STING agonist diABZI induces PANoptosis and DNA mediated acute respiratory distress syndrome (ARDS)

We first addressed the effect of the natural cyclic dinucleotide STING ligand cGAMP in the airways. Endotracheal administration of cGAMP (1, 3, or 10 µg) for 3 consecutive days induced airway inflammation within 24 h (Fig. 1A). The neutrophilic attracting chemokine CXCL1/KC was released in the bronchoalveolar lavage fluid (BALF; Fig. 1B), and this was accompanied by increased neutrophil recruitment (Fig. 1C) and myeloperoxidase (MPO), a marker of neutrophil presence and activation in the BALF (Fig. 1D) and in the lung tissue (Fig. 1E). cGAMP-induced airway dysfunction was evidenced by increased total protein extravasation (Fig. 1F), self-dsDNA release in the bronchoalveolar space (Fig. 1G), and IFNα and IFNβ in the airways (Fig. 1H, I).

Thus, the natural CDN STING ligand cGAMP is a potent inducer of neutrophilic inflammation in the airways that may cause cell damage and lung dysfunction. Indeed, airway cGAMP resulted in acute alveolitis with thickened alveolar septae containing inflammatory cells, partial disruption of the alveolar membrane and rarefaction of alveoli, acute bronchiolitis with cell death and a beginning of remodeling with peribronchial infiltrates, visible 24 h after 1 µg, and further increased after 3-10 µg cGAMP endotracheal administration (Fig. 1J–N), but no mucus secretion (PAS staining). These are typical features following a massive injury with inflammation and incipient emphysema, repair, and fibrosis (Fig. 1J–N).

STING activation triggers several inflammatory pathways including IRF3 and NF-κB activation. We show that cGAMP administration at 1 to 10 µg increased STING dimers, TBK1 phosphorylation (pTBK1), IRF3, cGAS, NF-κB, and NF-κB p65 subunit phosphorylation (p-p65 NF-κB) in the lung tissue (Fig. 1O and Supplemental Fig. S5A). We next asked whether both pathways were functionally activated following the administration of cGAMP in the airways. Indeed, the chemokine Cxcl10 mRNA, a typical Interferon-stimulated gene (ISG) downstream of IRF3/type I IFN, was overexpressed in the lung tissue after administration of 3-10 µg cGAMP in the airways (Supplemental Fig. S1A). CXCL10/IP-10 was increased in the BALF after 10 µg cGAMP administration (Supplemental Fig. S1B), as were TNFα and IL-6, both triggered in response to NF-κB pathway (Supplemental Fig. S1C, D). Interestingly, the expression of STING gene Tmem173 doubled in the lung 1 day after cGAMP administration (Supplemental Fig. S1E).

Thus, airway cGAMP administration induces strong lung inflammation and injury, associated with dsDNA release, type I IFN and inflammatory cytokines in the lung, and overexpression of STING transcript.

Since airway cGAMP administration induced extracellular dsDNA release in the BALF, we hypothesized that this dsDNA might act as a danger signal, activate DNA sensors, and induce further type I IFN response.

We analyzed the kinetics of cGAMP-induced responses in vitro in bone marrow-derived macrophages (Fig. 2). Type I IFNα and IFNβ release at 4 h, further increased 16 h after macrophage activation by cGAMP (Fig. 2A, B). CXCL10 was also strongly increased at 4–16 h (Fig. 2C). Ifnβ gene was already overexpressed 2 h after cGAMP addition (Fig. 2D), and phosphorylation of STING, TBK1 and IRF3 occurred already at 1 h, while STAT1 phosphorylation was delayed to 2 h (supplemental Fig S2 and S5G). ISG includes DNA sensors such as cGAS, AIM2, or IFI204, and we verified their expression. There was overexpression of cGAS gene Mb21d1, Aim2, and Ifi204 genes at 4–16 h (Fig. 2E–G), while Ddx41 and Nlrp3 genes were barely affected (Fig. 2H, I). STING gene Tmem173 itself was overexpressed at 16 h (Fig. 2J). CXCL1 and IL-10 were increased at 16 h (Fig. 2K, L). We verified that the effects of cGAMP was indeed mediated by STING using Tmem173-deficient (STING^−/−) macrophages. As expected, Ifnβ, Ifna2, and Ifna4 genes expression, and IFNα and IFNβ release were abrogated in STING^−/− macrophages (Fig. 2M–Q).

Thus, activation of STING by the natural CDN agonist cGAMP triggers a well-coordinated expression of type I IFNs and ISGs including several DNA sensors that may contribute to downstream chemokines and cytokines release.

To further characterize the effects of triggering STING activation in the airways, we turned to the recently developed and highly potent synthetic, non-nucleotidyl molecule diABZI [14]. We first checked the response induced by diABZI on murine macrophages. DiABZI at 1 µM induced the release of IFNα, IFNβ, CXCL10, IL-6, TNFα, CXCL1, and IL-10 at 16 h (Fig. 3A–G), similar to the natural CDN cGAMP, and this was abrogated in STING^−/− macrophages (Fig. 3A–G).

DiABZI at 0.3–1 µM induced the phosphorylation of STING and downstream TBK1 kinase, together with STAT1 phosphorylation, similar to cGAMP (Fig. 3H and supplemental Fig. S5B). The relative reduction of STING protein in conditions when STING is phosphorylated is intriguing, not due to antibody artifact as it is seen in independent blots, and might be related to the STING protein turnover in cells in culture in vitro. DNA damage was revealed by phosphorylated-γH2AX (pγH2AX; Fig. 3H and supplemental Fig. S4A). Indeed, cell death was also induced, as documented by immunoblots with elevated caspase 3, cleaved Gasdermin D (GSDMD) as a marker of pyroptosis, elevated Mixed Lineage Kinase domain Likepseudokinase (MLKL), and phospho-MLKL as a marker of necroptosis (Fig. 3I and supplemental Fig. S4A). This was suggestive of PANoptosis, and confirmed as the Z-DNA-binding protein 1 (ZBP1) was also upregulated (Fig. 3I and supplemental Fig. S5B). These responses were reduced in STING^−/− mice (Fig. 3H, I), documenting the STING dependence of the induced responses.

We next characterized the response induced by diABZI on primary human airway epithelial cells (hAEC). DiABZI at 1–10 µM induced the expression of IFNβ, IL-1β, and IL-8 (Fig. 3J–L). Indeed, diABZI induced the phosphorylation of STING and downstream TBK1 kinase, as well as overexpression of cGAS, but it also induced cell death, documented by cleaved Caspase 3 as a marker of apoptosis and DNA damage with increased pγH2AX (Fig. 3M and supplemental Fig. S5C).

Thus, diABZI induced activation of the STING pathway and cell death with upregulation of markers of apoptosis, pyroptosis, and necroptosis, indicative of STING-dependent PANoptosis, a term recently introduced [13].

Endotracheal administration of diABZI for 3 consecutive days induced a strong airway inflammation within 24 h (Fig. 4A). CXCL1 was released in the bronchoalveolar space after 0.1–1 µg diABZI administration (Fig. 4B), alongside neutrophil recruitment (Fig. 4C) and increased MPO (Fig. 4D). Protein extravasation in the bronchoalveolar space (Fig. 4E), indicative of disruption of the respiratory barrier, was accompanied by self-dsDNA release (Fig. 4F), suggesting cell damage and/or cell death. As cell stress or death may release mitochondrial DNA in the cytosol or extracellularly, that acts as a strong danger signal recognized by nucleic acid sensors such as cGAS, we quantified nuclear and mitochondrial DNA released in the BALF in response to diABZI or cGAMP. While nuclear DNA was barely released, there was a strong increase in mtDNA after diABZI 1 µg and to a lesser extent after exposure to cGAMP 10 µg (Fig. 4G, H), with a 35- to 42-fold mtDNA/nDNA ratio after cGAMP and diABZI exposure, respectively.

We next asked whether both IRF3 and NF-κB pathways were activated following STING activation by diABZI administration. Although no phosphorylated TBK1, IRF3, or p65 NF-κB could be detected at this time point (not shown), there was a functional activation of these pathways. Indeed, IFNα and IFNβ were released in the BALF after administration of diABZI at 0.1–1 µg (Fig. 4I, J), as was CXCL10, downstream of type I IFN/IRF3, after 1 µg diABZI (Fig. 4K), and both IL-6 and TNFα, triggered in response of NF-κB pathway (Fig. 4L, M).

DiABZI activating neutrophilic responses, we further investigated the formation of NETs, a neutrophil innate effector mechanism. Indeed, NETs comprising neutrophil-derived DNA, citrullinated histone H3 (Cit-H3), and MPO were increased by diABZI administration in the airways (Fig. 4N–R).

Histologically, marked inflammatory cell infiltration in the lung parenchyma was visible 24 h after three doses of 0.01 µg, and furthermore 0.1–1 µg diABZI endotracheal administration (Fig. 4S–W) with signs of epithelial injury, alveolitis, and emphysema, but no mucus secretion revealed by PAS staining.

Thus, the synthetic, non-nucleotidyl, STING ligand diABZI is a potent inducer of neutrophilic inflammation in the airways that may cause cell damage, dsDNA release and NETosis, and lung dysfunction.

As STING may mediate apoptosis [12, 15], we characterized the mode of cell death induced by diABZI in the lung. Airway administration of diABZI at 0.1 and more at 1 µg induced caspase 3 cleavage in the lung tissue, indicative of STING activation-induced apoptosis (Fig. 5A and supplemental Fig S5D). The occurrence of pyroptosis and necroptosis in diABZI-induced pulmonary cell death was assessed by immunoblots of cleaved Gasdermin D (GSDMD) and phosphorylated MLKL. Indeed, GSDMD and MLKL were increased in the lung, with the presence of cleaved GSDMD and phosphorylated MLKL (54 kD lower band) after administration of diABZI at 0.1 and further augmented at 1 µg (Fig. 5A, B). These data were suggestive of PANoptosis induced in the lung after exposition to diABZI, and was confirmed as the Z-DNA-binding protein 1 (ZBP1) was also upregulated (Fig. 5A, B). DNA damage was documented as increased levels of pγH2AX after diABZI treatment (Fig. 5A, B). Cit-H3 detection provided evidence of pulmonary NETosis triggered by diABZI administration (Fig. 5A, B).

To further characterize PANoptosis we next analyzed ASC-Caspase 8-RIPK3 complex formation [16, 17]. Indeed, in sterile inflammation, the Caspase 8-RIPK3 shared by apoptosis and necroptosis is generally associated with ASC, the adapter protein of inflammasomes [18], these molecular interactions likely serving as the backbone for PANoptosome formation. Confocal microscopy analysis of immunofluorescence labeling shows increased Caspase 8, ASC, and RIPK3 protein expression in BAL cells of mice exposed to diABZI 1 µg as compared to saline controls, and colocalization of these PANoptosome components, indicative of PANoptopsis complex (Fig. 5C).

Therefore, triggering airway STING activation by local administration of a potent STING agonist, diABZI, induced lung cell death through apoptosis, necroptosis, and pyroptosis, with evidence of PANoptosis complex formation, indicating STING-induced PANoptosis.

We next asked whether triggering the STING pathway induced further DNA sensors expression. We report STING overexpression and activation visible by increased STING dimers in immunoblots of lung tissue after endotracheal administration of 0.1–1 µg diABZI (Fig. 5D, E and supplemental Fig S5E). cGAS protein expression was also strongly increased in the lung after the diABZI challenge (Fig. 5D, E). DNA sensors cGAS gene Mb21d1 and AIM2 are ISG, and their messages increased after the diABZI challenge, as were Tmem173, Ifi204, Nlrp3, less Ddx41, and >200-fold CxCl10 (Fig. 5F).

Therefore, triggering airway STING activation by local administration of the potent STING agonist diABZI induced DNA damage and overexpression of DNA sensors and inflammasomes that may fuel further inflammatory responses.

To evaluate the contribution of the self-dsDNA released after diABZI administration to the induced lung inflammation through secondary responses, we treated mice receiving local airway diABZI with DNase I (Fig. 6A). Pulmozyme DNase I treatment at 50 µg i.t. on the 3 consecutive days of diABZI administration efficiently abrogated dsDNA in the BALF (Fig. 6B), and reduced neutrophil recruitment and CXCL10 release in the airways (Fig. 6C, D), while other parameters such as IFNα, IFNβ (Fig. 6E, F), MPO, IL-6, or TNFα were barely affected (Supplemental Fig. S3A–C).

DNase I treatment resulted in the impairment of NETs formation as evidenced by DNA, MPO and citrullinated Histone 3 immunofluorescence analysis (Fig. 6G), and decreased ZBP1 protein in the airways (Fig. 6H).

The CD45⁺ inflammatory cells recruited in the airways after diABZI local administration showed increased frequency and absolute counts of necrotic Annexin V⁺ Propidium Iodide (PI)⁺ and apoptotic Annexin V⁺ PI⁻ cells, that were largely abrogated by the DNase I treatment (Fig. 6I, J and supplemental Fig. S3D).

To evaluate the interplay between STING activation and NETs formation, we treated mice receiving local airway diABZI with Cl-amidine, an inhibitor of peptidyl arginine deiminase 4 (PAD4), an enzyme essential for NET formation (Fig. 6K). Cl-amidine treatment at 200 µg i.p. on the 3 consecutive days of diABZI administration efficiently prevented neutrophil recruitment and NETs in the BALF (Fig. 6L, M). Treatment with Cl-amidine also reduced NETs formation as evidenced by Cit-H3 immunoblot, and the expression of STING protein in the lung (Fig. 6N, O).

To determine the link between NETosis and PANoptosis, we evaluated PANoptosis markers following Cl-amidine treatment in diABZI exposed mice. Cl-amidine treatment reduced diABZI-induced PANoptosis as evidenced by reduced cleaved caspase 3, caspase 3, cleaved GSDMD, GSDMD, pMLKL, MLKL, and ZBP1 proteins in lung, assessed by western blot (Fig. 6N, O and supplemental Fig S5F). Caspase 8, ASC, and RIPK3 protein expression in BAL cells was also reduced after Cl-amidine treatment, as seen by immunofluorescence (Fig. 6P). The diABZI-induced release of dsDNA in the BALF (Fig. 6Q), damaged DNA documented by immunoblot of pγH2AX in the lung (Fig. 6N, O), and overall lytic cell death measured by Lactate dehydrogenase (LDH) release in the BALF (Fig. 6R) were alleviated by Cl-amidine treatment.

Thus, the airway inflammatory response to diABZI local administration is in part dependent on the released dsDNA and NET formation.

As STING agonist cGAMP may activate alternative, noncanonical inflammasome priming, we first verified that the inflammatory airway response to diABZI administration is dependent on the STING pathway (Fig. 7A–D). Indeed, the recruitment of neutrophils and dsDNA release in the airways seen after endotracheal administration of 0.1–1 µg diABZI for 3 consecutive days in wild-type mice was absent in STING^−/− mice (Fig. 7A, B), as well as MPO release in BALF (Supplemental Fig. S4A). Further, IFNα and IFNβ release in the airways was absent in STING^−/− mice (Fig. 7C and Supplemental Fig. S4B), as were both CXCL10, downstream of IRF3/type I IFN (Fig. 7D), and NFκB-dependent IL-6 and TNFα (Supplemental Fig. S4C, D).

As dsDNA was released and cGAS expression was increased after diABZI administration, we then assessed the contribution of self-dsDNA-induced secondary responses, in cGAS-deficient mice. Interestingly, the airway inflammation induced by local diABZI administration was retained in cGAS^−/− mice, in terms of neutrophil recruitment, MPO, dsDNA, or cytokine release in the airways, including IFNα and IFNβ, CXCL10, IL-6, or TNFα, (Fig. 7E–H and supplemental Fig. S4E–H). This indicated a cGAS-independent response to diABZI and downstream dsDNA released. Similarly, while NLRP3 and AIM2 expression was increased after diABZI administration, the response to airway diABZI administration was unaffected in NLRP3^−/− and AIM2^−/− mice, in terms of neutrophil recruitment, MPO, dsDNA, or cytokine release in the airways, including IFNα and IFNβ, CXCL10, IL-6, or TNFα (Fig. 7I–P and supplemental Fig. S4I–P). As expected, the inflammatory responses to diABZI were abrogated in IFNAR^−/− mice (Fig. 7I–L and supplemental Fig. S4I–L), and the neutrophilic response was reduced in TLR9^−/− mice (Fig. 7R–U and supplemental Fig. S4R–U).

Thus, the airway inflammatory response to diABZI local administration is fully dependent on STING and IFNAR, and extracellular DNA and TLR9 are central to diABZI-induced neutrophilic response, independently of cGAS, NLRP3, and AIM2 pathways.

Female C57BL/6Rj mice were purchased from Janvier Laboratories (Le Genest St Isle, France). Wild-type mice and mice deficient for STING (STING^−/−) [23], cGAS (cGAS^−/−) [24], IFNAR (IFNAR^−/−) [25], TLR9 (TLR9^−/−) [26], NLRP3 (NLRP3^−/−) [27], AIM2 (AIM2^−/−) [28] were bred and housed under specific pathogen-free conditions at CNRS animal facility (TAAM UAR44, Orleans, France). They were maintained in a 12-h light-dark cycle with food and water ad libitum, following European and local legislation. Age-matched, 8- to 12- week-old mice were used for experiments. All animal experiments complied with the French Government animal experiment regulations and ARRIVE guidelines. The protocols were submitted to the “Ethics Committee for Animal Experimentation of CNRS Campus Orleans” (CCO) under numbers CLE CCO 2015-1087 and CLE CCO 2020-2018 and approved by the French Minister under APAFIS #19361 and #25360.

Mice were anesthetized with 2% Isoflurane (ISO-VET, Netherlands) and challenged intratracheally (i.t) with cGAMP (1–10 µg in 40 µL, Invitrogen, California) or diABZI, a STING agonist 3 (0.01 to 1 µg in 40 µL, Cayman chemicals, Michigan) or saline (NaCl 0.9%) for 3 consecutive days. Bronchoalveolar lavage (BAL) was performed 24 h after the last challenge by flushing lung tissue four times with 0.5 mL of cold NaCl 0.9% via a cannula inserted in the trachea. BALF was collected, cells counted and cytospins performed. The supernatant of the first lavage was collected after centrifugation and stored at −80 °C for dsDNA quantification. The left lung lobe was harvested for histology, the post caval lung for RNA extraction and qPCR analysis, and the right lobes for Western blots analysis and cytokines measurement. Protein extravasation in the BALF was measured by Pierce™ BCA Protein Assay (ThermoFisher^®, Massachusetts).

Lung tissues were homogenized in T-PER™ buffer supplemented with a cocktail containing protease and phosphatase inhibitors (halt™, ThermoFisher). Total protein was extracted and quantified by using Pierce™ BCA Protein Assay Kit (ThermoFisher®). Total protein (40 µg) was treated with NuPAGE™ LDS sample buffer and sample reducing agent (ThermoFisher®), and heated 10 min at 70 °C. Samples were resolved on 4–12% polyacrylamide gel (Bolt™ Mini protein gel, ThermoFisher) and run at 160 V for 45 min using the Mini gel Tank (ThermoFisher^®). Total proteins were electroblotted to 0.2 µm nitrocellulose membrane (Amersham™, UK) using a Trans-Blot SD Transfer System (Bio-Rad, California) at 100 V for 45 min. Successful protein transfer was confirmed by using Ponceau S staining. Membranes were blocked with 5% nonfat milk (Cell signaling, Massachusetts) in 1X TBS-T (20 mM Tris Base, 150 mM sodium chloride, and 0.05% Tween-20 pH 7.6) for 1 h at room temperature.

Full membranes or portions of membranes were incubated overnight using primary antibodies from rabbit anti-phospho-STING (#72971 1/500; Cell signaling), anti-STING (#13647 1/500; Cell signaling), anti-phospho-TBK1 (#5483 1/500; Cell signaling), anti-TBK1 (#3504 1/500; Cell signaling), anti-phospho-IRF3 (#4947 1/500; Cell signaling), IRF3 (#4302 1/500; Cell signaling), anti-phospho-NF-kB (#3031 1/500; Cell signaling), anti- NF-kB (#4764 1/500; Cell signaling), anti- cGAS (#31659 1/500; Cell signaling), anti-NLRP3 (#15101 1/500; Cell signaling), anti-MLKL (#37705 1/500; Cell signaling), anti-phospho-MLKL (#37333 1/500; Cell signaling), anti-GSDMD (#ab209845 1/500; Abcam), anti-cleaved GSDMD (#10137 1/500; Cell signaling), anti-cleaved Caspase 3 (#9661 1/500), anti-Caspase 3 (#9662 1/500; Cell signaling), anti-phospho-γH2AX (#9718 1/500; Cell signaling), anti-γH2AX (#7631 1/500; Cell signaling), anti-histone H3 (#ab5103 1/500; Abcam, UK), anti-IFI204 (#ab228512 1/500; Abcam), anti-phospho-STAT1 (#9167 1/500; Cell signaling), anti-STAT1 (#14994 1/500; Cell signaling), mouse anti-ZBP1 (#SC-271483 1/300; Santa Cruz, Texas), and anti actin (#A3854 1/10000; Sigma-Aldrich, Massachusetts). Membranes were washed in TBS-T three times for 10 min each at room temperature, and then incubated with goat anti-rabbit-IgG-HRP-conjugate (#7074 1/2000; Cell signaling) or horse anti-mouse-IgG-HRP-conjugate (#7076 1/2000; Cell signaling) diluted in 5% nonfat milk in TBS-T for 1 h at RT.

The membranes were washed three times in TBS-T. Protein bands were visualized following exposure of the membrane to Amersham ECL™ prime substrate solution (Cytiva, Massachusetts) on film (PXi gel doc system^®, India), and quantified by densitometry analysis using ImageJ software. Uncropped immunoblot gels are shown in Supplemental Fig. S5.

Lungs were fixed with 4% formalin for 72 h, embedded in paraffin, and sectioned at 3 µm. Lung sections were dewaxed and rehydrated, then heated 20 min at 80 °C in citrate buffer 10 mM pH = 6 for antigen retrieval (unmasking step). Lung sections were permeabilized in PBS 0.5% triton X-100, blocked with 5% FCS for 1 h at RT, and then incubated overnight with primary goat antibodies to MPO (1:40, R&D systems, Minneapolis) and rabbit anti-H3-citrulline (1:100, Abcam, UK) or mouse anti- ZBP1 (1:100, Santa Cruz, Texas). After washing, the sections were incubated with donkey anti-rabbit IgG secondary antibodies conjugated with AlexaFluor568 (1:500, Invitrogen, Massachusetts) and anti-goat IgG secondary antibodies conjugated with AlexaFluor488 (1:200, Invitrogen) or with goat anti-mouse IgG conjugated with AlexaFluor488 in 1% FCS. Following washing, lung sections were stained with DAPI (1:1000) for 10 min, washed with PBS, and mounted onto microscope slides (Fluoromount-G, Invitrogen).

Cytospin slides were fixed in 4% PFA. Cells were washed three times in TBS, incubated 15 min in TBS-0.3% Triton X-100, then washed three times in TBS, blocked in TBS-10% FCS for 45 min and incubated overnight at 4 °C with primary antibodies to MPO (1:40, R&D systems), and to histone H3-citrulline (1:100, Abcam) for NETs visualization. For PANoptosome visualization slides were incubated with primary antibodies from mouse anti-ZBP1 (1:100, Santa Cruz), rat anti-Caspase 8 (1: 100, Enzo Lifesciences, France), goat anti-ASC (1:50, Abcam), and mouse anti-RIPK3 (B-2) Alexa Fluor^® 647 (1:30, Santa Cruz Biotechnology). After washing, slides were incubated as described above for NETs formation and ZBP1 detection. For analyse of PANoptosome formation, slides were incubated with donkey anti-goat IgG secondary antibodies conjugated with AlexaFluor488 (1:200, Invitrogen) and goat anti-rat secondary antibodies Alexa Fluor 546. Slides were stained using DAPI for 10 min. For NETs and ZBP1 visualization, cells were observed using a Leica DM 6000B microscope (Leica Camera, Wetzlar, Germany), images were acquired using MetaMorph^® software, and were treated using ImageJ software. For PANoptosme visualization, cells were observed using a Zeiss LSM 980 microscope coupled with a Zeiss Airyscan 2 device (Carl Zeiss Co. Ltd., Jena, Germany). Images were acquired using Zeiss LSM Image Browser (Carl Zeiss Co. Ltd., Jena, Germany).

Bronchoalveolar lavage cells were counted and plated on a 96 well plate for extracellular staining. Different subsets of lung infiltrating cells or resident cells were detected by flow cytometry in BAL cell suspensions using a mix of the following fluorochrome-conjugated antibodies against mouse (CD45-PE-Cy7 1/200, eBiosciences, California), (Siglec F-PE-Cf59, 1/300, BD Bioscience, New Jersey), (Ly6G-FITC, 1/200), (CD3-AF700, 1/100, BD bioscience), (B220-BV711, 1/200), (Epcam-SB645, 1/100, Invitrogen, Massachusetts), (CD11c-BV605, 1/200), (F4/80-BV42, 1/200), and (Fc Block, 1/200) to avoid nonspecific binding. Early apoptosis to late apoptosis/necrosis in the detected cell populations was assessed by using Annexin V-APC and Propidium Iodide Apoptosis Detection Kit I (BD Bioscience). All staining reactions were performed at RT for 20 to 30 min. Flow cytometry analyses were performed on an LSR Fortessa X-20 flow cytometer (Becton Dickinson, New Jersey). The gating strategy was set up according to FMO control for all antibodies. Analysis and graphical output were performed using FlowJo™ software (Tree Star, Ashland, OR).

Bone marrow-derived macrophages (BMDMs) were generated by differentiating mouse bone marrow cells for 10 days in DMEM medium (Life Technologies) supplemented with 10% heat-inactivated fetal calf serum (FCS, Life Technologies, Massachusetts), 100 U/mL penicillin, 100 µg/mL streptomycin (Gibco, Australia), 2 mM glutamine (Sigma-Aldrich, Missouri), 20% horse serum, and 30% (v/v) L929 conditioned medium as a source of M-CSF. On day 10, BMDMs were seeded in 24 wells plate at 10⁶ cells/well and stimulated with cGAMP (14 µM) or diABZI (0.3 and 1 µM) in DMEM containing 0.2% heat-inactivated FCS.

Human airway epithelial cells (hAEC, Epithelix, Switzerland), passage 1, isolated from bronchial biopsies were maintained in 75 cm² flasks in the hAEC growth medium (Epithelix) and incubated at 37 °C and 5% CO2. hAEC (passage 8) were seeded in 24 wells plate at 5 × 10⁵ cells/well and stimulated after 4 h when they reached 80% confluence. Cells were stimulated with diABZI (1; 3 and 10 µM) and 2′,3′-cGAMP (14 µM) diluted in DMEM (Life Technologies, Massachusetts) supplemented with 2 mM l-glutamine, 2 mM of sodium pyruvate, 100 U/mL penicillin, 100 μg/mL streptomycin (Life Technologies), and 0.2% heat-inactivated FCS for 24 h at 37 °C and 5% CO2.

dsDNA was measured in the acellular fraction of the BALF using Quant-iT PicoGreen (Invitrogen, Massachusetts) according to the manufacturer’s protocol.

Mitochondrial and nuclear DNA were quantified as described previously [10]. Briefly, total DNA released in the BALF was purified using NucleoSpin Tissue Genomic DNA from tissue (Macherey-Nagel) and quantified by real-time PCR (10 ng/well). Primers for mouse mtDNA (mMitoF1 and mMitoR1) and mouse B2M (mB2MF1 and mB2MR1) that do not co-amplify nuclear mitochondrial insertion sequences (NumtS), mouse fragments of the mitochondrial genome present in the nuclear genome in the form of pseudogenes, were used [29]. The quantitative real-time PCR were performed in AriaMx Real-Time PCR System.

Cytotoxicity was determined by quantifying lactate dehydrogenase (LDH) released in the acellular fraction of the BALF using LDH-Glo^TM- Cytotoxicity assay kit (Promega) according to the manufacturer’s instructions.

Total mRNA was extracted using TRIzol (TRI-Reagent, Sigma-Aldrich, Germany) and reverse transcribed in cDNA with GoScript™ Reverse Transcription kit (Promega, Wisconsin). Genes mRNA expression were analyzed using GoTaq^®qPCR Master Mix (Promega). All primer sequences used were from Qiagen: Tmem173 (#QT00261590), Mb21d1 (#QT00131929), Nlrp3 (#QT00122458), Ifi204 (#QT01753535), Ddx41 (#QT00137130), Aim2 (#QT00266819), Cxcl10 (#QT00093436), Ifna2 (#QT00253092), Ifna4 (#QT01774353), and Ifnβ1 (#QT00249662). RNA expression was normalized to Rn18s expression (Qiagen, Maryland). Data were analyzed using the comparative analysis of relative expression by ^ΔΔCt methods.

Lung left lobe was removed and fixed in 4% formalin, embedded in paraffin, sectioned at 3 µm, stained with periodic acid-Schiff (PAS), and blindly scored by an anatomo-pathologist. Semi-quantitative scoring (0–5) of airway inflammation and damage including epithelial injury, peribronchial infiltrates, alveolitis, and emphysema severity was performed.

MPO, IL-6, TNFα, CXCL1, IL-10, and CXCL10 concentrations in BALF or cell culture supernatant were measured by ELISA (R&D System, Minneapolis). IFNα and IFNβ levels were measured in BALF or cell culture supernatant by multiplex immunoassay according to manufacturers’ instructions (ProcartaPlex, Life Technologies, Massachusetts). Data were acquired on Luminex equipment (MagPix, Bio-Rad, California) and analyzed using Bioplex Manager software (Bio-Rad).

Statistic analysis was performed with GraphPad Prism 8.0 software (San Diego). Statistical significance was determined by one-way ANOVA followed by Kruskal–Wallis multiple-comparisons tests or two-way ANOVA followed by Dunn post test as indicated in figure legends. P value < 0.05 was considered significant. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. All data were shown as mean ± SEM.

All datasets generated and analysed during this study are included in this published article and its Supplementary Information files. Additional data are available from the corresponding author on reasonable request.