Curcumin Suppresses IL-1β Secretion and Prevents Inflammation through Inhibition of the NLRP3 Inflammasome

Turmeric, a common oriental spice that gives curry powder its yellow color, is frequently used in Asian cooking, particularly Indian, Pakistani, and Thai cooking. Curcumin, a polyphenolic compound, is the principal curcuminoid derived from the rhizomes of turmeric (1). The other two curcuminoids derived from turmeric extract are desmethoxycurcumin (DMC, 15%) and bisdemethoxycurcumin (BDMC, 5%) (2). Turmeric has a long history of use in Ayurvedic medicine as a treatment for inflammatory conditions. Because of its numerous biological activities and health benefits, curcumin is of current interest, especially as an anti-inflammatory agent (3). In particular, both in vitro and in vivo studies have demonstrated that curcumin can suppress inflammation with no associated toxicities and plays a beneficial role in a variety of inflammatory diseases, including obesity, diabetes, cardiovascular diseases, bronchial asthma, and rheumatoid arthritis (4). However, despite the mounting evidence of the anti-inflammatory effects of curcumin, large gaps in knowledge still exist regarding its mechanisms of action.

Inflammasomes are multimeric protein complexes that orchestrate host defense mechanisms against pathogen-associated molecular patterns released by infectious agents and danger-associated molecular patterns released during noninfectious physiological damage (5). Assembly of the inflammasome complex is initiated by a nucleotide-binding domain and leucine-rich repeat receptors or absent in melanoma 2–like receptors. NOD-like receptors and AIM2-like receptors mediate host recognition of a diverse set of inflammation-inducing stimuli and control the production of highly proinflammatory cytokines IL-1β and IL-18 (6). Additionally, inflammasome activation causes a rapid, proinflammatory form of cell death called pyroptosis (7).

The NLRP3 inflammasome is the most characterized inflammasome. As an important innate immune sensor, it is activated by a wide range of signals of pathogenic, endogenous, and environmental origin and consists of the NLRP3 scaffold, the apoptosis-associated speck-like protein containing a CARD (ASC), and the inflammatory protease caspase-1 (8). Given the emerging role of the NLRP3 inflammasomes in type 2 diabetes, obesity, and other autoinflammatory diseases, identification of potential agents and their mechanisms that control NLRP3 inflammasome deactivation may provide insights into the control of several chronic inflammatory disorders (9–11).

Curcumin has been shown to inhibit the production of the proinflammatory cytokine IL-1β (12). This finding prompted us to investigate the possible role of curcumin in controlling inflammasome activity. In this study, we found that curcumin suppressed inflammation via strong inhibition of NLRP3-dependent caspase-1 activation and IL-1β secretion. Furthermore, curcumin attenuated the excessive IL-1β in high-fat diet (HFD)–treated wide-type (WT) mice but not in Nlrp3-deficient mice. Finally, curcumin prevented inflammation in a monosodium urate crystal (MSU)–mediated peritonitis model. These findings suggest that the NLRP3 inflammasome is an important mediator of the anti-inflammatory actions of curcumin.

Nlrp3^−/− mice were described previously (13). All mice were from a C57BL/6 background and were housed with an alternating 12-h light/12-h dark cycle. All animal experiments were performed according to the Guidelines for the Care and Use of Laboratory Animals and were approved by the Ethics Committee of Shandong University.

Curcumin, nigericin, MSU, PMA, poly(deoxyadenylic-deoxythymidylic) acid [poly(dA:dT)], celecoxib, 3-methyladenine, and glucose were purchased from Sigma-Aldrich (St. Louis, MO). Imiquimod (R837) and ultrapure LPS, MitoTracker Deep Red, and MitoSOX were obtained from InvivoGen (San Diego, CA). DMC and BDMC were from Shifeng Biological Technology (Shanghai, China). Tetrahydrocurcumin (THC) was from Natural Remedies (Bangalore, India). Salmonella typhimurium was a donated from the microbiology laboratory of the Institute of Basic Medical Sciences, Shandong Academy of Medical Sciences. The Ab to human IL-1β was from Abcam (Cambridge, MA). Anti–human caspase-1, anti–β-actin, anti–mouse tubulin, anti–α-tubulin (acetyl K40), and anti–LC3 Ab were from Cell Signaling Technology (Danvers, MA). The Ab against mouse IL-1β and the ELISA kits for human and mouse IL-1β, mouse TNF-α, and IL-18 were from R&D Systems (Minneapolis, MN). Anti-mouse caspase-1 (p20) and anti-NLRP3 were from Adipogen (San Diego, CA). The Abs targeting ASC, sirtuin 2 (SIRT2), and Tom20 were from PTG (Wuhan, China). Asante Potassium Green 2 (APG-2) was purchased from Teflabs (Austin, TX). The NAD/NADH quantification kit was from BioVision (Milpitas, CA). All tissue culture reagents were bought from Invitrogen.

THP-1 cells were grown in RPMI 1640 medium supplemented with 10% FBS and 50 mM 2-ME. THP-1 cells were differentiated for 3 h with 100 nM PMA. Mouse bone marrow–derived macrophages (BMDMs) were derived from the femurs and tibias of C57BL/6 mice and cultured in DMEM medium complemented with 10% FBS in the presence of L929 culture supernatants.

For inducing NLRP3 inflammasome activation, 1.0 × 10⁶ macrophages were plated in 12-well plates overnight, then the medium was changed to Opti-MEM, cells were primed with LPS (500 ng/ml) for 3 h, and curcumin or indicated compounds were added for another hour. Cells were then stimulated with MSU (150 μg/ml), aluminum salts (Alum; 300 μg/ml), and R837 (15 μg/ml) for 6 h or with nigericin (10 μM) for 40 min. For AIM2 inflammasome activation, poly(dA:dT) (0.5 μg/ml) was transfected using Lipofectamine according to the manufacturer’s instructions. After 6 h, the supernatants and cell lysates were collected and analyzed for caspase-1 and IL-1β activation by immunoblotting and for IL-1β levels by ELISA.

S. typhimurium was cultured overnight. On the following day, BMDMs were primed with LPS for 3 h, infected with the Salmonella (1:200) for 1 h, and then either untreated or treated with curcumin in the presence of gentamicin. Supernatants and cell lysates were collected 3 h after treatment.

Supernatants from cell culture or mouse sera were collected and stored at −80°C, then assayed for mouse IL-1β, mouse IL-18, mouse TNF-α, and human IL-1β, according to the manufacturer’s instructions.

Medium supernatants from treated macrophages were precipitated and whole-cell lysates were prepared. The samples were separated by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and hybridized with primary Abs. The immune complexes were incubated with HRP-conjugated secondary Ab. The membranes were scanned with an LAS-4000 luminescent image analyzer (Fujifilm, Tokyo, Japan).

LPS-primed BMDMs were plated at a density of 10⁵ cells per well in 96-well plates and treated with nigericin in the presence of curcumin. Culture supernatants were collected and assayed with a lactate dehydrogenase (LDH) cytotoxicity assay kit (Beyotime, Shanghai, China) according to the manufacturer’s instructions.

Macrophages were plated overnight on coverslips and stimulated as described above. After stimulation, cells were washed, fixed with 4% paraformaldehyde, and processed for immunocytochemistry. MitoSOX and MitoTracker Deep Red staining were used according to the manufacturer’s instructions. Genomic DNA was stained with DAPI. Samples were imaged with an LSM 710 confocal laser scanning microscope (Carl Zeiss). The fluorescence intensity in images was analyzed by ImageJ.

Intracellular NAD⁺ levels were measured by an NAD/NADH quantification kit. The NAD⁺ level in treated BMDMs was divided by the NAD⁺ level in the untreated control BMDMs to determine the relative concentration (presented in arbitrary units).

LPS-primed BMDMs were loaded with APG-2, a fluorescent indicator of intracellular K⁺, and simultaneously incubated with or without curcumin. Cells were then stimulated with nigericin. After washing, the cells were fixed with 4% paraformaldehyde and counterstained with DAPI. Confocal microscopy analyses were performed at 488 and 405 nm.

Additionally, K⁺ in BMDMs was also measured with inductively coupled plasma optical emission spectrometry (ICP-OES). Cells were digested in boiling nitric acid and samples were diluted to 3–8% (v/v) HNO₃ with dH₂O. The K⁺ was detected using an Optima 7300 DV ICP-OES. Samples and standards were run in duplicate.

BMDMs were cultured on chamber slides overnight, then primed with LPS and treated with nigericin in the presence or absence of curcumin. For ASC oligomerization detection, BMDMs were seeded on six-well plates (2 × 10⁶ cells per well) and treated with different stimuli. The cells were pelleted by centrifugation, resuspended in 1% Nonidet P-40 lysis buffer (0.5 ml, ice-cold), and lysed by shearing 10 times through a 21-gauge needle. Cell lysates were then centrifuged (5000 × g, 10 min, 4°C). Pellets were washed twice with PBS, resuspended in PBS (500 μl), crosslinked with fresh disuccinimidyl suberate (2 mM, 37°C, 30 min), and pelleted by centrifugation (5000 × g, 10 min). Crosslinked pellets were resuspended in SDS sample buffer (20 μl), separated using 12% SDS-PAGE, and immunoblotted using anti-mouse ASC Abs.

To measure ASC speckles, cells were fixed with 4% paraformaldehyde followed by ASC and DAPI staining. ASC speckles were quantified using ImageJ software. At least four distinct fields were analyzed.

For metabolic studies, male mice (6 wk old) were fed with normal diet (ND), HFD (D12492, Research diets), or HFD with curcumin supplementation for 12 wk. Curcumin (150 mg/kg daily) was administered by gavage during weeks 9–12. For glucose tolerance tests, mice were fasted overnight (16 h) and were i.p. injected with glucose (1.5 mg/g body weight). Blood samples were drawn 0, 30, 60, 90, and 120 min after injection. For insulin tolerance tests, mice were fasted for 2 h and then injected with recombinant human insulin (1.5 IU/kg body weight). Blood glucose was measured with a glucometer (Roche). At the end of each experiment, mice were euthanized. Serum cytokines were measured by ELISA. Liver tissues were isolated, washed in cold PBS supplemented with penicillin and streptomycin, and cultured in 12-well plates in Opti-MEM medium supplemented with penicillin and streptomycin. After 24 h, supernatants were collected and stored at −80°C until analyzed.

For peritonitis, mice were treated with curcumin (100 mg/kg body weight, i.p. injection). After 1 h, peritonitis was induced by injection of MSU (3 mg in 200 μl of sterile PBS). Mice were euthanized 6 h later and peritoneal cavities were washed with 10 ml of PBS. Lavage fluids were analyzed for IL-1β production by ELISA and for polymorphonuclear neutrophil recruitment by flow cytometry using the neutrophil marker Ly6G-PE. Serum IL-1β was also measured.

Samples were analyzed using a Student t test unless indicated in the figure legends. Differences between treatments with p < 0.05 were considered statistically significant.

To test whether curcumin affects inflammasome activation, we first examined whether curcumin could inhibit caspase-1 cleavage and IL-1β secretion. We pretreated LPS-primed BMDMs with curcumin for 1 h and then treated them with the NLRP3 inflammasome activator nigericin. We measured both caspase-1 activation and IL-1β maturation using immunoblots that detect the enzymatically active p20 subunit of caspase-1 and the biologically active p17 form of IL-1β, respectively. Curcumin dose-dependently blocked nigericin-induced cleavage of caspase-1 into p20 and the IL-1β maturation (Fig. 1A, 1B). Similarly, curcumin suppressed the secretion of IL-18, another inflammasome-dependent cytokine (Fig. 1C). However, curcumin had no effect on TNF-α production or pro–IL-1β expression (Fig. 1A, 1D). These results suggest that curcumin inhibited IL-1β maturation independently of priming in LPS-primed BMDMs.

Because the inhibition of NF-κB activation and TNF-α production has been implicated in the anti-inflammatory activity of curcumin, we sought to determine whether curcumin had an impact on LPS-induced priming for inflammasome activation (14). BMDMs were treated with curcumin for 3 h, then primed with LPS for 3 h, and finally stimulated with nigericin. Curcumin inhibited LPS-induced pro–IL-1β expression and TNF-α production (Fig. 1E, 1F). The caspase-1 activation and IL-1β secretion were also blocked (Fig. 1E, 1G). The inhibition of priming was not as strong as the inflammasome inhibition detected at similar concentrations of curcumin. These results suggest that curcumin can inhibit caspase-1 activation and IL-1β secretion by suppressing both LPS priming and inflammasome activation.

To identify the mechanisms underlying curcumin-induced inflammasome suppression, we treated BMDMs with curcumin after LPS priming in the subsequent experiments. To exclude the possibility that curcumin inhibition of IL-1β secretion was due to cell death, cell death was analyzed by the LDH release assay under similar experimental conditions. Curcumin had no effect on cell viability (Fig. 1H), suggesting that the inflammasome inhibition is not due to cell death.

Besides nigericin, NLRP3 inflammasome can be activated by R837 or crystalline substances, such as Alum and MSU (8). To determine whether curcumin specifically targets common signaling mechanisms in response to structurally diverse NLRP3 activators, we assessed other NLRP3 agonists. As found with nigericin, curcumin blocked inflammasome activation by R837, Alum, and MSU (Fig. 2A, 2B). We further investigated the specificity of curcumin to NLRP3 as compared with other inflammasomes. The LPS-primed BMDMs were either infected with S. typhimurium to activate the NLRC4 inflammasome or transfected with poly(dA:dT) to activate the AIM2 inflammasome. Our data showed that neither the NLRC4 inflammasome nor the AIM2 inflammasome was inhibited by curcumin (Fig. 2C, 2D). We also detected that with regard to expression of the NLRP3 inflammasome components, neither NLRP3 nor ASC is modified by curcumin (Supplemental Fig. 1). These results indicate that curcumin acts on a common signaling pathway specific to the NLRP3 inflammasome activated by broad proinflammatory activators.

To investigate whether curcumin exerts the same anti-inflammatory effects in humans, we primed PMA-differentiated THP-1 cells with LPS, then treated them with curcumin or vehicle for 1 h, and finally stimulated the cells with nigericin. Results obtained by immunoblotting indicated that the cells treated with curcumin showed markedly reduced processing of caspase-1 and secretion of mature IL-1β (Fig. 2E). Taken together, these findings demonstrate that curcumin can specifically inhibit the NLRP3 inflammasome in both mice and humans.

In addition to curcumin, the other curcuminoids in turmeric are two structurally related derivatives that lack either one (DMC) or both (BDMC) methoxy groups in the phenyl rings (Fig. 3A). We found that both DMC and BDMC blocked caspase-1 cleavage, IL-1β maturation, and IL-18 secretion at similar concentrations to curcumin when the primed BMDMs were stimulated with nigericin (Fig. 3B–D).

THC is the main metabolite of curcumin that lacks the α,β-unsaturated ketone groups (Fig. 3A) (15). THC pretreatment had no significant effect on nigericin-induced caspase-1 activation or IL-1β or IL-18 secretion in BMDMs (Fig. 3B–D). These results suggest that the α,β-unsaturated ketone groups, but not the methoxy group, are critical for curcuminoid-mediated inflammasome inhibition.

Recently, it was suggested that oxidative activation should be considered as a potential mechanism of action of curcumin and that enzymatic oxidation of curcumin occurs by cyclooxygenase-2 (COX-2) (16). We found that inhibition of COX-2 activity by celecoxib did not significantly alter the inhibitory effects of curcumin on the inflammasome (Fig. 3E, 3F). This result suggests that the inhibitory effects of curcumin on IL-1β secretion are not due to COX-2–modified enzymatic products.

Several molecular and cellular events have been proposed as the trigger for NLRP3 inflammasome activation. Potassium (K⁺) efflux is thought to act at or upstream of NLRP3 activation and it has emerged as a common event and plays a critical role in NLRP3 activation (17). To assess whether curcumin inhibited NLRP3 activation by suppressing K⁺ efflux, we incubated primed BMDMs with APG-2, a fluorescent indicator of cytosolic K⁺. Consistent with recent studies, curcumin prevented the decline of intracellular K⁺ in response to incubation with the NLRP3 activators nigericin and MSU (Fig. 4A) (18). Results were confirmed by ICP-OES (Fig. 4B, 4C).

In addition to K⁺ efflux, mitochondrial dysfunction and reactive oxygen species (ROS) have also been proposed to be important signals responsible for NLRP3 inflammasome activation. Studies have shown that NLRP3 inflammasome inducers cause mitochondrial-associated dysfunction and that the characteristics of mitochondrial damage are the production of ROS, lower mitochondrial membrane potential (ΔΨm), a reduction in intracellular NAD⁺, and morphological changes of mitochondria from string-shaped to dot-shaped structures (19, 20). We wanted to determine whether reduced mitochondrial impairment mediated curcumin’s effects on the NLRP3 inflammasome. First, we examined mitochondrial ROS (mtROS) production by MitoSOX, a mitochondrial superoxide indicator. Consistent with recent data, curcumin decreased mtROS production in BMDMs upon treatment with nigericin (Fig. 4D) (18). In contrast, curcumin did not show such reduction in BMDMs upon treatment with MSU. Furthermore, although mitochondrial complex I inhibitor rotenone robustly increased mtROS production (Fig. 4D), it failed to abrogate the suppressive effects of curcumin on nigericin-induced NLRP3 inflammasome activation significantly (Fig. 4E), indicating that reduction in mtROS was not a general feature of curcumin on NLRP3 inflammasome inhibition.

Additionally, we assessed the functional mitochondrial pool in macrophages using MitoTracker Deep Red, a fluorescent probe sensitive to the mitochondrial inner transmembrane potential. Pretreatment with curcumin decreased ΔΨm further in LPS-primed BMDMs upon nigericin stimuli (Fig. 4F). Consistent with the pattern of ΔΨm, a lower abundance of NAD⁺ was obtained in the presence of curcumin (Fig. 4G). Furthermore, curcumin did not significantly alter nigericin-induced morphological changes of mitochondria in BMDMs; mitochondria still changed into dot-like shapes. These results indicate that curcumin fails to block the nigericin-induced mitochondrial damage, ruling it out as the mechanism by which curcumin inhibits the inflammasome.

Autophagy is a negative regulator targeting mitochondria for degradation or other pathways (20, 21). Furthermore, induction of autophagy has also been linked to the biological effects of curcumin (22). We found that curcumin treatment did not significantly induce autophagy in primed BMDMs, and the autophagy inhibitor 3-methyladenine failed to abrogate the inhibitory effects of curcumin (Fig. 4H, 4I), ruling out a major role for autophagy in the effects of curcumin on the inflammasome.

Our data suggest that curcumin inhibits NLRP3 inflammasome activation by controlling an unknown upstream event that reduces K⁺ efflux from macrophages, and not by preventing mitochondrial destabilization.

It has been suggested that ASC on mitochondria moves to the perinuclear region and localizes together with NLRP3 on the endoplasmic reticulum during activation of the NLRP3 inflammasome. Moreover, this microtubule-driven spatial arrangement of mitochondria provides a platform for complex assembly and is necessary for NLRP3 inflammasome activation (19, 23). Additionally, tubulin has been reported to be a target of curcumin (24). To clarify the role of the subcellular localization of mitochondria in curcumin’s inhibition of NLRP3, we examined the subcellular location of Tom20 and α-tubulin in primed BMDMs pretreated with curcumin and then stimulated with nigericin. Colchicine, an inhibitor of tubulin polymerization, was used as a positive control, as it inhibits NLRP3 inflammasome activation by breaking the microtubule structure and subsequently disturbing the mitochondrial subcellular localization. Similar to colchicine, imaging revealed that curcumin blocked the nigericin-induced transportation of mitochondria to the perinuclear region (Fig. 5A).

Acetylated α-tubulin has been reported to mediate mitochondrial transport to promote activation of the NLRP3 inflammasome (23). However, curcumin promoted, rather than suppressed, the accumulation of acetylated α-tubulin during nigericin-induced activation of the NLRP3 inflammasome (Fig. 5B). Consistent results were obtained in BMDMs by immunoblotting (Fig. 5C). We further detected SIRT2, α-tubulin deacetylase specifically involved in the process for activation of the NLRP3 inflammasome, by immunoblotting. We found that curcumin did not alter the SIRT2 expression during nigericin-induced activation of the NLRP3 inflammasome in BMDMs (Fig. 5C). Studies have suggested that curcumin can act as a histone deacetylase inhibitor (25). However, we found that inhibition of histone deacetylases using trichostatin A (TSA) did not significantly affect NLRP3 inflammasome activation in LPS-primed and nigericin-treated macrophages despite robust induction of α-tubulin acetylation in macrophages (Fig. 5B, 5D). Our data suggest that the blocking of mitochondrial transport is independent of acetylated α-tubulin.

Besides caspase-1 activation and IL-1β release, an additional marker for NLRP3 inflammasome activation is the formation of ASC nucleation-induced polymerization or oligomerization, a large structure localized to a detergent-insoluble fraction in the cell that is thought to mediate caspase-1 activation (7, 26). Consistent with previous findings, immunofluorescent staining for endogenous ASC in BMDMs showed that the polymerization or oligomerization of ASC into a large protein speckle at the perinuclear region was triggered upon activation with nigericin (Fig. 5E). Treatment of LPS-primed BMDMs with curcumin markedly reduced the formation of nigericin-induced ASC speckles (Fig. 5E, 5F). Similar results were obtained by immunoblotting; curcumin prevented nigericin-induced or MSU-induced ASC oligomerization in primed BMDMs (Fig. 5G). These experiments confirm that curcumin inhibits the NLRP3 inflammasome also by blocking mitochondrial transport, ASC polymerization, and assembly of the inflammasome.

Increasing evidence suggests the benefits of curcumin in the prevention and treatment of diabetes and its associated disorders (4, 27). Additionally, it has been confirmed that the NLRP3 inflammasome can produce the proinflammatory cytokines and is critically involved in insulin resistance in pancreatic islets and in adipose, liver, and kidney tissues (10, 28). We investigated whether curcumin can improve insulin sensitivity via inhibition of NLRP3 inflammasome activation. Mice were fed an HFD for 12 wk to induce the emergence of insulin resistance. To assess whether curcumin can reverse insulin insensitivity, we performed glucose tolerance tests and insulin tolerance tests in HFD-treated mice with or without curcumin supplementation. Following i.p. injection of glucose, plasma glucose levels in HFD-treated WT mice were significantly higher compared with those in WT mice fed an ND (Fig. 6A). Curcumin treatment reduced the glucose elevation and the mice were more glucose tolerant than HFD-treated WT mice without curcumin administration. Meanwhile, by measuring the reduction in plasma glucose after insulin treatment, the insulin tolerance tests showed that insulin sensitivity was markedly improved by curcumin in HFD-treated WT mice (Fig. 6B). In contrast, in Nlrp3^−/− mice, curcumin did not significantly improve glucose tolerance or insulin sensitivity (Fig. 6A, 6B). Moreover, no significant weight changes of mice were observed after treatment of curcumin (Supplemental Fig. 2). Therefore, curcumin-mediated prevention of metabolic disorders may rely on inhibition of the NLRP3 inflammasome.

To confirm that curcumin prevents HFD-induced insulin resistance by inhibition of NLRP3 inflammasome activation, we examined NLRP3 inflammasome activation in vivo. Serum concentrations of IL-1β and IL-18 in HFD-treated mice were higher than in ND-fed mice, and the robust elevation was blocked by curcumin. However, Nlrp3 deficiency prevented HFD-induced IL-1β and IL-18 production (Fig. 6C, 6D). Importantly, Nlrp3 deficiency abrogated curcumin inhibition of IL-1β and IL-18 production (Fig. 6C, 6D).

In accord with the above results, liver tissue from HFD-treated mice showed higher IL-1β or IL-18 production compared with ND-treated mice. Moreover, curcumin administration and Nlrp3 deficiency both blocked HFD-induced IL-1β or IL-18 production in the liver, indicating that curcumin supplementation can suppress metabolic stress–induced NLRP3 inflammasome activation in HFD-treated mice (Fig. 6E, 6F). Taken together, these results indicate that curcumin can prevent HFD-induced insulin resistance by blocking NLRP3 inflammasome activation.

Next we verified whether delivery of curcumin can inhibit the NLRP3 inflammasome in mouse models of NLRP3-driven inflammation in vivo. The NLRP3 inflammasome was activated following i.p. injection of MSU crystals, resulting in an influx of neutrophils into the peritoneum and increased secretion of IL-1β by macrophages 6 h after injection. Compared to mice given vehicle, curcumin treatment inhibited MSU-induced IL-1β production and reduced neutrophil infiltration into the peritoneum, suggesting direct effects of curcumin on NLRP3-driven peritoneal inflammation in vivo (Figs. 7, 8).

In this study, we identified curcumin as a common NLRP3 inflammasome activation inhibitor. Curcumin inhibited both LPS-priming and NLRP3 inflammasome activation pathway in macrophages. The inhibitory effect of curcumin on inflammasome activation was specific to the NLRP3 inflammasome, and it did not affect the NLRC4 or AIM2 inflammasomes. Importantly, curcumin delivery could prevent HFD-induced insulin resistance and MSU-induced peritoneal inflammation by inhibition of the NLRP3 inflammasome in vivo. Taken together, our results demonstrate a previously unrecognized mechanism through which curcumin repress inflammation in diabetes and other NLRP3-driven diseases (Fig. 8).

The NLRP3 inflammasome is activated in a two-checkpoint activation mechanism. First, LPS or another TLR agonist induces NLRP3 and pro–IL-1β synthesis via the NF-κB pathway. Second, the NLRP3 inflammasome and the subsequent caspase-1 processing can be activated by nigericin, MSU crystals, or other stimuli. Dysregulated NLRP3 inflammasome activity is associated with a wide range of diseases, including diabetes and autoinflammation (10, 29, 30). Additionally, work from our laboratory has shown that omega-3 fatty acids and the endogenous neurotransmitter dopamine exerts anti-NLRP3 inflammasome effects (13, 31). Recently, various plant-derived polyphenols have been identified as inhibitors of NLRP3 inflammasome, including isoliquiritigenin, resveratrol, emodin, and epigallocatechin-3-gallate (32–35). It is well known that curcumin is one of the most studied polyphenol compounds. How, then, does it affect NLRP3 inflammasome activation? We confirmed the ability of curcumin to suppress NLRP3 inflammasome activation by various stimuli, including crystalline molecules such as MSU and Alum, that require phagocytosis for activation, as well as pore-forming toxins such as nigericin (18). Also, similar to most regulators of the NLRP3 inflammasome reported, curcumin can prevent both steps to inhibit the NLRP3 inflammasome. The inhibition of NF-κB activity by curcumin has been discussed extensively, and hence we focused our investigation of how curcumin inhibits NLRP3 inflammasome activation in the present study.

The generation of ROS was one of the first intermediates discovered to be common to various stimuli-induced NLRP3 activation (36). Since then, there have been many conflicting reports regarding the role of ROS in this process, creating much controversy in understanding the regulation of NLRP3 inflammasome activation (37, 38). Our data reveal that the effects of curcumin vary depending on the nature of the NLRP3 activator; curcumin decreased the effects of nigericin, but enhanced the effects of crystalline substances, such as MSU. Therefore, mtROS is likely not the common pathway by which curcumin inhibits the NLRP3 inflammasome. Moreover, these results demonstrate that only mtROS are not enough to activate the NLRP3 inflammasome. The separation of ROS between nigericin and MSU might due to the redox state in cells induced by curcumin and various challenges or the cooperative sensitized effects between curcumin and MSU. Besides, lower ΔΨm and reduction of intracellular NAD⁺ are also the characteristics of mitochondrial damage, which plays a critical role in the activation of the NLRP3 inflammasome. However, our findings revealed that instead of preventing mitochondrial damage, curcumin resulted in more damage to it, exacerbated depolarization of ΔΨm, and further diminished the amount of cellular NAD⁺. This suggests that maybe curcumin inhibits NLRP3 inflammasome activation by impairing the mitochondria in a manner that prevents mitochondrial from responding to nigericin.

In addition to mitochondrial damage, the efficient spatial arrangement of mitochondria has been shown to be a limiting step in NLRP3 inflammasome activation. The NLRP3 inflammasome is a multiprotein complex that is composed of a sensor protein NLRP3, the adaptor protein ASC, and the inflammatory protease caspase-1. ASC on mitochondria bridges NLRP3 and caspase-1 to form ternary inflammasome complexes. Mitochondria act as a platform to facilitate the molecular complexity of sensor and adapter interactions that promote effective NLRP3 inflammasome activation. Recently, microtubules have been suggested to mediate the transport of mitochondria to create optimal sites for activation of the NLRP3 inflammasome (23). In the present study, we indeed found that curcumin blocked transportation of damaged mitochondria induced by nigericin.

Additionally, the pathway has been proposed that NLRP3 agonists lead to an abundance of acetylated α-tubulin that drives mitochondria to the perinuclear area to promote the NLRP3 inflammasome activation (23). Curcumin-mediated inhibition of NLRP3 inflammasome activation appears to be independent of the pathway mentioned above because curcumin delivery enhances, rather than inhibits, the accumulation of acetylated α-tubulin. Moreover, TSA did not affect NLRP3 inflammasome activation despite significant induction of tubulin acetylation. The results indicate that there is another alternative, acetylated α-tubulin–independent mechanism involved in the upstream of transport of mitochondria that promotes NLRP3 activation. As the molecular mechanisms by which NLRP3 agonist–driven spatial arrangement of mitochondria promotes activation of the NLRP3 inflammasome are currently unknown, future studies are needed to further characterize this pathway.

Proper maintenance of the delicate balance between immune response and metabolism is crucial for health and has important implications for many pathological states such as obesity, diabetes, and other chronic noncommunicable diseases (39). Increasing evidence suggests that NLRP3 inflammasome functions as a sensor to detect danger signals and induce downstream chronic, low-grade, metabolic inflammatory signaling that contributes to obesity and associated disorders such as insulin resistance (10, 28, 40). Proinflammatory cytokines, including IL-1β and TNF-α, are significantly elevated in diabetes. Curcumin has been reported to exert antidiabetic effects via inhibition of diabetes-induced increases in IL-1β, TNF-α, and NF-κB activity (4). However, some animal studies and several clinical trials using TNF-α blockade have failed to prevent insulin resistance, suggesting that TNF-α may not be the primary target of curcumin to exert beneficial effects (41). In the current study, we present that curcumin prevents NLRP3 inflammasome-dependent IL-1β production and HFD-induced insulin resistance in WT mice. More importantly, these effects were abrogated in Nlrp3^−/− mice. Although some other targets for curcumin, such as VEGF and PPAR-γ, have also been proposed, our data support a critical role for the inhibition of the NLRP3 inflammasome underlying the beneficial effect of curcumin in inflammatory disorders, at least in the HFD-induced model we explored (42, 43).

During past decades, as a potential treatment for diabetes and its associated complications, the function of curcumin has been investigated and the mainstream view at present is that curcumin is safe, nontoxic, and improves most of the complications of diabetes (1–4). In contrast, other researchers claim that curcumin has cytotoxic effects in vitro and no significant effect on blood glucose in vivo (44). Perhaps various disease models and ways of drug delivery or different bioavailability at physiologically achievable concentrations can account for the differences among findings. We found that curcumin demonstrated a time- and dose-dependent cytotoxic effect in BMDMs (data not shown), and curcumin also failed to inhibit the caspase-1–dependent pyroptotic cell death induced by nigericin, despite this compound blocking the ASC speckle formation. Moreover, although curcumin indeed reduced NLRP3 inflammasome–dependent IL-1β production in mice, curcumin did not significantly improve blood glucose, serum insulin, cholesterol, triglyceride, low-density lipoprotein, or high-density lipoprotein (data not shown). Thus, despite a broad spectrum of potentially beneficial pharmacological activities, many questions remain regarding the fate of this compound in the mammalian organism. More investigation is required to improve our understanding of curcumin and facilitate successful translation to human diseases.

Collectively, our findings demonstrate a direct inhibitory effect of curcumin on NLRP3 inflammasome activation in macrophages. Our results further show that dietary curcumin can prevent HFD-induced insulin resistance via specific inhibition of the NLRP3 inflammasome in vivo. It may provide a low-cost, well-tolerated addition to a high-calorie diet for preventing chronic, low-grade, metabolic inflammation. Moreover, the ability of curcumin to block NLRP3 inflammasome makes it an attractive new candidate for even more clinical applications, including diabetes, gout, Alzheimer disease, or other NLRP3-driven disorders.

We thank Rongbin Zhou (School of Life Sciences and Medical Center, University of Science and Technology of China) for providing the Nlrp3^−/− mice, and Peng Li (Microbiological Research Department, Institute of Basic Medicine, Shandong Academy of Medical Sciences) for providing S. typhimurium. We thank Dr. Hua-Ming Yu (Instruments’ Center for Physical Science, University of Science and Technology of China) for ICP-OES.

The authors have no financial conflicts of interest.