NLRP3 inflammasome mediates oxidative stress-induced pancreatic islet dysfunction
Abstract
Inflammasomes are multiprotein inflammatory platforms that induce caspase-1 activation and subsequently interleukin (IL)-1β and IL-18 processing. The NLRP3 inflammasome is activated by different forms of oxidative stress, and, based on the central role of IL-1β in the destruction of pancreatic islets, it could be related to the development of diabetes. We therefore investigated responses in wild-type C57Bl/6 (WT) mice, NLRP3−/− mice, and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC) after exposing islets to short-term hypoxia or alloxan-induced islet damage. NLRP3-deficient islets compared with WT islets had preserved function ex vivo and were protected against hypoxia-induced cell death. Furthermore, NLRP3 and ASC-deficient mice were protected against oxidative stress-induced diabetes caused by repetitive low-dose alloxan administration, and this was associated with reduced β-cell death and reduced macrophage infiltration. This suggests that the beneficial effect of NLRP3 inflammasome deficiency on oxidative stress-mediated β-cell damage could involve reduced macrophage infiltration and activation. To support the role of macrophage activation in alloxan-induced diabetes, we injected WT mice with liposomal clodronate, which causes macrophage depletion before induction of a diabetic phenotype by alloxan treatment, resulting in improved glucose homeostasis in WT mice. We show here that the NLRP3 inflammasome acts as a mediator of hypoxia and oxidative stress in insulin-producing cells, suggesting that inhibition of the NLRP3 inflammasome could have beneficial effects on β-cell preservation.
INTRODUCTION
Type 1 (T1D) and type 2 (T2D) diabetes mellitus have traditionally been considered to have different pathologies. T1D is defined as an autoimmune disease characterized by a specific immune-mediated destruction of the insulin-producing β-cells (29). T2D is currently regarded as a condition driven by low-grade chronic inflammation in response to excessive nutrients and metabolic stress, and the innate immune system seems to be centrally involved in eliciting this metabolic inflammation (i.e., metaflammation) (3, 27). Notably, evidence also points to a central role of the innate system not only in the initiation of T1D but also directly in the destruction of β-cells (25).
The nucleotide-binding oligomerization domain (NOD), leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) is a central component of the innate immune system. The NLRP3 inflammasome is activated after assembly of the NLRP3 protein, and the adaptor apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC) resulting in the activation of caspase-1 that catalyzes the proteolytic activation of the inflammatory cytokines interleukin (IL)-1β and IL-18 (38). NLRP3 is thought to be a central sensor and mediator in the pathogenesis of metabolic diseases like obesity and nonalcoholic fatty liver disease, and the NLRP3 inflammasome has been implicated in the development of both T2D and T1D involving IL-1β-mediated destruction of pancreatic β-cells (13, 27, 41).
The NLRP3 inflammasome is activated in a two-step manner (38). After an initial priming signal by, for example, endotoxins or other activators of nuclear factor (NF)-κB, extracellular ATP, urate or cholesterol crystals, and metabolic (e.g., glucose or saturated fatty acids) and oxidative stress could function as signal 2 leading to NLRP3 inflammasome assembly and release of IL-1β and IL-18 (38). Augmented oxidative stress, involving increased generation of reactive oxygen species (ROS), plays a central part in the development of diabetes (37). Also, hypoxia is a potent inducer of oxidative stress and is suggested to be an important contributor to β-cell loss (2, 10). Oxidative stress seems also to play an important role in NLRP3 inflammasome activation (42), and hypoxia and oxidative stress could therefore be of particular relevance in relation to pancreatic NLRP3 inflammasome activation in diabetes.
Based on these data, we hypothesized that deficiency of NLRP3 inflammasome could protect pancreatic islet β-cells from oxidative stress-induced damage. This hypothesis was investigated by examining 1) effects of short-term hypoxia ex vivo on islets from NLRP3-deficient and wild-type (WT) mice and 2) the in vivo effects of alloxan, a ROS-inducing diabetogenic glucose analog (18, 33), in inflammasome-deficient and WT mice.
MATERIALS AND METHODS
Animals.
Animal experiments were approved by the Norwegian Animal Research Authority project license no. FOTS id 7333 and 13643. The animal experiments were performed in accordance with the European Directive 2010/63/EU and The Guide for the Care and Use of Laboratory Animals, 8th edition (NRC 2011, National Academic Press). The animals were housed under standard conditions in an approved unit and given free access to food except when fasting. Water was not restricted. The mice were handled by an experienced animal technician at all times, and all efforts were made to minimize suffering. The same technician monitored animal welfare in accordance to standardized requirements for the animal unit at Oslo University Hospital, administered injections, and performed the blood sampling. WT mice (C57BL/6J) acquired from the Jackson Laboratory (Bar Harbor, ME) and NLRP3−/− and ASC−/− mice on a C57BL/6J background (17, 24) were bred in our animal facility according to institutional guidelines.
Murine islet isolation, culture, and incubation.
Mouse islets were isolated from 16- to 20-wk-old male mice as previously described (34). Briefly, we injected 3.0 ml of Hanks’ balanced salt solution containing 0.8 mg/ml collagenase P from Clostridium histolyticum (Roche Diagnostics, Mannheim, Germany) in the pancreatic duct. The distended pancreas was subsequently removed and incubated at 37°C for 17 min followed by gradient purification. The islets were placed in 90-mm petri dishes (Sterilin, Thermo Fisher Scientific, Waltham, MA) in RPMI 1640 media (HyClone, South Logan, UT) supplemented with 10% heat-inactivated FBS, 1% penicillin/streptomycin, 10 mM HEPES, and 1% l-glutamine (Life Technologies, Carlsbad, CA) and cultured free floating overnight at 37°C (5% CO2). The next day, islets (WT and NLRP3−/−) were either further incubated for 3 h in normoxia (21% O2) or hypoxia (1% O2) in a humidified incubator providing atmosphere containing 21% O2 − 5% CO2 (normoxia) or using a closed hypoxia chamber (New Brunswick, Galaxy48 R; Eppendorf, Hauppauge, NY) flushed with a gas mixture of 1% O2 − 94% N2 − 5% CO2 (hypoxia) at 37°C followed by immediate in vitro assessments, as described below.
Glucose-stimulated insulin secretion assay.
Twenty islets were handpicked, transferred to Transwell trays (Costar, Cambridge, MA), and preincubated in Krebs-Ringer-bicarbonate buffer (25 mM HEPES, 24 mM NaHCO3, 5 mM KCl, 1 mM MgCl2·6H2O, 115 mM NaCl, and 0.1% human serum albumin) containing 1.67 mM glucose at 37°C (5% CO2) for 30 min before the islets were incubated for 40 min in fresh Krebs-Ringer-bicarbonate buffer containing 1.67 mM glucose (basal). Finally, the islets were incubated for 40 min in fresh Krebs-Ringer-bicarbonate buffer containing 20 mM glucose (stimulated) according to the guidelines of the National Institutes of Health Clinical Islet Transplantation Consortium that do not require adjustment of osmolality (22a). The supernatants were subsequently collected, and insulin secretion was measured using a mouse insulin enzyme immunoassay (EIA) (Mercodia, Uppsala, Sweden). The capacity for insulin release was expressed as a stimulation index (SI), calculated as the ratio of stimulated to basal insulin secretion.
Viability assays.
Intracellular ATP content was quantified using the ATPLite Luminescence Assay System (PerkinElmer, Waltham, MA). Apoptotic cell death was determined by the detection of DNA histone complexes present in the cytoplasmic fraction of the cells, using the Cell Death Detection ELISAPLUS (Roche Diagnostics). The assays were performed as described by the manufacturers.
Biochemical analyses.
Mouse IL-6 was measured in duplicate by commercially available ELISA (DuoSet Mouse IL-6; R&D Systems, Minneapolis, MN) according to the specifications of the manufacturer.
In vivo animal experimental procedures.
Male mice (16–24 wk old), WT, NLRP3−/−, and ASC−/−, were injected daily with a low dose of alloxan monohydrate (40 mg/kg iv; Sigma Aldrich) for three to five consecutive days to induce a diabetic phenotype. To deplete macrophages, WT mice were administered an intravenous injection (0.1 ml/10 g body wt) of clodronate-encapsulated liposomes (Liposomas, Amsterdam, The Netherlands) at days −5, −3, and −1. The dose of clodronate-liposomes was selected based on the recommendations of the manufacturer and previously published literature (35). Phosphate-buffered saline (PBS)-encapsulated liposomes (PBS-liposomes) were administered in the same manner to the control group. Blood was collected from the saphenous vein during the studies or by heart puncture in EDTA-coated tubes (Becton-Dickinson, Puls, Oslo, Norway) upon sacrifice. Blood glucose levels were measured using a glucometer (Accu-Chek Aviva Nano; Roche Diagnostics). Body weight and nonfasting blood glucose were measured on days 0, 1, 2, 3, 4, 7, 11, 14, and 18. Fasting blood glucose was measured 11 and 18 days after the first alloxan injection. An oral glucose tolerance test (OGTT) was performed on day 11. After 2–18 days of follow-up, the animals were euthanized, and the pancreata were harvested for histology and gene expression analyses, as described below.
Oral glucose tolerance test.
OGTT was performed after a 4-h fast and administration of oral (gavage) d-glucose (1.5 g/kg body wt; Fresenius Kabi, Oslo, Norway). Blood glucose was measured at the indicated time points. Insulin response was measured in whole blood during the OGTT at the indicated time points using the Ultra Sensitive Mouse Insulin ELISA kit (Crystal Chem, Zaandam, Netherlands) with a modified protocol using whole blood and adjusting the sample volume to 3 µl.
Quantification of gene expression.
For purification of total RNA from formalin-fixed, paraffin-embedded tissue sections, an RNeasy FFPE Kit was used (Qiagen, Hilden, Germany). cDNA was made using the qScript cDNA SuperMix (Quanta Biosciences, Beverly, MA). Quantification of gene expression was performed using the Mx3005P (Agilent Technologies, Cedar Creek, TX), Brilliant III Ultra-Fast SYBR Green Master Mix (Agilent Technologies), and sequence-specific PCR primers for NLRP3 designed using the Primer Express software, version 3.0 (Applied Biosystems, Foster City, CA). Gene expression of β-actin was used as reference for relative quantifications.
Immunohistochemistry and immunofluorescence.
Sections (5 μm) of paraffin-embedded mouse pancreata were deparaffinized in xylene, rehydrated in alcohol series, and immersed in distilled water, followed by high-temperature antigen retrieval in citrate buffer (pH 6) and blocked with 0.5% bovine serum albumin (Sigma-Aldrich, St. Louis, MO). For immunohistochemistry, the complete pancreas was harvested, paraffin embedded, and systematically sliced across from top to bottom in serial sections to identify β-cell populations within the islets. Every tenth section throughout the pancreata was stained and examined for insulin-positive islets. These slides were also stained with rat anti-mouse antibody against Mac-2 (1:750, CL8942AP; Cedarlane, Burlington, Canada) for 1 h at room temperature. After being washed, the slides were incubated for 30 min with peroxidase-conjugated secondary antibody (Impress-Vector; Vector Laboratories, Burlingame, CA), rinsed, and developed with chromogen for immunoperoxidase staining (DAB Plus; Vector Laboratories). The sections were counterstained with hematoxylin. Images were taken using a Nikon Eclipse E400 microscope. The number of Mac-2-positive cells was scored manually and expressed relative to the islet area in four separate pancreata for each genotype of mice (NIS-Elements BR; Nikon Software). Omission of the primary antibody was used as negative control. For immunofluorescence, the slides were stained with guinea pig anti-mouse antibody against insulin (1:400, A0564; Dako, Glostrup, Denmark), rat anti-mouse antibody against Mac-2 (1:50; Cedarlane), mouse anti-mouse antibody against NLRP3 (Cryo-2, 1:50, AG-20B-0014; Adipogen, San Diego, CA), and rabbit anti-mouse antibody against superoxide dismutase (SOD) 2 (1:100, ab13533; Abcam, Cambridge, UK) (31) for 1 h at room temperature (RT) or overnight incubation at 4°C and counterstained with Alexa Fluor 568 goat anti-guinea pig, Alexa Fluor 488 goat anti-rat, Alexa Fluor 633 goat anti-mouse, and Alexa Fluor 488 goat anti-rabbit (1:500; Life Technologies), respectively. Images were captured using a Zeiss LSM710/Elyra S1 confocal microscope (Jena, Germany) housing a Plan-Apochromat W40x/1.0 DIC M27 objective. Zeiss ZEN Lite software or Adobe Photoshop was used for processing images. For quantification of SOD2 in islets, the stained sections were scanned (AxioScan; Carl Zeiss, Oberkochen, Germany), and the level of SOD2 was estimated by using the Z9 platform (an in-house made software program for examining and quantitatively evaluating histological sections) calculated as mean intensity per pixel in islet.
Flow cytometry analysis.
Fresh whole blood (100 µl) was incubated with Mouse BD Fc Block (BD Biosciences, San Jose, CA) for 5 min at 4°C, after which 2.5 µl CD11b-FITC (0.2 mg/ml), Ly6C-APC (0.2 mg/ml), and Ly6G-PE (0.2 mg/ml) (BD Biosciences) were added. After a 30-min incubation in the dark at room temperature, FACS Lysing Solution (BD Biosciences) was added, and the samples were incubated for another 10 min at room temperature and subsequently centrifuged at 500 g for 5 min. The supernatant was discarded, and cell pellets were washed two times in buffer [PBS containing 0.5% bovine serum albumin (BSA; Sigma Aldrich)] and centrifuged (500 g/5 min). After the final centrifugation, supernatant was discarded, and flow cytometry analysis was performed using FACSVerse (BD Biosciences).
Statistical analysis.
All data are presented as means ± SE, and the GraphPad Prism 7.02 statistical software (La Jolla, CA) was used for analysis. In ex vivo islet experiments, results were analyzed with two-way ANOVA with the factors of genotype and hypoxia and post hoc tested with uncorrected Fisher’s least-significant difference test. In in vivo experiments, difference between groups was evaluated by two-tailed t-test (2 groups) or one-way ANOVA with Tukey’s multiple-comparisons test (3 groups). In time-dependent experiments, repeated-measures two-way ANOVA was used. A P value <0.05 was considered statistically significant.
RESULTS
NLRP3 deficiency preserves ex vivo islet function and inhibits hypoxia-induced islet cell death.
Oxidative stress, involving increased ROS generation, is a central feature in diabetes pathogenesis. Hypoxia-induced oxidative stress could lead to activation of NLRP3 inflammasomes (30). We therefore isolated pancreatic islets from NLRP3−/− and WT mice and exposed these to normoxia or hypoxia ex vivo. We first evaluated the effect on β-cell function by measuring basal and glucose-stimulated insulin secretion (Fig. 1A). Glucose stimulation increased insulin secretion both during normoxia and hypoxia, but with no significant differences between WT and NLRP3−/− islets. However, when calculating the SI, i.e., the ratio of glucose-stimulated to basal insulin secretion, we found that NLRP3-deficient islets had higher SI than WT islets independent of normoxia or hypoxia (2-way ANOVA, genotype effect: P = 0.02; Fig. 1B). Although we were not able to determine whether the changes in SI in the 2-way ANOVA analyses were caused by an increase in glucose-stimulated insulin secretion or a decrease in the basal secretion at low glucose, our results suggest that NLRP3 deficiency preserves ex vivo islet function.
Fig. 1.Nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) inflammasome deficiency preserves ex vivo islet function. Islets isolated from wild-type (WT; n = 5) and NLRP3−/− (n = 5) mice were exposed to normoxia (21% O2) or hypoxia (1% O2) for 3 h. Immediately thereafter, basal (1.67 mM glucose) and glucose-stimulated (20 mM glucose) insulin secretion was measured (A). Glucose-stimulation index was calculated as the ratio of glucose-stimulated to basal insulin secretion (P = 0.02 for effect of NLRP3 deficiency; 2-way ANOVA with the factors of genotype and hypoxia) (B). Results are expressed as means ± SE, and n indicates the no. of independent isolations of islet. *P < 0.05 vs. basal glucose (paired t-test).
Hypoxia increased cell death in WT islets (P = 0.01 vs. normoxia), which was significantly reduced in NLRP3−/− islets (P = 0.03 vs. WT islets with hypoxia; Fig. 2A). Moreover, two-way ANOVA indicated a significant hypoxia-genotype interaction (P = 0.03). Hypoxia also significantly decreased intracellular levels of ATP in both WT and NLRP3-deficient pancreatic islets (2-way ANOVA, hypoxia effect: P < 0.001; Fig. 2B). Thus, the hypoxic environment affected cellular metabolism to a similar extent in islets from both genotypes, but NLRP3-deficient islets were protected against hypoxia-induced cell death and also had preserved β-cell function during both normoxia and hypoxia.
Fig. 2.Nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) inflammasome deficiency protects against hypoxia-induced cell death. Islets isolated from wild-type (WT; n = 5) and NLRP3−/− (n = 5) mice were exposed to normoxia (21% O2) or hypoxia (1% O2) for 3 h. Immediately thereafter, apoptotic cell death (P = 0.01 for hypoxia-genotype interaction, A) and intracellular ATP content (P < 0.001 for hypoxia effect, B) were measured. IL-6 levels were determined by ELISA in islet culture supernatants (P = 0.03 for hypoxia × genotype interaction effect, C). Positive control (A) denotes a DNA-histone complex included in the Cell Death Detection ELISA kit. Results are expressed as means ± SE, and n indicates the no. of independent isolations of islet. P < 0.05 vs. WT (*) and normoxia (†) by uncorrected Fisher’s least-significant difference (LSD) test following 2-way ANOVA with the factors of genotype and hypoxia.
We were not able to determine islet secretion of IL-1β or IL-18 (and indeed, these cytokines can be difficult to detect even in cell supernatants). However, IL-6 has been implicated as a central downstream effector of IL-1β (4), and, when measuring IL-6 secretion in the islet media, we found a significant hypoxia-induced increase in IL-6 in WT islets (P = 0.04), with a lower level in NLRP3−/− islets (P = 0.02 vs. WT hypoxia; 2-way ANOVA interaction hypoxia-genotype: P = 0.03; Fig. 2C). We hypothesize that the reduced IL-6 levels with hypoxia in NLRP3−/− islets reflect impaired IL-1β bioactivity with attenuating damage in pancreatic islet cells as a consequence.
NLRP3 is associated with alloxan-induced macrophage infiltration in islets.
Our findings also indicate a detrimental role for the NLRP3 inflammasome on islet function and viability during hypoxic stress ex vivo. Alloxan is a toxic glucose analog that preferentially accumulates in pancreatic β-cells via the GLUT2 glucose transporter. The cytotoxic action of this diabetogenic agent is mediated by ROS (18), and a high dose of alloxan is commonly used for induction of T1D (1, 33). We injected WT mice with low alloxan doses (40 mg/kg) for four consecutive days that could mimic the situation in a prediabetic state (1, 15). Hyperglycemia was defined as blood glucose level >11.1 mmol/l (14). We found that nonfasting blood glucose was significantly increased in mice injected with multiple low doses of alloxan (Fig. 3A), associated with a marked increase in islet-associated Mac-2-positive macrophages (Fig. 3, B and C). Importantly, NLRP3 immunostaining of pancreatic sections of alloxan-treated mice revealed strong colocalization of NLRP3 with the macrophage marker Mac-2 (Fig. 3C, inset).
Fig. 3.Alloxan treatment induces islet-associated macrophage accumulation. Wild-type (WT) mice were injected with a daily iv dose of alloxan (40 mg/kg; n = 5) or PBS (control; n = 5) for four consecutive days. Blood glucose was measured, and pancreata were harvested for histological analysis. A: fasting blood glucose. B: quantification of Mac-2-positive macrophages in pancreatic islets. C: immunofluorescence images showing positive staining of Mac-2 (green) and nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3, red) within an insulin-producing islet (blue). Inset, NLRP3 and Mac-2 double-positive cells (yellow), shown by arrows. Results are expressed as means ± SE. **P < 0.01 vs. control determined by 2-tailed t-test. Scale bars: 50 µm.
Macrophages mediate alloxan-induced diabetes.
To examine if the macrophage accumulation was directly involved in diabetes development in alloxan-exposed mice, we depleted macrophages using clodronate-loaded liposomes before induction of a diabetic phenotype by multiple low-dose alloxan treatment. WT mice were injected intravenously with PBS-liposomes or clodronate-liposomes as indicated in Fig. 4A. Flow cytometry of peripheral blood verified that the cluster CD11b+ Ly6C++ Ly6G−, which is considered as the main monocyte immunoregulatory population, is significantly depleted after clodronate-liposome treatment (Fig. 5), associated with a statistically significant decrease in islet-associated macrophages in clodronate- and liposome-treated mice compared with PBS-liposomes as assessed by immunohistochemistry evaluation (Fig. 4, B and C). Pancreatic NLRP3 mRNA levels increased with alloxan treatment, whereas macrophage depletion with clodronate-liposomes decreased NLRP3 gene expression (Fig. 4D). Furthermore, the reduced pancreatic macrophage accumulation during clodronate-liposome treatment was associated with a significant reduction in blood glucose (Fig. 4E) and an increase in blood insulin levels (Fig. 4F), indicating that macrophages are centrally involved in the diabetogenic effect of alloxan, potentially involving NLRP3-mediated mechanisms.
Fig. 4.Depletion of islet-associated macrophages reduces diabetogenic effect of alloxan. Wild-type (WT) mice were injected iv with PBS-liposomes (PBS-Lipo; n = 8) or clodronate-liposomes (Clod-Lipo; 0.1 ml/10 g body wt; n = 10) at days −5, −3, and −1 followed by low doses of alloxan (40 mg/kg) at three consecutive days (A). Islet-associated Mac-2-positive macrophages in pancreas sections were identified using immunohistochemistry (scale bar: 150 µm, B) and quantified relative to islet cell area (C) (PBS-Liposome; n = 4 and Clod-Liposome; n = 4). Pancreatic gene expression of nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3, D) was determined from formalin-fixed paraffin-embedded tissue sections using real-time quantitative RT-PCR in untreated controls (n = 4) and in mice receiving alloxan with pretreatment with either PBS-Lipo (n = 8) or Clod-Lipo (n = 10) and is presented relative to levels of β-actin mRNA. Plasma glucose (E) and plasma insulin (F) were measured in fasting mice (PBS-Liposome; n = 8 and Clod-Liposome; n = 10). Results are expressed as means ± SE. *P < 0.05 and **P < 0.01 vs. PBS-liposome and †††P < 0.001 vs. control by t-test or 1-way ANOVA with Tukey’s multiple-comparisons test (D).
Fig. 5.Injections of clodronate-liposomes cause depletion of inflammatory monocyte and neutrophils in wild-type (WT) mice whole blood. WT mice were injected iv with PBS-liposomes (n = 4) or clodronate-liposomes (0.1 ml/10 g body wt) (n = 4) at days −5, −3, −1, and 0 followed by low doses of alloxan (40 mg/kg) at days 0 and +1. One day after the last injection of liposomes, whole blood was drawn from the heart, and a total of 20,000 single cell events were analyzed on a FACS Verse instrument. The CD11b-positive cluster (in % of total events) is then gated and used for further gating. A: representative FACS images are shown. Populations of inflammatory monocytes (B) and a neutrophil population (C) were quantified. Results are expressed as means ± SE. *P < 0.05 vs. PBS-liposome injections.
NLRP3 inflammasome deficiency protects against alloxan-induced diabetes in vivo.
To further elucidate the role of macrophages and NLRP3 in the diabetogenic effect of alloxan, WT, NLRP3−/−, and ASC−/− mice were injected with alloxan (40 mg/kg) for five consecutive days. Nonfasting blood glucose increased markedly in WT mice (Fig. 6A). This time-dependent increase in blood glucose was significantly attenuated in both NLRP3- and ASC-deficient mice [P < 0.001 vs. WT (repeated-measures 2-way ANOVA); Fig. 6A]. Compared with WT, the blood glucose AUC was also significantly lower in NLRP3−/− and ASC−/− mice (Fig. 6B). In contrast, there were no statistical differences in fasting blood glucose or insulin secretion between WT, NLRP3−/−, and ASC−/− mice in the absence of alloxan measured by two-way ANOVA (Fig. 7, A and B).
Fig. 6.Nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) inflammasome deficiency protects β-cell function from oxidative stress-induced diabetes after low-dose alloxan injections. Wild-type mice (WT, black circles; n = 11), NLRP3−/− mice (green triangles; n = 13), and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC−/−, red squares; n = 13) were injected with a daily low dose of alloxan (40 mg/kg) for five consecutive days, and nonfasting glucose was monitored for 18 days. A and B: nonfasting blood glucose levels at indicated time points (A) with corresponding area under the curve (AUC, B). Results are expressed as means ± SE. **P < 0.01 and ***P < 0.001 vs. WT by repeated-measures 2-way ANOVA (A) or 1-way ANOVA with Tukey’s multiple-comparison test (B).
Fig. 7.Healthy untreated WT mice, nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3-deficient mice (NLRP3−/−), and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC−/−) have similar fasting blood glucose and blood insulin. Blood glucose (A) and blood insulin (B) were measured in fasting wild-type (WT, n = 11), NLRP3−/− (n = 11), and ASC−/− mice (n = 11). Results are expressed as means ± SE.
To further study the effect of NLRP3 inflammasomes on β-cell function, we performed an OGTT in WT, NLRP3−/−, and ASC−/− mice starting at day 11 after the first alloxan injection. Fasting glucose levels were significantly lower in ASC−/− mice than in WT mice (Fig. 8A), whereas there was no difference between WT and NLRP3−/−. Similarly, OGTT was significantly improved in ASC−/− mice compared with WT mice (Fig. 8, B and C). There were no significant differences in fasting blood insulin levels but a trend toward higher levels in NLRP3−/− and ASC−/− (Fig. 8D), and, notably, NLRP3−/− mice had a significantly improved glucose-stimulated insulin secretion and insulinogenic index (Fig. 8, E and F) compared with WT mice, indicating preserved β-cell function. Although there were some differences in the effects of NLRP3 and ASC deficiency on these parameters, both genotypes seemed to have protective effects on islet function during alloxan exposure, and there were in fact no significant differences in nonfasting or fasting blood glucose levels or in the OGTT between alloxan-treated NLRP3−/− and ASC−/− mice. However, because ASC acts as a scaffold protein in several inflammasomes (20), differences between ASC and NLRP3 may still exist, but this study was not powered to reveal such a difference.
Fig. 8.Nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) inflammasome deficiency improves β-cell function. Wild-type mice (WT, black circles; n = 11), NLRP3−/− mice (green triangles; n = 13), and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC−/−, red squares; n = 13) were injected with a daily low dose of alloxan (40 mg/kg) for five consecutive days. Fasting plasma glucose (A) and fasting plasma insulin levels (D) are shown at day 11. An oral glucose tolerance test was performed at day 11 as described in materials and methods. B and C: blood glucose was measured (B) with the corresponding area under the curve (AUC, C). E and F: insulin response was measured (E), and the insulinogenic index was determined as AUCinsulin/AUCglucose (F). Results are expressed as means ± SE and are representative for n = 11–13. *P < 0.05 and **P < 0.01 vs. WT by 1-way ANOVA and Tukey’s multiple-comparisons test (A, C, D, and F) or repeated-measures 2-way ANOVA (B and E).
NLRP3 inflammasome deficiency reduces alloxan-induced islet oxidative stress.
Alloxan is used as an inducer of oxidative stress in β-cells. SOD2, an enzyme induced by O2·− and used as a marker of oxidative stress (31, 40), is linked to insulin resistance and islet dysfunction (10). We found a significant increase in SOD2 expression as assessed by fluorescent antibody staining in islets in WT mice treated with alloxan compared with untreated animals, and, notably, SOD staining was significantly lower in both NLRP3- and ASC-deficient mice treated with alloxan (Fig. 9).
Fig. 9.Alloxan treatment induces oxidative stress in islets. Wild-type (WT) mice were injected with PBS (Control; n = 9) and WT (n = 9) mice, nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3-deficient mice (NLRP3−/−, n = 9), and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC−/−, n = 4) were injected with a daily iv dose of alloxan (40 mg/kg) for four consecutive days. Pancreata were harvested for histological analysis at day 7 after the first alloxan injection. A: immunofluorescence images show positive staining of superoxide dismutase (SOD) 2 (green) within an insulin-producing islet (red). Nuclei are stained blue with DAPI. Scale bar: 50 µm. B: histograms showing quantification of positive SOD2 given as %saturation of SOD2 in islets. Results are expressed as means ± SE. P < 0.05 vs. control (*) and WT (†) determined by 1-way ANOVA and Tukey’s multiple-comparisons test.
NLRP3 inflammasome deficiency reduces β-cell loss and islet inflammation in alloxan-induced dibetes.
Pancreata from the three genotypes were then subjected to histological analyses, as described in materials and methods. Healthy untreated WT, NLRP3−/−, and ASC−/− mice had similar islet areas and numbers of resident macrophages (Fig. 10). However, in alloxan-treated mice, we found that both the NLRP3−/− and the ASC−/− genotype had significantly more insulin-positive staining islets (Fig. 11, A and B) and significantly reduced islet-associated macrophage accumulation compared with WT mice (Fig. 11, A and C). These findings suggest that a beneficial effect of NLRP3 inflammasome deficiency on oxidative stress-mediated β-cell damage could involve reduced macrophage infiltration.
Fig. 10.Healthy untreated wild-type (WT) mice, nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3-deficient mice (NLRP3−/−), and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC−/−) have a similar amount of islet-associated macrophages and islet area. Pancreata were harvested from 16-wk-old WT (n = 3), NLRP3−/− (n = 3), and ASC−/− (n = 3) mice and stained with Mac-2 antibody. Mac-2-positive macrophages in islets (A) and islet area/pancreata (B) were quantified. Results are expressed as means ± SE.
Fig. 11.Nucleotide-binding oligomerization domain, leucine-rich-containing family, pyrin domain-containing-3 (NLRP3) inflammasome deficiency reduces β-cell loss and islet inflammation in oxidative stress-induced diabetes. Wild-type mice (WT, n = 11), NLRP3−/− mice (n = 13), and mice deficient in apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC−/−, n = 13) were injected with a daily low dose of alloxan (40 mg/kg) for five consecutive days. After the first injection (18 days), pancreata were harvested for histological analysis. A: representative immunohistofluorescence images showing insulin staining in islets (scale bar 150 µm, top) and immunohistochemistry for Mac-2-positive macrophages (scale bar 50 µm, bottom). The no. of β-cell-positive areas/pancreas (B) and the no. of Mac-2-positive macrophages/islet area (C) were quantified. Results are expressed as means ± SE. ****P < 0.0001 vs. WT by 1-way ANOVA and Tukey’s multiple-comparisons test.
DISCUSSION
In this study we investigate the role of the NLRP3 inflammasome and macrophages in pancreatic islet function during oxidative stress-induced diabetes. We find that 1) deficiency of NLRP3 inflammasome reduces the attenuation of ex vivo islet cell function and viability during hypoxia and 2) deficiency of inflammasome components results in preserved β-cell function and attenuated β-cell loss after oxidative stress-induced islet destruction in vivo, potentially related to decreased islet accumulation of macrophages in NLRP3−/− and ASC−/− mice. Our findings suggest an important role for the NLRP3 inflammasome during oxidative stress-induced islet cell exposure, potentially representing a target for therapy in prediabetic patients.
The innate immune system, including NLRP3 inflammasome, is central in the pathogenesis of both T1D and T2D (25, 27). In models of T2D, the NLRP3 inflammasome has been linked to metabolic and oxidative stress-induced β-cell destruction (16, 22, 36, 41, 42). In a mouse model of T1D, NLRP3 deficiency was associated with reduced diabetes incidence because of dysregulated T cell chemotaxis (13). Our findings herein suggest that NLRP3 inflammasome is also involved in monocyte/macrophage chemotaxis in oxidative stress-exposed islet cells. In the in vivo model of oxidative stress-induced islet cell destruction, NLRP3- and ASC-deficient mice had less macrophage infiltration compared with WT. Jourdan et al. showed that the NLRP3 inflammasome is activated upon stress in islet infiltrating macrophages (16), and this has also been demonstrated by others (22, 26). Macrophages are involved in β-cell destruction during oxidative stress, and it is therefore not inconceivable that the beneficial effects of NLRP3 and ASC deletion on islet cell function in our study are related to impaired macrophage infiltration. The reduced number of infiltrating macrophages, that in addition are less activated in NLRP3−/− and ASC−/− mice than in WT mice, will again result in decreased chemotactic signals for additional monocyte/macrophage infiltration, representing an anti-inflammatory loop that could protect β-cells from destruction during oxidative stress. Importantly, we show that NLRP3 immunostaining of pancreatic sections of alloxan-treated mice revealed strong colocalization of NLRP3 with the macrophage marker Mac-2, but not with the insulin-producing cells. This is in accordance with Jourdan et al. (16) who showed that little colocalization was observed between IL-1β and the β-cell marker insulin, suggesting that islet resident and/or infiltrating macrophages are the major source of IL-1β in islets in vivo. However, in untreated control mice, deficiency in NLRP3 or ASC did neither affect pancreatic β-cell mass, the number of resident macrophages, nor fasting glucose and insulin levels, indicating that the major effects of NLRP3 and ASC deficiency are the induction and activation of infiltrating macrophages. However, our data on NLRP3 colocalization to macrophages are based on pancreatic sections, and the lack of data that demonstrate NLRP3 mRNA specifically in pancreatic islets is a limitation in our study.
Alloxan is widely used to induce oxidative stress-related diabetes (18). We used multiple low-dose (40 mg/kg) alloxan injections to mimic oxidative stress-induced onset of diabetes. From other studies, a higher dose of alloxan (120–180 mg/kg) is required to destroy all β-cells and induce severe diabetes in mice (23). SOD2 as a marker of oxidative stress (40) was increased in WT mice treated with multiple low doses of alloxan compared with untreated animals, and, importantly, islet SOD2 expression was reduced in NLRP3 inflammasome-deficient mice. It is from our studies not clear at what level oxidative stress interacts with NLRP3 inflammasome activation and β-cell death, but, based on the SOD2 expression data, it is tempting to hypothesize that it involves the combined effect of 1) attracting and activating monocytes/macrophages in the pancreas and 2) direct oxidative stress-mediated cytotoxicity on β-cells.
In the present study, we show that deletion of NLRP3 preserves ex vivo islet function and also protects islets from hypoxia-induced toxicity ex vivo. We were not able to determine at what level NLRP3 deficiency improved the islet function, i.e., reduced basal vs. increased glucose-induced insulin secretion. Moreover, the mechanisms for the improved islet function with NLRP3 deficiency are unclear, and we can only speculate that they involve protection from the stress and the hypoxic condition of the islets during isolation (2). Our findings contrast those from Wali et al., which did not find any role of NLRP3 in stress-induced pancreatic islet death (36). However, while we stressed islets using hypoxia, Wali et al. use different chemical stressors as well as glucose and palmitate (36). Thus, the nature of the stressor may be important when determining the role of NLRP3 in islet function ex vivo. Moreover, whereas metabolic stress is clearly relevant to the development of diabetes, oxidative stress is also an important feature of prediabetic states (7), making our model that was used in the present study relevant to clinical diabetes. Indeed, islets have a low antioxidant capacity and are therefore made particularly vulnerable to oxidative stress (19).
The anti-inflammatory effects of NLRP3 deletion in macrophages are linked to impaired IL-1β and IL-18 secretion (27). In particular, islets are explicitly sensitive to the toxicity of IL-1β because even extremely low concentrations of the cytokine will lead to β-cell dysfunction (21). If these effects are directly caused by IL-1β or if it reflects that IL-1β is a potent upstream mediator in the inflammatory cytokine cascade or a combination thereof is at present not clear. A limitation of our study is that we were not able to detect secretion of IL-1β or IL-18 in the ex vivo islet experiments. However, several effects of IL-1 seem to be mediated by IL-6 (4), and we show that IL-6 secretion increased significantly with hypoxia in WT islets, but not in NLRP3−/− islets, and we hypothesize that this could reflect impaired IL-1β bioactivity.
The deleterious effect of chronic activation of the IL-1β system on T2D and other metabolic diseases is well documented (8, 32, 39). However, recent reports also highlight roles of IL-1β in augmenting insulin secretion, i.e., in the postprandial stimulation of insulin secretion and in islet compensation to acute metabolic stress (5, 12). Although these latter studies point to important physiological properties of IL-1β, results from clinical studies that target IL-1β with anakinra or canakinumab do indicate beneficial effects on glucose metabolism in patients with prediabetes or T2D (11). Thus, at least in a setting of chronic metabolic disease, IL-1β might have predominantly detrimental effects on β-cell function. Importantly, however, although the recent CANTOS trial showed that treatment with canakinumab reduced the number of cardiovascular events in patients with previous myocardial infarction (28), canakinumab did not reduce the incidence of new-onset diabetes (9). These results represent an argument against a prominent role of IL-1β, at least in the development of T2D. However, canakinumab led to small decreases in HbA1C levels during the first 6–9 mo, and the study was not designed to detect differences in development of diabetes allowing lifestyle intervention and alteration in other antidiabetic drugs.
In conclusion, our results suggest that deletion of the NLRP3 inflammasome improves β-cell function and viability during hypoxia and oxidative stress, potentially related to anti-inflammatory effects, including attenuated macrophage infiltration in the islets. Our findings indicate a role for activation of the NLRP3 inflammasome during oxidative stress-induced β-cell dysfunction such as during the development and progression of T2D, potentially also representing a target for therapy in this disorder and prediabetic states in more general term.
GRANTS
This work was supported by grants from Helse Sør-Øst Regional Health Authority, Norway (Grant No. 2012037 to A. Yndestad) and the Norwegian Research Council (Grant No. 240099/F20 to P. Aukrust).
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.
AUTHOR CONTRIBUTIONS
M.S., A.S., E.L., P.A., A.Y., T.R., and H.S. conceived and designed research; M.S., A.S., M.H., J.x., and T.R. performed experiments; M.S., A.S., M.H., J.x., A.Y., T.R., and H.S. analyzed data; M.S., A.S., P.A., A.Y., T.R., and H.S. interpreted results of experiments; M.S., A.S., A.Y., and T.R. prepared figures; M.S., A.S., and T.R. drafted manuscript; M.S., P.A., A.Y., T.R., and H.S. edited and revised manuscript; M.S., A.S., M.H., J.x., E.L., P.A., A.Y., T.R., and H.S. approved final version of manuscript.
ACKNOWLEDGMENTS
We acknowledge Ellen Lund Sagen for technical assistance.
REFERENCES
- 1. . Long-term effects of alloxan in mice. Int J Pancreatol 17: 197–201, 1995. doi:10.1007/bf02788539.
Crossref | PubMed | Google Scholar - 2. . A preexistent hypoxic gene signature predicts impaired islet graft function and glucose homeostasis. Cell Transplant 22: 2147–2159, 2013. doi:10.3727/096368912X658728.
Crossref | PubMed | ISI | Google Scholar - 3. . Macrophage-mediated inflammation in metabolic disease. Nat Rev Immunol 11: 738–749, 2011. doi:10.1038/nri3071.
Crossref | PubMed | ISI | Google Scholar - 4. . Interleukin-1 in the pathogenesis and treatment of inflammatory diseases. Blood 117: 3720–3732, 2011. doi:10.1182/blood-2010-07-273417.
Crossref | PubMed | ISI | Google Scholar - 5. . Postprandial macrophage-derived IL-1β stimulates insulin, and both synergistically promote glucose disposal and inflammation. Nat Immunol 18: 283–292, 2017. doi:10.1038/ni.3659.
Crossref | PubMed | ISI | Google Scholar - 7. . Metabolic stress in insulin’s target cells leads to ROS accumulation - a hypothetical common pathway causing insulin resistance. FEBS Lett 581: 3734–3742, 2007. doi:10.1016/j.febslet.2007.06.044.
Crossref | PubMed | ISI | Google Scholar - 8. . Inflammation as a link between obesity, metabolic syndrome and type 2 diabetes. Diabetes Res Clin Pract 105: 141–150, 2014. doi:10.1016/j.diabres.2014.04.006.
Crossref | PubMed | ISI | Google Scholar - 9. . Anti-inflammatory therapy with canakinumab for the prevention and management of diabetes. J Am Coll Cardiol 71: 2392–2401, 2018. doi:10.1016/j.jacc.2018.03.002.
Crossref | PubMed | ISI | Google Scholar - 10. . The role of oxidative stress and hypoxia in pancreatic beta-cell dysfunction in diabetes mellitus. Antioxid Redox Signal 26: 501–518, 2017. doi:10.1089/ars.2016.6755.
Crossref | PubMed | ISI | Google Scholar - 11. . Therapeutic approaches targeting inflammation for diabetes and associated cardiovascular risk. J Clin Invest 127: 83–93, 2017. doi:10.1172/JCI88884.
Crossref | PubMed | ISI | Google Scholar - 12. . Interleukin-1 signaling contributes to acute islet compensation. JCI Insight 1: e86055, 2016. doi:10.1172/jci.insight.86055.
Crossref | PubMed | ISI | Google Scholar - 13. . NLRP3 deficiency protects from type 1 diabetes through the regulation of chemotaxis into the pancreatic islets. Proc Natl Acad Sci USA 112: 11318–11323, 2015. doi:10.1073/pnas.1513509112.
Crossref | PubMed | ISI | Google Scholar - 14. . Molecular target structures in alloxan-induced diabetes in mice. Life Sci 71: 1681–1694, 2002. doi:10.1016/S0024-3205(02)01918-5.
Crossref | PubMed | ISI | Google Scholar - 15. . Experimentally induced rodent models of type 2 diabetes. Methods Mol Biol 933: 161–174, 2012. doi:10.1007/978-1-62703-068-7_10.
Crossref | PubMed | Google Scholar - 16. . Activation of the Nlrp3 inflammasome in infiltrating macrophages by endocannabinoids mediates beta cell loss in type 2 diabetes. Nat Med 19: 1132–1140, 2013. doi:10.1038/nm.3265.
Crossref | PubMed | ISI | Google Scholar - 17. . Bacterial RNA and small antiviral compounds activate caspase-1 through cryopyrin/Nalp3. Nature 440: 233–236, 2006. doi:10.1038/nature04517.
Crossref | PubMed | ISI | Google Scholar - 18. . The mechanisms of alloxan- and streptozotocin-induced diabetes. Diabetologia 51: 216–226, 2008. doi:10.1007/s00125-007-0886-7.
Crossref | PubMed | ISI | Google Scholar - 19. . Low antioxidant enzyme gene expression in pancreatic islets compared with various other mouse tissues. Free Radic Biol Med 20: 463–466, 1996. doi:10.1016/0891-5849(96)02051-5.
Crossref | PubMed | ISI | Google Scholar - 20. . Structural mechanisms of inflammasome assembly. FEBS J 282: 435–444, 2015. doi:10.1111/febs.13133.
Crossref | PubMed | Google Scholar - 21. . Blockade of interleukin 1 in type 1 diabetes mellitus. Nat Rev Endocrinol 6: 158–166, 2010. doi:10.1038/nrendo.2009.271.
Crossref | PubMed | ISI | Google Scholar - 22. . Activation of the NLRP3 inflammasome by islet amyloid polypeptide provides a mechanism for enhanced IL-1β in type 2 diabetes. Nat Immunol 11: 897–904, 2010. doi:10.1038/ni.1935.
Crossref | PubMed | ISI | Google Scholar - 22a. . Functional assessment of purified human pancreatic islets: glucose stimulated insulin release by ELISA: a standard operating procedure of the NIH Clinical Islet Transplantation Consortium. CellR4 2: e900, 2014.
Google Scholar - 23. . Induction of alloxan diabetes in obese-hyperglycemic mice (genotype obob). Experientia 29: 1142–1143, 1973. doi:10.1007/BF01946770.
Crossref | PubMed | Google Scholar - 24. . Distinct roles of TLR2 and the adaptor ASC in IL-1beta/IL-18 secretion in response to Listeria monocytogenes. J Immunol 176: 4337–4342, 2006. doi:10.4049/jimmunol.176.7.4337.
Crossref | PubMed | ISI | Google Scholar - 25. . Inflammasome priming in sterile inflammatory disease. Trends Mol Med 23: 165–180, 2017. doi:10.1016/j.molmed.2016.12.007.
Crossref | PubMed | ISI | Google Scholar - 26. . Galectin-3 deficiency accelerates high-fat diet-induced obesity and amplifies inflammation in adipose tissue and pancreatic islets. Diabetes 62: 1932–1944, 2013. doi:10.2337/db12-0222.
Crossref | PubMed | ISI | Google Scholar - 27. . Fatty acids and NLRP3 inflammasome-mediated inflammation in metabolic tissues. Annu Rev Nutr 37: 77–102, 2017. doi:10.1146/annurev-nutr-071816-064836.
Crossref | PubMed | ISI | Google Scholar - 28. . Inhibition of interleukin-1β by canakinumab and cardiovascular outcomes in patients with chronic kidney disease. J Am Coll Cardiol 71: 2405–2414, 2018. doi:10.1016/j.jacc.2018.03.490.
Crossref | PubMed | ISI | Google Scholar - 29. . Immune modulation in humans: implications for type 1 diabetes mellitus. Nat Rev Endocrinol 10: 229–242, 2014. doi:10.1038/nrendo.2014.2.
Crossref | PubMed | ISI | Google Scholar - 30. . Mitochondria as a centrally positioned hub in the innate immune response. Biochim Biophys Acta 1863: 1090–1097, 2017. doi:10.1016/j.bbadis.2016.10.020.
Crossref | PubMed | ISI | Google Scholar - 31. . Disruption of SM22 promotes inflammation after artery injury via nuclear factor kappaB activation. Circ Res 106: 1351–1362, 2010. doi:10.1161/CIRCRESAHA.109.213900.
Crossref | PubMed | ISI | Google Scholar - 32. . The inflammasome puts obesity in the danger zone. Cell Metab 15: 10–18, 2012. doi:10.1016/j.cmet.2011.10.011.
Crossref | PubMed | ISI | Google Scholar - 33. . The mechanism of alloxan and streptozotocin action in B cells of the rat pancreas. Physiol Res 50: 537–546, 2001.
PubMed | ISI | Google Scholar - 34. . Murine pancreatic islet isolation. J Vis Exp 255: 255, 2007. doi:10.3791/255.
Crossref | PubMed | Google Scholar - 35. . The liposome-mediated macrophage ‘suicide’ technique. J Immunol Methods 124: 1–6, 1989. doi:10.1016/0022-1759(89)90178-6.
Crossref | PubMed | ISI | Google Scholar - 36. . Activation of the NLRP3 inflammasome complex is not required for stress-induced death of pancreatic islets. PLoS One 9: e113128, 2014. doi:10.1371/journal.pone.0113128.
Crossref | PubMed | ISI | Google Scholar - 37. . Autophagy regulates inflammation following oxidative injury in diabetes. Autophagy 9: 272–277, 2013. doi:10.4161/auto.23628.
Crossref | PubMed | ISI | Google Scholar - 38. . Mechanisms of NOD-like receptor-associated inflammasome activation. Immunity 39: 432–441, 2013. doi:10.1016/j.immuni.2013.08.037.
Crossref | PubMed | ISI | Google Scholar - 39. . A role for the NLRP3 inflammasome in metabolic diseases–did Warburg miss inflammation? Nat Immunol 13: 352–357, 2012. doi:10.1038/ni.2228.
Crossref | PubMed | ISI | Google Scholar - 40. . Induction of manganous superoxide dismutase by tumor necrosis factor: possible protective mechanism. Science 242: 941–944, 1988. doi:10.1126/science.3263703.
Crossref | PubMed | ISI | Google Scholar - 41. . Elimination of the NLRP3-ASC inflammasome protects against chronic obesity-induced pancreatic damage. Endocrinology 152: 4039–4045, 2011. doi:10.1210/en.2011-1326.
Crossref | PubMed | ISI | Google Scholar - 42. . Thioredoxin-interacting protein links oxidative stress to inflammasome activation. Nat Immunol 11: 136–140, 2010. doi:10.1038/ni.1831.
Crossref | PubMed | ISI | Google Scholar
