The role of caloric load and mitochondrial homeostasis in the regulation of the NLRP3 inflammasome

Abstract

Sterile inflammation is a cornerstone of immune activation in obesity and type 2 Diabetes Mellitus. The molecular underpinnings of this inflammation include nutrient excess-mediated activation of the innate immune NLRP3 inflammasome. At the same time, disruption of mitochondrial integrity is emerging as an integral control node in NLRP3 inflammasome activation and is also associated with caloric overload conditions including obesity and diabetes. Conversely, caloric restriction and fasting mimetic interventions alleviate these caloric excess-linked diseases and reduce inflammation and the NLRP3 inflammasome. The objective of this review is to integrate the findings linking mitochondrial integrity to the activation of the NLRP3 inflammasome and to evaluate how caloric restriction or caloric restriction mimetic compounds may play a role in attenuating the NLRP3 inflammasome and sterile inflammation.

Introduction

An inflammatory underpinning of nutrient excess and obesity is well established and is increasingly being recognized in obesity-associated diseases including type-2 diabetes, hepatosteatosis and asthma. Conversely, caloric restriction and/or intermittent fasting have been found to reverse these nutrient-overload conditions in parallel with the blunting of inflammation [1, 2]. Taken together, an emerging concept is that the regulation of intracellular nutrient-sensing events may play an important role in exacerbating and/or reversing the inflammatory component of nutrient-overload-linked diseases.

At the same time, mitochondria are being recognized as an important nutrient-sensing organelle, which functionally adapts in a nutrient-dependent manner [3]. Additionally, these nutrient-dependent metabolic changes initiate retrograde signaling from the mitochondria that plays an important role in immune activation [4]. Amalgamating these data, a developing hypothesis is that the mitochondria may function as the nutrient-sensing signaling organelle to integrate metabolic disease and immune activation.

Here, we will give a brief overview of inflammation linked to obesity and then focus on the role of the mitochondria as a ‘nutrient-sensing’ organelle and its role in the NLRP3 inflammasome as an apical component of nutrient excess-initiated inflammatory signaling. The potential therapeutic strategies that focus on these pathways will then be explored.

Inflammation in obesity-linked disease

Classically, inflammation is initiated as a coordinated and robust response to distinct harmful stimuli that when acutely activated, functions to restore basal homeostasis. The sustained robust activation of these inflammatory cascades can result in acute (systemic inflammatory response syndrome or sepsis) or more chronic inflammatory disease/s. In contrast, obesity activates a chronic low-level inflammatory program that is promulgated via activation of multiple stress-signals in a wide array of tissue types. Although multiple components of classic inflammatory programs are engaged as a component of obesity, their lower levels of activation rather play a role in exacerbating complications linked to obesity, instead of manifesting with overt evidence of signs or symptoms synonymous with more robust activation of the immune system. The consequences of this chronic low level of inflammation include the perpetuation of atherosclerosis, gout, insulin resistance and type 2 diabetes mellitus and to other obesity-linked diseases including, for example, non-alcoholic steatohepatitis (NASH) and asthma [5].

Obesity-associated inflammation includes the activation of both the innate and adaptive immune systems. Moreover, in contrast to other inflammatory diseases that can be linked to specific organs or locations in the body, obesity results in the infiltration and/or activation of inflammatory cells in various lipid-accumulating organs which in turn result in the production of cytokines and acute-phase reactants that confer systemic effects. The most well-studied organs include immune activation in adipocytes, the liver, pancreas, muscle, and hypothalamus. Triggers of the pro-inflammatory pathways include the excess nutrient intermediates including saturated fatty acids, ceramides and excess glucose levels that function through pattern recognition receptors on leukocyte or adipocyte cell membranes. These in turn activate resident leukocytes and initiate pathways that stimulate cytokines and chemokines to recruit circulating cells to perpetuate the process. These pathways have been reviewed in detail and are not discussed further in this review [5]. The role of these nutrients in activating a component of sterile inflammation, termed the Nod-like receptor family protein 3 (NLRP3) inflammasome pathway, is schematized in Fig. 1, and expanded upon below. In addition, the nutrient intermediate level changes associated with caloric restriction or fasting, which has an opposing effect on the NLRP3 inflammasome, are similarly discussed later in the review, and the nutrients themselves are also depicted in Fig. 1.

Fig. 1.

Nutrients linked to NLRP3 inflammasome activation. Schematic showing the effects of different nutrients on the NLRP3 inflammasome. Here, excess levels of nutrients associated with caloric overload including glucose and fatty acids and reactive lipids including, e.g., ceramide play a role in activating the NLRP3 inflammasome. Low-level circulatory endotoxemia (LPS), which is also evident in obesity, could also function to activate the NLRP3 inflammasome. In contrast, nutrient deprivation such as caloric restriction or fasting via the increase in NAD⁺ may also blunt the inflammasome. The mechanisms here have not been fully elucidated but appear to function in part through augmenting mitochondrial fidelity. The ketone body, β-hydroxybutyrate suppresses the inflammasome via the prevention of K⁺ efflux and by preventing oligomerization of ASC

The role of the NLRP3 inflammasome pathway that is both activated as an intracellular immune-surveillance program in obesity and amplifies subsequent inflammation will be reviewed in detail with the additional focus of the role of the mitochondria as a nutrient-sensing organelle in this inflammation activating program. This large multiprotein complex consists of the sensor protein NLRP3, the adaptor ASC (Apoptosis-associated Speck-like protein containing a Caspase-recruitment domain, also called PYCARD) and caspase-1, a cysteine protease which mediates the processing and release of IL-1β and IL-18. These cytokines are synthesized as inactive precursors, as the first signal of the inflammasome program, and require post-translational processing to complete NLRP3 inflammasome activation [6]. Sterile inflammation linked to obesity is mediated in part by the NLRP3 inflammasome [7]. The activation of this program, as a component of the innate immune system, similarly exacerbates obesity-linked diseases including insulin resistance, diabetes and asthma [8, 9]. The biological pathways driving this innate immune program are now well defined [10]. In the context of obesity, triggers that engage toll-like receptors (TLRs) to initiate transcriptional priming of the NLRP3 inflammasome include: adipose tissue hypertrophy with macrophage infiltration and cytokine secretion; elevated circulating saturated fatty acids; and/or obesity-linked endotoxemia [11–14]. These, in turn activate NF-κB-dependent transcription to upregulate genes encoding NLRP3 and pro-IL-1β. The transcriptional initiation of the inflammasome program is either termed priming or signal 1. Subsequent inflammasome activation/execution, which is also called signal 2, is inducted in response to additional pathogen-associated molecular patterns (PAMPs) or host cell-derived damage-associated molecular patterns (DAMPs). This second signal resulted in the assembly and self-oligomerization of inflammasome components. The NLRP3 complex subsequently promotes caspase-1 activation and cleavage of pro-IL-1β and pro-IL-18 into bioactive cytokines that amplify inflammation [6, 15]. Given the co-association of nutrient excess and NLRP3 activation, it is not surprising that distinct DAMPs are elevated during nutrient excess, to promulgate this pro-inflammatory environment [16]. These pathways are schematized in Fig. 2.

Fig. 2.

Schematic of NLRP3 inflammasome activation. NLRP3 priming usually results from activation of Toll-like receptor (TLR)-mediated activation of NF-κB, which then transactivates canonical NLRP3 complex components and associated cytokine pre-proteins including those encoding for IL-1β and IL-18. Subsequent pathogen- or molecular damage-associated patterns, through either the promotion of K⁺ efflux, or the disruption of intracellular organelles (with the subsequent generation of, e.g., mitochondrial ROS, ER stress or lysosomal accumulation of crystals) promotes oligomerization of NLRP3 supercomplexes. These trigger collectively function as pathogen- or disease-associated molecular patterns (PAMPS/DAMPS). This results in the activation of caspase-1, which cleaves pro-IL-1β and Pro-IL-18 to generate cytokines which initiate subsequent immune amplification. ASC apoptosis-associated Speck-like protein containing a caspase-recruitment domain, NLRP3 Nod-like receptor family protein 3

Regulation of mitochondrial function by caloric load

Mitochondria are pivotal double-membrane-bound organelles that function to sustain numerous homeostatic processes including energy and heat production, nutrient metabolism, calcium and redox homeostasis, numerous biosynthetic pathways and they play a role in regulating programmed cell death [17]. At the same time, mitochondria are unique intracellular organelles in that they retain their own genome and protein synthetic ribosomes. However, the mitochondrial genome only encodes a fraction of the genes and canonical proteins required to perform the myriad of mitochondrial functions [18]. To complete the full array of proteins within the mitochondria, there is coordinate regulation and control between the nuclear and mitochondrial genomes. At the same time, mitochondria are increasingly being recognized as signaling organelles where signals emanating from mitochondria orchestrate cellular signaling events and control nuclear gene regulation.

Nutrient restriction activates AMP-activated protein kinase (AMPK), a nutrient-sensing enzyme that is activated by decreases in the energy state (reflected by an increase in the AMP/ATP ratio). AMPK activates in turn several pathways that lead to an increase in mitochondrial function in an attempt to restore ATP levels, while it inhibits other pathways that consume ATP and are not required for immediate survival (synthesis of fatty acid and proteins, cell proliferation). Caloric restriction also leads to high intracellular nicotinamide adenine dinucleotide (NAD⁺) levels [19], which activate sirtuin proteins, a group of enzymes possessing deacylase activity that requires NAD⁺ as a cofactor to function, and thus are considered metabolic and energy sensors [20]. There are seven sirtuin family members in mammals: SIRT1, SIRT6, and SIRT7 are nuclear proteins, SIRT2 is mainly cytosolic, and SIRT3, SIRT4, and SIRT5, localize within the mitochondria [20]. Most sirtuins are lysine deacetylases, although SIRT5 has been shown to catalyze the desuccinylation, demalonylation and deglutarylation of lysines [21, 22].

Interestingly, AMPK and SIRT1, the best characterized member of the sirtuin family, reciprocally regulate each other during caloric restriction [23]: SIRT1 deacetylates and activates serine-threonine liver kinase B1 (LKB1) [24], which phosphorylates and activates AMPK, whereas AMPK further activates SIRT1 by increasing the levels of NAD⁺ [25, 26], either by inducing the transcription of the NAD⁺ biosynthetic enzyme nicotinamide phosphoribosyltransferase (Nampt) [25], or by other pathways [26]. AMPK and SIRT1 share a number of common targets, including key transcriptional activators of mitochondrial gene expression such as peroxisome proliferator-activated receptor coactivator-1α (PGC-1α) and forkhead box O (FoxO). Here, AMPK phosphorylates these regulators [27], which primes them for SIRT1-dependent deacetylation [19, 28], leading to their full activation. Remarkably, both AMPK and SIRT1 are also involved in the regulation of mitochondrial quality control programs: AMPK activates mitophagy via activation of Unc-51-like autophagy-activating kinase 1 (ULK1), a protein involved in the early steps of autophagosome biogenesis, by direct phosphorylation, and by inhibiting mammalian target of rapamycin complex 1 (mTOR), which normally acts as a suppressor of ULK1 [29]. SIRT1 directly interacts with and deacetylates canonical autophagy mediators, including Atg5, 7 and 8 [29]. By regulating both mitochondrial biogenesis and mitochondrial turnover, caloric restriction improves mitochondrial function. In addition, caloric restriction activates other sirtuins, including SIRT3, via an increase in the levels of mitochondrial NAD⁺ [30].

SIRT3 improves mitochondrial function by deacetylating key residues of several mitochondrial proteins, including proteins of the respiratory chain (such as mitochondrial complex I protein NDUFA9 [31]), proteins involved in fatty acid oxidation (such as long-chain acyl-CoA dehydrogenase, LCAD [32]), proteins that act as ROS scavengers (such as superoxide dismutase 2, SOD2 [33]) and proteins involved in cell death, such as the regulatory component of the mitochondrial permeability transition pore (mPTP) (cyclophilin D, CypD [34–36]). Thus, SIRT3 regulates mitochondrial activity (increases mitochondrial ATP synthesis capacity) and protection against ROS and cell death.

Conversely, chronic nutrient excess inhibits AMPK activity because of high intracellular ATP levels, dramatically decreases NAD⁺ levels [37, 38] and SIRT1 and Nampt expression [37], leading to hyperacetylated, inactive, PGC-1 α [39] and to decreased mitochondrial biogenesis. Nutrient overload also decreases SIRT3 expression, causing mitochondrial protein hyperacetylation and mitochondrial dysfunction [40]. Globally, this results in fatty acid accumulation, via decreased oxidation, and in the generation of ROS.

Mitochondrial dysfunction and immune activation

In recent years, it has become apparent that the mitochondria have a front row seat in the programs orchestrating inflammation and the study of this field is now termed immunometabolism. An example of metabolic contributions is evident in T cell polarity. Quiescent and dividing T cells embrace mitochondrial oxidative metabolism and glycolysis whereas differentiation into effector T cells downregulates oxidative metabolism with a preference for glycolysis [41]. Myeloid cells (including as monocytes and macrophages), which are the most NLRP3-enriched cells [42], also exhibit metabolic remodeling as a component of their inflammatory functioning. Here, activation of TLRs on quiescent myeloid cells, similarly divert the metabolic machinery away from oxidative phosphorylation towards glycolysis. In M1 (pro-inflammatory) macrophages, glycolysis and pentose phosphate pathway are strongly stimulated to provide rapid ATP and NADPH, respectively. NADPH is used for the synthesis of fatty acids as precursors of prostaglandins and other pro-inflammatory molecules. Additionally, the activation of Nox2, an NADPH oxidase, and a membrane-bound enzyme complex that faces the extracellular space, promotes the generation of superoxide to promote bacterial killing. Another main source of superoxide and reactive oxygen species (ROS) is the mitochondria, especially complex I of the electron transport chain (ETC) [43]. ROS production at complex I is favored by a high NADH/NAD⁺ ratio in the matrix [43], which happens in the context of an inhibited ETC in M1 macrophages and lipopolysaccharide (LPS)-treated cells [44]. ROS also have a signaling function, as they prevent the dephosphorylation of several mitogen-activated protein kinases (MAPKs) in LPS-treated cells, including c-Jun N-terminal kinase (JNK), extracellular signal-regulated kinase (ERK) and p38 MAPK, which may contribute to enhanced pro-inflammatory cytokine production, including that of IL-6 and TNF-α [45]. Importantly, inhibition of glycolysis blunts the inflammatory response in LPS-treated macrophages [46].

An additional metabolic control node in immune function involves the disruption of the tricarboxylic acid (TCA) cycle. Two bottlenecks in this cycle have been found to play a role in immune modulation including reductions in flux through isocitrate dehydrogenase (IDH) and succinate dehydrogenase (SDH) enzymes [47, 48]. At the first break point, IDH is reduced through transcriptional repression and this leads to a reduction in citrate conversion to alpha-ketoglutarate. The increased citrate levels are used for two different purposes: first, the enzyme product of the immunoresponsive gene 1, Irg1 encodes cis-aconitate decarboxylase. This enzyme promotes citrate decarboxylation to generate itaconic acid, an antimicrobial compound that can effectively inhibit growth of a variety of microorganisms. Second, the accumulation of citrate is used to synthesize fatty acids and pro-inflammatory intermediates. Here, the excess citrate is exported to the cytosol through the mitochondrial citrate carrier (CIC), where it is converted by citrate lyase to oxaloacetate and acetyl-CoA. Cytosolic acetyl-CoA is carboxylated by acetyl-CoA carboxylase to yield malonyl-CoA, which is subsequently used by fatty acid synthase in an NADPH-dependent manner. Export of citrate from the mitochondria to the cytosol for this purpose is stimulated in LPS-treated cells by increased CIC expression [49]. The disruption of SDH activity leads to the buildup of succinate which drives the stabilization, in a ROS-dependent manner, of the transcription factor hypoxia-inducible factor-1α (HIF-1α that in turn binds to the promoter of IL-1β and stimulates its expression [46]. Interestingly, the integration of metabolic remodeling during immune cell activation is also evident in that succinate accumulation is dependent on glycolytic activity and can be inhibited by 2-deoxyglucose, an inhibitor of this pathway [46]. In addition, given the incomplete TCA cycling, with the disruption in the conversion of isocitrate to alpha-ketoglutarate, glutamine-dependent anaplerosis becomes the major source of succinate, and the aspartate-arginosuccinate shunt (AASS), which feeds into the TCA cycle at malate, becomes the source of oxaloacetate and thus citrate [47].

The role of mitochondria in the regulation of the NLRP3 inflammasome is not limited to metabolic remodeling, but also due to the disruption of mitochondrial integrity itself. This is also evident in the NLRP3-linked conditions of obesity and diabetes that also associate with mitochondrial perturbations [50]. Here, the mitochondrial content itself functions as a DAMP to activate the NLRP3 inflammasome through the leak of mitochondrial content into the cytoplasm [51, 52], into the circulation [53] and/or via disruption of mitochondrial quality control programs [54]. The mechanisms whereby mitochondrial content extrusion into the cytoplasm function as a DAMP include: a role of the mitochondrial membrane cardiolipin via its direct interaction with NLRP3 [55]; via the release of mitochondrial reactive oxygen species [54] and/or due to the intrinsic composition of hypomethylated CpG motifs of mitochondrial DNA that resemble the immunogenic properties of bacterial CpG DNA motifs [56].

The role of potassium efflux versus mitochondrial sensing in the activation of the NLRP3 inflammasome

The amalgamation of events that results in NLRP3 inflammasome activation is extensive and not fully characterized. Nevertheless, common components or pathways in the activation (signal 2) of the immune surveillance program include the efflux of potassium out of the cell and the involvement, at multiple levels of mitochondrial biology. The studies underpinning these events are expanded upon below.

After a priming step triggered by engagement of TLRs, the NLRP3 inflammasome is activated by a large number of stimuli that are chemically and structurally distinct [57]. These include bacterial products or toxins (nigericin, bacterial RNA) which function as PAMPs and endogenous or exogenous DAMPS including biological intermediates such as ATP, fatty acids, monosodium urate (MSU) and cholesterol crystals and environmental triggers such as silica or alum. Given the diversity of these triggers, the identification of common downstream pathways would be necessary to be able to comprehensively manipulate NLRP3 inflammasome activation. The most well-characterized common denominator downstream of these stimuli is cellular potassium (K⁺) efflux. Efflux of this cation has been demonstrated in response to multiple triggers of the NLRP3 inflammasome and has shown to be countered by the elevation of extracellular K⁺ levels that inhibits inflammasome assembly and activation [58]. Data to show that K⁺ efflux is upstream of the NLRP3 is evident given that, in the gain-of-function Cryopyrin-Associated Periodic fever Syndrome (CAPS) NLRP3 mutant cells, priming can induce and activate the inflammasome without the requirement of a second signal or of K⁺ efflux [58]. Interestingly, the serine-threonine kinase NEK7 has been recently found to be a mediator of NLRP3–ASC complex formation and ASC oligomerization downstream of K⁺ efflux [59, 60]. NEK7 is also required for inflammasome activation in CAPS mutant cells [59], although and somewhat surprisingly, its kinase activity is dispensable for inflammasome formation [59, 61].

At the same time, emerging data support a robust link between mitochondrial dysfunction and the activation of the NLRP3 inflammasome. This was conceptually illustrated where chemical inhibitors of mitochondrial respiration and pharmacologic and genetic disruption of mitochondrial turnover through mitophagy have been shown to activate the NLRP3 inflammasome [51]. Here, the mitochondrial-emitted signals have included both excess mitochondrial ROS levels and a lower mitochondrial membrane potential [51, 54] and the cytosolic translocation of mitochondrial DNA (mtDNA) in response to inflammasome activation [54]. The close association between mitochondria and NLRP3 complex formation has also been shown where the NLRP3 relocates to the perinuclear space and co-localizes with the ER and the mitochondria in response to inflammasome induction [51]. In a similar vein, the mitochondrial antiviral signaling protein (MAVS) has been found to mediate recruitment of the NLRP3 to mitochondria after inflammasome induction by noncrystalline activators [62, 63]. The importance of mitochondrial DNA in inflammasome activation was reinforced in ρ⁰ (mtDNA-deficient) J774A.1 macrophages, which showed both impaired mitochondrial ROS generation and a strong attenuation of the inflammasome response, with no change in pro-IL-1β levels, IL-1β mRNA or TNF-α release [54]. Together these data suggested that inflammasome inducers, via the generation of mitochondrial ROS and mitochondrial damage could promote the release of mtDNA into the cytosol. Interestingly, transfected mtDNA binds to both the NLRP3 inflammasome and to the double-stranded DNA-sensing inflammasome called AIM2 [52]. However, NLRP3 inflammasome inducers preferentially increase the amount of oxidized mitochondrial DNA, as measured by an increase in 8-hydroxy-guanosine, which interacts exclusively with NLRP3 [52]. Indeed, transfection of DNA or native mtDNA to LPS-treated macrophages preferentially triggers the AIM2 inflammasome, but transfection of oxidized mtDNA triggers the NLRP3 inflammasome [52].

As the activation of the mPTP would facilitate mtDNA extrusion, investigators began to explore whether the modulation of the mPTP also modulates inflammasome activation. The involvement of the mPTP was supported by the reduction of mtDNA release into the cytosol following inflammasome induction in the presence of the mPTP inhibitor cyclosporine A (CsA) [54] and by the fact that increased Ca²⁺, a known inducer of the mPTP [64], activates the inflammasome in murine bone marrow-derived macrophages (BMDM) [65] and in human monocyte-derived macrophages [66]. Indeed, inhibition of store-operated Ca²⁺ entry (SOCE) or Ca²⁺ release from the endoplasmic reticulum to the cytosol, and Ca²⁺ chelation agents similarly inhibits NLRP3 activation without affecting NLRP3 priming [65, 66]. The physiological role of Ca²⁺-mediated effects was confirmed by Ca²⁺-chelation-mediated inhibition of constitutive inflammasome activation in CAPS mutation subject cells [67] and in other immune dysregulation syndromes [68]. Interestingly, high extracellular levels of K⁺ inhibit the influx of Ca²⁺ during inflammasome induction, suggesting a role of K⁺ during inflammasome activation may be, in part, via the control of intracellular Ca²⁺ levels [65, 69]. Ca²⁺ signaling inhibitors also block ATP-induced generation of mitochondrial ROS, loss of mitochondrial membrane potential and the release of mtDNA into the cytosol [65], suggesting that the requirement for Ca²⁺ may reflect the ability of this cation to trigger both mitochondrial damage and permeability transition. Mitochondrial damage with the extrusion of mitochondrial content is also evident in cell death programs including apoptosis and necrosis [64], and the overexpression of the anti-apoptotic protein Bcl-2 prevents NLRP3 inflammasome activation [52].

An additional mechanism whereby the integrity of mitochondria mediates NLRP3 inflammasome assembly and activation is the exposure of NLRP3 to the specialized inner mitochondrial membrane lipid cardiolipin [55]. The NLRP3 protein has found to associate with cardiolipin and as cardiolipin is found exclusively in the inner mitochondrial membrane, the fidelity of the outer membrane would need to be compromised for this interaction to occur [55]. At the same time, interfering with the synthesis of cardiolipin inhibited inflammasome activation [55]. Given the prokaryotic origins of mitochondria, this DAMP-like effect of the unique mitochondrial lipid is not surprising, although whether cardiolipin and mtDNA function in concert to regulate the NLRP3 inflammasome needs to be investigated further. The role of mitochondria and mitochondrial components in NLRP3 activation are shown in Fig. 3.

Fig. 3.

The role of mitochondria in NLRP3 inflammasome assembly and activation. The disruption of mitochondria can promote inflammasome assembly and activation through the release of ROS signaling and/or by the extrusion of cardiolipin (CL) or mitochondrial DNA (mtDNA) into the cytoplasm. Emerging data also support that mitochondria may function as a platform for the enucleation and oligomerization of inflammasome regulatory protein components to enable the activation of caspase-1. Not shown is the mitochondria-associated viral signaling (MAVS), which may function as the anchor to tether the inflammasome multiprotein complex to the mitochondria

Despite the multiple studies identifying various biochemical consequences of mitochondrial functioning and physical components of mitochondria that modulate the NLRP3 inflammasome, other studies have questioned the role of mitochondria in this innate immune surveillance program. The contrarian studies have questioned the role of Bcl-2 in attenuating [70] and of ROS production in activating the inflammasome [58]. The role of the mPTP has also been challenged [70], but here the manipulation was by depleting levels of a regulatory protein of the mPTP, i.e., CypD, rather than by manipulating the pore itself [71, 72]. The regulatory role of Ca²⁺ has also been questioned with data suggesting that BAPTA inhibits the inflammasome independent of its Ca²⁺ chelating properties [73]. A concept that is developing that may reconcile these discrepant findings suggests that the relative contribution of mitochondrial function and fidelity may be dependent on the duration and the magnitude of the inflammasome inducing stress [74]. This can be envisioned further when, for example, a stress that promotes strong K⁺ efflux might allow NLRP3 oligomerization without mitochondrial damage, and would, therefore, not be blocked by mPTP or ROS inhibitors, whereas a stress that triggers weaker K⁺ efflux might be more dependent on mitochondrial danger signals for NLRP3 to use the mitochondria as a platform for oligomerization [74]. Studies support this concept, as it was found that differing times and concentrations of LPS and ATP resulted in substantially different amounts of mtDNA release [74]. Similarly, unpublished studies from our laboratory show that the ability of the mitochondrial ROS scavenger mitoTEMPO to fully inhibit inflammasome activation in murine BMDMs depends on the concentration of ATP used (unpublished observations). Taken together, these studies suggest that some of these discrepancies in mechanism-defining studies may need to be interpreted using the same dose and temporal exposure to NLRP3 triggering and activation events.

Sirtuins as metabolic sensors that regulate inflammation

Since sirtuins are nutrient sensing and play an important role in the modulation of mitochondrial fidelity [20], a role for this enzymes in inflammasome regulation is currently being explored, especially given the diametrically opposed link between mitochondrial dysfunction, the NLRP3 inflammasome and caloric excess versus mitochondrial fidelity and anti-inflammatory effects of caloric restriction and/or fasting mimetic interventions [16]. Emerging evidence supports that these proteins, and also their cofactor NAD⁺, are involved in metabolic diseases associated with inflammation [75, 76].

SIRT1

SIRT1 deacetylates and inactivates the p65 subunit (RelA) of the nuclear NF-κB complex [77], which limits the expression of pro-inflammatory genes and thus inflammasome priming, as deacetylated RelA is unable to bind the promoter of its target genes. Thus, SIRT1 knockdown leads to a significant increase in the mRNA expression for TNF-α, IL-6 and IL-1β in adipocytes [78]. Although SIRT1 knockout mice rarely survive postnatally [79], inducible SIRT1 knockout mice have been generated and are highly susceptible to LPS-induced inflammatory kidney injury [80] or to sepsis-induced inflammatory lung injury associated with NLRP3 inflammasome activation [81]. Furthermore, myeloid-specific SIRT1 knockout mice display aggravated high-fat diet-induced hepatosteatosis and insulin resistance due to hyperacetylated RelA [82]. As stated above, SIRT1 also modulates mitochondrial metabolism by regulating the acetylation of PGC-1α and thus also may limit inflammation indirectly by promoting a metabolic switch. In fact, it has been proposed that SIRT1, along with SIRT6 [83] and SIRT3 [84], limits and shuts off acute inflammatory responses by generating a metabolic shift from glycolysis to fatty acid oxidation in monocytes to restore homeostasis [75].

SIRT2

SIRT2 in the cytoplasm deacetylates α-tubulin [85, 86], which results in microtubule destabilization. This, in turn, prevents: the localization of NLRP3 to the mitochondria; the approximation of ASC to NLRP3; the interaction of NLRP3 with mitochondrial products such as oxidized mitochondrial DNA or cardiolipin. Several NLRP3 activators, including ATP, nigericin, or MSU, but not activators of the AIM2 or NLRC4 inflammasomes, induce a depletion of NAD⁺, a fact that probably inhibits SIRT2, given the concurrent accumulation of acetylated tubulin. Interestingly, NAD⁺ supplementation during inflammasome induction inhibits the acetylation of tubulin and the activation of the inflammasome. AGK2, a potent inhibitor of SIRT2, stimulates the NLRP3 inflammasome [85]. Accordingly, inhibitors of tubulin polymerization, such as colchicine, suppress the activation of the NLRP3 inflammasome [85].

SIRT3

SIRT3 regulates mitochondrial ROS levels by deacetylating and increasing the activity of SOD2 [33, 87–89] and isocitrate dehydrogenase 2 [90]. We recently showed that fasting increases the activity and decreases the acetylation of SOD2 in human PBMCs and this correlates with decreased NLRP3 inflammasome activation [91]. It is possible that fasting is increasing the levels of mitochondrial NAD⁺ in these cells, as has been previously reported for other cell types [30], and this triggers an increase in the activity of SIRT3. The supplementation of the NAD⁺ precursor nicotinamide riboside (NR) to human leukocytes recapitulates the effects of fasting on SOD2 activity and inflammasome activation, suggesting that SIRT3 is involved in this process. Indeed, overexpression of SIRT3 in THP-1 cells inhibited the activation of the NLRP3 inflammasome, whereas its knockdown stimulated its activation [91]. Not surprisingly, SIRT3 inhibits the NLRP3 inflammasome in a mouse model of sepsis-induced acute kidney injury by attenuating ROS production [92], and is also protective in a model of LPS-induced lethality [93]. SIRT3 deficiency accelerates the development of metabolic syndrome in mouse, and single nucleotide polymorphisms in the human SIRT3 gene have been associated with this condition [40].

SIRT5

A direct role for SIRT5 on the NLRP3 inflammasome has not been shown. However, the addition of LPS to BMDM induces, in addition to an increase in succinate, an accumulation of succinylated proteins [46], which suggested a decrease in the activity of SIRT5. In fact, LPS stimulation decreased both NAD⁺ levels and SIRT5 expression [46]. The consequences of this increase in succinylated proteins or whether it contributes to the induction of the NLRP3 inflammasome has not been explored.

SIRT6

SIRT6 levels appear to increase with caloric restriction [94] and decrease on a high-fat diet [94]. However, there seem to be conflicting results on its role in inflammation. On one hand, SIRT6 is able to decrease the activity of the NF-κB pathway, as it is able to deacetylate and inactivate RelA [94] (possibly more strongly than SIRT1), and also able to deacetylate histone H3 at NF-κB target gene promoters [95], which causes their inactivation. Binding of SIRT6 to these promoters depends on its interaction with RelA [95]. In fact, exacerbated NF-κB-dependent gene expression contributes greatly to the shortened life span and degenerative symptoms observed in Sirt6 knockout mice, which die at around 4 weeks of age [96], as RelA heterozygosity attenuates them and extends life span to more than 3 months [95]. This modulation of the NF-κB pathway suggests that SIRT6 might modulate inflammasome priming. Indeed, SIRT6 overexpression and knockdown decreases and increases inflammation, respectively, in several models [97–99] with corresponding changes in the release of IL-1β. In parallel, SIRT6-overexpressing mice display improved healthspan and decreased adipose tissue inflammation [100]. Like SIRT1, SIRT6 may also modulate inflammation indirectly by regulating metabolism in the myeloid cell [83], since SIRT6 binds to the promoters of glycolytic genes and HIF-1α and deacetylates histone H3 which causes their inactivation [101]. Thus, SIRT6 turns off glycolysis after an acute inflammatory response, while SIRT1 turns on fatty acid oxidation in monocytes to restore homeostasis [83]. On the other hand, some reports have demonstrated a pro-inflammatory role for SIRT6, a fraction of which localizes to the endoplasmic reticulum where it promotes the secretion of TNF-α in MEFs, THP-1 cells and BMDM by removing the fatty acyl modification on TNF-α [102, 103]. Lysine fatty acylation helps to target TNF-α to lysosomes instead of to the endosomal/secretory pathway [104]. In addition, SIRT6 controls the differentiation and maturation of dendritic cells, as SIRT6 knockout mice display low frequencies of bone marrow dendritic cell precursors and low yields of bone marrow‐derived dendritic cells compared to wild-type mice, but increased frequencies of granulocytes and macrophages [105]. SIRT6 knockout dendritic cells were also shown to be immature. Interestingly, the capacity of those cells to secrete TNF-α and IL-6 was reduced in response to some Toll-like receptor ligands, but increased in response to others, suggesting that under some circumstances the inhibitory effects of SIRT6 on the NF-κB pathway may overcome the stimulatory effects of SIRT6 on direct TNF-α secretion [105].

A schematic depicting the roles of sirtuins in NLRP3 inflammasome activation in response to caloric load is shown in Fig. 4.

Fig. 4.

The caloric load-dependent role of sirtuins in the control of the NLRP3 inflammasome. Under conditions of caloric sufficiency, the relative NAD⁺ levels are reduced with diminished sirtuin activity in the nucleus (SIRT1 and 6), cytoplasm (SIRT2) and the mitochondria (SIRT3). The resultant acetylation of canonical sirtuin targets predisposes to upregulation of inflammasome encoding genes, generation of ROS and oxidized mitochondrial DNA (Ox mtDNA) in the mitochondria, and in the spatial trafficking of mitochondria towards the endoplasmic reticulum (ER) platform to trigger NLRP3 assembly. Conversely, with caloric restriction, the NAD⁺ levels are higher, with sirtuin activation and chromatin and protein deacetylation, resulting in counter-regulatory controlling events to blunt the NLRP3 inflammasome

Strategies and therapeutics to inhibit the NLRP3 inflammasome

The genetic elimination of the NLRP3 inflammasome in mice protects against cardiovascular [106], neurologic [107] and metabolic diseases [108] and protects against age-related functional decline [109]. Given these diverse ameliorative effects investigators are actively searching for NLRP3 antagonists. The most widely used class of inflammasome-associated therapeutics to date includes anakinra and related compounds. These agents directly target IL-1β signaling to blunt inflammasome-linked diseases [110]. This strategy, however, does not directly target the NLRP3 molecule or complex and, therefore, would not address IL-18 as the second inflammasome-activated cytokine.

Recently, MCC950 was identified as a direct small molecule inhibitor of the NLRP3 inflammasome [111]. This compound was found to block experimental diseases associated with NLRP3 inflammasome activation as well as ex vivo inhibition of the inflammasome in PBMCs from patients with CAPS-associated mutations [111]. Interestingly, the neurotransmitter dopamine has also been found to inhibit the NLRP3 inflammasome via dopamine D1 receptor-mediated E3-ligase activation that results in the proteasome-dependent degradation of NLRP3 [112]. Cyclic adenosine monophosphate (cAMP) acts downstream of dopamine D1 receptor to exert its inhibitory effects [112], and remarkably this metabolite has been associated with inflammasome inhibition in other studies [67, 113, 114]. In fact, a recent study has proposed that protein kinase A (PKA), a serine/threonine kinase that is activated by increasing intracellular concentrations of cAMP, inhibits NLRP3 by phosphorylating it on serine 295 of its nucleotide-binding domain, and that this mechanism of inhibition is disrupted in many CAPS-associated mutations [113]. Thus, this pathway seems interesting for future therapy.

The focus of this review has revolved around the mitochondria and metabolic perturbations that activate the NLRP3 inflammasome. In this context, a question arises as to whether metabolic intermediates, the regulation of metabolism and/or the modulation of mitochondrial integrity could also function to modulate inflammasome activation. A study by Youm et al. showed that the NLRP3 inflammasome activation was inhibited by the ketone body β-hydroxybutyrate (BHB) but not by acetoacetate or by related short-chain fatty acids butyrate and acetate [115]. BHB conferred this inhibition via the prevention of K⁺ efflux and by reducing ASC oligomerization and speck formation [115]. The veracity of the effect of BHB on the NLRP3 inflammasome was validated by the reduction in IL-1β production by human monocytes and in numerous murine models of NLRP3 inflammasome activation [115]. Interestingly, in this same study, the effects of NLRP3 inflammasome suppression by BHB did not appear to be dependent on the inhibition of ROS or other starvation-induced pathways including AMPK activation.

Interestingly, and in contrast to the Youm et al. study, a 24-h fasting and refeeding study in humans found that the fasting-mediated blunting of the NLRP3 inflammasome was associated with SIRT3 activation and the suppression of ROS [91]. At the same time, in that study refeeding was found to prime the NLRP3 inflammasome via the activation of NF-κB, but the identity of the dietary triggers has not been explored [91]. Furthermore, in the same study, the in vitro supplementation with the NAD⁺-salvage metabolite NR blunted refeeding activation effect on the NLRP3 inflammasome. Interestingly, other groups have found a similar inhibitory effect of NAD⁺ repletion by NR supplementation on hepatosteatosis, hepatic inflammation and NLRP3 inflammasome activation in mice [76, 116, 117]. Whether NR could function to suppress the NLRP3 inflammasome has the potential to be explored as a therapeutic agent, given its myriad of other beneficial effects in vivo [118–121]. Accordingly, NAD⁺ repletion by other compounds such as nicotinamide mononucleotide (NMN) similarly ameliorated age-associated adipose tissue inflammation [122], HFD-induced liver inflammation [37] and fructose-rich diet-induced pancreatic islet β cell IL-1β secretion and dysfunction in mice [123]. These effects of NR and NMN appear to be largely mediated by SIRT1 and/or SIRT3 [117, 121, 124]. Not surprisingly, resveratrol, the polyphenol present in the skin of grapes, another sirtuin activator, also inhibits NLRP3 inflammasome activation in mice by preserving mitochondrial function, and protects against hepatosteatosis, renal inflammation and LPS-induced lung injury [125–127]. Interestingly, recent evidence suggests that resveratrol activates sirtuins indirectly by increasing NAD⁺ levels in an AMPK-dependent manner [128]. Conversely, NAD⁺ depletion by the NAD⁺-consuming enzyme poly-ADP-ribose polymerase (PARP) is thought to facilitate inflammation, and thus PARP inhibitors such as 3-aminobenzamide (3-AB) inhibit IL-1β generation [129]. Collectively, this evidence suggests that the modulation of the levels of NAD⁺ should be explored as a therapeutic agent for the NLRP3 inflammasome.

The role of metabolic intermediates is further validated in that monounsaturated fatty acids also blunt the NLRP3 inflammasome in adipocytes and this effect appeared to be mediated, in part, via the activation of AMPK [130]. Given the role of mitochondrial integrity and signaling in the modulation of the NLRP3 inflammasome it is interesting to note that the anti-diabetic agent metformin, which blunts mitochondrial ROS production via the inhibition of complex I of the ETC has been found to have anti-inflammatory effects which include the suppression of IL-1β production [131]. The direct effect of metformin on the NLRP3 inflammasome was, however, not explored in that study. Other antioxidant compounds that blunt ROS production have been found to be inhibitors of the inflammasome in vivo, such as the mitochondrial ROS scavenger MitoQ, which suppresses activation of the NLRP3 inflammasome in a model of dextran sulfate-induced colitis [132], or natural compounds such as Aloe vera extract, emodin, ginseng or curcumin [133]. The broad-spectrum tetracycline antibiotic Minocycline also inhibits the NLRP3 inflammasome in several disease models by a mechanism that involves the dampening of ROS generation [134–136].

Conclusions

Caloric overload has long been recognized as a risk factor for insulin resistance and the development of type 2 diabetes. These disease states are associated with sterile inflammation which has now recognized to result from the harnessing of the intracellular NLRP3 inflammasome immune surveillance program. At the same time nutrients themselves have been identified to activate the NLRP3 inflammasome and could play an important role in the perpetuation of insulin resistance and inflammation. A parallel concept has been the recognition that the disruption of mitochondrial integrity is linked to insulin resistance and has been found to be an integral component in inflammasome execution. Given that there are active investigations to identify therapeutic strategies to enhance mitochondrial functioning, may not only be a useful target for direct mitochondrial linked pathologies, but may also be important in blunting the inflammasome, with the possible additional benefit of ameliorating insulin resistance and other sterile inflammation-linked diseases. This exciting possibility remains to be directly explored.