Triptolide-loaded MSC membrane nanoparticles with enhanced pulmonary accumulation for the treatment of acute lung injury: from network pharmacology to experimental validation

Abstract

Acute lung injury (ALI) and its severe form acute respiratory distress syndrome (ARDS) are critical conditions with few effective therapies. Triptolide (TPL), a bioactive compound from Tripterygium wilfordii, exhibits potent anti-inflammatory effects but its clinical use is restricted by poor solubility, rapid clearance, and hepatotoxicity. Using network pharmacology and molecular docking as hypothesis-generating approaches, we prioritized NLRP3 inflammasome activation as a key candidate pathway mediating the therapeutic effects of TPL against ALI. Based on this, we developed mesenchymal stem cell membrane-coated TPL-loaded biomimetic nanoparticles (CM@TPL-NPs) comprising a PLGA core loaded with TPL and coated with mesenchymal stem cell membranes. This coating improved biocompatibility, reduced hepatic sequestration, and enabled preferential lung accumulation, while retaining inherent anti-inflammatory activity. CM@TPL-NPs showed favorable physicochemical properties, sustained release, and improved biocompatibility. Mechanistically, they suppressed NLRP3 inflammasome activation in LPS-stimulated macrophages, while preserving endothelial barrier integrity. In a murine ALI model, both intravenous and intranasal CM@TPL-NPs administration reduced inflammation, cytokine levels, oxidative stress, and histopathological damage. Intranasal (IN) instillation demonstrated enhanced pulmonary accumulation and therapeutic efficacy, positioning it as a promising non-invasive strategy. Importantly, the nanoformulation significantly reduced hepatotoxicity compared with free TPL. Overall, CM@TPL-NPs provide a potent, lung-accumulating platform for ALI therapy via modulation of the NLRP3 inflammasome, supporting further preclinical investigation.

Graphical abstract

1. Introduction

Acute lung injury (ALI) and its fatal form acute respiratory distress syndrome (ARDS) are driven by cytokine release syndrome (CRS)-a hyperinflammatory state sustained by macrophage-endothelial cell crosstalk (Abate et al., 2026; Andrade et al., 2016; Mehta et al., 2020). Current pharmacological interventions, including corticosteroids and monoclonal antibodies, lack pulmonary selectivity and often cause systemic immunosuppression, underscoring the urgent need for targeted therapeutic strategies. Triptolide (TPL), a bioactive diterpenoid triepoxide from Tripterygium wilfordii, exhibits potent anti-inflammatory activity through inhibition of the NF-κB/MAPK signaling axes, yet its clinical translation is severely constrained by poor solubility, rapid clearance, and dose-limiting hepatorenal toxicity (Ballicu et al., 2026; Boregowda et al., 2025; Gupta et al., 2020; Mahajan et al., 2020). Although poly(lactic-co-glycolic acid) (PLGA) encapsulation has improved TPL bioavailability, our prior data demonstrated that conventional PLGA nanoparticles predominantly accumulate in the liver rather than in inflamed lungs (Chuang et al., 2026; Zhang et al., 2019) (Fig. S1)-a targeting deficit that fundamentally limits both efficacy and safety.

To address this limitation, we turned to cell-membrane biomimetic nanotechnology. Among various membrane sources-macrophage, neutrophil, platelet, and mesenchymal stem cell (MSC) membranes-we selected MSC membranes for two compelling reasons that confer unique dual functionality. First, MSC membranes are thought to naturally overexpress homing-related surface molecules, including chemokine receptors, adhesion receptors, and selectin ligands, which may enable multi-receptor synergistic recognition of multiple upregulated ligands in the inflamed pulmonary microenvironment—a feature that we hypothesize contributes to superior pulmonary accumulation compared to single-receptor-dependent systems. Second, and more importantly, MSC membranes intrinsically carry endogenous immunomodulatory factors (PGE2, TGF-β, IL-10, NO) that directly suppress macrophage activation and attenuate inflammatory responses-a feature absent from neutrophil or platelet membrane coatings. Thus, the MSC membrane may serve as both a “targeting ligand library” and a “therapeutic active component,” potentially outperforming other membrane sources that offer targeting alone. In a preliminary screening experiment, MSC membrane-coated nanoparticles at 1 μg/kg (1/50th of the conventional TPL dose) already significantly reduced vascular permeability and inflammatory infiltration (Fig. S2), demonstrating the potency gain conferred by the biomimetic coating. It should be noted that this 1 μg/kg dose refers specifically to the preliminary screening experiment, whereas the main efficacy study used a TPL-equivalent dose of 100 μg/kg, administered as a single dose after LPS challenge. The 100 μg/kg dose was selected based on our previous studies with free TPL (0.2–1 mg/kg) and published TPL nanoformulation studies, aiming to provide robust efficacy for the IV versus IN comparison while remaining below the hepatotoxic threshold of free TPL (Cao et al., 2025).

Equally critical to our design is the recognition that administration route fundamentally dictates therapeutic outcome. Intravenous (IV) injection suffers from hemodilution, reticuloendothelial system (RES)-mediated clearance, and first-pass hepatic metabolism, limiting pulmonary deposition and increasing off-target risks. Intranasal (IN) instillation deposits nanoparticles directly onto the respiratory mucosa, potentially achieving high local drug concentrations with lower distribution to off-target organs while bypassing hepatic first-pass metabolism, (Chang and Xiong, 2020). Importantly, no previous MSC membrane nanoparticle study has systematically compared IV versus IN administration(Liu et al., 2025). Our head-to-head comparison reveals that IN instillation achieves approximately 3-fold higher lung fluorescence intensity than IV administration, with superior suppression of pro-inflammatory cytokines and further improved hepatotoxicity profiles (Fig. S3).

A further innovation lies in the synergistic therapeutic mechanism between TPL and the MSC membrane. Previous MSC membrane-coated nanoparticles have predominantly delivered natural products whose effects are limited to the payload's single anti-inflammatory activity. In contrast, our system achieves a “dual-hit” strategy: TPL directly inhibits NOD-like receptor family pyrin domain containing 3 (NLRP3) inflammasome assembly and caspase-1 activation, blocking IL-1β/IL-18 maturation at the pyroptotic cascade's upstream source, while the MSC membrane concurrently modulates the local immune microenvironment via surface-bound immunoregulatory molecules. This dual mechanism allows CM@TPL-NPs to outperform both free TPL and uncoated TPL-NPs at equivalent concentrations.

To rationally guide our experimental design, we employed network pharmacology and molecular docking as hypothesis-generating tools. These analyses predicted that TPL engages well-established inflammatory pathways, with the NOD-like receptor (NLRP3 inflammasome) signaling pathway ranking as the top enriched target (P < 0.001), thus directing our focus toward NLRP3-mediated NLRP3 inflammasome activation. Molecular docking further suggested favorable binding between TPL and key inflammasome components, providing a computational basis for subsequent experimental validation. The principal novelty and strength of this study, however, lie in the rigorous experimental validation of a biomimetic nanodelivery platform that translates these pathway insights into potent therapeutic efficacy against ALI.

Collectively, the differentiated advantages of our CM@TPL-NPs platform can be summarized across five dimensions. Regarding formulation strategy, MSC membranes provide dual functionality and intrinsic anti-inflammatory functionality, outperforming macrophage, neutrophil, or platelet membranes that offer targeting alone. In terms of targeting capability, multi-receptor synergistic recognition achieves superior pulmonary accumulation compared to single-receptor-dependent systems. Concerning therapeutic efficacy, the TPL-membrane synergy enables “upstream blockade + microenvironment modulation” with enhanced suppression of NLRP3 inflammasome activation. With respect to pharmaceutical advantages, our systematic IN versus IV comparison-absent from prior literature-demonstrates that IN delivery is non-invasive, offers high local bioavailability, and achieves∼3-fold higher lung retention with superior cytokine suppression. Finally, in terms of translational potential, the reduced hepatotoxicity of TPL at therapeutic doses alleviates the primary barrier to clinical translation (Jin et al., 2024; Lin et al., 2025).

In this study, we present a comprehensive investigation integrating network pharmacology-guided target identification, molecular docking, and experimental validation of CM@TPL-NPs for therapy of acute lung injury (ALI). We fully characterized the physicochemical properties, in vitro anti-inflammatory activity, and in vivo biodistribution of the formulation, systematically compared intravenous (IV) and intranasal (IN) administration routes, and evaluated its therapeutic efficacy and systemic safety. Our findings demonstrate that CM@TPL-NPs, through modulation of the NLRP3 inflammasome, represent an effective, safe, and lung-accumulating therapeutic platform for ALI, providing a solid foundation for future translational research.

2. Materials and methods

2.1. Reagents and materials

TPL (HPLC purity >98%) was purchased from Chengdu Herb-Purify Co., Ltd. (Chengdu, China). PLGA-PEG (50:50 lactide: glycolide, MW 38,000; PEG 5000) and polyvinyl alcohol (PVA, 87–89% hydrolyzed) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Lipopolysaccharide (LPS, Escherichia coli O111:B4), l-ascorbic acid, penicillin-streptomycin solution, dimethyl sulfoxide (DMSO), 4′,6-diamidino-2-phenylindole (DAPI) fluorescent dye, Evans Blue dye, and paraformaldehyde were purchased from Sigma-Aldrich. High-glucose Dulbecco's Modified Eagle Medium (DMEM), fetal bovine serum (FBS), trypsin-EDTA solution, and phosphate-buffered saline (PBS) were obtained from Thermo Fisher Scientific (Waltham, MA, USA). TRIzol reagent, SYBR Green Master Mix, and HiScript reverse transcription kits were sourced from Thermo Fisher Scientific (or Vazyme, Nanjing, China) for RNA extraction and quantitative real-time PCR. Primary antibodies against IL-1β, IL-6, TNF-α, and β-actin, as well as horseradish peroxidase (HRP)-conjugated secondary antibodies, were acquired from Abcam (Cambridge, UK). Enzyme-linked immunosorbent assay (ELISA) kits for IL-6, TNF-α, and IL-1β were obtained from R&D Systems (Minneapolis, MN, USA). Hematoxylin and eosin (H&E), Masson's trichrome, and Sirius Red staining kits were purchased from Servicebio (Wuhan, China) or Solarbio (Beijing, China). The bicinchoninic acid (BCA) protein assay kit and radioimmunoprecipitation assay (RIPA) lysis buffer were obtained from Beyotime Biotechnology (Shanghai, China). All other chemicals and reagents were of analytical grade.

2.2. Network pharmacology analysis

2.2.1. Target prediction for TPL

The chemical structure of TPL (PubChem CID: 107985) was retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). Drug-likeness and oral bioavailability were evaluated using the SwissADME web server (http://www.swissadme.ch/), which integrates multiple predictive models including Lipinski's Rule of Five and the Veber criteria. TPL-associated targets were predicted through the SwissTargetPrediction database (http://www.swisstargetprediction.ch/), PharmMapper server (http://www.lilab-ecust.cn/pharmmapper/), and the Comparative Toxicogenomics Database (CTD, http://ctdbase.org/). Targets with probability scores greater than 0.2 were retained for subsequent analysis.

2.2.2. ALI-associated target identification

Disease-associated targets were retrieved from the GeneCards (https://www.genecards.org/), OMIM (https://www.omim.org/), DisGeNET (https://www.disgenet.org/), and Therapeutic Target Database (TTD, http://db.idrblab.net/ttd/) using “acute lung injury” and “acute respiratory distress syndrome” as search terms. The resulting target lists were consolidated, and redundant entries were removed.

2.2.3. Protein-protein interaction (PPI) network and enrichment analysis

The intersection of TPL targets and ALI-associated targets was defined as potential therapeutic targets. The STRING database (https://string-db.org/, version 12.0) was employed to construct a PPI network with a confidence score threshold of 0.9. The PPI data were imported into Cytoscape software (version 3.8.0) for network visualization and topology analysis using the CytoNCA plugin. Node centrality metrics including betweenness centrality (BC), closeness centrality (CC), and degree centrality (DC) were computed to identify hub targets. Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed using the DAVID (https://davidbioinformatics.nih.gov/) and Weishengxing (https://www.bioinformatics.com.cn/) with a false discovery rate (FDR)-adjusted p-value threshold of 0.05.

2.3. Molecular docking studies

Three-dimensional crystal structures of key target proteins were retrieved from the Protein Data Bank (PDB, https://www.rcsb.org/). The protein structures were preprocessed using PyMOL (version 2.5) by removing water molecules, adding hydrogen atoms, and optimizing charge states. The 3D structure of TPL was prepared using AutoDockTools 1.5.6, with Gasteiger charges assigned and rotatable bonds defined. Molecular docking was performed using AutoDock Vina 1.2.0 with an exhaustiveness parameter of 32. The optimal binding poses were selected based on the lowest binding free energy and analyzed for intermolecular hydrogen bonds and hydrophobic interactions using PyMOL.

2.4. Preparation and characterization of CM@TPL-NPs

2.4.1. Preparation of TPL-loaded PLGA nanoparticles

TPL-loaded PLGA nanoparticles (TPL-NPs) were fabricated using a single-emulsion solvent evaporation method. Briefly, the organic phase was prepared by dissolving 50 mg of PLGA and 5 mg of TPL in 5 mL of dichloromethane (DCM). This solution was added dropwise to 20 mL of 2% (w/v) PVA aqueous solution under sonication at 200 W for 5 min (2-s pulse, 2-s pause) in an ice bath using a probe sonicator (Scientz-IID, Ningbo, China). The resulting oil-in-water emulsion was stirred at room temperature for 4 h to remove the organic solvent. Nanoparticles were collected by centrifugation at 12,000 ×g for 20 min, washed three times with deionized water, and resuspended in PBS. Blank PLGA nanoparticles were prepared identically without TPL.

2.4.2. MSC membrane extraction and coating

Mouse mesenchymal stem cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin at 37 °C in a 5% CO₂ humidified atmosphere. Cells at 80–90% confluence were harvested using trypsin-EDTA and washed three times with ice-cold PBS. Cell membranes were extracted using a hypotonic lysis buffer (10 mM Tris-HCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, pH 7.4) followed by repeated freeze-thaw cycles (3 cycles at −80 °C/37 °C). The lysate was sonicated for 5 min and then centrifuged at 800 ×g for 10 min to remove nuclei and cell debris. The supernatant was further centrifuged at 12,000 ×g for 30 min at 4 °C to pellet the membrane fraction. The membrane vesicles were resuspended in PBS, quantified using the BCA protein assay kit, and stored at −80 °C until use.

2.4.3. Membrane coating via extrusion

TPL-NPs were mixed with MSC membrane vesicles at a weight ratio of 1:1 (NP protein: membrane protein) and incubated at room temperature for 30 min with gentle agitation. The mixture was then sequentially extruded through polycarbonate membranes with pore sizes of 400 nm, 200 nm, and 100 nm using an Avanti Mini-Extruder (Avanti Polar Lipids, Alabaster, AL, USA) for 11 passes per membrane size. The resulting CM@TPL-NPs were purified by centrifugation at 12,000 ×g for 20 min to remove uncoated NPs and excess membrane material (Jiang et al., 2019).

2.4.4. Determination of membrane coating efficiency

The membrane coating efficiency was determined using a BCA protein assay. Coating efficiency is defined as the percentage of MSC membrane proteins successfully coated onto the nanoparticle surface relative to the total membrane proteins added.

The procedure was as follows: MSC membrane vesicles were extracted and their total protein content was quantified using a BCA kit; TPL-NPs were mixed with membrane vesicles at a 1:1 (w/w) ratio, extruded to promote coating, and then centrifuged at 12,000 ×g for 20 min to separate coated nanoparticles from free proteins; the supernatant was collected and the amount of uncoated free membrane protein was determined again by the BCA assay. Three independently prepared batches (n = 3) were measured. Bare TPL-NPs without membrane vesicles and membrane vesicles without nanoparticles were processed in parallel as background and recovery controls, respectively.

$$\mathit{Coating\; efficiency}\left(\%\right)=\left[\left(W\_\mathit{total}-W\_\mathit{free}\right)/W\_\mathit{total}\right]\times 100\%$$

Where $W\_\mathit{total}$ is the total amount of MSC membrane proteins initially added, and $W\_\mathit{free}$ is the amount of uncoated free membrane proteins in the supernatant after centrifugation.

2.4.5. Physicochemical characterization

The hydrodynamic diameter, polydispersity index (PDI), and zeta potential of TPL-NPs and CM@TPL-NPs were determined by dynamic light scattering (DLS) using a Zetasizer Nano ZS instrument (Malvern Panalytical, Malvern, UK). All measurements were carried out at 25 °C with samples dispersed in water (refractive index: 1.33, pH 7.0), and an equilibration time of 30 s was allowed prior to each measurement. Morphological diameter and surface morphology were examined by transmission electron microscopy (TEM) and scanning electron microscopy (SEM), respectively (TEM, JEM-1400, JEOL, Japan), following negative staining with 2% phosphotungstic acid. The encapsulation efficiency (EE%) and drug loading capacity (DLC%) of TPL were quantified by high-performance liquid chromatography (HPLC, Agilent 1260 Infinity II) using a C18 reverse-phase column (250 × 4.6 mm, 5 μm). The mobile phase consisted of acetonitrile: water (35:65, v/v) at a flow rate of 1.0 mL/min, with UV detection at 218 nm. EE% and DLC% were calculated as follows:

Encapsulation efficiency (EE%):

$$\mathrm{EE}\%=\frac{W_{\text{total}}-W_{\text{free}}}{W_{\text{total}}}\times 100\%$$

Drug loading capacity (DLC%):

$$\mathrm{DLC}\%=\frac{W_{\text{total}}-W_{\text{free}}}{W_{\text{nanoparticles}}}\times 100\%$$

Where $W_{\text{total}}$ is the total amount of TPL added, $W_{\text{free}}$ is the amount of free TPL determined in the supernatant, and $W_{\text{nanoparticles}}$ is the total mass of the nanoparticles (lyophilized).

2.4.6. In vitro drug release study

The in vitro release profile of TPL from TPL-NPs was evaluated using a dialysis method. The TPL-NP suspension (equivalent to 1 mg of TPL) was transferred to a dialysis bag (molecular weight cut-off 14 kDa) and dialyzed against 50 mL of PBS (pH 7.4) supplemented with 0.5% Tween 80 at 37 °C under continuous agitation at 100 rpm. To maintain sink conditions throughout the release study, the solubility of TPL in the release medium (PBS containing 0.5% Tween 80, pH 7.4) was first determined to be 500 μg/mL. The maximum theoretical TPL concentration in the release medium upon complete drug release (1 mg/50 mL = 20 μg/mL) was confirmed to be less than 30% of the saturation solubility, ensuring sink conditions were maintained. The stability of TPL in the release medium was confirmed by incubating TPL at 37 °C for 96 h and quantifying by HPLC at 0, 24, 48, 72 and 96 h; recovery remained above 95%. To exclude adsorption of TPL to the dialysis membrane, a mass-balance experiment was performed using a TPL solution of known concentration dialyzed under identical conditions; the overall recovery was 97.2% ± 1.1%, indicating negligible membrane adsorption. At predetermined time intervals (0, 0.5, 1, 2, 4, 8, 24, 48, 72 and 96 h), 1 mL aliquots of the release medium were withdrawn and replaced with an equal volume of fresh medium. Cumulative release was calculated using the standard correction formula Q(t) = C(t) × V + v × ΣC(i), where C(t) is the concentration at time t, V is the total medium volume, v is the withdrawn volume and C(i) are the concentrations at previous sampling points. TPL concentration in the samples was quantified by HPLC as described above. The release study was performed with three independently prepared nanoparticle batches (n = 3), and data are presented as mean ± SD.

2.4.7. Stability assessment

The colloidal stability of TPL-NPs was monitored over 7 days of storage at 4 °C, 25 °C, 40 °C by periodically measuring particle size, PDI, and Zeta potential using DLS.

2.4.8. Serum stability

The stability of CM@TPL-NPs in physiological environments was assessed by incubating the nanoparticles in DMEM containing 10% FBS at 37 °C. Particle size was monitored at 0, 2, 4, 8, 16, and 24 h. CM@TPL-NPs maintained stable particle size throughout the 24-h incubation period, with no significant aggregation observed, indicating favorable serum stability.

2.4.9. Sterilization and endotoxin testing

Endotoxin levels in the final CM@TPL-NP formulation were quantified using the Limulus Amebocyte Lysate (LAL) kinetic turbidimetric assay (Charles River Laboratories) according to the Chinese Pharmacopoeia (2020 edition, General Chapter 1143) and USP <85>. The detection limit of the assay was 0.03 EU/mL. To minimize nanoparticle interference with the LAL reaction, samples were diluted with endotoxin-free water, and spike-recovery validation was performed (recovery within 90–110%, confirming the absence of interference). The measured endotoxin concentration of the administered formulation was 0.08 ± 0.02 EU/mL (mean ± SD, n = 3 independent batches), which is below the acceptance criterion of 0.5 EU/mL recommended for parenteral nanoparticle formulations. Prior to in vivo administration, all formulations were sterilized by filtration through a 0.22 μm sterile syringe filter (Millipore), and only batches passing the endotoxin criterion were used in animal experiments.

2.4.10. HPLC method validation

An HPLC method was established and validated for the quantification of TPL. The calibration curve showed good linearity over the concentration range of 2–400 μg/mL (R² = 1). The precision RSD was 0.216% (n = 6), indicating good instrument precision. The recoveries of spiked samples ranged from 95.0% to 105.0% with RSD < 2.0%, confirming satisfactory accuracy. The LOD and LOQ were 0.02 μg/mL and 0.05 μg/mL (S/N = 3:1 and 10:1, respectively). These results demonstrate that the HPLC method is sensitive, accurate, and reliable for TPL quantification.

2.5. In vitro cell experiments

2.5.1. Cell culture

RAW264.7 mouse macrophages, EA.hy926 human pulmonary microvascular endothelial cells, and mouse mesenchymal stem cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin at 37 °C in a 5% CO₂ humidified incubator(Thermo Fisher Scientific，Waltham, MA，USA). Cells were passaged at 80–90% confluence using standard protocols. EA.hy926 cells were obtained from the American Type Culture Collection (ATCC, CRL-2922) and authenticated by short tandem repeat (STR) profiling. All cell lines were routinely tested and confirmed to be free of mycoplasma contamination.

2.5.2. Cell viability assay

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8) assay. RAW264.7 macrophages and EA.hy926 endothelial cells were seeded in 96-well plates at a density of 5 × 10³ cells per well and incubated overnight. Cells were then treated with varying concentrations of free TPL (0–20 ng/mL), TPL-NPs, or CM@TPL-NPs (equivalent TPL concentrations) for 24 h. For the LPS stimulation model, cells were pre-treated with LPS (1 μg/mL) for 1 h prior to nanoparticle administration. Following treatment, 10 μL of CCK-8 solution was added to each well and incubated for 2 h at 37 °C. Absorbance was measured at 450 nm using a microplate reader (Thermo Fisher Scientific). Cell viability was expressed as a percentage relative to the untreated control group.

2.5.3. Hemolysis assay

The hemocompatibility of CM@TPL-NPs was evaluated using fresh mouse blood collected in heparinized tubes. Erythrocytes were isolated by centrifugation at 800 ×g for 10 min, washed three times with PBS, and resuspended to a 2% (v/v) hematocrit. The TPL nanoparticles of various groups were incubated with red blood cell suspensions at 37 °C for 4 h. PBS and deionized water served as negative and positive controls, respectively. After incubation, samples were centrifuged at 800 ×g for 10 min, and the absorbance of the supernatant was measured at 540 nm. The hemolysis rate was calculated as:

$$\text{Hemolysis}\left(\%\right)=\left({\mathrm{A}}_{\text{sample}}-{\mathrm{A}}_{\mathrm{neo}}\right)/\left(\mathrm{A}\_\mathrm{pos}-\mathrm{A}\_\mathrm{neo}\right)\times 100\%$$

2.5.4. Immunofluorescence assay

RAW264.7 macrophages were seeded on glass coverslips in 24-well plates at 2 × 10⁴ cells per well and cultured overnight. Cells were pre-treated with LPS (1 μg/mL) for 1 h and subsequently treated with free TPL, TPL-NPs, or CM@TPL-NPs (equivalent TPL 0.1 μg/mL) for 24 h. Following treatment, cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 10 min, and blocked with 3% bovine serum albumin (BSA) for 30 min at room temperature. Cells were then incubated with primary antibodies against NLRP3 (1:200), caspase-1 (1:200), ASC (1:200), IL-1β (1:200), and IL-18 (1:200) overnight at 4 °C. After washing with PBS, cells were incubated with Alexa Fluor-conjugated secondary antibodies (1:500) for 1 h at room temperature. Nuclei were counterstained with DAPI (1 μg/mL) for 5 min. Fluorescence images were captured using a confocal laser scanning microscope (Leica TCS SP8, Wetzlar, Germany).

2.5.5. Endothelial barrier function assessment

The integrity of the pulmonary microvascular endothelial barrier was evaluated using an in vitro Transwell model. EA.hy926 cells were seeded on Transwell permeable supports (0.4 μm pore size, polyester membrane, Corning, NY, USA) at 1 × 10⁵ cells per insert and cultured until a confluent monolayer was established. The transendothelial electrical resistance (TEER) was measured using an EVOM2 epithelial voltohmmeter (World Precision Instruments, Sarasota, FL, USA) to confirm monolayer integrity. LPS (1 μg/mL) was added to the upper chamber to induce barrier disruption, followed by treatment with free TPL, TPL-NPs, or CM@TPL-NPs (equivalent TPL 0.1 μg/mL). TEER values were measured at 0, 6, 12, 18, and 24 h. The apparent permeability coefficient (Papp) of fluorescein isothiocyanate-dextran (FITC-dextran, 4 kDa, 0.1 mg/mL) across the endothelial monolayer was determined at 24 h by measuring fluorescence intensity in the lower chamber (excitation: 485 nm, emission: 530 nm). Additionally, the expression of the adherens junction protein VE-cadherin was assessed by immunofluorescence staining analysis.

2.5.6. Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from RAW264.7 macrophages using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. RNA purity and concentration were assessed using a NanoDrop One spectrophotometer (Thermo Fisher Scientific). Complementary DNA (cDNA) was synthesized from 1 μg of total RNA using the PrimeScript RT Reagent Kit (Takara Bio, Kusatsu, Japan). qRT-PCR was performed on a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA) using SYBR Green PCR Master Mix (Takara Bio). The following primer sequences were used: IL-1β forward 5’-TGCCACCTTTTGACAGTGATG-3′, reverse 5’-AAGGTCCACGGGAAAGACAC-3′; IL-6 forward 5’-ACAAGTCGGAGGCTTAATTACACTA-3′, reverse 5’-TTGCCATTGCACAACTCTTTTC-3′; and TNF forward 5’-CTGTAGCCCACGTCGTAGC-3′, reverse 5’-TTTGAGATCCATGCCGTTG-3′. Relative gene expression was calculated using the 2 − ΔΔCt method, with GAPDH as the internal reference.

2.6. Establishment of the murine ALI model

Male C57BL/6 mice (8–10 weeks old, 22–25 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Mice were housed under standard conditions (22 ± 2 °C, 55 ± 5% humidity, 12-h light/dark cycle) with free access to food and water.

Animal Grouping, Blinding, and Sample Size Determination. A total of 35 mice were initially enrolled and randomly allocated to seven groups (n = 5 per group): (1) Control, (2) LPS, (3) LPS + free TPL, (4) LPS + TPL-NPs (IV), (5) LPS + CM@TPL-NPs (IV), (6) LPS + TPL-NPs (IN), and (7) LPS + CM@TPL-NPs (IN). No animals died or were excluded during the study, and all five animals per group were included in every endpoint analysis. Sample size was determined a priori based on our previous ALI model studies, with TNF-α level in BALF as the primary outcome measure. Assuming α = 0.05 (two-sided), power = 0.80, and an estimated effect size of 30%, a minimum of 4–5 mice per group was required. To adhere to the 3R principles of animal welfare, we included 5 mice per group. In the main efficacy study, TPL-equivalent doses were administered as a single dose 3 h after LPS challenge. The TPL-equivalent dose used for all formulations was 100 μg/kg. This dose was selected based on: (i) our previous studies, in which free TPL at 0.2–1 mg/kg produced anti-inflammatory efficacy but also dose-limiting hepatotoxicity in murine inflammation models; (ii) our preliminary screening experiment (Fig. S2), in which MSC membrane-coated TPL nanoparticles at as low as 1 μg/kg already significantly reduced pulmonary vascular permeability; and (iii) published TPL nanoformulation studies using comparable doses. Although 1 μg/kg already reduced vascular permeability in the preliminary screening experiment, a higher dose of 100 μg/kg was selected for the main efficacy study to ensure robust and reproducible therapeutic effects for the head-to-head IV versus IN comparison. The selected dose was intended to remain below the hepatotoxic threshold of free TPL.

2.7. In vivo fluorescence imaging

To evaluate the in vivo distribution and lung accumulation capability of the nanoparticles, CM@TPL-NPs and TPL-NPs were labeled with the near-infrared fluorescent dye indocyanine green (ICG). ICG was physically encapsulated into the PLGA core during nanoparticle preparation (10 μg ICG per mg PLGA) rather than adsorbed onto the surface. Unbound dye was removed by three centrifugation/washing cycles (12,000 ×g, 20 min), and the labeling efficiency was 95.2% ± 1.8% as determined by spectrofluorometry against an ICG standard curve. The stability of the ICG-labeled nanoparticles was assessed by incubating them in PBS and in DMEM containing 10% FBS at 37 °C for 24 h; more than 90.5% ± 2.1% of the initial fluorescence remained particle-associated after ultrafiltration, indicating negligible dye release during the imaging window. In addition, a free-ICG control group (equivalent dye dose, administered intravenously) was included in the imaging experiment; free ICG showed rapid clearance with minimal pulmonary signal at 4 h and 24 h, confirming that the lung fluorescence observed in the nanoparticle groups represents nanoparticle-associated material rather than released or free dye. For the biodistribution study, three mice per group (n = 3) were used at each imaging time point. ALI mice were administered ICG-labeled nanoparticles via IV or IN routes. At 1, 2, 4, 8, 12, and 24 h post-administration, mice were anesthetized and imaged using the IVIS Spectrum In Vivo Imaging System (PerkinElmer, Waltham, MA, USA) with excitation/emission wavelengths of 780/820 nm. Following the final imaging time point, major organs (heart, liver, spleen, lungs, and kidneys) were excised, rinsed with PBS, and subjected to ex vivo fluorescence imaging. Fluorescence intensities were quantified using Living Image software (PerkinElmer). For tissue-section imaging and flow cytometry, Coumarin-6 (C6) was used as the fluorescent probe because its excitation/emission profile is compatible with the confocal microscope and flow cytometer; the same purification and stability procedures were applied. The dye used for each experiment is explicitly stated in the Methods and in each figure legend.

2.8. Assessment of lung inflammation and injury

2.8.1. Bronchoalveolar lavage fluid (BALF) analysis

BALF was collected by performing three sequential lavages of the lungs with 0.5 mL of sterile PBS through a tracheal cannula. The collected BALF was pooled and centrifuged at 500 ×g for 10 min at 4 °C. The supernatant was stored at −80 °C for cytokine analysis. Total protein concentration in BALF was measured using the BCA protein assay kit. The levels of tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), IL-1β in BALF were quantified using commercial enzyme-linked immunosorbent assay (ELISA) kits (MultiSciences, Hangzhou, China) according to the manufacturer's instructions. The cell pellet was resuspended in PBS, and total cell counts were determined using a hemocytometer. Differential cell counts were performed on cytospin preparations stained with Wright-Giemsa stain.

2.8.2. Histopathological examination

Lung tissues were fixed in 4% paraformaldehyde for 48 h, dehydrated through a graded ethanol series, cleared in xylene, and embedded in paraffin. Sections of 4 μm thickness were prepared and stained with hematoxylin and eosin (H&E) for general histopathological evaluation. The severity of lung injury was assessed using a semi-quantitative scoring system (0–12 scale) based on the following criteria: alveolar congestion, hemorrhage, neutrophil infiltration into the alveolar spaces and interstitium, and alveolar wall thickness. Scoring was performed by a pathologist blinded to the tested groups.

2.8.3. ELISA analysis of lung tissue cytokines

Lung tissue homogenates were prepared as described above, and the levels of TNF-α, IL-6 and IL-1β, were measured using commercially available ELISA kits (MultiSciences) according to the manufacturer's instructions. Cytokine concentrations were normalized to total protein content and expressed as pg/mg protein.

2.9. Hepatotoxicity assessment

Blood samples were collected via cardiac puncture, allowed to clot at room temperature for 30 min, and centrifuged at 3000 ×g for 15 min to obtain serum. The activities of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in serum were measured using commercial assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's protocols. Liver tissues were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned at 4 μm, and stained with H&E for histopathological examination.

2.10. Statistical analysis

All data are presented as mean ± SD. Biological replicates were defined as independent animals for in vivo experiments or independently seeded and treated cell cultures for in vitro experiments. The biological replicate was treated as the experimental unit for all statistical analyses. Technical replicates (e.g., triplicate wells or repeated measurements of the same sample) were averaged to yield a single value per biological replicate before statistical testing. Normality was assessed using the Shapiro–Wilk test (p > 0.05 considered normally distributed). For multiple-group comparisons, one-way ANOVA followed by Tukey's post-hoc test was applied, which controls the family-wise error rate for all pairwise comparisons. For two-group comparisons, an unpaired two-tailed Student's t-test was used. Statistical analyses were performed using GraphPad Prism version 9.0. A p-value <0.05 was considered statistically significant. Significance symbols are uniformly defined as: ns (p ≥ 0.05), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact n values (number of biological replicates) are stated in each figure legend and for every experimental endpoint: n = 5 mice per group for efficacy endpoints; n = 3 mice per group for biodistribution imaging; n = 3 independent cell-culture experiments for in vitro assays. Exact P values were not available for all comparisons; P value ranges and the statistical test used are provided in Supplementary Table S5.

3. Results

3.1. Network pharmacology and molecular docking analysis

3.1.1. ADME properties and target prediction

SwissADME analysis revealed that TPL exhibits favorable drug-likeness characteristics, conforming to Lipinski's Rule of Five (molecular weight: 360.4 g/mol; hydrogen bond donors: 1; hydrogen bond acceptors: 8; lipophilicity: 1.84). The oral bioavailability was predicted to be 63%, indicating reasonable intestinal absorption. A total of 183 TPL-associated targets were identified by integrating data from SwissTargetPrediction, PharmMapper, and CTD databases. Concurrently, 10478 ALI-related targets were retrieved from GeneCards, OMIM, DisGeNET, and TTD. The intersection analysis yielded 167 overlapping targets, which were designated as potential therapeutic targets mediating the anti-ALI effects of TPL (Fig. 1A).

Fig. 1.

Network pharmacological profiling of triptolide (TPL) against acute lung injury (ALI): target identification and interactome mapping. (A) Venn diagram showing the intersection between TPL putative targets and ALI-related disease genes. A total of 167 overlapping targets were selected as candidate therapeutic genes for subsequent network analysis. (B) Full-scale protein–protein interaction (PPI) network constructed from the 167 overlapping targets. Nodes are visualized with colour and size scaled by connectivity degree, reflecting their topological centrality in the network. (C) First-neighbor expanded PPI network centered on key targets, providing an integrated view of the TPL-modulated regulatory landscape in ALI. (D—F) Core hub modules and subnetworks extracted via multi-parameter topological filtering (degree and betweenness centrality). These highly interconnected hub genes represent the principal molecular effectors mediating the therapeutic effects of TPL against ALI.

3.1.2. PPI network analysis

The PPI network constructed using STRING and visualized in Cytoscape comprised 167 nodes, indicating extensive functional interactions among the therapeutic targets. Topology analysis identified the top 10 hub genes based on combined degree, betweenness, and closeness centrality scores: AKT1, TNF, IL6, TP53, CASP3, BCL2, JUN, ACTB and STAT3. Notably, AKT1 exhibited the highest connectivity (degree = 39), followed by TNF (degree = 37) and IL6 (degree = 36), underscoring the central roles of these signaling nodes in TPL-mediated anti-ALI activity (Fig. 1B-F).

3.1.3. Functional enrichment analysis

GO enrichment analysis revealed significant enrichment in biological processes related to cellular response to oxidative stress, positive regulation of cytokine production, regulation of inflammatory response, and cell death (Fig. 2A). KEGG pathway analysis identified the TNF signaling pathway, IL17 signaling pathway, and the NOD-like receptor signaling pathway as the most significantly enriched pathways (Fig. 2B). The NOD-like receptor signaling pathway, which directly governs NLRP3 inflammasome activation, ranked among the top pathways with the highest enrichment significance (p < 0.001), strongly suggesting that NLRP3 inflammasome activation represents a critical target for TPL in ALI treatment. Drug–target–pathway–disease network (Fig. 2C).

Fig. 2.

Systematic characterization of TPL targets in ALI via functional enrichment and network analysis. (A) GO functional annotation of putative targets across three domains: biological process (BP), cellular component (CC), and molecular function (MF). Bars are colored according to −log₁₀(P-value) significance, with lengths corresponding to gene enrichment counts. (B) KEGG pathway enrichment profiles. The chord diagram (left) maps hub genes to their respective signaling pathways, while the bubble plot (right) visualizes pathway significance and enrichment depth (gene ratio). (C) Construction of the TPL–target–pathway network to elucidate the polypharmacological landscape. Green nodes (ovals) represent core targets, blue nodes (rectangles) represent ALI-associated pathways, and edges signify predicted interactions. The network highlights a coordinated, multi-dimensional mode of action underlying the therapeutic efficacy of TPL. In the bubble plot, each bubble's size reflects the count of core targets (from the screened set) that are enriched in the corresponding pathway-larger bubbles signify more core targets. The colour scale indicates the P-value, ranging from blue through purple to red, where a deeper red denotes greater significance. The enrichment score and gene ratio represent the fraction of core targets in that pathway out of the total 9 screened core targets. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.1.4. Molecular docking prediction

To verify the interaction between TPL and key components of the inflammasome pathway through computational methods, molecular docking was performed using AutoDock Vina to obtain binding energies. TPL exhibited favorable binding affinities with TNF (−9.06 kcal/mol), Caspase-3 (−6.84 kcal/mol), and IL-6 (−7.5 kcal/mol), with binding energies lower than −6.0 kcal/mol, indicating strong binding interactions. Detailed interaction analysis revealed that TPL forms hydrogen bonds with key residues within each active site, while extensive hydrophobic interactions contribute to overall binding stability (Fig. 3A-B). These docking results provide computational predictions suggesting potential binding between TPL and inflammasome-related targets; these predictions are hypothesis-generating and require further experimental validation.

Fig. 3.

In silico validation of TPL binding affinity toward core targets via molecular docking. (A) Structural visualization of TPL (purple sticks) occupying the catalytic or allosteric sites of eight candidate targets. Magnified views elucidate the intermolecular interaction landscapes, specifically the hydrogen-bond networks (yellow dashed lines) and spatial residue orientations that facilitate stable complex formation. (B) Summary of docking energetics. The integrated table details the target parameters and calculated binding free energies (kcal/mol). All eight core targets exhibited binding energies below the −5.0 kcal/mol threshold, demonstrating a high degree of thermodynamic favorability and robust binding affinity between TPL and its therapeutic targets. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. Preparation and characterization of CM@TPL-NPs

3.2.1. Optimization of TPL-NPs formulation

TPL-loaded PLGA nanoparticles were prepared using the single-emulsion solvent evaporation method. Schematic diagram of the preparation process and administration routes of the nanodrug (Fig. 4A-B).

Fig. 4.

Schematic diagram of the preparation and multi-pronged therapeutic mechanism of CM@TPL-NPs. (A) Schematic workflow for the synthesis of TPL-NPs and their subsequent biomimetic functionalization with cell membranes (CM) to yield CM@TPL-NPs. (B) Mechanistic overview of CM@TPL-NPs in the treatment of ALI. The nanoparticles accumulate in the inflamed lungs, where they neutralize ROS and inhibit NLRP3 inflammasome activation, thereby attenuating the pro-inflammatory cytokine storm.

Systematic optimization results demonstrated that the optimal balance between encapsulation efficiency (EE%) and drug loading capacity (DLC%) was achieved at a drug-to-polymer ratio of 1:10 (w/w). The particle size, PDI, zeta potential, EE%, and DLC% of the optimized TPL-NPs are presented in the figures. Under these optimized conditions, the average hydrodynamic diameter of TPL-NPs was approximately 214 nm, with a narrow size distribution (PDI = 0.187 ± 0.012) and negative surface charge (Zeta = −17 ± 7.2 mV), consistent with effective nanoparticle formation. A series of physicochemical and morphological analyses confirmed the successful coating of MSC membranes onto TPL-NPs. After membrane extrusion, the hydrodynamic diameter of CM@TPL-NPs increased to approximately 217 nm (PDI = 0.159 ± 0.021), consistent with the addition of an approximately 10 nm membrane shell. The zeta potential shifted to a more negative value (Zeta = −75 ± 8.1 mV), reflecting the presence of membrane surface proteins (Fig. 5A-B). Scanning electron microscopy revealed a distinct core-shell nanostructure with a uniform dark PLGA core surrounded by a lighter MSC membrane layer (Fig. 5C). UV–Vis spectra, Western blotting, and Fourier-transform infrared spectroscopy all indicated that the cell membranes had been successfully coated onto the drug-loaded nanoparticles (Fig. 5D-G). Collectively, these results confirmed the successful preparation of CM@TPL-NPs with a well-defined core-shell structure. UV–Vis spectra, together with complementary characterizations including DLS, SEM, Western blotting, and FT-IR spectroscopy, collectively confirmed the successful coating of MSC membranes onto the drug-loaded nanoparticles. Fourier-transform infrared spectroscopy (FTIR, Fig. 5F) was systematically analyzed with full peak assignment and band shift interpretation to validate TPL encapsulation and MSC membrane coating. For free TPL, prominent hydroxyl vibration at 3420 cm⁻¹ and epoxy fingerprint peaks at 1260, 930 cm⁻¹ were clearly observed; these signature bands nearly vanished in TPL-NPs, accompanied by a blue shift of the hydroxyl band to 3360 cm⁻¹ and red shift of PLGA C O peak from 1748 to 1742 cm⁻¹, demonstrating strong intermolecular hydrogen bonds between TPL and PLGA polymer matrix and verifying TPL was embedded inside nanoparticles rather than physically mixed on particle surface. After MSC membrane extrusion, CM@TPL-NPs displayed two unique protein characteristic peaks at 1640 cm⁻¹ (Amide I) and 1542 cm⁻¹ (Amide II), which were absent in both PLGA and bare TPL-NPs. Meanwhile, the carbonyl peak further red-shifted to 1736 cm⁻¹, and a broad Amide A band centered at 3290 cm⁻¹ emerged, originating from abundant peptide bonds on MSC membrane proteins. Additional weak absorption within 980–1050 cm⁻¹ assigned to membrane phospholipid phosphate groups further confirmed the complete coating of intact MSC membrane on nanoparticle surface. Orthogonal spectral changes including loss of free TPL fingerprint peaks, sequential carbonyl red-shifts and appearance of paired amide protein bands collectively confirmed the successful fabrication of core-shell CM@TPL-NPs. While UV–Vis spectroscopy alone provides limited evidence for membrane coating, its combined use with these orthogonal techniques enables robust validation of the core-shell nanostructure. Transmission electron microscopy (TEM) images (Fig. 5G) revealed that the prepared nanoparticles exhibited a regular spherical shape and uniform morphology, with no obvious aggregation observed. The TPL-NPs had an average morphological diameter of approximately 210 nm. High-resolution TEM images of CM@TPL-NPs further revealed a clear core-shell structure with a shell thickness of about 10 nm, which is consistent with our findings. To directly assess the physical state of TPL within the nanoparticles, differential scanning calorimetry (DSC) was performed on PLGA, free TPL, TPL-NPs, and CM@TPL-NPs (Fig. 5H). Free TPL exhibited a sharp endothermic peak at 239.8 °C, corresponding to its crystalline melting. This characteristic peak completely disappeared in the TPL-NPs, which instead showed a broad endotherm in the range of 104.7–122.8 °C, indicating that TPL was molecularly dispersed or present in an amorphous state within the PLGA matrix, with intermolecular interactions between the drug and polymer. The CM@TPL-NPs exhibited a new endothermic peak at approximately 148.7 °C (ΔH = 936.37 J/g), which may be attributed to thermal denaturation of MSC membrane proteins and alterations in PLGA crystallization behavior induced by membrane coating. These DSC results therefore provide direct evidence that TPL is not present as a crystalline phase within the nanoparticle matrix but is instead molecularly dispersed or amorphous.The in vitro release profile demonstrated that the drug was almost completely released from the nanoparticles within 96 h (Fig. 5I). Western blot analysis confirmed the successful coating of the nanodrug with the cell membrane, as evidenced by the presence of CD105, a specific membrane marker. (Fig. 5J), Western blot results also provided certain evidence for the membrane coating efficiency. Although CD105 retention and SDS-PAGE protein profiling indicate successful membrane protein retention on CM@TPL-NPs, these data do not establish complete preservation of the MSC membrane proteome or its biological functionality. Additional representative MSC membrane markers, such as CD44, CD73, CD90 and CXCR4, were not evaluated in the present study, and further experiments are required to confirm functional preservation. The hydrodynamic diameter measured by DLS (approximately 214–218 nm) is larger than the morphological diameter observed by TEM/SEM, because DLS measures the solvated particle including the hydration layer and PEG corona, whereas electron microscopy visualizes the dried particle core.

Fig. 5.

Fabrication, characterization, and multi-pronged therapeutic rationale of CM@TPL-NPs. (A–E) Morphological and physical assessment: (A) DLS-measured hydrodynamic diameter distribution, (B) surface charge (Zeta potential), (C) spatial morphology (SEM), and (D) drug loading verification via UV–Vis absorption spectra. (E) Protein profiling by SDS-PAGE, demonstrating the retention of essential membrane proteins on the CM@TPL-NPs. (F) FT-IR characterization of constituent materials-including mannose, chondroitin sulfate, and PLGA-mPEG-confirming the successful assembly and chemical integrity of the nanoparticle components. (G) TEM images showing morphological diameter. DLS measures the hydrodynamic diameter (including hydration layer and PEG corona), whereas SEM/TEM visualize the dried morphological diameter; the two are complementary. (H) DSC thermograms (I) In vitro drug release results. Data are presented as mean ± SD (n = 3 independent nanoparticle batches). (J) Western blot analysis of CD105 as a representative MSC membrane marker. This indicates membrane protein retention rather than complete preservation of the MSC membrane proteome or functionality. Data are presented as mean ± SD (n = 3 independent batches). Statistical significance was determined by one-way ANOVA followed by Tukey's post-hoc test; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.2.2. Drug loading and encapsulation efficiency

The optimized CM@TPL-NPs exhibited an encapsulation efficiency of 62.45% ± 2.43%, a membrane coating efficiency of 86.78% ± 2.14%, and a drug loading capacity of 7.80% ± 0.19%. These values were essentially unchanged compared with uncoated TPL-NPs, indicating that negligible drug loss occurred during the membrane coating process.

3.2.3. Stability assessment under different temperatures

The colloidal stability of TPL-NPs was evaluated by monitoring particle size, PDI, and Zeta potential over 7 days at 4 °C, 25 °C, and 40 °C. TPL-NPs stored at 4 °C maintained the smallest hydrodynamic diameter and lowest PDI throughout the observation period, indicating superior stability under refrigeration. In contrast, nanoparticles stored at elevated temperatures (25 °C and 40 °C) exhibited progressive increases in particle size and PDI, with the most pronounced aggregation observed at 40 °C, suggesting that temperature elevation accelerates nanoparticle destabilization (Fig. 6 A-C), Quantitative comparison of absolute Zeta potential values further confirmed that 4 °C storage preserved the surface charge stability of TPL-NPs (Fig. 6 D). These results demonstrate that 4 °C is the optimal storage condition for maintaining the physicochemical integrity of TPL-NPs. The particle size was monitored at various time intervals under physiological conditions, and it remained relatively stable over a 24-h period.

Fig. 6.

Colloidal stability of TPL-NPs under different storage temperatures. (A) Hydrodynamic diameter changes of TPL-NPs stored at 4 °C, 25 °C, and 40 °C over 7 days; (B) Size distribution changes of TPL-NPs stored at 4 °C, 25 °C, and 40 °C over 7 days; (C) Representative polydispersity index (PDI) profiles of TPL-NPs under different temperature conditions over 7 days; (D) Quantitative comparison of the absolute Zeta potential values over 7 days to evaluate the relative stability of the nanoparticles. (E) Physiological stability of nanoparticles. TPL-NPs were stored at 4 °C, 25 °C, and 40 °C for 7 days. Data are presented as mean ± SD (n = 3 independent batches). Statistical significance was determined by one-way ANOVA.

3.2.4. Summary tables

The Supplementary Materials provide detailed physicochemical characterization (Table S1), key in vitro and in vivo experimental results (Table S2), and a comparative analysis with representative ALI biomimetic nanocarriers (Table S3) (Zhao et al., 2024).

3.3. In vitro anti-inflammatory activity

3.3.1. Cytotoxicity and hemocompatibility

The cytotoxicity of free TPL, TPL-NPs, and CM@TPL-NPs was evaluated in RAW264.7 macrophages and EA.hy926 endothelial cells using the CCK-8 assay. Free TPL exhibited dose-dependent cytotoxicity, with IC₅₀ values of approximately 0.5 μg/mL and 0.3 μg/mL in RAW264.7 and EA.hy926 cells, respectively. Both TPL-NPs and CM@TPL-NPs demonstrated significantly reduced cytotoxicity compared to free TPL at equivalent drug concentrations, likely due to the sustained release of TPL from the nanoparticle matrix and reduced peak plasma concentrations (Fig. 7A-B). The hemolysis assay confirmed the excellent hemocompatibility of CM@TPL-NPs, with hemolysis rates below 3% across all tested concentrations, well within the acceptable threshold (<5%) for IV administration (Fig. 7C-D).

Fig. 7.

Biocompatibility evaluation of TPL, TPL-NPs, and CM@TPL-NPs. (A) Cell viability of RAW264.7 cells treated with TPL, TPL-NPs, and CM@TPL-NPs at different concentrations. (B) Cell viability of EA.hy926 cells treated with TPL, TPL-NPs, and CM@TPL-NPs at different concentrations. (C) Hemolysis extent of different groups (Water, PBS, TPL, TPL-NPs, CM@TPL-NPs), with the inset showing the visual hemolysis results. (D) Microscopic images of red blood cells treated with H₂O, PBS, TPL, TPL-NPs, and CM@TPL-NPs. Scale bar = 200 μm. Concentration range (0–20 ng/mL), incubation time (24 h), and statistical method (one-way ANOVA). Cells were treated with TPL, TPL-NPs, or CM@TPL-NPs at 0–20 ng/mL for 24 h. Data are presented as mean ± SD (n = 3 independent cell-culture experiments). Statistical significance was determined by one-way ANOVA. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.3.2. NLRP3 inflammasome inhibition

The effect of CM@TPL-NPs on inflammasome activation was investigated in lipopolysaccharide (LPS)-stimulated RAW264.7 macrophages. Quantitative RT-PCR analysis further confirmed that CM@TPL-NPs significantly inhibited the transcriptional upregulation of IL-6, IL-1β, and TNF-α (Fig. 8A). Immunofluorescence staining revealed that CM@TPL-NPs treatment enhanced the integrity of endothelial junctions compared to the endotoxin-treated group (Fig. 8B). Notably, CM@TPL-NPs prevented LPS-induced stress fiber formation and morphological abnormalities in EA.hy926 cells (Fig. 8C).

Fig. 8.

CM@TPL-NPs attenuate inflammation and maintain endothelial barrier integrity in vitro. (A) Transcriptional levels of IL-6, TNF-α, and IL-1β in EA.hy926 cells subjected to various treatments, as quantified by RT-qPCR. CM@TPL-NPs significantly downregulate LPS-induced pro-inflammatory signaling. LPS stimulation concentration (1 μg/mL), CM@TPL-NPs concentration (0.1 μg/mL TPL equivalent), and treatment duration (24 h). Data are presented as mean ± SD (n = 3 independent cell-culture experiments). P-values were calculated using one-way ANOVA. (B) Evaluation of endothelial junctional integrity via immunofluorescence staining of VE-cadherin (green). The disrupted VE-cadherin network induced by LPS was effectively restored by CM@TPL-NP treatment. Nuclei were labeled with DAPI (blue). (C) Visualization of the F-actin cytoskeleton (red) by phalloidin staining. The results demonstrate that CM@TPL-NPs prevent LPS-induced stress fiber formation and morphological distortion in EA.hy926 cells. Nuclei were labeled with DAPI (blue). Scale bar = 20 μm. Cells were stimulated with LPS (1 μg/mL) and treated with TPL, TPL-NPs, or CM@TPL-NPs (0.1 μg/mL TPL equivalent) for 24 h. Data are presented as mean ± SD (n = 3 independent cell-culture experiments). P-values were calculated using one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4. In vivo therapeutic efficacy

3.4.1. Lung-accumulating biodistribution

The in vivo biodistribution of ICG-labeled nanoparticles was monitored using near-infrared fluorescence imaging. Following IV administration, both TPL-NPs and CM@TPL-NPs exhibited substantial lung accumulation, with CM@TPL-NPs demonstrating significantly higher lung fluorescence intensity compared to TPL-NPs at all time points (Fig. 9A-C). Ex vivo fluorescence imaging of lung sections further confirmed the superior pulmonary retention of CM-coated NPs (Fig. 9D–E). Furthermore, flow cytometry analysis revealed that CM@C6-NPs significantly enhanced cellular uptake efficiency in the injured lungs, as evidenced by the increased percentage of C6-positive cells and CD31⁺/C6⁺ double-positive endothelial cells (Fig. 9F-I).

Fig. 9.

In vivo pulmonary accumulation and cellular uptake of CM@TPL-NPs in LPS-induced ALI mice. (A1, A2) Schematic experimental timelines for (A1) intravenous (I.V.) and (A2) intranasal (I.N.) instillation. The workflow includes ALI induction, formulation administration, IVIS imaging, and subsequent tissue analysis. (B1, B2) Representative ex vivo fluorescence images of major organs (heart, lung, liver, spleen, and kidney) at 4 h post-administration. Formulations were labeled with Indocyanine Green (ICG) for whole-organ visualization. (C1, C2) Quantitative analysis of mean fluorescence intensity (radiant efficiency) in lung tissues, showing significantly enhanced lung accumulation/retention of CM-coated NPs compared with bare NPs via both routes. (D1, D2) Immunofluorescence staining of lung sections. Nuclei were stained with DAPI (blue), vascular endothelial cells were labeled with CD31 (red), and nanoparticles were tracked using Coumarin-6 (C6, green). C6-labeled nanoparticles were prepared using the same purification and stability procedures as described for ICG labeling. Scale bars = 100 μm. (E1, E2) Quantification of C6 fluorescence intensity within lung sections, confirming that CM-coating facilitates higher accumulation in the pulmonary parenchyma. (F1, F2) Representative flow cytometry histograms illustrating the percentage of C6-positive cells among total lung cells. (G1, G2) Statistical comparison of C6-positive cell percentages, demonstrating superior lung delivery of the biomimetic CM@C6-NPs. (H1,H2) Representative flow cytometry dot plots of CD31 and C6, identifying the population of endothelial cells that internalized the nanoparticles (CD31⁺/C6⁺). (I1, I2) Quantitative analysis of CD31⁺/C6⁺ double-positive cells, confirming the preferential uptake and endothelial localization of CM@C6-NPs in the injured lungs. Mice received 100 μg/kg TPL-equivalent formulations via IV or IN instillation. Data are presented as mean ± SD (n = 3 per group at each imaging time point). Statistical significance was determined by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4.2. IN instillation and lung retention

IN instillation of CM@TPL-NPs resulted in rapid and sustained pulmonary localization, with fluorescent signals detectable within 3 h post-administration and persisting for up to 24 h (Fig. 10A). Ex vivo imaging of Evans Blue Albumin (EBA) staining revealed strong fluorescence in the lungs, and quantitative analysis of EBA extravasation demonstrated that CM@TPL-NPs treatment significantly reduced the intense blue coloration in lung tissues, indicating amelioration of pulmonary microvascular hyperpermeability (Fig. 10B-C). LPS-induced ALI was associated with a marked increase in total protein concentration in BALF, suggesting increased vascular permeability and protein extravasation into the alveolar space. CM@TPL-NPs treatment effectively reduced BALF protein levels. In the ALI model, the total cell count in BALF was also significantly elevated, primarily due to neutrophil infiltration. CM@TPL-NPs significantly reduced the recruitment of inflammatory cells into the alveolar space, with IN instillation showing superior therapeutic efficacy. Total leukocyte counts in bronchoalveolar lavage fluid (BALF) demonstrated suppressed inflammatory cell recruitment (Fig. 10D-E). These findings highlight the IN route as an effective and non-invasive strategy for achieving high local drug concentrations in the lungs while reducing distribution to off-target organs. Histopathological examination of lung tissues by H&E staining revealed severe lung injury in endotoxin-induced mice, characterized by extensive alveolar wall thickening, marked neutrophil infiltration, intra-alveolar hemorrhage, and pulmonary edema. Treatment with free TPL or TPL-NPs provided modest histopathological improvements. In contrast, treatment with CM@TPL-NPs, particularly via the IN route, significantly improved lung architecture, reduced inflammatory cell infiltration, alleviated alveolar wall thickening, and preserved alveolar structure. IN instillation of CM@TPL-NPs achieved the most pronounced protective effect (Fig. 10F–G).

Fig. 10.

Comparative therapeutic efficacy of CM@TPL-NPs via intravenous (I.V.) and intranasal (I.N.) instillation in LPS-induced ALI. (A1, A2) Schematic illustration of the experimental protocols for (A1) I.V. and (A2) I.N. delivery, outlining the timeline for ALI induction, formulation administration, and barrier function assessment at 24 h post-LPS challenge. (B1, B2) Representative macroscopic images of Evans Blue Albumin (EBA)-stained lungs, highlighting the reduction in intensive blue coloration following CM@TPL-NP treatment. (C1, C2) Quantitative analysis of EBA extravasation (μg/g lung), demonstrating that CM@TPL-NPs effectively attenuate pulmonary microvascular hyperpermeability via both administration routes. (D1, D2) Total leukocyte counts in bronchoalveolar lavage fluid (BALF), illustrating suppressed inflammatory cell recruitment. (E1, E2) Total protein concentrations in BALF, confirming the protection of the alveolar-capillary barrier integrity. (F1, F2) Representative H&E-stained lung sections at low (top) and high (bottom) magnification, showing the preservation of alveolar architecture and attenuation of septal thickening and inflammatory infiltration. (G1, G2) Quantitative assessment of H&E-stained sections. Mice received 100 μg/kg TPL-equivalent formulations via IV or IN instillation. Data are presented as mean ± SD (n = 5 mice per group). Statistical significance was determined by one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. the LPS-treated group. Scale bar = 200 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4.3. Attenuation of lung inflammation

The therapeutic efficacy of CM@TPL-NPs was evaluated in the LPS-induced murine ALI model. LPS challenge caused profound inflammatory responses, evidenced by markedly elevated levels of TNF-α, IL-6 and IL-1β in BALF and lung tissue homogenates. Treatment with CM@TPL-NPs, administered via either IV or IN routes, significantly reduced the levels of all measured proinflammatory cytokines compared to the LPS-treated group (Fig. 11A). Notably, IN CM@TPL-NPs exhibited the most potent anti-inflammatory effects, reducing TNF-α, IL-6, IL-1β, and IL-18 levels in BALF by approximately 65%, 70%, 75%, and 72%, respectively, compared to the LPS group. This superior efficacy is consistent with the enhanced pulmonary accumulation of intranasally delivered nanoparticles. Representative IHC images of lung sections demonstrated that CM@TPL-NPs significantly decreased the expression levels of NLRP3, Caspase-1, and IL-1β via both administration routes, confirming its superior anti-inflammatory efficacy and effective inhibition of the NLRP3/Caspase-1/IL-1β signaling axis in the injured lungs (Fig. 11B-G).

Fig. 11.

Comparative anti-inflammatory and NLRP3 inflammasome inhibitory effects of CM@TPL-NPs via intravenous (I.V.) and intranasal (I.N.) administration in ALI mice. (A1, A2) Relative mRNA expression levels of pro-inflammatory cytokines (IL-6, TNF-α, and IL-1β) in lung tissues following (A1) I.V. and (A2) I.N. treatments, determined by RT-qPCR. (B1–D2) Representative immunohistochemical (IHC) images of lung sections illustrating the protein expression of (B1, B2) NLRP3, (C1, C2) Caspase-1, and (D1, D2) IL-1β under both administration routes (Scale bars = 200 μm). (E1–G2) Statistical quantification of IHC results, expressed as the number of (E1, E2) NLRP3⁺, (F1, F2) Caspase-1⁺, and (G1, G2) IL-1β⁺ cells per 2000 total cells. These data demonstrate that CM@TPL-NPs effectively attenuate the NLRP3/Caspase-1/IL-1β signaling axis in the injured lungs regardless of the delivery pathway. Mice received 100 μg/kg TPL-equivalent formulations via IV or IN instillation. Data are presented as mean ± SD (n = 5); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. the LPS-treated group. Scale bar = 200 μm.

3.5. Hepatotoxicity assessment

A primary concern limiting the clinical translation of TPL is its dose-limiting hepatotoxicity. To evaluate whether the nanoformulation mitigates systemic toxicity, serum ALT and AST levels were measured as biochemical indicators of hepatic function. Mice treated with free TPL exhibited significantly elevated serum ALT and AST activities compared to the control group, indicating substantial hepatic injury (Fig. 12A-D). In stark contrast, mice treated with CM@TPL-NPs at the equivalent TPL dose displayed ALT and AST levels comparable to the control group, with no statistically significant differences observed. Histopathological examination of liver tissues confirmed these findings, with free TPL-treated mice showing hepatocellular necrosis, inflammatory cell infiltration, and sinusoidal congestion, whereas mice treated with CM@TPL-NPs exhibited normal hepatic architecture. These results demonstrate that, under the tested experimental conditions, the MSC membrane-coated nanoformulation markedly attenuated TPL-associated hepatic injury, as evidenced by serum ALT/AST levels and hepatic histology comparable to those of control animals. This improvement is likely attributable to reduced systemic drug exposure and the hepatoprotective effects of MSC membrane components. Free TPL exhibits potential nephrotoxic and splenotoxic effects, possibly causing tubular injury and apoptosis of splenic lymphocytes; therefore, monitoring of liver, kidney, and spleen functions is required clinically.

Fig. 12.

Systemic biosafety profile of CM@TPL-NPs following intravenous (I.V.) and intranasal (I.N.) instillation in ALI mice. (A, C) Representative H&E-stained histological sections of major metabolic and immune organs (liver, kidney, and spleen) following (A) I.V. and (C) I.N. treatments. No discernible histopathological damage or inflammatory infiltration was observed across all treatment groups, indicating that the nanoformulations do not induce organ-specific toxicity (Scale bars = 200 μm). (B, D) Biochemical assessment of liver function markers, specifically serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST), after (B) I.V. and (D) I.N. administration. All measured parameters remained within normal physiological ranges without significant elevation compared to the control groups, confirming the excellent systemic biocompatibility and safety of the biomimetic CM@TPL-NPs. Mice received 100 μg/kg TPL-equivalent formulations via IV or IN instillation. Data are presented as mean ± SD (n = 5). Statistical method: one-way ANOVA; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

4. Discussion

ALI is a critical syndrome with high mortality and limited therapies, motivating extensive research into its inflammatory mechanisms. Here, we employed network pharmacology and molecular docking primarily as screening tools to prioritize the NLRP3 inflammasome axis, and subsequently established a biomimetic nanodelivery platform through comprehensive experimental validation. MSC membrane-coated PLGA nanoparticles loaded with TPL (CM@TPL-NPs) effectively overcame the pharmaceutical limitations of free TPL, achieving potent pulmonary-preferential anti-inflammatory activity. Our data suggest that NLRP3 inflammasome inhibition is a key contributing mechanism.

The NLRP3 inflammasome orchestrates the inflammatory response in ALI, linking pathogen sensing to downstream cascades. Upon endotoxin recognition, NLRP3 assembles and activates caspase-1, which processes pro-IL-1β and cleaves GSDMD to induce pyroptosis, amplifying inflammation (Morihiro et al., 2026). Our network pharmacology prioritized NLRP3 inflammasome activation as a key TPL-associated pathway, consistent with recent reports. Molecular docking predicted favorable binding of TPL to NLRP3, caspase-1, IL-6, IL-1β and TNF-α; these computational predictions were subsequently supported by pharmacological evidence in vitro and in vivo, although direct biophysical binding assays (e.g., surface plasmon resonance or cellular thermal shift assay) will be required to definitively confirm direct molecular interaction.

Developing effective ALI therapies has been hindered by the challenge of achieving therapeutic lung concentrations without systemic toxicity. Conventional approaches yield poor pulmonary bioavailability and off-target effects (Salvado et al., 2025). Our biomimetic strategy may address this via two complementary mechanisms: the PLGA core enhances TPL solubility, reduces hepatic sequestration, and enables sustained release; the MSC membrane coating is thought to provide biological functionalities—adhesion receptors, chemokine receptors (CXCR4, CXCR7), and immunomodulatory factors—that synthetic modifications cannot easily replicate. We hypothesize that these functionalities mediate preferential homing to inflamed sites. Biodistribution studies confirmed CM@TPL-NPs achieved significantly greater lung accumulation than uncoated NPs via both IV and IN routes, supporting their therapeutic superiority.

The administration route critically influences pulmonary delivery outcomes. IV injection suffers from RES clearance, first-pass metabolism, and limited lung targeting. We systematically compared IV and IN delivery in LPS-induced ALI. IN instillation achieved ∼3-fold higher lung fluorescence intensity than IV, with lower distribution to off-target organs, translating into superior anti-inflammatory effects, oxidative stress reduction, and histopathological improvement(Andrade et al., 2016). These findings align with evidence supporting IN instillation as an efficient, non-invasive strategy for achieving high local drug concentrations with reduced distribution to off-target organs (Ballicu et al., 2026). Direct respiratory mucosal deposition, combined with MSC membrane glycoprotein mucoadhesion, likely contributes to the enhanced lung retention observed with IN instillation.

An important mechanistic observation is that the MSC membrane coating itself appears to confer synergistic anti-inflammatory effects beyond drug delivery. At equivalent TPL concentrations, CM@TPL-NPs consistently outperformed free TPL and uncoated TPL-NPs in suppressing NLRP3 inflammasome activation and preserving endothelial barrier integrity. This advantage likely stems from a dual action: encapsulated TPL directly inhibits NLRP3 inflammasome components, while the MSC membrane may present endogenous immunomodulators (PGE2, TGF-β, IL-10, NO) that suppress macrophage activation, promote Treg expansion, and attenuate inflammation. Although the relative contribution of each component remains to be quantified using blank membrane controls, retention of these immunomodulatory components on the nanoparticle surface may contribute to the enhanced anti-inflammatory activity observed. (Boregowda et al., 2025). Compared with existing MSC-based therapeutic strategies, our platform offers distinct translational advantages. Live MSC therapies face challenges including batch-to-batch heterogeneity, tumorigenic risk, immune rejection, and complex cell expansion procedures. MSC-derived extracellular vesicles, despite their immunomodulatory activity, suffer from poor drug loading efficiency, short half-life, and rapid clearance. In contrast, CM@TPL-NPs as a cell-free nano-platform combine high drug loading, sustained release, excellent batch-to-batch consistency, and scalability, while potentially preserving the native homing and immunomodulatory functions of the MSC membrane. More importantly, the dual-hit synergy—TPL directly blocking NLRP3 inflammasome and the MSC membrane simultaneously modulating the local microenvironment—may provide a combinatorial benefit that is difficult to achieve with live cells or EVs alone.

Preserving pulmonary microvascular endothelial barrier integrity is critical in ALI. Barrier disruption, marked by VE-cadherin loss, increases vascular permeability, edema, and impairs gas exchange (Komarova et al., 2017). CM@TPL-NPs effectively preserved VE-cadherin expression and barrier function in LPS-stimulated EA.hy926 cells, as shown by restored TEER and reduced paracellular permeability. These protective effects are likely mediated, at least in part, through inhibition of NLRP3 inflammasome activation and subsequent reduction in inflammatory cytokine release, which directly compromise endothelial integrity. By simultaneously suppressing NLRP3 inflammasome activation and preserving barrier function, CM@TPL-NPs offer a comprehensive therapeutic strategy for ALI.

TPL hepatotoxicity remains the principal barrier to its clinical use. Free TPL caused significant liver injury (elevated ALT/AST and histopathological damage), whereas CM@TPL-NPs at the same dose exhibited an improved hepatic safety profile under the tested conditions, with liver markers and architecture comparable to controls. This improvement is likely attributed to reduced hepatic accumulation of TPL via nanoencapsulation and lung delivery, although systemic exposure was not directly measured, with additional hepatoprotective effects from MSC membrane components via antioxidant and anti-inflammatory mechanisms (Pedersbaek et al., 2021). The favorable safety profile of CM@TPL-NPs, combined with potent efficacy, represents a step toward improving the therapeutic index of TPL that warrants further preclinical safety evaluation.

The integration of network pharmacology with biomimetic nanotechnology provided a rational starting point for target selection. While our computational predictions were largely consistent with TPL's known engagement with inflammatory pathways (e.g., NLRP3, NF-κB), they effectively guided our experimental focus toward the inflammasome axis. The principal value of this study lies not in novel mechanistic discovery from in silico analysis (Liu et al., 2026), but in the robust experimental validation, where our biomimetic delivery system translated these pathway predictions into potent therapeutic outcomes against ALI. This successful translation of computational insights into in vitro and in vivo efficacy underscores the value of an integrated approach for accelerating targeted therapeutic development (Tong et al., 2022).

This study has several limitations. The murine ALI model does not fully reflect the complexity of clinical ALI/ARDS. Pharmacokinetic parameters of TPL (plasma concentration, systemic exposure, half-life, clearance and AUC) were not evaluated in this study. The fluorescence-based biodistribution data reflect the tissue distribution of the nanoparticle carrier rather than the systemic pharmacokinetics of TPL; therefore, statements regarding circulation time and systemic exposure should be interpreted with caution. The current membrane characterization is limited to CD105 and SDS-PAGE protein profiling; therefore, the data demonstrate membrane protein retention rather than complete preservation of the MSC membrane proteome and biological functionality. The receptor-mediated mechanisms underlying MSC membrane-enhanced lung accumulation require further elucidation through receptor-blocking and competition experiments; the present data demonstrate enhanced pulmonary accumulation rather than receptor-specific precision targeting. Additionally, while our data demonstrate robust inhibition of NLRP3, caspase-1, IL-1β, and IL-18—which are upstream events of NLRP3 inflammasome activation—we acknowledge that direct markers of pyroptosis, including cleaved Gasdermin D (GSDMD-N), cleaved caspase-1 (p20), ASC speck formation, and LDH release, were not evaluated in the current study. Therefore, we interpret our findings as suppression of NLRP3 inflammasome activation and its downstream inflammatory cascade, rather than definitive blockade of the pyroptosis pathway per se. These direct mechanistic validations are important priorities for our future investigations (Lu et al., 2025). Beyond the NLRP3 pathway, other TPL targets (e.g., NF-κB, MAPK, AKT1) warrant further investigation. The safety evaluation in this study was limited to serum ALT/AST activities and histological examination of the liver, kidney and spleen at a single time point after treatment. Comprehensive hematological analysis, renal biochemical markers (blood urea nitrogen and creatinine), repeated-dose toxicity, immunogenicity assessment and long-term safety evaluation were not performed. Given the known systemic toxicity of triptolide, these aspects must be systematically addressed before clinical translation, and the safety conclusions of the present study should therefore be considered preliminary. Moreover, residual dichloromethane was not quantitatively analyzed; therefore, complete removal of the organic solvent cannot be claimed based on evaporation time alone. In addition, a blank MSC membrane-coated nanoparticle control (without TPL payload) was not included in the in vivo efficacy study. Therefore, the contribution of the MSC membrane itself cannot be fully separated from that of improved TPL delivery, and the therapeutic improvements observed for CM@TPL-NPs should not be attributed exclusively to enhanced TPL delivery. No clinically relevant positive control (e.g., dexamethasone) was included in this study; therefore, no conclusions regarding superiority over existing clinical therapies can be drawn, and the translational potential of CM@TPL-NPs should be interpreted as preliminary (Chou and Lin, 2024). In addition, direct head-to-head experimental comparisons between CM@TPL-NPs and live MSCs or MSC-EVs were not performed in this study; such systematic comparisons represent an important direction for our future translational validation.

Beyond experimental validation, the clinical translation of CM@TPL-NPs requires attention to several aspects. Regarding manufacturing complexity and scalability, the single-emulsion method faces challenges in size uniformity, batch-to-batch consistency, and coating efficiency during scale-up; microfluidic mixing or high-pressure homogenization may offer potential solutions. Regarding membrane-source variability and membrane protein composition, donor-to-donor and passage-dependent variability of MSC membrane proteins should be carefully controlled, and quality-control criteria (e.g., marker panel verification by Western blot and SDS-PAGE profiling for each membrane batch) are warranted. Regarding sterilization, residual solvent, and endotoxin control, 0.22 μm filtration, residual dichloromethane monitoring, and quantitative LAL testing should be included as critical quality attributes in future manufacturing. Regarding storage stability, liquid formulations require lyophilization for long-term storage; cryoprotectants such as trehalose have enabled ≥3-month stability of stem cell-hybrid biomimetic nanoparticles with preserved membrane protein function. It should be noted that intranasal instillation in mice deposits the formulation in the nasal cavity and upper airways, with subsequent distribution to the lungs, and does not fully reproduce clinically relevant inhalation or nebulized pulmonary delivery in patients. Translation of the IN route to the clinic will therefore require the development of nebulized or dry-powder inhalation formulations, and the stability of CM@TPL-NPs under nebulization shear stress will need to be evaluated. Vibrating-mesh nebulizers and protective excipients represent promising strategies in this regard. Regarding regulatory considerations, as a complex nanomedicine, CM@TPL-NPs require comprehensive physicochemical characterization, batch-to-batch reproducibility, membrane source quality control, sterility/endotoxin testing, and immunogenicity evaluation; the FDA-approved status of PLGA-PEG may facilitate regulatory engagement (Zhang et al., 2024). Regarding the translational roadmap, priorities include GMP protocols, lyophilized formulation development, nebulization optimization, large-animal toxicology/PK studies, and early regulatory dialogue; the non-invasive IN route offers strong potential for patient compliance. The favorable efficacy and safety profile of CM@TPL-NPs provides a foundation for further preclinical investigation (Lopes et al., 2026; Lu et al., 2026).

In conclusion, this study successfully developed a biomimetic nanodelivery platform, CM@TPL-NPs, that addresses the critical pharmaceutical limitations of TPL while achieving preferential pulmonary accumulation and modulation of the NLRP3 inflammasome in ALI. Through a combination of network pharmacology-guided target identification, molecular docking validation, and comprehensive experimental characterization, we demonstrated that CM@TPL-NPs effectively suppress NLRP3 inflammasome activation and the associated inflammatory responses, preserve pulmonary endothelial barrier integrity, and exert potent anti-inflammatory effects in a murine ALI model. While our data strongly support the NLRP3 inflammasome as a major effector pathway, we acknowledge that other TPL targets (e.g., NF-κB, MAPK) may also contribute to the overall therapeutic efficacy. The remarkable improvement in the safety profile, the superiority of IN delivery, and the synergistic anti-inflammatory effects conferred by the MSC membrane coating collectively position CM@TPL-NPs as a promising therapeutic candidate warranting further preclinical investigation before clinical translation for the treatment of ALI and related inflammatory lung diseases (Zhou et al., 2025).

5. Conclusion

In this study, we successfully developed a novel biomimetic nanodelivery system, CM@TPL-NPs, that integrates the therapeutic potential of TPL with the biological functionalities of mesenchymal stem cell membranes for the treatment of ALI. Through network pharmacology and molecular docking as hypothesis-generating tools, we prioritized the NLRP3 inflammasome pathway as a key target of TPL in ALI, which was subsequently validated through experimental studies. The resulting CM@TPL-NPs exhibited favorable physicochemical properties, sustained drug release, and excellent colloidal stability. In vitro studies demonstrated potent inhibition of NLRP3 inflammasome activation and effective preservation of pulmonary microvascular endothelial barrier integrity. In a murine LPS-induced ALI model, both IV and IN administration of CM@TPL-NPs significantly attenuated lung inflammation, reduced inflammatory cytokine production, mitigated oxidative stress, and improved histopathological outcomes, with IN delivery demonstrating superior pulmonary accumulation and therapeutic efficacy. Importantly, the nanoformulation substantially reduced the hepatotoxicity associated with free TPL under the tested conditions. Collectively, these findings establish CM@TPL-NPs as an effective, safe, and lung-accumulating therapeutic platform for ALI, with NLRP3 inflammasome modulation serving as a central mechanism.

CRediT authorship contribution statement

Jiahao Liu: Writing – review & editing, Writing – original draft, Visualization, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Yue Zhao: Investigation. Yixin Zhuang: Investigation. Bowen Zhou: Conceptualization. Xuerong Sun: Visualization. Ning Wei: Methodology. Xianglan Zeng: Data curation. Jiaying Li: Software. Huihui Liu: Software. Yinqiang Fan: Methodology. Xin Wang: Conceptualization. Hua Jin: Writing – original draft, Visualization, Validation, Supervision, Software, Resources.

Ethics declaration

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. This study was approved by the guangdong medical university.

Experimental procedures using mice in this study were reviewed and approved by the ethical review board of Guangdong Medical University, and all the experiments were performed in accordance with relevant guidelines and regulations of Animal Ethics Committee of Guangdong Province, China. Animals and protocol were approved by the Ethics Committee of Guangdong Medical University (GDY2602334).

Funding

This work was supported by Shenzhen Bao'an Hospital - Clinical and Basic Research Project of Guangdong Medical University (GDMULCJC2025192, GDMULCJC2025198, GDMULCJC2025184), Guangdong Medical University Innovation & Entrepreneurship Base Project (JDXM2024006, S202510571042).

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

Supplementary material

Data availability

The data supporting the findings of this study are available within the article and its supplementary materials. Raw data and additional experimental details are available from the corresponding author upon reasonable request.