Gasdermin D and Gasdermin E Are Dispensable for Silica-Mediated IL-1β Secretion from Mouse Macrophages Open Access

Silicosis is an occupational lung disease caused by the inhalation of silica dust and is characterized by pulmonary fibrosis and persistent inflammation, ultimately leading to respiratory failure (1). As of now, there is no effective therapy for silicosis. Although the immunological processes that drive silicosis pathology remain incompletely characterized, the NLR family pyrin domain containing 3 (NLRP3) inflammasome and IL-1β have been suggested to play a role in the inflammatory response to silica (2–5).

Activation of the NLRP3 inflammasome activation is a two-step process requiring both a priming step and an activation signal (6). A diverse array of molecules, including ATP, nigericin, silica, alum, and monosodium urate, can serve as the activation signal and are thought to activate a common pathway culminating in NLRP3 inflammasome activation (6). NLRP3 activation results in the assembly of NLRP3, the adaptor molecule apoptosis-associated speck-like protein containing CARD (ASC), and the cysteine protease caspase-1 into a multiprotein inflammasome complex. Procaspase-1 undergoes autocatalysis to form active caspase-1, which then cleaves pro–IL-1β and pro–IL-18 into their mature secreted forms. Caspase-1 also cleaves gasdermin D (GSDMD), with the cleaved N-terminal domain of GSDMD being recruited to the plasma membrane, triggering the formation of a GSDMD pore (7–9). The formation of GSDMD pores can result in the release of cytokines and the initiation of pyroptotic programmed cell death.

IL-1β lacks a signal peptide and hence does not undergo conventional secretion from the cell via the endoplasmic reticulum–Golgi (10). The release of small proteins such as IL-1β, IL-18, Rho-GTPases, and galectin-1 is thought to occur via direct passage through the GSDMD transmembrane pore or secondarily after plasma membrane rupture during pyroptotic cell death (11–14). However, recent studies have found that GSDMD is dispensable for the release of IL-1β from macrophages in response to the NLRP3 agonists monosodium urate and alum, suggesting that additional pathways for IL-1β secretion exist (15). Monteleone et al. found that cleavage of pro–IL-1β causes the relocation of mature IL-1β to phosphatidylinositol 4,5-bisphosphate (PIP2)-enriched plasma membrane ruffles, resulting in GSDMD-independent secretion of IL-1β (16).

In this study, we demonstrate that GSDMD is dispensable for the secretion of IL-1β from macrophages in response to silica. We further found that gasdermin E (GSDME) did not compensate for the absence of GSDMD in mediating IL-1β secretion in response to silica in vitro. Finally, GSDMD and GSDME were also not necessary for silica-induced acute lung injury in vivo.

Wild-type (WT) C57BL/6NCI mice were purchased from Charles River Laboratories and used as WT control animals unless otherwise stated. The generation of Asc^−/− (17), Il1r1^−/− (18), Gsdmd^−/− (8), and Gsdme^−/− (19) mice has been described previously. Both male and female mice 6–12 wk of age were used; however, mice were sex and age matched for individual experiments. The institutional animal care and use committee at Cedars-Sinai Medical Center approved all protocols used in this study.

Oropharyngeal aspiration of 50 µl of 10 mg/ml silica (Min-U-Sil-5; Pennsylvania Glass Sand) suspension in PBS was performed on isoflurane-anesthetized mice as described (20). Eight days after silica challenge, bronchoalveolar lavage (BAL) was performed by delivering 1 ml cold PBS into the airway via a tracheal cannula and gently aspirating the fluid. The BAL was repeated three times.

Cells from bronchoalveolar lavage fluid (BALF) were blocked with anti-mouse CD16/32 (clone 2.4G2; Tonbo Biosciences) in FACS staining buffer (Rockland) for 20 min at 4°C. Following blocking, cells were stained in FACS staining buffer with fluorochrome-conjugated CD11b (clone M1/70; eBioscience) and Ly6G (clone 1A8; eBioscience) Abs in the dark for 30 min at 4°C. Cells were then resuspended in PBS and analyzed on a Sony SA3800 cell analyzer.

Bone marrow–derived macrophages (BMDMs) were generated as described previously (21). BMDMs were either left unstimulated or primed with 500 ng/ml LPS (InvivoGen) for 3 h and then exposed to the indicated concentrations of nigericin (MilliporeSigma), ATP (Sigma-Aldrich), silica, and alum (Thermo Scientific). At 2, 4, 8, and 24 h after challenge, supernatants were collected, and cytokines were quantified using mouse ELISA kits for IL-1β (R&D Systems), TNFα (Thermo Fisher Scientific), and IL-18 (MBL International) according to the manufacturers’ instructions. Lactate dehydrogenase (LDH) release into supernatant was measured using the CytoTox 96 Non-Radioactive Cytotoxicity Assay kit (Promega) according to the manufacturer’s instructions.

Lysates were prepared in radioimmunoprecipitation lysis buffer (Cell Signaling Technology) with a protease inhibitor mixture that included 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride, aprotinin, bestatin, E-64, leupeptin, and pepstatin A in DMSO (Thermo Fisher Scientific). Proteins were separated on a NuPAGE gel (Invitrogen) and transferred to a polyvinylidene difluoride membrane using the XCell II blotting system (Invitrogen). Membranes were blocked with 5% BSA and incubated with primary Ab overnight at 4°C. Primary Abs used were as follows: IL-1β (AF-401-NA; R&D Systems), caspase-1 (AG-20B-0042-C100; AdipoGen Life Sciences), GSDMD (46451S; Cell Signaling Technology), and β-actin (4970S; Cell Signaling Technology). Following washing, membranes were incubated with HRP-tagged anti-mouse IgG (7076P2; Cell Signaling Technology), anti-rabbit IgG (7074S; Cell Signaling Technology), or anti-goat IgG (705-036-147; Jackson ImmunoResearch). Membranes were developed using SuperSignal West Pico or Femto substrate (Thermo Fisher Scientific).

Data were graphed and the indicated statistical tests performed using GraphPad Prism software. If not otherwise stated, analysis was performed with multiple t tests (Mann–Whitney U test) with correction for multiple comparisons using the Holm–Sidak method.

To determine if GSDMD contributes to pulmonary inflammation induced by silica, we administered silica by oropharyngeal aspiration to WT and Gsdmd^−/− mice. Eight days after challenge, the total number of neutrophils in the BALF was determined. The total number of neutrophils in the BALF of Gsdmd^−/− mice was similar to that observed in WT mice (Fig. 1A), suggesting that GSDMD is not required for the acute inflammatory response to silica in vivo. In contrast, mice deficient in the inflammasome adaptor molecule ASC or IL-1R1 had significantly fewer neutrophils in the BALF than WT mice following silica challenge (Fig. 1B, 1C), consistent with a requirement for the inflammasome and IL-1R1 signaling for the acute inflammatory response.

Consistent with previous studies (2, 4), we found that silica treatment of LPS-primed WT BMDMs resulted in the rapid activation of caspase-1, as indicated by the processing of procaspase-1 into the p20 fragment (Fig. 2A). Pro–IL-1β was also processed, and the mature IL-1β p17 fragment was secreted (Fig. 2A). Interestingly, silica challenge of LPS-primed BMDMs from Gsdmd^−/− mice did not ablate the secretion of mature IL-1β into the supernatant, as determined by both immunoblot analysis and ELISA (Fig. 2A, 2B). Furthermore, the absence of GSDMD did not stop silica-induced IL-18 secretion or prevent cell death as measured by LDH release (Fig. 2C, 2D). At 24 h after stimulation, we did observe diminished IL-1β and IL-18 secretion from Gsdmd^−/− BMDMs compared with WT (Fig. 2B, 2C). In contrast to IL-1β p17, secretion of caspase-1 p20 from silica-challenged BMDMs required GSDMD (Fig. 2A). The secretion of IL-1β from silica-treated LPS-primed BMDMs was fully dependent on the NLRP3 inflammasome because BMDMs deficient in either NLRP3 or ASC failed to secrete IL-1β into the supernatant (Fig. 2E, 2F). Secretion of TNFα was not affected by the absence of NLRP3, ASC, or GSDMD (Supplemental Fig. 1A–C). The cell death in BMDMs observed following silica challenge was also independent of NLRP3 and ASC (Supplemental Fig. 1D, 1E).

Taken together, these findings suggest that although silica can induce a robust activation of the NLRP3 inflammasome resulting in the secretion of mature IL-1β, GSDMD is dispensable for the silica-induced secretion of IL-1β from macrophages in vitro.

We next asked if other NLRP3-inflammasome activators required GSDMD for secretion of IL-1β from macrophages in vitro. Consistent with previous studies (9, 12), we found that the early secretion of IL-1β from LPS-primed BMDMs in response to ATP or nigericin was dependent on GSDMD (Fig. 3A–3C). However, following a 20-min pulse with ATP or nigericin, IL-1β secretion from LPS-primed BMDMs was not lost, despite the absence of GSDMD (Fig. 4A, 4B). Consistent with Rashidi et al. (15), we also found that cells lacking GSDMD had intact IL-1β secretion in response to the particulate NLRP3 agonist alum (Fig. 4C). We did observe diminished IL-1β secretion from Gsdmd^−/− BMDMs compared with WT BMDMs at 2 h after stimulation with ATP and nigericin and at 2 and 4 h after stimulation with alum (Fig. 4A–4C), suggesting a role for GSDMD in early IL-1β secretion. IL-1β secretion from Gsdmd^−/− BMDMs at later time points after stimulation, 4, 8, and 24 h for ATP and 24 h for nigericin and alum, were elevated compared with WT BMDMs; however, the reason for this is unclear (Fig. 4A–4C). TNF-α secretion was not affected by GSDMD deficiency (Supplemental Fig. 2A–2C). Negligible cytotoxicity was observed when macrophages were challenged with a short pulse of ATP or nigericin (Fig. 4A, 4B). The absence of GSDMD did not prevent alum-mediated cell death at 24 h as measured by LDH release (Fig. 4C). These data suggest that although GSDMD contributes to IL-1β secretion from BMDMs at early time points, there appears to be a GSDMD-independent mechanism through which IL-1β secretion can occur.

The gasdermin family consists of gasdermin A, gasdermin B, gasdermin C, GSDMD, GSDME, and pejvakin. Once activated and cleaved from their full-length proteins, their N-terminal domains share the ability to form membrane pores (22). Because macrophages only express Gsdmd and Gsdme (23), we assessed if GSDME played a role in silica-induced IL-1β secretion from BMDMs. LPS-primed BMDMs from WT and Gsdme^−/− mice were challenged with silica and IL-1β secretion into the supernatant assessed. Similar amounts of IL-1β secretion in response to silica were observed in BMDMs from Gsdme^−/− mice compared with WT (Fig. 5A). To address the possibility that GSDMD and GSDME may be compensating for one another, we further assessed silica-induced IL-1β secretion from BMDMs from Gsdmd^−/−Gsdme^−/− mice deficient in both GSDMD and GSDME. IL-1β secretion in response to silica was not ablated in BMDMs from Gsdmd^−/−Gsdme^−/− mice, although it was somewhat diminished compared with WT BMDMs (Fig. 5A). No significant difference in cytotoxicity was observed between LPS-primed BMDMs from WT and Gsdmd^−/−Gsdme^−/− mice in response to silica (Fig. 5B). Cytotoxicity was slightly increased in BMDMs from Gsdme^−/− mice in response to silica, although the physiological relevance of this difference is unclear (Fig. 5B).

To determine if GSDME contributed to pulmonary inflammation induced by silica in vivo, we administered silica to WT, Gsdme^−/−, and Gsdmd^−/−Gsdme^−/− mice. Eight days after challenge, the total number of neutrophils in the BALF was determined. The total number of neutrophils in the BALF of Gsdme^−/− and Gsdmd^−/−Gsdme^−/− mice was similar to that observed in WT mice (Fig. 5C, 5D), suggesting that in addition to GSDMD, GSDME is not required for the acute inflammatory response to silica in vivo. These results suggest that GSDME is not involved in the early neutrophil accumulation in the lungs during silicosis and is not essential for silica-induced IL-1β secretion. To determine if gasdermins play a role in pulmonary inflammation at later time points, we assessed neutrophils in the BALF of Gsdmd^−/− and Gsdmd^−/−Gsdme^−/− mice 14 d after silica challenge. Interestingly, at this later time point, there were fewer neutrophils in the BALF of Gsdmd^−/− and Gsdmd^−/−Gsdme^−/− mice than observed in WT mice (Supplemental Fig. 3A, 3B), suggesting that GSDMD and GSDME may play a role in the later inflammatory response to silica in vivo.

This work suggests that GSDMD does not play a role in the acute inflammatory response to inhaled silica. We report no differences in the number of neutrophils recruited to the lungs following silica challenge in Gsdmd^−/− mice compared with WT mice. Our findings contrast with recent reports suggesting that inhibiting GSDMD is essential for mitigating pulmonary inflammation to inhaled silica and subsequent development of fibrotic silicosis (24, 25). Both studies show less inflammation in the lungs or BALF in Gsdmd^−/− mice that have received silica. However, the experimental design of these studies is different from ours with respect to the concentration of silica, the length of the administration, and the time points at which inflammation is assessed. In contrast to our question, these studies address the relevance of GSDMD to the eventual development of fibrotic silicosis rather than focusing purely on the early inflammatory response to inhaled silica. Consistent with these studies, we also observed diminished neutrophils in the BALF of Gsdmd^−/− and Gsdmd^−/−Gsdme^−/− mice at 14 d after silica challenge (24, 25). Taken together, these studies suggest that GSDMD plays a role in the development of silicosis after a longer time, but it is also true that GSDMD does not have an essential role in the early acute inflammatory response to inhaled silica, as we now show.

We report that macrophages have GSDMD-independent IL-1β secretion after stimulation with silica. These data, along with our findings showing GSDMD-independent IL-1β secretion after stimulation with alum, support the idea that IL-1β release following stimulation with a particle agonist for the NLRP3 inflammasome does not require GSDMD (15). A GSDMD-independent secretion pathway for IL-1β, involving PIP2-enriched plasma membrane ruffles, has also been suggested for delayed protein secretion (16). We also report GSDMD-independent IL-1β secretion after stimulation with the nonparticulate agonists nigericin and ATP. Although many studies have shown that GSDMD is required for IL-1β secretion with similar canonical inflammasome ligands (8, 9, 12), our data suggest that the extent to which GSDMD is required for IL-1β secretion depends on the circumstances of the stimulation of cells, such as agonist concentration and time, rather than the type of agonist.

Following cleavage by caspase-3, the N-terminal fragment of GSDME oligomerizes to generate pores in the plasma membrane (19). In the absence of GSDMD, GSDME has been shown to release IL-1β from the cell after activation of the NLRP3 inflammasome pathway (26, 27). In a separate study, Aizawa et al. found that when caspase-1–mediated cleavage of GSDMD is inhibited, GSDME permits the selective release of IL-1α (28). Blocking GSDME-dependent pyroptotic pathways has been reported to be essential in preventing the development of silicosis (25). However, in our study, GSDME plays no role in neutrophil accumulation in the lung at 8 d following silica challenge in vivo or IL-1β secretion after silica stimulation of macrophages in vitro. In addition, in our study IL-1β secretion induced by silica in Gsdmd^−/−Gsdme^−/− macrophages was only slightly reduced, but not ablated, compared with WT cells.

In our model, neither GSDMD nor GSDME was required for silica-induced lung inflammation at 8 d after challenge, nor were they essential to IL-1β secretion. Hence, the mechanism by which IL-1β is released from macrophages after silica stimulation independently from GSDMD and GSDME remains unclear. A number of alternative mechanisms for IL-1β secretion have been proposed that are likely dependent on the strength and duration of the stimulus. Monteleone et al. described a lower GSDMD-independent secretion that relies on the localization of mature IL-1β in PIP2-enriched membrane microdomains (16). Transmembrane emp24 domain-containing protein 10 (TMED10) has also been implicated in the secretion of mature IL-1β (29). TMED10 facilitates the translocation of leaderless cargos, such as IL-1β, into the endoplasmic reticulum–Golgi intermediate compartment. The leaderless cargo triggers oligomerization of TMED10, which forms a protein channel facilitating cargo translocation into the vesicle (29). Another possible route of secretion is through secretory autophagy, in which leaderless cytosolic proteins are released through autophagosomes (30–32). It is known that lysosomal damage can initiate the secretory autophagy process (33), and because silica’s activation of the NLRP3 inflammasome is believed to involve lysosomal damage, secretory autophagy could be involved in GSDMD-independent IL-1β secretion. Tripartite motif containing 16 (TRIM16) has been shown to bind directly to IL-1β and act as a specialized secretory autophagy cargo receptor necessary for the release of IL-1β and other cytosolic leaderless proteins (33). Future studies are needed to examine whether alternative pathways such as TRIM16, TMED10, and PIP2 membrane microdomains play roles in the GSDMD-independent release of IL-1β after silica activation of macrophages.

The authors have no financial conflicts of interest.