TLR2 inhibition attenuates NET-driven inflammation and restenosis during poly-lactic acid bioresorbable scaffold degradation

Abstract

Poly-lactic acid (PLA) bioresorbable scaffolds provide temporary vascular support but trigger late-stage inflammation during polymer dissolution, frequently driving neointimal hyperplasia before complete dissolution. While acute biomaterial responses are well documented, the innate immune mechanisms dictating chronic vascular remodeling remain poorly defined. We demonstrate that late-stage PLA degradation actively instructs a microenvironment dominated by neutrophil extracellular trap (NET). Combining transcriptomic profiling of stented porcine arteries, and longitudinal spatial histology, we mapped a sustained immune response directly coupled to polymer breakdown. Mechanistically, PLA degradation products induce NET extrusion via mitochondrial calcium uniporter-mediated calcium overload. These extracellular structures subsequently function as potent paracrine signals, driving human vascular smooth muscle cells (VSMCs) toward a proliferative, synthetic phenotype. We identified a toll-like receptor 2 (TLR2) signaling pathway as the primary transducer of this NET-driven VSMC reprogramming. To apply these findings therapeutically, we engineered a customized PLA scaffold coated exclusively with the selective TLR2 inhibitor C29. In a rabbit iliac artery model, this targeted immunomodulatory device successfully disrupted the local inflammatory loop, significantly reducing neutrophil accumulation and mitigating neointimal hyperplasia. This study establishes the biomaterial-driven NET-TLR2 axis as a central mechanism of adverse vascular remodeling and offers a precise bioactive strategy to preserve long-term luminal patency.

Graphical abstract

Highlights

• •The study identifies neutrophil extracellular trap (NET) formation as the dominant immune response during the late-stage degradation of polylactic acid (PLA) scaffolds, rather than the acute inflammatory response typically seen with permanent metallic stents.
• •Mechanistic investigations reveal that PLA degradation products trigger NET release via mitochondrial calcium overload; these NETs subsequently drive vascular smooth muscle cell (VSMC) phenotypic switching and proliferation through a STAT3-dependent TLR2/NF-κB signaling pathway.
• •By engineering a customized bioresorbable scaffold functionalized with the TLR2 inhibitor C29, the research demonstrates that targeted disruption of the NET-TLR2 paracrine loop effectively reduces neointimal hyperplasia and preserves vascular patency in vivo.

1. Introduction

Bioresorbable scaffolds (BRS) manufactured from poly-lactic acid (PLA) represent a major shift in cardiovascular intervention [1], designed to provide transient mechanical support before safely degrading within the vessel wall [2]. However, the physical and chemical dissolution of PLA polymers fundamentally alters the local vascular microenvironment over time [3,4]. Recent polyester-based biomaterial studies further suggest that degradation kinetics and degradation-derived products can be engineered to modulate tissue remodeling and biological responses [5,6]. Unlike permanent metallic drug-eluting stents (DES), which primarily trigger acute, injury-induced inflammation, BRS degradation generates a sustained, active immune response driven directly by the continuous release of polymer breakdown products [7,8].

At the biomaterial-tissue interface, most efforts to understand host immunomodulation have focused heavily on macrophage polarization[[9], [10], [11], [12]]. Yet, the distinct role of innate immune cells, particularly neutrophils, during the advanced stages of scaffold remodeling remains largely overlooked. While traditionally considered early acute responders, emerging evidence suggests neutrophils actively shape chronic tissue remodeling[[13], [14], [15], [16]]. Specifically, the expulsion of neutrophil extracellular trap (NET), the web-like structures of decondensed chromatin and granule proteins, emerges as a critical variable in the host response to degrading polymers[[17], [18], [19]].

The ultimate success of a vascular implant depends on its ability to maintain luminal patency, which is governed by the behavior of adjacent vascular smooth muscle cells (VSMCs)[[20], [21], [22]]. Pathological VSMC proliferation and migration, characterized by a shift from a contractile to a pro-inflammatory synthetic phenotype, are the primary drivers of neointimal hyperplasia [23]. Defining the receptor-mediated pathways linking biomaterial-induced NET formation to VSMC phenotypic switching is critical for optimizing immunomodulatory scaffold design [[24], [25], [26]].

In this study, we map the long-term host-biomaterial interactions following PLA BRS implantation and present a targeted bioactive strategy to resolve advanced stage vascular inflammation. Using a combination of transcriptomic profiling in stented porcine arteries, and longitudinal spatial histology, we demonstrate that advanced PLA degradation establishes a NET-dominant immune microenvironment. Mechanistically, we reveal how PLA degradation products induce NET release via mitochondrial calcium overload, which in turn drives VSMC phenotypic switching through a STAT3-dependent TLR2/NF-κB signaling axis. To explore the in vivo therapeutic potential of this mechanism, we engineered a customized PLA scaffold functionalized with the selective TLR2 inhibitor C29. In a rabbit iliac paired implantation model, this immunomodulatory device reduced neointimal hyperplasia and attenuated adverse vascular remodeling, providing proof-of-concept support for local TLR2-targeted intervention.

2. Methods

2.1. Preparation and characterization of customized bioresorbable scaffolds

Customized poly-lactic acid (PLA) bioresorbable scaffolds were manufactured and supplied by Shandong Huaan Biotechnology to evaluate local immunomodulation by the targeted intervention. To avoid confounding effects from conventional anti-proliferative drugs (e.g., rapamycin), experimental scaffolds were coated exclusively with the toll-like receptor 2 (TLR2) inhibitor C29. Given the high lipophilicity of C29, the coating solution was prepared by dissolving 1 mg C29 per scaffold in acetone together with a PLA carrier polymer to form a drug-polymer matrix suitable for deposition. The formulation was applied to the outer surface of the PLA scaffold using a proprietary ultrasonic spraying process performed by the manufacturer to achieve uniform coating coverage and a reproducible drug-elution profile.

Surface morphology and coating integrity were examined by high-resolution scanning electron microscopy (Zeiss, GeminiSEM 560, Germany) after balloon expansion of the C29-coated PLA scaffold operated at an accelerating voltage of 5.0 kV. SEM imaging confirmed the formation of the intended phase-separated surface layer on C29-coated scaffolds, consistent with a microdomain architecture designed to support degradation-coupled release behavior.

In vitro C29 release kinetics were evaluated by immersing individual C29-coated PLA scaffolds in 1000 mL HBSS at 37 °C under shaking conditions with 200 rpm. At predefined time points (0, 20 min, 40 min, 1 h, 2 h, 4 h, 8 h, 1 day, 2 days, 3 days, 4 days, 5 days, 7 days, 9 days, 11 days, 14 days, 18 days and 30 days), 600 μL of release medium was collected and replaced with an equal volume of fresh medium. The released C29 concentration was quantified using Triple Quad 7500 LC-MS/MS System (SCIEX, Framingham, MA, USA).

2.2. Animal models and surgical interventions

All procedures were performed in accordance with institutional ethical guidelines at the Gateway Medical Innovation Center and were approved by the Gateway Medical Innovation Center Institutional Animal Care and Use Committee. To investigate long-term transcriptomic remodeling after stent implantation, a 9-month coronary implantation model was established in miniature pigs (35-45 kg). Animals were anesthetized with a Zoletil 50 based regimen combined with isoflurane to maintain adequate anesthesia throughout the intervention. Vascular access was obtained via the right femoral artery for device delivery. Miniature pigs were randomly assigned to receive a commercially available bioresorbable scaffold provided by Shandong Huaan Biotechnology, a permanent metallic drug-eluting stent, or plain old balloon angioplasty in the target coronary artery (n = 3 biological replicates per group). At 9 months, animals were euthanized, and the stented coronary segments were dissected free of perivascular adipose tissue and immediately snap-frozen in liquid nitrogen to preserve RNA integrity for downstream transcriptome sequencing.

For longitudinal time-course analyses of scaffold degradation and local inflammation, commercially available bioresorbable scaffolds provided by Shandong Huaan Biotechnology were implanted into the coronary arteries of an independent cohort of miniature pigs using the same anesthetic regimen and right femoral access strategy. Animals were euthanized at 1, 3, 6, and 12 months post-implantation (n = 6 per time point). Stented vessel segments were harvested, fixed in paraformaldehyde overnight, and paraffin-embedded for sectioning.

To evaluate the in vivo therapeutic performance of customized C29-coated scaffolds, an independent 6-month bilateral iliac artery stenting model was established in rabbits (3.0-3.5 kg). Following induction of general anesthesia with Zoletil 50 and isoflurane, the right neck was surgically prepared to expose the right carotid artery, and a 6F vascular sheath (Terumo, RS*A60G07SQ, Japan) was introduced to establish arterial access. Under angiographic guidance, target segments within the bilateral iliac arteries were selected for implantation. Each rabbit received one bare PLA bioresorbable scaffold and one C29-coated PLA scaffold deployed into contralateral iliac arteries (left/right) to enable an internally paired comparison. After 6 months in vivo, bilateral iliac angiography and intravascular optical coherence tomography (OCT; Forssmann, China) were performed to assess luminal patency and structural remodeling. OCT pullback was conducted at 20 mm/s to acquire contiguous cross-sectional images, and quantitative analysis was performed using OCTplus (Pulse Medical Imaging Technology, China) to calculate luminal stenosis rate. Animals were then euthanized, and stented arterial segments were excised, fixed overnight in paraformaldehyde, and paraffin-embedded for subsequent histopathology and immunofluorescence analyses.

2.3. Immunofluorescence assessment

Multiplex immunofluorescence was performed to resolve the spatial organization of immune-cell infiltration and associated vascular remodeling within the peri-strut microenvironment. Sections underwent antigen retrieval and were incubated with a validated primary antibody panel for lineage and activation mapping: anti-myeloperoxidase (Proteintech, 22225-1-AP; 1:1000), anti-CD19 (Proteintech, 27949-1-AP; 1:200), anti-CD3 (Proteintech, 81324-1-RR; 1:200), and anti-CD14 (Proteintech, 17000-1-AP; 1:200). Neutrophil populations and NET-associated epitopes were probed using anti-Rabbit T/Neutrophil antibody (Santa Cruz Biotechnology, sc-59376; 1:20) and anti-citrullinated histone H3 (Cell Signaling Technology, 97272; 1:1000). To delineate vascular smooth muscle cell architecture and interrogate the candidate mechanistic receptor target, sections were stained with anti-α-smooth muscle actin (Boster, BM0002; 1:1000) and anti-toll-like receptor 2 (Bioss, bs-1019R; 1:100). Due to species-of-origin conflicts among the primary antibodies, and to improve sensitivity and spatial resolution, particularly for low-abundance targets within the peri-strut regions, selected sections were processed using tyramide signal amplification techniques.

To relate scaffold strut degradation to local innate immune activation, degradation status was classified directly from immunofluorescence images using predefined morphology-based criteria [27]. Because fully resorbed struts cannot be reliably delineated by fluorescence microscopy within remodeled tissue, the “dissolved box” stage was not quantified and was excluded from statistical analyses. Accordingly, struts were categorized as either “preserved box” or “open box.” Preserved box struts displayed a typical box-like profile with sharply defined, continuous borders and preserved geometric integrity, consistent with minimal in vivo degradation. In contrast, open box struts exhibited disrupted or irregular contours with partial loss of the box geometry. To capture progression within this phenotype, open box struts were further stratified into early and advanced stages based on the extent of cellular and/or extracellular matrix infiltration into the strut void: early open box showed limited focal invasion, whereas late open box showed extensive invasion with broad occupation of the degrading polymer space.

2.4. Cell culture cellular differentiation and in vitro model preparation

Primary human aortic smooth muscle cells (VSMCs; ScienCell, 6110, USA) were cultured under standard conditions in smooth muscle cell medium (ScienCell, 1101) supplemented with 2% fetal bovine serum, 1% smooth muscle cell growth supplement, and 1% penicillin/streptomycin. Cells were used between passages 5-10 to minimize phenotypic drift across functional assays. For stimulation experiments, the VSMCs were treated with purified NET with the final concentration of 500 ng/mL for 24 h [28].

To evaluate innate immune readouts, peripheral blood samples were collected from healthy volunteers who provided informed consent, in accordance with protocols approved by the Institutional Review Board. Primary human neutrophils were subsequently isolated from these samples using Percoll density-gradient centrifugation [29]. Briefly, whole blood was diluted 1:1 with RPMI 1640, layered onto a discontinuous Percoll gradient (62% and 75%), centrifuged, and the neutrophil layer was collected, washed, and immediately resuspended for experiments. In parallel, National Collection of Authenticated Cell Cultures-sourced HL-60 cells were cultured with IMDM with 20% fetal bovine serum and 1% penicillin/streptomycin then differentiated into neutrophil-like cells (dHL-60) by treatment with all-trans retinoic acid (Sangon, A506728, 2 μM) for 5 days.

To model the sustained local chemical microenvironment associated with bioresorbable scaffold breakdown, PLA degradation products were generated using an accelerated hydrolysis protocol. PLA particles were fully immersed in phosphate-buffered saline (PBS) and incubated under gentle agitation at 57 °C for 12 weeks. The suspension was then centrifuged (100 × g, 5 min) to remove residual polymer, and the supernatant containing soluble degradation products was collected and stored at −80 °C for subsequent stimulation studies.

NETs were prepared from primary neutrophils to probe NET-mediated paracrine effects [30]. Freshly isolated neutrophils were seeded and stimulated with phorbol 12-myristate 13-acetate (PMA; Beyotime, S1819; 500 nM) for 6 h to induce NET release. After stimulation, cultures were washed three times with sterile PBS to remove residual agonist, and extracellular matrices were gently scraped and collected. NET suspensions were purified by sequential centrifugation (450 × g, 10 min, 4 °C) to remove intact cells/debris, followed by high-speed centrifugation (18,000 × g, 4 °C) to pellet NET material. Pellets were resuspended in sterile PBS (100 μL per 1×10^7 neutrophils seeded). NET yield was quantified as double-stranded DNA using a dsDNA Quantitation Reagent (Yeason, C12641ES01) according to the manufacturer's instructions, and DNA concentration was normalized across experiments.

2.5. In vitro functional and mechanistic cellular assays

Proliferation ability of VSMC was quantified using a 5-ethynyl-2′-deoxyuridine (EdU) incorporation assay (Vazyme, A411-01) according to the manufacturer's instructions. VSMC motility was evaluated using complementary migration readouts. For collective migration, a wound-healing scratch assay was performed by mechanically disrupting confluent monolayers; phase-contrast images were acquired at 0 and 24 h, and wound closure was quantified in ImageJ as: [(A0 - A24)/A_0]×100%, where A denotes scratch area. For chemotactic migration, a transwell assay was conducted using permeable inserts (Corning, 3422). Cells were seeded in the upper chamber and allowed to migrate for 24 h toward complete medium supplemented with 10% fetal bovine serum in the lower chamber, after which migrated cells were fixed, stained, and counted microscopically.

VSMC contractility was assessed using a collagen gel contraction assay. Briefly, 5 × 10^4 cells VSMCs were mixed with collagen and culture medium to achieve a final collagen concentration of 1 mg/mL, followed by neutralization with 1 M NaOH. The mixture was plated into 24-well plates, allowed to polymerize at room temperature for 20 min, overlaid with complete medium containing the indicated treatments, and gently detached from the well edge. After 24 h, gels were imaged under a stereomicroscope, stained with 1% crystal violet for 10 min, washed with PBS, and photographed for contraction quantification.

To probe immune-vascular cell interactions relevant to localized inflammation, a neutrophil adhesion assay was performed. Primary human neutrophils were labeled with DiI Red (Beyotime, C1036; 10 μM) for 20 min at 37 °C and washed twice with PBS. 10^5 Labeled neutrophils were then added to VSMC monolayers that were either unstimulated or preconditioned with purified NET, with or without C29 treatment for 24 h. After 2 h of co-culture, non-adherent cells were removed, monolayers were washed twice, and adherent neutrophils were imaged by fluorescence microscopy and quantified.

Mitochondrial depolarization in dHL-60 cells was assessed to interrogate mechanisms linked to NET formation. dHL-60 cells were exposed to dPLA products (stock lactate concentration 0.5 mM with a pH 6.8) diluted 1:100 in culture medium; PBS served as the vehicle control. After 24 h, mitochondrial membrane potential was measured using a JC-1 probe kit (Beyotime, C2006) following the manufacturer's protocol. Cells were stained for 20 min, washed twice, and analyzed on a BD FACSAria III flow cytometer by detecting JC-1 aggregates and monomers in the PE and FITC channels, respectively. Flow cytometry data were processed in FlowJo (BD, v10.8).

Immunofluorescence staining of primary human neutrophils was performed after stimulation with dPLA products. Briefly, isolated dPLA products (stock lactate concentration: 0.5 mM with a pH 6.8) were diluted 1:100 in complete RPMI 1640 medium (Gibco, 11875119) containing 10% FBS and applied to neutrophils for 10 h. Cells were then fixed in paraformaldehyde for 10 min at room temperature, washed three times with PBS, blocked with blocking buffer (Beyotime, P0102) for 60 min, and incubated overnight at 4 °C with an anti-citrullinated histone H3 (H3Cit) primary antibody (Abcam, ab5103; 1:1000). After washing, cells were incubated with the appropriate fluorescent secondary antibody and counterstained with DAPI for nuclear visualization. Fluorescence images were acquired using an Olympus FV3000 confocal microscope.

2.5.1. Molecular biology, gene silencing, and pharmacological interventions western blotting

To profile protein expression and signaling activation, cells and stented arterial tissues were lysed in RIPA buffer (Beyotime, P0013B) supplemented with protease and phosphatase inhibitors. Protein concentration was determined using a BCA assay (Beyotime, P0012). Equal amounts of protein were separated by SDS-PAGE and transferred to PVDF membranes. After blocking, membranes were incubated overnight at 4 °C with primary antibodies against VSMC phenotypic markers, NET-related proteins, mitochondrial Ca²⁺ uniporter components, and inflammatory signaling mediators: osteopontin (Proteintech, 22952-1-AP; 1:1000), SM22α (Proteintech, 10493-1-AP; 1:5000), α-smooth muscle actin (Proteintech, 67735-1-Ig and 14395-1-AP; both 1:3000); PADI4 (Proteintech, 17373-1-AP; 1:1000; Abcam, ab128086; 1:2000), citrullinated histone H3 (Abcam, ab5103; 1:1000; Cell Signaling Technology, 97272; 1:1000), and histone H3 (Proteintech, 17168-1-AP; 1:5000); MCU (Proteintech, 26312-1-AP; 1:1000), MICU1 (Immunoway, YM3430; 1:1000), and MICU2 (Huabio, HA721742; 1:1000); TLR2 (Bioss, bs-1019R; 1:1000) and ICAM-1 (Proteintech, 10831-1-AP; 1:5000). Canonical NF-κB signaling was assessed using p-p65 (Affinity, AF2006; 1:1000) for VSMCs and p-p65 (Cell Signaling Technology, 3033, 1:1000) for rabbit vascular tissues), total p65 (Abcam, ab16502; 1:2000) for VSMCs and mouse monoclonal p65 antibody (Cell Signaling Technology, 6956, 1:1000) for vascular tissues, p-IκB (Affinity, AF2002; 1:1000), and total IκB (Affinity, AF5002; 1:1000). STAT3 activation was examined using p-STAT3 (Tyr705; Cell Signaling Technology, 9145; 1:1000) for VSMCs and the vascular tissues, p-STAT3 (Ser727; Selleck, F1212; 1:1000), and total STAT3 (Cell Signaling Technology, 9139; 1:1000) for VSMCs and the vascular tissues. β-actin was used as a loading control (Proteintech, 66009-1-Ig; 1:5000; and Proteintech, 81115-1-RR; 1:5000). After primary incubation, membranes were washed and probed with HRP-conjugated secondary antibodies, while rabbit monoclonal p-p65 antibody on the rabbits’ tissues, HRP-conjugated light-chain specific secondary antibody (Proteintech, SA00001-7L, 1:10000) was used, then followed by chemiluminescent detection.

2.5.2. RNA extraction and qRT-PCR

Total RNA was isolated from treated VSMCs using an RNA isolation kit (Vazyme, RC102-01), reverse-transcribed to cDNA (Takara, RR036A), and quantified by real-time PCR using a SYBR master mix (Takara, RR820A). Primer sequences were as follows: TLR2 (F: CTTCACTCAGGAGCAGCAAGCA; R: ACACCAGTGCTGTCCTGTGACA), ICAM1 (F: AGCGGCTGACGTGTGCAGTAAT; R: TCTGAGACCTCTGGCTTCGTCA), IL6 (F: AGACAGCCACTCACCTCTTCAG; R: TTCTGCCAGTGCCTCTTTGCTG), CSF3 (F: TCCAGGAGAAGCTGGTGAGTGA; R: CGCTATGGAGTTGGCTCAAGCA), CCL20 (F: AAGTTGTCTGTGTGCGCAAATCC; R: CCATTCCAGAAAAGCCACAGTTTT), and CXCL8 (F: GAGAGTGATTGAGAGTGGACCAC; R: CACAACCCTCTGCACCCAGTTT). Relative mRNA expression was calculated using the 2^-ΔΔCt method and normalized to ACTB (F: CACCATTGGCAATGAGCGGTTC; R: AGGTCTTTGCGGATGTCCACGT).

For qRT-PCR analysis of in vivo stented arterial tissues, total RNA was extracted from harvested rabbit iliac artery segments using RNA isolation kit (Vazyme, RC122-01) according to the manufacturer's instructions. RNA was also reverse-transcribed into cDNA using kit (Takara, RR036A), and quantitative PCR was performed using SYBR master mix (Takara, RR820A). Relative mRNA expression of IL6, ICAM1, and TLR2 was calculated using the 2^-ΔΔCt method and normalized to ACTB. Rabbit-specific primer sequences were: IL6 (F: CCCTGAGGCCAAAGGTCAAG; R: CGCAGGATGAAGTGGATCGT), ICAM1 (F: ACAGCAAGGACACGGTCAG; R: TTGGGAGCTGGGAAGCTGTA), TLR2 (F: CGGGTTCCCCAGGTTGGA; R: CACAAGACAGAGAGGGCAGG), ACTB (F: CTGGAACGGTGAAGGTGACA, R: CGGCCACATTGCAGAACTTT).

2.5.3. siRNA-mediated gene silencing

For targeted knockdown, VSMCs were transfected with pooled siRNAs (each pool comprising three duplexes mixed at equimolar ratios) at a final siRNA concentration of 50 nM in the culture medium, against TLR2 and STAT3 using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015) in Opti-MEM (Gibco, 31985070) according to the manufacturer's protocol. TLR2 siRNA sequences were: siTLR2-1 (sense: GCGAUAAAGUCCUAGGUUCdTdT; antisense: GAACCUAGGACUUUAUCGCdTdT), siTLR2-2 (sense: CAAUCCGGAGGCUGCAUAUdTdT; antisense: AUAUGCAGCCUCCGGAUUGdTdT), and siTLR2-3 (sense: GACCGCAAUGGUAUCUGCAdTdT; antisense: UGCAGAUACCAUUGCGGUCdTdT). STAT3 siRNA sequences were: siSTAT3-1 (sense: CACAAUCUACGAAGAAUCAdTdT; antisense: UGAUUCUUCGUAGAUUGUGdTdT), siSTAT3-2 (sense: GUCAUUAGCAGAAUCUCAAdTdT; antisense: UUGAGAUUCUGCUAAUGACdTdT), and siSTAT3-3 (sense: CCAACAAUCCCAAGAAUGUdTdT; antisense: ACAUUCUUGGGAUUGUUGGdTdT).

2.5.4. Pharmacological interventions

For pathway interrogation, small-molecule inhibitors were applied at the indicated final concentrations: the mitochondrial calcium uniporter inhibitor MCU-i4 (MedChemExpress, HY-138620; 10 μM), the NF-κB p65 translocation inhibitor JSH-23 (Selleck, S7351; 10 μM), and the TLR2 inhibitor C29 (Selleck, S6597; 0.5 μM) for evaluating coating-related therapeutic modulation under activated VSMC conditions.

Cell viability after C29 exposure was assessed using a Cell Counting Kit-8 assay. VSMCs were seeded into 96-well plates at 2000 cells per well and treated with C29 at concentration of 0, 10 nM, 0.1 μM, 0.5 μM, 1 μM, 10 μM, 100 μM, 200 μM and 500 μM for 24 h. CCK-8 reagent (C0037, Beyotime) was then added and incubated for 1 h at 37 °C. Absorbance was measured at 450 nm using a microplate reader, and relative cell viability was calculated after normalization to the vehicle-treated control group.

2.6. Bioinformatics and transcriptomic data analysis

Bulk RNA sequencing. RNA extraction, library construction, and high-throughput sequencing was performed for stented porcine coronary arteries and human primary VSMCs in collaboration with LC-Bio Technology. For cross-species, multiscale integration, porcine transcripts were mapped to human orthologs using the Pig RNA Atlas ortholog conversion framework to enable direct comparison with human in vitro datasets [31]. After read-level quality control and alignment, differential expression analysis was conducted in R using DESeq2 (v1.38.3); genes with |log2(fold change)| > 1 and adjusted p < 0.01 were defined as differentially expressed. Transcriptomic shifts were visualized using volcano plots (EnhancedVolcano, v1.16.0) and hierarchical clustering heatmaps (ComplexHeatmap, v2.14.0) to highlight cohort-specific expression patterns, including immune-lineage marker modules. Functional interpretation was performed using Gene Ontology and KEGG enrichment analyses implemented in clusterProfiler(v4.0.5), and pathway-directionality was evaluated by gene set enrichment analysis (GSEA), reporting normalized enrichment scores and false discovery rates for NET-related and neutrophil chemotaxis programs. To quantify local innate immune burden at the sample level, a NET score was calculated by single-sample GSEA (ssGSEA) using a predefined NET gene signature (MPO, ELANE, PRTN3, CTSG, CEACAM8, S100A8, S100A9, S100A12, LTF, MMP9). To prioritize candidate receptors linking biomaterial-associated inflammation to VSMC dysfunction, we intersected upregulated genes from stented porcine vessels and NET-treated human VSMCs with GO receptor-complex annotations to prioritize candidate receptors (visualized by Venn diagram).

2.7. JASPAR prediction and dual-luciferase reporter assay

To infer upstream transcriptional control of the identified inflammatory axis, the 2000-bp upstream promoter region of the human TLR2 locus was scanned using the JASPAR database to predict high-affinity STAT3 binding motifs.

Dual-luciferase reporter assays were performed to validate the transcriptional activity of predicted STAT3-binding motifs within the human TLR2 promoter. The wild-type TLR2 promoter fragment (−1000 to +50 bp), containing the predicted STAT3-binding sites, was amplified and cloned into the pGL4.1 vector. Site-directed mutagenesis was used to generate Mut1 (−185 to −177 bp, 5′-TGCTGGGAA-3′ to 5′-TGCTGAACC-3′), Mut2 (−145 to −137 bp, 5′-TGCCTGGAA-3′ to 5′-TGCCTAACC-3′), and Mut1&2, which carried simultaneous mutations at both predicted STAT3-binding motifs. Cells cultured in 96-well plates were co-transfected with 50 ng wild-type or mutant TLR2 promoter reporter plasmids, 10 ng pEnter-STAT3 or pEnter-empty vector plasmids, and 2 ng pGL4.73 using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015) in Opti-MEM (Gibco, 31985070) according to the manufacturer's protocol. After 36 h, firefly and Renilla luciferase activities were measured using a dual-luciferase reporter assay system (Vazyme, DD1205) according to the manufacturer's instructions. Relative promoter activity was calculated by normalizing firefly luciferase activity to Renilla luciferase activity.

2.8. Statistical analysis and data presentation

All quantitative data from in vitro and in vivo experiments were analyzed and visualized using GraphPad Prism (v10.4.1) and R (v4.2.2). Continuous variables are presented as mean ± SEM. Normality was assessed using the Shapiro-Wilk test, and homogeneity of variance was evaluated using the Brown-Forsythe and Bartlett's tests. For datasets meeting assumptions of normality and equal variance, two-group comparisons were performed using a two-tailed unpaired Student's t-test, and multi-group comparisons were analyzed by one-way ANOVA followed by Tukey's post hoc test. When variance homogeneity was violated, Welch's t-test was used for two-group comparisons and Welch's ANOVA followed by Dunnett's T3 multiple-comparisons test was applied for multi-group analyses. A two-tailed p < 0.05 was considered statistically significant for all analyses.

3. Results

Transcriptomic profiling and histological validation identify neutrophil extracellular trap formation as a dominant immune response following bioresorbable scaffold implantation.

To define the long-term, localized immune microenvironment induced by bioresorbable scaffolds, we performed bulk RNA sequencing on stented porcine coronary arteries harvested nine months after stent implantation. To enable comprehensive functional interpretation and strengthen clinical translational relevance, porcine gene expression profiles were humanized by ortholog conversion prior to downstream computational analyses.

Differential expression analysis comparing the PLA BRS and plain balloon angioplasty (CON) groups identified 438 upregulated inflammation-associated genes (Fig. 1A). Because sustained immune infiltration is central to advanced stage scaffold degradation and vascular remodeling, we generated a heatmap of representative immune lineage markers across CON, metal drug-eluting stents (DES), and BRS groups (Fig. 1B). Neutrophil, T-cell, and B-cell markers were higher in the BRS group, indicating that PLA scaffold degradation was accompanied by a broad immune-cell response rather than activation of a single leukocyte lineage.

Fig. 1.

Poly-lactic acid (PLA) bioresorbable scaffolds elicit a pronounced neutrophil-driven inflammatory response and promote neutrophil extracellular trap (NET) formation. (A) Volcano plot illustrating differentially expressed genes (DEGs) in stented vessels. Red and blue points represent significantly upregulated (438 genes) and downregulated (124 genes) genes, respectively, in the bioresorbable scaffold (BRS) group compared to controls. Horizontal and vertical dashed lines indicate the significance threshold (q-value cutoff is 0.01 and |log2FC| cutoff is 1). (B) Hierarchical clustering heatmap demonstrating the relative transcriptomic expression (Z-score) of canonical marker genes associated with major immune cell populations across the control (CON), drug-eluting stent (DES), and BRS groups. (C-F) Representative immunofluorescence images of stented arterial cross-sections evaluating immune cell infiltration into the vascular smooth muscle cell layer following scaffold degradation. Sections are co-stained for α-SMA (green) to identify vascular smooth muscle cells, DAPI (blue) for nuclei, and specific immune markers (red): (C) CD19 (B cells), (D) CD3 (T cells), (E) CD14 (monocytes), and (F) MPO (neutrophils). Scaffold struts are visible as distinct rectangular voids. Scale bars: 200 μm. (G) Dot plot of KEGG pathway enrichment analysis for the upregulated DEGs, highlighting Neutrophil extracellular trap formation (red text) as a highly enriched signaling cascade. Node size corresponds to gene count, and color gradient indicates the adjusted P value. (H) Gene Set Enrichment Analysis (GSEA) plot confirming the significant positive enrichment of the NET formation pathway (hsa04613) in the BRS group (Normalized Enrichment Score [NES] = 1.484, FDR = 0.025, P = 0.012). (I) Violin plot quantifying the relative NET scores across the CON, DES, and BRS groups. The BRS group exhibits a significantly elevated NET score. Data are analyzed via one-way ANOVA and Tukey's multiple comparison test (* indicates P < 0.05).

We then validated these transcriptomic findings at the tissue level by immunofluorescence staining of arterial cross-sections to map the distribution of leukocyte subsets. Staining confirmed peri-strut accumulation of adaptive immune cells and innate phagocytes after bioresorbable scaffold implantation; however, myeloperoxidase, reflecting neutrophil accumulation and neutrophil extracellular trap (NET) formation, was the most prominent and visually abundant marker within the adventitia and neointima, as illustrated in Fig. 1C through 1F.

To investigate the mechanisms driving this innate immune response, we performed pathway enrichment analysis on the transcriptomic data. Kyoto Encyclopedia of Genes and Genomes analysis of the upregulated gene set identified “neutrophil extracellular trap formation”, together with other key inflammatory cascades, among the most significantly enriched pathways, as shown in Fig. 1G. Enrichment results for downregulated genes (Fig. S1A) included “cytoskeleton in muscle cells”, suggesting ongoing remodeling. Gene set enrichment analysis further corroborated strong induction of the NET formation signature in the bioresorbable scaffold group relative to controls, yielding a highly significant normalized enrichment score in Fig. 1H.

Consistently, a relative NET score derived by single-sample gene set enrichment analysis (ssGSEA) was significantly increased in the BRS group compared with CON group and permanent metal DES group, as depicted in Fig. 1I. These data indicate that advanced PLA degradation is associated with neutrophil recruitment and increased NET formation.

3.1. Progressive degradation of the bioresorbable scaffold is associated with neutrophil infiltration and local neutrophil extracellular trap formation

To delineate temporal kinetics of innate immune activation after bioresorbable scaffold implantation, we performed longitudinal histological analyses of myeloperoxidase-positive neutrophil infiltration at 1, 3, 6, and 12 months post-procedure. As shown in Fig. 2A, MPO-positive neutrophil infiltration was minimal at 1 and 3 months, became apparent at 6 months, and was most prominent at 12 months, coinciding with the period of maximal structural degradation of the PLA polymer. Quantification confirmed delayed inflammatory kinetics, with a statistically significant expansion of myeloperoxidase-positive area at later time points compared with unstented controls (Fig. 2B).

Fig. 2.

Neutrophil accumulation and systemic NET release correlate with the progressive degradation of polylactic acid (PLA) bioresorbable scaffolds. (A) Representative immunofluorescence images of stented arterial cross-sections over a 12-month time course following implantation, compared to unstented controls. Tissues are stained for the neutrophil marker myeloperoxidase (MPO, red) and nuclei (DAPI, blue) to track neutrophil infiltration during scaffold degradation. Scale bars: 500 μm (Overview, top row) and 200 μm (Magnified, middle and bottom rows). (B) Quantification of the MPO-positive area percentage across the 12-month time course. Statistical significance was determined using Welch ANOVA (*P < 0.05, ****P < 0.0001). n = 6 for each group. (C) High-magnification immunofluorescence images detailing in situ NET formation across distinct morphological stages of scaffold degradation (Preserved Box, n = 8, Early Open Box, n = 7, and Late Open Box, n = 13). Sections are co-stained for MPO (red) and the specific NET marker citrullinated histone H3 (H3Cit, green). Scale bar: 100 μm. (D) Quantitative analysis of NET formation, calculated as the MPO/H3Cit double-positive area per field across the progressive degradation stages. Data were analyzed via Welch ANOVA (**P < 0.01) and presented as mean ± SEM.

Because the longitudinal MPO analysis showed that neutrophil accumulation became evident at the later observation time points, we next focused on peri-strut regions with clear MPO-positive infiltration and performed co-staining for myeloperoxidase and citrullinated histone H3 to determine whether these infiltrating neutrophils were associated with NET formation. MPO was used to localize neutrophil-associated regions, whereas H3Cit was used as a NET-associated marker of histone citrullination and chromatin decondensation; therefore, MPO/H3Cit dual positivity was interpreted as evidence of NET-associated inflammatory activity rather than neutrophil infiltration alone [32]. Because the scaffold presents a three-dimensional architecture in vivo whereas histological assessment relies on two-dimensional sections, we selected large strut cross-sections with clearly defined surrounding structures to enable consistent morphological classification and robust quantification. Based on strut morphology in immunofluorescence images, degradation status was classified into three stages, termed preserved box, early open box, and late open box, the latter representing near-complete dissolution. Preserved boxes exhibited minimal degradation with intact architecture and sharp boundaries, early open boxes showed moderate degradation with irregular contours and blurred borders accompanied by cellular or extracellular matrix invasion, and late open boxes exhibited severe degradation with substantial cellular ingrowth and highly irregular residual boundaries. Within these peri-strut regions, colocalization of myeloperoxidase and citrullinated histone H3 showed that NET-associated signals were most prominent around advanced degradation morphologies (Fig. 2C), and quantification of the dual-positive area showed a significant increase in late open boxes relative to preserved boxes. (Fig. 2D).

3.2. Degraded poly-lactic acid promotes neutrophil extracellular trap formation via mitochondrial calcium uniporter-mediated calcium overload

The in vivo data suggested a close association between PLA degradation and both neutrophil accumulation and NET formation. Given the fragility, short lifespan, and donor-dependent variability of primary human neutrophils, we generated PLA degradation products and co-cultured them with HL-60 cells differentiated into neutrophil-like cells using all-trans retinoic acid to establish a reproducible model for biochemical assays. After 24 h of stimulation, western blotting demonstrated significant upregulation of key enzymatic drivers of NET formation, including PADI4 and citrullinated histone H3, relative to unstimulated controls (Fig. 3A and B).

Fig. 3.

Degraded poly-lactic acid (dPLA) induces neutrophil extracellular trap (NET) formation via activation of the mitochondrial calcium uniporter (MCU). (A) Representative Western blot analysis of H3Cit and PADI4 protein expression in neutrophils treated with vehicle control or degraded PLA (dPLA). Total H3 and β-Actin served as loading controls. (B) Densitometric quantification of relative H3Cit and PADI4 protein levels. Data were analyzed using Welch's t-test (H3Cit, PADI4) (*P < 0.05, **P < 0.01, n = 6 for each group). (C) Representative immunofluorescence images of NET formation following dPLA exposure. Cells were stained for the NET marker H3Cit (green) and nuclei (DAPI, blue). Decondensed DNA and extracellular web-like structures are distinctly visible in the dPLA group. Scale bar: 100 μm. (D) Quantitative analysis of the percentage of cells exhibiting decondensed nuclei with positive H3Cit staining. Data were analyzed via unpaired t-test (****P < 0.0001, n = 6 for each group). (E) Flow cytometry analysis evaluating mitochondrial membrane potential using JC-1 dye. The shift in the FITC/PE fluorescence ratio indicates mitochondrial depolarization during dPLA-induced NETosis. (F) Quantification of the percentage of PE + FITC + cells from the flow cytometry analysis. Data were analyzed via unpaired t-test (***P < 0.001, n = 6 for each group). (G) Representative Western blot of mitochondrial calcium uniporter complex proteins (MCU, MICU1, and MICU2) in control and dPLA-treated cells. (H) Densitometric quantification of relative MCU, MICU1, and MICU2 expression. Data were analyzed using an unpaired t-test (****P < 0.0001; ns, not significant, n = 6 for each group). (I) Representative Western blot illustrating the effect of pharmacological MCU inhibition on NETosis markers. Cells were pre-treated with the MCU inhibitor MCU-i4 or vehicle (DMSO) prior to dPLA exposure. (J) Densitometric quantification of relative H3Cit and PADI4 levels. The addition of MCU-i4 significantly attenuated dPLA-induced upregulation of NET markers. Statistical significance across multiple groups was determined using Welch ANOVA (**P < 0.01, ***P < 0.001, ****P < 0.0001, n = 6 for each group). Data in all bar graphs are presented as mean ± SEM.

We then validated classical morphological hallmarks of NET formation using primary neutrophils isolated from healthy donor peripheral blood. Immunofluorescence imaging showed marked nuclear swelling, increased extracellular DNA structures, and elevated citrullinated histone H3-positive signals in degradation product-treated cells compared with controls, with quantification provided in Fig. 3C and D. These complementary results indicate that scaffold breakdown products can directly trigger chromatin decondensation and active NET release.

To explore intracellular mechanisms, we assessed mitochondrial status using the JC-1 fluorescent probe. Flow cytometry (Fig. S1B) revealed a significant loss of mitochondrial membrane potential upon degradation product stimulation, evidenced by a shift from red JC-1 aggregates to green monomers (Fig. 3E and F).

Given the central role of ion homeostasis in mitochondrial integrity, we examined expression of the mitochondrial calcium uniporter complex. Western blotting showed significant upregulation of the pore-forming subunit MCU, while the regulatory subunits MICU1 and MICU2 were not significantly altered, consistent with a state of mitochondrial calcium overload (Fig. 3G and H). To establish functional necessity, we treated cells with the MCU inhibitor MCU-i4 at 10 μM, which markedly suppressed induction of PADI4 and citrullinated histone H3 and rescued cells from degradation product-induced NET formation (Fig. 3I and J).

3.3. Neutrophil extracellular traps directly promote proliferation, migration, and phenotypic switching of human vascular smooth muscle cells

To define how neutrophil recruitment and NET formation contribute to vascular remodeling, we established a controlled in vitro model using purified NET generated from human primary neutrophils. Isolated NETs were co-cultured with human aorta vascular smooth muscle cells (VSMC) at 500 ng/mL for 24 h, after which we evaluated proliferative and migratory behaviors that underlie neointimal hyperplasia. NET-exposure significantly increased VSMC proliferation, as demonstrated by enhanced DNA synthesis in the EdU incorporation assay (Fig. 4A and B). Contractility assays further showed that NET-treated VSMCs exhibited impaired contractile capacity, as reflected by a significantly larger residual gel area compared with control cells (Fig. 4C and D). Migration was also accelerated, supported by faster wound closure in scratch assays (Fig. 4E and F) and increased cell traversal in transwell assays (Fig. 4G and H).

Fig. 4.

Neutrophil extracellular trap (NET) drive the proliferation, migration, and phenotypic switching of vascular smooth muscle cells. (A) Representative fluorescence images of EdU incorporation in vascular smooth muscle cells treated with vehicle control or isolated NET. Proliferating cells are indicated by EdU uptake (green), and nuclei are counterstained with DAPI (blue). Scale bar: 200 μm. (B) Quantification of the percentage of EdU-positive cells, demonstrating enhanced proliferation in the presence of NET. Data were analyzed via unpaired t-test (***P < 0.001, n = 6 for each group). (C) Representative images of the collagen gel contraction assay showing the contracted gel area in vascular smooth muscle cells treated with vehicle control or isolated NET. (D) Quantitative analysis of gel area percentage, indicating that NET treatment significantly altered gel contraction. Data were analyzed via unpaired t-test (P < 0.01, n = 6 for each group). (E) Representative bright-field images of the wound-healing (scratch) assay at 0 and 24 h, illustrating cell migration into the denuded area. Red outlines delineate the leading edge of the migrating cells. Scale bar: 200 μm. (F) Quantitative analysis of relative wound closure percentage at 24 h. Data were analyzed via unpaired t-test (****P < 0.0001, n = 6 for each group). (G) Representative images of crystal violet-stained cells that successfully migrated through Transwell inserts in response to control or NET-conditioned media. Scale bar: 300 μm. (H) Quantification of the number of migrated cells per field in the Transwell assay. Data were analyzed via unpaired t-test (****P < 0.0001, n = 6 for each group). (I) Representative Western blot analysis assessing the expression of key phenotypic switching markers. Cells treated with NET exhibit upregulation of the synthetic marker osteopontin (OPN) and concomitant downregulation of the contractile markers SM22α and α-SMA. β-Actin served as an internal loading control. (J) Densitometric quantification of relative OPN, SM22α, and α-SMA protein levels. Statistical significance was determined using an unpaired t-test for OPN and α-SMA, and Welch's t-test for SM22α (**P < 0.01, ***P < 0.001, n = 6 for each group). Data in all bar graphs are presented as mean ± SEM.

These functional changes suggested a shift toward a pathological synthetic phenotype. Western blotting showed increased osteopontin and reduced α-SMA and SM22α in NET-treated VSMCs (Fig. 4I and J). Collectively, these data suggest that NET promote VSMC proliferation, migration, and a shift toward a synthetic phenotype, linking scaffold-associated inflammation to vascular remodeling.

3.4. Neutrophil extracellular traps activate the TLR2 and NF-κB axis to induce a pro-inflammatory synthetic phenotype in vascular smooth muscle cells

To define the signaling programs underlying NET-induced smooth muscle cell reprogramming, we performed bulk RNA sequencing on VSMCs after stimulation with NET. Gene ontology enrichment analysis identified cell division and cell cycle pathways among the most significantly enriched terms, consistent with our functional evidence of increased proliferation, while immune and inflammatory response pathways were also prominently enriched, indicating conversion toward a pro-inflammatory synthetic phenotype (Fig. 5A). Among the top enriched pathways, neutrophil chemotaxis emerged as a notable signature, and gene set enrichment analysis confirmed upregulation of this chemotactic program in NET-treated cells (Fig. 5B), suggesting initiation of a maladaptive positive feedback loop that could recruit additional innate immune cells to the vascular lesion.

Fig. 5.

Neutrophil extracellular trap (NET) drive vascular smooth muscle cell (VSMC) proliferation and inflammatory crosstalk via the TLR2/NF-κB signaling axis. (A) Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEGs) in VSMCs following exposure to NET. The bubble plot highlights the top 5 enriched terms alongside other key immune and proliferative pathways. Bubble size represents the number of genes, and color gradient indicates the P value. (B) Gene Set Enrichment Analysis (GSEA) demonstrating significant positive enrichment of the Neutrophil Chemotaxis pathway in NET-treated VSMCs (Normalized Enrichment Score [NES] = 2.507, FDR <0.001, P < 0.001). (C) Venn diagram identifying TLR2 as the crucial overlapping candidate gene that is upregulated in vivo (stented vessels), upregulated in vitro (NET-treated VSMCs), and categorized under the Receptor Complex GO term (GO:0043235). (D) Representative in situ immunofluorescence images of stented arterial cross-sections confirming TLR2 expression in the neointima. Sections are co-stained for TLR2 (red), α-SMA (green, identifying VSMCs), and DAPI (blue, nuclei). Scale bar: 100 μm. (E) Volcano plot illustrating transcriptomic alterations in NET-treated VSMCs compared to controls. Red and blue points represent significantly upregulated (198 genes) and downregulated (291 genes) transcripts, respectively. Key inflammatory mediators, chemokines, and receptors (e.g., TLR2, IL6, CXCL8, IL1B) are highlighted. (F) Representative Western blot analysis validating the upregulation of TLR2 and the activation of the downstream NF-κB signaling pathway in VSMCs following NET exposure. Blots demonstrate increased phosphorylation of P65 (p-P65) and IκB (p-IκB) relative to total P65 and IκB. β-Actin served as an internal loading control. (G) Representative fluorescence images of EdU incorporation evaluating VSMC proliferation. Cells were transfected with either control siRNA (si-Control) or TLR2-targeted siRNA (si-TLR2) prior to vehicle or NET treatment. TLR2 knockdown successfully abrogates NET-induced EdU uptake (green). Nuclei are counterstained with DAPI (blue). Scale bar: 200 μm.

Because NET comprise complex extracellular ligands, we hypothesized that receptor-mediated signaling is required for intracellular transduction. We therefore performed a multi-layer intersection analysis integrating upregulated genes from NET-stimulated smooth muscle cells, upregulated genes from in vivo stented tissues in Fig. 1A, and gene ontology annotations for receptor complexes. This approach identified toll-like receptor 2 (TLR2) as the only candidate satisfying all criteria (Fig. 5C). Immunofluorescence staining of stented tissues confirmed robust localized expression of TLR2 within the vascular wall, supporting its in vivo relevance (Fig. 5D). In the in vitro model, the transcriptomic volcano plot showed significant induction of TLR2 along with downstream inflammatory genes including IL6, CXCL8, CCL20, CSF3, and ICAM1 (Fig. S4), consistent with activation of a pro-inflammatory synthetic state (Fig. 5E). Western blotting validated increased TLR2 protein expression and demonstrated activation of canonical downstream NF-κB signaling, as reflected by enhanced phosphorylation of key pathway components (Fig. 5F, Fig. S2B).

To test functional necessity, we suppressed TLR2 expression using targeted siRNA for 48 h before NET stimulation, which significantly attenuated NET-induced smooth muscle cell proliferation in EdU assays, establishing TLR2 as an essential mediator of the remodeling program initiated by NETs (Fig. 5G, Fig. S2A).

3.5. The selective TLR2 inhibitor C29 mitigates neutrophil extracellular NET-induced smooth muscle cell activation by uncoupling inflammatory transcription

To refine the regulatory hierarchy controlling the pro-inflammatory synthetic phenotype, we interrogated signaling downstream of TLR2. Consistent with earlier findings, siRNA-mediated reduction of TLR2 blunted NET-induced upregulation of intercellular adhesion molecule 1(ICAM1) and reversed the pathological decrease in the contractile marker SM22 (Fig. 6A and C).

Fig. 6.

The STAT3/TLR2/NF-κB signaling axis mediates NET-induced vascular smooth muscle cell phenotypic switching, proliferation, and neutrophil adhesion, which is reversed by pharmacological TLR2 inhibition. (A) Representative Western blot analysis of ICAM1, SM22, and TLR2 expression in VSMCs transfected with control siRNA (si-Control) or TLR2 siRNA (si-TLR2) followed by NET stimulation. β-Actin served as the loading control. (B) Representative Western blot evaluating the effect of the NF-κB inhibitor JSH-23 on NET-induced ICAM1 and TLR2 upregulation. β-Actin served as the loading control. (C, D) Densitometric quantification of the Western blots in (A) and (B), respectively. Data were analyzed using one-way ANOVA (*P < 0.05, ****P < 0.0001, n = 6 for each group). (E) Representative Western blot demonstrating that genetic knockdown of STAT3 (si-STAT3) abrogates NET-induced TLR2 upregulation in VSMCs. β-Actin served as the loading control. (F) Representative Western blot evaluating the therapeutic efficacy of the specific TLR2 inhibitor C29. Pre-treatment with C29 prevents NET-induced upregulation of inflammatory markers (ICAM1, TLR2) and rescues the expression of contractile markers (SM22, α-SMA). β-Actin served as the loading control. (G) Densitometric quantification of TLR2 levels from (E). (H) Densitometric quantification of protein levels from (F). Statistical significance for ICAM1 and TLR2 was determined using Welch ANOVA (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, n = 6 for each group). (I) Representative images of the cell adhesion assay. Neutrophils were labeled with DiI Red fluorescent dye and co-cultured with VSMCs under indicated treatments. Phase-contrast and fluorescence images were merged to visualize adhered neutrophils. Scale bar: 400 μm. (J) Quantitative analysis of DiI Red-labeled neutrophils adhered to VSMCs per field. Data were analyzed using Welch ANOVA (*P < 0.05, **P < 0.01, n = 6 for each group). (K) Representative fluorescence images of EdU incorporation evaluating VSMC proliferation in the presence of NET and/or the TLR2 inhibitor C29. Proliferating nuclei are marked by EdU (green) and total nuclei by DAPI (blue). Scale bar: 200 μm. (L) Quantification of the percentage of EdU-positive cells. Data were analyzed via one-way ANOVA (****P < 0.0001, n = 6 for each group). Data in all bar graphs are presented as mean ± SEM.

Because TLR2 commonly signals through NF-κB, we next tested pharmacologic blockade of p65 nuclear translocation with JSH-23 at 10 μM. Although this intervention restored ICAM1 protein expression toward baseline, it unexpectedly increased TLR2 protein levels, suggesting that NF-κB is not the principal transcriptional driver of TLR2 and that distal NF-κB inhibition may trigger compensatory feedback (Fig. 6B and D).

Guided by prior literature indicating that NET can activate STAT3 [33], we interrogated STAT3 as an upstream transcriptional regulator. Using JASPAR, we identified two high-scoring STAT3 binding motifs within the TLR2 promoter region, as detailed in Fig. S3A. Dual-luciferase reporter assays further demonstrated that STAT3 overexpression significantly enhanced WT TLR2 promoter activity, whereas simultaneous mutation of both predicted STAT3-binding motifs abolished this transcriptional activation, supporting the functional relevance of these STAT3-responsive elements (Fig. S3B). Western blotting confirmed robust activation of STAT3 phosphorylation at Ser727 and Tyr705 in response to NET stimulation (Fig. S3C–D). Importantly, STAT3 knockdown by siRNA markedly suppressed TLR2 expression (Fig. 6E and G), establishing STAT3 as a necessary transcriptional mediator of NET-induced TLR2 upregulation. Because STAT3 is challenging to inhibit without off-target toxicity and our preliminary experiments suggested that direct STAT3 inhibitors can induce apoptosis in VSMCs (data not shown), we instead pursued pharmacologic blockade of TLR2 with C29, a lipophilic specific TLR2 antagonist. CCK-8 assays showed that C29 did not affect VSMC viability at low to moderate concentrations, whereas reduced viability was observed only at higher concentrations, particularly 200–500 μM, supporting the use of a non-cytotoxic C29 concentration for subsequent functional experiments (Fig. S2C).

Treatment with C29 for 24 h suppressed induction of ICAM1, inhibited TLR2 expression, and restored contractile markers in stimulated cells (Fig. 6F and H). Quantitative PCR further demonstrated broad reduction of inflammatory mediators, as shown in Fig. S4. Functionally, C29 reduced pathological adhesion between VSMCs and infiltrating neutrophils in a DiI-labeled neutrophil adhesion assay (Fig. 6I and J) and markedly decreased NET-induced smooth muscle cell proliferation in EdU assays (Fig. 6K and L), indicating that C29 effectively interrupts the feed-forward inflammatory circuit linking NET to vascular remodeling.

3.6. A C29-coated bioresorbable scaffold suppresses neutrophil accumulation and attenuates neointimal hyperplasia in vivo

To obtain initial in vivo proof-of-concept evidence while avoiding confounding anti-inflammatory effects of rapamycin, we engineered an PLA bioresorbable scaffold coated exclusively with C29. Because C29 is highly lipophilic, we used acetone-based dissolution followed by ultrasonic spray coating to achieve uniform drug deposition on the scaffold surface. Gross inspection showed a characteristic yellow hue consistent with C29 coating (Fig. 7B), and scanning electron microscopy performed after balloon expansion revealed a rough but largely non-porous surface, without obvious coating cracking, peeling, or delamination, suggesting that the coating retained gross mechanical integrity during scaffold deployment (Fig. 7C). Given C29 lipophilicity, partitioning into PLA-rich domains is expected to reduce burst release and favor elution during later stages of matrix degradation. Consistent with this design rationale, in vitro release profiling demonstrated a gradual and sustained C29 elution pattern, with limited early burst release followed by progressive drug release over 30 days, reaching approximately 82% cumulative release by the end of the observation period (Fig. S5A).

Fig. 7.

Local TLR2 inhibition via C29-coated scaffolds attenuates in-scaffold restenosis and reduces peristrut neutrophil accumulation in vivo. (A) Schematic representation of the in vivo rabbit iliac artery stenting model, utilizing a right carotid artery entry point for scaffold deployment. (B) Macroscopic photograph of the bioresorbable scaffold coated with the specific TLR2 inhibitor C29. The coating was prepared via an acetone dissolution method and applied without the standard anti-proliferative drug rapamycin to exclusively isolate the therapeutic effects of C29. (C) Representative Scanning Electron Microscopy (SEM) images detailing the surface morphology of the bare scaffold and the C29-coated scaffold after balloon expansion. Scale bars: 400 μm (Overview) and 100 μm (Magnified). (D) Representative Hematoxylin and Eosin (H&E) stained cross-sections of stented arteries evaluating the extent of neointimal hyperplasia. Scale bar: 400 μm. (E) Quantification of the relative stenosis percentage derived from H&E histomorphometry. Data were analyzed via paired t-test (P < 0.05, n = 6 for each group). (F) Representative Optical Coherence Tomography (OCT) images acquired in vivo to assess lumen patency and scaffold strut integration. Scale bar: 1000 μm. (G) Quantification of lumen area stenosis percentage derived from the OCT volumetric imaging. Data were analyzed via paired t-test (P < 0.05, n = 5 for each group). (H) Representative Western blot analysis of STAT3 and NF-κB signaling pathway activation in arterial tissues from the bare scaffold and C29 scaffold groups. Protein expression of phosphorylated STAT3 (p-STAT3), total STAT3, phosphorylated P65 (p-P65), total P65, and β-Actin was assessed. (I) Densitometric quantification of relative p-STAT3 and p-P65 protein levels. Data were analyzed via paired t-test (P < 0.01, n = 5 for each group). (J) Quantitative RT-PCR analysis of inflammatory gene expression, including IL6, ICAM1, and TLR2 mRNA levels, in arterial tissues from the bare scaffold and C29 scaffold groups. Data were analyzed via paired t-test (P < 0.05, n = 5 for each group). (K) Representative in situ immunofluorescence images evaluating local immune cell recruitment following scaffold implantation. Sections are co-stained for neutrophils (red), the vascular smooth muscle cell marker α-SMA (green), and nuclei (DAPI, blue). Scale bars: 500 μm (Overview) and 100 μm (Magnified). (L) Quantitative analysis of the neutrophil-positive area percentage within the stented vessel segments. Data were analyzed via paired t-test (P < 0.05, n = 5 for each group). In one rabbit, the iliac artery on the bare-scaffold side was completely occluded; therefore, bilateral OCT imaging could not be performed. Accordingly, this animal was excluded from both the bilateral OCT analysis and the immunofluorescence-based neutrophil analysis. Data in all bar graphs are presented as mean ± SEM.

We then implemented a randomized bilateral paired design in six rabbits, in which each animal received one bare PLA scaffold and one C29-coated PLA scaffold in contralateral iliac arteries, thereby enabling an internally controlled comparison and reducing inter-animal variability (Fig. 7A). At six months after implantation, a time point selected based on the emergence of MPO-positive inflammation in the longitudinal porcine model, we performed intraluminal imaging and harvested tissues to quantify vascular remodeling. Histological analysis demonstrated that C29-coated scaffolds significantly reduced neointimal hyperplasia and decreased relative stenosis compared with bare scaffolds, as shown in Fig. 7D and E. In vivo optical coherence tomography performed before harvest corroborated these findings by revealing larger lumen area and lower stenosis in C29-treated vessels (Fig. 7F and G). Consistent with these structural improvements, Western blot analysis of stented arterial tissues showed that C29 coating attenuated inflammatory signaling, as evidenced by down-regulated STAT3 expression and decreased p-p65 levels compared with bare scaffolds (Fig. 7H and I) while the p-STAT3 levels and the p65 expression remains no significant changes (Fig. S5B–C). qPCR analysis further confirmed that C29-coated scaffolds markedly suppressed the expression of inflammatory and TLR2-related transcripts, including IL6, ICAM1, and TLR2, relative to bare scaffolds (Fig. 7J).

Finally, to test whether therapeutic benefit aligned with the proposed immunologic mechanism, we quantified neutrophil accumulation within the neointima by immunofluorescence staining. C29 coating markedly reduced local neutrophil infiltration relative to the robust inflammatory accumulation observed in bare scaffolds, as demonstrated and quantified in Fig. 7K and L.

Together, these in vivo results demonstrate that selective TLR2 inhibition via C29 coating disrupts neutrophil-driven inflammation, reduces neointimal hyperplasia, and preserves vascular patency in a rabbit iliac model during a degradation-associated inflammatory phase.

4. Discussion

The present study identifies NET formation as a major degradation-associated immune program associated with PLA BRS degradation and links this response to vascular smooth muscle cell reprogramming and neointimal hyperplasia. By combining bulk transcriptomics from porcine coronary arteries, spatial histology, and mechanistic cell experiments, we provide convergent evidence that the immune microenvironment surrounding degrading polymeric struts is distinct from that observed with permanent metallic drug-eluting stents. The transcriptomic analyses at nine months demonstrated broad activation of inflammatory pathways with particularly strong enrichment of NET-related signatures, and tissue staining confirmed prominent myeloperoxidase-positive accumulation around struts. Together, these findings support the interpretation that innate immunity remains active long after implantation and may even intensify as degradation proceeds.

A key implication of these data is that the timing of inflammatory risk with BRS may differ fundamentally from that of permanent metallic devices. Rather than an early transient leukocyte response that resolves with endothelial recovery, our longitudinal staining suggests a delayed increase in neutrophil infiltration that becomes most apparent during advanced degradation stages. The observation that degraded strut morphology correlates with increased colocalization of myeloperoxidase and citrullinated histone H3 further argues that the physical and chemical state of the polymer is not merely permissive, but associated with a local niche permissive to, and potentially promoting, NET formation. This temporal coupling aligns with the concept that degradation products, altered mechanical microgeometry, and persistent foreign-body signaling can create an immunologically active niche that is amplified rather than extinguished during late remodeling.

Neutrophils are often regarded as predominant effectors of the acute phase of inflammation; however, Binet and colleagues reported that NET may also participate during chronic stages of disease progression [34]. Consistent with this expanded view, our data indicate that neutrophils contribute to long-term vascular remodeling after scaffold implantation. Importantly, the vascular inflammation considered in the present work represents a sustained, chronic process driven by the continuous generation of PLA degradation products. As such, this remodeling-associated inflammatory trajectory is not directly comparable to conventional acute-chronic “phase” frameworks described in other settings, and the temporal staging of inflammation cannot be assumed to be equivalent across contexts.

Mechanistically, our in vitro findings support a causal link between PLA degradation products and NET induction and point to mitochondrial dysfunction as a proximal intracellular event. Loss of mitochondrial membrane potential after exposure to degradation products, together with increased expression of the mitochondrial calcium uniporter pore component and the ability of MCU inhibition to suppress PADI4 induction and histone citrullination, suggests that calcium overload driven by mitochondrial calcium uniporter activity may be a critical trigger that lowers the threshold for NET release. This mechanism is consistent with established biology in which excessive mitochondrial calcium promotes metabolic stress [35] and reactive signaling that facilitates chromatin decondensation [36] and extracellular trap extrusion [37]. In the context of implanted scaffolds, these data imply that local accumulation of polymer breakdown products can directly prime neutrophils, providing a cellular explanation for why NET formation becomes more prominent as degradation advances.

Beyond identifying NET formation as a degradation-associated inflammatory hallmark [38], our study provides a functional bridge between NET and the VSMC behaviors that drive restenosis. Exposure of human VSMCs to purified NET increased proliferation and migration and promoted a phenotypic switch away from contractile markers toward a synthetic state. The transcriptomic response of VSMCs revealed enrichment of cell-cycle programs alongside inflammatory pathways, indicating that NET act not only as cytotoxic or antimicrobial structures but also as potent paracrine signaling assemblies capable of reshaping vascular wall biology. The emergence of chemotaxis-related signatures in NET-stimulated smooth muscle cells suggests a feed-forward loop in which NETs amplify leukocyte recruitment cues, potentially sustaining a chronic inflammatory circuit within the neointima and adventitia.

NETs consist of extracellular DNA, histones, and neutrophil granule proteins, and different NET components may contribute to inflammatory signaling depending on the disease context. Among these components, extracellular histones, particularly H3 and H4, have been reported as plausible NET-associated DAMPs involved in TLR2-dependent activation[[39], [40], [41]]. HMGB1 may also participate in NET-related TLR2 signaling under certain inflammatory conditions [42]. In the present study, our findings demonstrate that NET accumulation is associated with activation of the TLR2-related downstream inflammatory pathway in the scaffold microenvironment. However, the specific NET-derived component responsible for TLR2 activation was not directly identified, and this remains an important direction for future investigation.

Our receptor-level analyses nominate TLR2 as a central transducer of NET-driven smooth muscle cell activation, with downstream engagement of canonical inflammatory signaling. The observation that TLR2 was the sole receptor candidate at the intersection of in vivo upregulated genes, in vitro VSMCs induction, and receptor complex annotations strengthens its plausibility as a mechanistic hub rather than a coincidental marker [43]. Inhibition experiments further support functional necessity, since TLR2 knockdown attenuated NET-induced proliferative responses and reduced induction of inflammatory markers. These data suggest STAT3 drives TLR2 upregulation, which may explain why NF-κB blockade shows compensatory effects, whereas direct TLR2 inhibition more consistently suppresses the response. The paradoxical increase in TLR2 expression after JSH-23 treatment also indicates that distal NF-κB blockade may induce compensatory receptor-level feedback rather than fully suppressing the upstream NET-sensing program. The molecular basis of this feedback remains unresolved, and future studies should examine whether NF-κB-dependent negative regulators, STAT3-NF-κB crosstalk, altered TLR2 protein turnover, or additional transcriptional and post-transcriptional mechanisms contribute to sustained TLR2 expression during NET stimulation.

The therapeutic arm of our study provides proof of concept that selective modulation of this innate immune axis can mitigate restenosis without relying on rapamycin. A C29-coated BRS reduced neointimal hyperplasia and decreased local neutrophil accumulation in vivo, supporting the idea that interrupting TLR2-driven signaling can attenuate this inflammatory circuit that connects degradation-induced NET formation to smooth muscle cell activation. This strategy targets a specific inflammatory pathway instead of broadly inhibiting proliferation, which may better match the advanced inflammation that coincides with polymer breakdown [44]. The in vitro release profile supports sustained C29 elution over the tested period, and serum LC-MS analysis showed that circulating C29 was undetectable at the 6-month endpoint. Nevertheless, the current surface-coated design does not fully define tissue-level pharmacokinetics throughout the entire scaffold degradation process, and transient systemic exposure after implantation cannot be fully excluded because serial pharmacokinetic measurements at earlier time points were not performed. Future scaffold designs may incorporate C29 into the PLA scaffold matrix or degradation-responsive coating domains to improve synchronization between drug availability and degradation-associated inflammatory activation. Because systemic inhibition of innate immune sensing may be undesirable, local scaffold-based delivery may offer a practical route to enrich drug activity at the peri-strut interface while limiting systemic immune modulation. However, because C29 is not currently an approved clinical drug, future translational development will require GLP toxicology, serial systemic pharmacokinetics, local tissue-retention analysis, infection-risk assessment, and coating-stability testing during scaffold expansion and degradation [45].

Several limitations should be considered when interpreting these results. Bulk RNA sequencing captures average signals across heterogeneous vascular compartments and cannot resolve the precise cellular sources of specific transcripts; although histology supports prominent neutrophil/NET-associated signals [46], single-cell or spatial transcriptomic approaches would sharpen cell-type attribution and clarify communication networks among neutrophils, macrophages, lymphocytes, and vascular cells [47]. Transcriptomic profiling was performed only at the 9-month time point in the porcine coronary model. Although this time point enabled identification of inflammatory programs during an established degradation-associated remodeling phase, it does not provide a full dynamic characterization of the NET-TLR2 axis across the entire degradation process. Our longitudinal histological analyses provide complementary temporal evidence of delayed neutrophil infiltration and NET-associated staining, but they cannot replace molecular profiling at earlier and later stages. Future studies incorporating bulk RNA-seq, single-cell RNA-seq, or spatial transcriptomics at multiple time points such as 3, 6, 9, and 12 months will be necessary to define the temporal sequence of neutrophil recruitment, NET formation, TLR2 activation, and interactions with other immune-cell populations during progressive PLA scaffold degradation.

Regarding the in vitro studies, although accelerated hydrolysis allowed us to generate a standardized PLA degradation-product mixture for mechanistic testing, this preparation does not fully recapitulate the gradual, heterogeneous, and locally buffered degradation process that occurs in vivo [48]. Future studies should characterize the specific molecular-weight distribution, acidity, lactic acid and oligomer composition, and tissue-level accumulation of degradation products recovered from implanted scaffolds. Although our data support MCU-mediated mitochondrial calcium dysregulation as a downstream mechanism linking PLA degradation products to NET formation, the upstream sensing mechanism remains to be defined [49]. PLA degradation products may alter neutrophil metabolism through acidification, lactate/oligomer transport, proton-sensitive ion channels, or mitochondrial metabolic stress. Future studies using purified PLA degradation components, pH-matched controls, transporter inhibition, and upstream calcium-channel screening will be required to define the proximal sensor that connects PLA degradation products to MCU activation and NET release [50,51]. In addition, the 500 ng/mL NET concentration used to treat VSMCs was selected to establish a reproducible in vitro vascular inflammation model [28,52], future dose-response experiments and direct measurement of local NET burden in peri-strut tissue would further improve the physiological calibration of this model.

Several aspects of the in vivo therapeutic mechanism also require cautious interpretation. Direct in vivo quantification of NET suppression by the C29-coated scaffold in rabbit explants remains technically challenging. Although C29 coating reduced peri-strut neutrophil accumulation and neointimal hyperplasia, these findings should be interpreted as being consistent with, rather than definitive proof of, disruption of the NET-TLR2 paracrine loop in vivo. Species-compatible NET markers, orthogonal assays such as MPO-DNA complex detection, and in vivo genetic or pharmacological validation of TLR2 signaling will be necessary to directly confirm whether local C29 delivery suppresses NET-driven TLR2 activation within the scaffold microenvironment. Moreover, although the present in vitro experiments support TLR2 as a major mediator of NET-induced VSMC activation, and CCK-8 assay did not support a direct inhibitory effect of C29 on basal VSMC viability/proliferation under the tested conditions, the current in vivo data cannot fully exclude TLR2-independent contributions to the protective effect of C29. Potential alternative mechanisms may include direct modulation of VSMC behavior, effects on macrophages or other immune-cell populations, or broader changes in the peri-scaffold inflammatory microenvironment. Therefore, the in vivo efficacy of C29-coated scaffolds should be interpreted as being consistent with attenuation of the NET-associated TLR2 inflammatory program rather than definitive proof of exclusive pathway specificity.

Another important limitation is that the present study focuses primarily on the NET-TLR2 axis, whereas the immune microenvironment generated during PLA scaffold degradation is likely to involve multiple interacting leukocyte populations. Transcriptomic and histological analyses indicated broader immune activation, including monocyte/macrophage and lymphocyte-associated signals, and these cells may cooperate with neutrophils or respond secondarily to NET-rich inflammation. Therefore, our findings should not be interpreted as excluding the contribution of macrophages, T/B cells, dendritic cells, complement activation, or other innate immune pathways. Instead, the NET-TLR2 axis represents one functionally validated mechanism within a broader degradation-associated immune network. Future single-cell, spatial transcriptomic, and cell-specific intervention studies will be needed to define how neutrophils/NETs interact with other immune-cell populations and to determine their relative contributions to scaffold-associated vascular remodeling.

The porcine coronary model [53] and rabbit iliac model [53] each capture important aspects of vascular injury and remodeling, yet neither perfectly replicates human coronary atherosclerosis and its immune landscape [54]; additional validation in diseased vessels would strengthen clinical generalizability. In addition, the rabbit iliac validation study was performed in a modest-sized bilateral paired design intended for mechanistic in vivo proof-of-concept rather than definitive translational efficacy assessment. Although this internally controlled design reduces inter-animal variability and limits animal use, larger studies with independent treatment groups and longer follow-up will be necessary to confirm robustness, quantify effect size more precisely, and strengthen translational generalizability. The rabbit iliac model also evaluated a 6-month endpoint and therefore cannot determine whether C29 coating prevents later rebound inflammation after further scaffold resorption. Moreover, endothelial recovery was not directly quantified in the present study by endothelial marker staining or en face assessment; therefore, whether C29 coating affects re-endothelialization remains unresolved and should be evaluated in future studies.

Despite these limitations, we propose a model in which advanced polymer degradation enriches neutrophils, increases NET formation via mitochondrial calcium dysregulation, and promotes VSMC proliferation/migration through a STAT3-TLR2 pathway. This framework expands the understanding of how bioresorbable scaffold degradation can influence long-term vascular remodeling and points to selective innate immune targeting as a rational strategy to improve late outcomes[[55], [56], [57]]. Future studies that combine high-resolution spatial profiling with longitudinal clinical sampling and optimized local delivery of TLR2 antagonists will be valuable for determining whether controlling NET-driven signaling can safely preserve the intended benefits of scaffold resorption while minimizing advanced stage inflammatory remodeling.

5. Conclusion

Degradation of PLA bioresorbable scaffolds is not a neutral end stage of implantation. As the polymer undergoes chemical and structural breakdown, the peri-strut immune niche shifts toward sustained neutrophil activation and persistent NET release. We link this response to the buildup of PLA degradation products, which perturbs mitochondrial function and promotes calcium overload through the mitochondrial calcium uniporter, thereby lowering the threshold for NET extrusion. In this environment, NETs function as paracrine signals that promote a synthetic, proliferative phenotype in adjacent vascular smooth muscle cells via TLR2 signaling. Importantly, interrupting this loop with a local TLR2 inhibitor coating mitigated neointimal hyperplasia in vivo. Together, these findings support a materials-guided view of late outcomes after bioresorbable implants, in which polymer dissolution actively shapes the immune-vascular interface. Designing scaffolds that modulate this specific innate immune axis may represent a promising strategy for improving long-term patency while preserving the benefits of resorption, although further validation in larger and longer-term preclinical studies will be required.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Zhongshan Hospital. All animal experiments were approved by Gateway Medical Innovation Center Institutional Animal Care and Use Committee (approval no. SH2022-06009, SH2025-07017).

CRediT authorship contribution statement

Changyi Zhou: Conceptualization, Data curation, Formal analysis, Software, Visualization, Writing – original draft. Rui Wang: Funding acquisition, Investigation, Methodology, Project administration, Validation, Writing – review & editing. Jiasheng Yin: Conceptualization, Funding acquisition, Methodology, Validation, Writing – review & editing. Yue Yu: Formal analysis, Software. Wenhui Lin: Conceptualization, Methodology. Yang Lyu: Formal analysis, Visualization. Chengpeng Liu: Software. Qiang Xu: Conceptualization. Guanzhao Zhang: Methodology. Han Chen: Conceptualization. Tongtong Zang: Conceptualization. Yizhe Wu: Funding acquisition, Supervision. Li Shen: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing. Junbo Ge: Funding acquisition, Project administration, Supervision.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

The mRNA-sequencing data generated from the pig stented coronary samples and the in vitro cell experiments have been deposited in the China National Center for Bioinformation and are publicly accessible under accession numbers OMIX016606 and OMIX016607.