NIR-responsive prussian Blue/Fer-1 co-loaded nanozymes for synergistic therapy of acute lung injury

Abstract

Sepsis-associated acute lung injury (ALI) is a critical clinical problem owing to the complex interactions between inflammation and oxidative stress. It is often associated with high mortality rates and a lack of therapeutic options. To simultaneously combat the interconnected inflammatory and oxidative stress pathways in sepsis-associated ALI, we designed near-infrared (NIR)-responsive nanomedicine by co-loading ferroptosis inhibitor Ferrostatin-1 (Fer-1) with Prussian blue nanozymes (denoted as Fer-1@PB) to achieve synergistic therapy and enhanced efficacy. Fer-1@PB combines the multi-enzyme mimetic properties of PB nanoparticle-based nanozyme (PBzyme) for reactive oxygen species (ROS) scavenging and the ferroptosis-inhibitory function of Fer-1 with spatiotemporally controlled burst release under NIR irradiation. The 0.05), and the conclusion of colloidal stability remains completely unchanged. All the above revisions are limited to typographical adjustments, wording accuracy, and minor typo corrections. No experimental data, results, figures, or cistance. Sincerely, Xiaoran Liu (on behalf of all authors) Corresponding author"?> Fer-1@PB nanozymes exhibited strong therapeutic efficacy in cellular and animal sepsis-related ALI models, with an increased survival rate, reduced pulmonary edema and histopathological damage, and recovery of lung function. Mechanistic investigations using transcriptomic analysis (RNA-seq) and experimental validation revealed that the therapeutic effects were mediated by the modulation of the CXCL2/ARF6/NCF1 signaling axis. This study introduces Fer-1@PB as a highly efficient and biosafe nanotherapeutic approach for ALI via synergistic anti-oxidative and anti-inflammatory effects. It also outlines a new molecular mechanism of action, yielding new insights into the precise nanomedicine for inflammatory lung disease.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04440-z.

Introduction

Acute lung injury (ALI), a common and severe complication of sepsis, is associated with inflammatory cell infiltration, cytokine storms, increased vascular permeability, and impaired gas exchange [1–3]. Clinically manifested as refractory hypoxemia and respiratory distress, ALI is a significant cause of mortality in intensive care units, with mortality rates ranging from 30 to 40% [4]. The pathogenesis involves a vicious circle of oxidative stress and uncontrolled inflammation, resulting in redox imbalance, epithelial damage, and neutrophil activation [5]. Currently, no effective drugs are available for the treatment of ALI [6, 7], and survivors often suffer from long-term lung dysfunction, which affects their quality of life and increases the healthcare burden of the disease. Therefore, there is an urgent need to identify innovative strategies that can simultaneously address the intertwined oxidative and inflammatory pathways in ALI, and nanomedicine presents a promising platform for multi-target intervention.

Current pharmacological treatments for ALI are hindered by poor pharmacokinetics, nonspecific side effects, low pulmonary barrier penetration, and poor control of disease heterogeneity, leading to suboptimal treatment outcomes [8]. Given the complexity of the disease and the rapidity of its progression, precise control of drug dosage and timing is critical; however, balancing efficacy and side effects remains a challenge. Nanotechnology is currently being widely used in medicine, especially for targeted drug delivery and anti-inflammatory medicines, which show great potential [9]. Recent cutting-edge studies have further enriched the toolbox of ALI nanotherapeutics: cerium-luteolin nanocomplexes have been proven to be capable of scavenging reactive oxygen species (ROS), inhibiting NF-κB activation, and ameliorating ALI through synergistic anti-oxidant and immuno-modulatory effects [10], and magnetically navigated nano-PROTACs have been demonstrated to achieve lung-targeted delivery, reprogramming of macrophage phenotypes, and drastic improvement of survival rates of ALI [11]. Moreover, nanomaterial-mediated photothermal therapy (PTT) has been proven to be an innovative therapeutic approach with several advantages, including noninvasiveness, high biosafety, and repeatability [8, 12]. PTT also shows great anti-infective potential by converting light to heat to cause membrane disruption, protein denaturation and irreversible bacterial damage [13, 14].

Prussian Blue Nanoparticles (PBNPs) have garnered significant attention from researchers due to their high photothermal conversion efficiency and notable antioxidant properties. They have advantages such as good biocompatibility, high blood stability, low cytotoxicity, biodegradability, and tunable size and morphology [15–17]. PBNPs can be triggered by near-infrared (NIR) light to induce localized thermal effects and exhibit enzyme-like activities. In particular, PBNPs exhibit SOD/CAT-mimetic activities, promoting efficient ROS removal and mitigating oxidative stress [18–20].

Ferroptosis is a form of regulated cell death (RCD) that is predominantly driven by iron-dependent lipid peroxidation, distinguishing it from classical necrosis, apoptosis, apoptosis, and autophagy [21, 22]. It has been regarded as the inducement of multitudinous diseases, including ALI [23–25], pulmonary fibrosis [26], cardiac dysfunction [27], and acute myocardial infarction [28]. During oxidative stress, ROS are generated, which initiate the Fenton reaction in the presence of intracellular ferrous ions. This oxidizes membrane-bound polyunsaturated fatty acids (PUFAs), producing lipid hydroperoxides (LOOH), the accumulation of which disrupts membrane integrity and promotes ferroptosis [29]. In vitro studies have demonstrated that the frequently-used Ferrostatin-1 (Fer-1) suppresses lipopolysaccharide (LPS)-induced injury and preserves BEAS-2B cell viability. In vivo, Fer-1 administration reduces proinflammatory cytokines (IL-6 and TNF-α) in bronchoalveolar lavage fluid (BALF), mitigates pulmonary inflammation, and alleviates ALI progression [30].

Given that oxidative stress and inflammation are two major pathological elements in sepsis-associated ALI, the investigation of PB nanoparticle-based nanozyme (PBzyme) for the treatment of this condition is justified. Li et al. [31] confirmed the protective effects of PBzyme in sepsis-induced ALI. However, the combination of PBzyme-based drug delivery and PTT for ALI remains unexplored. To achieve a synergistic enhancement of anti-oxidant and anti-inflammatory efficacy and enable targeted drug delivery to pulmonary inflammatory tissues, we developed NIR-responsive nanozymes, denoted as Fer-1@PB. This system was constructed by loading the ferroptosis inhibitor, Fer-1, onto hollow mesoporous PBNPs. Leveraging the unique properties of PBNPs, Fer-1@PB facilitates NIR-driven, thermally controlled release of Fer-1 for targeted lung therapy following nebulization (Scheme 1). Combined in vitro and in vivo studies demonstrated that Fer-1@PB alleviates sepsis-associated ALI through its anti-inflammatory and anti-oxidant effects, mediated by modulation of the CXCL2/ARF6/NCF1 signaling axis.

Scheme 1.

Schematic illustration of (A) the preparation procedure of Fer-1@PB nanocomposites and (B) the therapeutic mechanism of the temperature-controlled release formulation

Materials and methods

Materials and reagents

Potassium ferricyanide (K₃[Fe(CN)₆], AR, > 99.5%, Prod. No. P111564) and polyvinylpyrrolidone (PVP, K30, AR, molecular weight = 30000, Prod. No. P274371) were obtained from Aladdin Co., Ltd. (Shanghai, China). Fer-1 (AR, > 99.5%, Cat. No. S7243) was purchased from Selleck Chemicals (Houston, TX, USA). Carbon-coated copper grids (300 mesh, Cat. No. D11023) were purchased from Beijing Daji Scientific Instrument Co., Ltd. (Beijing, China). Phosphate-buffered saline (PBS, 1×, Cat. No. SH30256.01) was obtained from Cytiva (Marlborough, MA, USA). Lipopolysaccharide (LPS, AR, > 99.5%, Cat. No. L2630) was purchased from Sigma-Aldrich (St. Louis, MO, USA). All chemical agents and reagents were used as received.

Synthesis of Fer-1@PB

Hollow mesoporous PB nanocomposites were synthesized using a previously reported method with some modifications [32]. Briefly, a homogeneous mixture of PVP (3.0 g) and K₃[Fe(CN)₆] (120 mg) was dissolved in HCl (20.0 mL, 0.01 M) under sonication and then subjected to thermal treatment at 85 °C for 20 h with constant magnetic stirring (300 rpm). The resulting dark blue products were collected by centrifugation (12,500 × g, 15 min), washed alternately three times with ultrapure water and anhydrous ethanol, and then redispersed in an aqueous solution (10 mL) of HCl (1 M) containing PVP (100 mg) for hydrothermal treatment at 140 °C for 4 h to form hollow structures. For drug loading, hollow PB nanoparticles (20 mg) were incubated with Fer-1 (10 mg) in ethanol under sonication for 30 min, after which the ethanol was removed using a rotary evaporator. The resulting solid was mixed with molten 1‑tetradecanol (50 mg) at 60 °C, emulsified in PBS (2 mL, pH 7.4) via sonication, and rapidly cooled in an ice bath (0 °C). The final products were collected by centrifugation at 12,500 × g, alternately washed three times with ultrapure water and absolute ethanol, and freeze-dried to obtain the Fer-1@PB nanozymes.

Characterization of Fer-1@PB

The morphological characterization of the as-prepared nanozymes was conducted using scanning electron microscopy (SEM) and high-magnification transmission electron microscopy (TEM). The crystal structure and elemental valence states were analyzed using X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS), respectively. The size distribution and colloidal stability of the nanozymes in standard PBS were assessed using dynamic light scattering (DLS), and the surface charge dynamics were quantified using zeta potential measurements.

Drug loading and releasing

Fer-1 (5 µM) was added dropwise to the hollow PBNPs (80 µg/mL) under vigorous stirring, and the supernatant was collected by centrifugation at preset time points (0, 0.5, 1, 2, 6, 12, and 24 h). The concentration of unloaded Fer-1 in the supernatant was determined by ultraviolet-visibles spectroscopy (UV-vis) spectrophotometry at 324 nm based on a pre-established calibration curve (R² = 0.99). The adsorption kinetic curve was plotted using the measured values, followed by Langmuir and Freundlich isotherm model fittings and plotting of the adsorption isotherms of Fer-1 on PBzyme. The drug loading content (DLC) and encapsulation efficiency (EE) were calculated as follows: DLC (µg/mg) = (weight of loaded Fer-1 / weight of nanoparticles) × 1000; and EE (%) = (weight of loaded Fer-1 / total weight of added Fer-1) × 100%. For in vitro release studies, 1 mg of Fer-1@PB was dispersed in 2 mL of PBS (pH 7.4, 10 mM), transferred to a dialysis bag (MWCO = 10 kDa), and immersed in 20 mL of PBS buffer (pH 7.4, 10 mM). The release study was performed in the dark under shaking at 100 rpm. To monitor the drug release behavior, 500 µL of buffer was collected at predetermined time points and replaced with an equal volume of fresh buffer. The released Fer-1 was quantified by spectrophotometry at 324 nm, as previously described [33–35].

Photothermal performance evaluation

The photothermal properties of the PBzyme were evaluated by monitoring the temperature change under NIR laser irradiation. An aqueous dispersion of PBzyme (80 µg/mL) was irradiated with an 808 nm NIR laser at 1.0 W/cm². The temperature increase was recorded every 15 s for 5 min to generate a heating curve. The photothermal stability was investigated over “on-off” five cycles of laser irradiation (5 min on, 11 min off), during which the temperature was recorded at 25-second intervals.

In vitro cellular experiments

Routine culture of human bronchial epithelial (HBE) and RAW264.7 cells was performed in DMEM (10% FBS) at 37 °C under 5% CO₂ until the log phase was reached. To produce an inflammatory injury model, the cells were treated with LPS (10.0 µg/mL) for 24 h. Therapeutic strategies included the delivery of Fer-1@PB with NIR laser treatment (λ = 808 nm, 1.0 W/cm²) with each irradiation cycle performed for 5 min and repeated thrice. The cytocompatibility and toxicological response to Fer-1@PB at different concentrations were evaluated in both cell types using the CCK-8 assay.

In vivo animal experiments

Sepsis-induced ALI was induced in SPF male Sprague-Dawley (SD) rats (7–8 weeks, 220 ± 15 g) by cecal ligation and puncture (CLP). At 24 h post-operation, the rats were treated with Fer-1@PB (2 mg/kg) via nebulization. The NIR group underwent intrapulmonary laser therapy (λ = 808 nm, 1.5 W/cm²) with a 15-minute irradiation protocol starting 0.5 h post-administration. Post-treatment, lung specimens were harvested at 24 h and either fixed in 4% paraformaldehyde (PFA) or cryopreserved at − 80 °C for further analyses. This study was approved by the Ethics Committee of Hainan Medical University (Approval No.: HYALL-2025-049).

Biosafety evaluation

On days 3 and 7 post-intervention, hematological profiles of SD rats were assessed using an automated hematology system. Liver and renal function biomarkers were evaluated using an automated clinical biochemical system. For hemocompatibility assessment, fresh anticoagulated blood was collected from healthy SD rats and centrifuged (3000 rpm, 15 min, 4 °C). The erythrocytes were washed with 0.9% saline and resuspended in a 2% (v/v) solution. Different doses of PB were mixed with red blood cell (RBC) suspensions and incubated for 1 h at 37 °C. After centrifugation, the absorbance of the supernatant was measured at 540 nm. The percentage of hemolysis was calculated using the following formula: (OD_sample - OD_PBS) / (OD_ddH2O - OD_PBS) × 100%.

Oxidative stress evaluation

Oxidative stress markers were analyzed using multiple assays: SOD activity was measured using the WST-1 method (Nanjing Jiancheng, China), and catalase (CAT) activity and glutathione (GSH) levels were determined colorimetrically (Elabscience and Nanjing Jiancheng kits) according to the protocols. Lipid peroxidation was assessed using the TBARS assay (Biosharp, China) to quantify malondialdehyde (MDA) at 535 nm. Intracellular ROS levels were detected using an H₂DCFDA probe (Beyotime, China) and analyzed using flow cytometry. All measurements were performed using a microplate reader (BioTek, USA), and the enzyme activities were calculated using standard curves.

Histological analysis

Lung specimens harvested 24 h post-intervention were fixed in 4% PFA, followed by successive paraffin embedding and sectioning at a thickness of 4 μm for subsequent hematoxylin-eosin (H&E) staining. Two pathologists independently performed blinded semi-quantitative analysis using the classical Smith scoring system to evaluate: (i) alveolar wall thickening, (ii) inflammatory cells, (iii) alveolar edema, and (iv) hemorrhage severity (each parameter scored 0–4, with a maximum total score of 16, where higher scores indicated more severe injury). For fibrosis assessment on day 7, Masson’s trichrome-stained sections were analyzed by calculating the collagen volume fraction (CVF%) from five randomly selected 200× fields using ImageJ. Frozen sections were subjected to immunofluorescence staining with anti-CXCL2/NCF1 primary antibodies, followed by Alexa Fluor 488-conjugated secondary antibody. Confocal microscopy (Olympus FV3000) was used to capture fluorescence signals.

Transcriptomic analysis

Transcriptomic analysis integrated RNA-seq (DNBSEQ/BGI), Gene Set Enrichment Analysis (GSEA), and protein-protein interaction (PPI) network construction. Total RNA from rat lungs was sequenced, and differentially expressed genes (DEGs) (|log2FC| > 1, p < 0.05) were identified in the control, model, and Fer-1@PB + NIR groups. GO and KEGG analyses were performed to map the biological processes and pathways studied. GSEA was used to evaluate oxidative stress pathway enrichment, while STRING-derived PPI networks and key driver analysis (KDA) identified key regulators (CXCL2, ARF6, NCF1) of immune-inflammatory signaling.

Molecular mechanism investigation

The expression levels of CXCL2, ARF6, and NCF1 were evaluated using qRT-PCR and western blotting. RNA was isolated using a column-based TaKaRa kit, reverse-transcribed using PrimeScript RT reagent (Takara), and quantified by SYBR Green qRT-PCR (QuantStudio 3, Thermo Fisher) with GAPDH as the normalization control. Relative quantification analysis was performed using the 2⁻ΔΔCt method (primer sequences are listed in Table S1). Protein lysates were separated by SDS-PAGE, transferred to PVDF membranes, and probed with specific primary antibodies (anti-CXCL2, anti-ARF6, and anti-NCF1) and horseradish peroxidase-conjugated secondary antibodies. The bands were visualized using ECL reagent. To validate the signaling pathway, functional rescue experiments were performed using the CXCR2 inhibitor SB225002 (MCE, China) and ARF6 overexpression mediated by the Chariot cell-penetrating peptide transfection reagent (Active Motif, Carlsbad, CA, USA). Chariot-ARF6 complexes were introduced into HBE cells, and the regulatory relationships among CXCL2, ARF6, and NCF1 were further validated using western blotting.

Statistical analysis

All experiments and computational analyses were conducted with a minimum of three replicates, and the results are reported as mean ± standard deviation (SD). Statistical evaluations were performed using GraphPad Prism 10.3.1 and Origin 2021. Intergroup comparisons were performed using one-way ANOVA followed by Tukey’s post-hoc test, and survival probabilities were analyzed using Kaplan-Meier estimation and log-rank testing. Statistical significance was set at P < 0.05 (*/#).

Results and Discussion

Preparation and characteristics of Fer-1@PB

Hollow mesoporous PBNPs were synthesized via an optimized method and co-loaded with Fer-1 and 1-tetradecanol to construct Fer-1@PB (Scheme 1A). SEM (Fig. 1A-B), and TEM (Fig. 1C) revealed a uniform cubic morphology with a diameter of 112.98 ± 4.24 nm and a hollow mesoporous structure, providing a high surface area for drug loading [36–38]. The TEM elemental mapping (Fig. 1D) further confirmed the uniform distribution of C, N, Fe, and O in Fer-1@PB, verifying the successful incorporation of Fer-1 into the PB matrix.

Fig. 1.

Morphology, composition, and pore structure of PB and Fer‑1@PB. (A, B) SEM images of Fer‑1@PB. (C) TEM image of Fer‑1@PB. (D) TEM mapping of Fer-1@PB and its elemental composition (HAADF, C, N, Fe, and O images). (E, F) N₂ adsorption-desorption isotherms and corresponding BJH pore size distribution curves (inset) of blank PB (E) and Fer-1@PB (F). (G) FTIR spectra. (H) XPS survey spectra. (I-L) XPS spectra of PBNPs and Fer-1@PB: C 1s (I), Fe 2p (J), O 1s (K), and N 1s (L)

BET adsorption-desorption experiments were used to quantify the pore structures of PB and Fer-1@PB (Fig. 1E-F). Blank PB exhibited a BET surface area of 133.85 m²/g and a single-point adsorption total pore volume of 0.2224 cm³/g, whereas Fer-1@PB showed a reduced BET surface area (83.14 m²/g) and total pore volume (0.1785 cm³/g). This reduction in surface area and pore volume indicates that Fer-1 was successfully encapsulated into the mesoporous channels of PB, which is consistent with the hollow mesoporous morphology observed by electron microscopy.

Fourier-transform infrared (FTIR) spectroscopy (Fig. 1G) confirmed the formation of composites by retaining the characteristic Fer-1 peaks (C–H at 2880–2900 cm⁻¹, C = O at 1730.3 cm⁻¹) and intact PB framework signals (C ≡ N at 2210.2 cm⁻¹, Fe–N at 560 cm⁻¹), indicating successful Fer-1 loading without disturbing the PB matrix. XPS was used to analyze the chemical composition and valence states of Fer-1@PB. The XPS survey spectrum (Fig. 1H) confirmed the presence of C, N, Fe, and O in both PBNPs and Fer-1@PB. High-resolution C 1s spectra (Fig. 1I) revealed a new C = O peak (287.71 eV) and a shifted C ≡ N peak (285.98 eV) for Fer-1@PB, respectively. Fe 2p spectra (Fig. 1J) retained the characteristic Fe(II)/Fe(III) coordination peaks, retaining the redox active center of PB. The new O 1s (531.15 eV, Fig. 1K) and N 1s (398.5/399.59 eV, Fig. 1L) peaks also confirmed the successful Fer-1 incorporation.

UV-vis spectroscopy revealed a characteristic absorption peak at approximately 740 nm for PBNPs (Fig. 2A), which was attributed to the Fe²⁺-CN-Fe³⁺ electron transfer that imparts photothermal properties. The UV-vis spectrum of Fer-1@PB (Fig. 2B) revealed the typical peaks of PBNPs and Fer-1, further confirming successful loading. Zeta potential measurements (Fig. 2C) showed similar surface charges between Fer-1@PB (-32.67 ± 1.27 mV) and PBNPs (-34.57 ± 1.97 mV, P = 0.2331) to show the maintained colloidal stability after drug loading. XRD analysis (Fig. 2D), widely applicable for determining crystallinity changes, showed that Fer-1@PB preserved the same main diffraction peaks as PBNPs, suggesting that the lattice structure was not broken.

Fig. 2.

Spectroscopic, thermal, and drug‑loading/release properties. (A) UV‑vis absorption spectra of PBzyme in PBS. (B) UV‑vis spectra of PBNPs, free Fer‑1, and Fer‑1@PB. (C) Zeta potentials. (D) XRD patterns of the samples (E) TGA and DTG curves of the samples. (F) Adsorption kinetics of Fer‑1 on PBzyme. (G-H) Adsorption isotherms fitted using Freundlich (G) and Langmuir (H) models. (I) Temperature-dependent release profiles of Fer‑1 from Fer‑1@PB. (J) Korsmeyer-Peppas model fitting of the release data at 38.5 °C. (K) NIR‑triggered on‑off release profiles over six cycles (2 min on/2 min off). Data are presented as mean ± SD (n = 3)

The thermogravimetric analysis (TGA) and derivative thermogravimetry (DTG) results (Fig. 2E) confirmed the good thermal stability of both PB and Fer-1@PB, with less than 5% weight loss below 100 °C, making them good candidates for PTT applications. Adsorption studies showed time-dependent Fer-1 loading, with rapid adsorption within 2 h and equilibrium reached by 24 h (Fig. 2F). The adsorption process fitted better to the Freundlich model (R² = 0.98) (Fig. 2G) compared to the Langmuir model (Fig. 2H), indicating multilayer adsorption of the molecules. The calculated maximum loading capacity was 8.58 µg/mg, and the EE was 89%.

Fer-1@PB showed good photothermal-controlled release characteristics, with a 14-fold higher cumulative release at 38.5 °C (76.06%) than that at 25 °C (5.27%) over 24 h (Fig. 2I). Korsmeyer-Peppas model fitting of the 38.5 °C release curve (Fig. 2J) gave a correlation coefficient (R²) of 0.9544, a rate constant (k) of 33.95, and a diffusion exponent (n) of 0.2719. When n < 0.45, the release mechanism was dominated by Fickian diffusion, indicating that Fer-1 release was primarily driven by molecular diffusion from the nanocarrier under the photothermal conditions. NIR on-off cycling experiments (Fig. 2K) confirmed precise pulsatile release control with rapid laser-triggered pulses and minimal background release, enabling targeted drug activation for sepsis-associated ALI.

Notably, the 89% EE of Fer-1@PB was comparable to that of advanced lung-targeting liposomes for ALI (90% EE) [39] and higher than the reported EE of ionizable lipid nanoparticles (48.57%) [40] for ALI. Meanwhile, its photothermally triggered cumulative release efficiency (76.06%) exceeded those of previously reported ALI nanomaterials, including inhalable nano-astaxanthin (40%) [41] and poly(citrate-curcumin) nanomaterials (48%) [42]. While the absolute drug loading capacity may be lower than that of some specialized carriers (e.g., ~ 68–79% for certain nanocrystals) [43, 44], Fer‑1@PB uniquely integrates multi-enzyme anti-oxidant activity (SOD-/CAT-/GPx‑like) and NIR‑triggered pulsatile release within a single platform.

Dispersion, stability, and enzyme-like activity of NPs

The Fer-1@PB nanozymes exhibited excellent colloidal stability in PBS, 10% FBS/DMEM, and 3 mM H₂O₂/PBS over 72 h, showing no visible aggregation (Fig. 3A), making it suitable for these applications. TEM imaging confirmed a uniform particle size (Fig. 3B), and the DLS results after 72 h in PBS showed a hydrodynamic diameter of 165 ± 2.58 nm, which was slightly increased but without significant aggregation (Fig. 3C). This expected difference is primarily attributed to the hydration layer and solvent effects captured by DLS, which measures the effective size of nanoparticles in dispersion, whereas TEM provides the physical dimensions of the dry-state core. The UV-vis spectra under oxidative conditions showed no significant decrease in absorbance (Fig. 3D), further supporting the stability of the Fer-1@PB solution.

Fig. 3.

Dispersion stability, enzyme-mimetic activity, and photothermal properties of Fer-1@PB. (A) Dispersibility over 0–72 h in PBS, 10% FBS/DMEM, and 3 mM H₂O₂/PBS (left to right). (B) Particle size distribution (TEM). (C) Particle size distribution (DLS in aqueous media). (D) Absorbance of Fer-1@PB in PBS, 10% FBS/DMEM and 3 mM H₂O₂/PBS. (E) SOD-like activity at different concentrations (pH 5.5, 6.5, 7.4). (F) CAT-like activity at different concentrations (pH 5.5, 6.5, 7.4). (G) GPx-like activity at different concentrations (pH 5.5, 6.5, 7.4). (H-J) EPR spectra of Fer-1@PB for ROS-scavenging: (H) ·OH, (I) O₂⁻, and (J) H₂O₂. (K) Schematic representation of the multi-enzyme-like anti-oxidant activity of PBzymes. (L) Temperature response of PBzyme (0–320 µg/mL) under NIR (1.0 W/cm²). (M) Thermal profiles of PBzyme (80 µg/mL) under varying NIR intensities (0.8–1.4 W/cm²). (N) Photothermal stability of PBzyme (80 µg/mL): 5 NIR on/off cycles (1.0 W/cm²). (O) Time-resolved infrared thermography of PBzyme (80 µg/mL) under NIR (1.0 W/cm²). (P) Phase transition of 1-tetradecanol with increasing temperature. (Q) Phase transition of 1-tetradecanol from solid to liquid as temperature increases. (R) Photograph of PTT in SD rats. Data are presented as mean ± SD (n = 3)

Mechanistic studies have revealed that the Fe²⁺/Fe³⁺ redox pairs of PBzymes efficiently catalyze O₂^- disproportionation (2O₂^- + 2 H⁺ → H₂O₂ + O₂) [45]. PBzyme exhibited concentration-dependent multi-enzyme-mimetic activity, efficiently scavenging O·^- and H₂O₂ via SOD-, CAT-, and GPx-like mechanisms, which is consistent with previous reports [32]. To assess the functional adaptability of these peptides to the acidic inflammatory lung microenvironment, we characterized their activities using a pH gradient (5.5, 6.5, and 7.4). Optimal catalytic activity was recorded at a physiological pH of 7.4 owing to improved structural stability and electron transfer efficiency. Of particular interest, it maintained significant enzyme-like activity at acidic pH 5.5 and 6.5, showing flexibility in the acidic microenvironment of pulmonary inflammation (Fig. 3E-G). Furthermore, quantitative EPR analysis confirmed that PBzyme effectively scavenged key ROS subtypes implicated in ALI, with efficiencies of 83.53%, 83.83%, and 25.8% against OH, O₂^-, and H₂O₂, respectively (Fig. 3H-J).

Collectively, Fer-1@PB nanozymes demonstrated exceptional biostability and a “triple-enzyme” anti-oxidant cascade (Fig. 3K): SOD-like activity converts O₂^- to H₂O₂, which is subsequently decomposed by CAT-like activity [46, 47], whereas GPx-like activity reduces the oxidized lipids. This integrated defense network defines the redox-modulating capabilities of these molecules in inflammatory conditions.

Dual-responsive nanomaterials

To achieve precise drug release, the Fer-1@PB nanosystem was designed to be dual-responsive to photothermal conversion and temperature-sensitive phase-change control. Under NIR irradiation, the PBzyme demonstrated concentration-dependent photothermal heating, reaching nearly 60 °C at 320 µg/mL within 5 min (Fig. 3L), which is consistent with the reported high photothermal efficiency [48, 49]. Even at 80 µg/mL and 0.8 W/cm², the temperature increased beyond 45 °C (Fig. 3M), proving successful heating at clinically relevant temperature. PBzyme also exhibited stable performance over several on/off cycles (Fig. 3N), indicating excellent photothermal stability [50]. Thermal imaging was used to visually verify the increase in temperature (Fig. 3O), supporting its potential for an in vivo triggered release.

In this study, 1-tetradecanol (melting point ≈ 38 °C) was used as a thermally responsive gatekeeper [51, 52]. The phase transition behavior displayed gradient-regulation properties, which are the physical basis for temperature-controlled release mechanisms. At room temperature, 1-tetradecanol is solid, which helps effectively seal the drug and prevent premature leakage. At 36.0–38.5 °C, it enters the semi-solid phase to create partial pores for extended drug release. At 36.0–38.5 °C, complete melting is achieved, resulting in interconnected channels that initiate rapid Fer-1 release. This phase-dependent behavior allows for accurate NIR-triggered drug release (Fig. 3P).

To validate the in vivo photothermal responsiveness of Fer-1@PB, SD rats which had been treated with nebulized Fer-1@PB were exposed to 808 nm NIR laser irradiation. The treatment procedure is shown in Fig. 3Q, and the real-time thermal imaging and quantified temperature profiles are shown in Fig. 3R. The results showed that the local temperatures rapidly increased to approximately 42 °C in a short irradiation time, and no notable heat diffusion was observed, indicating that Fer-1@PB has excellent localized photothermal control.

This study innovatively integrates the photothermal effect of PBzyme and phase-change material properties for ALI therapy, achieving synergistic mild PTT and precise drug release. Unlike traditional PTT, which requires temperatures > 45 °C, our system induces drug release at a lower phase transition temperature (38.0 °C) with a dual benefit: mild PTT at 40–42 °C favors drug release without tissue damage, and local heating generates a “hot-spring effect” that modulates the inflammatory microenvironment. Previous studies have confirmed the anti-inflammatory effect of mild PTT, which shows accelerated wound healing [53] and therapeutic “hot-spring” effects in atherosclerosis at 42–45 °C without triggering apoptosis [54]. Building on this rationale, NIR-based photothermal approaches for ALI treatment have been vigorously explored. Studies have shown that NIR-driven nanoplatforms can control redox homeostasis in the lungs and inhibit the NF-κB pathway through photothermal effects [55] or improve ALI by enhancing ROS-scavenging [56]. Collectively, these advances provide a strong theoretical basis for the use of mild PTT in ALI treatment.

In vitro cell viability and migration ability

Subsequently, HBE and RAW 264.7 cells were used as in vitro model systems for functional evaluation of the extracts. HBE cells are part of the airway epithelial barrier and are among the first cells to be targeted in ALI. They are actively involved in the release of inflammatory cytokines, oxidative stress responses and epithelial repair. RAW 264.7 cells play a crucial role in the immune response, rapidly responding to pathogenic stimuli and activating inflammatory pathways during the early stages of ALI pathogenesis. These complementary cell models were strategically selected to evaluate the biological effects and safety profiles of the Fer-1@PB nanozymes under inflammatory conditions in the lungs.

As shown in Fig. 4A and S1, Fer-1 showed no significant cytotoxicity in HBE cells in the tested concentration range. Further safety evaluation (Figs. 4B-C and S2) showed that PBzyme showed no significant toxicity to either HBE or RAW 264.7 cells at a concentration of 240.0 µg/mL, indicating a favorable biocompatibility profile for the nanosystem.

Fig. 4.

Assessment of cell behavior and migration ability in vitro. (A) Relative cell viability of HBE cells treated with Fer-1 for different durations. (B) Relative viability of HBE cells treated with PBzyme for various durations. (C) Relative viability of RAW 264.7 cells treated with PBzyme for different durations. (D) Live/dead staining images of HBE cells following the different treatments. (E) Relative viability of HBE cells treated with LPS for different time periods. (F) Relative cell activity of RAW 264.7 cells treated with LPS for different durations. (G) Apoptosis in HBE cells was measured using flow cytometry across the treatment groups. (H) Corresponding quantitative results are shown. (I) Wound healing assay assessing HBE cell migration and (J) corresponding wound closure rate. Data are presented as mean ± SD (n = 3)

Under LPS-induced cellular injury (Fig. S3), Fer-1@PB exhibited excellent protective effect. Live/Dead staining (Fig. 4D) showed that cell death was significantly reduced following Fer-1@PB + NIR treatment. CCK-8 assay (Figs. 4E-F and S4-S5) further verified that this group exhibited the most significant promoting effect on cell viability recovery in both HBE and RAW 264.7 cell lines (P < 0.001). Flow cytometric analysis of apoptosis (Fig. 4G–H) indicated that the apoptosis rate in the Fer-1@PB + NIR group (5.24 ± 0.35%) was markedly lower than that in the LPS model group (29.04 ± 1.22%) and superior to those in the Fer-1-alone (15.04 ± 0.95%) and Fer-1@PB-only (10.71 ± 0.83%) groups, suggesting that photothermal triggering synergistically enhances the intracellular delivery and efficacy of Fer-1.

In the cell scratch assay (Fig. 4I–J), Fer-1@PB + NIR treatment significantly accelerated the migration rate of HBE cells, indicating a facilitative role in airway epithelial repair. Given the critical importance of airway epithelial integrity in ALI, these findings provide preliminary evidence for the potential of this strategy to promote overall pulmonary epithelial repair in patients with ALI.

In summary, Fer-1@PB protected HBE cells from LPS-induced damage by improving cell viability, suppressing apoptosis, and enhancing migration of HBE cells. This study is the first to demonstrate that PBzyme and Fer-1 synergistically enhance protection, with coordinated anti-oxidative and photothermal-responsive mechanisms playing critical roles in achieving this outcome.

Anti-inflammatory Effect and anti-oxidant effects in vitro

To assess the anti-oxidative and anti-inflammatory properties of Fer-1@PB, key cytokines and oxidative stress markers were measured. ELISA (Fig. 5A–C) indicated that LPS stimulation significantly increased IL-6, TNF-α, and IL-1β levels in the cells. Fer-1@PB treatment suppressed these cytokines, with additional reduction observed under NIR irradiation. qRT-PCR (Fig. 5D-F, Table S1) showed consistent mRNA trends, supporting the significant anti-inflammatory activity of Fer-1@PB at both the transcriptional and translational levels.

Fig. 5.

Anti-inflammatory and anti-oxidant effects in vitro. Inflammatory markers: (A) TNF-α, (B) IL-6, and (C) IL-1β in the supernat Anti-inflammatory capacity at the cellular level. Inflammation-related genes (D) TNF-α, (E) IL-6, and (F) IL-1β expression levels in treated cells, as determined by qRT-PCR. Oxidative markers in HBE cells after treatment: (G) SOD, (H) GSH, (I) CAT, and (J) MDA. (K) ROS staining (H₂DCFDA) in HBE cells, (L) flow cytometry of ROS intensity, and (M) quantitative analysis. (N) GPX4 and ACSL4 protein expression in HBE cells (Western blot, GAPDH as internal control). (O) Quantitative analysis of GPX4 and ACSL4 relative protein expression in HBE cells. (P) 4-HNE levels in HBE cells. (Q) Intracellular Fe²⁺ levels in HBE cells. Data are presented as mean ± SD (n = 3)

Fer-1@PB exhibited a “three-pronged” protective effect in terms of anti-oxidant capacity, including an increase in endogenous anti-oxidant enzyme activity (Fig. 5G-I) and direct elimination of intracellular ROS (Fig. 5K-M), and inhibition of lipid peroxidation (Fig. 5J). LPS stimulation depleted antioxidants (SOD, CAT, and GSH) and increased MDA levels, indicating extensive oxidative damage. In contrast, Fer-1@PB treatment significantly increased the levels of SOD, CAT, and GSH and decreased the MDA content. Notably, under NIR irradiation, the oxidative stress markers almost fell to baseline levels, which may indicate that Fer-1@PB has potent ROS-scavenging properties.

To visualize intracellular ROS levels, H₂DCFDA staining was used [57]. As shown in Fig. 5K, LPS stimulation caused significant accumulation of ROS, which was effectively circumvented by Fer-1@PB under NIR exposure. Quantitative analysis was performed using flow cytometry (Fig. 5L-M) further demonstrated that the Fer-1@PB + NIR group had the lowest ROS fluorescence intensity (P < 0.001), which was similar to that of the control group, indicating its excellent anti-oxidative efficacy.

Consistent with the reduction in overall oxidative stress, ferroptosis was significantly inhibited in both HBE and RAW264.7 cells by Fer-1@PB + NIR treatment. In HBE cells (Fig. 5N-Q), we found an upregulation of GPX4, downregulation of ACSL4, reduced 4-HNE accumulation, and a lower labile iron pool, suggesting good inhibition of lipid peroxidation-driven cell death. Similar anti-ferroptotic effects were observed in RAW264.7 macrophages (Fig. S6). These data indicate that the anti-oxidant activity of PBzyme, combined with the direct ferroptosis-inhibiting function of Fer-1, cooperatively protects cells from LPS-induced oxidative damage.

Biocompatibility of Fer-1@PB in vivo

Excellent biocompatibility and targeted delivery are of basic importance for nanomaterial-based therapy. We systematically tested the biodistribution and safety of Fer-1@PB nanoparticles. In vivo imaging system (IVIS) showed major accumulation of Cy5.5‑labeled Fer‑1@PB in the lungs within 0.5 h after nebulization and similar signal intensity at 1 h (Fig. 6A-B), indicating rapid pulmonary deposition. Inductively coupled plasma mass spectrometry (ICP-MS) quantitative analysis by Fe element tracing further proved the pulmonary-targeting characteristics of Fer-1@PB. Following intratracheal administration, Fer-1@PB showed low systemic exposure (blood concentration peaked at ~ 0.28 mg/L) and a long duration of lung retention (Fig. S7A-B). The lung-to-non-target organ ratios were significantly high (lung %ID/g > 5‑fold that of the liver at all time points), indicating the pulmonary-targeting advantage of nebulization (Fig. S7C). These pharmacokinetic results determined 0.5 h after administration as the optimal therapeutic window for NIR irradiation to ensure optimal drug distribution in target organs and maximum photothermal synergistic effects.

Fig. 6.

In Vivo Biocompatibility, Targeting, and Therapeutic Efficacy of Fer‑1@PB Nanoparticles in ALI. (A) In vivo fluorescence imaging of major organs. (B) Quantification of lung fluorescence intensity. (C) Changes in body weight over 7 days. (D) Hemolysis assay of PBzyme. (E) H&E staining of major organs at 7 days post-nebulization (scale bar: 100 μm; n = 5). (F) Schematic representation of the ALI model and its treatment. (G) Macroscopic lung images. (H) H&E staining of lung sections (scale bar: 200 μm). (I) Masson’s staining of lung sections (scale bar: 200 μm). (J) Smith score of lung tissue based on hematoxylin and eosin staining. (K) Collagen fiber content in the airway and lung parenchyma (%). (L) Survival curves (n = 10). (M) Lung wet/dry weight ratio. (N) Inflammation-related hematological parameters in rats after administration of different treatment doses. (O) PLT of the treated rats. (P) Serum total bilirubin levels in the treated rats were measured. (Q) Serum creatinine levels in the treated rats. (R‑S) Lung function: (R) Forced vital capacity (FVC) and (S) inspiratory capacity (IC). Data are presented as mean ± SD (n = 5)

Furthermore, 7-day body weight monitoring (Fig. 6C) showed stable weight gain across all treatment groups, with no significant loss induced by Fer-1@PB, even when combined with NIR irradiation. Hemolysis assays (Fig. 6D) revealed < 5% hemolysis at 320 µg/mL PB, far below the ISO 10993-4 safety threshold, confirming excellent hemocompatibility [58]. The 7-day hematological analysis of the control, Fer-1@PB (1, 2, and 4 mg/kg), and Fer-1@PB + NIR groups showed that white blood cells (WBC), RBC, hemoglobin (HGB), and platelets (PLT) were within the normal range (P > 0.05) (Fig. S8E-H), indicating no hematological toxicity from Fer-1@PB alone or in combination with NIR. Coagulation parameters (prothrombin time, PT and activated partial thromboplastin time, APTT) also remained normal in all the groups (Fig. S8I-J). Serum biochemistry assays showed that the hepatic (alanine aminotransferase, ALT and aspartate aminotransferase, AST) and renal (UREA and creatinine, CRE) indices in all groups were comparable to those of the controls (P > 0.05) (Fig. S8K-N). H&E staining of major organs (heart, lung, kidney, liver, and spleen) at 7 days post-nebulization inhalation (control, Fer-1@PB 1/2/4 mg/kg, Fer-1@PB ± NIR) revealed intact tissue architecture without inflammation or necrosis (Fig. 6E), confirming the favorable biocompatibility of Fer-1@PB ± NIR.

The in vivo therapeutic effect of Fer-1@PB with dual responsiveness on ALI

The therapeutic efficacy of the dual-responsive Fer-1@PB nanozyme was evaluated in a CLP-induced sepsis-associated ALI rat model (Figs. 6F, S9). We opted for the CLP model over intratracheal LPS administration because it more faithfully mimics polymicrobial infection and systemic inflammatory response inherent to clinical sepsis-associated ALI. Gross examination of lung tissues (Fig. 6G) showed severe congestion and hemorrhage in the model group, whereas the Fer-1@PB + NIR group exhibited markedly improved pulmonary morphology. H&E staining (Fig. 6H) further confirmed these findings, with the model group displaying alveolar destruction and inflammatory infiltration, whereas the Fer-1@PB + NIR group maintained near-normal alveolar architecture. Smith score quantification (Fig. 6J) demonstrated the significant therapeutic effect of Fer-1@PB under NIR activation. Masson staining and collagen quantification (Figs. 6I and 9k) showed that the Fer-1@PB + NIR group exhibited 29.84% less collagen deposition than the model group, indicating the potential to reduce late-stage fibrotic remodeling in ALI.

Fer-1@PB + NIR treatment significantly enhanced the 7-day survival rate from less than 30% to 80% in ALI rats (Fig. 6L). Pulmonary edema was mitigated, as shown by the decrease in the lung W/D ratio (from 6.05 ± 0.31 to 4.7 ± 0.2, P < 0.001; Fig. 6M). Key inflammatory biomarkers (WBC, NEU, and LYM; Fig. S.6 N, S10), and PLT (Fig. 6O) counts were normalized post-treatment, whereas reduced serum TBIL and CRE levels suggested the mitigation of multi-organ inflammation (Figs. 6P-Q). Lung function tests were used as physiological proof of treatment effectiveness. Profound increases in forced vital capacity (FVC, P < 0.01; Fig. 6R) and inspiratory capacity (IC, P < 0.0011; Fig. 6S) showed substantial recovery of both lung volume and pulmonary compliance, respectively, suggesting amelioration of the restrictive ventilatory deficit that is characteristic of ALI.

Meanwhile, the in vivo results showed a significant decrease in IL-6, TNF-α, and IL-1β in the Fer-1@PB + NIR group (Fig. S11, Table S1, Fig. S12), which indicated that the anti-inflammatory effect of the nanodrug was potent. Concurrently, oxidative stress markers demonstrated significant improvement, with elevated SOD, CAT, and GSH activities and reduced MDA levels (Fig. S13), implying effective reduction of oxidative damage in lung tissue. Furthermore, the therapeutic effect was observed in the inhibition of ferroptosis in the lung tissue. We found a significant restoration of GPX4 expression, reduction of ACSL4 and 4-HNE, and a marked reduction of Fe²⁺ levels in the Fer-1@PB + NIR group (Fig. S14), completing the in vivo evidence for the synergistic anti-oxidation–anti-inflammation–ferroptosis inhibition’ effect.

Transcriptional profiles and DEGs PPI network analysis

To further clarify the molecular mechanisms by which Fer-1@PB alleviates ALI under NIR stimulation, transcriptomic analysis (RNA-seq) was performed on lung tissues obtained from SD rat. This study systematically compared DEGs and related biological functions in the normal, ALI model, and Fer-1@PB + NIR treatment groups. RNA-seq analysis of lung tissues from rats showed high intra-group reproducibility (Fig. 7A) and good data quality (Fig. S15), with good separation between the groups (Venn diagram, Fig. S16). DEG analysis (Figs. 7B, S17) showed that compared to normal controls, ALI models had 384 upregulated and 290 downregulated genes (|log2FC| > 1, P < 0.05), whereas Fer-1@PB + NIR treatment mostly reversed these changes and resulted in upregulation of 341 genes and downregulation of 319 genes when compared with the model group (Fig. 7C-D), suggesting significant reprogramming of the transcriptome by Fer-1@PB + NIR treatment. Functional enrichment analysis highlighted immune-related processes, including “cellular process” and “response to stimulus” (GO, Fig. 7E–F), with KEGG classification further emphasizing roles in “immune system” and “signaling molecules and interaction” (Figs. 7G–H, S18–19).

Fig. 7.

RNA-seq DEGs analysis of ALI lung tissues. (A) RNA-seq Pearson heatmap of the ALI lung tissues. (B) Venn diagram of DEGs: Control, Model, and Treatment groups. (C) Volcano plot of normal vs. model groups. (D) Volcano plot of the model and treatment groups. (E) GO annotation classification of DEGs comparing the original control and final model groups. (F) GO enrichment bubble plot and corresponding bar chart of DEGs between the original control and final model groups. (G) KEGG pathway annotation classification of DEGs between the control and model groups. (H) KEGG pathway enrichment bubble plot

To further delineate the functional pathways and core regulators underlying ALI pathogenesis, we first performed PPI network analysis of inflammation-related pathways, identifying 16 shared DEGs across the model vs. normal and treatment vs. model comparisons (Fig. 8A), and subsequent KDA identified CXCL2 and IL-6 as core inflammatory regulators linked to cytokine-cytokine receptor and chemokine signaling pathways (Fig. 8B). Targeted GSEA analysis of the Model vs. Normal group, with quantitative evaluation of the normalized enrichment score (NES) and false discovery rate (FDR), revealed activation of two biologically relevant pathways closely tied to ALI pathology: the Th17 cell differentiation pathway (NES = 1.89, FDR < 0.25), a classical pro-inflammatory pathway mediating neutrophil recruitment and lung tissue damage, and the Fc gamma R-mediated phagocytosis pathway (NES = 1.72, FDR < 0.25) that regulates oxidative stress via NADPH oxidase-dependent ROS generation (Fig. 8C). These pathways are closely intertwined with the classical NF-κB signaling axis. Th17 cell differentiation is transcriptionally driven by NF-κB-dependent cytokine production, whereas Fc gamma R-mediated phagocytosis promotes ROS accumulation to amplify NF-κB activation. Our findings suggest that Fer-1@PB suppresses this pro-inflammatory and oxidative stress positive feedback loop through the CXCL2/ARF6/NCF1 axis, which is further supported by the construction of an oxidative stress-related PPI network construction (Fig. 8D) and KDA analysis (Fig. 8E) and identified NCF1 and ARF6 as the key regulatory nodes.

Fig. 8.

Verification of the CXCL2/ARF6/NCF1 signaling axis-mediated molecular mechanism. (A) PPI network of inflammation-related DEGs. (B) KDA was used to identify key inflammatory regulators. (C) GSEA of enriched inflammatory/oxidative stress pathways. (D) PPI network of oxidative stress-related DEGs. (E) KDA was used to identify key oxidative stress regulators. (F–H) qRT-PCR analysis of (F) CXCL2, (G) ARF6, and (H) NCF1 mRNA in HBE cells (normalized to GAPDH) (n = 3). (I) Western blot analysis of CXCL2, ARF6, and NCF1 expression in HBE cells. (J) Relative expression levels (n = 3). (K) Western blot analysis of lung tissues. (L) Relative expression levels (n = 3). (M) Quantification of CXCL2 and NCF1 fluorescence intensities in lung tissues (n = 5). (N) Immunofluorescence staining of CXCL2 (red) and NCF1 (green) in lung sections (scale bar: 50 μm; n = 5). (O) Western blotting of HBE cells after CXCR2 inhibition (SB225002). (P) Corresponding quantifications (n = 3). (Q) Western blotting of lung tissues after CXCR2 inhibition. (R) Corresponding quantification of data (n = 3). Data are presented as mean ± SD

To verify the reliability of the RNA-seq data, we performed qRT-PCR validation of three core inflammation/oxidative stress-related DEGs (CXCL2, NCF1, and ARF6) in the lung tissues of SD rats ( Fig. S20, Table S1). The results confirmed that the expression trends of these genes at the transcriptional level were consistent with the RNA-seq data: CXCL2 was significantly downregulated, whereas NCF1 and ARF6 were upregulated in the ALI model group. These patterns were effectively reversed after treatment with Fer-1@PB + NIR.

The identified key genes perform distinct yet complementary roles, with CXCL2 functioning in inflammation [59] and NCF1 in oxidative stress [60], both of which operate in chemokine signaling networks. This coordination establishes the CXCL2/ARF6/NCF1 axis as a potential therapeutic target for Fer-1@PB + NIR therapy in ALI.

Dual-responsive Fer-1@PB regulate CXCL2/ARF6/NCF1-mediated immune response and inflammatory reaction

Based on preliminary RNA-seq and PPI network analyses, CXCL2, ARF6, and NCF1 were identified as co-regulated key factors in inflammatory and oxidative stress processes, as demonstrated in Fer-1@PB treatment both in vitro (HBE cells) and in vivo (ALI lung tissue). qRT-PCR (Fig. 8F–H, Table S1) and Western blot (Fig. 8I–L) confirmed that Fer-1@PB + NIR effectively suppressed the LPS/CLP-induced upregulation of CXCL2, ARF6, and NCF1 at both transcriptional and translational levels. Immunofluorescence staining showed high expression of CXCL2 and NCF1 primarily within inflammatory infiltrates of ALI lungs, while their fluorescence intensities were markedly reduced in the Fer-1@PB + NIR treatment group (Fig. 8M–N), further supporting the role of Fer-1@PB in suppressing local pulmonary inflammation and oxidative stress.

To clarify the hierarchical relationship within the signaling pathway, a “suppression-rescue” dual experiment strategy was employed. To determine whether CXCL2 functions upstream in this pathway, the CXCL2-CXCR2 inhibitor SB225002 [61, 62] was used. The data revealed that the expression of CXCL2, ARF6, and NCF1 was significantly reduced by SB225002 in both the in vitro LPS and in vivo CLP models (Fig. 8O–R), suggesting that ARF6 and NCF1 may be regulated by the CXCL2–CXCR2 signaling axis. To further confirm the mediating role of ARF6 in this pathway, ARF6 was overexpressed in HBE cells by the Chariot transfection system. The Chariot-ARF6 complex intervention was applied to assess the expression levels of ARF6 and NCF1. The results showed that overexpression of ARF6 rescued SB225002-suppressed ARF6 and NCF1 proteins and returned their levels to those of the LPS model group (Fig. S21).

Collectively, these data revealed ARF6 as a central hub integrating the signaling pathways of CXCL2 and NCF1. This two-way validation explains the mechanism of the CXCL2/ARF6/NCF1 pathway and uncovers new mechanisms of action connecting inflammation and oxidative stress. CXCL2 stimulates the receptor CXCR2 to attract neutrophils (NEU) and trigger inflammation [63, 64]. ARF6 (a small GTPase) is important for regulating various processes, such as ROS generation and cell migration [65–67]. NCF1 is a core subunit of NADPH oxidase that catalyzes the reduction of oxygen to produce ROS [68]. This study is the first to systematically establish the hierarchical regulatory role of the CXCL2/ARF6/NCF1 axis in ALI pathogenesis. These results indicate that Fer-1@PB exerts its therapeutic advantage by dual targeting of CXCL2 and its downstream oxidative stress pathways. Notably, Fer-1@PB showed therapeutic superiority by dual targeting of CXCL2 and the downstream oxidative stress pathway, and the axis-mediated ROS inhibition further provided a microenvironment unfavorable for the initiation of ferroptosis, complementing the direct ferroptosis-inhibitory effect of Fer-1 and forming a multi-faceted nanozyme therapeutic strategy.

While the rescue experiment using ARF6 overexpression confirmed the central regulatory role of ARF6 in this axis, future studies using more direct genetic intervention approaches, such as ARF6 stable knockdown in vitro cell models and conditional knockout in vivo animal models, will be critical to further consolidate this mechanistic framework. Such approaches will enable a much more precise delineation of ARF6-mediated control of NCF1 expression and ROS production downstream of CXCL2.

Conclusion

In this study, a novel NIR-responsive PB/Fer-1 co-loaded nanozyme (Fer-1@PB) was successfully developed with thermally controlled drug release and photothermal responsiveness for synergistic ALI therapy. These nanozymes not only had PBzyme-like enzymatic activity but also helped with pulmonary delivery of Fer-1. Comprehensive evaluations have shown its therapeutic potential and have uncovered the underlying therapeutic mechanisms in an ALI model. Material characterization proved that the hollow mesoporous PB had a good Fer-1 loading capacity, and the obtained Fer-1@PB nanozymes had good structural stability, good biocompatibility, and heat-triggered release ability. In vitro, Fer-1@PB with NIR irradiation was effective in reducing LPS-induced damage. In vivo, after inhalation of aerosol, Fer-1@PB was accumulated in the lungs of SD rats, and under NIR stimulation it effectively improved the manifestations of CLP-induced ALI, including reduction of inflammatory infiltration, oxidative damage, impaired lung function, and collagen deposition. Furthermore, RNA-seq and PPI network analysis linked the therapeutic efficacy of Fer-1@PB-NIR to the regulation of the CXCL2/ARF6/NCF1 axis, mediating immune-inflammatory responses, which was validated using qRT-PCR, western blotting, and immunofluorescence assays. Collectively, this study is the first to elucidate the therapeutic potential of Fer-1@PB in ALI. Its unique photothermal responsiveness, combined with its multi-target regulatory effects, provides strong theoretical evidence for the innovative advancement of novel treatment strategies for ALI.

Author contributions

H. L.: Review & editing, Original draft, Visualization, Methodology, Investigation, Data curation. M.L.: Writing-original draft, Formal analysis. W.F.: Validation, Formal analysis. Z.L.: Visualization, Methodology, Formal analysis. J.C.: Supervision, Formal analysis, Data curation. R.L.: Software, Methodology, Conceptualization. P.L.: Validation, Visualization, Resources, Investigation. D.L.: Supervision, Resources, Project administration, Funding acquisition. J.L.: Writing-review & editing, Project administration, Funding acquisition. X.L.: Writing-review & editing, Supervision, Funding acquisition, Conceptualization. All authors reviewed and approved the final manuscript.

Funding

This research was funded by National Key R&D Program of China (Grant 2023YFC3011805), National Natural Science Foundation of China (Grant 81960351, 82560012), Graduate Innovation Research Project of Hainan Province (Grant Qhyb2024-182), Academic Enhancement Support Program of Hainan Medical University (Grant No. XSTS2026055).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal experiments were approved by the Animal Ethics Committee of Hainan Medical University (Approval No.: HYALL-2025-049).

Competing interests

The authors declare no competing interests.