Main

Apoptosis is considered as the predominant, programmed pathway of cell death; however, recently several pathways of regulated cell death have been shown to operate in cells that are considered highly inflammatory.1, 2 Although apoptosis has a major role during fetal development,3 regulated necrotic cell death does not appear to influence fetal development.4 Interestingly, receptor interacting protein kinase 1 (RIPK1) deficiency cause embryonic lethality, suggesting a key role of RIPK1 in host survival.5 Besides the kinase activity, RIPK1 appears to have a scaffolding function, which may influence immune homeostasis.6, 7 Depending on the interacting partners and posttranslational modifications, RIPK1 has a multifaceted role in cell signaling and cell survival.8 Following TNF-R1 signaling, RIPK1 transitions between pro-survival and pro-cell death signaling complexes.9, 10

Necroptosis is a form of regulated necrosis of cells that operates in the absence of caspase activity9 and is initiated by the engagement of various TLRs or cytokine receptors.11 The first cardinal signaling step in necrosome signaling is the phosphorylation of RIPK1, which leads to RIPK1–RIPK3 interaction and phosphorylation of RIPK3.12 Although necroptotic signaling has been shown to be dependent on the homotypic interaction of RIPK1 and RIPK3, in certain cases RIPK3 can induce necroptosis independently of RIPK1.13 Indeed, the engagement of TNF-R, IFNAR1, TLR3 or TLR4 leads to phosphorylation of RIPK1 without induction of necroptosis.14

In contrast, phosphorylation of RIPK3 and the subsequent phosphorylation of the mixed lineage kinase-like (MLKL) protein is essential for induction of necroptosis. Phosphorylation of MLKL results in oligomerization of MLKL, which then translocate to cell membrane and disrupt cellular integrity by forming pores in the cell membrane.15, 16, 17

Irrespective of the relative protein interactions that lead to a regulated cell death, RIPK1 also interacts with various adaptor proteins. Through its RHIM domain, RIPK1 interacts with TRIF and DAI18, 19 and consequently influences downstream signaling following ligation of pathogen-recognition receptors. In this report, we have used a kinase-dead mutant of RIPK1 (K45A mutation)20 and evaluated the impact on regulated cell death and cytokine signaling in macrophages and the consequent impact on inflammatory response in vivo.

Results

Lysine-45 of RIPK1 promotes necroptosis of macrophages induced by various stimuli

We have previously shown that lipopolysaccharide (LPS), TNFα and IFNβ can induce necroptosis of macrophages in the presence of zVAD.14 We used a similar approach to evaluate whether the kinase activity of RIPK1 is important in promoting necroptosis of macrophages. Using mice with a mutation in K45 of RIPK1,20 our results indicate that RIPK1K45A macrophages survive better than the wild-type (WT) macrophages when stimulated with LPS/zVAD, TNFα/zVAD or IFNβ/zVAD (Figures 1a–c). For pharmacological inhibition of RIPK1 kinase activity, we also used Nec-1s (the stable variant of necrostatin-1 (Nec-1)).21 Nec-1s blocked the cell death of WT macrophages when treated with LPS/zVAD, TNFα/zVAD or IFNβ/zVAD (Figure 1a), indicating that kinase activity of RIPK1 was responsible for the cell death of macrophages. The resistance of RIPK1K45A macrophages was greatest upon stimulation by TNFα/zVAD in comparison to LPS/zVAD, which is in agreement with previous studies.20, 22 Furthermore, the resistance of RIPK1K45A macrophages to necroptosis was not as great as that of RIPK3-deficient macrophages (Figure 1a), indicating that there may be additional mechanisms besides the K45 of RIPK1 that promotes cell death.

Figure 1
figure 1

Lysine-45 of RIPK1 promotes necroptosis of macrophages induced by various stimuli. (a) Cell survival of macrophages at 24 h posttreatment with LPS (100 ng/ml), TNFα (50 ng/ml), IFNβ (100 U/ml)+/−zVAD-fmk (50 μM) and/or Nec-1s (10 μM), as measured by MTT assay. (b) Images of macrophages treated as in panel (a) and co-stained with propidium iodide (PI) and Hoechst, scale 200 μm. (c) Numbers of PI-positive cells were quantified as in panel (b). CTRL, non-treated cells. ANOVA with Bonferroni post-test, *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; NS, not significant. Error bars are S.E.M. Graphs show the percentage of surviving cells relative to the cells treated in the absence of zVAD-fmk. Experiments were carried out at least three times

Full size image

The kinase activity of RIPK1 promotes trimerization of MLKL

We addressed the mechanisms through which the kinase activity of RIPK1 promotes necroptosis. The phosphorylation of RIPK1 itself was significantly reduced in RIPK1K45A macrophages upon treatment with TNFα/zVAD but not after treatment with LPS/zVAD (Figures 2a–d). We performed mass spectrometric analysis of immuno-precipitated RIPK1 from BMDMs treated with LPS/zVAD or TNFα/zVAD for 3 h in order to identify phosphorylation sites within RIPK1. This analysis showed that Thr-235 and Ser-313 of RIPK1 are phosphorylated in both treatments (Supplementary Figure S1). Furthermore, Thr-235 phosphorylation was significantly attenuated in RIPK1K45A macrophages following LPS/zVAD treatment and abolished following TNFα/zVAD treatment. To our knowledge phosphorylation at Thr-235 of RIPK1 has not been reported before. Relative comparison of phosphorylation of Ser-313 was not influenced by K45 of RIPK1 (Supplementary Figure S1F). However, it is not clear whether the phosphorylation of Thr235 or Ser313 represents a 'necroptosis-specific' modification.

Figure 2
figure 2

K45 of RIPK1 promotes necroptotic signaling. (a, b, e–h) WT or RIPK1K45A macrophages were treated as indicated in figure panels with LPS (100 ng/ml) and TNFα (50 ng/ml) with or without zVAD-fmk (50 μM), and samples were evaluated by western blotting. Panels (a) and (b) shows western blotting at 3 h posttreatment. Panel (h) shows western blotting at 4 h posttreatment. (c, d, i–k) Densitometry of pooled western blottings from several experiments. ANOVA with Bonferroni post-test was used to measure statistical significance, *P<0.05; **P<0.01; error bars are S.E.M. Experiments were carried out at least three times

Full size image

Phosphorylation of RIPK3, a bench mark for necroptosis induction, was also poor in RIPK1K45A macrophages (Figures 2a and b). By dephosphorylating the total proteins from a cell lysate, we show that RIPK3 upper bands are the phosphorylated forms of RIPK3 (Figure 2b). RIPK1K45A macrophages displayed poor trimerization of MLKL in response to TNFα/zVAD (Figure 2e), LPS/zVAD (Figure 2f) or IFNβ/zVAD (Figure 2g) stimulation. These results clearly demonstrate that the K45 of RIPK1 has a key role in promoting necroptosis of macrophages. Concomitantly, RIPK1K45A macrophages displayed slightly increased activation of NFκβ, as assessed by the phosphorylation of p65 upon TNFα/zVAD treatments (Figures 2h and i). Increased activation of NFκβ in RIPK1K45A macrophages following TNFR1 ligation is an indication of enhanced cell survival signaling.23 In response to LPS/zVAD treatment, phospho-p65 was not modulated by RIPK1K45A (Figures 2h and j). As the cellular inhibitors of apoptosis proteins (cIAP1/2) participate in cell signaling following TNF-R engagement, we evaluated whether the levels of cIAPs were modulated in RIPK1K45A macrophages. Both WT and RIPK1K45A macrophages expressed similar levels of cIAPs, which is in line with previous work in MEFs24 (Figures 2h and k).

RIPK1 kinase activity modulates signaling downstream of TNF-R

As TNFα signaling following receptor engagement involves activation (phosphorylation) of STAT1,25, 26, 27 we measured pS727-STAT1 levels following LPS/zVAD or TNFα/zVAD treatment of macrophages. Phosphorylation of STAT1 was reduced in RIPK1K45A macrophages following LPS/zVAD treatment (Figures 3a and b). More interestingly, pS727-STAT1 levels after TNFα/zVAD were blunted in RIPK1K45A macrophages (Figures 3c and d) even as early as 1 h poststimulation. This differential phosphorylation of S727-STAT1 between WT and RIPK1K45A macrophages is not observed during non-necroptotic stimulations (Supplementary Figures S2A and B; treatment with LPS or TNFα in the absence of zVAD).

Figure 3
figure 3

RIPK1K45A impacts JNK activation, STAT1 phosphorylation at position S727 and chemokine production by macrophages. WT or RIPK1K45A macrophages were treated as indicated in the panels and tested for STAT1 and JNK phosphorylation (a, c, e) by western blotting analysis. (b, d, f) Densitometric analysis of the representative western blottings. Chemokine production by macrophages was measured (g) at various time intervals or (h) at 6 h poststimulation of cells. (i) mRNA levels of the chemokines measured in panels (g and h) were analyzed under same conditions. The mRNA levels were normalized to rpp30. ANOVA with Bonferroni post-test was used to measure statistical significance, **P<0.01; ***P<0.001; ****P<0.0001; NS, not significant; error bars are S.E.M. Experiments carried out at least three times

Full size image

Furthermore, TNF-R signaling also initiates a rapid and transient induction of c-Jun N-terminal kinase (JNK).28 In this regard, TNFα/zVAD treatment of RIPK1K45A macrophages resulted in diminished levels of p-JNK (Figures 3e and f). These results pinpoint to a mechanism wherein the phosphorylation by RIPK1 impacts TNFα signaling. Thus we measured cytokine or chemokine production following treatment of macrophages with TNFα or TNFα/zVAD. The levels of chemokines MIP-1α and RANTES were significantly induced in WT macrophages following stimulation of cells with TNFα alone (Figure 3g) or TNFα/zVAD (Figure 3h) for 6 h. In contrast, RIPK1K45A macrophages showed reduced expression of these chemokines. In addition to this, low mRNA levels of MIP-1α after TNFα/zVAD stimulation are indicative of a RIPK1-dependent transcriptional regulation of MIP-1α (Figure 3i). When these BMDMs were treated with LPS/zVAD for 6 h, the expression of IL-1α, TNFα and IL-10, but not IL-6, were reduced in RIPK1K45A macrophages (Supplementary Figure S2C).

In the absence of PAMPs, K45 of RIPK1 is necessary for auto-phosphorylation of RIPK1, which is inhibited by cIAPs

Treatment with a mimetic of the second mitochondrial activator of apoptosis (SMAC) induces regulated cell death of macrophages even in the absence of any PAMP stimulation, and this is induced by rapid degradation of cIAPs by SMAC.29 Stimulation of macrophages with the SMAC mimetic, birinapant (BP) resulted in a significant loss of cell viability, which was further enhanced by co-treatment with zVAD, and was rescued by treatment with Nec-1 (Figure 4a). Interestingly, RIPK1K45A macrophages displayed potent resistance to cell death in this model (Figure 4a). Similar results were noted when cells were stained with PI/Hoechst (Figure 4c and Supplementary Figure S3A), suggesting that the kinase activity RIPK1 has an even greater role in PAMP-independent necroptosis. The cell death of WT macrophages upon treatment with BP and zVAD was detectable as early as 2 h posttreatment (Figure 4b), while BP alone failed to cause cell death at that time point. BP induced cell death was dependent on TNF-R signaling (Figure 4d). Treatment of cells with the SMAC mimetic (SM164 or BP) resulted in rapid degradation of cIAPs as expected, but the phosphorylation of RIPK1 was totally abrogated in RIPK1K45A macrophages (Figures 4e and f). Treatment of WT macrophages with SMAC mimetics resulted in a significant reduction of RIPK1 levels, which was significantly inhibited in RIPK1K45A macrophages (Figure 4g). Furthermore, trimerization of MLKL was also abrogated in RIPK1K45A macrophages treated with SMAC+zVAD (Figure 4h). We then tested whether the death of macrophages induced by SMAC mimetic results in the activation of caspase 3 and caspase 8. The cleaved forms of caspase 3 (p17 and p20) were detectable by western blotting in SMAC mimetic-treated WT macrophages but not in RIPK1K45A macrophages (Figure 4i). Although the cleaved form (p18) of caspase 8 was undetectable by western blotting (Figure 4i), caspase 8 activity was readily detectable in a luminescence assay, which was inhibited by caspase 8-specific inhibitor (Figure 4j) or by N-benzyloxycarbonyl-Val-Ala-Asp(O-Me) fluoromethyl ketone (zVAD-fmk; Supplementary Figure S3B). Treatment of macrophages with the SMAC mimetic resulted in significant induction of caspase 8 activity relative to control in both genotypes. However, caspase 8 activity was induced stronger in WT than in the RIPK1K45A macrophages (Figure 4j, Supplementary Figure S3B). In the absence of zVAD treatment, SMAC mimetic induced the activation of caspase 3, caspase 8 and cell death of WT but not in RIPK1K45A macrophages.

Figure 4
figure 4

Lysine-45 of RIPK1 is necessary for RIPK1 auto-phosphorylation following depletion of cIAPs. Cell survival of macrophages was measured by MTT assay after (a) 24 h and (b) 2 or 4 h posttreatment of cells with SMAC mimetic (BP, 10 μM) with or without zVAD-fmk (50 μM) or Nec-1 (10 μM). Graphs show the percentage of surviving cells relative to the corresponding vehicle control. These graphs are representative of three biological replicates each carried out in triplicate. (c) Cell death of macrophages after 24 h treatment with BP as measured microscopically by PI/Hoechst staining. (d) WT or TNF-R1,2−/− macrophages were treated as in panel (a) and assessed for survival by MTT assay. (e, h, i) WT or RIPK1K45A macrophages were prepared for western blotting at the indicated times (i – 4 h) posttreatment with BP (10 μM) with or without zVAD-fmk (50 μM) or SMAC 164 (5 μM). (f, g) Densitometry data of pooled three western blotting experiments as in panel (e) at 4-h treatment was used to assess p-RIPK1/RIPK1 ratio (f) or total RIPK1 levels (g). (j) Caspase 8 activity was measured with the Luciferase Kit from Promega at 4 h posttreatment; C8I, caspase 8 inhibitor. ANOVA with Bonferroni post-test was used to measure statistical significance, **P<0.01; ***P<0.001; ****P<0.0001; error bars are S.E.M. NS, not significant. Experiments were repeated at least three times

Full size image

RIPK1K45A mice display enhanced survival against endotoxin shock and poor control of Salmonella typhimurium (ST) in vivo

RIPK1K45A mice displayed enhanced survival following endotoxin shock in comparison to WT controls (Figures 5a and b). IL-1β and TNFα levels in the sera of endotoxin-injected mice were lower in RIPK1K45A mice (Figure 5c), and the lower level of TNFα may be responsible for decreased expression of MIP-1α (Supplementary Fig. S4). Dysregulation in cytokine response was also detected in macrophages stimulated in vitro with LPS (Figure 5d), which correlated with reduction in the mRNA levels of IL-1β and TNFα (Figure 5e). These results prompted us to test the impact of K45A mutation during challenge with a virulent intracellular bacterium. We challenged mice with low dose of ST and measured the bacterial burden in the spleens of infected mice at day 5 postintravenous infection. Our results show that RIPK1K45A mice harbor higher burden of ST (Figure 6a and Supplementary Figures S5A and B) that is associated with increased splenomegaly (Figure 6b). The numbers of spleen cells were also increased in infected RIPK1K45A mice (Figure 6c). RIPK1K45A macrophages displayed reduced level of reactive oxygen species (ROS) following ST infection in vitro, as measured by DCF staining (Figure 6d) and reduced release of glucose-6-phosphate dehydrogenase (G6PD), indicative of reduced cell death (Figure 6e). RIPK1K45A macrophages infected with ST in vitro expressed reduced IL-1β in comparison to WT macrophages (Figure 6f). Upon infection with ST, caspase 8 activity was reduced, and this reduction was even more pronounced in RIPK1K45A macrophages (Figure 6g). Taken together, our results reveal the impact of the kinase domain of RIPK1 on necroptosis and cytokine signaling during inflammatory (PAMPs, cytokines) and non-inflammatory stimuli (SMAC).

Figure 5
figure 5

The kinase activity of RIPK1 promotes susceptibility to endotoxic shock. (a) Enhanced survival of RIPK1K45A mice after intraperitoneal injection with LPS (50 mg/kg). Graph represents pooled data of three separate experiments. Mantel–Cox test was used to evaluate the statistical significance. (b) RIPK1K45A mice maintain higher body temperature than the WT mice following LPS injection as described in panel (a). ANOVA was used to evaluate statistical significance. (c) Cytokine levels in sera of WT or RIPK1K45A mice after 5 h of intraperitoneal (i.p.) injection with LPS (50 mg/kg). Graphs are pooled data of three separate experiments. Statistical significance was measured by t-test. (d) Cytokine production by macrophages at 24 h poststimulation (in vitro) by LPS (100 ng/ml). (e) mRNA levels of cytokines in macrophages at 6 h posttreatment (normalized to rpp30). *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001, error bars are S.E.M.

Full size image
Figure 6
figure 6

Kinase activity of RIPK1 promotes inflammatory response against ST. (a) ST burden in the spleen of WT or RIPK1K45A mice was evaluated at day 5 postintravenous injection of 200 bacteria. Graph represents pooled data of three separate experiments. Mann–Whitney test was used to measure statistical significance, (b) image of the spleens at day 5 postinfection, (c) number of total splenocytes after infection as in panel (a). (d) ROS levels were measured at 24 h postinfection of macrophages with ST (10 MOI) in vitro. (e) Glucose 6-phosphate dehydrogenase (G6PD) release after ST infection at 24 h postinfection. The plot shown is representative of two biological repeats. (f) Cytokine levels were measured by ELISA at 24 h postinfection (in vitro) of macrophages with ST. Data are pooled from three separate experiments, all carried out in triplicate. The IL-1β cytokine levels from panel (f) were also represented as fold-reduction relative to arithmetic mean of the cytokine levels in WT cells. Line at 1 is the arithmetic mean of WT cell cytokine levels. (g) Caspase 8 activity after ST infection at 3 h postinfection, measured using a Luciferase Kit from Promega. The plot shown is representative of two biological repeats. *P<0.05; **P<0.01; ***P<0.001, error bars are S.E.M. RFU, relative fluorescence unit

Full size image

Discussion

RIPK1 interacts with various proteins such as cIAPs and TRAFs that are associated with the TNF-R.9 Dissociation of RIPK1 from this complex results in interaction of RIPK1 with caspase 8, FADD and RIPK3, which promotes cell death signaling by apoptosis or necrosis.9 RIPK1-deficient mice die within the first week of birth5 as a result of uncontrolled activation of necrosis and apoptosis mediated by RIPK3 and caspase 8,6, 30 suggesting that RIPK1, RIPK3 and caspase 8 regulate the overactivation of each other. As RIPK3-deficient mice are viable,31 this suggests that RIPK1 has additional roles besides necroptosis. We have evaluated the role of the kinase region of RIPK1 in macrophages and its consequent impact in vivo. Our results indicate that RIPK1 kinase-dead (RIPK1K45A) mice do not have any impairment in survival; however, they are highly resistant to necroptosis induced by various stimuli, and they display reduced TNF-R signaling, caspase 8 activation and IL-1β expression, which results in significant protection against endotoxin shock and compromised control of an infection with a virulent intracellular bacterium.

Our results indicate that RIPK1K45A macrophages are significantly, but not completely, resistant to necroptotic stimulation by LPS. In contrast, RIPK1K45A macrophages displayed a greater resistance to necroptosis upon stimulation by TNFα or IFNβ. As both RIPK1 and RIPK3 can promote necroptosis as well apoptosis,22, 32, 33 reduced oligomerization of MLKL is indicative of reduced necroptosis in RIPK1K45A macrophages because oligomerization of MLKL has been considered to be mainly associated with necroptosis.34 We show that MLKL trimer formation is dependent on the RIPK1 kinase activity in all necroptotic stimuli. Here again, we noted differences in the extent of inhibition of signaling in RIPK1K45A macrophages that was dependent on the stimulus used. There was total inhibition of MLKL oligomerization following treatment with SMAC/zVAD or TNFα/zVAD in comparison to stimulation with LPS/zVAD where MLKL oligomerization was reduced. This also correlated well with the extent of cell death. Interestingly, K45-dependent auto-phosphorylation of RIPK1 appears to be the main mechanism of RIPK1 phosphorylation when macrophages are treated with SMAC or TNFα/zVAD. However, when cells were treated with LPS, RIPK1 phosphorylation did not appear to be impacted by K45 of RIPK1. This is in line with the notion that engagement of TLR4 may lead to multiple and redundant activation of signaling proteins,35 notably the TAK1 (TGFβ-activated kinase 1). Activated TAK1 can activate RIPK3–MLKL axis independently of RIPK1 leading to regulated necrotic cell death.36

Interestingly, Nec-1s treatment results in a slight but significant enhancement of survival of RIPK1K45A macrophages following LPS, TNFα or IFNβ treatment (Figure 1a). As Nec-1s, in contrast to Nec-1, specifically inhibits the kinase activity of RIPK1,21 these results suggest that Nec-1s may inactivate other regions of RIPK1, which may impact necroptosis further.

When macrophages were treated with TNFα/zVAD, the kinase activity of RIPK1 seemed to inhibit the activation of NFκβ. Although RIPK1 was shown to previously promote NFκβ activation5, 37 in MEFs, our results indicate that the kinase activity (K45) of RIPK1 may negatively regulate NFκβ signaling in macrophages. It is not clear whether these differences are related to the cell type. As RIPK1 participates in various signaling complexes following TNF-R engagement, it is conceivable that the transitioning of RIPK1 between TNF-R1 complex I and complex 2b may be delayed in RIPK1K45A macrophages, leading to enhanced signaling through complex I for NFκβ.

Interestingly, the K45A mutation of RIPK1 significantly impacted cytokine signaling as revealed by poor p-STAT1 phosphorylation at S727 during necroptotic stimulation, which in turn may reflect the reduction in caspase activity. Phosphorylation of STAT1 at Y701 is required to decouple the STAT1 from TNF-R,26 but phosphorylation at S727 is required for maximal STAT1-dependent gene transcription.25 Our results indicate an impairment in chemokine expression in RIPK1 kinase-dead macrophages, which correlated with S727-STAT1 phosphorylation. Signaling through the TNF-R and IFNA-R is mediated through STAT1 phosphorylation.38 However, the phosphorylation of STAT1 is blunted early on (1 h upon stimulation) in RIPK1K45A macrophages pointing to an early impairment in phosphorylation most likely localized at the TNF–receptor complex. We also noted that the kinase activity of RIPK1 promotes rapid induction of JNK following TNFα/zVAD treatment. JNK has been shown to promote ROS production,39 and reduction in ROS levels may impact necroptosis.2 These results suggest that K45 of RIPK1 has a key role in promoting chemokine signaling, which is indicative of a mechanism through which RIPK1 promotes inflammatory response.

We also noted that, in the absence of cIAPs, the stability of RIPK1 is dependent on the kinase activity (K45) of RIPK1. It is therefore likely that the kinase activity of RIPK1 inhibits signaling that promotes RIPK1 degradation. It is also possible that the lysine-45 of RIPK1 is itself the target for ubiquitination leading to degradation, hence a mutation of this amino acid impacts RIPK1 stability. Furthermore, the cell survival after cIAP depletion is dependent on RIPK1 kinase activity and requires TNF-R signaling, similarly to previous reports.40

Recently, in a mouse model of inflammation in the skin that is induced by SHARPIN-deficiency, RIPK1 kinase-dead mice displayed significant rescue,20 whereas in the models of gut inflammation the kinase activity of RIPK1 did not promote severe colitis.24, 41 Further, it was shown that RIPK1 and RIPK3 are found in a complex with caspase 8 and caspase 1 where the kinase activity of RIPK1 is dispensable.42, 43 However, in vivo deficiency of RIPK1 kinase activity is shown to be deleterious for the host after Vaccinia virus infection.40 Our results highlight the impact of the kinase domain of RIPK1 in two models of inflammation. RIPK1K45A mice display enhanced survival against endotoxin shock, yet they are highly susceptibility to a challenge with ST. Enhanced survival of RIPK1K45A mice following endotoxin shock is not as strong as seen in mice with disrupted inflammasome signaling,44, 45, 46 suggesting that RIPK1K45A may not modulate the dominant pathway of inflammasome signaling, which is mediated by caspase 11.47 However, the impact of RIPK1K45A mice on the control of ST appears to be similar to what has been published in inflammasome-deficient mice.48 These results indicate that the impact of the kinase activity of RIPK1 in promoting inflammation is dependent on the experimental model. Interestingly, in macrophages, our in vitro results also point to a transcriptional regulation of cytokine activation that is dependent on RIPK1 kinase activity. Although the kinase activity promotes the production of IL-1β, TNFα and IL-10, it impedes IL6 production, similarly to a previous report on IL6 production in MEFs.37 Intriguingly, we noted reduced levels of only IL-1β and TNFα in RIPK1K45A mice.

RIPK1K45A mice were highly susceptible to infection by ST, which correlated with reduced expression of IL-1β, ROS and reduced cell death of RIPK1K45A macrophages. The 16-fold enhanced burden of ST in RIPK1K45A mice is similar in magnitude to what has been shown previously in mice lacking inflammasome signaling.48 Although the processing of the IL-1 cytokine family has been shown to promote protection against pathogens,48 cell death has also been to limit pathogen spread through release of intracellular bacteria for uptake and killing by neutrophils, independently of IL-1.49 Although caspase 1/11 signaling promotes canonical inflammasome signaling, caspase 8 and other members of the ripoptosome complex have also been shown to promote inflammasome activation.50, 51, 52, 53, 54 Our results show that caspase 8 activity was reduced in RIPK1K45A macrophages, which may limit IL-1β processing in RIPK1K45A macrophages. We observed that the activity of caspase 8 was reduced in macrophages following infection with ST. Interestingly, caspase 1, which is activated during infection with ST, has been previously shown to limit the processing of various caspases.55

Although the molecular mechanisms that govern the inflammasome and necrosome activation are quite distinct, the end result in both cases is membrane rupture and release of damage-associated molecular patterns, leading to excessive inflammation.9, 56 Although RIPK3 is a key component of the necrosome complex, and we have previously shown that RIPK3 is activated following ST infection,57 RIPK3 has also been shown to promote inflammasome signaling.11, 50, 51, 53 More recently, MLKL, another member of the necrosome, has also been shown to promote inflammasome signaling,58 indicating that there is significant overlap in inflammasome and necrosome signaling. We have previously shown that WT and RipK3−/− mice harbor similar bacterial burden at day 5 postinfection with ST,57 whereas, in our current study, the bacterial burden in RIPK1K45A is much higher than in WT mice (Figure 6a). It is quite likely that this is related to the participation of RIPK1 in additional signaling platforms such as the Ripoptosome and the impact of RIPK1 that we have shown on TNFα signaling. Our results indicate that the kinase activity of RIPK1 promotes inflammatory cell death and processing of various inflammatory cytokines. Expression of inflammatory cytokines is akin to a double-edged sword wherein the expression of these cytokines promotes protection against pathogens but promotes lethality during sterile injury. Impairment in the processing and expression of various inflammatory cytokines by RIPK1K45A macrophages would result in poor control of virulent pathogens such as ST but better protection against endotoxin shock. It is also conceivable that RIPK1K45A mice show better protection in other models of sterile inflammation. Delineation of the mechanisms through which the kinase region of RIPK1 influences host outcome would lead to the development of therapeutics against inflammatory diseases.

Materials and Methods

Mice

RIPK1K45A mice20 and WT litter mates were maintained in our animal facility. RIPK3−/− were a kind gift from Dr Vishva Dixit (Genentech, San Francisco, CA, USA).31 TNF-R1,2−/− mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Experiments were performed in accordance with the Canadian Council on Animal Care guidelines and the Ethics Board and/or the Animal Care Committee at the University of Ottawa, Ottawa, ON, Canada.

In vivo endotoxin shock model

Mice (litter mates only) were age and sex matched, their weight was measured and the average mass of all mice per experiment was used for calculating the concentration of intraperitoneally injected LPS (50 mg/kg) (LPS from Escherichia coli 055:B5, L4005, Sigma, St. Louis, MO, USA). For survival experiments, mice were killed when the rectal temperature would drop <26 °C. Sera were prepared by standard procedures.

Macrophages

Whole bone marrow was differentiated to macrophages (BMDM) in the presence of macrophage colony-stimulating factor (416- ML, R&D, Minneapolis, MN, USA). The R8 media (R8 – RPMI 1640 with 8% fetal bovine serum and 50 μM β-Mercaptoethanol) was changed every second day, and the floater cells were discarded. The macrophages were used for experiments between day 7 and day 9 and were plated at 106 cells/ml. Reagents used were: LPS (from E. coli 0111:B4, L3024, Sigma), TNFα (410-MT, R&D), IFNβ (12400-1, PBL Interferon, Piscataway, NJ, USA), zVAD-fmk (627610, Millipore, Billerica, MA, USA), BP (S7015, Selleckchem, Cedarlane, Burlington, ON, Canada), SMAC 164 (gift from Shaomeng Wang), Nec-1 (9037, Sigma), and Nec1s (gift from Dr. Katey Rayner, University of Ottawa). MTT (1-(4,5-dimethylthiazol-2-yl)-3,5-diphenylformazan) assay was performed as described earlier.14

Cytokine and chemokine measurements

Cytokines were measured using commercial kits: IL-1β (559603, BD OptEIA, San Diego, CA, USA), IL-12 (p70) (555256, BD OptEIA), IL-10 (555252, BD OptEIA), IL-6 (14-7061, 13-7062, eBioscience), IL-1α (16-7011, 13-7111, eBioscience, San Diego, CA, USA), and CBA from BD was used for TNFα, (558299; BD, San Diego, CA, USA). Chemokines were measured by CBA from BD (MIP-1α, 558449, RANTES, 558345).

Western blotting

BMDMs at the end point of treatment were washed twice with PBS and lysed in RIPA buffer in the presence of protease inhibitors (04693159001, Roche, Mannheim, Germany) and cocktail of phosphatase inhibitors (P5726, Sigma Aldrich) while on ice. Total protein per sample was estimated by the Micro BCA Kit (Thermo Scientific Pierce, cat. number: 23235, Rockford, IL, USA) and continued with the standard western blotting procedure. The samples were mixed with Laemmli loading buffer and heat denaturated at 95 °C for 5 min and then separated by Laemmli-Discontinuous SDS-PAGE in SDS-TRIS-Glycine buffer, pH 8.8. The protein transfer to a PVDF membrane was carried out in TRIS-Glycine-Methanol buffer. With exception for samples prepared for detecting RIPK1 and RIPK3, where BMDMs were directly lysed in 1 × Laemmli loading buffer while on ice without previous washing steps and continued as above. In addition, cells prepared for MLKL western blotting after the washing steps were lysed in M2 buffer (20 mM Tris, pH 7, 0.5% NP40, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA and 50 mM NaF) supplemented with protease and phosphatase inhibitors as above. The samples were mixed in sample buffer (161-0737, Bio-Rad, Hercules, CA, USA) in non-reducing condition (without β-Mercaptoethanol). Samples lysed in M2 buffer were separated by same procedures as above for RIPA-prepared samples. Antibodies used were: anti-p65 (8242, Cell Signaling, Danvers, MA, USA), anti-SAPK/JNK (9258, Cell Signaling), anti-phospho-SAPK/JNK (4668, Cell Signaling), anti-p-p65 (3033, Cell Signaling), anti-caspase 3 (sc-7148, Santa Cruz, Dallas, TX, USA), anti-RIPK1 (610458, BD), anti-RIPK3 (2283, ProSci Inc., Poway, CA, USA), anti-actin (sc-81178, Santa Cruz), anti-STAT1 (9172, Cell Signaling), anti-p-STAT1-S727 (9167, Cell Signaling), anti-caspase 8 (ALX-804-447-C100, Enzo, Farmingdale, NY, USA), anti-MLKL (MABC604, EMD Millipore, Billerica, MA, USA) and anti-cIAP1/2 (CY-P1041, CycLex, Ina, Japan). Densitometric analysis was performed with the GE Healthcare ImageQuantTL software (Piscataway, NJ, USA). Dephosphorylation of the total protein was performed with 50U of Alkaline Phosphatase, Calf Intestinal – CIP, (M0290, NEB, Ipswich, MA, USA) in RIPA-EDTA free, supplemented with protease inhibitors but not phosphatase inhibitors, and 1 × CutSmart NEB buffer.

Immunoprecipitation and mass spectrometry

BMDMs were treated for 3 h with LPS/zVAD or TNFα/zVAD before proceeding with immunoprecipitation. Anti-RIPK1 (610458, BD) antibody was coupled to Dynabeads (14321D, Life Technologies, Carlsbad, CA, USA) per the manufacturer's instructions. The elute of the immunoprecipitation reaction was analyzed by mass spectrometry (SPARC BioCentre, The Hospital for Sick Children, Toronto, ON, Canada).

Caspase 8 activity assay

The Caspase-Glo 8 Luminescence Kit from Promega (G8200, Madison, WI, USA) was used to assess the caspase 8 activity. The BMDMs were plated in 96-well plate at 106 cells/ml and handled as indicated in the corresponding figures, caspase 8 inhibitor (C8I) – (zIETD-fmk, 1064-20C, BioVision, Milpitas, CA, USA).

Salmonella infection

ST SL1344 (or ST) was grown in BHI broth to OD600nm 0.8–0.9 and 1 ml aliquots were frozen at −80 °C in 20% glycerol–BHI broth. Colony-forming units (CFU) were assessed and used to start infection at the desired multiplicity. For BMDM infection, first the bacteria were washed twice in a room temperature R8 media, then added to the 96- or 24-well plated macrophages and spun for 10 min at 400 × g at room temperature and placed at 37 °C in a humidified CO2 incubator. After 45 min, the extracellular bacteria were washed 3 times for 96-well plates or 5 times for 24-well plates with room temperature R8 media supplemented with gentamycin at 50 μg/ml. After additional 2 h of infection, the macrophages were washed again in R8 with gentamycin at 10 μg/ml; this time the infection was carried on until the desired end point. For in vivo tail vein infection, the frozen stock was thawed and washed in ice-cold 0.89% NaCl solution, and mice were infected with 200 CFU via the lateral tail vein. The mice were age and sex matched. Release of G6PD following cell death was measured by Vybrant Cytotoxicity Assay Kit (V-23111, Invitrogen, Molecular Probes, Eugene, OR, USA) as recommended by the manufacturer. For caspase 8 activity measurements, where appropriate the cells were pretreated for 30 min with 20 μM zVAD or DMSO (as control), after each washing step zVAD or DMSO was replenished.

Quantitative RT-PCR

Quantitative RT-PCR was performed as described elsewhere.14 Briefly, total RNA was isolated using TRIzol reagent (15596-026, Life Technologies). Isolated RNA was reverse transcribed into cDNA with SuperScript IV First-Strand Synthesis System (18091050, Invitrogen) per the manufacturer's instruction. The cDNA was analyzed with SYBR Green (Life Technologies) method performed on an Applied Biosystems 7500 quantitative RT-PCR system (Foster City, CA, USA). The primers used for qPCR are as follows: Rpp30 FW: 5′-TGACGTGGCAAACTTAGGACT-3′; Rpp30 REV: 5′-ATGGCCGTGGTTTCTTCACT-3′ (for Ribonuclease P/MRP 30 subunit); RANTES FW: 5′-CTCACCATATGGCTCGGAC-3′; RANTES REV: 5′-CTTGGCGGTTCCTTCGAGT-3′; MIP-1α FW: 5′-CATATGGAGCTGACACCCCG-3′; MIP-1α REV: 5′-GTCAGGAAAATGACACCTGGC-3′; IL-10 FW: 5′-GGTTGCCAAGCCTTATCGGA-3′; IL-10 REV: 5′-GGGGAGAAATCGATGACAGC-3′; IL-1b FW: 5′-TGCCACCTTTTGACAGTGATG-3′; IL-1b REV: 5′-TGATGTGCTGCTGCGAGATT-3′; IL-6 FW: 5′-CACGGCCTTCCCTACTTCAC-3′; IL-6 REV: 5′-TGCAAGTGCATCATCGTTGT-3′; TNFα FW: 5′-ACGTCGTAGCAAACCACCAA-3′; and TNFα REV: 5′-ATAGCAAATCGGCTGACGGT-3′.

Reactive oxygen species

ROS levels in macrophages were measured on a plate reader following infection with ST as described above. Cells were stained with 2',7'-dichlorodihydrofluorescein diacetate (H2DCFDA; D-399, Molecular Probes–Life Technologies) and fluorescence was measured on a microplate reader.