Targeting NLRP3 inflammasome: a novel strategy for improving immune checkpoint inhibitor-associated pneumonitis

Abstract

Background

Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment but are frequently accompanied by immune-related adverse events (irAEs) affecting multi-organ. Checkpoint inhibitor-associated pneumonitis (CIP) is a serious irAE whose mechanistic underpinnings remain poorly understood, limiting the clinical application of ICIs. The objective of this study was to investigate the role of the NOD-like receptor family pyrin domain-containing protein 3 (NLRP3) inflammasome in the pathogenesis of CIP and to explore potential targeted therapies.

Methods

A CIP mouse model was established via regulatory T cell (Treg) depletion and anti-programmed cell death protein 1 (PD-1) antibody treatment. Lung tissue gene expression was analyzed using RNA-sequencing, quantitative polymerase chain reaction (qPCR), and Western blot. Therapeutic interventions with NLRP3 inflammasome inhibitors (MCC950 and dapansutrile) and a macrophage-depleting agent (clodronate liposomes) were evaluated by micro-computed tomography (CT), histopathology, and serum inflammatory cytokine assays. Validation was performed via single-cell transcriptomic analysis of bronchoalveolar lavage fluid from 13 non-small cell lung cancer patients with or without ICI-induced CIP.

Results

The findings revealed: (I) significant upregulation of NLRP3 inflammasome pathway-related genes and downstream factors (IL-1β, IL-18) in CIP mouse lungs, alongside increased macrophage infiltration; (II) pneumonitis injury was markedly alleviated and serum inflammatory cytokine levels were reduced by both NLRP3 inflammasome inhibition and macrophage depletion; (III) dapansutrile downregulated NLRP3 expression via inhibition of the NF-κB pathway; (IV) enriched macrophage subpopulations (Mac-IL1B) expressing high levels of NLRP3 were identified in CIP patients.

Conclusions

This study provides the first evidence that NLRP3 inflammasome activation constitutes a key upstream mechanism in CIP pathogenesis, offering novel therapeutic strategies for targeted intervention.

Highlight box.

Key findings

• This study identifies NOD-like receptor family pyrin domain-containing protein 3 (NLRP3) inflammasome activation in pulmonary macrophages as a key driver of checkpoint inhibitor-associated pneumonitis (CIP). In a CIP mouse model (regulatory T cell-depleted + anti-programmed death receptor-1), NLRP3 pathway genes and cytokines were significantly upregulated. Both NLRP3 inhibitors (MCC950, dapansutrile) and macrophage depletion (clodronate liposomes) reduced lung inflammation, serum cytokines, and tissue damage. Dapansutrile uniquely inhibited NLRP3 priming via NF-κB suppression and inflammasome assembly. Single-cell RNA-sequencing of human CIP patients revealed enriched NLRP3-high macrophage subsets (Mac-IL1B) in bronchoalveolar lavage fluid.

What is known and what is new?

• CIP is a life-threatening immune-related adverse event of ICIs with limited treatment options.

• While NLRP3 is implicated in other inflammatory diseases, this study first demonstrates its upstream role in CIP pathogenesis via macrophage pyroptosis. It also discovers the therapeutic efficacy of dapansutrile’s dual inhibition mechanism and identifies Mac-IL1B as a novel cellular mediator in human CIP.

What is the implication, and what should change now?

• Targeting NLRP3 offers a precision strategy to mitigate CIP without broad immunosuppression. Dapansutrile, with its oral bioavailability and favorable safety profile (unlike MCC950-induced hepatotoxicity), presents a clinically translatable option to preserve anti-tumor immunity while managing pneumonitis.

• Clinical trials should evaluate NLRP3 inhibitors (especially dapansutrile) for CIP prophylaxis/therapy in ICI-treated patients.

• Diagnostic approaches could incorporate Mac-IL1B or NLRP3 pathway markers in BALF as potential CIP biomarkers.

• Treatment guidelines may evolve beyond corticosteroids to include targeted NLRP3 blockade, improving ICI safety and continuity.

Introduction

Over the past few decades, immune checkpoint inhibitors (ICIs) have emerged as a groundbreaking advancement in tumor immunotherapy, revolutionizing the clinical landscape of cancer treatment (1). Currently, clinically deployed ICIs primarily target three key immune checkpoint molecules: (I) programmed death receptor-1 (PD-1), with agents including nivolumab and pembrolizumab; (II) programmed death-ligand 1 (PD-L1), represented by avelumab, atezolizumab, and durvalumab; and (III) cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4), with inhibitors such as tremelimumab and ipilimumab (2). However, while these agents activate T cells by blocking immune checkpoint molecules, immune-related adverse events (irAEs) may be induced, affecting multiple organ systems (3,4). The clinical manifestations of irAEs are highly heterogeneous and can involve multiple organs and tissues, including the digestive system (colitis, hepatitis) (5,6), respiratory system (pneumonitis) (7), endocrine system (thyroiditis, type 2 diabetes) (8,9), and skin (dermatitis) (10,11).

Checkpoint inhibitor-associated pneumonitis (CIP) is one of the most frequent adverse events associated with ICI therapy, with its risk profile established in multiple cohort studies (12,13). CIP is characterized by a spectrum of respiratory symptoms, including cough and dyspnea, alongside radiographic pulmonary parenchymal abnormalities. In severe cases, progression to acute respiratory failure or death may occur (14). Notably, CIP accounts for approximately 35% of immunotherapy-related fatalities in patients receiving PD-1/PD-L1 inhibitors (15). Current management strategies vary according to disease severity, though limited evidence supports the efficacy of subsequent therapies following CIP diagnosis. Given that CIP restricts the application of tumor immunotherapy, elucidating its pathogenesis is critically important.

In recent years, the in-depth analysis of patient-derived biological samples has significantly advanced our understanding of the pathogenesis of both irAEs and CIP. For instance, key cytokines, including interleukin (IL)-6, have been identified as central drivers of pathogenesis and have subsequently been validated as clinical targets. Notably, anti-IL-6 therapies, such as tocilizumab, have demonstrated significant efficacy in the treatment of severe or refractory cases of CIP (16,17). However, clinical translation remains challenging due to several factors: not all patients exhibit a response to current therapeutic regimens, certain individuals present with contraindications or experience adverse reactions, and the heterogeneity in underlying disease mechanisms suggests the involvement of critical pathogenic pathways beyond those mediated by IL-6. Therefore, the exploration of novel therapeutic targets and candidate agents, particularly those directed against distinct pathological mechanisms or intended for use in combination regimens, continues to be of significant clinical and scientific importance.

Macrophages polarize into a proinflammatory phenotype and activate the NOD-like receptor family pyrin domain-containing protein 3 (NLRP3) inflammasome, thereby triggering inflammatory responses (18). Structurally, the NLRP3 inflammasome comprises three core components: (I) the receptor protein NLRP3, which contains an N-terminal pyrin domain (PYD) responsible for homodimerization, a central NACHT domain exhibiting ATPase activity, and a C-terminal leucine-rich repeat (LRR) domain involved in ligand recognition and autoinhibition; (II) the adaptor protein ASC, which bridges NLRP3 to downstream effectors through its PYD and caspase activation and recruitment domain (CARD); and (III) the effector protease pro-caspase-1. Upon activation, pro-caspase-1 cleaves gasdermin D (GSDMD), a substrate of inflammatory caspases, the N-terminal fragment of which oligomerizes to form membrane pores that mediate pyroptosis—inducing this lytic, proinflammatory form of programmed cell death while concurrently promoting the maturation and secretion of proinflammatory cytokines such as IL-1β and IL-18 (19-21). Emerging evidence demonstrates that the NLRP3 inflammasome serves as a central regulator in ICI-induced myocarditis (22). Given its established role in pulmonary inflammatory diseases and involvement in ICI-related toxicity, we speculate that NLRP3 inflammasome activation may similarly contribute to the pathogenesis of ICI-related CIP.

This study was conducted to preliminarily elucidate the key role of the NLRP3 inflammasome in the pathogenesis of CIP, and to explore a novel intervention strategy. It was found that following treatment with anti-PD-1 antibodies, the NLRP3 inflammasome in lung macrophages, particularly within the Mac-IL1B subset, was activated, subsequently triggering a cascade of inflammatory responses. In therapeutic investigations, the NLRP3 inflammasome inhibitor dapansutrile demonstrated promising potential. Building on the previous analysis, this drug was not only found to effectively inhibit the progression of pneumonitis, but also demonstrated a reduced impact on the liver compared to MCC950. Notably, dapansutrile was further observed to exhibit a ‘dual inhibition’ effect, simultaneously interfering with both NLRP3 inflammasome assembly and the NF-κB signaling pathway. Although the present findings are derived primarily from animal experiments, analysis of clinical patient-derived single-cell sequencing data lends supporting evidence. This body of work suggests that targeting the ‘NLRP3/IL-1β axis’ represents a novel upstream oral therapeutic strategy for CIP, one which holds potential for combination with existing IL-6 blockade therapies. We present this article in accordance with the ARRIVE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1240/rc).

Methods

Chemicals and reagents

Diphtheria toxin (DT; Sigma-Aldrich, St. Louis, USA), anti-PD-1 antibody (clone RMP1-14), recombinant mouse interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α), IL-6, and IL-1β enzyme-linked immunosorbent assay (ELISA) kits were obtained from Thermo Fisher Scientific (Waltham, USA). The NLRP3 inflammasome inhibitors MCC950 and dapansutrile were acquired from MedChemExpress LLC (Shanghai, China). Clodronate liposomes were provided by Liposome BV (Amsterdam, Netherlands). Mouse serum amyloid A (SAA) ELISA kits, along with mouse IL-18 ELISA kits, were procured from Abcam (Cambridge, UK).

Animal treatments

A murine model mimicking the clinical and pathological manifestations of irAEs was successfully established by combining regulatory T cell (Treg) cell depletion—which compromises self-tolerance—with ICI therapy, adapting a previously reported approach. In Foxp3-diphtheria toxin receptor (DTR) mice, transient Treg depletion was induced via DT administration followed by anti-PD-1 monoclonal antibody treatment, resulting in severe irAEs (23,24).

Specific pathogen-free (SPF) inbred C57BL/6 Foxp3-DTR-GFP mice were obtained from Nanmo Bio (Shanghai, China) and maintained at the Chinese PLA General Hospital Animal Care Center under SPF conditions with a 12-hour light/dark cycle and ad libitum access to food and water. All animals were acclimated to the environment for at least 7 days prior to the start of the experiment. A veterinarian regularly monitored the animals’ health to ensure the absence of infectious diseases or abnormal signs. The animals were confirmed to be in good health before the experiment began. Mice were stratified into three experimental cohorts randomly: control, DT, and anti-PD-1. All Foxp3-DTR mice received subcutaneous implantation of 8×10⁵ Lewis lung carcinoma (LLC) cells on day 0. Tumor dimensions were monitored using digital calipers, with treatment initiation occurring when mean tumor area reached 40–50 mm². Humane endpoints were applied when tumors exceeded 150 mm². For Treg depletion, 300 ng/mL DT was administered intraperitoneally every 72 hours for three total doses. The anti-PD-1 cohort subsequently received 250 µg/mL anti-PD-1 antibody intraperitoneally every 72 hours for five doses. Pulmonary inflammatory lesions were evaluated by micro-computed tomography (micro-CT) and quantitative histopathological analysis.

To evaluate whether NLRP3 inflammasome inhibition ameliorates pneumonitis induced by Treg depletion combined with anti-PD-1 therapy, two experimental cohorts received concurrent interventions: one group was administered 25 mg/kg body weight (bw) MCC950 intraperitoneally, while another received 120 mg/kg bw dapansutrile via daily oral gavage alongside anti-PD-1 treatment. A third cohort received 10 mL/kg bw liposomal clodronate via tail vein injection during the first, third, and fifth anti-PD-1 treatment sessions to assess macrophage involvement in this pneumonitis model (25). The MCC950 dosage was selected based on established efficacy in prior studies (26,27), whereas the dapansutrile dosage was determined through pharmacokinetic profiling and published protocols (28,29). Following cervical dislocation euthanasia, lung tissue and blood specimens were collected for subsequent analysis.

We used individual animals as experimental units, with each experimental group comprising 5 units (n=5). Following the 3Rs principle (reducing, reusing, and recycling), the minimum sample size meeting statistical requirements was employed while ensuring detection of anticipated significant biological effects. Total number of animals = (number of baseline model groups × 5) + (number of intervention study groups × 5) = (3 × 5) + (3 × 5) =30 animals.

All experiments utilized 8- to 12-week-old mice and were performed under a project license (approval No. 2023-X19-105) granted by the Animal Welfare and Ethics Committee of the Chinese PLA General Hospital, in compliance with relevant Chinese regulations and institutional guidelines for the care and use of animals, consistently following the 3Rs principle to minimize animal suffering and discomforts. A protocol was prepared before the study without registration.

In vivo micro-CT system

In accordance with the approved protocol, in vivo pulmonary micro-CT was performed using the PerkinElmer Quantum GX imaging system (PerkinElmer Inc., Waltham, USA). Mice were maintained under anesthesia with 2% isoflurane, while physiological parameters—including respiratory rate (80–120 breaths per minute) and core body temperature (37.0±0.5 °C)—were continuously monitored. Respiratory-gated image acquisition was conducted during the expiratory phase at 90 kVp, 88 µA, and 72 µm isotropic resolution, with the cumulative radiation dose constrained to <200 mGy.

Measurement of the IFN-γ, TNF-α, IL-6, IL-1β, IL-18 and SAA levels in mouse serum

Serum samples from mice were prepared by centrifugation of whole blood at 3,000 ×g for 10 minutes. Concentrations of IFN-γ, TNF-α, IL-6, IL-1β, IL-18, and SAA in murine serum were quantified using ELISA kits (Thermo Fisher Scientific; Abcam) with absorbance measurements performed on Multiskan microplate readers (Thermo Fisher Scientific).

Total RNA extraction and reverse transcription quantitative polymerase chain reaction (RT-qPCR) analyses

Total RNA was isolated from mouse lung tissue utilizing TRIzol reagent (Thermo Fisher Scientific). RNA concentration and purity were assessed with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), and only samples exhibiting an A260/A280 ratio between 1.9 and 2.1 were retained for downstream analyses. First-strand complementary DNA (cDNA) synthesis was subsequently performed using the RevertAid^TM First Strand cDNA Synthesis Kit (Thermo Fisher Scientific; Cat# K1622), with 1 µg of total RNA in a 20-µL reaction volume. RT-qPCR was conducted using Hieff UNICON^TM qPCR SYBR Green Master Mix (YeastOne, Thermo Fisher Scientific; Cat# 11198), with gene expression normalized to 18S ribosomal RNA. Primer sequences are provided in Table S1.

RNA-sequencing (RNA-seq) and data analysis

Lung tissue samples (10 mg per mouse) were harvested, flash-frozen at −80 °C, and homogenized in TRIzol^TM reagent (Thermo Fisher Scientific; Tokyo, Japan). Samples were pooled into three experimental groups (control, DT, DT + PD-1) and submitted to Novogene (Nagoya, Japan) for total RNA extraction. RNA purity and integrity were verified by Novogene prior to library preparation. RNA-seq was performed on the Illumina platform. Raw count data were analyzed using the edgeR package (v3.4) in R to identify differentially expressed genes (DEGs), with significance thresholds set at a false discovery rate (FDR)-adjusted P<0.05 and absolute log₂-fold change >1. Functional annotation of aggregated DEGs was conducted using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases to investigate potential therapeutic mechanisms of intermittent exercise. Enrichment analysis of GO terms and KEGG pathways was performed via the clusterProfiler R package, with FDR-adjusted P<0.05 defining statistical significance. Gene Set Enrichment Analysis (GSEA) was executed using the Java-based GSEA platform, with gene sets derived from the Molecular Signatures Database (MSigDB).

Hematoxylin and eosin (H&E) and immunofluorescence (IF)

Lung tissue specimens were fixed, sectioned at 5 µm thickness, and processed for H&E staining and IF analysis. Sections were dewaxed in xylene followed by a graded ethanol series, then stained with hematoxylin and eosin according to standard protocols. Antigen retrieval was performed via microwave-mediated heating in 10 mM sodium citrate buffer (pH 6.0). After incubation with primary antibodies (Table S2), IF staining was conducted using Alexa Fluor^TM 594-conjugated goat anti-rabbit IgG or Alexa Fluor^TM 633-conjugated goat anti-rat IgG secondary antibodies. Representative images were acquired using an optical microscope (Nikon, Shinagawa-ku, Japan) and a laser-scanning confocal microscope (Zeiss LSM 880 with Airyscan, Germany). Fluorescence intensity was quantified using ImageJ software (NIH).

Pulmonary inflammation scores

According to the previous study (30), 1ung tissues were fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned for H&E staining. Three random high-power fields (400×) per slide were scored on a 0–4 scale: 0 (normal alveoli, no pathology); 1 (focal mild infiltration in ≤10% of field); 2 (patchy infiltration/minimal exudate in 11–30% of field); 3 (confluent infiltration/consolidation in 31–60% of field); 4 (diffuse infiltration/extensive consolidation/necrosis in >60% of field). The pneumonitis score per mouse was calculated as the mean of all fields.

Western blot analyses

Lung tissue proteins were extracted using RIPA lysis buffer and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Resolved proteins were transferred onto nitrocellulose membranes, which were blocked with 5% (w/v) non-fat dry milk for 1 hour at room temperature. Membranes were subsequently incubated overnight at 4 °C with primary antibodies (1:1,000 dilution), followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:2,000 dilution) for 1 hour at room temperature. Protein signals were detected using enhanced chemiluminescence (ECL) reagents and visualized with a Tanon^TM 5200 Ultra-Sensitive Multifunctional Chemiluminescence Imager (Biotton; Shanghai, China). Primary antibodies employed in this study are cataloged in Table S2.

Single-cell sequencing

Single-cell RNA-sequencing (scRNA-seq) data were obtained from public repositories (31). Raw scRNA-seq data from 13 clinical samples [7 CIP (+) and 6 CIP (−) cases] were sourced from the Genome Sequence Archive (GSA)-Human database under accession number HRA002094. Processed datasets necessary for replication of analytical results and figure generation are available through the OMIX repository (China National Center for Bioinformation) under accessions OMIX001006 and OMIX004420. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Statistical analysis

Quantitative data are expressed as mean ± standard deviation (SD). Statistical comparisons between two groups were performed using two-tailed Student’s t-tests, while comparisons across multiple groups employed one-way analysis of variance (ANOVA) followed by appropriate post hoc testing. A threshold of P<0.05 was applied to determine statistical significance. All statistical analyses and graphical representations were generated using GraphPad Prism^® version 10.0 (GraphPad Software, San Diego, CA, USA).

Results

Treg depletion combined with anti-PD-1 antibody therapy induces upregulation of NLRP3 inflammasome signaling pathway expression in pneumonitis tissues

Foxp3-DTR mice received subcutaneous implantation of LLC cells, and those receiving anti-PD-1 antibody treatment exhibited significantly enhanced survival rates relative to controls, confirming therapeutic efficacy (Figure S1A). Following euthanasia of CIP model mice established through Treg depletion combined with anti-PD-1 treatment, pneumonitis severity was evaluated by pulmonary CT and histopathological examination (H&E staining). Specifically, CT imaging revealed characteristic pneumonitis manifestations, including pneumonitis-like, ground-glass opacities, and interlobular septal thickening (Figure 1A). Histological analysis demonstrated features of interstitial pneumonitis, notably inflammatory cell infiltration, alveolar wall thickening, fibrotic interstitial expansion, and tissue consolidation (Figure 1B). Quantitative assessment confirmed significantly elevated pulmonary inflammation scores in the anti-PD-1 group versus control and DT groups (Figure 1C). Serum analysis detected significantly increased IFN-γ, TNF-α, and IL-6 levels in anti-PD-1-treated mice compared to controls, whereas the DT group showed no significant cytokine alterations (Figure 1D). Mechanistic investigation via RNA-seq of lung tissues from control, DT, and anti-PD-1 groups identified DEGs (Figure S1B). GSEA and GO enrichment analysis revealed activation of NLRP3 inflammasome-related pathways and assembly pathways in CIP mouse lung tissue, whereas no significant alterations were observed in the Treg depletion-only group (Figure 1E-1G). Collectively, these data suggest that anti-PD-1 antibody treatment promotes CIP through NLRP3 inflammasome pathway activation.

Figure 1.

Treg depletion combined with anti-PD-1 antibody therapy induces upregulation of NLRP3 inflammasome signaling pathway expression in pneumonitis tissues. Foxp3-DTR mice (n=5 per group) were assigned to: control, DT, and DT + anti-PD-1 cohorts. (A) Representative pulmonary CT scans. (B) H&E-stained lung sections (scale bar: 100 µm). (C) Quantitative assessment of pulmonary inflammation scores. (D) IFN-γ, TNF-α, and IL-6 concentrations. RNA-sequencing was performed on lung tissues from subsets of control (n=3), DT (n=2), and anti-PD-1 (n=3) mice to identify DEGs. (E-G) Differentially expressed genes associated with NLRP3 inflammasome GESA and GO pathway functional enrichment maps. Data represent mean ± SD; ns, no statistically significant; **, P<0.01; ***, P<0.001; ****, P<0.0001 versus control group. CT, computed tomography; DEGs, differentially expressed genes; DT, diphtheria toxin; H&E, hematoxylin and eosin; GESA, Gene Set Enrichment Analysis; GO, Gene Ontology; IFN-γ, interferon-gamma; IL-6, interleukin-6; NLRP3, NOD-like receptor family pyrin domain-containing protein 3; NLRP3, NOD-like receptor family pyrin domain-containing protein 3; ns, not statistically significant; PD-1, programmed death receptor-1; SD, standard deviation; TNF-α, tumor necrosis factor-alpha.

NLRP3 inflammasome activation in macrophages of DT + PD-1-induced mouse pneumonitis

To further investigate the transcriptomic differences, we collected lung and blood samples from control, DT, and anti-PD-1 groups with lung and blood samples for RNA-seq. Specifically, the RNA-seq data analysis revealed significant upregulation of NLRP3 inflammasome assembly-associated genes, including Gbp5, Casp1, and Pycard (Figure 2A), in anti-PD-1 group. Quantitative PCR validation confirmed concordant expression patterns (Figure 2B). Anti-PD-1 treatment additionally elevated mRNA levels of Nlrp3, Gsdmd, and Il1b (Figure 2C). IF analysis demonstrated substantially increased macrophage infiltration and ASC speck formation in lung tissues from anti-PD-1-treated mice relative to controls (Figure 2D-2G). Transcriptional profiling further indicated enrichment of NLRP3 effector pathways governing downstream inflammatory mediators (IL-1β and IL-18) in both pulmonary tissue and systemic circulation of CIP mice (Figure 2H,2I). Elevated serum concentrations of IL-1β and IL-18 were biochemically confirmed in the anti-PD-1 cohort (Figure 2J). SAA levels were similarly elevated in treatment groups (Figure S2), with emerging evidence suggesting SAA may function as a NLRP3 activator, though mechanistic validation remains necessary. Collectively, these data indicate that anti-PD-1 therapy promotes pneumonitis through NLRP3 inflammasome activation and downstream inflammatory cascades, with potential SAA involvement.

Figure 2.

NLRP3 inflammasome activation in macrophages in DT + PD-1-induced mouse pneumonitis. (A) Heatmap of NLRP3 inflammasome assembly-related genes. (B,C) Quantitative mRNA expression of Gbp5, Casp1, Pycard, Nlrp3, Gsdmd, Il1b, and Il18. (D) Representative immunofluorescence images of F4/80⁺ macrophages (green) and ASC specks (red) in lung tissue (scale bar: 20 µm; nuclei counterstained with DAPI). (E,F) Quantification of F4/80⁺ macrophage infiltration and ASC speck formation, respectively. (G) Co-localization analysis of ASC⁺ macrophages. (H,I) GO and GSEA plots demonstrating NLRP3 effector pathways in pulmonary and systemic compartments. (J) IL-1β and IL-18 concentrations. Data represent mean ± SD. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001 versus control group. a.u., arbitrary unit; DT, diphtheria toxin; GESA, Gene Set Enrichment Analysis; GO, Gene Ontology; IL, interleukin; NLRP3, NOD-like receptor family pyrin domain-containing protein 3; ns, not statistically significant; PD-1, programmed death receptor-1; SD, standard deviation.

NLRP3 inflammasome inhibitors and macrophage-depleting agents improve anti-PD-1 antibody therapy-induced pneumonitis

To delineate whether NLRP3 inflammasome activation functions as an upstream driver or concomitant event in Treg depletion-associated anti-PD-1-induced pneumonitis, therapeutic intervention was performed using two NLRP3 inhibitors (MCC950 and dapansutrile) and a macrophage-depleting agent (clodronate liposomes) (Figure 3A). MCC950 selectively inhibits NLRP3-dependent ASC oligomerization (32), while dapansutrile targets NLRP3 ATPase activity (33). Clodronate liposomes was employed to deplete tissue-resident and circulating macrophages—the primary cellular platform for NLRP3 inflammasome assembly. No significant differences in body weight dynamics or tumor progression were observed among MCC950-, dapansutrile-, and clodronate-treated cohorts relative to the anti-PD-1 group (Figure 3B,3C). However, pulmonary CT revealed attenuated radiological pathology (Figure 3D), and H&E staining confirmed reduced lung injury across all intervention groups (Figure 3E,3F). Concomitant reductions in serum IFN-γ, TNF-α, and IL-6 levels were observed (Figure 3G). These findings demonstrate that NLRP3 blockade and macrophage depletion comparably ameliorate anti-PD-1-induced pneumonitis, supporting NLRP3 inflammasome activation as a principal upstream pathogenic mechanism.

Figure 3.

NLRP3 inflammasome inhibitors and macrophage-depleting agents improve anti-PD-1 antibody therapy-induced pneumonitis. (A) Experimental design: mice (n=5/group) received MCC950 and dapansutrile or clodronate liposomes (macrophage depletion) alongside anti-PD-1 therapy. (B) Longitudinal body weight changes. (C) Tumor area kinetics. (D) Representative pulmonary CT scans. (E) H&E-stained lung sections (scale bar: 100 µm). (F) Quantitative pulmonary inflammation scores. (G) Serum IFN-γ, TNF-α, and IL-6 concentrations. Data represent mean ± SD. ***, P<0.001; ****, P<0.0001 versus control group. CT, computed tomography; DT, diphtheria toxin; H&E, hematoxylin and eosin; IFN-γ, interferon-gamma; IL-6, interleukin-6; LLC, Lewis lung carcinoma; NLRP3, NOD-like receptor family pyrin domain-containing protein 3; PD-1, programmed death receptor-1; SD, standard deviation; TNF-α, tumor necrosis factor-alpha.

Dapansutrile demonstrates superior efficacy and safety in inhibiting NLRP3 inflammasome activation compared with other inhibitors

According to the above research, administration of MCC950, dapansutrile, and clodronate liposomes significantly attenuated anti-PD-1 antibody-induced pulmonary macrophage infiltration. Both MCC950 and dapansutrile suppressed ASC oligomerization, whereas clodronate liposomes exhibited no significant effect on this process (Figure 4A-4C). All three therapeutic interventions reduced serum concentrations of IL-1β and IL-18 (Figure 4D). Safety assessments revealed hepatic inflammatory cell infiltration in MCC950-treated mice and splenic structural abnormalities in the clodronate liposomes group. In contrast, dapansutrile administration induced no significant histopathological alterations in major organs, demonstrating an improved safety profile (Figure S3). Mechanistically, dapansutrile downregulated NLRP3 expression via NF-κB pathway inhibition while concurrently suppressing caspase-1 activation, IL-1β secretion, and GSDMD cleavage (Figure 4E,4F). Collectively, dapansutrile effectively mitigates anti-PD-1-induced pneumonitis through dual-pathway inhibition of NLRP3 priming and inflammasome activation, presenting a favorable safety-efficacy profile for clinical translation.

Figure 4.

Dapansutrile inhibits NLRP3 inflammasome activation. (A) Representative immunofluorescence images of F4/80⁺ macrophage infiltration (green) and ASC specks (red) in lung sections (scale bar: 20 µm). (B,C) Quantitative analysis of F4/80⁺ area and ASC⁺ foci density derived from (A). (D) Serum IL-1β and IL-18 concentrations measured by ELISA. (E) Representative immunoblots of NF-κB pathway (NF-κB, p-NF-κB) and NLRP3 inflammasome components (NLRP3, cleaved caspase-1, GSDMD, cleaved-IL-1β) with β-actin loading control. (F) Densitometric quantification of protein expression normalized to β-actin. Data represent mean ± SD. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001 versus control group. DT, diphtheria toxin; ELISA, enzyme-linked immunosorbent assay; GSDMD, gasdermin D; IL, interleukin; NLRP3, NOD-like receptor family pyrin domain-containing protein 3; ns, not statistically significant; PD-1, programmed death receptor-1; SD, standard deviation.

NLRP3 inflammasome-associated macrophages are increased in the bronchoalveolar lavage fluid (BALF) of CIP patients

Utilizing BALF samples from a cohort of 13 non-small cell lung cancer (NSCLC) patients treated with ICIs, we systematically characterized immune cell populations. Participants were stratified into two groups: seven who developed ICI-associated pneumonitis [CIP(+) group, comprising one grade 1 and six grade 2 cases] and six without pneumonitis [CIP(−) group]. Integrated bioinformatics analysis identified six major cell types (Figure 5A,5B). Subsequent myeloid subclustering delineated eight distinct subsets: CD1C⁺ dendritic cell (DC-CD1C), LAMP3⁺ dendritic cell (DC-LAMP3), XCR1⁺ dendritic cell (DC-XCR1), APOC2⁺ macrophage (Mac-APOC2), FCN1⁺ macrophage (Mac-FCN1), FOLR²⁺ macrophage (Mac-FOLR2), IL1B⁺ macrophage (Mac-IL1B), and PLA2G16⁺ macrophage (Mac-PLA2G16) (Figure 5C,5D), with differentiation trajectories mapped (Figure 5E,5F). Notably, NLRP3 expression was significantly upregulated in the Mac-IL1B subpopulation (Figure 5G). Single-cell transcriptomics revealed substantially greater Mac-IL1B infiltration in CIP(+) BALF versus CIP(−) controls (Figure 5H). Pathway enrichment analysis demonstrated specific activation of the NLRP3 inflammasome pathway and downstream proinflammatory factor secretion (e.g., IL-1β, IL-18) within this subset (Figure 5I). These results confirm the pivotal role of NLRP3 inflammasome activation in CIP pathogenesis. It should be noted that only mild-to-moderate (grades 1–2) pneumonitis cases were included, potentially excluding features of severe disease. While preclinical modeling successfully recapitulated the pathological progression of severe (grades 3–4) CIP, the relevance of these findings to critically ill patients requires further investigation. Collectively, clinical samples validate that Mac-IL1B cells drive pneumonitis through NLRP3 pathway activation, consistent with preclinical models, though mechanistic insights into severe disease remain limited.

Figure 5.

NLRP3 inflammasome-associated macrophages are increased in the bronchoalveolar lavage fluid of CIP patients. (A) Unbiased clustering of BALF cells was performed, with (B) subsequent cell type annotation. (C) Myeloid cell subclustering revealed eight distinct subpopulations, while (D) a dot plot illustrates scaled differential gene expression across these myeloid subclusters. (E,F) Cellular differentiation trajectories were reconstructed. (G) NLRP3 gene expression levels were quantified across the eight myeloid subclusters. (H) The relative abundance of Mac-IL1B cells was compared across clinical cohorts, with statistical significance assessed using Student’s t-test. (I) Pathway enrichment analysis demonstrated significant activation of NLRP3 inflammasome signaling pathways specifically within Mac-IL1B cells. Data represent mean ± SD. **, P<0.01; ****, P<0.0001. BALF, bronchoalveolar lavage fluid; BC, bronchoalveolar lavage fluid from CIP (−) patients; BP, bronchoalveolar lavage fluid from CIP (+) patients; CIP, checkpoint inhibitor-associated pneumonitis; DC, dendritic cell; IL, interleukin; Mac, macrophage; NLRP3, NOD-like receptor family pyrin domain-containing protein 3; SD, standard deviation; UMAP, Uniform Manifold Approximation and Projection.

Discussion

ICIs constitute a pivotal breakthrough in cancer immunotherapy. By blocking inhibitory signals on T-cell surfaces-notably PD-1 and CTLA-4—these agents overcome tumor-mediated immune suppression, thereby restoring antitumor T-cell activity (34). The discovery and clinical implementation of ICIs have fundamentally transformed therapeutic paradigms for advanced malignancies, offering durable clinical responses, including long-term survival and potential cure in selected patient cohorts. Among diverse ICI classes, anti-PD-1/PD-L1 inhibitors have emerged as the dominant therapeutic approach in clinical oncology, owing to their broad activity spectrum, sustained efficacy, and manageable toxicity profiles (35). However, treatment-related adverse events necessitate vigilant monitoring and management. The NLRP3 inflammasome, a core component of the innate immune system, recognizes both exogenous pathogens and endogenous damage signals, playing an essential role in the pathogenesis of immune-mediated and inflammatory disorders (19,36).

A murine model of ICI-associated pneumonitis was developed using DTR transgenic mice treated with anti-mouse PD-1 antibodies. Pulmonary CT scans revealed inflammatory exudation, tissue fibrosis, with concomitant elevation of serum inflammatory cytokines in experimental groups. RNA-seq of murine lung and blood specimens revealed transcriptomic pathways dysregulated in CIP following PD-1 blockade. This analysis demonstrated that anti-PD-1 therapy induced upregulation of macrophage activation and chemotaxis pathways, pyroptosis-related genes, and NLRP3 inflammasome assembly components, indicating NLRP3 inflammasome activation. Western blot and IF analyses confirmed significantly increased protein expression of core NLRP3 inflammasome constituents in lung tissues of anti-PD-1-treated mice. These collective findings establish that PD-1 blockade triggers NLRP3 inflammasome activation within pulmonary tissue.

Recent research utilizing human data has successfully established IL-6 as a critical therapeutic target for CIP, with its inhibitors demonstrating encouraging efficacy in clinical practice. This advancement underscores a central research question: are there alternative, equally effective, or complementary therapeutic avenues beyond IL-6-targeted intervention? MCC950 and dapansutrile represent therapeutically validated NLRP3 inflammasome inhibitors characterized by potent bioactivity and favorable bioavailability. MCC950 directly inhibits ASC oligomerization, while dapansutrile suppresses NLRP3 ATPase activity through distinct molecular mechanisms (37,38). Both inhibitors conferred significant protection against CIP development. Crucially, while both agents substantially reduced serum pro-inflammatory mediators (IFN-γ, TNF-α, IL-6) and attenuated pulmonary inflammatory infiltration, their mechanisms diverged fundamentally. Dapansutrile demonstrated a unique dual mechanism: directly inhibiting NLRP3 inflammasome assembly while transcriptionally downregulating NLRP3 expression via NF-κB pathway suppression. Critically, divergent safety profiles were demonstrated. While MCC950 exhibited potent inhibitory activity in preclinical models, the current investigation revealed its capacity to elicit histologically confirmed hepatic inflammatory infiltration, consistent with prior reports (39). Conversely, dapansutrile treatment elicited no significant organ toxicity. This discrepancy is attributed to fundamental mechanistic differences: MCC950 achieves irreversible inhibition via covalent modification of NLRP3’s Walker B motif, a process that may disrupt cellular homeostasis and contribute to off-target effects (40). In contrast, dapansutrile functions as a competitive ATP analog that reversibly binds NLRP3, demonstrating enhanced target specificity. This finding indicates that for specific CIP subtypes primarily driven by the NLRP3/IL-1β axis, shifting the therapeutic target upstream to the inflammasome itself may constitute a viable and promising investigative pathway. Specifically, the clinical value of dapansutrile may be observed across several dimensions. First, as an orally administered agent, it offers significant practical convenience for patients compared to biologic therapies requiring intravenous infusion. Second, its target, the NLRP3 inflammasome, is positioned upstream of IL-1β and exhibits partial overlap with, while maintaining relative independence from, the IL-6 signaling pathway. This distinct mechanistic profile suggests that dapansutrile may provide a novel therapeutic option for patients exhibiting a suboptimal response to existing IL-6 inhibitors. In the long term, the potential to combine it with established therapies, such as anti-IL-6 agents, to achieve synergistic inhibition across multiple inflammatory pathways holds promise as a novel strategic approach for managing the most challenging refractory cases.

M1 macrophages are recognized as the principal pulmonary cells capable of assembling NLRP3 inflammasomes (41,42). Prior studies established that damage-associated molecular patterns (DAMPs) activate the NLRP3 inflammasome in alveolar macrophages, triggering caspase-1-mediated maturation and release of IL-1β and IL-18, thereby amplifying the inflammatory cascade (43). Clinical investigations have identified significant accumulation of NLRP3-expressing macrophages in BALF from patients with severe coronavirus disease 2019 (COVID-19)-associated acute respiratory distress syndrome (ARDS), particularly those requiring mechanical ventilation (44). Given the established role of NLRP3 inflammasome activation in driving anti-PD-1 antibody-induced CIP, we hypothesized that targeted depletion of pulmonary macrophages would mitigate pneumonitis development. To determine macrophage dependency in CIP pathogenesis, we administered clodronate liposomes, a well-validated macrophage-depleting agent, to achieve complete pulmonary macrophage ablation prior to anti-PD-1 challenge. Histopathological examination and inflammatory cytokine concentration assays demonstrated that pretreatment with clodronate liposomes significantly attenuated CIP severity. Consequently, these findings definitively establish that anti-PD-1 antibodies trigger NLRP3 inflammasome assembly specifically within pulmonary macrophages.

In this study, single-cell transcriptomic analysis was performed on BALF obtained from NSCLC patients undergoing ICI treatment. A distinct macrophage subset characterized by high expression of Il1b and Nlrp3 was identified and designated as the Mac-IL1B subset. Notably, this subset demonstrated significant transcriptomic similarity to the pathogenic macrophage population—previously identified in human CIP patients—which exhibits elevated expression of both Il1b and Il6 (45,46). These findings directly implicate NLRP3 signaling’s pathogenic role in lung toxicity documented in preclinical models (47). This mechanistic distinction further supports the translational promise of targeting the NLRP3/IL-1β axis in immunotherapy toxicity. For instance, efficacy of anakinra (an IL-1 receptor antagonist) against immunotherapy-related toxicities has been reported (48,49), and this discovery provides theoretical justification for its application in CIP. Notably, Mac-IL1B defines a distinct macrophage/monocyte subpopulation exhibiting a pathogenic pro-inflammatory signature. Its elevated IL1B expression aligns with classical inflammatory monocytes/macrophages (50), while selective NLRP3 pathway upregulation implicates microenvironment-driven pathological activation within CIP.

The mouse model of pneumonitis employed in this study was based on previous research, which incorporates Treg depletion alongside anti-PD-1 treatment, indicates a potentially critical role for the NLRP3 inflammasome in this pathogenic process. However, it must be acknowledged that this model possesses certain inherent limitations. The experimental model utilized strong dual immune perturbation to simulate a pneumonitis pathology driven by a severe immune imbalance. In clinical practice, however, the etiology of immune-related pneumonitis is considerably more complex, as the immunological profiles induced by different therapeutic agents vary significantly. Not all patients concurrently experience pronounced Treg dysfunction alongside PD-1 pathway blockade. Therefore, this model may be more representative of severe CIP or pneumonitis induced by specific combination immunotherapy regimens, and may not fully encompass the entire spectrum of CIP manifestations. Furthermore, inherent immunological differences between mice and humans necessitate caution when directly extrapolating these findings to the clinical setting. Further investigation is required to validate and expand upon these mechanistic insights using clinical patient samples, alternative CIP models, or humanized mouse systems. Despite these limitations, under defined conditions, this model can still function as a mechanism-driven and highly controllable experimental platform, providing valuable insights into the fundamental processes underlying CIP pathogenesis. Experimental data demonstrate that NLRP3 inflammasome inhibition effectively attenuates pneumonitis manifestations in the model, providing a rationale for the hypothesis that this pathway holds potential as a therapeutic target. However, this finding must be interpreted with caution. On one hand, direct evidence establishing a precise causal role for the NLRP3 inflammasome in the pathogenesis of human CIP remains to be obtained. On the other hand, clinically validated targets, such as IL-6, are themselves regulated by upstream pathways, including NF-κB. This suggests that therapeutic intervention at multiple nodes within the inflammatory cascade—whether at the relatively upstream NLRP3/IL-1β axis or at the validated downstream target IL-6—may each yield beneficial effects in the treatment of CIP. Therefore, dapansutrile can be considered a potential alternative or complementary therapeutic strategy, distinguished by its targeting of a distinct molecular node and its practical advantage of oral administration, alongside the currently successful IL-6-targeted approach. Nevertheless, its precise clinical value and role in the management of human CIP require further elucidation through subsequent clinical investigations.

Conclusions

In summary, this study not only confirms the central mechanistic role of the NLRP3 inflammasome in the development of CIP but also provides translational evidence supporting its potential as a target for clinical intervention. Using multiple lines of experimental evidence—including a mouse model combining DTR-mediated cell depletion with anti-PD-1 treatment, as well as clinical patient-derived single-cell sequencing data—the following pathological scenario was progressively delineated: the immune imbalance induced by anti-PD-1 antibodies activates a specific pro-inflammatory macrophage subset (Mac-IL1B), which subsequently amplifies pulmonary inflammation via the NLRP3/IL-1β axis. In the context of therapeutic intervention, our findings also yield clinically relevant insights. While IL-6 blockers are currently established as standard therapy, dapansutrile may offer distinct potential advantages. This agent not only presents a favorable safety profile and the convenience of oral administration but also targets NLRP3, a component positioned upstream in the inflammatory cascade. This mechanistic profile positions dapansutrile as a promising candidate within a stratified treatment framework for CIP. Whether employed as monotherapy for mild to moderate cases or as a complementary strategy in combination with IL-6 inhibitors for refractory disease, this approach merits further investigation. However, a significant translational gap remains between murine models and the complexity of human disease, necessitating caution in the interpretation and extrapolation of these findings. Overall, however, this study suggests that targeting key components of innate immunity, such as the NLRP3 inflammasome, may represent a viable strategy for restoring pulmonary immune homeostasis following the disruption induced by immune checkpoint blockade. Future clinical studies are required to evaluate whether such an ‘upstream intervention’ strategy can indeed improve the quality of life for cancer survivors.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All animal experiments were performed under a project license (approval No. 2023-X19-105) granted by the Animal Welfare and Ethics Committee of the Chinese PLA General Hospital, in compliance with relevant Chinese regulations and institutional guidelines for the care and use of animals.

Data Sharing Statement

Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1240/dss

tlcr-15-02-26-dss.pdf ^{(141.4KB, pdf)}

DOI: 10.21037/tlcr-2025-aw-1240