MLKL-Driven Inflammasome Activation and Caspase-8 Mediate Inflammatory Cell Death in Influenza A Virus Infection

ABSTRACT Influenza A virus (IAV) triggers multiple programmed cell death pathways, including MLKL-dependent necroptosis, caspase-8-dependent apoptosis, and caspase-1-dependent pyroptosis in myeloid cells. All three pathways share common upstream regulators, namely, ZBP1 and RIPK3. Yet, the molecular mechanism underlying IAV-induced inflammasome activation remains unclear. Here, we demonstrate that MLKL promotes inflammasome activation and IL-1β processing in IAV-infected macrophages. MLKL drives NLRP3 inflammasome activation through potassium efflux. In the absence of the MLKL-inflammasome axis, caspase-8 coordinates the maturation and secretion of IL-1β. MLKL alone is dispensable for host inflammatory responses to IAV in vivo. Taken together, MLKL and caspase-8 serve as redundant mechanisms by which to drive an inflammatory form of cell death in response to an IAV infection.

IL-1b/18 into bioactive products (12)(13)(14). The GSDMD-N fragment forms pores on the plasma membrane, leading to pyroptosis, the secretion of IL-1b/18, and a plethora of other alarmins or danger signals. In contrast to MLKL and FADD/caspase-8, whether the inflammasome components directly interact with ZBP1-RIPK3 during an IAV infection is unclear. Although it has been proposed that the pyroptosis pathway acts in parallel with apoptosis and necroptosis or downstream of caspase-8, the molecular mechanisms underlying inflammasome activation during an IAV infection remain to be defined.
Emerging evidence suggests that apoptosis and necroptosis pathways can crosstalk to pyroptosis. Caspase-8 can directly cleave GSDMD and IL-1b to promote inflammatory pyroptotic cell death in fungal and bacterial infections (15)(16)(17). In these cases, GSDMD pores can cause potassium efflux and can activate the NLRP3 inflammasome, thereby contributing to the release of IL-1b (13,15,18). Caspase-3/7 downstream of caspase-8 in the apoptosis pathway can negatively regulate pyroptosis by cleaving and inactivating GSDMD (19). Active MLKL in the necroptosis pathway can also trigger the activation of the NLRP3 inflammasome via potassium efflux and can promote IL-1b secretion independently of GSDMD (20). Given that all three pathways are induced in IAV-infected cells and share upstream regulators, we investigated whether and how apoptosis and necroptosis pathways regulate the pyroptosis pathway.

RESULTS
MLKL promotes IAV-induced inflammasome activation. Bone marrow-derived macrophages (BMDMs) express the key molecules that coordinate apoptosis, necroptosis, and pyroptosis. As a result, these cells are useful in the study of the interplay of cell death pathways during IAV infection. Although IAV itself provides a priming signal that is sufficient for inflammasome activation (21), we pretreated the BMDMs with the TLR1/2 agonist Pam3CSK4 to enhance the upregulation of pro-IL-1b, which allowed for a more robust detection of mature IL-1b (Fig. S1). This also mimics the condition of IAV and bacterial coinfection, such as Streptococcus pneumoniae (22).
To investigate whether the necroptosis pathway is involved in IAV-induced inflammasome activation, we tested small molecules that block RIPK3 kinase activity (GSK'872) or RIPK3-MLKL interaction (Necrosulfonamide [NSA]) (4), and measured the release of IL-1b and lactate dehydrogenase (LDH) in the supernatant. RIPK3 inhibition and NSA blocked the release of IL-1b in a dose-dependent manner ( Fig. 1A and B). However, neither affected IAVinduced cell death, as indicated by comparable LDH release ( Fig. 1D and E). The potassium ionophore nigericin activates the canonical NLRP3 inflammasome, in which GSDMD is the primary cell death executor and is required for IL-1b secretion (13,14). Canonical NLRP3 inflammasome-mediated cell death and the release of IL-1b were inhibited by NSA, which also targets GSDMD (23), but not by the RIPK3 inhibitor GSK'872 ( Fig. 1C and F). These results suggest that active MLKL, possibly phosphorylated by RIPK3, is involved in IAVinduced inflammasome activation but not in canonical inflammasome activation.
We compared WT cells and cells lacking RIPK3, MLKL, or caspase-1/11 for their responses to IAV infections ( Fig. 2A and B). Real-time cell death was determined using automated imaging and was indicated by the percentage of SYTOX Orange permissive cells in the total cells (Fig. 2B). Consistent with the results of previous reports that show RIPK3 as the master upstream regulator (1, 2), we observed a significant delay of cell death and diminished IL-1b production from RIPK3 KO BMDMs ( Fig. 2A and B). A deficiency of caspase-1/11 or MLKL led to a slight delay in cell death at the early stage of infection, but they were largely dispensable at later stages (Fig. 2B). As expected, caspase-1/11 DKO cells produced significantly less IL-1b, compared to the WT at 12 h postinfection ( Fig. 2A). Consistent with the cells that were treated with MLKL inhibitors, MLKL KO cells had significantly less IL-1b release at this time point ( Fig. 2A). The remaining level of IL-1b in MLKL KO cells was comparable to that of caspase-1/11 DKO cells. Moreover, both MLKL KO cells and caspase-1/11 DKO cells were still capable of releasing IL-1b, and their IL-1b levels were close to those of WT cells at 24 h ( Fig. 2A).
Indeed, MLKL KO cells had reduced processing of caspase-1, IL-1b, and GSDMD at 12 and 24 h postinfection (Fig. 2C). The remaining levels of IL-1b p17 and GSDMD p30 were comparable to those of the caspase-1/11 DKO cells at 12 h, suggesting that IAVinduced inflammasome activation requires MLKL at the early stage (Fig. 2C). In contrast, canonical NLRP3 inflammasome-mediated IL-1b secretion and cell death were dependent on caspase-1/11 but not on RIPK3 or MLKL ( Fig. 2D and E). Necroptosis triggered by Pam3CSK4 and the pan-caspase inhibitor Z-VAD-FMK required RIPK3 and MLKL but not caspase-1/11 (Fig. 2F). Together, these results support a model in which MLKL promotes inflammasome activation at the earlier stages of IAV infection.
Caspase-8 can be activated through the TLR and TNFR pathways, which also signal to MLKL when caspase-8 is inhibited. In most of these cases, MLKL activation and caspase-8 activation require RIPK1 enzymatic activity (24)(25)(26). However, in IAV infection, TNFR1 KO and RIPK1 K45A kinase-inactive cells had normal IL-1b production and cell death, supporting the claim that caspase-8 and MLKL downstream of ZBP1-RIPK3 but not TNFR or RIPK1 phosphorylation are involved ( Fig. 3B and C).
Deleting both MLKL and caspase-8 further blocked the release of IL-1b, compared to MLKL KO cells, and rescued IAV-induced cell death, phenocopying RIPK3 KO cells ( Fig. 3D and E). However, caspase-8 has an essential scaffold function regulating NF-k B signaling (27)(28)(29)(30) and is required for the transcriptional upregulation of pro-IL-1b (Fig. 3F). To minimize this effect, we used small molecules inhibiting caspase-8 and caspase-3, respectively, in the background of MLKL deficiency after Pam3CSK4 priming ( Fig. 3G and H). Treating MLKL KO cells with the caspase-8 inhibitor Z-IETD-FMK recapitulated the phenotype of the MLKL/ IAV-Induced Inflammatory Cell Death by MLKL and CASP8 mBio caspase-8 DKO cells, in that IL-1b release was abolished and cell death was rescued ( Fig. 3G and H). In contrast, the RIPK3 inhibition and caspase-3 inhibitor Z-DEVD-FMK did not affect IAV-induced cell death in MLKL KO cells (Fig. 3H). The inhibition of caspase-3, which is downstream of caspase-8 in the apoptosis pathway, partially reduced the remaining IL-1b levels in MLKL KO cells (Fig. 3G). RIPK3 inhibition did not repress IL-1b in MLKL KO cells, confirming its specificity for MLKL (Fig. 3G). Together, those results support that caspase-8 and MLKL serve as two major pathways for IL-1b production and cell death during IAV infection. MLKL-mediated K + efflux drives NLRP3 inflammasome activation. IAV infection has been shown to activate the NLRP3 inflammasome (7)(8)(9). Given that MLKL-inflammasomes acts in the same axis, we hypothesized that MLKL drives NLRP3 inflammasome activation. We compared IL-1b levels and cell death in WT, MLKL KO, NLRP3 KO, and GSDMD KO cells after IAV infections ( Fig. 4A-C). NLRP3 KO cells produced significantly less IL-1b at 12 h; however, the remaining IL-1b level was higher than that observed in the MLKL KO cells (Fig. 4A). These data suggest that NLRP3 partially accounts for the MLKL-mediated inflammasome activation. GSDMD alone, which lies , and Casp1/11 2/2 BMDMs were treated with Pam3 and Z-VAD-FMK (zVAD) for 24 h. Cell death was determined by measuring the release of LDH in the supernatant. The data show one representative experiment of at least two independent experiments and are presented as the mean with the SD. N = 3 (replicate wells). The data were analyzed via a two-way ANOVA. ns, not significant; *, P , 0.0332; **, P , 0.0021; ***, P , 0.0002; ****, P , 0.0001. IAV-Induced Inflammatory Cell Death by MLKL and CASP8 mBio downstream of caspase-1/11 and caspase-8 activation, was dispensable for caspase-1 cleavage, IL-1b cleavage, and IL-1b release ( Fig. 4A-C). This suggests that inflammasome activation is not driven by the GSDMD pores that are downstream of caspase-8 and that IL-1b is released through GSDMD-independent mechanisms (Fig. 4C). Together, these observations further support that MLKL acts upstream of inflammasome activation. Caspase-8 cleavage remained intact in MLKL KO, NLRP3 KO, and GSDMD KO cells, consistent with the fact that those cells had largely normal cell deaths ( Fig. 4B and C). Deleting both MLKL and GSDMD did not alter IAV-induced IL-1b release or cell death, compared to MLKL KO cells ( Fig. 4D and E), which suggests that additional effectors, other than GSDMD, are required downstream of caspase-8.
One converged trigger for the NLRP3 inflammasome is potassium efflux (31). MLKL pores that execute necroptosis can cause potassium efflux and can activate the NLRP3 inflammasome (20). We investigated whether this mechanism contributes to inflammasome activation in IAV infection and tested it by blocking the ion efflux with extracellular ions (Fig. 4F-I). As expected, extracellular K 1 but not Na 1 at as low of a concentration as 25 mM shut down the canonical NLRP3 inflammasome that was activated by nigericin, a K 1 -H 1 ionophore. IAV-induced inflammasome activation, indicated by the release of IL-1b, was also inhibited by extracellular K 1 , but this only occurred at 100 mM, which is a much higher threshold than that observed with the nigericin-stimulated NLRP3 inflammasome. Although it remains Rip3 IAV-Induced Inflammatory Cell Death by MLKL and CASP8 mBio unclear why the thresholds were different, these data suggest that MLKL activates the inflammasome partially through necroptosis-mediated K 1 efflux, which activates NLRP3. MLKL is dispensable for host responses to IAV infection. Next, we investigated the role of MLKL in vivo by using a mouse model of IAV infection. We observed that MLKL KO mice were comparable with WT mice in terms of body weight loss and survival at a sublethal dose (Fig. 5A and B). The production of IL-1b and IL-18 as well as the recruitment of neutrophils are all hallmarks of inflammasome-mediated inflammatory responses. To examine these responses, we used a lethal dose of IAV and analyzed the mice 48 h after infection. Similar amounts of viral genome were detected in the lungs of WT and MLKL KO mice (Fig. 5C). The IL-18, IL-1b, and IFN-b levels in the bronchoalveolar lavage fluid from MLKL KO mice were comparable to those of the WT mice (Fig. 5D).  Fig. 1-4. The data show one representative experiment of at least two independent experiments and are presented as the mean with the SD. N = 3 (replicate wells). The data were analyzed via a two-way ANOVA. ns, not significant; *, P , 0.0332; **, P , 0.0021; ***, P , 0.0002; ****, P , 0.0001.

IAV-Induced Inflammatory Cell Death by MLKL and CASP8 mBio
The neutrophil numbers were also largely comparable (Fig. 5E). Together, these data support that MLKL serves as a redundant mechanism downstream of RIPK3 and that MLKL alone is dispensable for host inflammatory responses against IAV in vivo.

DISCUSSION
IAV infection leads to multiple forms of cell death, including apoptosis, necroptosis, and pyroptosis, through the common regulator RIPK3. Here, we demonstrated that the three death pathways have a hierarchy and interplay with each other in IAV-infected cells. Apoptosis and necroptosis are the primary pathways downstream of RIPK3. They act redundantly to promote not only cell death but also the processing and release of the proinflammatory cytokine IL-1b. At the earlier stages of infection, MLKL-mediated necroptosis drives inflammasome activation, contributing to the maturation of IL-1b. In the absence of the MLKL-inflammasome axis, the production of IL-1b is delayed. However, active caspase-8 is sufficient for IL-1b processing and secretion at these later stages. In these cases, necroptosis and caspase-8-dependent pathways eventually lead to an inflammatory form of cell death.
Inflammatory cell death is highly immunogenic. It is crucial in the antiviral response in that it recruits neutrophils and monocytes to the site of infection, which activates adjacent immune cells for effector functions and promotes the development of adaptive immune responses (7-9, 32, 33). Additionally, the cell deaths of infected cells can halt viral replication and can the prevent further dissemination of the virus (34). The interplay of the cell death pathways in IAV-infected cells, especially myeloid cells that express components of all three pathways, ensures the immunogenic outcome and viral restriction, even when viruses inhibit parts of the pathways. Indeed, pandemic but not seasonal H1N1 strains blocked cell death in human dendritic cells, suggesting that successful viral evasion correlates with enhanced pathogenicity (35).

(C-E) WT and
Mlkl 2/2 mice (age-and sex-matched, cohoused) were intranasally infected with PR8 (8,000 PFU) for 48 h. The viral RNA in the total lung tissues was determined via qPCR (C). The cytokine levels in the bronchoalveolar lavage fluid were determined via ELISA (D), and the cellularity was analyzed via flow cytometry (E). The data were combined from at least two independent experiments. Each dot represents an individual mouse. The mean with the SD is presented for the body weight data, and the median is presented for the other data. The data were analyzed via a two-way ANOVA. ns, not significant; *, P , 0.0332; **, P , 0.0021; ***, P , 0.0002; ****, P , 0.0001. IAV-Induced Inflammatory Cell Death by MLKL and CASP8 mBio supports our hypothesis that the redundant caspase-8 pathway is sufficient for driving inflammatory cell death. On the other hand, Casp8 DA/DA Mlkl 2/2 mice but not Casp8 DA/DA mice were more susceptible to IAV infection (38). Casp8 DA/DA mutants had impaired apoptosis without affecting necroptosis and NF-k B signaling. These data support that the MLKL-inflammasome pathway is also sufficient in the absence of apoptosis. The importance of inflammasome components for IAV infections was first identified in vivo. ASC KO, caspase-1/11 DKO, and IL-1R KO mice were more susceptible to IAV infections, whereas NLRP3 KO mice had moderate to normal susceptibility, compared with WT mice (7)(8)(9)32). We also found that NLRP3 KO BMDMs had higher remaining levels of IL-1b, compared to MLKL KO cells, suggesting that NLRP3-independent inflammasomes are also involved (Fig. 4). However, the milder phenotype of the MLKL KO mice, compared to mice that were deficient in inflammasome components, conflicted with the in vitro observations that MLKL KO BMDMs had impaired inflammasome activation. One potential reason could be that non-myeloid cells have different mechanisms of activating inflammasomes independently of MLKL and NLRP3. In support of this idea, a recent report showed that MxA, rather than NLRP3, induces ASC oligomerization and caspase-1 activation in respiratory epithelial cells (39).
Taken together, the results of our study provided a novel mechanism for IAVinduced inflammasome activation and dissected the redundant roles of the apoptosis and necroptosis pathways in driving inflammatory cell death.
The Influenza A/Puerto Rico/8/1934 H1N1 virus (PR8) was grown in SPF embryonated chicken eggs and was purchased from Charles River Laboratories (10100374). The mice were anesthetized using isoflurane and were intranasally inoculated with 25 mL diluted PR8 in PBS or PBS control. The mice were monitored for body weight and disease daily. Alternatively, bronchoalveolar lavage (BAL) fluids and whole lung tissues were collected at 2 days postinfection.
Cell culture. Bone marrow cells were harvested from approximately 6 to 24-week-old mice and were removed of red blood cells using red blood cell lysis buffer (Sigma). Cells were cultured in DMEM supplemented with 10% (vol/vol) fetal calf serum (FCS), 1Â Penicillin-Streptomycin (Corning), and L929 conditioned media containing M-CSF. Fresh medium was added every 3 days. BMDMs were split and used for experiments after at least 6 days of differentiation.
BMDMs were seeded the day before the stimulations. Cells were primed with 100 ng/mL Pam3CSK4 (Invivogen, tlrl-pms) for approximately 2 to 3 h. For the IAV infections, cells were washed twice with PBS and incubated with PR8 diluted in 0.05% (wt/vol) bovine serum albumin (BSA)/PBS for 2 h at 37°C. X-Vivo 10 media (Lonza, 04-743Q) containing FCS was added to a final concentration of 3% (vol/vol) FCS. Supernatant, cells, or a mixture of supernatant and cells were collected at various time points for downstream analysis.
Cell death assays. For real-time live cell imaging to monitor cell death, BMDMs seeded in 96-well plates were incubated with 1 mg/mL Hoechst (ThermoFisher) and 0.25 mM Sytox Orange (ThermoFisher) for 30 min. For the IAV infections, dyes were added after 2 h of incubation with the virus. For the nigericin treatment, dyes were added after priming and 30 min before the nigericin was added. The plates were read using a Cytation 5 instrument (BioTek) with the temperature controlled at 37°C and the CO 2 controlled. The percentage of Sytox Orange-positive cells in the Hoechst positive-cells was calculated using the Gen5 software package.
Alternatively, the percentage of maximum lactate dehydrogenase (LDH) release was determined using a CytoTox 96 Nonradioactive Cytotoxicity Assay (Promega).
cDNA synthesis and real-time PCR. Mouse whole lungs were harvested in QIAzol (Qiagen) and were homogenized. Total RNA was extracted using chloroform, per the manufacturer's instructions. The RNA was quantified via NanoDrop. 500 ng of RNA were reversed transcribed using an iScript cDNA Synthesis Kit (Bio-Rad). Real-time PCR was performed using SYBR green supermix (Bio-Rad) with detection on a CFX96 system (Bio-Rad). The primers were ordered from Integrated DNA Technologies. The relative expression was determined via 2-dCt, using Gapdh as the reference gene.
Flow cytometry. Cells were centrifuged from BAL fluids and were resuspended in FACS buffer (PBS, 5% FCS, 2 mM EDTA, and 0.05% sodium azide). Cells were stained for the viability dye and surface antigens on ice for 30 min. After washing with FACS buffer, the cells were fixed on ice for 15 min using 1% paraformaldehyde (PFA). Fixed cells were supplemented with PKH25 reference microbeads (Sigma) and were analyzed on a Cytek Aurora spectral cytometer. The flow data were analyzed using FlowJo. Antibodies and dyes: CD45.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. FIG S1, PDF file, 0.9 MB.

ACKNOWLEDGMENTS
We thank F. Humphries, P. Orning, and N. Ketelut-Carneiro in the Fitzgerald Lab for their discussions and advice on experimental techniques. We also thank the staff at the University of Massachusetts Chan Medical School Animal Facility for maintaining the mouse colonies.
This work was supported by funding from the National Institutes of Health to K.A.F. (AI067497), to E.L. (AI146855), and by funding from the Research Council of Norway (Center of Excellence Funding Scheme 223255/F50) to E.L. Y.C. was supported by a Charles King postdoctoral fellowship.
X.L. conceived the project, designed the study, performed the experiments, analyzed the data, and wrote the manuscript. Y.C. performed the animal experiments. E.L. provided critical resources and advice for developing the project. K.A.F. supervised the project, provided critical resources and guidance to develop the project, and wrote the manuscript.