Hypothermia protects against ventilator-induced lung injury by limiting IL-1β release and NETs formation

To better understand the mechanisms underlying ARDS and VILI, we subjected C57BL/6 mice to intratracheal instillation of LPS and MV with high-volume ventilation (HVV) or low-volume ventilation (LVV). At 30 and 150 min of MV, arterial blood gases were measured, and at the end of 180 min, the animals were euthanized and the bronchoalveolar lavage fluid (BALF) was collected (Figure 1A). The combination of LPS and HVV (LPS-HVV) caused a prominent reduction in the partial pressure of oxygen (Figure 1B) in the arterial blood (PaO₂) of mice when compared to the other controls receiving normal saline (NS) and/or LVV instillation, indicative of lung dysfunction. The requirement of both LPS and HVV for hypoxemia development was confirmed by the strong interaction (p<0.001) between these two factors found by the statistical analysis, and it occurred without significant change in carbon dioxide partial pressure (PaCO₂), pH, and base excess in the arterial blood (Figure 1—figure supplement 1A–D). Hypoxemia was accompanied by increased neutrophil migration (Figure 1C) compared with NS+HVV without significant change in the number of macrophages (Figure 1D) as determined by flow cytometry (Figure 1—figure supplement 2) and in the total cell number (Figure 1—figure supplement 1E) in BALF. To evaluate the induction of local inflammatory responses in the alveoli, we assessed the levels of several inflammatory markers in BALF. Only LPS, but not HVV, was required for the increase of IL-6, TNFα, and CXCL1. On the other hand, only HVV was sufficient to increase the IL-18 release (Figure 1—figure supplement 1F–I). Although these increased inflammatory mediators do not indicate interactions between LPS and HVV in the proposed model, LPS+HVV increased albumin levels in the BALF (Figure 1E), indicating increased vascular permeability. This injury was associated with increased IL-1β (Figure 1F), IL-1α (Figure 1G), CXCL2 (Figure 1H), plasminogen (Figure 1—figure supplement 1J), and fibrinogen (Figure 1—figure supplement 1K) in the BALF. We next sought to identify if the neutrophils were activated differently between LPS+LVV and LPS-HVV since there was no significant difference in neutrophil migration between the two models (Figure 1C), but only LPS-HVV developed hypoxemia (Figure 1B). We measured the concentration of myeloperoxidase (MPO) (Figure 1I) and neutrophil elastase (NE) (Figure 1J), two neutrophil granule enzymes which are also known to contribute to the formation of NETs (Papayannopoulos et al., 2010; Metzler et al., 2014), as well as the presence of histone-DNA (Figure 1K) and MPO-DNA (Figure 1L) complexes, validated methods for estimating general cell death and NETs (Ishii et al., 2000; Yoo et al., 2014), respectively. We found that LPS-HVV significantly increased all these NETs-associated parameters, indicating the involvement of NETs in the alveolar space with the local severe ALI, resulting in hypoxemia observed in this model.

To further investigate the role of neutrophils in this LPS-HVV-induced hypoxemia, we depleted neutrophils in vivo by treating C57BL/6 mice with antibodies against Ly6G (Carr et al., 2011) 18 hr before starting MV (Figure 2A). Neutrophil depletion prevented hypoxemia during LPS-HVV, while mice receiving isotype control antibody developed hypoxemia as before (Figure 2B). The absence of neutrophils in the alveoli was confirmed by the reduced total cell number and nearly a complete absence of neutrophils in BALF (Figure 2C and D). Moreover, in mice depleted of neutrophils, the number of macrophages was quite similar (Figure 2E). The absence of neutrophils significantly reduced the concentration of albumin and IL-6 in BALF (Figure 2F and H), but did not affect the amount of IL-1β and TNFα (Figure 2G and I). We also found higher concentrations of the chemokine CXCL2 (Figure 2J) in neutrophil-depleted mice. As expected, neutrophil depletion also resulted in lower MPO and NE concentrations (Figure 2K and L), as well as reduced detection of cell death and NETs (Figure 2M and N) in the BALF.

To investigate the functional role of NETs in the developing hypoxemia in the LPS-HVV-induced severe ALI, we used two different approaches. In the first, we used neutrophil-specific PAD4-deficient (Padi4^Δ/ΔS100a8^cre) mice and controls (Padi4^fl/fl) (Figure 3A). PADs are required for NETs formation (Rohrbach et al., 2012). Neutrophil-specific PAD4 deletion significantly inhibited the hypoxemia development compared with control mice (Figure 3B) but without affecting the numbers of neutrophils (Figure 3C) and macrophages (Figure 3—figure supplement 1A and B) in the alveoli. It also resulted in reduced levels of BALF albumin but did not change the IL-1β amount in the BALF (Figure 3D and E). PAD4 deficiency in neutrophils did not alter the levels of MPO and NE in BALF (Figure 3—figure supplement 1C and D) but did reduce cell death as measured by histone DNA (Figure 3—figure supplement 1E) and NETs formation as measured by MPO-DNA complexes (Figure 3F). We next treated C57BL/6 mice with DNase I (Figure 3G) aiming to eliminate NET structures (Czaikoski et al., 2016). This intervention also significantly prevented hypoxemia (Figure 3H) with no impact on neutrophil migration (Figure 3I) and number of macrophage (Figure 3—figure supplement 1G). It also reduced the albumin leakage (Figure 3J) without altering IL-1β levels (Figure 3K) in the alveoli. We observed that DNase I treatment resulted in similar levels of MPO and NE (Figure 3—figure supplement 1H and I) and resulted in efficiently reduced cell death as measured by histone DNA (Figure 3—figure supplement 1J) and NETs formation as measured by MPO-DNA complexes (Figure 3L).

We previously reported that the activation of NLRP3 inflammasome and IL-1β are required for LPS and MV-induced two-hit model of ALI in mice (Jones et al., 2014; Nosaka et al., 2020). To confirm the participation of IL-1 receptor type 1 (IL-1R1) downstream signaling in the severe ALI development in this two-hit model, we submitted wild-type (WT) and Il1r1^-/- mice to LPS-HVV (Figure 4A). Il1r1^-/- mice did not develop hypoxemia, as shown with preserved PaO₂ values while WT mice have reduced PaO₂ (Figure 4B). Although IL1R1 deficiency did not affect the number of neutrophils and macrophages in BALF (Figure 4C and D), it prevented the development of edema resulting from increased vascular permeability, as measured by albumin in BALF (Figure 4E). As expected, IL-1β levels (Figure 4F) were not affected by loss of IL-1R1, demonstrating that there is no feedback loop required for IL-1β production in this model. Loss of this receptor also reduced the release of IL-6 in the BALF (Figure 4G) but did not alter the levels of TNFα and CXCL2 (Figure 4H and I). Even though the lack of IL-1R1 did not affect neutrophil migration, its loss led to a reduction in BALF MPO and NE levels (Figure 4J and K), as well as cell death (histone DNA) and the NETs formation as measured by MPO-DNA complexes in the BALF (Figure 4L and M). IL-1β has been found to promote NETs formation in some studies (Mitroulis et al., 2011; Meher et al., 2018). To confirm the participation of IL-1β in NETs formation in vitro, we used neutrophils from a variety of tissues, including bone marrow neutrophils (BMN), alveolar neutrophils (AN), circulating neutrophils (CN), and peritoneal neutrophils (PN). The purity of these neutrophils was evaluated by flow cytometry (CD45.2⁺CD11b⁺ Ly6G⁺). Isolated BMN, AN, CN, and PN were approximately 88%, 99%, 86%, and 89% pure, respectively (Figure 4—figure supplement 1A). Moreover, AN nuclei look more segmented than the others (Figure 4) and present higher expression of CD11b and Ly6G (Figure 4—figure supplement 1C–E). Neutrophil quality was tested by the baseline of NETs formation (Figure 4—figure supplement 2A) with no stimulus after 4 hr of incubation. BMN and AN baseline NETs formation was up to 10%, while CN and PN presented about 20% and 60%, respectively, indicating lower stability under cell culture (Figure 4—figure supplement 2B and C). We then evaluated the NETs formation in response to IL-1β, LPS, and ionomycin (ION) by neutrophils under cell culture with concentration-response curves, as well as co-stimulation experiments (Figure 5A). With BMN and AN, we evaluated the concentration-response with single stimuli and co-stimulation of LPS or IL-1β with ION (Figure 5B–G). The concentration-response curves were also evaluated with CN and PN (Figure 4—figure supplement 2D and E). In all samples, the amounts of LPS and IL-1β used were not sufficient to induce NETs formation. Given the better stability of BMN and AN under cell culture, we used these cells for additional experiments. BMN were more resistant to ION, requiring a concentration of 10 μM to induce a 50% NETs formation response, while AN only required 3 μM (Figure 5C and F). The sensitivity difference between BMN and AN NETs formation may be related to their maturity state, as AN have highly segmented nuclei and higher CD11b and Ly6G expressions compared with BMN (Figure 4—figure supplement 1B–E). As LPS and IL-1β did not induce NETs formation by themselves, we next investigated if they could alter the response to ION. For this, neutrophils were stimulated with 10 μg/mL of LPS or 100 ng/mL of IL-1β, and 10 μM (BMN) or 3 μM (AN) of ION, and NETs formation was evaluated. We found that IL-1β, but not LPS, enhanced ION-induced NETs formation of both BMN (Figure 5C and D) and AN (Figure 5F and G). These data demonstrate that IL-1 signaling is pivotal for hypoxemia development and can modulate NETs formation in LPS+HVV ALI model.

During the course of our mouse studies, we observed that maintaining normal body temperature was important in obtaining consistent results. Several studies have proven that TH can modulate the inflammatory response controlling the release of a variety of inflammatory mediators including IL-1β (Diao et al., 2020). Thus, we decided to investigate whether hypothermia treatment might provide a feasible way to modulate the development of LPS-HVV-mediated severe ALI. We then subjected C57BL6 to the LPS+HVV model under controlled body temperature of 37±1°C or 32 ± 1°C, designated as normothermia and hypothermia, respectively (Figure 6A and B). Hypothermia provided strong protection against hypoxemia throughout the time course (Figure 6C). While similar neutrophil and macrophage counts were observed between the two groups (Figure 6D and E), hypothermia resulted in reduced levels of BALF albumin, IL-1β, IL-6, TNFα, MPO, and NE (Figure 6F–K). GSDMD-mediated pore formation impacts not only IL-1β release by macrophages but was also described to play a role in NETs formation (Sollberger et al., 2018). We therefore evaluated the soluble GSDMD concentration in the BALF by ELISA (Figure 6L) and found it significantly diminished in mice ventilated under hypothermia. We observed that hypothermia also significantly inhibited the presence of cell death and NETs formation in the BALF in mice subjected to LPS-HVV (Figure 6M and N). Finally, neither alkalosis (blood pH>7.45) nor acidosis (blood pH<7.35) was observed during hypothermia treatment (Figure 6—figure supplement 1A and B).

Since we found that hypothermia inhibited two-hit-induced acute respiratory failure with reduced IL-1β in the airways, we next evaluated the ability of bone marrow-derived macrophages (BMDMs) to release IL-1β under hypothermia. BMDMs were primed with LPS for 3 hr at 37°C, incubated at 37°C or 32°C for 30 min prior to adenosine triphosphate (ATP) or nigericin (NIG) treatment and incubated for another 30 min (Figure 7A). Macrophages incubated at 32°C released significantly less IL-1β compared with those incubated at 37° (Figure 7B and E). The mechanism by which hypothermia inhibits IL-1β release seems to be independent of caspase-1 activation, as there was no difference in caspase-1 activity assay by FLICA between 37°C or 32°C treated macrophages (Figure 7C and D), but we found that hypothermia resulted in reduced caspase-1 release in the supernatant (Figure 7E). Cleavage of GSDMD is a late limiting step for inflammasome-mediated IL-1β release because its N-terminal fragment forms pores on macrophages' plasma membrane where the intracellular cytokine crosses into the extracellular compartment (He et al., 2015). Thus, we investigated the effect of 37°C or 32°C temperature on GSDMD expression and cleavage in BMDMs by immunofluorescence and observed that at 32°C BMDMs express less GSDMD with reduced GSDMD cleavage (Figure 7F–H). Furthermore, 32°C treatment inhibited GSDMD secretion. These data corroborated our observation that while hypothermia inhibited the IL-1β release in the supernatant, there was a concomitant accumulation of unreleased mature IL-1β in the macrophages incubated at 32°C (Figure 7E). Another mechanism by which the NLRP3 inflammasome activity is regulated is by autophagy (Nosaka et al., 2020). To identify whether hypothermia regulates IL-1β release by inducing autophagy, we isolated BMDMs from Atg16l1^fl/fl or Atg16l1^fl/flLyz2^Cre+ mice and repeated the previous experiment. Hypothermic condition (32°C temperature) again significantly inhibited IL-1β release even in macrophages with impaired autophagy, suggesting that the hypothermia effect is independent of autophagy (Figure 7—figure supplement 1) and predominantly through diminished activation of GSDMD.

Finally, we investigated the impact of hypothermia on NETs formation in vitro. BMN or AN were preconditioned at 37°C or 32°C 1 hr prior to stimulation, and then stimulated for 4 hr with 10 or 3 μM of ION, respectively (Figure 7—figure supplement 2A). We observed significantly reduced NETs formation in both BMN and AN stimulated at 32°C when compared to experiments conducted at 37°C (Figure 7—figure supplement 2B and C). Taken together, these data demonstrated that hypothermia could be a therapeutic strategy to modulate both IL-1β release and NETs formation for preventing the development of severe acute respiratory failure.

C57BL/6, S100a8^Cre, Il1r1^-/-, Lyz2^Cre mice on C57BL/6 background were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Padi4^fl/fl mice were provided by Dr. Kelly Mowen (Scripps Research, San Diego, CA, USA). Atg16l1^fl/fl mice were provided by Dr. Shih (Cedars-Sinai Medical Center, Los Angeles, CA, USA). All in vivo experiments were performed in mice at 8–12 weeks of age. All animal studies presented here have been approved by the Institutional Animal Care and Use Committee of the Cedars-Sinai Medical Center. All rodent experimental procedures were conducted under approved Institutional Animal Care and Use Committee protocols.

Male mice were anesthetized with isoflurane (Piramal Healthcare, Bethlehem, PA, USA), and orotracheally intubated with an intravenous catheter (BD Insyte Autoguard, 20GA 1.00 in., Becton Dickinson Infusion Therapy Systems Inc, Sandy, UT, USA). LPS from Escherichia coli (LPS-EB Ultrapure, tlrl-3pelps, Invivogen) was diluted in sterile NS (0.9% sodium chloride, NS) to a concentration of 0.1 mg/mL, and 2 microliters per gram of body weight (µL/g) of LPS or NS were administered intrathracheally (i.t.) to mice. Two hours after LPS administration, the mice were intraperitoneally (i.p.) anesthetized with ketamine (Ketaved Ketamine HCl, NDC 50989-161-06, Vedco) and dexmedetomidine (Dexdomitor dexmedetomidine hydrochloride, 122692-5, Zoetis) mixture prepared in NS, 50 and 1.0 milligrams per kilogram of body weight (mg/kg), respectively, and orotracheally re-intubated with attention to catheter insertion length and ventilated using a small animal ventilator system (VentElite, Harvard Apparatus) for 180 min with a tidal volume of 10 milliliter per kilogram of body weight (mL/kg) in a respiratory rate (RR) of 150 breaths per minute (bpm), LVV, or 30 mL/kg of body weight in an RR of 35 breaths per minute, HVV, and 3 centimeters of water (cmH₂O) for PEEP. For hemodynamic support, 500 μL of sterile phosphate-buffered saline (PBS) were given to each mouse at the onset of MV. A complementary dose of 25 mg/kg of ketamine and 0.5 mg/kg of dexmedetomidine was administered i.p. 90 min after starting MV, or earlier as needed. Body temperature was maintained at either 37 ± 1°C or 32 ± 1°C using a heating pad (Heated Hard Pad, Hallowell EMC) and measured rectally by using a temperature probe (08D2, DeltaTrack).

The arterial blood was collected from anesthetized mice via tail artery by nicking the ventral side of the tail with a blade. Approximately 100 µL of whole blood was collected using a Heparinized Micro-Hematocrit Capillary Tube (Fisherbrand). Arterial blood gas was analyzed at 30 and 150 min after starting MV using i-STAT1 Analyzer and the i-STAT G3+Cartridges (Abbott, IL, USA), which provided the values of partial pressure of oxygen (pCO₂) and partial pressure of carbon monoxide (pCO₂), in millimeters of mercury (mmHg), pH and base excess as milliequivalent per liter (mEq/L).

BALF was obtained after 180 min of MV with 0.5 mL of cold PBS with 2 mM of EDTA by inserting a standard disposable intravenous catheter (BD Insyte Autoguard, 20GA 1.00 in., Becton Dickinson Infusion Therapy Systems Inc) into the trachea. A small portion of BALF was stained with ViaStain AOPI staining solution, prepared in Cellometer cell counting chambers, and the cells were quantified in the Cellometer Auto 2000 (Nexcelom Bioscience, Lawrence, MA, USA). The supernatant was isolated for ELISA, and the remaining cells were stained with PE anti-mouse Ly6G (50-1276-U100, 1:200, Tonbo Bioscience), FITC anti-human/mouse CD11b (35-0112-U100, 1:200, Tonbo Bioscience), APC anti-mouse F4/80 (20-4801-U100, 1:200, Tonbo Bioscience), violetFluor 450 anti-mouse CD11c (75-0114-U100, 1:200, Tonbo Bioscience), and APC/Cyanine7 anti-mouse CD45.2 (109824, 1:200, BioLegend). The percentage of neutrophils (CD11b⁺ Ly6G⁺) and macrophages (CD11c⁺ F4/80⁺) in the gate of CD45.2⁺ cells were determined by flow cytometry in the Sony SA3800 spectral cell analyzer (Sony Biotechnology) and analyzed using FlowJo software (FlowJo LLC 10.10.0, Becton Dickinson).

Enzyme-linked immunosorbent assay (ELISA) kits were used for quantifying albumin (Mouse Albumin ELISA kit, 99-134, Bethyl Laboratories), IL-1β (IL-1 beta Mouse Uncoated ELISA Kit, 88-7013-88, Invitrogen), IL-1α (ELISA MAX Deluxe Set Mouse IL-1α, 433404, BioLegend), CXCL-2 (Mouse CXCL2/MIP-2 DuoSet ELISA, DY452, R&D Systems), MPO (Mouse Myeloperoxidase DuoSet ELISA, DY3667, R&D Systems), NE (Mouse Neutrophil Elastase/ELA2 DuoSet ELISA, DY4517-05, R&D Systems) cell death (Cell Death Detection ELISA, 11544675001, Roche Life Sciences), IL-6 (Mouse IL-6 ELISA Set, 555240, BD Biosciences), TNFα (TNF alpha Mouse Uncoated ELISA Kit, 88-7324-88, Invitrogen) CXCL-1 (Mouse CXCL1/KC DuoSet ELISA, DY453, R&D Systems), IL-18 (Mouse IL-18 Matched ELISA Antibody Pair Set, SEK50073, Sino Biological) plasminogen (Mouse PLG/Plasmin/Plasminogen ELISA Kit, LS-F10445, Life Span Biosciences), fibrinogen (Mouse Fibrinogen ELISA Kit, LS-F10440, Life Span Biosciences) and GSDMD (Mouse GSDMD ELISA Kit, ab233627, Abcam). In addition, we developed an ELISA based on MPO associated with DNA as previously described (Caudrillier et al., 2012) with some modifications. For the capture antibody, 800 ng/mL of anti-MPO capture mAb (Mouse Myeloperoxidase DuoSet ELISA, DY3667, R&D Systems) was coated onto 96-well plates overnight at room temperature (RT). After blocking and washing, 25 μL of BALF was added to the wells with 75 μL incubation buffer (1% BSA/PBS) and incubated for 2 hr at RT. After washing, incubation buffer containing a peroxidase-labeled anti-DNA mAb (Cell Death Detection ELISA, 11544675001, Roche Life Sciences, dilution 1:10). The plate was incubated for 1 hr at RT. After washing, the peroxidase substrate (TMB) was added, and after 15 min at RT in the dark, the reaction was stopped by adding H₂SO₄ solution and the absorbance at 450 nm wavelength.

Neutrophils were isolated from 8- to 12-week-old male WT C57BL/6 mice. The purification was made using two different density gradient protocols made with Percoll (Percoll, GE Healthcare, GE17-0891-01) prepared in Hanks' balanced salt solution (HBSS). The Percoll gradient 1 (PG1) consists of a three-layer gradient made with 75%, 57%, and 52% of Percoll, and the Percoll gradient 2 (PG2) consists of two layers, 68% and 52% of Percoll. The gradients containing 1 mL of cell suspension on the top were centrifuged for 30 min, 1500 × g at RT, with soft acceleration and deceleration. Neutrophils were located above the 75% or 68% Percoll layers for PG1 and PG2, respectively. To obtain BMNs, mice were euthanized, and the rear leg bones were removed and flushed with HBSS with 2 mM of EDTA (HBSS-EDTA). The cells were centrifuged for 5 min, 450 × g at RT, and the red blood cells (RBCs) were lysed by adding 10 mL of 0.2% NaCl, gently mixing for 30 s. The salt balance was recovered by adding 5 mL of 2.3% NaCl. The cells were centrifuged for 5 min, 450 × g at RT, and resuspended in 1 mL of HBSS-EDTA. The cell suspension was carefully added to the PG1, and the BMN isolation was performed as above. For ANs, as described in the two-hit ALI model, mice were anesthetized with isoflurane, orotracheally intubated with intravenous catheter, and LPS 0.2 mg/kg was i.t. instilled. Three days later, BALF was obtained with 5 mL of HBSS-EDTA divided into five lavages with 1 mL of HBSS-EDTA. The cells were centrifuged for 5 min, 450 × g at RT, and resuspended at 1 mL of HBSS-EDTA. The cell suspension was carefully added to PG2, and the AN were isolated. CNs were obtained from the blood collected 6 hr after LPS 0.2 mg/kg i.p. injection. The total whole blood was collected by retro-orbital bleeding, in a 5 mL tube containing 2.5 mL of EDTA (2 mg/mL in PBS). The RBCs were lysed by adding the blood (maximum 5 mL, 2.5 mL of blood+2.5 mL of EDTA) to 30 mL of 0.2% NaCl, gently mixing for 30 s. The salt balance was recovered by adding 15 mL of 2.3% NaCl. The lysis step was repeated one time in case remaining RBCs were observed in the pellet formed after 5 min of centrifugation, 450 × g at RT. After total elimination of RBCs, the pellet was resuspended in 1 mL of HBSS-EDTA. The cell suspension was added to the PG2, and the CN were purified. For peritoneal neutrophil isolation, mice received i.p. injections with 1.5 mL of 3% thioglycolate, and after 16 hr, the animals were euthanized, and the peritoneal cavity was washed with 5 mL of HBSS-EDTA. The cells were centrifuged for 5 min, 450 × g at RT, and resuspended in 1 mL of HBSS-EDTA. The cell suspension was added to PG2, and the neutrophils were isolated. The cells were stained with ViaStain AOPI staining solution, prepared in Cellometer cell counting chambers, quantified in the Cellometer Auto 2000, and resuspended in RPMI/neutrophils (RPMI 1640, 10-040-CV, Corning supplemented with 1% of Penicillin-Streptomycin Solution, 30-0002 CL, Corning; 2% of Fetal Bovine Serum, FB-02, Omega Scientific; and 1% of MEM Non-essential Amino Acid Solution, Sigma-Aldrich, M-7145). Neutrophil purity was evaluated by flow cytometry by staining the cells with PE anti-mouse Ly6G (50-1276-U100, 1:200, Tonbo Bioscience), FITC anti-human/mouse CD11b (35-0112-U100, 1:200, Tonbo Bioscience), and APC/Cyanine7 anti-mouse CD45.2 (109824, 1:200, BioLegend). The percentage of neutrophils (CD11b⁺ Ly6G⁺) was determined in the gate of CD45.2⁺ cells, analyzed using FlowJo software (FlowJo LLC 10.10.0, Becton Dickinson). The cells were also observed using cytospin (Cytospin 4, Thermo Scientific) prepared slides stained with a rapid staining of blood smear (Hemacolor Rapid staining of blood smear, 111661, Sigma-Aldrich).

2.0×10⁴ Neutrophils were labeled with 2 µM Hoechst 33342 (Immunochemistry Technology, 639) and seeded in flat-bottom 96-well plates. We made three sets of experiments to investigate NETs formation under different conditions, using RPMI/neutrophils to prepare all the treatments. In the first set, after resting the cells for 1 hr at 37°C, the cells were treated with the following: LPS (30–100,000 ng/mL); IL-1β (recombinant mouse IL-1β protein, ab259421, Abcam) (0.3–1000 ng/mL); and ION (ionomycin calcium salt 10634, Sigma-Aldrich) (10–100,000 nM). In the second set, combined stimuli were given to BMN and AN. The cells were first rested at 37°C for 30 min, LPS (10 µg/mL) or IL-1β (100 nM) was added, and after 30 min, the cells were treated with the concentration of ION sufficient to induce about 50% of NETs formation, which is 10 µM for BMN and 3 µM for AN. In the third set, BMN and AN were placed at 37°C or 32°C for 1 hr, and then treated with ION 10 µM for BMN and 3 µM for AN. In all of them, the cells were incubated for 4 hr at 37°C or 32°C, maintaining the initial resting temperature setting in each experiment, to allow NETs induction. The cells were then stained with 5 µM of SYTOX Green (SYTOX Green Nucleic Acid Stain, 57020, Invitrogen) prepared in sterile PBS and centrifuged for 5 min, 500 × g RT. The images were obtained on a Keyence BZ-9000 (Keyence Corporation of America) microscope at ×20 magnification. The NETs forming neutrophils percentage was evaluated manually by investigators blinded to sample identity. Quantification was based on the total cell number in the field, determined by the sum of Hoechst (white) and SYTOX Green (green) positive cells, as counted using the Keyence BioAnalyzer software (Keyence Corporation of America).

Bone marrow was obtained from 8- to 12-week-old male WT C57BL/6, Atg16l1^fl/fl or Atg16l1^Δ/Δ Lyz2^Cre mice, and BMDMs were differentiated in RPMI/macrophages (RPMI 1640, 10-041-CV, Corning; supplemented with 1% of Penicillin-Streptomycin Solution, 30-0002-CL, Corning; 50 μM of 2-mercaptoethanol; and 10% of Fetal Bovine Serum, FB-02, Omega Scientific) containing 15% of L929 cell conditioned medium (LCM) for 7 days, supplementing the medium with extra 15% of LCM every 3 days. Cells were washed with PBS, and non-adherent cells were removed; adherent cells were then collected and seeded in a 96-well plate 1 day before stimulation. BMDMs were primed with LPS 1 μg/mL for 3 hr, followed by ATP 5 mM (adenosine 5′-triphosphate disodium salt, A2383-1G, Sigma-Aldrich) or NIG 10 μM (Nigericin sodium salt, BML-CA421-0005, Enzo Life Sciences) stimulation for 30 min at either 37°C or 32°C, and the supernatants were collected for ELISA measurements and western blotting, and the cells were lysed for western blotting. For evaluating caspase-1 activity and GSDMD expression by immunofluorescence, LPS-primed BMDMs were stimulated with 5 mM ATP for 15 min or 10 μM NIG for 30 min. For caspase-1 activity, a commercial kit (FAM-FLICA Caspase-1 Assay Kit, 98, Immunochemistry Technology) was used. Immunoblots were performed using antibodies against IL-1β (anti-IL-1 beta antibody, 2 μg/mL, ab9722; Abcam), caspase-1 (recombinant anti-pro caspase-1+p10+p12 antibody, 1:1000, ab179515; Abcam), and GSDMD (recombinant anti-GSDMD antibody, 1:1000, ab209845, Abcam). The GSDMD antibody used in the immunoblot was also used for immunofluorescence staining, which was mounted with fluorescence mounting medium with DAPI (Mounting Medium With DAPI, ab104139, Abcam). The FAM-FLICA caspase-1 images and the other immunofluorescence images were obtained on a Keyence BZ-9000 (Keyence Corporation of America) microscope at ×20 magnification.

All data were analyzed using Prism 9 (GraphPad Software Inc, La Jolla, CA, USA). Normality within each group was assessed using the Shapiro-Wilk and Kolmogorov-Smirnov tests. For comparisons between two groups, the Mann-Whitney U test was used for non-normally distributed data, while the unpaired Student's t-test was applied to data that met the assumption of normality. One-way ANOVA (with a single independent factor), two-way ANOVA (with two independent factors), and three-way ANOVA (with three independent factors) were performed for comparisons involving more than two groups, followed by Tukey's post hoc test. A p-value of less than 0.05 was considered statistically significant.