Innate immune priming in the absence of TAK1 drives RIPK1 kinase activity–independent pyroptosis, apoptosis, necroptosis, and inflammatory disease

Inflammasomes are multimeric protein complexes consisting of a cytoplasmic sensor, an apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC) adaptor, and the cysteine protease caspase-1 (Man and Kanneganti, 2016; Poudel and Gurung, 2018). Assembly of an inflammasome complex activates caspase-1, which has two major cellular consequences: production of the proinflammatory bioactive cytokines IL-1β and IL-18 and cell death induced by the pore-forming N-terminal fragment of gasdermin D (Kayagaki et al., 2015; Shi et al., 2015). The NLRP3 inflammasome is one of the most studied inflammasomes; it is activated in response to a wide array of pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns. Conventional activation of the NLRP3 inflammasome requires two signals: a priming signal often provided through TLRs and an activation signal provided by molecules such as ATP, nigericin, alum, silica, or pore-forming toxins (Karki and Kanneganti, 2019). Although inflammasome activation is critical for protection against infection, this system comes with the cost of a potential predisposition to unwanted inflammatory diseases through excessive inflammasome activation. In line with this possibility, we recently reported the existence of a transforming growth factor-β (TGF-β)–activated kinase 1 (TAK1)–mediated mechanism of NLRP3 inflammasome quiescence under steady-state conditions; therefore, TAK1 deficiency provokes spontaneous NLRP3 activation and cell death in macrophages in the absence of any external stimulation (Malireddi et al., 2018).

TAK1 regulates cellular homeostasis and proinflammatory signaling to activate NF-κB and MAPKs downstream of several well-known receptors, including the TLRs, IL-1 receptor, and TNF receptor (Landström, 2010). In the TNF receptor–induced signaling cascade, receptor-interacting protein kinase 1 (RIPK1) and TAK1 are present in the same complex with TNF receptor type 1–associated death domain protein (also known as complex-I), which initiates NF-κB and ERK signaling (Ea et al., 2006). TAK1 is tightly regulated in cells, as both TAK1 inactivation (or deficiency) and TAK1 hyperactivation (or overexpression) promote cell death and inflammation (Guo et al., 2016; Mihaly et al., 2014). How the presence or absence of TAK1 regulates RIPK1 function and subsequent cell death has been an area of intense debate, and the detailed cellular and molecular pathways involved remain incompletely understood. Here we show that TLR priming by PAMPs bypasses the requirement for RIPK1 kinase activity in promoting cell death and NLRP3 inflammasome activation during the physiologically relevant condition of TAK1 deficiency in myeloid cells. Furthermore, TLR priming results in recruitment of RIPK1 to the caspase-8 and FADD complex in TAK1-deficient macrophages. These data highlight the formation of a novel RIPK1 kinase activity–independent ripoptosome-like complex with caspase-8 and FADD that promotes cell death and NLRP3 inflammasome activation, driving neutrophilia and severe sepsis-like shock in mice.

We recently highlighted the role of TAK1 in regulating cellular homeostasis (Malireddi et al., 2018). In contrast to its known role in promoting NF-κB and ERK activation (Landström, 2010), TAK1 restricts the tonic activation of NF-κB and ERK and the production of TNF under steady-state conditions (Malireddi et al., 2018). This homeostatic control mechanism prevents spontaneous RIPK1 kinase–driven activation of the NLRP3 inflammasome and cell death (Malireddi et al., 2018). Because microbes are ubiquitous and have evolved to inhibit host TAK1 signaling (Human Microbiome Project Consortium, 2012; Meinzer et al., 2012;Orth et al., 1999; Peterson et al., 2017), we investigated the cellular and molecular consequences of microbial (or TLR) priming on the role of RIPK1 under the conditions of TAK1 inactivation. TAK1 deficiency in bone marrow–derived macrophages (Lyz2^cre+ × Tak1^f/f BMDMs) induces spontaneous cell death that depends on RIPK1 kinase activity (Malireddi et al., 2018). Lyz2^cre+ × Tak1^f/f BMDMs underwent spontaneous cell death, as observed by SYTOX Green dye uptake after 2 and 6 h of culture (Fig. 1, A and B). Consistent with previous findings (Malireddi et al., 2018), the cell death observed in Lyz2^cre+ ×Tak1^f/f BMDMs was rescued in BMDMs isolated from a newly generated RIPK1 kinase-dead (Ripk1^KD/KD) mouse strain (Fig. 1, A and B; and Fig. S1, A and B). This result extends previous findings that RIPK1 kinase activity is crucial for inducing cell death in response to a wide variety of stimuli (Weinlich and Green, 2014) and demonstrates the important role of RIPK1 kinase activity in the induction of spontaneous cell death.

It has been shown that RIPK1 kinase activity is necessary for cell death and inflammasome activation induced by Yersinia infection and TAK1 inhibition (Malireddi et al., 2018;Orning et al., 2018; Sarhan et al., 2018). However, LPS priming of Lyz2^cre+ ×Tak1^f/f BMDMs promoted the induction of cell death independent of RIPK1 kinase activity (Fig. 1, C and D). Similarly, Yersinia enterocolitica infection of Lyz2^cre+ ×Tak1^f/f BMDMs induced cell death independent of RIPK1 kinase activity (Fig. 1, A and B; and Fig. S2, A and B). To examine whether the RIPK1 kinase activity–independent cell death inLyz2^cre+ ×Tak1^f/f mice was specific to LPS/TLR4 stimulation, we stimulated Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs with the TLR2 agonist Pam3CSK4 and the TLR3 agonist poly(I:C) (polyinosinic:polycytidylic acid). Cell death was observed in Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs in response to both TLR ligands tested (i.e., TLR2 and TLR3 ligands; Fig. 1, A, B, E, and F; and Fig. S2 A). Consistent with previous findings (Malireddi et al., 2018), the spontaneous cell death was not rescued inLyz2^cre+ ×Tak1^f/f ×Nlrp3^−/− BMDMs; however, the TLR priming enhanced the progression of cell death (Fig. 1, A–F). Thus, TLR priming through either MyD88- or TIR-domain-containing adapter-inducing IFN-β (TRIF)–signaling pathways bypasses the need for RIPK1 kinase activity to induce cell death in TAK1-deficient macrophages.

TAK1 deficiency induces spontaneous inflammasome activation that is promoted by RIPK1 kinase activity (Malireddi et al., 2018). Thus, we examined the consequences of TLR stimulation on inflammasome activation, using caspase-1 activation as the readout, inLyz2^cre+ ×Tak1^f/f BMDMs that lacked RIPK1 kinase activity. As demonstrated previously, we observed spontaneous activation of caspase-1 (as evidenced by caspase-1 p20 bands) in Lyz2^cre+ × Tak1^f/f BMDMs that depended on RIPK1 kinase activity (Fig. 2 A). TLR stimulation with LPS, PAM3CSK4, or poly(I:C) or Y.enterocolitica infection resulted in robust caspase-1 cleavage in Lyz2^cre+ ×Tak1^f/f BMDMs that was independent of RIPK1 kinase activity (Fig. 2 B and Fig. S2, C–E). TLR4 stimulation also induced proteolytic processing of caspase-3, caspase-8, and gasdermin D inLyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs, suggesting RIPK1 kinase activity–independent activation of apoptotic and pyroptotic pathways in TAK1-deficient cells (Fig. 2, A and B). The NLRP3 inflammasome is responsible for RIPK1 kinase activity–dependent caspase-1 activation in TAK1-deficient macrophages (Malireddi et al., 2018). To determine whether TLR-induced caspase-1 cleavage was also NLRP3 inflammasome-dependent, we stimulatedLyz2^cre+ ×Tak1^f/f ×Nlrp3^−/− BMDMs with LPS, Pam3CSK4, or poly(I:C). Although the TLR stimulation ofLyz2^cre+ ×Tak1^f/f ×Nlrp3^−/− BMDMs enhanced the progression of cell death (Fig. 1, A–F), caspase-1 activation was still not detected in these cells (Fig. 2, C–F). These results demonstrate that the requirement of NLRP3 for inflammasome activation is not affected by TLR priming and that TAK1-deficient BMDMs can undergo cell death independent of inflammasome activation. Together, these data also demonstrate that TLR priming alleviates the requirement of RIPK1 kinase activity to activate the inflammasome and cell death in TAK1-deficient macrophages.

Since TLR priming bypassed the need for RIPK1 kinase activity in promoting inflammasome activation and cell death in TAK1-deficient BMDMs, we next examined whether RIPK1 is totally dispensable under similar conditions, using RIPK1-deficient cells. To this end, we generated WT andRipk1^−/− macrophages from fetal liver cells (Ripk1^−/− mice are perinatally lethal; Kelliher et al., 1998). TAK1 inhibitor (TAK1-i) treatment of macrophages for 12 h induced inflammasome activation, as demonstrated by caspase-1 cleavage in WT but not Ripk1^−/− macrophages (Fig. 3 A). Moreover, LPS or TNF stimulation of TAK1-i–treated Ripk1^−/− macrophages failed to activate caspase-8 and the NLRP3 inflammasome, suggesting the requirement of full-length RIPK1 (RIPK1-scaffold function; Fig. 3 B). In concurrence with the lack of caspase-1 and caspase-8 activation,Ripk1^−/− macrophages were resistant to cell death induced by TAK1 inactivation, both in the presence and in the absence of TNF or LPS priming (Fig. 3, C and D). However, consistent with our genetic data, we also observed that pharmacological inhibition of TAK1 followed by LPS priming induced cell death and inflammasome activation in RIPK1 kinase-dead (Ripk1^KD/KD) BMDMs (Fig. 3, E and F). Together, these observations suggest that TAK1 deficiency, when combined with TLR priming, promotes RIPK1-dependent (scaffold) but kinase activity–independent cell death and NLRP3 inflammasome activation.

In TAK1-deficient BMDMs, RIPK1 kinase activity promoted spontaneous NLRP3 inflammasome activation in the absence of external stimuli, but upon LPS priming, RIPK1 kinase activity–independent function promoted NLRP3 inflammasome activation and cell death. To investigate the role of RIPK1 recruitment into the cell death complex, we used fluorescence microscopy ofLyz2^cre+ ×Tak1^f/+ (control) andLyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD (TAK1-deficient and RIPK1 kinase-dead) BMDMs after LPS stimulation and probed for the presence of RIPK1, ASC, and caspase-8. RIPK1 colocalized with ASC and caspase-8 in a single punctum in LPS-primed Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs but not in unprimed cells, suggesting that the RIPK1 complex formed independently of RIPK1 kinase activity (Fig. 4 A). Coimmunoprecipitation studies showed that in the absence of external stimuli, RIPK1 did not pull down NLRP3 or ASC in control BMDMs but pulled down both inLyz2^cre+ ×Tak1^f/f BMDMs, suggesting a spontaneous interaction between RIPK1 and the NLRP3 inflammasome in TAK1-deficient macrophages (Fig. 4 B, left panel). Furthermore, this spontaneous RIPK1 interaction with NLRP3 and ASC was abrogated in Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs, suggesting that RIPK1 kinase activity drove spontaneous NLRP3 inflammasome assembly in unprimed TAK1-deficient BMDMs (Fig. 4 B, left panel). After LPS priming, the association of RIPK1 with NLRP3 and ASC was observed in Lyz2^cre+ ×Tak1^f/f BMDMs; however, this interaction was also observed in Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs (Fig. 4 B, right panel). Together, these studies demonstrate that TLR priming of Tak1^f/f ×Lyz2^cre+ BMDMs switches the kinase-dependent association of RIPK1 with NLRP3 and ASC to a kinase activity–independent process of assembling a noncanonical ripoptosome-like cell death complex.

In TAK1-deficient macrophages, we observed that RIPK1 was recruited to the caspase-8–containing cell death complex (Fig. 4 A). Moreover, RIPK1 is a potent upstream activator of the RIPK3–caspase-8–mediated cell death pathway (Weinlich and Green, 2014). We therefore sought to address the effect of priming on the role of the RIPK3–caspase-8 axis in cell death and inflammasome activation in TAK1-deficient cells. Caspase-8 and FADD deficiency are embryonically lethal in mice, but the addition of RIPK3 deficiency rescues this embryonic lethality (Dillon et al., 2012; Kaiser et al., 2011; Oberst et al., 2011; Zhang et al., 2011). Therefore, to study caspase-8 functions, we used BMDMs fromRipk3^−/− mice as controls and compared their responses with those ofRipk3^−/− ×Casp8^−/− andRipk3^−/− ×Fadd^−/− BMDMs following TAK1-i treatment in the presence or absence of TLR priming. Consistent with earlier studies (Orning et al., 2018; Sarhan et al., 2018), stimulation by TAK1-i or LPS + TAK1-i induced caspase-1 cleavage and cell death inRipk3^−/− BMDMs but not inRipk3^−/− ×Casp8^−/− orRipk3^−/− ×Fadd^−/− BMDMs (Fig. S3, A–C). In addition, Pam3CSK4 and poly(I:C) also failed to induce cell death and caspase-1 activation in TAK1-i–treated Ripk3^−/− × Casp8^−/− andRipk3^−/− ×Fadd^−/− macrophages (Fig. S3, D–F). Moreover, we also confirmed the specificity of these findings by genetic deletion of TAK1, RIPK3, and caspase-8 using BMDMs prepared fromLyz2^cre+ ×Tak1^f/f ×Ripk3^−/− ×Casp8^−/− mice (Fig. S4, A and B). Thus, under the conditions of TAK1 deficiency, TLR priming through either the TLR adapter MyD88 (Pam3CSK4) or TRIF (poly(I:C)) is sufficient to promote assembly of a RIPK1 kinase activity–independent cell death complex with FADD and caspase-8, which can drive potent inflammatory cell death and inflammasome activation.

Given that TLR stimulation–induced cell death inLyz2^cre+ ×Tak1^f/f BMDMs was independent of RIPK1 kinase activity, we next investigated the requirement of RIPK1 kinase activity for promoting necroptosis. As expected, LPS + z-VAD (a prototypical necroptotic stimulus) promoted necroptosis in Lyz2^cre+ × Tak1^f/+ (control) BMDMs, and this cell death was abrogated in kinase-dead RIPK1 (Lyz2^cre+ ×Tak1^f/+ ×Ripk1^KD/KD) BMDMs or cells deficient in TAK1, RIPK3, and caspase-8 (Lyz2^cre+ ×Tak1^f/f ×Ripk3^−/− ×Casp8^−/−; Fig. 5 A and Fig. S5 A). However, the same LPS + z-VAD stimulation induced necroptotic cell death in Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs (Fig. 5 A), revealing that TAK1 deficiency alleviates the requirement for RIPK1 kinase activity to induce necroptosis. To further corroborate the RIPK1 kinase activity–independent form of necroptotic cell death, we analyzed BMDMs for MLKL phosphorylation and oligomerization, the biochemical markers of necroptosis. Western blot analysis showed phosphorylation and oligomerization of MLKL in Lyz2^cre+ × Tak1^f/f ×Ripk1^KD/KD after LPS + z-VAD or TNF + z-VAD stimulation (Fig. 5 B and Fig. S5 B), demonstrating the occurrence of RIPK1 kinase activity–independent necroptosis in TAK1-deficient macrophages.

Since our data from the Lyz2^cre+ ×Tak1^f/f ×Ripk3^−/− ×Casp8^−/− andRipk3^−/− ×Casp8^−/− mice (Fig. 5, Fig. S3, and Fig. S5, A and B) showed a requirement for RIPK3 and capase-8 to drive cell death under the conditions of TLR priming and caspase inactivation, we further examined the role of RIPK3 in driving necroptosis in TAK1-deficient cells. To determine whether RIPK3 is required for necroptosis induction in TLR-primed TAK1-deficient cells, we stimulated BMDMs with LPS + z-VAD, which induced robust cell death and MLKL phosphorylation inLyz2^cre+ ×Tak1^f/+,Lyz2^cre+ ×Tak1^f/f, andLyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD cells (Fig. 6, A and B). However, cell death and MLKL activation were not detected in BMDMs from Lyz2^cre+ × Tak1^f/f ×Ripk3^−/− orLyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD ×Ripk3^−/− BMDMs. As expected, TLR stimulation with LPS resulted in robust cell death (Fig. 6 A) and also induced proteolytic activation of caspase-1, caspase-8, and gasdermin D inLyz2^cre+ ×Tak1^f/f,Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD, andLyz2^cre+ ×Tak1^f/f ×Ripk3^−/− BMDMs (Fig. 6 C). Moreover, TLR stimulation also promoted robust cell death and caspase-1, caspase-8, and gasdermin D activation in TAK1-deficient BMDMs that lack both RIPK3 and RIPK1 kinase activity (Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD ×Ripk3^−/−; Fig. 6, A and C). Together, these results demonstrate that the requirement of RIPK3 for necroptosis activation is not altered by TLR priming and that TAK1-deficient cells can be rescued from cell death by combined inactivation of necroptosis and caspase-mediated cell death pathways.

Lyz2cre-mediated cell-specific deletion of TAK1 in mice induces myeloid proliferation that is hallmarked by an increase in the neutrophil population in the blood (Ajibade et al., 2012; Lamothe et al., 2012; Malireddi et al., 2018). This neutrophilia is partially rescued in RIPK1 kinase-dead, TAK1-deficient mice (Malireddi et al., 2018). However, it is not clear if the hyperactivated RIPK3 and/or caspase-8 drive this disease in TAK1-deficient mice. Moreover, to further extend our in vitro observations, we investigated the role of the RIPK3–caspase-8 axis in driving the peripheral blood neutrophilia seen in TAK1-deficient mice. Consistent with other studies (Orning et al., 2018; Sarhan et al., 2018) and our in vitro cell death observations, RIPK3 deficiency in TAK1-deficient mice (Lyz2^cre+ ×Tak1^f/f ×Ripk3^−/−) did not rescue this neutrophilia; however, RIPK3 and caspase-8 double-deficiency (Lyz2^cre+ ×Tak1^f/f ×Ripk3^−/− ×Casp8^−/−) restored the neutrophil frequency and numbers to those of control mice (Fig. 7 A). Hyperinflammation is often associated with the pathophysiology of neutrophilic diseases, and inactivation of MAPKs such as TAK1 exacerbate inflammatory responses that promote neutrophilia and sepsis (Ajibade et al., 2012; Meinzer et al., 2012; Orth et al., 1999; Peterson et al., 2017). To study this phenomenon, we developed a physiologically relevant model of TAK1-i–induced sepsis (Fig. 7 B). To examine the physiological relevance of RIPK1 kinase activity and caspase-8 in vivo, we treated WT, Ripk1^KD/KD,Ripk3^−/−, andRipk3^−/− ×Casp8^−/− double-knockout (DKO) mice with an extremely mild, nonlethal dose of LPS (10 µg/mouse) in the presence of TAK1-i. This combination resulted in sensitization and susceptibility of WT mice to low-dose LPS treatment and the development of sepsis-like shock resulting in the death of 60% of the WT mice. However,Ripk1^KD/KD mice showed significant protection from TAK1-i + LPS-induced septic shock (Fig. 7 C), indicating a role for RIPK1 kinase activity in acute inflammatory responses in this model. In concurrence with the in vitro findings,Ripk3^−/− ×Casp8^−/− DKO mice were fully protected from this sepsis-like shock (Fig. 7 C). Liver sections fromRipk3^−/− ×Casp8^−/− DKO mice consistently lacked signs of inflammation and were protected from neutrophilic infiltration (Fig. S5 C). Deficiency of FADD, an upstream activating adaptor of caspase-8, also provided similar protection from organ damage during TAK1-i + LPS treatment in mice (Fig. S5 C). Furthermore, we found that lack of TAK1 in myeloid cells (Lyz2^cre+× Tak1^f/f) enhanced susceptibility to LPS-induced septic shock (Fig. 7 D). Lack of RIPK1 kinase activity resulted in partial rescue of the mice, while the combined deletion of RIPK3 and caspase-8 provided full protection from sepsis-induced death (Fig. 7 D). These results support that TAK1 inactivation in myeloid cells relieves the requirement for RIPK1 kinase activity but does require RIPK3 and caspase-8 to drive cell death and inflammation. Together, these studies demonstrate that ripoptosome components caspase-8 and FADD play central roles in promoting myeloid proliferation and severe sepsis in the absence of TAK1.

TAK1 is required for NF-κB and MAPK signaling downstream of several pattern-recognition receptors, growth receptors, and cytokine receptors (Ajibade et al., 2013; Dai et al., 2012). We previously described a paradoxical role for TAK1 in restricting RIPK1 kinase activity–dependent NLRP3 inflammasome activation and regulation of cellular homeostasis (Malireddi et al., 2018).

Recent studies have further affirmed the occurrence of RIPK1 kinase activity–dependent pyroptosis under the conditions ofYersinia-mediated TAK1 inactivation (Orning et al., 2018; Sarhan et al., 2018). Previously, we have also demonstrated that RIPK1 kinase activity promotes neutrophilic dermatosis in the spontaneousPtpn6^spin mouse model of the inflammatory syndrome (Lukens et al., 2013). However, here we investigated the molecular mechanisms that govern cell death and inflammation under TAK1-deficient conditions that would occur in the presence of PAMPs to model pathogenic insults. We found an association of RIPK1 with NLRP3 and ASC and a priming-independent role for caspase-8 and FADD in driving inflammasome activation and cell death in TAK1-deficient BMDMs. We also discovered the existence of RIPK1 kinase activity–independent inflammatory cell death and inflammasome activation. Our findings suggest that RIPK1 kinase inhibition may not successfully treat disorders associated with TAK1 dysfunction when TLRs are stimulated (such as by the microbiome, infection, or damage). Although microbes have evolved to inactivate central nodes of inflammatory signaling, such as TAK1 and other MAPKs, to modulate the host immune response, host coevolution has opted to sense these intracellular insults to promote alternative modes of immune activation and inflammatory cell death (Ajibade et al., 2012; Meinzer et al., 2012;Orth et al., 1999; Peterson et al., 2017). Our current findings show an alternate pathway of ripoptosome-like, but RIPK1 kinase activity–dependent and –independent mechanisms to drive inflammatory cell death, which may also contribute to the pathophysiology of some inflammatory diseases and sepsis.

The differences in the kinase and scaffold functions of RIPK1 have been well studied. The scaffolding function (through K63-linked polyubiquitination) is required to recruit TNF receptor–associated factor 2 (TRAF2), cellular inhibitor of apoptosis proteins, TAK1, and TGF-β–activated kinase and MAP3K7 binding protein 2 (TAB2) to stabilize complex-I, which promotes the activation of NF-κB and ERK signaling (Ea et al., 2006; Lee et al., 2004). In the absence of complex-I components, ligation of TNFR1 or TLRs leads to RIPK1 kinase-dependent recruitment of FADD and caspase-8, which assembles complex-II to promote apoptosis; however, as a backup mechanism of cell death, complex-II can also drive RIPK3-MLKL–mediated necroptosis when caspases are inhibited (Feoktistova et al., 2011; Tenev et al., 2011). Thus, the RIPK1 scaffold function promotes prosurvival signaling, whereas RIPK1 kinase activity is important for inducing RIPK1-mediated cell death. RIPK1 plays central roles in regulating cell death and inflammation, which is particularly important to promote immune defense against pathogenic microbes such as Yersinia species. Our data provide evidence for the presence of a novel cell death mechanism and support the concept that cells have evolved to sense TAK1 inactivation as a danger signal to drive RIPK1 kinase activity–independent modes of cell death to bypass the pathogen-mediated immune evasion strategies. However, under these conditions, it is also possible that microbial PAMPs may prime an unknown serine/threonine kinase that substitutes for RIPK1 kinase activity in activating downstream target molecules for cell death induction. Alternatively, TLR priming may promote direct RIPK1 phosphorylation, ubiquitination, or a transition to an open conformation that facilitates its role as a scaffold protein to induce cell death. In line with this hypothesis, we found that TLR priming accelerated cell death in TAK1-deficient BMDMs, but more importantly, it promoted cell death of TAK1-deficient BMDMs even in the absence of RIPK1 kinase activity. Specifically, LPS stimulation promoted cell death in Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs. A recent report described that TAK1-deficient macrophages undergo a partially RIPK1 kinase activity–independent form of inflammatory cell death (Sanjo et al., 2019), complementing our results. We observed a significant role for RIPK1 kinase activity in TAK1-i + LPS–induced septic shock, supporting its role in driving acute inflammatory responses in this model. These findings also reflect the delayed cell death observed in TAK1-i + LPS–treated Ripk1^KD/KD BMDMs, suggesting that the delayed myeloid cell death may contribute to the protection from septic shock. Alternatively, it is also highly likely that the TAK1-i targets other nonmyeloid cells, which would not be affected in the myeloid-specificLyz2^cre+ ×Tak1^f/f mice. Therefore, the hypersusceptibility to sepsis in this case is a result of combined responses coming from both the myeloid and nonmyeloid cells. However, we observed that genetic deletion of TAK1 in myeloid cells (Lyz2^cre+ × Tak1^f/f) augmented the susceptibility to LPS-induced septic shock and was only partially dependent on RIPK1 kinase activity compared with the combined deletion of RIPK3 and caspase-8, which provided complete protection in this model. These results support that TAK1 inactivation in myeloid cells relieves the requirement for RIPK1 kinase activity but not RIPK3 and caspase-8 to drive cell death and inflammation. We further extended these findings using RIPK1-deficient fetal liver–derived macrophages, which demonstrated a requirement for RIPK1 scaffold function, while RIPK1 kinase activity is dispensable for cell death and inflammasome activation under TAK1-deficient conditions. The upstream mechanisms acting on RIPK1 to drive the scaffold-mediated cell death observed in LPS-stimulatedLyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs are not known and warrant further investigation. Regardless, our data highlight a RIPK1 kinase activity–independent, ripoptosome-like cell death complex that drives inflammasome activation and cell death in TAK1-deficient cells.

TAK1-deficient BMDMs undergo spontaneous forms of cell death that include apoptosis, pyroptosis, and the alternative backup process of necroptosis (Malireddi et al., 2018). Recent findings have highlighted the importance of homeostatic regulation of RIPK1 in preventing cell death and the development of inflammatory diseases such as amyotrophic lateral sclerosis, and multiple MAPKs coordinate to restrain spontaneous RIPK1 activation by mediating inhibitory phosphorylation of this kinase (Dondelinger et al., 2015; Geng et al., 2017; Jaco et al., 2017; Xu et al., 2018). Pharmacological inactivation of TAK1 also drives RIPK1 kinase-dependent apoptosis (Amin et al., 2018). We observed that caspase-8 was associating with inflammasome components in TAK1-deficient BMDMs and was important for inflammasome activation under both unprimed and primed conditions. Importantly, TLR priming bypassed the requirement for RIPK1 kinase activity and promoted cell death and inflammasome activation inLyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD BMDMs. Moreover, a recent study showed that caspase-8–mediated cleavage of RIPK1 is essential for limiting aberrant cell death and loss of developmental homeostasis (Newton et al., 2019). Interestingly, they observed thatRipk1^{D138N,D325A/D138N,D325A} mice that lack both RIPK1 kinase activity and sites for cleavage by caspase-8 were runted and did not survive to weaning age, suggesting a kinase activity–independent role for RIPK1 in cell death, which complements our findings. Thus, both the kinase-dependent and -independent functions of RIPK1 drive inflammatory cell death.

In summary, using TAK1-deficient cells, we have uncovered a previously unknown role for RIPK1 kinase activity–dependent and –independent functions in promoting multifaceted cell death complex formation, which drives activation of the NLRP3 inflammasome, pyroptosis, apoptosis, and necroptosis, referred to as PANoptosis (Malireddi et al., 2019). The occurrence of RIPK1 kinase activity–independent necroptosis in TAK1-deficient macrophages provides a novel model to further explore the alternative cellular mechanisms of MLKL and/or RIPK3 activation with potential implications for inflammatory cell death and disease. Future studies should address how TLR priming during TAK1 deficiency bypasses the requirement for RIPK1 to induce necroptosis. Together, our findings highlight the complexity in regulating inflammatory cell death and identify potential factors that could be targeted to treat myeloid proliferation and sepsis.

Ripk1^−/− (Kelliher et al., 1998);Ripk3^−⁄− (Newton et al., 2004);Mlkl^−⁄−,Ripk3^−⁄− ×Fadd^−⁄−, andRipk3^−⁄− ×Casp8^−⁄− (Gurung et al., 2014; Oberst et al., 2011);Nlrp3^−⁄− (Kanneganti et al., 2006); andLyz2^cre+ ×Tak1^f/f ×Nlrp3^−⁄− (Malireddi et al., 2018) mouse strains were all described previously. Tak1^f/f mice were bred with Lyz2^cre+ (B6.129P2-Lyz2^tm1(cre)Ifo/J; The Jackson Laboratory) mice to generate conditional TAK1-knockout mice. TAK1-knockout mice were bred with Ripk1^KD/KD(K45A) mice (generated in this study and described below),Mlkl^−⁄−,Ripk3^−⁄−, orRipk3^−⁄− ×Casp8^−⁄− mice to generate Lyz2^cre+ ×Tak1^f/f ×Ripk1^K45A (Ripk1^KD/KD),Lyz2^cre+ ×Tak1^f/f ×Mlkl^−⁄−,Lyz2^cre+ ×Tak1^f/f ×Ripk3^−⁄−,Lyz2^cre+ ×Tak1^f/f ×Ripk1^KD/KD ×Ripk3^−⁄−, andLyz2^cre+ ×Tak1^f/f ×Ripk3^−⁄− ×Casp8^−⁄− mice. C57BL/6 WT (The Jackson Laboratory) and littermate controls were bred at St. Jude Children’s Research Hospital (St. Jude). Animal studies were conducted under protocols approved by the St. Jude Animal Care and Use Committee.

Ripk1^K45A mice were generated using CRISPR/Cas9 technology. Staff in the St. Jude Transgenic/Gene Knockout Shared Resource injected pronuclear-staged C57BL/6J zygotes with a single-guide RNA (sgRNA; Ripk1_K45A_Guide 01, 5′-TCC​TGA​AAA​AAG​TAT​ACA​CA-3′; 125 ng/µl) designed to introduce a DNA double-stranded break into exon 2 of the Ripk1 gene, human codon-optimizedCas9 mRNA transcripts (50 ng/µl), and a 200-nucleotide single-stranded DNA molecule containing the desired mutations (Ripk1–K45A–HDR [homology-directed repair], 5′-AGG​CTT​CGG​GAA​GGT​GTC​CTT​GTG​TTA​CCA​CAG​AAG​CCA​TGG​ATT​TGT​CAT​CCT​GGC​GAA​AGT​CTA​CAC​CGG​TCC​CAA​CCG​CGC​TGA​GTG​AGT​TGG​GGG​CAT​AAA​GGG​CTT​GGC​TTT​TTG​CTA​GCT​GAC-3′;Fig. S1). To facilitate the identification of founder mice and genotyping, silent substitutions generating an AgeI restriction site (5ʹ-ACCGGT-3ʹ) were also introduced. Zygotes were surgically transplanted into the oviducts of pseudopregnant CD1 females, and newborn mice carrying theRipk1^K45A allele were identified by PCR, AgeI-restriction digestion, and Sanger sequencing using primers Ripk1-K45A-F51 (5′-CGG​TCC​TTT​TGC​CCT​GAG​AC-3′) and Ripk1-K45A-R52 (5′-AAA​AAG​GCG​CCC​CTC​TCA​A-3′). The sgRNAs and Cas9 mRNA transcripts were designed and generated as described previously (Pelletier et al., 2015). The target site for each sgRNA is unique in the mouse genome, and no potential off-target site with fewer than three mismatches was found using the Cas-OFFinder algorithm (Bae et al., 2014).

BMDMs were differentiated from bone marrow cells as detailed previously (Gurung et al., 2012). In brief, cells isolated from bone marrow were maintained in culture in L929 cell-conditioned IMDM supplemented with 10% fetal bovine serum, 1% nonessential amino acids, and 1% penicillin–streptomycin for 5 d to differentiate into macrophages. Fully differentiated BMDMs were collected on day 5 and seeded into 12-well cell culture plates. As indicated for pharmacologic inhibition, BMDMs were treated before or after stimulation with chemical inhibitors of inflammatory signaling or cell death. In other experiments, BMDMs were treated with the TAK1-i 5Z-7-oxozeaenol (0.1 µM), Pam3CSK4 (10 µg/ml), poly(I:C) (10 µg/ml), or mTNF (50 ng/ml) or infected with Y. enterocolitica (ATCC 700822) at a multiplicity of infection (MOI) of 5 to study inflammasome activation and cell death.

Y. enterocolitica was grown on 2× yeast extract tryptone (YT) agar plates overnight under aerobic conditions at 37°C. Next, single colonies were cultured in the 2× YT broth overnight at 26°C. Then the bacteria were subcultured (1:10) in fresh 2× YT broth for 2 h and washed and resuspended in PBS for infection of BMDMs. Extracellular bacteria were killed by adding gentamicin (100 µg/ml) 1 h after infection.

The neutrophilia phenotype was assessed in vivo by a standard complete blood count performed on an automated laboratory instrument. Peripheral blood samples were collected in EDTA (1% wt/vol) fromLyz2^cre+ ×Tak1^f/+,Lyz2^cre+ ×Tak1^f/f,Lyz2^cre+ ×Tak1^f/+ ×Ripk3^−/− ×Casp8^+/−,Lyz2^cre+ ×Tak1^f/f ×Ripk3^−/− ×Casp8^+/−, andLyz2^cre+ ×Tak1^f/f ×Ripk3^−/− ×Casp8^−/− mice. For in vivo septic shock treatments, WT or genetically manipulated knockout mice were intraperitoneally injected with TAK1-i (50 mg/kg body weight) followed by LPS (10 µg) 2 h later, or alternatively mice were subjected to standard LPS shock (25 mg/kg body weight). Tissue samples were collected 6 h after completion of TAK1-i treatment for histological analysis of inflammation and cell death, or the mice were monitored for an extended period (50 h) for survival analysis.

Protein samples for Western blotting were prepared by collecting and combining cell lysates with cell culture supernatants. Samples were denatured by boiling for 10 min at 100°C in loading buffer containing 10% SDS and 100 mM dithiothreitol. SDS-PAGE–separated proteins were transferred to Amersham Hybond P polyvinylidene difluoride membranes (10600023; GE Healthcare Life Sciences) and immunoblotted with primary antibodies against caspase-1 (mouse mAb, Casper-1, AG-20B-0042; AdipoGen Life Sciences), caspase-3 (rabbit polyclonal antibody [pAb], 9662; Cell Signaling Technology), cleaved caspase-3 (rabbit pAb, 9661; Cell Signaling Technology), gasdermin D (rabbit mAb, ab209845; Abcam), ASC (rabbit pAb, AL177, AG-25B-0006; AdipoGen Life Sciences), RIPK1 (rabbit mAb, 3493; Cell Signaling Technology or mouse anti-RIP clone 38/RIP, 610458; BD Biosciences), caspase-8 (rat mAb, clone 1G12, AG-20T-0137; AdipoGen Life Sciences), cleaved caspase-8 (rabbit mAb, Asp391, 9496; Cell Signaling Technology), MLKL (rabbit pAb, AP14272b; Abgent), phospho-MLKL (rabbit mAb, 37333; Cell Signaling Technology), NLRP3 (mouse mAb, Cryo-2, AG-20B-0014; AdipoGen Life Sciences), or β-actin (rabbit mAb, 13E5, 4970; Cell Signaling Technology) followed by secondary anti-rabbit, anti-mouse, or anti-rat HRP antibodies (Jackson ImmunoResearch Laboratories), as previously described (Kanneganti et al., 2006).

For protein–protein interaction studies, cell lysates prepared in NP-40 lysis buffer (1.0% NP-40, 150 mM NaCl, and 50 mM HEPES) containing complete protease inhibitors and phospho-STOP (Roche) were subjected to immunoprecipitation with 3 µg/sample of the indicated primary antibodies that were cross-linked to Dynabeads (Superparamagnetic Dynabeads M-270; Thermo Fisher Scientific) on a rocking platform for 12 to 16 h at 4°C. After the incubation, the beads were centrifuged and washed three times with the lysis buffer. Immunoprecipitates were eluted in sample buffer after three washes in the 2× SDS loading buffer and then subjected to immunoblotting analysis.

BMDMs from WT and the indicated knockout mice were seeded in 12-well cell culture plates and were either left untreated (control) or treated with TAK1-i or different inhibitors of inflammation or cell death for the indicated times. Light microscopy images were obtained using an Olympus CKX41 microscope with a 40× objective lens. Digital image recording and analysis were performed with the INFINITY ANALYZE software (Lumenera Corporation). The images were processed and analyzed with ImageJ software. For histological analysis, livers were fixed in 10% formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. A board-certified pathologist analyzed the stained sections for the presence of inflammation and signs of tissue damage in a blinded manner.

Unprimed or primed BMDMs were fixed for 5 min at room temperature in 4% ChemCruz paraformaldehyde (sc-281692; Santa Cruz Biotechnology). The cells were permeabilized and blocked for nonspecific binding with 10% normal goat serum (Sigma-Aldrich) in PBS containing 0.1% Triton X-100. Cells were incubated with the following primary antibodies overnight at 4°C or for 2 h at room temperature: anti-ASC (rabbit pAb AL177; diluted 1:200; AdipoGen Life Sciences), anti-RIPK1 (mouse anti-RIP clone 38/RIP; diluted 1:50; BD Biosciences), or anti–caspase-8 (rat mAb clone 1G12; diluted 1:100; AdipoGen Life Sciences). The secondary antibodies used were an Alexa Fluor 594–conjugated antibody against rat immunoglobulin G (A-11007; diluted 1:250; Life Technologies), an Alexa Fluor 568–conjugated antibody against rabbit immunoglobulin G (A-11011; diluted 1:250; Life Technologies), and an Alexa Fluor 488–conjugated antibody against mouse immunoglobulin G (R37120; diluted 1:250; Life Technologies). Cells were counterstained with DAPI mounting medium (Vector Laboratories), and confocal microscopy was performed using an AxioObserverZ.1 inverted microscope (Zeiss) equipped with a CSU-X spinning disc (Yokagawa), 100× (1.45 numerical aperture) objective, and Prime 95B complementary metal oxide semiconductor camera (Teledyne Photometrics). Images were acquired and analyzed using Slidebook software (Intelligent Imaging Innovations).

Time-course analysis of cell death assays was performed using a two-color IncuCyte Zoom in incubator imaging system (Essen Biosciences). Fully differentiated BMDMs were seeded in 12-well (1.0 × 10⁶ cells/well) or 24-well (0.5 × 10⁶ cells/well) tissue culture vessels in the presence of 100 nM of the cell-impermeable DNA binding fluorescent dye SYTOX Green (S7020; Life Technologies), which rapidly enters dying cells during membrane permeabilization. A series of images were acquired with a 20× objective and analyzed using the IncuCyte S3 software, which allows precise analysis of the number of SYTOX Green–positive cells present in each image. A minimum of three images per well was taken for the analysis of each time point. Dead-cell events for each strain of BMDMs were acquired via SYTOX Green staining and plotted using GraphPad Prism version 5.0 software.

GraphPad Prism version 5.0 software was used for data analysis. Data are shown as mean ± SEM. Statistical significance was determined by t tests (two-tailed) for two groups, log-rank (Mantel-Cox) test for survival analysis, and one-way ANOVA (with Dunnett’s or Tukey’s multiple comparisons tests) for three or more groups. The number of repeats for each experiment is indicated in the corresponding figure legends; n in the figure legends represents the number of animals used in the experiments. P < 0.05 was considered statistically significant and is represented as *; ns, not significant.

Fig. S1 shows generation ofRipk1^KD/KD mice by using CRISPR/Cas9 technology.Fig. S2 shows that priming via TLR2 or TLR3 or by Y. enterocolitica bypasses the requirement for RIPK1 kinase activity in promoting cell death and caspase-1 activation. Fig. S3 shows that during TAK1 inactivation, FADD and caspase-8 are required for cell death and inflammasome activation, both in the presence and in the absence of priming.Fig. S4 shows that genetic deletion of RIPK3 and caspase-8 protects TAK1-deficient cells from cell death, both in the presence and in the absence of priming. Fig. S5 shows that genetic deletion of caspase-8 rescues TAK1 inactivation–induced inflammatory cell death of myeloid cells and that FADD deletion rescues TAK1-deficient mice from the pathophysiology of sepsis.

We thank Dr. Hongbo Chi (St. Jude Children’s Research Hospital, Memphis, TN) for generously supplying TAK1-mutant mice and Rebecca Tweedell, PhD, for scientific editing and writing support.

T.-D. Kanneganti is supported by funding from the National Institutes of Health (grants AI101935, AI124346, AR056296, and CA163507) and the American Lebanese Syrian Associated Charities. P. Gurung is supported by the National Institute of Allergy and Infectious Diseases (grant K22AI127836) and the National Institute of Environmental Health Sciences (P30ES005605 pilot grant) and by University of Iowa startup funds.

Author contributions: R.K.S. Malireddi, P. Gurung, and T.-D. Kanneganti designed the study. R.K.S. Malireddi, P. Gurung, S. Kesavardhana, P. Samir, A. Burton, H. Mummareddy, S. Pelletier, and S. Burgula performed the experiments. R.K.S. Malireddi, P. Gurung, and T.-D. Kanneganti analyzed the data with input from the other authors. P. Vogel conducted the histological analysis. R.K.S. Malireddi, P. Gurung, and T.-D. Kanneganti wrote the manuscript with input from the other authors. T.-D. Kanneganti oversaw the project.