Influenza virus

The IAV H3N2 strain HKx31, a high-yielding reassortment of A/Puerto Rico/8/34 A/PR/8 (PR8; H1N1) with the surfaces of A/Hong Kong/1/68 (H3N2), was used in both in vivo and in vitro models of infection. Human influenza strains A/Brazil/11/78 (BR; H1N1), A/Solomon Island/3/2006 (SI; H1N1), A/Tasmania/2004/2009 (Tas; pandemic H1N1), and A/Perth/16/2009 (Perth; H3N2) were used in in vitro experiments. IAVs were propagated in 10-day old embryonated chicken eggs, and viral stocks were stored at -80 °C. Viral titers were quantified on Madin-Darby Canine Kidney (MDCK) cells using a standard plaque assay.

In vitro influenza virus infection

Normal immortalized human bronchial epithelial cells (HBEC3-KT; ATCC, Virginia, USA) were cultured in complete bronchial epithelial growth medium (BEGM; Lonza, Basel, Switzerland) and bovine collagen-coated flasks (Gibco, Thermo Fisher Scientific, Waltham, USA). HBEC3-KT cells in hydrocortisone-free BEGM medium were plated at 1 × 10⁵ cells/well into collagen-coated 12-well plates. Cells were incubated overnight and then infected with various strains of IAV at a multiplicity of infection (MOI) of 3 for 1 h. Supernatants were then discarded and replaced with complete BEGM. At the indicated time points, cell supernatants were harvested, and cell lysates were collected using RIPA buffer (50 mM Tris-HCl pH 8, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA) pH 8, 1% v/v Igepal, 0.5% w/v sodium deoxycholate, 0.1% v/v SDS, 10 mM sodium fluoride, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and cOmplete Protease Inhibitor Cocktail (Roche, Basel, Switzerland)) and stored at −80 °C. Cell supernatants were assayed for levels of lactate dehydrogenase (LDH) using a CytoTox 96 Non-radioactive Cytotoxicity Assay (Promega, Madison, USA), according to the manufacturer’s instructions. Levels of IL-1β and IL-18 in cell supernatants were determined by ELISA (R&D Systems, Minneapolis, USA), according to the manufacturer’s instructions.

Examination of protein expression by immunoblot

Supernatants from HBEC3-KT were concentrated using StrataClean Resin (Agilent, Santa Clara, USA) in loading buffer (250 mM Tris pH 6.8, 12% v/v SDS, 50% glycerol, 0.025% bromophenol blue) containing 5 mM dithiothreitol. HBEC3-KT cell lysates were sonicated and centrifuged. Protein levels were quantified using the Pierce^TM BCA Protein Assay Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Proteins were separated on a 4-12% NuPAGE^TM Bis-Tris Protein Gel in 1X NuPAGE^TM MES running buffer (Thermo Fisher Scientific) and transferred onto Immobilon®-FL PVDF membrane (Merck Millipore, Massachusetts, USA). Membranes were blocked with 5% bovine serum albumin (BSA) in either TBS 0.1% Tween-20 (TBST), 5% skim milk in TBST, or Intercept (PBS) blocking buffer (LI-COR Biotech, Nebraska, USA) for 30 min and incubated overnight at 4 °C with primary antibodies: GSDME (Abcam, Cambridge, UK; EPR19859), GSDMD (Cell Signalling Technology; 69469), caspase-1 (Cell Signalling Technology; 3866 or 83383), caspase-3 (Cell Signalling Technology; 9662 or 9664), GAPDH (Cell Signalling Technology; 97166), and α-tubulin (Abcam; YL1/2). The following day, membranes were washed twice in TBST before being incubated with horseradish peroxidase (HRP), Alexa 680 (Thermo Fisher), or Dylight 800 (Rockland Immunochemicals, Pennsylvania, USA)-conjugated secondary antibodies for 2 h at room temperature. Membranes were washed with TBST before being visualized on the Bio-Rad ChemiDoc MP Imaging System (Bio-Rad, Hercules, USA) or Odyssey® LI-COR Imaging System (LI-COR Biotech) via chemiluminescence or fluorescence. All original blots are included in the Supplementary Information.

siRNA silencing of GSDMD and GSDME in vitro

In the indicated experiments, siRNA knockdown of GSDMD and GSDME in HBEC3-KT cells was performed individually in triplicate wells. Briefly, siRNA specific to GSDMD (two different target sequences combined: #1, sense 5′-AGCUGGUUAUUGACUCUGAtt-3′, antisense 5′-CUGAGUCAAUAACCAGCUgg-3′; #2, sense 5′-GGAACUCGCUAUCCCUGUUtt-3′, antisense 5′-AACAGGGAUAGCGAGUUCCgg-3′; Silence Select, Ambion, Life Technologies), GSDME (two different target sequences combined: #1, sense 5′-GAGAGAACAAUAAAUCUGAtt-3′, antisense 5′-UCAGAUUUAUUGUUCUCUCgg-3′; #2, sense 5′-GAAAAGAUACAGAAAAGGUUUtt-3′, antisense 5′-AAACCUUUCUGUAUCUUUCag-3′; Ambion, Life Technologies) or non-targeting control siRNA (Silence Select Negative Control #1, Ambion) was transfected using Lipofectamine™ RNAiMAX Transfection Reagent (Thermo Fisher; 7.5 pM siRNA complexed with 0.75 μL RNAiMax per well). At 48 h following siRNA transfection, triplicate wells were infected with Brazil/78 or HKx31 IAV (MOI 3), as described above. At 24 h post-IAV infection, cell lysates were collected using RIPA buffer and pooled from triplicate wells. Suppression of GSDMD and/or GSDME expression was confirmed by immunoblot, as described above. Cell supernatants from each triplicate well were assayed for levels of lactate dehydrogenase (LDH) and IL-1β, as described above. Levels of infectious virus in cell supernatants from each triplicate well were determined using a standard plaque assay on MDCK cells. Data was normalized to NT control, where indicated, to control for inherent experimental variation.

Influenza virus infection of mice

6-12-week-old male and female Gsdme^−/− mice and wildtype littermates on the C57BL/6N background were maintained in the Specific Pathogen Free Physical Containment Level 2 (PC2) Animal Research Facility at the Monash Medical Centre (Clayton, Victoria, Australia). Gsdme^−/− mice were kindly supplied by VM Dixit (Genentech, South San Francisco, USA). All procedures were approved by the Hudson Animal Ethics, and experimental procedures were carried out in accordance with approved guidelines.

For in vivo infection studies, cages of mice were randomly allocated to groups. Mice were lightly anesthetized with isoflurane and intranasally inoculated with 10⁴ plaque-forming units (PFU) of HKx31 IAV in 50 µl of PBS to induce severe influenza disease, as previously described [16, 18, 19]. Mice were weighed daily and assessed for clinical signs of disease on a scale of 0–3 (0 = no visible signs; 1 = slight ruffling of fur; 2 = ruffled fur, reduced mobility; and 3 = ruffled fur, reduced mobility, and rapid breathing). Animals that lost 20% of their original body weight and/or displayed severe clinical signs of disease (clinical score of 3) were immediately euthanized. In the survival cohort, mice were monitored for up to 10 days.

In another cohort, at the indicated time points, mice were sacrificed via intraperitoneal injection of sodium pentobarbital. Blood was immediately harvested via cardiac puncture, and serum was collected using Serum Gel microtubes (Sarstedt, Nümbrecht, Germany). Bronchoalveolar lavage (BAL) was obtained by flushing the lungs three times with 1 ml of cold PBS and was placed on ice. BAL fluid was collected following centrifugation. BAL fluid and serum were stored at −80 °C until further analysis was performed. Lung tissues were excised following BAL collection and snap-frozen in liquid nitrogen before titration on MDCK cells to measure levels of infectious virus using a standard plaque assay. A final cohort of mice was sacrificed via intraperitoneal injection of sodium pentobarbital, and lung tissues were immediately inflated and fixed with 10% neutral buffered formalin (NBF).

Cytokine and chemokine analysis of mouse BAL fluids and serum

Levels of IL-6, CCL2, IFNγ, IL-10, IL-12p70, and TNF proteins were determined in BAL and serum by cytokine bead array (CBA) using the mouse inflammation kit (BD Biosciences, San Jose, USA). Mouse IL-1β, IL-18, IL-1α, CXCL1, and CXCL2 were quantified by ELISA (R&D Systems). Levels of IFNβ and IFNα in BAL were measured by ELISA, as previously described [20].

Flow cytometry analysis of mouse BAL cells

Cells in the BAL were separated by centrifugation at 1600 rpm for 5 min at 4 °C. BAL cells were treated with a red blood cell lysis buffer (Sigma Aldrich, St. Louis, USA) at room temperature. After 5 min, FACS buffer (PBS containing 2% v/v fetal calf serum (FCS, Thermo Fisher Scientific) and 2 mM EDTA (Promega)) was added to quench the reaction. BAL cells were stained with fluorescently labeled antibodies in the FACS buffer, including Fc receptor-blocking monoclonal antibodies against CD16/CD32 (clone 93, Thermo Fisher Scientific) to prevent non-specific antibody binding at 4 °C for 20 min. BAL cells were also resuspended with a standard amount of blank calibration particles (ProSciTech, Kirwan, Australia) to determine cell count. The monoclonal antibodies used to stain the various infiltrated immune BAL cells are Siglec-F (clone E50-2440, BD Biosciences), NK1.1 (clone PK136, BioLegend, San Diego, USA), CD3ε (clone 145-2C11), CD11c (clone HL3, BD Biosciences), CD64 (clone X54-5/7.1 BioLegend), Ly6C (clone AL-21, BD Biosciences), Ly6G (clone 1A8, BD Biosciences), I-A^b (clone AF6-120.1, BD Biosciences), and the Zombie Aqua or Zombie NIR viability dye (BioLegend). Total live cells (Zombie Aqua^– or NIR viability dye^–), neutrophils (Ly6G⁺ Ly6C^int), natural killer (NK) cells (NK1.1⁺ CD3^–), T cells (NK1.1^– CD3⁺), inflammatory monocytes/macrophages (IM; Ly6G^– Ly6C^hi), alveolar macrophages (AM; Ly6C^int CD11c⁺ Siglec-F⁺), and dendritic cells (DC; CD11c⁺ I-A^b+) were quantified by flow cytometry using an Aurora flow cytometer (Cytek Biosciences, Fremont, USA) and analyzed using FlowJo^TM 10 analysis software (BD Biosciences). Cells were enumerated using a standard amount of blank calibration particles (ProSciTech) as determined using a hemocytometer.

For flow cytometric analysis of cell death, BAL cells were incubated with AF647-conjugated Annexin V (BioLegend) and 5 µg/ml propidium iodide (PI; Thermo Fisher Scientific) resuspended in binding buffer (10 mM HEPES pH 7.4, 150 mM NaCl, and 2.5 mM CaCl₂ buffer) and analyzed by flow cytometry using an Aurora flow cytometer (Cytek Biosciences) and FlowJo^TM 10 analysis software (BD Biosciences).

Immunofluorescence and immunohistochemical staining of lung tissue sections

NBF-inflated and fixed lung tissues were submerged in NBF for 48 h before being embedded in paraffin wax. Longitudinal lung tissue sections (4 µm) mounted on histology slides were then dewaxed and rehydrated. Sections were then microwaved for 15 min (high power) and 5 min (low power) in EDTA buffer (1 mM EDTA, pH 8) for heat-induced antigen retrieval. Slides were then blocked with CAS-Block Histochemical Reagent (Thermo Fisher Scientific) for 1 h before incubation with a cocktail of primary antibodies: cleaved GSDME (38821S, Cell Signalling Technology), E-cadherin (AF648SP, R&D Systems), and CD45⁺ (clone 30-F11, BD Biosciences) at 4 °C overnight. Slides were then washed with PBS containing 0.01% Tween 20 and incubated with anti-goat (Abcam), anti-rabbit, or anti-rat (Thermo Fisher Scientific) secondary antibodies. Slides were then washed, stained with Hoechst 33342 nuclear staining (Thermo Fisher Scientific), and mounted using Fluorescence Mounting Medium (Agilent). Lung sections were examined and imaged at 20x magnification using a Nikon A1R confocal microscope (Nikon, Tokyo, Japan). Five random fields of view (FOV) were analyzed per mouse using HALO software (Indica Labs, New Mexico, USA). Four fluorescent colors were used in the image analysis. Individual cells were identified using the nuclear Hoechst signal, and the expression of cleaved GSDME, E-cadherin, and CD45 was analyzed. Parameter settings in the HighPlex FL algorithm were verified by manual visual inspection of positive cells from randomly selected images. The number and percentage of cells that were positive for cleaved GSDME and double positive for E-cadherin/cleaved GSDME and CD45/cleaved GSDME were quantified per FOV.

Terminal deoxynucleotidyl transferase-mediated dUDP nick-end labeling (TUNEL) assay was performed on lung tissue sections using the ApopTag Peroxidase In Situ Apoptosis Detection Kit (Merck Millipore), according to the manufacturer’s instructions. Lung sections were counterstained with hematoxylin and cover-slipped with D.P.X. mounting medium (Sigma Aldrich). Lung sections were viewed using an Olympus DP74 microscope (Olympus, Tokyo, Japan), and five random single-plane images per mouse section were imaged with an Olympus DP74 color camera using Olympus cellSens Dimension software at 10x magnification. ImageJ software (National Institutes of Health, USA) was used to quantify TUNEL labeling intensity (percentage positive pixel intensity per FOV), color deconvolution was performed, and a threshold was set on 3,3′-diaminobenzidine (DAB) intensity, which was applied to all lung sections.

Assessment of lung damage and pathology

Longitudinal lung tissue sections (4 µm) were prepared and stained with hematoxylin and eosin (H&E) by the Monash University Histology Platform. Tissues were graded for alveolitis and peribronchial inflammation on a subjective scale of 0 to 5 (0 = no inflammation, 1 = very mild, 2 = mild, 3 = moderate, 4 = marked, and 5 = severe inflammation), as previously performed [16, 19]. Sections were also scored for features of epithelial damage such as the presence of debris in the airspace, epithelial denudation, and thickening of the epithelial wall (0 = no obvious damage, 1 = mild, 2 = moderate, 3 = marked, and 4 = severe). Sections were blinded and randomized, and samples corresponding to the least severe and most severe were assigned scores of 0 and 4/5, respectively, and five random fields per mouse were graded by three independent researchers. Lung sections were viewed on an Olympus BX60 microscope and photographed at 10x magnification with an Olympus DP74 color camera using Olympus cellSens Dimension software.

As further assessment of lung damage, levels of adenosine triphosphate (ATP) and total protein levels in BAL fluid were determined, as previously described [19], using the Cell Titer Glo 2.0 Cell Viability Assay (Promega) and Pierce^TM BCA Protein Assay Kit (Thermo Fisher Scientific), respectively.

Data and statistical analysis

Data is presented as mean ± standard deviation (SD). Data from each individual animal are shown. Sample sizes were used based on previous extensive experience in the laboratory within similar studies. The investigator was blinded to the group allocation for the assessment of histology but was not blinded for other experiments. Data were tested for normality and analyzed by GraphPad Prism Version 9 software (Graphstats Technologies, Bangalore, India). A Student’s t test (two-tailed unpaired) was used when comparing two values. When comparing three or more sets of values, a one-way or two-way analysis of variance (ANOVA) was used with either Tukey’s or Dunnett’s multiple comparisons post-hoc test. Survival proportions were compared using the Mantel–Cox log-rank test. A P value < 0.05 was considered statistically significant.