volume 11 number 2 february 2010 nature immunology
Active, mature interleukin 1β (IL-1β) is produced by cleavage of the
inactive pro-IL-1β precursor by caspase-1, which is activated in a
large multiprotein complex called the inflammasome1,2. The NLRP3
inflammasome, composed of the Nod-like receptor protein NLRP3
(also called cyopyrin or NALP3), CARDINAL, the adaptor protein
ASC and caspase-1 (ref. 3), is vital for the production of mature IL-1β
in response to a variety of signals. The NLRP3 inflammasome
is activated by bacterial toxins4 or pathogen-associated molecular
patterns, such as muramyldipeptide5. NLRP3 can also detect endog-
enous stress-associated danger signals, such as ATP4, monosodium
urate crystals (MSU)6 or β-amyloid7. Although reactive oxygen spe-
cies (ROS), together with low cytoplasmic concentrations of potas-
sium (K+), are known to be prerequisites for activation of the NLRP3
inflammasome8,9, the signaling pathways that lead to NLRP3 activa-
tion are still poorly understood.
Three distinct signaling pathways for inflammasome activation
have been proposed. In one model, activators enter the cytoplasm
via pannexin 1, a channel protein that is associated with and opened
by the purinergic receptor P2X7 (ref. 10). In a second model, cathe-
psin B released from ruptured lysosomes after phagocytosis of large
particles such as silica or β-amyloid cleaves an unidentified sub-
strate and triggers inflammasome activation7,11. In a third model,
ROS, which are generated by all known NLRP3 activators, induce a
conformational change in an unidentified protein that subsequently
activates the inflammasome1. To clarify the relative contribution
of these proposed signaling pathways, we searched for previously
unknown NLRP3 binding partners. One of the proteins we identi-
fied was thioredoxin-interacting protein (TXNIP; also called VDUP1
or TBP-2), which in conditions of oxidative stress binds NALP3 and
leads to its activation.
TXNIP is an NLRP3-binding protein
A yeast two-hybrid screen using the leucine-rich repeats (LRRs) of
NLRP3 as bait has identified TXNIP12,13 as a potential binding part-
ner of NLRP3. We confirmed the interaction between TXNIP and
NLRP3 in human embryonic kidney (HEK293) T cells. Overexpressed
TXNIP (Fig. 1a) and endogenous TXNIP (Fig. 1b) bound full-length
NLRP3 and the LRR region of NLRP3, as expected. The nucleotide-
binding NACHT domain of NLRP3 also interacted with TXNIP.
Endogenous TXNIP did not bind other members of the NLR family
or other proteins containing LRRs (Fig. 1c), which suggested that the
TXNIP interaction is a specific feature of NLRP3. TXNIP contains
two arrestin domains and a carboxy (C)-terminal extension with no
apparent homology to other proteins (Fig. 1a). The C-terminal arres-
tin domain ‘preferentially’ interacted with NLRP3 (Fig. 1d).
Activation of the NLRP3 inflammasome is dependent on the
generation of ROS14. In fact, all known NLRP3 activators generate
ROS and, conversely, inhibitors of ROS block inflammasome acti-
vation14. Consistent with that, we found that the addition of H2O2
was sufficient to trigger the processing and secretion of IL-1β in an
NLRP3–ASC–caspase-1–dependent manner in THP-1 human macro-
phages (Fig. 2a and Supplementary Fig. 1a). In contrast, we still
observed IL-1β processing in the absence of the caspase-1 activa-
tor Ipaf (Supplementary Fig. 1b), which indicated that Ipaf did not
participate in the ROS-mediated inflammasome activation. TXNIP
has been identified as a binding partner of reduced thioredoxin
(TRX)15. TXNIP directly interacts with the redox-active domain of
TRX and is thought to function as a negative regulator of the TRX
reductase activity. At high concentrations of H2O2, TXNIP dissoci-
ated from TRX (Fig. 2b). We also observed dissociation with other
1Department of Biochemistry and 2Center of Integrative Genomics, University of Lausanne, Epalinges, Switzerland. 3Korea Research Institute of Bioscience and
Biotechnology, Yusong, Taejon, Republic of Korea. Correspondence should be addressed to J.T. (email@example.com).
Received 27 August; accepted 11 November; published online 20 December 2009; doi:10.1038/ni.1831
Thioredoxin-interacting protein links oxidative stress
to inflammasome activation
Rongbin Zhou1, Aubry Tardivel1, Bernard Thorens2, Inpyo Choi3 & Jürg Tschopp1
The NLRP3 inflammasome has a major role in regulating innate immunity. Deregulated inflammasome activity is associated
with several inflammatory diseases, yet little is known about the signaling pathways that lead to its activation. Here we show
that NLRP3 interacted with thioredoxin (TRX)-interacting protein (TXNIP), a protein linked to insulin resistance. Inflammasome
activators such as uric acid crystals induced the dissociation of TXNIP from thioredoxin in a reactive oxygen species (ROS)-
sensitive manner and allowed it to bind NLRP3. TXNIP deficiency impaired activation of the NLRP3 inflammasome and
subsequent secretion of interleukin 1 (IL-1). Akin to Txnip –/– mice, Nlrp3 –/– mice showed improved glucose tolerance and
insulin sensitivity. The participation of TXNIP in the NLRP3 inflammasome activation may provide a mechanistic link to the
observed involvement of IL-1 in the pathogenesis of type 2 diabetes.
© 2010 Nature America, Inc. All rights reserved.
Reagents. Nigericin, uric acid crystals (MSU), ATP, phorbol 12-myristate
13-acetate, APDC ((2R,4R)-4-aminopyrrolidine-2,4-dicarboxylate), insu-
lin, uricase, glibenclamide, silica, poly(dA:dT) and glucose were from Sigma;
Imject Alum (aluminum hydroxide adjuvant) was from Pierce; H2O2 was from
Applichem; DPI was from Alexis; and R-837, zymosan and ultrapure lipopoly-
saccharide were from Invivogen. S. typhimurium was a gift from R.V. Bruggen.
Antibody to human pro-IL-1β (anti–human pro-IL-1β) was produced ‘in house’.
Monoclonal mouse anti-NLRP3 (Cryo-2) was from Adipogen. Anti–mouse
IL-1β was a gift from R. Solari. Anti–mouse caspase-1 (p20) was a gift from
P. Vandenabeele. The following antibodies were also used: anti–human cas-
pase-1 (sc-622; Santa Cruz), anti–human IL-1β, cleaved (D116; Cell Signaling),
anti-ASC (AL177; Alexis), anti-Ipaf (3107; ProSci), anti-Flag (F7425; Sigma),
anti-VSV (V4888; Sigma), anti-tubulin (T5168; Sigma), rabbit anti-TXNIP
(40-3700; Invitrogen), mouse anti-TXNIP (JYI; MBL) and anti-TRX (ab26928;
Abcam). All tissue culture reagents were from Invitrogen.
Mice. Nlrp3–/–, Txnip–/– and P2rx7–/– mice have been described6,28. All mice
were on a C57BL/6 background. All animal experiments were approved by a
local ethics committee.
Cell culture. Human THP-1 cells were grown in RPMI-1640 media supple-
mented with 10% (vol/vol) FBS and 50 µM 2-mercaptoethanol. THP-1 cells
were differentiated for 3 h with 100 nM phorbol-12-myristate-13-acetate. Bone
marrow macrophages were derived from tibia and femoral bone marrow pro-
genitors as described elsewhere36 and were cultured in DMEM complemented
with 10% (vol/vol) FCS, 1 mM sodium pyruvate and 2 mM L-glutamine36.
Macrophages were primed for 16–18 h with ultrapure lipopolysaccharide
For IL-1β induction, 1 × 106 macrophages were plated in 12-well plates
overnight. The medium was changed to Opti-MEM the next morning and
cells were treated for 6 h with MSU (150 µg/ml), alum (200 µg/ml), silica
(200 µg/ml), R-837 (15 µg/ml) or H2O2 (0.1–10 mM) or for 30 min with
ATP (5 mM). For transfection of poly(dA:dT) into cells, Lipofectamine 2000
(4 µl/ml) was used according to the manufacturer’s protocol (Invitrogen).
Cell extracts and precipitated supernatants were analyzed by immunoblot.
Islets from male C57BL/6 mice were isolated by collagenase digestion and
Histopaque density-gradient purification and were suspended in RPMI-1640
medium containing glucose (11.1 mmol/l) supplemented with 10% (vol/vol)
FBS and 2 mM l-glutamine37. The next morning, the medium was replaced
with RPMI-1640 medium containing 2.5 mM glucose and 10% (vol/vol) FBS,
followed by incubation for 2 h. Islets (400 islets/ml) were then incubated for
24 h or 48 h in the presence of 5 mM or 33 mM glucose, and supernatants and
cell extracts were collected.
Enzyme-linked immunosorbent assay. Cell culture supernatants or tissue
lysates were assayed for mouse IL-1β, tumor necrosis factor (R&D Systems),
IL-6 (eBioscience) and interferon-β (PBL InterferonSource) according to the
Transfection and immunoprecipitation. Constructs were transfected into
HEK293T cells by the calcium-phosphate method; after 18 h, cells were col-
lected and resuspended in lysis buffer (50 mM Tris, pH 7.8, 50 mM NaCl, 0.1%
(vol/vol) Nonidet-P40, 5 mM EDTA and 10% (vol/vol) glycerol). Extracts
were immunoprecipitated with anti-Flag agarose beads and then were assessed
by immunoblot. THP-1 cells (1 × 107) were differentiated with phorbol
12-myristate 13-acetate and then were treated for 2 h with MSU (150 µg/ml),
R-837 (15 µg/ml) or H2O2 (10 mM) with or without 30 min of pretreatment
with APDC (100 µM). THP-1 cells were then resuspended in lysis buffer and
proteins were immunoprecipitated from extracts with anti-TXNIP (C40-3700;
Invitrogen) or anti-TRX (ab26320; Abcam).
Knockdown or overexpression in THP-1 cells. THP-1 cells stably expressing
short hairpin RNA specific for NLRP3, caspase-1, ASC, Ipaf or TRX were
obtained as described38. THP-1 cells were transfected with small interfering
RNA by the Amaxa Nucleofector system and the Cell Line Nucleofector kit V
according to the manufacturer’s recommendations. TXNIP was cloned into the
lentiviral vector pRDI292 for overexpression of TXNIP in THP-1 cells.
Salmonella infection. S. typhimurium was precultured with tryptic soy broth
the day before the experiment, then was diluted 1:10 and incubated for 2 h.
Lipopolysaccharide-primed BMDMs were infected for 1 h with various dilu-
tions of the S. typhimurium culture (multiplicity of infection, 10 or 100). Cells
were washed and then incubated for 1 h in OptiMEM containing gentamycin
Peritonitis. MSU- or zymosan-induced peritonitis has been described6; this
was induced by intraperitoneal injection of 1 mg MSU crystals or 0.2 mg
zymosan dissolved in 0.5 ml sterile PBS. After 6 h, mice were killed by exposure
to CO2 and peritoneal cavities were washed with 10 ml cold PBS. Peritoneal
lavage fluid was assessed by flow cytometry with the neutrophil markers Ly6G
and CD11b (BD Pharmingen) for analysis of the recruitment of polymorpho-
Metabolic studies. Mice 6 weeks of age were fed a normal diet or high-fat diet
(D12331; Researchdiets) for 8 weeks. For glucose-tolerance tests, mice were
fasted for 8 h and then were injected intraperitoneally with glucose at a dose
of 1.5 mg per g body weight. Blood samples were obtained at various time
points for glucose measurements with a Glucometer (Roche) or for insulin
measurements by enzyme-linked immunosorbent assay.
ROS assay. THP-1 cells were loaded for 10 min with the ROS-specific
fluorescent probe H2DCFDA (2′,7′-dichlorofluorescin diacetate; 10 mM;
BioChemika or Fluka), then were washed twice with PBS and exposed to MSU.
Fluorescence was assessed with a Titertek Fluoroskan II microtiter fluores-
cence plate reader with a fluorescein isothiocyanate filter (excitation, 485 nm;
emission, 538 nm).
Statistical analysis. Samples were analyzed by Student’s t-test for comparison of two
groups and analysis of variance (ANOVA) for comparison of multiple groups.
36. Didierlaurent, A. et al. Tollip regulates proinflammatory responses to interleukin-1
and lipopolysaccharide. Mol. Cell. Biol. 26, 735–742 (2006).
37. Gotoh, M. et al. Reproducible high yield of rat islets by stationary in vitro digestion
following pancreatic ductal or portal venous collagenase injection. Transplantation
43, 725–730 (1987).
38. Papin, S. et al. The SPRY domain of Pyrin, mutated in familial Mediterranean fever
patients, interacts with inflammasome components and inhibits proIL-1β processing.
Cell Death Differ. 14, 1457–1466 (2007).
© 2010 Nature America, Inc. All rights reserved.