. Nonparametric alternatives (Mann–Whitney U or Kruskal–Wallis with Dunn's post hoc) were used when assumptions were violated. Statistical significance thresholds were set as P < 0.05 (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001), ns denotes no statistically significant difference.

3. Results and discussion

3.1. Morphology and heterostructure formation of MXene@Cu-MOF

DLS analysis of Cu-MOF in water at room temperature after brief sonication showed a unimodal, intensity-weighted hydrodynamic size distribution with a mean diameter of 78.6 ± 3.4 nm (Fig. S1). SEM and TEM revealed that polyhedral Cu-MOF nanocrystals were uniformly anchored on ultrathin Ti₃C₂T_x (MXene) nanosheets, producing a continuous 2D/3D contact interface rather than a loose physical mixture (Fig. 1A and B). In TEM, the Cu-MOF domains appeared as discrete nanocrystals distributed over the MXene basal plane and edges, consistent with heterogeneous nucleation on oxygen-terminated sites. Quantitative analysis showed that the MXene@Cu-MOF nanocrystals have an average diameter of approximately 85.91 ± 0.73 nm with a relatively narrow size distribution (Fig. S2), which is consistent with the visual impression of uniform nucleation and growth of Cu-MOF on the MXene sheets observed in TEM images. AFM showed increased surface height and roughness for MXene@Cu-MOF compared to pristine MXene, consistent with a conformal MOF overlayer (Fig. 1C). Such MOF “spacer” domains are known to mitigate MXene restacking and to maintain accessible surface area, a recurring advantage of MOF/MXene heterostructures. The spatial distributions of constituent elements are illustrated by EDS mapping: C, N, O, and Cu in the Cu-MOF (Fig. S3), and C, O, F, and Ti in the MXene (Ti₃C₂T_x) (Fig. S4). Elemental mapping of the MXene@Cu-MOF heterostructure showed co-localized C/N/O/Cu signals with the MOF domains and Ti/F delineating the MXene framework, supporting intimate intergrowth at the nanoscale (Fig. 1D). This interfacial assembly aligns with recent reports that in-situ growth of MOFs on MXene yields intimate contact and avoids phase segregation often observed in post-mixing approaches [26]. Taken together, the microscopy suite substantiates formation of a single heterostructure in which MOF nanocrystals are rooted on the MXene scaffold rather than merely adsorbed or embedded at a distance.

Fig. 1.

Structural and morphological characterization of the synthesized MXene@Cu-MOF heterostructure. (A) SEM images of Cu-MOF, MXene (Ti₃C₂T_x), and MXene@Cu-MOF heterostructure. (B) TEM images of Cu-MOF, MXene, and MXene@Cu-MOF, showing MOF nanocrystals uniformly anchored on MXene nanosheets. (C) AFM 2D height images and corresponding 3D surface profiles of Cu-MOF, MXene, and MXene@Cu-MOF. (D) EDS elemental maps of the MXene@Cu-MOF heterostructure showing the spatial distribution of N, Ti, C, F, O, and Cu.

At 37 °C in PBS, the swelling ratio (Q) of all hydrogel formulations increased rapidly over time and reached a plateau at approximately 4–8 h (Fig. S5). Pristine GelMA exhibited the highest equilibrium swelling ratio, whereas incorporation of MXene, Cu-MOF, or MXene@Cu-MOF led to a slight decrease in Q, suggesting that the nanofillers introduce additional physical crosslinking points and reinforce the network, thereby moderately restraining swelling. Nevertheless, the equilibrium Q values of all composite hydrogels remained within the typical range reported for wound dressings, providing adequate water uptake and moisture retention without compromising mechanical stability due to excessive swelling.

As shown in Fig. S6A–B, the storage modulus (G′) is markedly higher than the loss modulus (G″) for all formulations, indicating that the hydrogels exhibit predominantly elastic, solid-like behavior. Compared with pristine GelMA, incorporation of MXene and/or Cu-MOF increases both G′ and G″, with MXene@Cu-MOF/GelMA displaying the highest moduli, consistent with a reinforced and more structurally stable hydrogel network.

The adhesive strength of the GelMA-based hydrogels was evaluated by lap-shear tests on model substrates. As shown in Fig. S7, pristine GelMA exhibited an adhesive strength of approximately 7–8 kPa. Incorporation of MXene or Cu-MOF slightly increased the interfacial strength, whereas MXene@Cu-MOF/GelMA showed the highest adhesive strength among all groups. These results indicate that the MXene@Cu-MOF heterostructure does not compromise, and may even enhance, the adhesion of GelMA to tissue-mimicking substrates, which is attributable to the functional groups of GelMA and the increased surface roughness and interfacial interactions introduced by the nanofillers.

FTIR spectrum of the heterostructure retained hallmark bands from both components while displaying subtle shifts and intensity changes in linker-related modes (e.g., carboxylate or imidazolate vibrations, depending on MOF) and in MXene termination vibrations (Fig. S8). Specifically, MXene showed a broad band at ∼3425 cm⁻¹ corresponding to surface -OH/-O- terminations and a band near 1618 cm⁻¹ attributable to surface C=O/C-O groups, whereas Cu-MOF exhibited characteristic ligand-related bands at ∼1605 cm⁻¹ (COO⁻/C=C), ∼1375 cm⁻¹ (COO⁻/C-O), and ∼1015 cm⁻¹ (C-N stretching). In MXene@Cu-MOF, these Cu-MOF ligand bands are preserved, while the MXene-OH/C=O features become attenuated and slightly shifted, accompanied by the appearance/enhancement of metal-oxygen vibrations. Such spectral evolution is consistent with coordination between Cu nodes and oxygenated MXene terminations at the interface, as described for MOF/MXene hybrids prepared by in-situ routes [26].

XPS spectra of Cu-MOF showed the expected C/N/O/Cu elements (Fig. S9A). High-resolution spectra of Ti₃C₂T_x revealed Ti 2p components attributable to Ti–C and surface-terminated Ti–O_x/Ti–F_x species; C 1s and O 1s features further supported the carbide backbone and terminations. These assignments follow recommended fitting protocols for Ti₃C₂T_x (Fig. S9B). In addition, the Cu 2p region of Cu-MOF displays well-defined Cu²⁺ main peaks together with characteristic shake-up satellite features, confirming the predominance of divalent copper in the framework. The presence of Ti-O/-OH terminations on MXene, as indicated by the Ti 2p and O 1s components, provides plausible anchoring and coordination sites for Cu-MOF growth at the heterointerface. In MOF contexts, Cu 2p spectra commonly exhibit shake-up satellites characteristic of Cu²⁺ in paddlewheel nodes, and mixed Cu²⁺/Cu⁺ states can appear depending on synthesis history; these spectral motifs frame interpretation of the Cu environments in Cu-MOFs [27].

Powder XRD of delaminated Ti₃C₂T_x showed a strong (002) reflection and higher-order harmonics, consistent with expanded interlayer spacing after exfoliation (Fig. S9C). MOF-on-MXene architectures are widely reported to suppress re-aggregation relative to physical blends, which helps preserve these structural features in composites [23].

Electrophoretic light-scattering showed distinct ζ-potential signatures for each component and a reproducible shift for MXene@Cu-MOF relative to its parents, indicating interfacial assembly alters surface charge (Fig. S10).

Collectively, FTIR, XPS, and XRD corroborate an interface-engineered MXene@Cu-MOF heterostructure in which coordination at oxygenated MXene sites stabilizes MOF attachment and mitigates MXene restacking. Two design weaknesses of physically mixed dressings are poor component dispersion and limited coupling between modalities. In contrast, the present heterostructure shows uniform MOF anchoring and short heat/mass-transfer paths at the MXene–MOF interface. Multiple recent studies emphasize that MOF domains can act as spacers to keep MXene sheets apart while MXene grants electrical/thermal conductivity to the MOF phase; this mutualism is difficult to realize by simple mixing [28].

Several recent reviews and primary studies on MXene–MOF hybrids concur that in-situ growth offers superior interfacial contact and dispersion compared with solution blending, with MOF spacers counteracting MXene restacking and enhancing transport. The present structural dataset fits this consensus and extends it to a Cu-MOF/MXene system aimed at infected diabetic wounds, a setting where microenvironment-sensitive delivery and on-demand escalation are particularly valuable [23,26].

3.2. In vitro antioxidant capacity of MXene@Cu-MOF heterostructure

We next asked whether MXene@Cu-MOF preserves and enhances the intrinsic ROS-scavenging capacity of Ti₃C₂T_x in chemical and cellular settings. MXene@Cu-MOF exhibited concentration-dependent scavenging of ABTS•⁺ and DPPH• across 0.1–1.0 mg/mL, with percent inhibition increasing monotonically relative to radical controls. The ABTS readout in aqueous medium showed robust quenching, and DPPH in alcoholic medium displayed a similar dose–response, consistent with the known solvent and kinetic differences between the two assays (Fig. 2A and B). These outcomes align with the established use of ABTS and DPPH as complementary, single-electron transfer–dominated methods for ranking antioxidant capacity [29].

Fig. 2.

In vitro antioxidant activity of the MXene@Cu-MOF heterostructure. (A) UV–Vis spectra and quantification of ABTS•⁺ scavenging by MXene@Cu-MOF at 0 (Control), 0.1, 0.2, 0.5, and 1.0 mg/mL (B) UV–Vis spectra and quantification of DPPH• scavenging by MXene@Cu-MOF at 0 (Control), 0.1, 0.2, 0.5, and 1.0 mg/mL (C) UV–Vis spectra and quantification of superoxide (O₂•^-) scavenging by MXene@Cu-MOF at 0 (Control), 0.1, 0.2, 0.5, and 1.0 mg/mL (D) UV–Vis spectra and quantification of hydroxyl radical (•OH) scavenging by MXene@Cu-MOF at 0 (Control), 0.1, 0.2, 0.5, and 1.0 mg/mL. (E) Representative fluorescence images of intracellular ROS in RAW264.7 macrophages stained with DCFH-DA (red) under the following treatments: PBS (blank), H₂O₂ only, H₂O₂ + GelMA, H₂O₂ + MXene/GelMA, and H₂O₂ + MXene@Cu-MOF/GelMA. (F) Flow cytometric analysis of cellular ROS (DCF fluorescence) for the groups in (E). (G) Representative DCFH-DA staining (red) of intracellular ROS in RAW264.7 macrophages treated with: PBS (blank), LPS alone, LPS + GelMA, LPS + MXene/GelMA, and LPS + MXene@Cu-MOF/GelMA. (H) Flow cytometric analysis of cellular ROS (DCF fluorescence) for the groups in (G). Data are presented as mean ± SD (n = 3). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

In chemical systems that generate reactive oxygen intermediates, MXene@Cu-MOF reduced O₂•^- and •OH signals in a dose-dependent manner, indicating broad-spectrum radical mitigation beyond chromogenic ABTS/DPPH surrogates (Fig. 2C and D). Given the electron-rich Ti₃C₂T_x surface and oxygenated terminations, MXenes have been reported to engage redox species and modulate ROS, which is consistent with the present multi-assay profile [30].

Because different radical assays report distinct chemistries and may be differentially sensitive to matrix and solvent, applying a panel of orthogonal tests is recommended to avoid over-interpreting any single metric; the present agreement across ABTS, DPPH, superoxide, and hydroxyl assays strengthens the conclusion that MXene@Cu-MOF possesses intrinsic antioxidant capacity [31].

In RAW264.7 macrophages, H₂O₂ exposure elevated intracellular ROS as measured by DCFH-DA fluorescence. Incorporation of pristine GelMA produced minimal attenuation, whereas MXene/GelMA reduced the fluorescent signal. MXene@Cu-MOF/GelMA achieved the most pronounced decrease, yielding a significant reduction compared with H₂O₂ only and GelMA controls (Fig. 2E and F; Fig. S11A). These data indicate that integrating Cu-MOF onto MXene does not diminish the ROS-quenching behavior attributed to Ti₃C₂T_x; rather, the heterostructure embedded in GelMA preserves or strengthens this cellular effect. Reports surveying MXene-based antimicrobials and wound materials similarly describe ROS modulation and cytoprotective trends in immune and stromal cells [30].

Under LPS stimulation, which models inflammatory ROS bursts in macrophages, GelMA again had limited impact, MXene/GelMA reduced the DCF signal, and MXene@Cu-MOF/GelMA produced the largest decline. Flow-cytometric quantification mirrored microscopy, with significant differences between MXene@Cu-MOF/GelMA and control groups (Fig. 2G and H; Fig. S11B). This pattern is consistent with the view that controlling intracellular oxidative stress can dampen downstream inflammatory signaling relevant to chronic and diabetic wounds [32].

The DCFH-DA assay is widely used yet prone to artifacts related to probe loading, esterase activity, photo-oxidation, and microenvironmental pH and O₂. In line with recent guidelines and protocol refinements, assays here were performed with matched exposure, consistent gating on viable single cells, and independent repeats. Interpreting DCF signals alongside orthogonal chemical assays (ABTS/DPPH/superoxide/•OH) addresses common concerns about over-reliance on a single probe [33].

The multi-assay antioxidant activity is plausibly rooted in the electronic structure and surface terminations of Ti₃C₂T_x MXene, which can engage ROS through electron or hydrogen-atom transfer and by modulating catalytic redox cycles at the solid–liquid interface. Recent syntheses and reviews describe MXenes as photothermally active, redox-interactive materials with emerging “nanozyme-like” behaviors that are beneficial in infected wound microenvironments characterized by oxidative stress [34]. In the present heterostructure, Cu-MOF domains are in intimate contact with the MXene surface; although copper can participate in redox chemistry that is context dependent, contemporary analyses of Cu-based MOFs underscore their capacity to modulate oxidative stress and provide antimicrobial action when release is controlled by microenvironmental cues [18].

Excess ROS in diabetic wounds sustains inflammation, damages extracellular matrix and cells, and impairs angiogenesis. By reducing extracellular radical surrogates and intracellular DCF signals in macrophages, MXene@Cu-MOF/GelMA addresses an upstream driver of non-healing states. This antioxidant function complements the antibacterial and pro-angiogenic roles already established for copper-containing biomaterials and dressings in preclinical and clinical contexts, suggesting that the same platform could coordinate early inflammation control with later vascular support when tuned appropriately [35].

Many reported antioxidant hydrogels rely on phenolic polymers, small-molecule scavengers, or enzymatic additives; these can be effective yet sometimes suffer from rapid depletion or limited integration with antibacterial modalities. MXene-containing composites have shown ROS modulation and antimicrobial activity, and Cu-MOF systems have demonstrated antibacterial effects. The present heterostructure leverages both in a single, interface-engineered construct that exhibits chemical radical scavenging and cellular ROS suppression at baseline. In light of literature describing MOF-MXene synergies and the therapeutic versatility of HKUST-1-based composites, an integrated MXene@Cu-MOF within GelMA offers a materials path toward wound dressings that are simultaneously antioxidant, antimicrobial, and, with appropriate dosing, pro-angiogenic [36].

Because ABTS/DPPH readouts can be affected by light absorption and scattering from particulate suspensions, confirming trends across multiple assays, as done here, aligns with best practices. Future work should quantify rate constants or equivalent metrics for radical quenching, expand to cell types central to granulation (fibroblasts, endothelial cells), and profile cytokine outputs alongside ROS to confirm immunomodulation. Following recent community guidance on ROS measurements, additional probes and genetically encoded sensors could triangulate intracellular redox states, while release-controlled copper dosing can be tuned to maximize antimicrobial benefit without provoking oxidative damage [33].

3.3. Macrophage polarization and inflammatory readouts in vitro

In LPS-stimulated RAW264.7 macrophages, immunostaining revealed strong CD86 signal with relatively weak CD206, consistent with an M1-skewed state. As an assay control, IL-4 treatment produced the expected M2-like phenotype dominated by CD206. Incorporation of pristine GelMA produced only minor changes relative to LPS alone, whereas MXene/GelMA visibly reduced CD86 and increased CD206. The MXene@Cu-MOF/GelMA group showed the most pronounced shift, with lower CD86 and higher CD206 than the MXene/GelMA group. Quantification (normalized mean fluorescence intensity or percentage of positive cells) mirrored the images (Fig. S12A–B). These marker choices and controls align with accepted practice in RAW264.7 polarization assays, where LPS (±IFN-γ) induces M1 programs and IL-4 drives M2 features, and where CD86 and CD206 are widely used surface readouts [37].

Western blots from the same treatment sets showed elevated IL-6, TNF-α, and IL-1β in the LPS group, with densitometry normalized to the loading control. GelMA alone produced modest attenuation, MXene/GelMA reduced band intensities further, and MXene@Cu-MOF/GelMA yielded the largest decreases relative to LPS and GelMA controls (Fig. S12C–D). These trends are consistent with canonical LPS responses in RAW264.7 cells and provide protein-level corroboration of the immunofluorescence readouts [38].

Overall, the data indicate that embedding the heterostructure in GelMA preserves a biocompatible matrix while conferring a stronger anti-inflammatory, pro-resolving polarization signature than GelMA or MXene alone under identical assay conditions.

Two complementary features of the MXene@Cu-MOF/GelMA platform likely underlie the observed anti-inflammatory effects. First, Ti₃C₂T_X MXene has documented capacity to dampen oxidative stress in inflammatory settings; by lowering intracellular ROS, it can indirectly reduce NF-κB-driven transcription of inducible Nitric Oxide Synthase (iNOS) and pro-inflammatory cytokines. Recent reviews in wound-care contexts show Ti₃C₂-based materials decreasing intracellular ROS and protecting cells from oxidative damage, consistent with the reduced IL-6, TNF-α, and IL-1β observed here [39]. Second, copper released from Cu-based biomaterials can modulate macrophage phenotypes when dosed judiciously. In vitro studies report that sub-to-low-micromolar Cu²⁺ promotes expression of M2-related genes and suppresses M1-associated markers, while broader reviews highlight copper's immunoregulatory role in tissue repair. The greater CD206 upregulation and CD86 suppression in the MXene@Cu-MOF/GelMA group are consistent with this literature [40].

MXene-containing composites have been reported to attenuate inflammatory signaling and contribute antimicrobial or photothermal functions in wound models, whereas copper-containing biomaterials can reduce pro-inflammatory cytokines and support pro-regenerative macrophage phenotypes. By integrating Cu-MOF directly on MXene within a single hydrogel dressing, the present approach consolidates these behaviors in one platform and, importantly, does so under identical assay conditions where GelMA alone shows little effect. This interface-engineered design is coherent with current trends toward immunomodulatory biomaterials that shape macrophage responses to accelerate resolution and tissue repair [41].

3.4. Photothermal heating characteristics in dispersions and hydrogels

Under 808-nm NIR irradiation, MXene@Cu-MOF dispersions exhibited rapid and sustained temperature elevation, whereas PBS and GelMA alone showed minimal heating. Embedding the nanoplatform in GelMA preserved robust photothermal behavior, with a slightly moderated heating rate compared with the dispersion, consistent with heat diffusion and partial optical attenuation by the hydrogel matrix (Fig. 3A and B). The temperature rises of MXene@Cu-MOF/GelMA increased with nanosheet concentration at a fixed irradiance and increased with irradiance at a fixed concentration, indicating controllable thermal dosing (Fig. 3C and D).

Fig. 3.

Photothermal behavior of the MXene@Cu-MOF/GelMA platform. (A) Infrared thermal images of Control (PBS), MXene@Cu-MOF, GelMA, and MXene@Cu-MOF/GelMA (all at 1.0 mg/mL, based on nanosheets) under NIR irradiation (1.5 W/cm², 10 min). (B) Corresponding temperature–time profiles for the groups in (A). (C) Temperature rise of MXene@Cu-MOF/GelMA at different nanosheet concentrations (0.5, 1.0, and 1.5 mg/mL) under NIR irradiation (1.5 W/cm²). (D) Temperature rise of MXene@Cu-MOF/GelMA at a fixed concentration (1.0 mg/mL) under different irradiances (0.5, 1.0, and 1.5 W/cm²). (E) Heating–cooling curve of MXene@Cu-MOF/GelMA (1.0 mg/mL) during 10 min NIR irradiation followed by 10 min natural cooling (1.5 W/cm²). (F) Photothermal stability of MXene@Cu-MOF/GelMA: repeated heating–cooling cycles (10 cycles) at 1.0 mg/mL under NIR irradiation (1.5 W/cm²). (G) In vitro Cu²⁺ release kinetics from MXene@Cu-MOF with or without NIR irradiation. (H) In vitro Cu²⁺ release kinetics from MXene@Cu-MOF/GelMA with or without NIR irradiation. (I) In vitro degradation profiles of MXene@Cu-MOF/GelMA in PBS (pH 7.4), with or without NIR irradiation. (J) pH-dependent Cu²⁺ release from Cu-MOF archetypes and the effect of NIR heating. Data are presented as mean ± SD (n = 3).

Heating–cooling traces showed fast on–off thermal responses and reproducible profiles across multiple cycles, indicating photothermal stability of the heterostructure within the hydrogel network (Fig. 3E and F). Photothermal cycling fidelity has been emphasized as a practical requirement for dressings that may be irradiated in repeated sessions; recent MXene-based hydrogels have demonstrated comparable stability over multiple NIR cycles [42].

In vitro release studies showed baseline Cu²⁺ liberation in the dark that increased under NIR irradiation for both the free heterostructure and the hydrogel-embedded form (Fig. 3G and H). The magnitude and rate of release were higher with irradiation, consistent with thermally accelerated framework degradation and faster diffusion. NIR-triggered acceleration of MOF degradation and cargo release has been demonstrated in MOF composites designed for light-programmed delivery [43].

MXene@Cu-MOF/GelMA displayed gradual mass loss in PBS (pH 7.4), with modest additional softening or erosion upon intermittent NIR exposure (Fig. 3I). Such behavior is in line with reports that GelMA is largely NIR-transparent yet can experience matrix relaxation as local temperature rises during photothermal operation in composite hydrogels [44].

Cu-MOF archetypes showed greater Cu²⁺ release under mildly acidic conditions than at neutral pH, and NIR exposure further increased release rates (Fig. 3J). The acid-lability of HKUST-1 in aqueous environments and the pH-responsive dissolution of ZIF-8 are well documented; higher local temperature further accelerates ligand exchange and hydrolysis, providing a mechanistic basis for the observed thermo-chemical coupling [45].

The robust, programmable heating observed across dispersions and hydrogels is consistent with the broadband NIR absorption and photothermal conversion of Ti₃C₂T_x MXene. Reviews and applications in infection control and wound repair emphasize MXene's high extinction coefficients and stable cycling behavior, which support its use as a repeatable heat source inside moist dressings. The present results extend these findings to an interface-engineered heterostructure in which the MXene framework and Cu-MOF domains remain effective after encapsulation in GelMA [46].

A key functional readout of the heterostructure is the NIR-amplified Cu²⁺ release. In acidic microenvironments that mimic infected niches, HKUST-1 and related Cu-MOFs undergo faster hydrolysis, while ZIF-8 frameworks dissolve under mildly acidic conditions; superimposing NIR heating increases the release rate by accelerating bond cleavage and diffusion. Recent studies have directly leveraged NIR to destabilize MOF carriers and program release kinetics, supporting the interpretation that photothermal energy can be converted into a gated chemical output at the MXene–MOF interface [47].

The observation that GelMA alone exhibits negligible NIR heating, whereas MXene@Cu-MOF/GelMA heats effectively, matches prior reports of NIR-responsive hydrogels where the polymer matrix primarily serves as a transparent, conformal scaffold. This is advantageous for clinical translation because heating can be spatially confined to nanofiller-rich regions while preserving the bulk hydrogel's integrity and moisture retention. The cycling stability demonstrated here parallels other photothermal hydrogel systems developed for antibacterial therapy [46].

MXene-only photothermal dressings deliver reliable heating but lack an on-demand chemical antibacterial component. Cu-MOF systems provide pH-responsive copper but may release passively without external control. By integrating Cu-MOF onto MXene and embedding the heterostructure in GelMA, the present platform realizes a controllable photothermal source and a heat-amplified copper reservoir within a single construct, consistent with emerging MOF–photothermal “synergy” paradigms reported for infection control and related indications [43].

3.5. In vitro antibacterial effects

Under standardized 808-nm NIR irradiation, PBS and GelMA controls exhibited abundant colony formation on agar and predominantly green Live/Dead fluorescence, indicative of high viability. MXene@Cu-MOF reduced colony coverage and shifted Live/Dead readouts toward reduced viability. The MXene@Cu-MOF/GelMA group showed the most pronounced effect, with sparse colonies and predominance of non-viable staining; quantitative analysis confirmed significant decreases in colony area and viable fraction for both E. coli and S. aureus relative to controls (Fig. 4A–D). Representative SEM images revealed intact, smooth cell envelopes in control groups, whereas MXene@Cu-MOF/GelMA plus NIR produced envelope wrinkling, pitting, and partial collapse in both species, morphological hallmarks of membrane compromise and thermal stress (Fig. 4E).

Fig. 4.

In vitro antibacterial activity of the MXene@Cu-MOF/GelMA platform under NIR irradiation against E. coli and S. aureus. (A) Representative agar plates of E. coli colonies and corresponding Live/Dead fluorescence images after treatment with PBS (blank), GelMA, MXene@Cu-MOF, and MXene@Cu-MOF/GelMA. (B) Quantification of E. coli colony area (plates) and bacterial viability (Live/Dead assay). (C) Representative agar plates of S. aureus colonies and corresponding Live/Dead staining under the same treatments as in (A). (D) Quantification of S. aureus colony area and bacterial viability. (E) Representative SEM images of E. coli and S. aureus after the indicated treatments, red arrows indicate damaged bacteria. Data are presented as mean ± SD (n = 3); ∗∗∗∗P < 0.0001; ns, not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The data support a thermo-chemical antibacterial mechanism in which MXene's broadband NIR absorption produces localized hyperthermia while the Cu-MOF component supplies bioactive copper. Hyperthermia disrupts membranes, denatures proteins, and sensitizes bacteria; copper ions further damage membranes, catalyze redox cycling and ROS production, and interact with nucleic acids and essential enzymes. The combination enhances lethality against both E. coli and S. aureus, as reflected in plate counts, Live/Dead staining, and SEM. Recent reviews and studies emphasize that photothermal materials can rapidly eradicate bacteria upon NIR exposure, and that copper's multifaceted mechanisms provide robust action against diverse pathogens [48].

Contemporary antibacterial strategies increasingly integrate photothermal heating with Cu-based chemistries or Cu -containing nanocomposites to achieve synergistic effects. Cu -oxide or Cu -chalcogenide systems activated by NIR display enhanced antibacterial efficacy, and photothermal hydrogels embedding Cu species demonstrate amplified killing compared with either modality alone. The present MXene@Cu-MOF/GelMA results align with this literature and add an interface-engineered route to combine a strong NIR transducer (MXene) with a pH-labile Cu reservoir (Cu-MOF) in a single hydrogel chassis [49].

Differences in envelope architecture can modulate stress responses, yet both Gram-negative E. coli and Gram-positive S. aureus were susceptible to MXene@Cu-MOF/GelMA under NIR, in line with reports that photothermal hydrogels and copper-based agents act across species. This breadth is advantageous for infected diabetic wounds, where polymicrobial communities predominate is common [50].

While Fig. 4 focuses on planktonic assays, recent work clarifies that photothermal heating increases bacterial susceptibility; copper further disrupts redox homeostasis within bacterial. The robust plate, Live/Dead, and SEM outcomes here therefore establish a physical–chemical foundation for subsequent antibacterial experiments with the same platform [48].

3.6. Endothelial proliferation, migration, and network formation in vitro

HUVECs exposed to MXene@Cu-MOF exhibited higher EdU labeling than PBS and GelMA controls under otherwise identical culture and NIR conditions (Fig. 5A and B). Embedding the MXene@Cu-MOF heterostructure in GelMA maintained the percentage of EdU⁺ nuclei, indicating that the hydrogel chassis does not impede, and may help sustain, the pro-proliferative stimulus. In monolayer scratch assays, PBS and GelMA groups displayed baseline closure over time. MXene@Cu-MOF and MXene@Cu-MOF/GelMA accelerated wound gap reduction (Fig. 5C and D). On growth-factor-reduced Matrigel, MXene@Cu-MOF and MXene@Cu-MOF/GelMA groups increased capillary-like network complexity relative to PBS and GelMA, as seen by greater numbers of branch points and longer total tube length (Fig. 5E–G). RT–qPCR revealed up-regulation of VEGF, eNOS, HIF-1α, and FGF2 in HUVECs treated with MXene@Cu-MOF, and MXene@Cu-MOF/GelMA groups, showing the largest increases compared with controls. Normalization to housekeeping genes and to the Blank group was applied uniformly (Fig. 5H–K). The selected targets encompass growth factor production (VEGF, FGF2), nitric-oxide signaling (eNOS), and hypoxia-responsive transcriptional control (HIF-1α), which collectively drive endothelial proliferation, migration, and tubulogenesis.

Fig. 5.

In vitro angiogenic evaluation of the MXene@Cu-MOF/GelMA platform under NIR irradiation. (A) EdU staining of HUVECs and (B) quantification of the percentage of EdU⁺ nuclei in the following groups: Blank (PBS), GelMA, MXene@Cu-MOF, and MXene@Cu-MOF/GelMA. (C) Scratch (wound-healing) assay of HUVEC monolayers and (D) quantification of wound closure (%) at the indicated time points. (E–G) Matrigel tube-formation assay: representative images (E) with quantification of branch points (F) and total tube length (G). (H–K) RT–qPCR analysis of angiogenesis-related genes (HIF-1α, FGF2, eNOS, and VEGF) in HUVECs. Data are presented as mean ± SD (n = 3); ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001; ns, not significant.

Cu is a well-recognized modulator of angiogenesis in vitro and in vivo, acting through stabilization and transcriptional activation of HIF-1α and subsequent VEGF up-regulation, as well as through pathways that converge on eNOS-mediated nitric-oxide production. Recent reviews and translational studies in wound care report that copper-releasing dressings enhance granulation and neovascularization, and that copper-containing biomaterials promote endothelial outgrowth in multiple models. The pattern observed here—enhanced EdU incorporation, accelerated migration, increased tube metrics, and elevated VEGF/eNOS/HIF-1α/FGF2—is therefore coherent with copper-dependent pro-angiogenic signaling [51]. Mechanistically, copper has been shown to stabilize HIF-1α protein and activate downstream VEGF expression in endothelial and tumor contexts, lending molecular plausibility to the gene expression changes detected in HUVECs [52].

In this platform, Cu-MOF is in intimate contact with a photothermal MXene framework and is encapsulated in a hydrophilic GelMA network. Prior results (Fig. 3) established that NIR irradiation increases copper release from the heterostructure; such NIR-amplified availability of Cu²⁺ during the assay window likely accounts for the stronger pro-angiogenic readouts in the MXene@Cu-MOF/GelMA group. This interpretation is consistent with studies showing that light-responsive or photothermal composites can program on-demand ion or cargo release, and with reports that Cu(II)@MXene hydrogels accelerate endothelialization and vascular ingrowth in infected wound models [53,54].

Under controlled NIR parameters that avoid cytotoxic temperatures, mild photothermal stimulation can promote endothelial repair, increase local perfusion in vivo, and engage heat-shock responses that favor vascular remodeling. While the present experiments are in vitro, the same irradiation protocol used to trigger copper release may also provide a permissive thermal microenvironment that facilitates migration and tubulogenesis, as supported by recent analyses of mild photothermal therapy in regenerative contexts [55].

MXene-based hydrogels have been explored mainly for photothermal antibacterial action, with emerging evidence of pro-angiogenic benefits when combined with bioactive cues. Copper-containing hydrogels independently promote angiogenesis but often lack external programmability. The present interface-engineered MXene@Cu-MOF/GelMA combines a strong NIR transducer with a pH- and heat-responsive copper reservoir in a clinically familiar hydrogel, aligning with recent directions that emphasize multifunctional, on-demand platforms for vascularized wound repair [42].

3.7. In vivo wound closure

One week after model induction, the fasting blood glucose level of diabetic mice was 21.8 ± 2.4 mmol/L (n = 30), which was significantly higher than that of non-diabetic control mice (5.1 ± 0.7 mmol/L, n = 6; P < 0.001), confirming the successful establishment of the diabetic model.

In a bacteria-infected diabetic wound model, PBS and GelMA controls showed gradual reduction of wound area over two weeks, whereas MXene/GelMA and Cu-MOF/GelMA accelerated closure. The MXene@Cu-MOF/GelMA group exhibited the most pronounced contraction at days 5, 9, and 14 (Fig. 6A). Planimetric overlays captured a steeper healing trajectory for MXene@Cu-MOF/GelMA relative to all comparators, indicating faster progression out of the inflammatory phase toward tissue formation (Fig. 6B).

Fig. 6.

Therapeutic efficacy of the MXene@Cu-MOF/GelMA platform in a bacteria-infected diabetic wound model. (A) Representative macroscopic images of wounds at days 5, 9, and 14 for the following groups: PBS (Control), GelMA, MXene/GelMA, Cu-MOF/GelMA, and MXene@Cu-MOF/GelMA. Wound areas were traced under identical imaging conditions. (B) Wound-healing trajectories compiled from serial planimetry, showing the overlay of wound boundaries over time. (C) H&E staining of wound sections on days 9 and 14, highlighting re-epithelialization and granulation tissue formation. (D) Quantification of epithelial thickness for the groups in (C), measured at comparable distances from the wound center. (E) Masson's trichrome staining on days 9 and 14 to visualize collagen deposition and extracellular-matrix remodeling. (F) Image-analysis quantification of collagen area fraction (%) for the sections in (E). (G) IHC for collagen III on days 9 and 14 to assess early matrix deposition. (H) Quantification of collagen III-positive area fraction (%) corresponding to (G). (I) IHC for collagen I on days 9 and 14 to evaluate mature matrix formation. (J) Quantification of collagen I-positive area fraction (%) corresponding to (I). Data are presented as mean ± SD (n = 6); ∗P < 0.05.

H&E sections at days 9 and 14 revealed more continuous epithelial tongues and thicker neo-epidermis in MXene@Cu-MOF/GelMA than in controls, with MXene/GelMA and Cu-MOF/GelMA showing intermediate improvement (Fig. 6C). Quantification of epithelial thickness corroborated these observations (Fig. 6D). Enhanced re-epithelialization at this window is typical of effective pro-healing interventions in diabetic models.

Masson's trichrome staining showed greater collagen-positive area fractions in MXene@Cu-MOF/GelMA at days 9 and 14 (Fig. 6E and F), and IHC images indicated higher collagen III at day 9 with a shift toward collagen I by day 14, relative to controls (Fig. 6G–J).

FISH imaging at days 3 and 6 showed dense bacterial signals in PBS and GelMA, reduced signals in MXene/GelMA and Cu-MOF/GelMA, and the lowest burden in MXene@Cu-MOF/GelMA (Fig. 7A). Field-wise counts confirmed the trend (Fig. 7B). DHE staining at days 9 and 14 indicated high superoxide-linked fluorescence in PBS and GelMA, reduced signal in MXene/GelMA and Cu-MOF/GelMA, and the lowest levels in MXene@Cu-MOF/GelMA (Fig. 7C). Quantified mean fluorescence supported the images (Fig. 7D). CD206 immunofluorescence (M2-associated) increased in MXene@Cu-MOF/GelMA relative to controls at days 9 and 14, while CD68⁺ cell counts (pan-macrophage) were reduced or normalized toward homeostatic levels compared with inflamed controls (Fig. 7E–H). This profile suggests a shift toward a pro-resolving phenotype rather than simple depletion. CD31 staining revealed more numerous microvessels in MXene@Cu-MOF/GelMA than in controls at days 9 and 14, consistent with enhanced angiogenesis (Fig. 7I and J).

Fig. 7.

In vivo assessment of antibacterial, antioxidant, macrophage-polarization, and pro-angiogenic effects in an infected diabetic-wound model. (A) Representative FISH images of bacterial burden in wound sections from PBS (Control), GelMA, MXene/GelMA, Cu-MOF/GelMA, and MXene@Cu-MOF/GelMA groups on days 3 and 6; nuclei were counterstained with DAPI. (B) Quantification of bacterial load per field from (A), obtained by automated segmentation of FISH-positive signals (values averaged over multiple fields per animal). (C) DHE fluorescence staining of superoxide in wound tissue on days 9 and 14 for the indicated groups, with DAPI nuclear counterstain. (D) Mean fluorescence intensity (MFI) of DHE from (C), normalized as specified in the panel. (E) Immunofluorescence staining for CD206 (M2-associated macrophages) on days 9 and 14. (F) Quantification of CD206⁺ cells per field corresponding to (E). (G) Immunofluorescence staining for CD68 (pan-macrophage marker) on days 9 and 14. (H) Quantification of CD68⁺ cells per field corresponding to (G). (I) Immunofluorescence staining for CD31 to visualize microvessels on days 9 and 14. (J) Quantification of vessel numbers per field corresponding to (I). Data are presented as mean ± SD (n = 6); ∗P < 0.05.

In addition, H&E staining of major organs (heart, liver, spleen, lung, and kidney) did not reveal apparent pathological alterations in the MXene@Cu-MOF/GelMA-treated groups compared with controls, suggesting that no overt systemic toxicity occurred under the current dosing and treatment schedule (Fig. S13).

PCA of variance-stabilized counts showed clear segregation between experimental (EXP; MXene@Cu-MOF/GelMA-treated) and control (CON; GelMA-treated) wounds (Fig. 8A), and volcano plus heatmap views identified DEGs meeting adjusted significance cutoffs (Fig. 8B and C). Up-regulated programs in EXP included antibacterial defense, controlled inflammatory regulation, angiogenesis, and extracellular-matrix organization (Fig. 8D).

Fig. 8.

Whole-transcriptome RNA-seq analysis of repaired skin tissue at 14 days. (A) PCA of global transcriptomic profiles for CON and EXP groups. (B) Volcano plot of DEGs between EXP and CON; significance defined as |log₂ fold change| ≥ 1 (fold change ≥2) and Benjamini–Hochberg–adjusted P < 0.05. (C) Heatmap of significant DEGs across samples, illustrating relative expression patterns between groups. (D) Relative expression of representative genes associated with antibacterial defense, inflammatory regulation, angiogenesis, and antioxidant in CON vs. EXP. (E) GO and KEGG pathway enrichment analyses of upregulated genes associated with antibacterial defense, inflammatory regulation, angiogenesis, and antioxidant in EXP vs. CON (top enriched terms shown). (F) GSEA enrichment plots for selected gene sets related to antibacterial/immune defense, inflammatory response, angiogenesis, and antioxidant, showing positive enrichment in EXP relative to CON. Data are presented as mean ± SD (n = 3); ∗P < 0.05.

GO/KEGG over-representation (Fig. 8E) and GSEA (Fig. 8F) showed positive enrichment of immune-defense, angiogenic, and antioxidant pathways in EXP relative to control, matching the histologic and immunofluorescence readouts. While MXene@Cu-MOF/GelMA treatment upregulated pathways associated with anti-inflammatory, angiogenesis, and antioxidant effects, we did not observe significant enrichment of gene signatures indicative of pathological fibrosis, uncontrolled angiogenesis, or other adverse remodeling processes in the RNA-seq data at this time point. In diabetic wounds, transcriptomic surveys often show the opposite pattern—sustained inflammation with inadequate vascular and matrix programs—highlighting the functional significance of the EXP shift.

The strongest antibacterial effect occurred in the MXene@Cu-MOF/GelMA group, consistent with the platform's ability to combine photothermal heating from MXene with pH- and heat-responsive copper release from Cu-MOF. Photothermal dressings weaken microbial membranes, and copper ions exert multi-target toxicity while minimizing resistance; clinical and preclinical literature on copper-impregnated dressings reports faster closure with improved infection control. The FISH reduction at early time points and lower DHE signals later support a scenario in which early bacterial burden and oxidative stress are both curtailed [42].

MXene-containing composites can dampen ROS, indirectly reducing pro-inflammatory signaling, while controlled copper dosing has been reported to bias macrophages toward reparative phenotypes in wound settings. The increased CD206⁺ cells with normalized CD68⁺ counts in MXene@Cu-MOF/GelMA are therefore biologically plausible and in line with studies demonstrating the necessity of M2-skewed macrophages for efficient resolution and tissue repair [56].

The CD31 increase, together with transcriptomic enrichment of angiogenic pathways, is consistent with copper's role in stabilizing or potentiating HIF-1α/VEGF signaling and supporting endothelial function; copper-releasing wound dressings have shown enhanced neovascularization and faster closure. The histologic transition from collagen III dominance toward collagen I by day 14, plus higher trichrome-positive area, indicates advanced matrix organization, paralleling reports that effective therapies accelerate collagen deposition and remodeling in this time window [51].

Embedding Cu-MOF directly on MXene shortens heat and diffusion distances so that NIR input not only provides hyperthermia but also amplifies copper flux [57]. GelMA offers a conformal, largely NIR-transparent matrix that maintains moisture while localizing therapy—consistent with reports that composite GelMA hydrogels deliver spatially confined heating and sustained ion or cargo release [58]. The convergence of early antibacterial and antioxidant control with later angiogenic and matrix programs explains the steeper healing trajectory seen macroscopically [42].

FISH visualizes intact microbes organization in tissue where culture may underreport burden; DHE requires attention to controls and imaging parameters but remains a validated readout of superoxide-linked stress in-situ; CD31, CD68, and CD206 are widely accepted markers for vessels and macrophage states in wound tissue. Using multiple orthogonal readouts—microbiology, redox, immune polarization, vascularization, histology, and transcriptomics—reduces the chance of single-assay bias and aligns with current recommendations for comprehensive wound-healing assessment [59].

Compared with previously reported MOF-based and photothermal dressings, which typically focus on either antibacterial activity or single-stage modulation of the wound microenvironment, our interface-engineered MXene@Cu-MOF/GelMA platform provides a more integrated and cascade-optimized strategy for infected diabetic wounds. MOF-based hydrogels have shown promising antibacterial effects through metal ion release, and photothermal composites can efficiently eradicate planktonic bacteria under NIR irradiation; however, they often lack coordinated regulation of oxidative stress, inflammation, and angiogenesis within one system. In contrast, the MXene@Cu-MOF heterostructure couples MXene-mediated ROS scavenging and photothermal heating with pH-responsive Cu²⁺ release, enabling early antioxidation/immunomodulation, mid-stage photothermal sterilization, and late-stage Cu-driven pro-angiogenesis in a single hydrogel dressing. In addition, embedding this heterostructure into a mechanically and adhesively suitable GelMA matrix, together with multi-level validation (macrophage polarization, endothelial function, and infected diabetic wound healing), highlights the translational potential of this platform relative to state-of-the-art MOF-based or photothermal wound dressings.

Beyond the promising preclinical efficacy, several translational challenges should be acknowledged. First, although the solution-based etching/delamination of Ti₃C₂Tx and in-situ Cu-MOF growth are, in principle, scalable, clinical translation will require further optimization of batch-to-batch uniformity, process robustness, and standardized sterilization/packaging of the hydrogel dressings. Second, our current 14-day window and organ histology suggest no obvious systemic toxicity, but the long-term fate and biosafety of MXene fragments and Cu-containing degradation products (including chronic exposure, tissue retention, and clearance) must be evaluated in extended follow-up and biodistribution studies. Third, while NIR light sources are increasingly available in clinical settings and we used moderate irradiation parameters to avoid overheating, the practical implementation of photothermal therapy in real wounds will need careful consideration of irradiation area, penetration depth, treatment frequency, and integration with standard wound-care workflows, potentially in conjunction with miniaturized or wearable NIR devices. Unlike many MOF-based or photothermal dressings that target only one or two phases of wound healing, MXene@Cu-MOF/GelMA integrates early antioxidation and immunomodulation, on-demand photothermal-chemical antibacterial activity, and late-stage copper-driven angiogenesis within a single, interface-engineered platform.

4. Conclusion

By integrating MXene and Cu-MOF into a cohesive 2D/3D nano-architecture and translating that architecture into a hydrogel dressing, this study advances beyond co-loaded blends to establish true modality coupling: photothermal heating actively drives therapeutic ion generation instead of merely coexisting with it. The platform delivers timed, on-demand actions matched to disease stage while maintaining material simplicity at the point of care (a single nanoplatform within a standardizable hydrogel). Given the prevalence of infected diabetic wounds and the limitations of current dressings in controlling bacterial and stimulating angiogenesis, an interface-engineered, cascaded system with programmable intensity and a self-regulated messenger offers clear translational promise. More broadly, the heterostructure concept presented here provides a blueprint for converting physical “mixtures” into functional partnerships in which energy transduction at designed interfaces programs chemical release for precision wound therapy.

CRediT authorship contribution statement

Qianming Li: Data curation, Conceptualization. Jianxiang Zhu: Formal analysis, Data curation. Jiawei Mei: Investigation, Formal analysis. Qiong Li: Project administration, Methodology. Fanyu Meng: Software, Resources. Xianfei Xie: Software, Resources. Lin Tao: Supervision. Fuqian Lei: Data curation, Formal analysis. Xiangyang Xu: Writing – original draft, Visualization. Ming Ni: Project administration, Methodology. Quan Liu: Investigation, Funding acquisition. Tao Yu: Writing – review & editing, Writing – original draft.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is/are the supplementary data to this article:

Data availability

Data will be made available on request.