Critical role for calcium mobilization in activation of the NLRP3 inflammasome

Abstract

The NLRP3 (nucleotide-binding domain, leucine-rich-repeat-containing family, pyrin domain-containing 3) inflammasome mediates production of inflammatory mediators, such as IL-1β and IL-18, and as such is implicated in a variety of inflammatory processes, including infection, sepsis, autoinflammatory diseases, and metabolic diseases. The proximal steps in NLRP3 inflammasome activation are not well understood. Here we elucidate a critical role for Ca²⁺ mobilization in activation of the NLRP3 inflammasome by multiple stimuli. We demonstrate that blocking Ca²⁺ mobilization inhibits assembly and activation of the NLRP3 inflammasome complex, and that during ATP stimulation Ca²⁺ signaling is pivotal in promoting mitochondrial damage. C/EPB homologous protein, a transcription factor that can modulate Ca²⁺ release from the endoplasmic reticulum, amplifies NLRP3 inflammasome activation, thus linking endoplasmic reticulum stress to activation of the NLRP3 inflammasome. Our findings support a model for NLRP3 inflammasome activation by Ca²⁺-mediated mitochondrial damage.

Inflammation must be tightly regulated to meet the demands of host defense and stress adaptation while limiting immunopathology and the detrimental effects of chronic inflammation. Sensors of the innate immune system, such as pattern recognition receptors, couple detection of proinflammatory triggers to induction of inflammation in both adaptive and maladaptive settings, so the regulation of this interaction is absolutely critical in control of inflammation (1). A particularly intriguing “sensor” is the NLRP3 (nucleotide-binding domain, leucine-rich-repeat-containing family, pyrin domain-containing 3) inflammasome complex, which detects cellular stress in a variety of infectious and “sterile” settings and links this recognition to the induction of a critical inflammatory pathway (2–5).

The NLRP3 inflammasome is a cytoplasmic complex consisting of the regulatory subunit NLRP3, the adaptor ASC, and the effector subunit caspase-1. Stimulus dependent assembly of the complex activates the proteolytic activity of caspase-1, which is required for the processing and secretion of the inflammatory cytokines IL-1β and IL-18 (2–5). In recent years, many different molecules have been shown to activate the NLRP3 inflammasome complex. These molecules include: extracellular ATP released from dying cells (6); alum, the adjuvant activity in human vaccines (7); nigericin, a bacterial toxin with potassium ionophore activity (6); and cholesterol crystals (8), which contribute to the development of atherosclerotic plaques. Such findings underscore the critical role of the NLRP3 inflammasome in inducing inflammation in a variety of diseases.

How these diverse stimuli activate the NLRP3 inflammasome is not clear, but the prevailing view is that they generate some signal of cellular stress that is often referred to as signal 2. Although signal 1 is important for transcriptional up-regulation of the NLRP3 subunit and is typically provided by signaling through Toll-like receptors (9), three cellular processes have been suggested to provide signal 2 (2, 3, 10). The first is K⁺ efflux, which has been linked to most/all NLRP3 inflammasome activators, but just how K⁺ efflux can activate the inflammasome in a specific manner is not obvious; moreover, modulating the ionic milieu in other ways also affects NLRP3 inflammasome activation (11, 12). Second, crystals trigger phagolysosomal membrane damage/rupture (13); but how this is linked to the proximal steps in NLRP3 inflammasome activation is not clear, nor is its relevance for stimuli that are not phagocytosed. Finally, other studies implicate mitochondrial damage, including increased mitochondrial reactive oxygen species (mROS) production, loss of membrane potential (ΔΨ), and release of mtDNA into the cytosol (14–16). Importantly, how NLRP3 inflammasome activators trigger mitochondrial damage has not been characterized.

Ca²⁺ mobilization controls diverse cellular processes, including proliferation and differentiation, transcription, cellular metabolism, and cell death (17). Such Ca²⁺ mobilization is made possible by maintaining low levels of cytoplasmic Ca²⁺ at basal state, so that signal-dependent influx of Ca²⁺ from the extracellular space and intracellular Ca²⁺ stores (e.g., endoplasmic reticulum, ER) leads to a rapid increase in cytoplasmic Ca²⁺ levels, activation of Ca²⁺ binding proteins, and induction of various cellular responses. A key player in Ca²⁺ signaling is the mitochondria, which take up Ca²⁺ from the ER or the extracellular space to regulate spatiotemporal patterns of Ca²⁺ signaling (18–20). However, excessive or sustained mitochondrial Ca²⁺ uptake can lead to mitochondrial damage and cell death (21–23).

In this article, we show that activation of the NLRP3 inflammasome requires Ca²⁺ signaling. Several NLRP3 inflammasome activators mobilize Ca²⁺, disruption of which inhibits NLRP3 inflammasome activation. During ATP stimulation, the crucial role of Ca²⁺ mobilization is in inducing mitochondrial damage. C/EPB homologous protein (CHOP), a protein known to regulate Ca²⁺ release from the ER during ER stress, amplifies NLRP3 inflammasome activation.

Results

Ca²⁺ Signaling Is Critical for NLRP3 Inflammasome Activation by Extracellular ATP.

Extracellular ATP activates the NLRP3 inflammasome through P2X7R (24), a ligand-gated ion channel of the purinergic receptor family. Using Ca²⁺-sensitive fluorometric probes and analysis by time-lapse microscopy to measure cytosolic Ca²⁺ levels, we confirmed previous findings (25) that Ca²⁺ is mobilized from both extracellular and intracellular pools during ATP stimulation (Fig. 1A). To determine which pool contributes to NLRP3 inflammasome activation, we used experimental manipulations common in Ca²⁺ signaling research: treatment with thapsigargin (Tg), an inhibitor of the SERCA (sarcoplasmic/ER Ca2⁺-ATPase) pump to deplete ER Ca²⁺ stores, or incubation in Ca²⁺ free media, to block extracellular Ca²⁺ entry. We found that brief Tg pretreatment (Fig. 1B) and incubation in Ca²⁺-free media (Fig. 1C) both attenuate NLRP3 inflammasome activation by ATP stimulation (as measured by caspase-1 processing to the active p10 subunit), IL-1β processing, and release of processed caspase-1 and IL-1β into culture supernatants. This finding suggests that ER Ca²⁺ release and influx of extracellular Ca²⁺ are both required to activate the NLRP3 inflammasome.

Fig. 1.

Extracellular ATP mobilizes Ca²⁺ to activate the NLRP3 inflammasome. (A) LPS-primed BMDMs were loaded with Fura-2 followed by stimulation with 1 mM ATP and analysis of Ca²⁺ flux by time-lapse microscopy. Ca²⁺ is released from intracellular stores during the initial stimulation in Ca²⁺-free buffer, and exchange to 1 mM Ca²⁺-containing buffer permits extracellular Ca²⁺ influx. (B and C) LPS-primed BMDMs were treated or not with Tg for 30 min to deplete ER Ca²⁺ (B) or switched to Ca²⁺ free (−) or Ca²⁺-containing (+) media immediately before ATP stimulation (C). NLRP3 inflammasome activation was assessed by Western blotting of lysates and supernatants (supes) as well as IL-1β ELISA. (D–F) BMDMs were pretreated with XeC, U73122, or 2-APB, followed by ATP stimulation and analysis of Ca²⁺ flux (D) and NLRP3 inflammasome activation (E and F).

We sought to further characterize the mechanisms that regulate Ca²⁺ mobilization during ATP stimulation. Phospholipase C (PLC) proteins are key players in Ca²⁺ signaling through their production of the second messengers inositol triphosphate (IP₃) and diacylglycerol (17). Using macrophages expressing a PLC reporter (consisting of the PLC pleckstrin homology domain fused to GFP) (26), we showed that the protein translocates from the plasma membrane to the cytosol during ATP stimulation, indicative of PLC activation (Fig. S1A). Importantly ATP-induced Ca²⁺ flux could be blocked by the PLC inhibitor U73122 (Fig. 1D). PLC activation can be coupled to ER Ca²⁺ release through the IP₃ receptor (IP₃R), and indeed treatment with the IP₃R inhibitor XeC attenuates ATP-mediated Ca²⁺ signaling (Fig. 1D). Finally, ER Ca²⁺ release can trigger extracellular Ca²⁺ influx through store-operated Ca²⁺ entry (SOCE) (27). The chemical 2-aminoethoxydiphenyl borate (2-APB), which blocks SOCE at multiple steps (including oligomerization and translocation of the STIM proteins that sense ER Ca²⁺ depletion, as well as the activity of the Ca²⁺ channels ORAI), also blocks Ca²⁺ mobilization (Fig. 1D). Inhibition of Ca²⁺ flux was strongest with 2-APB and weaker with XeC and U73122, suggesting that other Ca²⁺-releasing second messengers or ER Ca²⁺ release mechanisms may be triggered by P2X7R.

These findings indicate that during ATP stimulation, Ca²⁺ mobilization may be regulated by PLC proteins, IP₃R-mediated Ca²⁺ release, and SOCE. Next, we addressed the role of such Ca²⁺ mobilization in activation of the NLRP3 inflammasome, using multiple read-outs of NLRP3 inflammasome activity. We found that XeC, U73122, and 2-APB blocked ATP-induced caspase-1 and IL-1β processing and release of processed caspase-1 and IL-1β into culture supernatants (Fig. 1 E and F). As with Ca²⁺ flux (Fig. 1D), inhibition was strongest with 2-APB and weakest with XeC. Finally, Ca²⁺ signaling inhibitors blocked assembly of the NLRP3 inflammasome complex as visualized by immunofluorescence (Fig. S2). Although many ATP-stimulated cells (>25%) accumulate foci corresponding to activated NLRP3 inflammasome complexes, cotreatment with Ca²⁺ signaling inhibitors restricts formation of such foci (<5%). This finding demonstrates that Ca²⁺ signaling regulates NLRP3 inflammasome activation at an upstream step, proximal to complex assembly.

Of note, XeC, U73122, and 2-APB had no effect on the transcriptional induction of NLRP3 (signal 1) or pro–IL-1β by LPS (Fig. S1B), indicating that Ca²⁺ mobilization does not regulate expression of the NLRP3 inflammasome or LPS signaling. We also rule out a role for the inhibitors in modulating cell viability (Fig. S1C). In conclusion, Ca²⁺ signaling—probably regulated by PLC, ER Ca²⁺ release through the IP₃R, and SOCE—is important for NLRP3 inflammasome activation by extracellular ATP.

Ca²⁺ Signaling May Have a Central Role in NLRP3 Inflammasome Activation.

Next, we asked if Ca²⁺ signaling plays a role in NLRP3 inflammasome activation by other stimuli. We found that nigericin stimulation led to Ca²⁺ mobilization (Fig. 2A) (28), as did monosodium uric acid crystals (MSU) (Fig. 2B). We were initially not successful in detecting alum-induced Ca²⁺ flux, but because this effort could be confounded by high rates of Ca²⁺ efflux from the cytosol, we took an alternative experimental strategy (29) and asked if alum pretreatment could block well-established mechanisms of ER Ca²⁺ release. Indeed, alum pretreatment inhibited Ca²⁺ mobilization during P2Y receptor activation, strongly suggesting that it triggered ER Ca²⁺ release and depletion of ER Ca²⁺ stores (Fig. S3A). Thus, multiple stimuli that activate the NLRP3 inflammasome induce Ca²⁺ signaling.

Fig. 2.

Ca²⁺ signaling may be a common mechanism for NLRP3 inflammasome activation. (A and B) LPS-primed BMDMs stimulated with nigericin (A) or MSU (B) were analyzed by Ca²⁺ imaging. In B, buffer exchange to remove (0 Ca²⁺) and replace Ca²⁺ (1 mM Ca²⁺) indicates contribution of store-operated Ca²⁺ entry to Ca²⁺ flux. (C–F) BMDMs stimulated as indicated were examined for NLRP3 inflammasome activation.

Next, we asked if such Ca²⁺ mobilization plays a role in NLRP3 inflammasome activation. Indeed XeC, U73122, and 2-APB blocked caspase-1 processing and IL-1β production during stimulation with nigericin, MSU, and alum (Fig. 2 C–F and Fig. S3 B and C). Incubation in Ca²⁺-free media also attenuated NLRP3 inflammasome activation by MSU and alum stimulation (Fig. S4). In contrast, activation of the NLRC4 inflammasome by FliC was attenuated by the PLC inhibitor U73122 but not by the IP₃R or SOCE inhibitors (Fig. S5). This result may reflect “noncanonical” functions of PLC (e.g., nuclear PLC) in settings where the roles of IP₃R and SOCE are not well-defined (30, 31). U73122 is unlikely to have off-target effects during ATP stimulation because it blocked Ca²⁺ mobilization (Fig. 1D) concomitantly with NLRP3 inflammasome activity (Fig. 1E); moreover, the inactive analog U73343 did not have such effects (Fig. S1D). Collectively, these findings indicate that Ca²⁺ signaling may be a common step in NLRP3 inflammasome activation.

Ca²⁺ Signaling Promotes Mitochondrial Damage.

Next, we asked how the requirement for Ca²⁺ signaling could be integrated with current models of NLRP3 inflammasome activation. K⁺ efflux has been implicated by studies showing that high extracellular K⁺ blocks NLRP3 inflammasome activation by multiple stimuli (Fig. 3A) (10). K⁺ efflux is not directly linked to activation of specific signaling pathways but can regulate Ca²⁺ signaling through its effects on plasma membrane polarization, and during extracellular Ca²⁺ entry K⁺ efflux counteracts membrane depolarization, thus promoting further Ca²⁺ influx. Indeed, we found that ATP-mediated Ca²⁺ flux is strongly reduced in the presence of high extracellular K⁺ (Fig. 3B), suggesting that K⁺ efflux may promote Ca²⁺ influx to activate the NLRP3 inflammasome.

Fig. 3.

Ca²⁺ signaling promotes mitochondrial damage during ATP stimulation. (A) BMDMs were stimulated with ATP in the presence of 40 mM K⁺ (hK⁺) or Na⁺ (hNa⁺), followed by analysis of NLRP3 inflammasome activation. (B) Ca²⁺ imaging of ATP-stimulated BMDMs in the presence of hK⁺ or hNa⁺ buffers. Fold-increase represents the ratio of maximal fluorescence to baseline fluorescence. Ionomycin stimulation was included as a control for Fura-2 loading. (C) BMDMs were pretreated with Leu-Leu-OMe or not (control), followed 1 h later by stimulation with 100 μM ATP in Ca²⁺-free HBSS (white arrow) to induce ER Ca²⁺ release through the P2Y receptor. Stimulation with the Ca²⁺ ionophore Ionomycin (black arrow) mobilized Ca²⁺ from distinct Ca²⁺ stores resistant to Leu-Leu-OMe and P2YR activation and served as a positive control for Fura-2 loading. (D and E) BMDMs, stimulated as indicated, were examined for NLRP3 inflammasome activation. (F and G) BMDMs stimulated as indicated were stained with MitoSOX (F) or Mitotracker Green and Deep Red (G) followed by flow cytometry analysis.

Phagolysosomal rupture has been posited to be crucial in NLRP3 inflammasome activation by crystals. In support of this, a lysosomotropic peptide (Leu-Leu-OMe) that induces phagolysosomal rupture through a completely distinct mechanism—osmotic stress of the lysosome—can activate the NLRP3 inflammasome (13). Importantly, the lysosomal compartment constitutes an intracellular Ca²⁺ store that can be released in a signal-dependent way to trigger ER Ca²⁺ release, indicative of functional coupling between these Ca²⁺ stores (32). Thus, we hypothesized that phagolysosomal rupture may induce Ca²⁺ mobilization to activate the NLRP3 inflammasome. Indeed, we demonstrated that Leu-Leu-OMe stimulation induces ER Ca²⁺ release, because it blocks subsequent Ca²⁺ mobilization by P2YR activation (Fig. 3C). Furthermore, Ca²⁺ signaling inhibitors block NLRP3 inflammasome activation by Leu-Leu-OMe, indicating that such Ca²⁺ mobilization is critical (Fig. 3 D and E). Thus, phagolysosomal rupture may induce Ca²⁺ mobilization to activate the NLRP3 inflammasome.

Finally, mitochondrial damage has been linked to NLRP3 inflammasome activation. ATP stimulation leads to increased mROS production, loss of membrane potential (ΔΨ), and release of mtDNA into the cytosol (14–16), and manipulations that block any of these processes reduce inflammasome activation. Importantly, mitochondria play a key role in shaping the spatiotemporal dynamics of Ca²⁺ signaling by uptake and release of cytosolic Ca²⁺ (18–20). However, excessive or sustained Ca²⁺ uptake can lead to mitochondrial damage characterized by increased production of mROS, mitochondrial permeability transition, and eventually rupture of the mitochondria (21–23). Thus, we hypothesized that the requirement for Ca²⁺ mobilization in activation of the NLRP3 inflammasome may reflect the ability of Ca²⁺ signaling to trigger mitochondrial damage. Consistent with this idea, Ca²⁺ signaling inhibitors block mROS production during ATP stimulation as measured by MitoSOX, a specific probe of mROS (Fig. 3F). Ca²⁺ signaling inhibitors also rescue ATP-mediated loss of membrane potential as measured by staining with Mitotracker Deep Red, a dye that accumulates in a ΔΨ-dependent manner (Fig. 3G). Moreover, Ca²⁺ signaling inhibitors block release of mtDNA into the cytosol during ATP stimulation (Fig. S6). Thus, we propose that Ca²⁺ mobilization is required for inducing mitochondrial damage. Together with recent studies, our results support the model whereby Ca²⁺ signaling critically regulates NLRP3 inflammasome activation by triggering mitochondrial damage.

CHOP Amplifies Activation of the NLRP3 Inflammasome.

Our results implicate a critical role for Ca²⁺ mobilization in NLRP3 inflammasome activation, thus we hypothesized that regulators of this process may modulate NLRP3 inflammasome activity. CHOP is a transcription factor that can regulate ER Ca²⁺ release through the IP₃R, and during ER stress CHOP deficiency leads to attenuated ER Ca²⁺ release concomitant with reduced oxidative stress and improved cellular survival (33–35). To test if CHOP plays any role in modulating NLRP3 inflammasome activity, we examined NLRP3 inflammasome activation in WT and Chop^−/− bone marrow-derived macrophages (BMDMs). We observed a consistent decrease in NLRP3 inflammasome activation in Chop^−/− BMDMs in response to ATP stimulation (Fig. 4A). In contrast, LPS-mediated IL-6 and TNF-α production (Fig. S7B); NLRP3 and pro–IL-1β up-regulation (Fig. S7C); and NLRC4 inflammaction activation were normal (Fig. S7D), suggesting no general defect in inflammatory responses. Expression of Asc, P2X7R, and cathepsin B was also normal (Fig. 4 and Fig. S7A).

Fig. 4.

CHOP amplifies NLRP3 inflammasome activation. (A) WT and Chop^−/− BMDMs were examined for inflammasome activation. (B) Ca²⁺ flux in ATP-stimulated WT and Chop^−/− BMDMs was examined by Ca²⁺ imaging. (C) WT and Chop^−/− BMDMs were pretreated with XeC or not followed by ATP stimulation. (D) WT and Chop^−/− mice were injected with LPS to induce sepsis, followed by analysis of serum cytokine levels. ns, not significant.

These findings suggested a specific involvement of CHOP in modifying NLRP3 inflammasome activity, and we asked whether such a role could be related to control of ER Ca²⁺ release via the IP₃R. First, we showed that ATP-mediated Ca²⁺ mobilization is reduced in Chop^−/− BMDMs (Fig. 4B). Second, we reasoned that if CHOP and IP₃R function in the same pathway, XeC should block NLRP3 inflammasome activity in WT but not Chop^−/− BMDMs. Indeed, although XeC treatment and CHOP deficiency both inhibited ATP-induced IL-1β production by ∼30–40% in WT BMDMs, no further block of IL-1β production was observed upon XeC addition to Chop^−/− BMDMs (Fig. 4C). Thus, CHOP may amplify NLRP3 inflammasome activity by regulating IP₃R-mediated Ca²⁺ release.

To demonstrate physiological relevance, we turned to a mouse model of sepsis in which IL-1β production is dependent on the NLRP3 inflammasome (6, 36). WT and Chop^−/− mice were injected with high dose LPS (40 mg/kg), followed by analysis of serum IL-1β levels. Significantly reduced IL-1β secretion was found in Chop^−/− mice compared with WT mice, but levels of TNF-α and IL-6 were similar (Fig. 4D). Thus, CHOP has a specific role in regulation of the NLRP3 inflammasome in vivo.

Discussion

In this study, we elucidate a critical role for Ca²⁺ signaling in activation of the NLRP3 inflammasome. Previous studies relying mainly on 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA)-AM treatment or incubation in Ca²⁺-free media have made somewhat inconsistent conclusions regarding the role of Ca²⁺ in NLRP3 inflammasome activation (28, 37–39). Given the pleitropic effects of BAPTA-AM on many cellular processes including LPS signaling (40), we wished to more carefully examine this issue by defining specific regulators of Ca²⁺ mobilization. Using XeC, U73122, and 2-APB, we implicate PLC proteins, IP₃R-mediated Ca²⁺ release, and SOCE as critical regulators during ATP stimulation (Fig. 1D). Although not well-characterized, induction of such a pathway by stimuli that induce Ca²⁺ influx across the plasma membrane is not without precedence (41, 42). Definitive demonstration of the role of this pathway requires genetic models, but pharmacological inhibitors are commonly used in the field of Ca²⁺ signaling because of redundancy in regulation (i.e., multiple PLCs) or, conversely, lethality of cells and mice lacking critical regulators (i.e., IP₃R1). Most importantly, we correlate block of Ca²⁺ mobilization (Fig. 1D) with NLRP3 inflammasome activity (Fig. 1 E and F), and use independent methods to implicate ER Ca²⁺ release and extracellular Ca²⁺ entry (Fig. 1 B and C). Ca²⁺ signaling inhibitors do not regulate signal 1 (Fig. S1B) or affect cell viability (Fig. S1C), but block NLRP3 inflammasome activation at every step examined, including complex assembly (Fig. S2), caspase-1 processing, IL-1β processing, and IL-1β release to culture supernatants (Fig. 1 E and F). Collectively, these data indicate a role for Ca²⁺ mobilization in regulating the upstream, proximal steps in NLRP3 inflammasome activation (although not ruling out additional effects more distally).

How can our findings be integrated into existing models for activation of the NLRP3 inflammasome? K⁺ efflux and phagolysomal rupture have been implicated in NLRP3 inflammasome activation but the critical intermediate events have not been defined. Our results are consistent with the idea that high K⁺ buffers perturb Ca²⁺ mobilization and that K⁺ efflux may regulate Ca²⁺ signaling (Fig. 3B), while rupture of phagolysosomes leads to ER Ca²⁺ release (Fig. 3 C–E), likely as a result of Ca²⁺ mobilization from the lysosomal compartment. Thus, we propose that K⁺ efflux and phagolysosomal rupture promote Ca²⁺ mobilization to activate the NLRP3 inflammasome. Conversely, we report that Ca²⁺ signaling inhibitors block ATP-mediated mitochondrial Ca²⁺ damage, as measured by mROS production (Fig. 3F), loss of ΔΨ (Fig. 3G), and release of mtDNA into the cytosol (Fig. S6). Because these processes are required for NLRP3 inflammasome activation, this finding suggests that the critical role of Ca²⁺ signaling, at least during ATP stimulation, is to mediate mitochondrial damage. Consistent with this theory, activation of the NLRP3 inflammasome by transfection of oxidized mitochondrial DNA (14) is not blocked by Ca²⁺ signaling inhibitors (Fig. S8).

Additionally, we find that Ca²⁺ is mobilized during stimulation with alum, MSU, and nigericin, and that such Ca²⁺ signaling is required for NLRP3 inflammasome activation (Fig. 2, and Figs. S3 and S4). This finding suggests that Ca²⁺ mobilization is a common, proximal step in activation of the NLRP3 inflammasome. We propose that the multitude of stimuli that activate the NLRP3 inflammasome can be rationalized in light of the many mechanisms for mobilizing Ca²⁺, including G protein-coupled receptors, receptor tyrosine kinases, immunoreceptor tyrosine-based activation motif-coupled receptors, and Ca²⁺ entry channels (17). However, we note that Ca²⁺ signaling cannot be sufficient for NLRP3 inflammasome activation, as indicated by the inability of the Ca²⁺ ionophore ionomycin to induce IL-1β production (Fig. S9), despite triggering Ca²⁺ mobilization and mitochondrial Ca²⁺ uptake (43). Presumably, many other stimuli that induce “physiological” Ca²⁺ signaling would also not activate the NLRP3 inflammasome. Importantly, Ca²⁺ signaling is well-established to be necessary but not sufficient for mitochondrial damage (21–23), and we propose that only stimuli that mobilize Ca²⁺ in a manner leading to mitochondrial damage would activate the NLRP3 inflammasome (Fig. 5).

Fig. 5.

Model. Our results support a model in which Ca²⁺ signaling is critical for NLRP3 inflammasome activation. In response to ATP and perhaps other stimuli, Ca²⁺ mobilization from ER stores and the extracellular space triggers mitochondrial damage, including increased mROS production, loss of membrane potential, and release of mtDNA into the cytosol.

Finally, we wondered about a potential role for the NLRP3 inflammasome in “sensing” dysregulated Ca²⁺ homeostasis and inducing inflammation in chronic diseases, such as neurodegeneration, atherosclerosis, and obesity-associated metabolic diseases (44–46). In considering this possibility, we focused on CHOP, an effector of the ER-stress pathway (35). ER stress is closely linked to dysregulated Ca²⁺ homeostasis and CHOP can regulate ER Ca⁺ levels (33); moreover, a recent report links ER stress to NLRP3 inflammasome activation (47). Indeed, CHOP deficiency attenuates ATP-mediated NLRP3 inflammasome activation (Fig. 4A) and IL-1β production during sepsis (Fig. 4D), but other inflammatory responses were not affected (Fig. 4D and Fig. S7). The underlying mechanism is likely to be attenuated Ca²⁺ release from the IP₃R, as supported by Ca²⁺ imaging studies (Fig. 4B) and the inability of IP₃R inhibition to further block inflammasome activation (Fig. 4C). Somewhat surprisingly, alum stimulation of Chop^−/− BMDMs leads to normal NLRP3 inflammasome activation (Fig. S10 A and B), but it also depletes ER Ca²⁺ stores comparably in WT and Chop^−/− BMDMs (Fig. S10C). Less ROS is generated by alum stimulation compared with ATP (Fig. S10 D and E), and CHOP has been proposed to regulate ER Ca²⁺ release in a manner dependent on ER redox status, so one possible interpretation of these results is that CHOP amplifies NLRP3 inflammasome activation in a manner dependent on ROS levels. Our results reveal CHOP as a modulator of the NLRP3 inflammasome pathway and support the idea that in many chronic diseases, CHOP and ER stress may amplify NLRP3 inflammasome activity to augment inflammation.

Materials and Methods

Mice.

B6 and Chop^−/− mice were from the Jackson Laboratories and Nlrp3^−/−, Asc^−/−, and Caspase-1^−/− mice were gifts of Visha Dixit (Genentech, San Francisco, CA) and Ruslan Medzhitov (Yale University, New Haven, CT). All mice are on the B6 background. Animals were bred and maintained at Harvard Medical School and all animal experiments were done with approval by and in accordance with regulatory guidelines and standards set by the Institutional Animal Care and Use committee of Harvard Medical School.

Cells and Stimulations.

To activate the NLRP3 inflammasome, BMDMs (48) were primed with LPS before stimulation with ATP (1–5 mM, 30 min), alum (200 μg/mL, 4 h), nigericin (20 μM, 1 h), MSU (100 μg/mL, 4 h), and Leu-Leu-OMe (1 mM, 5 h). To activate the NLRC4 inflammasome, LPS-primed BMDMs were transfected with FliC (80 ng/0.7 × 10⁶ cells) using the Lipofectamine 2000 reagent (8 μL) 5 h before harvest. Transfection of oxidized mtDNA was done as previously described (14). U73122 and U73343 (10 μM), 2-APB (100 μM), XeC (5 μM), and YVAD (40 μM) were added following LPS priming, 30 min before addition of the NLRP3 inflammasome activators. Ca²⁺ free media (Invitrogen) was added 10 min before addition of NLRP3 inflammasome activators, as was KCl or NaCl (40 mM final concentration). To transiently express PLC-PH-GFP, BMDMs were transfected with DNA using Nucleofector technology (Lonza) and rested for 4 h before stimulation; 100 μM ATP was used to activate the P2YR.

Antibodies.

Antibodies were obtained from Santa Cruz (caspase-1), Alexis (NLRP3), Sigma (Tubulin), R&D Systems (IL-1β), and National Cancer Institute-Frederick Biological Resources Branch Repository (IL-1β).

Western Blotting and ELISA.

In some Western blotting experiments, supernatants were concentrated by addition of TCA (20% final volume, 10 min at 4 °C), followed by centrifugation (17,000 × g, 5 min at 4 °C) and two washes of the protein pellet with ice-cold acetone. IL-1β ELISA kit was from eBioscience, and IL-6 and TNF-α ELISA kits were from Biolegend.

Ca²⁺ Imaging.

BMDMs were loaded with the Ca²⁺-sensitive indicator dye Fura-2-AM (1.5 μM, 30 min, 37 °C). Ca²⁺ flux was measured by time-lapse microscopy, and fluorescence images were collected using either a Yokogawa Spinning Disk microscope equipped with a Sutter DG-4 illuminator and a 40× objective or a Nikon TE200 inverted microscope with Nomarski optics. The 340/380 ratio was calculated using ImageJ (National Institutes of Health).

Flow Cytometry.

MitoSOX, MitoTracker Green and Red, and DCFDA staining were done according to manufacturer's instructions (Invitrogen). Data were acquired with a FACSCalibur flow cytometer (BD Biosciences) and analyzed with FlowJo analytical software (TreeStar).

Measuring Cytosolic mtDNA.

Measurements of cytosolic mtDNA were done following an established protocol (15).

Sepsis.

Mice age 6–10 wk were injected intraperitoneally with 40 mg/kg LPS. After 3 h, mice were killed and blood was obtained by cardiac puncture. Serum levels of IL-1β, TNF-α, and IL-6 were analyzed by ELISA.

Statistical Analysis.

Statistics were calculated using GraphPad Prism 5. Comparisons of two groups were analyzed using two-tailed t test, and comparisons of multiple groups were analyzed using one-way or two-way ANOVA as indicated. P values <0.05 were considered significant.