A near-infrared regulated programmable multi-mode periosteum scaffold for sequential healing of infected bone defects

Abstract

Infected bone defects present a major clinical challenge, requiring precise sequential therapy that transitions from antibacterial activity to bone regeneration. Piezoelectric materials can transduce external stimulation into bioelectrical cues, providing a promising controllable handle to regulate antibacterial and osteogenic processes. However, conventional piezoelectric scaffolds often lack the capacity to distinctly separate these multifunctional roles, making it difficult to meet the therapeutic needs of different stages in the treatment of infected bone defects. Here, we develop a programmable NIR-responsive periosteum scaffold featuring a Janus bilayer architecture, in which a PDA-rich photothermal side and a non-photothermal piezoelectric side enable spatially separated antibacterial and osteogenic functions. This system integrates a thermoresponsive hydrogel with a piezoelectric polyvinylidene fluoride/barium titanate (PVDF/BT) electrospun membrane, achieving three switchable functional modes: continuous NIR irradiation for antibacterial therapy, intermittent irradiation for immunomodulation, and no irradiation for mechanical support and Ca²⁺ release to facilitate mineralization. In vitro studies demonstrated effective antibacterial efficacy and enhanced pro-healing M2 macrophage polarization, while comprehensive in vivo studies in rat models of 8 mm infected bone defect models achieved over 60% bone healing rate through programmable sequential therapy at 4 weeks. This programmable multi-mode periosteum scaffold provides a promising strategy for temporally orchestrating complex microenvironments, advancing the regenerative medicine for infected bone repair.

Graphical abstract

Highlights

• •A programmable NIR-responsive periosteum scaffold was developed to achieve sequential therapy of infected bone defects, integrating photothermal antibacterial activity and piezoelectric immune-osteogenic modulation.
• •Continuous NIR irradiation enabled rapid on-demand photothermal sterilization, achieving efficient MRSA clearance and establishing early infection control.
• •Intermittent NIR stimulation generated piezoelectric signals that promoted macrophage M2 polarization and angiogenic/osteogenic cytokine release, reconstructing a pro-regenerative microenvironment.
• •In vivo studies confirmed accelerated vascularized bone regeneration and superior defect healing compared with single-mode or non-programmed treatments.

1. Introduction

Infected bone defects remain among the most challenging complications in orthopedic surgery, where bacterial contamination severely disrupts natural healing processes. Persistent infection triggers excessive inflammation, degrades extracellular matrix, impairs osteoblast activity, and increases the likelihood of delayed union or nonunion [1,2]. Clinically, treating infected bone defects requires not only effective bacterial eradication but also precise modulation of the local immune environment to transition toward tissue repair [[2], [3], [4], [5], [6]]. However, these two therapeutic goals are inherently conflicting: aggressive antibacterial interventions often induce cytotoxicity and inflammation, while premature regenerative stimulation risks exacerbating infection. Therefore, the key unmet clinical need lies in developing therapeutic strategies that can achieve infection elimination first and then guide a controlled transition into immune resolution and osteogenesis [7].

To overcome these obstacles, various biomaterials have been explored, including antibiotic-loaded scaffolds, photothermal agents, bioactive ions, and composite hydrogels [[8], [9], [10], [11], [12], [13], [14]]. These strategies enable antibacterial effects, angiogenesis, and osteogenic stimulation to varying degrees. Nevertheless, most of them operate in a single-function or simultaneous-action mode, lacking the capacity to align therapeutic cues with the dynamic pathological stages of infected bone healing [[15], [16], [17], [18]]. Infection eradication, immune modulation, angiogenesis, and bone regeneration do not occur concurrently, but rather follow a time-dependent sequence [19,20]. Simultaneous stimulation can lead to antagonistic effects—for example, antibacterial agents interfering with osteoblast viability or regeneration signals diminishing necessary inflammatory responses during infection control [[21], [22], [23]]. Hence, a staged and programmable treatment platform capable of delivering sequential therapy—early antibacterial activity followed by immunomodulation and late-stage osteogenesis—is urgently needed.

Piezoelectric materials offer a compelling avenue to achieve such staged intervention due to their ability to convert external physical stimuli into electrical signals [24]. This property enables noninvasive, controllable activation that can modulate cell behaviors relevant to bone healing, including macrophage polarization, angiogenesis, and osteoblast differentiation [11,25,26]. Piezoelectricity also holds potential to generate reactive oxygen species or enhance antibacterial efficacy when coupled with additional stimuli [27]. However, conventional piezoelectric scaffolds often lack the capacity to distinctly separate antibacterial and osteogenic functions, as similar stimulation conditions may activate both simultaneously [28]. Furthermore, achieving reversible and switchable control over multiple therapeutic modes remains challenging, limiting their application for infected bone defects where staged interventions are essential. From our view, the primary barrier is not the intrinsic biological activity of piezoelectric materials, but the lack of programmable external stimuli and material architectures that enable precise sequential control.

Herein, we present a near-infrared (NIR) regulated, programmable periosteum-mimicking scaffold that enables staged antibacterial–immune–osteogenic therapy for infected bone defects. By integrating a thermoresponsive hydrogel with a bilayer PVDF/BT electrospun membrane, our system employs continuous NIR irradiation to induce photothermal antibacterial activity during the early infection phase and intermittent irradiation to generate piezoelectric signals for immune modulation and osteogenesis during later stages. As a result, three functional modes were achieved through different NIR irradiation strategies (Scheme 1): photothermal effects for antibacterial (continuous irradiation), electrical stimulation for inducing M2 macrophage polarization (intermittent irradiation) and mechanical support and Ca²⁺ ion release to facilitate mineralization (without irradiation). In vitro experiments, these three functional modes of periosteum scaffolds were characterized and validated separately. In rat models of methicillin-resistant Staphylococcus aureus (MRSA)-infected critical-sized bone defect, the scaffold achieved efficient infection clearance within one week and near-complete bone regeneration within four weeks. Table 1 summarizes recent advances in periosteum scaffolds with dual antibacterial and osteogenic capabilities, highlighting the absence of programmable sequential therapy in previous reports.

Scheme 1.

Schematic illustration of the mechanisms of NIR-controlled sequential treatment for infected bone repair.

Table 1.

Comparison of different antibacterial and osteogenic dual-functional scaffolds.

Anti-bacterial effects	Enhancement of osteogenesis	Control method	Sequence control	Reference
Photothermal treatment	Magnesium-modified black phosphorus	✕	✕	[59]
Quaternized chitosan	Nano-hydroxyapatite	✕	✕	[60]
Piezoelectric ES	Piezoelectric ES	Ultrasound	✕	[61]
Self-promoted ES	Self-promoted ES	✕	✕	[62]
Bioactive ions release	Bioactive ions release	✕	✕	[63]
Magnetic hyperthermia	Piezoelectric ES	magnetic fields/Ultrasound	✕	[64]
Photothermal treatment/Ag Ions release	Mild photothermal treatment	NIR	✕	[65]
Photothermal treatment	Piezoelectric ES	NIR	✓	This work

2. Experimental section/methods

2.1. Materials

Poly(vinylidene fluoride) (PVDF, average Mw ≈ 530,000), barium titanate (BaTiO₃, BT) nanoparticles, dopamine hydrochloride (≥98%), N,N-dimethylformamide (DMF, anhydrous, ≥99.8%), acetone (≥99.5%), Tris-HCl buffer, ammonium persulfate (APS), N,N,N′,N'-tetramethylethylenediamine (TEMED), N-isopropylacrylamide (NIPAm), acrylamide (AM), N,N′-methylenebisacrylamide (MBAA), sodium alginate, calcium chloride (CaCl₂), and polydimethylsiloxane (PDMS, Sylgard 184) were purchased from Sigma–Aldrich (USA). Polyacrylonitrile (PAN) was obtained from Aladdin (China). Indium tin oxide (ITO) conductive films were purchased from Zhuhai Kaivo Optoelectronic Technology Co., Ltd. (China). Silver paste was obtained from Ted Pella Inc. (USA).Dulbecco's Modified Eagle Medium (DMEM), α-Minimum Essential Medium (α-MEM), fetal bovine serum (FBS), penicillin–streptomycin solution, and TRIzol reagent were purchased from Invitrogen (USA). Cell Counting Kit-8 (CCK-8) was obtained from Dojindo (Japan). Calcein-AM, propidium iodide (PI), annexin V-FITC/PI apoptosis detection kit, and BCA protein assay kit were purchased from Beyotime (China). Primary antibodies against CD86, CD206, Runx2, osteocalcin (OCN), and corresponding fluorescent secondary antibodies were purchased from Abcam (UK) or Proteintech (China), as detailed in the respective experimental sections. All reagents were used as received without further purification. RAW264.7 murine macrophages and bone marrow–derived mesenchymal stem cells (BMSCs) were obtained from Pricella Biotechnology Co., Ltd. (Wuhan, China). Methicillin-resistant Staphylococcus aureus (MRSA, ATCC 43300) and Escherichia coli (ATCC 25922) were purchased from the American Type Culture Collection (ATCC).

2.2. Preparation of BT/polydopamine (PDA) nanoparticles

Barium titanate (BT) nanoparticles (0.5 g) were dispersed in 50 mL of Tris-HCl buffer (pH 8.5) and sonicated for 2 h to ensure homogeneous suspension. Subsequently, dopamine hydrochloride (0.25 g) was added to the dispersion under constant stirring, and the reaction was allowed to proceed for 12 h in the dark to facilitate dopamine self-polymerization. The resulting dark-colored suspension was collected by centrifugation and washed several times with deionized water to remove residual reactants. Finally, the BT/PDA nanoparticles were obtained by freeze-drying.

2.3. Preparation of anisotropic bilayer PVDF/BT@ PVDF/BT/PDA electrospun membrane

2g of the PVDF powder and 0.5g of BT nanoparticles were dissolved in a 10 mL of 6:4 mixture of N, N–Dimethylformamide (DMF, anhydrous, ≥99.9%) and acetone (v/v) under strong stirring. The well-dispersed solution was placed into a 10 ml syringe and inserted into a syringe pump from KD Scientific Inc., MA. The flow rate of the solution was adjusted to 1.8 ml/h and delivered through a flat-tipped 20-gauge needle (SAI Infusion Technologies, CA) while applying a voltage of 15 kV. Subsequently, the polarized solution was carefully deposited onto aluminum foil, which was wrapped around a grounded aluminum drum rotating at a speed of 1500 rpm (rotations per minute). The entire electrospinning process was carried out in an atmosphere with a relative humidity of 70 ± 15% at ambient temperature. Furthermore, 2g of the PVDF powder and 0.5g of BT/PDA nanoparticles were dissolved in a 10 mL of 6:4 mixture of DMF and acetone (v/v) under strong stirring and followed by the same electrospinning step onto previous membrane. As a control group, the PAN@PAN/PDA electrospun membrane was fabricated using the same process, but polyacrylonitrile (PAN) was used to replace PVDF. Additionally, PDA powder was added directly in place of BT nanoparticles.

2.4. Synthesis of double network (DN) hydrogel encapsulated membrane scaffold

2 g of sodium alginate and 0.1 g of calcium chloride were completely dissolved in 100 mL of deionized water while vigorously stirring at 60 °C. Once a clear solution was obtained, 2 g of N-Isopropylacrylamide and acrylamide with varying mass ratios were dissolved into 10 mL of the prepared solution. To get the precursor solution, 0.05 g of initiator ammonium persulfate and 0.01 g of crosslinking agent N,N′-Methylenebisacrylamide were further dissolved. Then, the precursor solution was dropwise added onto the electrospun membrane, and the solution was thoroughly filled into the pore of the electrospun membrane by using vacuum suction. Finally, 5ul of accelerator TEMED (N,N,N′,N′-Tetramethylethylenediamine) was added onto its surface and spread evenly. A glass slide was then placed on top to obtain a flat surface. The complete polymerization was achieved overnight. As control group, single network (SN) hydrogel was prepared without adding sodium alginate.

2.5. Measurement of the piezoelectric properties of electrospun membranes

Electrospun membranes were sandwiched between two layers of aluminum foil, serving as electrodes. The assembled structure was then encapsulated with polyimide tape to prevent air exposure and maintain electrode stability. The open-circuit voltage generated by the device was recorded using a digital oscilloscope (DPO3014, Tektronix, CA, USA). Mechanical stimulation was applied to the device using a universal testing machine (Instron, MA, USA) at a compression speed of 10 mm/s to assess its piezoelectric response.

2.6. Photothermal effect of DN hydrogel/electrospun membrane

The photothermal performance of the DN hydrogel/electrospun membrane was evaluated under NIR irradiation using a therapeutic diode laser system (9W Pilot®, A.R.C. Laser, UT, USA). During irradiation, the temperature changes of the samples were continuously monitored and recorded using a portable thermal imaging camera (KKmoon, China). All experiments were conducted at ambient temperature unless otherwise specified.

2.7. Measurement of the piezoelectric property of DN hydrogel/electrospun membrane under NIR irradiation

Transparent indium tin oxide (ITO) films were used as electrodes to sandwich both surfaces of the PVDF/BT@PVDF/BT/PDA electrospun membrane. Silver paste was applied to enhance adhesion between the ITO films and the electrospun membrane. The assembled structure was subsequently immersed in the hydrogel precursor solution, followed by a degassing process to eliminate air bubbles. After complete gelation, excess hydrogel was carefully removed to expose the device structure. Given the significantly lower electrical resistance of the hydrogel compared to the PVDF electrospun membrane, the generated current could be directly measured to assess piezoelectric performance. A precision source/measure unit (B2902A, Keysight Technologies, CA, USA) was employed to record the output current under mechanical stimulation. Compression was applied using a universal testing system (Instron, MA, USA) at a speed of 10 mm/s. Simultaneously, near-infrared (NIR) irradiation was provided using a therapeutic diode laser system (9W Pilot®, A.R.C. Laser, UT, USA) to trigger the thermoresponsive behavior of the scaffold.

2.8. Cell culture

RAW264.7 murine macrophage cell line and BMSCs were obtained from Pricella Biotechnology Co., Ltd. (Wuhan, China). RAW264.7 cells were maintained in high-glucose Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (P/S). BMSCs were cultured in α-Minimum Essential Medium (α-MEM) with identical supplementation. All cells were incubated at 37 °C in a humidified atmosphere containing 5% CO₂. For co-culture experiments, RAW264.7 cells were seeded onto programmable periosteum scaffolds and divided into four groups based on treatment conditions (n = 3): NC (negative control, no scaffold treatment); M (scaffold without irradiation); M + CI (scaffold with continuous NIR irradiation), and M + II (scaffold with intermittent NIR irradiation).

2.9. Biocompatibility tests

RAW264.7 macrophages and BMSCs (1 × 10⁴ cells/well) were respectively seeded in 96-well plates and cultured with periosteum scaffolds under four different NIR irradiation conditions (n = 3): NC (no treatment), M (scaffold without NIR), M + CI (continuous irradiation), and M + II (intermittent irradiation). Continuous NIR irradiation was applied for 30 min with the temperature maintained at approximately 42 °C. Intermittent irradiation was performed for 30 min, during which the temperature cyclically fluctuated between 35 and 42 °C. These irradiation parameters were consistently adopted in the subsequent antibacterial experiments, in vitro cell studies, and in vivo treatments. For cell viability assessment, after 1, 2, and 3 days of culture, cells were incubated with 10% Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto, Japan) solution for 2 h. Absorbance at 450 nm was recorded using a microplate reader (Synergy Neo2, BioTek Instruments, Winooski, VT, USA). For live/dead staining, RAW264.7 cells and BMSCs (1 × 10⁴ cells/well) were respectively cultured with scaffolds for 24 h, then stained with Calcein-AM (2 μM) and propidium iodide (PI, 1 μg/mL) (Beyotime, Shanghai, China). Fluorescence images were acquired using a confocal laser scanning microscope (SP8, Leica Microsystems, Wetzlar, Germany). For apoptosis analysis, RAW264.7 cells and BMSCs (1 × 10⁶ cells/well) were respectively treated with scaffolds for 24 h and subsequently stained with annexin V-FITC/PI apoptosis detection kit (Solarbio, Beijing, China). Apoptotic cells were quantified by flow cytometry (FACScalibur, BD Biosciences, San Jose, CA, USA).

2.10. RNA-sequencing (RNA-seq) and data analysis

RAW264.7 cells were divided into 4 groups (NC, M, M + CI and M + II, n = 3) according to the irradiation modes performed on RAW264.7 cells co-cultured with periosteum scaffolds system, and total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) after 48 h of culture. mRNA enrichment, fragmentation, reverse transcription, library construction and data analysis were performed by Oebiotech Co. Ltd. (Shanghai, China). |log2 (Fold change)| >1 and q < 0.05 (q-value is the corrected value of p-value) were used as the threshold criteria for screening differentially expressed genes (DEGs). Functional and signaling pathway enrichment analysis of DEGs was performed using GO and KEGG databases.

2.11. Macrophage polarization analysis

After 48 h of treatment, RAW264.7 cells were collected for macrophage polarization analysis. For flow cytometry, cells were stained with anti-CD86 (Proteintech, Cat# 98025-1-RR, Wuhan, China) and anti-CD206 (BioLegend, Cat# 141711, San Diego, CA, USA) antibodies. Fluorescence intensity was quantified using a FACScalibur flow cytometer (BD Biosciences, San Jose, CA, USA). For immunofluorescence staining, cells were fixed with 4% paraformaldehyde, incubated with anti-CD86 (Proteintech, Cat# 13395-1-AP) and anti-CD206 (Abcam, Cat# ab64693, Cambridge, UK) antibodies, and visualized by confocal laser scanning microscopy (SP8, Leica Microsystems, Wetzlar, Germany). For quantitative real-time PCR (RT-qPCR) analysis, total RNA was extracted using TRIzol reagent (Invitrogen), and reverse transcription was performed using a commercial kit (Toyobo, Osaka, Japan). RT-qPCR amplification was conducted on a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Primer sequences for M1-and M2-associated markers, including Nos2, Tnf, Il1b, Arg1, Il10 and Tgfb1, are listed in Table S1 (Supporting Information).

2.12. Osteogenic differentiation of BMSCs

Conditioned media (CM) from RAW264.7 cultures were collected and mixed with osteogenic induction medium (1:2 ratio) to treat BMSCs. After 7 days, alkaline phosphatase (ALP) activity was measured using a commercial kit (Beyotime, Shanghai, China), and mineralization was assessed via Alizarin Red S (ARS) staining at 21 days. Runx2 immunofluorescence was performed at day 3 using anti-Runx2 antibody (Proteintech Cat No. 82636-2-RR). RT-qPCR analyzed Runx2 (day 3) and Spp1 (day 7) expression.

2.13. Western blot analysis

Western blotting was performed to evaluate the protein expression of macrophage polarization and osteogenic differentiation markers. RAW264.7 cells were used to detect macrophage-related proteins, including IL-1β (Proteintech Cat No. 26048-1-AP), ARG1(Proteintech Cat No. 16001-1-AP), iNOS (Proteintech Cat No. 22226-1-AP), and TGF-β (Proteintech Cat No.80805-3-RR). BMSCs were used to detect osteogenic markers, including RUNX2 (Proteintech Cat No. 82636-2-RR) and OPN (Proteintech Cat No.22952-1-AP). After treatment, total cellular proteins were extracted using RIPA lysis buffer, separated by SDS-PAGE, and transferred onto PVDF membranes. The membranes were incubated with primary antibodies overnight at 4 °C followed by appropriate HRP-conjugated secondary antibodies. Protein bands were visualized using an enhanced chemiluminescence system.

2.14. In vitro antibacterial assays

Methicillin-resistant Staphylococcus aureus (MRSA, ATCC 43300) and Escherichia coli (E. coli, ATCC 25922) were cultured in Luria–Bertani (LB) broth. Bacterial suspensions (1 × 10⁶ CFU/mL) were co-incubated with periosteum scaffolds under four different conditions (n = 3): NC (no treatment), M (scaffold without irradiation), M + CI (scaffold with continuous NIR irradiation), and M + II (scaffold with intermittent NIR irradiation). Near-infrared (NIR) irradiation (808 nm, 0.8 W/cm²) was applied for 30 min (continuous mode) or five cycles of on/off switching (intermittent mode). During continuous NIR irradiation, the temperature of the surrounding solution was maintained at approximately 42–43 °C, as monitored by infrared thermal measurement. For antibacterial evaluation, treated bacterial suspensions were serially diluted, plated onto LB agar, and incubated overnight. Colony-forming units (CFUs) were subsequently counted to calculate bacterial survival rates. For morphological analysis, bacterial samples were fixed with 2.5% glutaraldehyde, dehydrated through graded ethanol series, and visualized using a field-emission scanning electron microscope (FE-SEM, S-4800, Hitachi, Tokyo, Japan). For viability assessment, live/dead staining was performed using SYTO9 (green, live) and propidium iodide (PI, red, dead) dyes (Thermo Fisher Scientific, Waltham, MA, USA), and fluorescence images were acquired using confocal laser scanning microscopy. For biofilm viability evaluation, bacterial biofilms formed on scaffold surfaces were similarly stained and analyzed via confocal microscopy. For protein leakage assays, bacterial suspensions were centrifuged after treatment, and supernatant protein concentrations were determined using a BCA protein assay kit (Beyotime, Shanghai, China) according to the manufacturer's instructions.

2.15. In vitro and in vivo degradation

In vitro and in vivo degradation of the scaffolds was evaluated over 4 weeks. For the in vitro study, scaffolds were immersed in phosphate-buffered saline (PBS, pH 7.4) at 37 °C. For the in vivo study, scaffolds were implanted into rat calvarial defects. After 4 weeks, the samples were retrieved, gently rinsed with distilled water to remove residual tissue, and weighed. The degradation rate was determined by comparing the initial weight with the remaining weight after the degradation period.

2.16. In vivo infected bone defect regeneration

The animal experiment protocol was approved by the Institutional Animal Care and Use Committee (IACUC), Sun Yat-Sen University (SYSU-IACUC-2025-B0740) and were conducted in accordance with relevant guidelines and regulations. Critical-sized (8 mm) MRSA-infected skull defects were established in Sprague-Dawley rats. During surgery, the native periosteum covering the skull was carefully elevated and removed at the defect site prior to scaffold implantation to accommodate the engineered periosteum scaffold. Rats were randomly assigned into six groups (n = 3): NC (no treatment); M (periosteum scaffolds without irradiation); M + CI (scaffolds with continuous NIR irradiation, 30 min/day); M + II (scaffolds with intermittent NIR irradiation, 5 cycles/day); M + CI&II (combined continuous and intermittent irradiation); M + CIII (continuous irradiation for 1 week followed by intermittent irradiation). Before implantation, all scaffolds were sterilized by Co⁶⁰ gamma irradiation, and all surgical procedures were conducted under sterile conditions following standard aseptic protocols. The in vivo biosafety of the scaffolds was evaluated after implantation. Major organs, including the heart, liver, spleen, lung, and kidney, were harvested and fixed in 4% paraformaldehyde, followed by paraffin embedding, sectioning, and hematoxylin–eosin (H&E) staining for histological observation. In addition, blood samples were collected for routine hematological analysis to assess potential systemic toxicity. For bacterial clearance assessment, exudates from defect sites were collected at 1 week for colony-forming unit (CFU) assays. Bone regeneration was evaluated at 4 and 8 weeks post-surgery by harvesting femurs for micro-computed tomography (micro-CT) analysis (SIEMENS, Berlin, Germany), quantifying bone volume/total volume (BV/TV) and bone mineral density (BMD). Histological evaluation was performed using hematoxylin and eosin (H&E) and Masson's trichrome staining to assess tissue repair. For immunohistochemical (IHC) analysis, expression of CD206 (M2 macrophage marker) and osteocalcin (OCN, osteogenic marker; Abcam, Cat# ab93876, Cambridge, UK) was detected at 1 and 4 weeks, respectively, using specific primary antibodies.

2.17. Statistical analysis

Data are presented as mean ± Standard Deviation (SD). Statistical comparisons between groups were performed using one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. Analyses were conducted using OriginPro 2022 software (OriginLab Corporation, Northampton, MA, USA). The value of P < 0.05 was considered statistically significant.

3. Results and discussions

3.1. Fabrication and characterization of PVDF/BT and PVDF/BT/PDA electrospun membrane

Through the self-polymerization of dopamine, a uniform PDA coating was formed on the surface of the BT nanoparticles. Fig. S1 illustrated the morphology of BaTiO₃ (BT) and BT/PDA nanoparticles. After self-polymerization, the average diameter of nanoparticles increased from 52.13 ± 2.13 to 99.71 ± 3.63 nm, corresponding to an estimated PDA shell thickness of approximately 24 nm. To further confirm the successful modification, FTIR analysis was performed (Fig. S2). Compared with pristine BT, the BT/PDA sample exhibited additional absorption bands associated with catechol and amine-related functional groups derived from polydopamine. Later, BT and BT/PDA nanoparticles were introduced to the PVDF solution separately for further electrospinning. By fine-tuning the electrospinning parameters, bead-free and uniform nanofibers were produced shown in Fig. S3. To further verify the distribution of BT nanoparticles within the electrospun fibers, additional TEM imaging and corresponding EDS elemental mapping were performed (Fig. S4). The elemental maps of Ba and Ti demonstrate homogeneous distribution throughout the nanofibrous structure, confirming effective incorporation and dispersion of BT within the PVDF matrix. The structural characteristics of the electrospun ss were further analyzed via X-ray diffraction (XRD) (Fig. S5). For reference and comparison, the XRD spectrum of pristine BT nanoparticles is provided in Fig. S6. The presence of distinct BT diffraction peaks verified the successful incorporation of BT and BT/PDA nanoparticles into the PVDF matrix. Importantly, a new reflection at 20.84° was observed after electrospinning, indicative of a significant phase transition from the α-phase to the β-phase of PVDF. The formation of the β-phase is known to enhance piezoelectric performance. Moreover, the introduction of BT and BT/PDA nanoparticles did not disrupt the β-phase structure, preserving the piezoelectric properties of the membrane. The piezoelectric properties of the electrospun membranes were evaluated under periodic compressive loading. To investigate the pressure dependence of the electromechanical response, the output signals were measured under a range of compressive stresses (10, 20, 30, 40, and 50 kPa) (Fig. S7). The output voltage/current increased progressively with increasing applied pressure, demonstrating a typical pressure-dependent piezoelectric behavior. For clarity and consistency in the main experiments, 30 kPa was selected as an intermediate representative pressure for subsequent characterization (Fig. S8).

All samples exhibited typical piezoelectric responses, while the output performance varied depending on the incorporated fillers. Compared to the pure PVDF membrane, the addition of BT nanoparticles significantly increased the voltage output from 48.38 mV to 86.73 mV. When BT/PDA nanoparticles were further introduced, the voltage increased to 101.23 mV, consistent with previous reports [29]. These results collectively demonstrate that PDA-modified BT nanoparticles synergistically enhance both the β-phase formation and piezoelectric output of PVDF-based electrospun membranes.

3.2. Fabrication and characterization of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane

The bilayer PVDF/BT@PVDF/BT/PDA electrospun membrane with the Janus structure was fabricated via a two-step electrospinning process (Fig. 1A). A second layer of PVDF/BT/PDA nanofibers was directly deposited onto the pre-formed PVDF/BT membrane, forming a bilayer heterogeneous structure. Cross-sectional imaging (Fig. 1B) showed a distinct interface without discernible gaps, indicating strong interfacial bonding between the two layers. The prepared electrospun membrane was first soaked in a precursor solution with a single network of calcium-alginate, followed by the in-situ polymerization of N-isopropylacrylamide (NIPAM) and acrylamide (AM) to create a second network for the hydrogel. Post-polymerization, the hydrogel was uniformly filled with the fibrous scaffold, as evidenced by SEM imaging (Fig. 1C).

Fig. 1.

Fabrication and characterization of the double-network (DN) hydrogel/electrospun membrane. A. Schematic illustration depicts the fabrication of bilayer PVDF/BT@PVDF/BT/PDA electrospun membrane via a two-step electrospinning process, followed by the encapsulation of a DN hydrogel within the fibrous membrane. B. Cross view of bilayer PVDF/BT@PVDF/BT/PDA electrospun membrane C. Top view of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane. D. Absorbance change at a 550 nm wavelength in UV−vis spectra with increasing temperature from 20 to 45 °C. According to previous reports, the temperature at which the change of absorbance increases by 50% is considered the LCST [58]. E. Mechanical properties of different membranes F. Swelling performance of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membranes. G. The adhesion strength of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membranes to various substrates. H. Piezoelectric current output of the bilayer PVDF/BT@PVDF/BT/PDA fibrous membranes under cyclic 30 kPa pressure. I. Piezoelectric current output of the DN hydrogel/electrospun membrane under cyclic 30 kPa pressure. J. The corresponding thermal images of DN hydrogel/PVDF/BT/PDA electrospun membrane and DN hydrogel/PVDF/BT electrospun membrane under 808-nm laser irradiation (3W/cm²). K. NIR-induced bending of the DN hydrogel/electrospun membrane. The PDA-containing brown side increases in temperature under NIR causing the thermoresponsive hydrogel to contract and bend toward the brown side. L. Reproducibility and stability of NIR responsive deformation that after 10 cycles of repeated irradiation, the bending angle of the DN hydrogel/electrospun membrane remained consistent.

Poly N-isopropylacrylamide (PNIPAm) hydrogel, as the most well-known and extensively studied thermoresponsive hydrogel, undergoes a reversible phase transition around its lower critical solution temperature (LCST), typically close to 32 °C. Here, the copolymer poly(N-isopropylacrylamide-co-acrylamide) poly(NIPAm-co-AM) was employed, and its LCST can be adjusted [30]. Fig. 1D illustrated the LCST of poly(NIPAm-co-AM) hydrogel with different monomer ratios, where as the acrylamide content increased, there was a gradual elevation in the LCST. Considering the human body temperature is approximately 37 °C, we chose the 0.85:0.15 ratio, and LCST was 39 °C. The figure also indicated that the introduction of additional sodium alginate will not affect the LCST. Meanwhile, the integration of electrospun membrane with a hydrogel can significantly enhance the mechanical properties of the composite material owing to a remarkable synergistic effect [31]. Mechanical testing revealed that the DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane exhibited an exceptionally high Young's modulus (768.0 ± 71.2 kPa), substantially outperforming its individual components (∼40 kPa) (Fig. 1E). This enhancement is particularly advantageous for mimicking native bone tissue properties and supporting cell proliferation [32]. Moreover, owing to this composite structure, the membrane maintained low swelling ratios and stable macroscopic dimensions (Fig. 1F), which is crucial for structural integrity during bone repair. Additionally, Fig. 1G illustrated that this membrane demonstrated good tissue adhesive capability, effectively adhering to both the pig femur (15.93 kPa) and skin surfaces (8.07 kPa) which provided a valuable advantage for surgical operation, making it easier for doctors to implant [33].

3.3. The piezoelectric properties of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane

Direct voltage measurements of the DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane using an oscilloscope failed to yield clear piezoelectric voltage outputs under external pressure. This is primarily due to the high dielectric constant and ionic conductivity of the hydrogel matrix, which introduces significant capacitive and leakage effects [34,35]. As a result, the generated piezoelectric charges are rapidly dissipated or redistributed, and the oscilloscope detects signals that resemble capacitive charging behavior rather than distinct piezoelectric voltage pulses. To more accurately assess the piezoelectric response, current measurement was employed instead. Unlike voltage detection, current sensing captures the transient charge generation directly and is less influenced by capacitive attenuation, thereby providing a reliable method to evaluate the intrinsic piezoelectric performance of the composite membrane under mechanical stimulation [36]. To validate this programmable therapy, we characterized both thermal responses and piezoelectric outputs under different irradiation strategies. We first evaluated the piezoelectric response of the PVDF/BT@PVDF/BT/PDA membrane by measuring its output current, as shown in Fig. 1H. The periodic current pulses exhibited the same waveform trend as their corresponding voltage response, confirming that current sensing reliably reflects the piezoelectric behavior of the membrane. Clear and repeatable current peaks of approximately 2.8 nA were recorded under cyclic loading. To determine whether this piezoelectric effect could be maintained after integration with the DN hydrogel, cyclic compression tests were subsequently performed on the DN hydrogel/PVDF/BT@PVDF/BT/PDA composite membrane (Fig. 1I). A Polyacrylonitrile (PAN) electrospun membrane, which lacks intrinsic piezoelectricity, was included as a control to exclude any potential signal contributions from the hydrogel matrix itself. Unlike the DN hydrogel/PAN@PAN/PDA sample, which produced negligible output, the DN hydrogel/PVDF/BT@PVDF/BT/PDA membrane generated distinct and periodic current signals under the same mechanical loading conditions. Notably, a current of ∼0.35 nA was obtained at 30 kPa, confirming that the observed signals originated from deformation of the embedded PVDF/BT nanofibers.

Thus, these results verify that the composite membrane retains measurable piezoelectric performance even after incorporation of the DN hydrogel, highlighting its feasibility for further electrical therapy.

3.4. NIR controlled mechanical deformation of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane

Due to the strong photothermal properties of PDA, the PVDF/BT/PDA and the Janus structure design electrospun membrane exhibited significantly enhanced light-to-heat conversion compared with PVDF/BT electrospun membrane. As shown in Fig. 1J, under NIR irradiation, the DN hydrogel/PVDF/BT/PDA electrospun membrane rapidly increased in temperature, whereas the DN hydrogel/PVDF/BT electrospun membrane exhibited a negligible thermal response. Moreover, both the heating rate and the maximum temperature could be finely tuned by adjusting the NIR power (Fig. S9), a critical feature for clinical applications. For example, suitable heating can inhibit bacterial proliferation, offering significant advantages during the early stage of implantation [9,37,38].

The bilayer heterogeneous electrospun membrane design enabled the formation of an in-pane temperature gradient under NIR exposure. Consequently, localized heating induced contraction of the thermoresponsive poly(NIPAm-co-AM) hydrogel on the irradiated side, resulting in bending of the entire membrane. This behavior was experimentally validated (Fig. 1K): under NIR irradiation, the membrane consistently bent towards the PDA coated side (brown side), reaching an average bending angle of approximately 15°. To eliminate the possibility that the temperature difference was due to the brown surface's proximity to the NIR irradiation, the DN hydrogel/electrospun membrane was flipped, and the experiment was repeated. The results were consistent, showing the membrane bending toward the brown surface at an angle of approximately 15°. Furthermore, to demonstrate the mechanical stability of this NIR-induced deformation, repeated irradiation cycles were performed, demonstrating the robust and repeatable photothermally driven actuation of the DN hydrogel/electrospun membrane Fig. 1L and Video S1. To further understand the deformation mechanism, finite element simulations were conducted using COMSOL Multiphysics to analyze the thermo-mechanical response of the bilayer scaffold under NIR-induced temperature changes. The simulation results (Fig. S10) indicate that the photothermal heating generates a temperature gradient across the bilayer structure, leading to internal stress and bending deformation of the membrane. These results are consistent with the experimentally observed actuation behavior.

3.5. NIR-induced multi-mode of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane

An ideal scaffold for treating infected bone defects should enable controllable transitions between antibacterial and osteogenic functions to achieve sequential repair. Here, we employ different NIR radiation strategies to program the function mode of the scaffold dynamically. After implantation, continuous NIR radiation was initially applied to maintain the high temperature of membrane, promoting antibacterial effects shown in Fig. 2A. Subsequently, intermittent NIR radiation was used to induce cyclic mechanical deformation, generating ES to upregulate M2 polarization for osteogenesis shown in Fig. 2B.

Fig. 2.

Characterization of the DN hydrogel/electrospun membrane controlled by different NIR irradiation strategies. A. Schematic diagram of continuous NIR irradiation (strong photothermal effects for antibacterial stage). B. Schematic diagram of intermittent NIR irradiation (piezoelectric effect for bone regeneration, electrical stimulation promotes M2 polarization). C. Continuous NIR irradiation and temperature changes of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane. D. Piezoelectric current of the membranes under continuous NIR irradiation. E. Intermittent NIR irradiation and temperature changes of DN hydrogel/PVDF/BT@PVDF/BT/PDA electrospun membrane. F. Piezoelectric current of the membranes under intermittent NIR irradiation. G. Schematic diagram of mechanical support and mineralization without NIR irradiation. H. The water contact angle of different membranes. I. Calcium ions release profile of SN hydrogel and DN hydrogel.

Under continuous NIR irradiation, membrane temperature rapidly increased and stabilized around 55 °C (Fig. 2C), while piezoelectric current was only detected during the initial and final heating transitions (Fig. 2D). This behavior was attributed to heat diffusion: once both sides of the membrane exceeded the hydrogel's LCST, no further asymmetric deformation occurred, halting current generation. Thus, continuous NIR enables a pure photothermal antibacterial phase without electrical interference, which is critical during early-stage infection control.

Later, under intermittent NIR irradiation, the temperature fluctuated between 35 and 44 °C (Fig. 2E). This sustained a persistent temperature gradient across the membrane, leading to continuous bending and recovery cycles, and thereby maintaining consistent piezoelectric current output (Fig. 2F and Video S2). The deformation-recovery cycles subjected the piezoelectric fibers to periodic compression and relaxation, generating stable piezoelectric charges. Furthermore, mild photothermal stimulation, as supported by previous studies [13,39], can synergistically enhance M2 macrophage polarization. Therefore, intermittent NIR irradiation enables the scaffold to transition from an antibacterial to an osteogenic function via combined piezoelectric and mild photothermal effects. Meanwhile, the control group DN hydrogel/PAN@PAN/PDA electrospun membrane under both stimuli could hardly generate any current, further confirming that the generated current originated from the piezoelectric effect of PVDF/BT.

Furthermore, even without NIR irradiation, the scaffold's intrinsic characteristics supported osteogenesis (Fig. 2G). Water contact angle measurements (Fig. 2H) confirmed that hydrogel incorporation significantly enhanced the surface hydrophilicity of the composite membrane, which, together with its mechanical robustness and favorable surface topography, provides an optimal interface for osteoblast adhesion and proliferation. Surface roughness analysis (Fig. S11) revealed an average Ra of 4.99 μm, offering micro‐scale cues conducive to cell attachment and spreading [40].

Importantly, calcium ions release assays demonstrated that DN hydrogel exhibited significantly slower ion release compared to single-network (SN) hydrogels. After four weeks, DN hydrogels released only 19.32% of the initial calcium content, compared to 96.5% release with 24 h of SN hydrogels Fig. 2I. This sustained calcium release mimics physiological bone mineralization kinetics, promoting gradual osteoblast differentiation and robust new bone formation [41].

Collectively, these results demonstrate that the NIR-regulated multi-mode scaffold dynamically orchestrates sequential antibacterial, immunomodulatory, and osteogenic phases, offering a powerful platform for the effective repair of infected bone defects.

3.6. In vitro antimicrobial performance of multi-mode periosteum scaffolds

Although the membrane can reach higher temperatures under NIR irradiation, the irradiation parameters for both antibacterial assays and in vitro cell experiments were deliberately controlled at ∼42 °C to ensure biosafety while maintaining antibacterial efficacy. To investigate the antimicrobial efficacy of multi-mode periosteum scaffolds, colony formation experiments were conducted under different NIR irradiation strategies. As shown in Fig. 3A–C, the number of colonies formed by MRSA (the common Gram-positive bacterium) and Escherichia coli (E. coli) (the common Gram-negative bacterium) was significantly reduced in the M + CI group. This enhanced bactericidal activity was attributed to the strong photothermal effect induced by continuous NIR.

Fig. 3.

Antimicrobial activity of multi-mode periosteum scaffolds with different NIR irradiation strategies. A. Representative images of MRSA and E. coli colonies with different treatments. Quantitative analysis of MRSA (B) and E. coli (C) colonies in (A), respectively (n = 3). Results are presented as means ± SD. Samples were subjected to one-way ANOVA with Tukey's post hoc test. Ns (p > 0.05), ∗∗p < 0.01. D. Representative SEM images of MRSA and E. coli colonies with different treatments. Scale bar: 1 μm. E. Live/dead staining of MRSA and E. coli colonies in different treatment groups. Scale bar: 100 μm. F. Confocal microscopy image of MRSA biofilm. Scale bar: 100 μm. Protein leakage statistics for MRSA (G) and E. coli (H), respectively (n = 3). Results are presented as means ± SD. Samples were subjected to one-way ANOVA with Tukey's post hoc test. ∗∗ Significant differences of group M + CI vs. NC, M, and M + II, p < 0.01.

Further morphological analysis using SEM revealed the substantial membrane disruption in both MRSA and E. coli after treatment with M + CI group, evidence by visible surface cracks and structural collapse (Fig. 3D). Consistent with these findings, live/dead bacterial staining demonstrated an obvious increase in dead bacterial populations (red fluorescence) in the M + CI group (Fig. 3E), confirming effective bacterial killing. The anti-biofilm capability of the scaffolds was further assessed using MRSA biofilms (Fig. 3F). Continuous NIR irradiation resulted in almost complete biofilm disruption, demonstrating the scaffold's potent ability to eradicate bacterial communities, an essential feature for treating persistent infections.

To further elucidate the underlying antibacterial mechanism, bacterial protein leakage assays were performed (Fig. 3G and H). A significantly higher protein release was observed in the M + CI group compared to other groups, indicating extensive membrane damage, consistent with SEM observations. Collectively, these results demonstrate that the multi-mode periosteum scaffolds, under continuous NIR irradiation, exhibit robust antimicrobial performance by inducing membrane disruption and biofilm eradication through photothermal effects.

3.7. Macrophage polarization induced osteogenic mechanism of multi-mode periosteum scaffolds in vitro

To evaluate the in vitro biocompatibility of the multi-mode periosteum scaffolds, co-culture experiments with RAW264.7 macrophages and bone marrow-derived mesenchymal stem cells (BMSCs) were conducted, respectively (Fig. S12 and S13). Cell counting kit-8 (CCK-8) assays demonstrated excellent cell viability across all groups, irrespective of the NIR irradiation strategy (Fig. S12A and S13A). Consistently, live/dead staining and flow cytometry apoptosis analysis confirmed the favorable biocompatibility of the scaffolds under all tested conditions (Figs. S12B, S12C, S13B and S13C).

To explore the molecular mechanisms by which the multi-mode periosteum scaffolds modulate macrophage polarization under different NIR irradiation strategies, RNA sequencing (RNA-seq) was applied to evaluate their effects on RAW264.7. Differentially expressed genes (DEGs) between the M + CI group and M + II groups were visualized via volcano plots (Fig. 4A). To further reveal the role of these DEGs in specific biological processes, Kyoto encyclopedia of genes and genomes (KEGG) signaling pathway enrichment analysis was performed. As shown in Fig. 4B, signaling pathways associated with inflammation were significantly enriched, including the nucleotide-binding oligomerization domain-like receptor (NLR) signaling pathway, cytokine-cytokine receptor interaction, Janus activated kinase-signal transducer and activator of transcription (JAK-STAT) signaling pathway, ‌nuclear factor kappa-B (NF-κB) signaling pathway and tumor necrosis factor (TNF) signaling pathway [[42], [43], [44], [45]]. In addition, Fig. 4C and D demonstrated the top 10 significantly up- or down-regulated gene ontology (GO) terms for the two irradiation strategies. We found that the M + II group significantly downregulated inflammation-related GO terms. and the results similarly indicated that the intermittent irradiation had an inhibitory effect on inflammation. (Fig. 4E). Heatmap analysis of M1/M2 polarization-related genes showed that intermittent irradiation significantly upregulated M2-associated genes while downregulating M1 markers (Fig. 4F). Notably, ES generated by intermittent irradiation suppressed M1 polarization, a critical step in infection resolution, underscoring the importance of carefully timing ES application in infected bone defect treatments [46].

Fig. 4.

Multi-mode periosteum scaffolds modulated immune-related gene expression related to macrophage polarization and inflammatory signaling at the transcriptomic level. A. The volcano plot of differential gene expression in M + II vs M + CI. Relatively up-regulated representative GO enriched terms. B. KEGG enrichment analysis of differentially expressed genes in M + II vs M + CI. Relatively up-regulated representative GO enriched terms (C) and down-regulated representative terms (D) in M + II vs M + CI. E. GSEA analysis of genes related to the cytokine-cytokine receptor interaction, the JAK-STAT signaling pathway and the TNF signaling pathway. F. Heatmap of M1 and M2 related gene expression among M + II, M + CI, M and NC groups.

To confirm the property of regulating macrophage polarization under different irradiation strategies, we selected CD206 (classic biomarker for M2 macrophages) and CD86 (biomarker for M1 macrophages) to evaluate the macrophage phenotypes of RAW264.7 using flow cytometry. As shown in Fig. 5A–C, after NIR intermittent irradiation, up to 61.3% ± 2.3 of RAW264.7 expressed CD206, while the proportion of RAW264.7 expressing CD86 significantly decreased to 18.6% ± 1.7. This result was attributed to the intermittent irradiation induced the generation of charges on the surface of the periosteum scaffolds, promoting the polarization of macrophages towards the M2 type. Immunofluorescence assays confirmed enhanced CD206 expression and suppressed CD86 expression under intermittent irradiation (Fig. 5D). Consistently, RT-qPCR results demonstrated upregulation of M2-associated genes (arginase 1 (Arg1), interleukin 10 (Il10), and transforming growth factor beta 1 (Tgfb1)) and downregulation of M1-associated genes (inducible nitric oxide synthase 2 (Nos2), Tnf, and interleukin 1b (Il1b)) further validating the polarization shift induced by intermittent NIR stimulation (Fig. 5E–H and S14). Consistently, Western blot analysis showed decreased expression of iNOS and IL-1β and increased expression of Arg-1 and TGF-β at the protein level, further confirming macrophage M2 polarization (Fig. S15).

Fig. 5.

Multi-mode periosteum scaffolds induced macrophage M2 polarization and regulated osteogenic differentiation of BMSCs in vitro. (A) Expression of CD206 and CD86 in RAW264.7 treated with different NIR irradiation and co-cultured with multi-mode periosteum scaffolds for 48 h. Quantitative statistics of cells expressing CD206 (B) and CD86 (C) in RAW 264.7 co-cultured with multi-mode periosteum scaffolds for 48 h and treated with different NIR irradiation strategies. D. Immunofluorescence staining of CD206 and CD86 in RAW 264.7 after 48 h of different treatments. Scale bar: 100 μm. RT-qPCR analysis of mRNA expression levels of Il10(E), Tgfb1(F), Tnf(G) and Il1b(H) after 48 h of different treatments of RAW 264.7. Results are presented as means ± SD (n = 3). Samples were subjected to one-way ANOVA with Tukey's post hoc test. Ns (p > 0.05), ∗∗P < 0.01. I. Diagram of the indirect co-culture of RAW264.7 and BMSCs. RAW264.7 and the multi-mode periosteum scaffolds were co-incubated for 48 h under different NIR irradiation strategies, and the conditioned medium was used to culture BMSCs. J. ALP staining and Alizarin Red S staining of BMSCs co-cultured with different conditioned medium and multi-mode periosteum scaffolds for 7 days (ALP) and 21 days (ARS). K. ALP activity of BMSCs co-cultured with different conditioned medium and multi-mode periosteum scaffolds for 7 days. Results are presented as means ± SD (n = 3). Samples were subjected to one-way ANOVA with Tukey's post hoc test. ∗∗ denote significant differences of group M + II vs M and M + CI, P < 0.01. L. Quantitative analysis of Alizarin Red S in BMSCs co-cultured with different conditioned medium and multi-mode periosteum scaffolds for 21 days. Results are presented as means ± SD (n = 3). Samples were subjected to one-way ANOVA with Tukey's post hoc test. ∗∗ denote significant differences of group M + II vs M and M + CI, P < 0.01. M. Immunofluorescence staining for detection of Runx2 in BMSCs after 3 days of the different conditioned medium. Scale bar: 100 μm.

Since M2 macrophages are known to promote osteogenic differentiation of BMSCs via secretion of pro-regenerative cytokines, we next assessed the osteogenic effects of conditioned medium derived from RAW264.7 cultures treated with scaffolds under different irradiation strategies [47]. In an indirect co-culture system (Fig. 5I), BMSCs exhibited elevated alkaline phosphatase (ALP) after 7 days in the M + II group (Fig. 5J and K). After 21 days, alizarin red staining revealed the abundant mineralized calcium nodules in the M + II group (Fig. 5J and L), indicating enhanced calcium deposition. Moreover, immunofluorescence staining showed significantly increased expression of the osteogenesis-related protein marker runt related transcription factor 2 (Runx2) in BMSCs exposed to conditioned medium from the M + II group (Fig. 5M). RT-qPCR further confirmed the upregulation of Runx2 and secreted phosphoprotein 1 (Spp1) mRNA levels at 3 and 7 days post-treatment, respectively, while a significant increase in osteocalcin (OCN) expression was observed at 2 weeks (Fig. S16), substantiating the enhanced osteogenic differentiation. Western blot analysis further verified elevated Runx2 and OPN protein expression, consistent with the transcriptional results and indicating enhanced osteogenic differentiation (Fig. S17). These results demonstrate that intermittent NIR irradiation of the multi-mode periosteum scaffolds dynamically promotes M2 macrophage polarization, which in turn enhances the osteogenic differentiation of BMSCs. Meanwhile, the sustained release of Ca²⁺ ions provide essential mineral components that support extracellular matrix mineralization during bone formation. This orchestrated immune modulation and pro-osteogenic microenvironment formation underscore the potential of the scaffolds for sequential repair of infected bone defects.

3.8. In vivo assessment of the programmable treatment efficacy

Prior to the in vivo therapeutic study, the degradation behavior of the patch was evaluated under both in vivo and in vitro conditions. After 4 weeks, no obvious macroscopic structural changes were observed. Quantitative analysis showed a mass loss of 6.9% in vivo and 3.2% in vitro, indicating good structural stability during the therapeutic period and suitability for implant applications requiring sustained mechanical support (Fig. S18). Histological evaluation of major organs (heart, liver, spleen, lung, and kidney) revealed normal tissue architecture without detectable pathological abnormalities, and routine blood analysis showed all hematological parameters within normal physiological ranges, confirming favorable in vivo biosafety during the implantation period (Fig. S19).

Given that different NIR irradiation strategies conferred different biological functions to the scaffolds, we further evaluated their therapeutic efficacy in a rat model of MRSA-infected critical-sized skull defects (Fig. 6A). Four programmed treatment regimens were designed based on distinct NIR irradiation strategies: continuous irradiation alone (M + CI), intermittent irradiation alone (M + II), simultaneous continuous and intermittent irradiation (M + CI&II), and sequential continuous irradiation for 1 week followed by intermittent irradiation (M + CIII). A detailed summary of the experimental groups and corresponding NIR irradiation protocols is provided in Table S2 for clarity. To validate the in vivo temperature profiles of the scaffolds under these strategies, infrared thermal imaging was performed. Continuous irradiation rapidly elevated scaffold temperature and stabilized at 42 °C, whereas intermittent irradiation caused periodic fluctuations between 35 °C and 42 °C, consistent with in vitro observations (Fig. 6B–D). Therefore, depending on the programmed NIR irradiation strategy, the multi-mode periosteum scaffold can deliver distinct therapeutic functions: continuous irradiation primarily induces photothermal antibacterial activity, intermittent irradiation generates piezoelectric stimulation to promote osteogenesis, while combined or sequential irradiation strategies integrate these effects to achieve stage-specific antibacterial and regenerative therapy.

Fig. 6.

In vivo establishment of an infected cranial bone defect model and photothermal performance of multi-mode periosteum scaffolds under different NIR irradiation strategies. A. Establishment of infected bone defect models in the rat skull and application of different NIR irradiation strategies. B. Photothermal images of multi-mode periosteum scaffolds in rats when subjected to different NIR irradiation strategies. Temperature variation patterns for continuous irradiation (C) and intermittent irradiation (D).

To assess the infection control, colony-formation assays were conducted on wound exudates. As shown in Fig. 7A and B, groups employing continuous irradiation (M + CI, M + CI&II, and M + CIII) exhibited significant reductions in bacterial counts, consistent with the in vitro antimicrobial results. Interestingly, the M + CIII sequential irradiation group achieved superior antibacterial efficacy corresponded to the M + CI&II group, suggesting that early-stage ES may impair antimicrobial efficiency by prematurely suppressing M1 macrophage activity, highlighting the critical importance of sequential therapy in infected bone defect treatment.

Fig. 7.

Multi-mode periosteum scaffolds for repairing MRSA infected bone defects in the rat skull. A. Representative images of MRSA colonies in the different groups. B. Quantitative analysis of MRSA colonies in (A). Results are presented as means ± SD (n = 3). Samples were subjected to one-way ANOVA with Tukey's post hoc test, ∗p < 0.05, ∗∗∗∗p < 0.0001. C. Micro-CT analysis of MRSA infected skull defects in SD rats at 4 weeks. D. Quantification of BV/TV at bone defect sites for 4 weeks. Results are presented as means ± SD (n = 3). Samples were subjected to one-way ANOVA with Tukey's post hoc test. ∗∗p < 0.01, ∗∗∗∗p < 0.0001. E. Comparison of BV/TV of different material systems for repairing critical bone defects. F. H&E and Masson staining of bone defect sites in SD rats at 4 weeks. G. Immunohistochemical staining of CD206 for 1 week and OCN for 4weeks in the bone defect sites at 4 weeks.

Bone regeneration was subsequently evaluated via micro-computed tomography (micro-CT) at 4 weeks post-surgery. Three-dimensional reconstruction revealed the most extensive new bone formation in the M + CIII group (Fig. 7C). Quantitative analysis demonstrated significantly higher bone volume/total volume (BV/TV) and bone mineral density (BMD) values in the M + CIII group compared to other groups (Fig. 7D and S20). To the best of our knowledge, among the reported periosteum scaffolds used for critical bone defects in rats, our work had the highest BV/TV at 4 weeks (Fig. 7E) [[48], [49], [50], [51], [52], [53], [54], [55], [56], [57]]. These results suggest that sequential NIR irradiation enables an initial antibacterial phase via photothermal effects, followed by an osteogenic phase facilitated by piezoelectric stimulation, thereby orchestrating effective bone repair. Histological analyses further corroborated these findings. H&E staining showed substantial cortical bone formation exclusively in the M + CIII group, while Masson's trichrome staining showed more mature bone tissue (red staining), compared to other groups (Fig. 7F). Immunohistochemical staining demonstrated robust CD206 expression at 1 week, indicating enhanced M2 macrophage polarization, and strong OCN expression at 4 weeks, signifying advanced osteogenesis at the at the defect site (Fig. 7G). All the above results suggested that the periosteum scaffolds, when combined with programmable NIR irradiation, achieve efficient infection clearance and robust bone regeneration in vivo, offering a promising strategy for the treatment of critical-sized (8 mm) bone defects.

4. Conclusion

In summary, we developed a NIR-programmable scaffold to enable sequential therapy for infected bone defects. By integrating a thermoresponsive PNIPAm-based hydrogel with a bilayer PVDF/BT@PVDF/BT/PDA electrospun membrane, the scaffold achieved on-demand switching of antibacterial and osteogenic activities through tailored NIR irradiation modes. Continuous NIR irradiation rapidly eradicated bacterial infections via photothermal mechanisms, while subsequent intermittent irradiation generated localized electrical stimulation, promoting M2 macrophage polarization and facilitating osteogenesis. In vivo studies demonstrated that the scaffold effectively cleared MRSA infections within one week and seamlessly transitioned to a regenerative phase via immunomodulation. Compared with single-mode or combined simultaneous irradiation, the sequential strategy achieved superior outcomes, including a bone healing rate of more than 60% at four weeks, increased bone mineral density (BMD), elevated CD206⁺ macrophage infiltration at one week, and enhanced osteogenic marker (OPN) expression at four weeks. Our work establishes a programmable, phase-specific therapy with significant translational relevance for complex, antibiotic-resistant bone infections. Clinically, the scaffold can be implanted into defect sites and triggered transcutaneously by NIR to sequentially sterilize infection and promote regeneration, eliminating the need for implant replacement to switch therapeutic phases. This strategy is particularly applicable to post-traumatic osteomyelitis, implant-associated infections, and infected critical-sized defects requiring rapid infection control followed by tissue regeneration. More broadly, it supports the development of smart biomaterials that dynamically tune material cues to match stage-specific host responses during healing in precision regenerative medicine.

CRediT authorship contribution statement

Ying Yin: Writing – original draft, Validation, Investigation, Formal analysis, Conceptualization. Yuting Cai: Writing – review & editing, Writing – original draft, Validation, Investigation, Conceptualization. Pengrui Dang: Writing – review & editing, Validation, Investigation, Writing – original draft, Formal analysis. Wenyi Zeng: Visualization, Software, Methodology, Investigation. Lu Wang: Validation, Investigation, Funding acquisition. Xu Yan: Funding acquisition, Formal analysis. Zhengtang Luo: Supervision, Methodology. Wenwen Liu: Writing – review & editing, Supervision, Funding acquisition. Yangzhi Zhu: Writing – review & editing, Supervision, Project administration. Lili Chen: Writing – review & editing, Supervision. Chenguang Zhang: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

Data availability statement

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

Ethics approval and consent to participate

The animal experiment protocol was approved by the Institutional Animal Care and Use Committee (IACUC), Sun Yat-Sen University (SYSU-IACUC-2025-B0740) and were conducted in accordance with relevant guidelines and regulations. As this study did not involve human participants, consent to participate was not applicable.

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 are the Supplementary data to this article.