NLRP3 inflammasome activation and interleukin-1β release in macrophages require calcium but are independent of calcium-activated NADPH oxidases

2.1. Purification of human monocytes and macrophage differentiation

Human subjects provided written informed consent for participation under the guidelines of NIH and UGA IRB-approved protocols.[24, 25] Whole blood was drawn at the Transfusion Medicine Branch of the National Institutes of Health or the University of Georgia Health Center. 40 mL blood was anticoagulated by ACD (anticoagulant citrate dextrose) or heparin. Red blood cells were removed by Dextran sedimentation. White blood cells were layered on 5-step Percoll (Sigma) gradient: 85%-80%-75%-70%-65% and centrifuged (800 xg, 30 min, 13° C). PBMCs containing monocytes (PBS/65% Percoll interface) were separated, washed, suspended in RPMI-1640 medium (Life Technologies, Grand Island, NY) and added to 6-well plates (BD Primeria, Swedesboro, NJ). After 30 min, adherent cells were washed four times with HBSS. Cells were differentiated into macrophages by incubating them in RPMI-1640 medium supplemented with 10% pooled human heat-inactivated serum and 50 ng/ml human macrophage colony stimulating factor (hM-CSF, R&D Systems, Minneapolis, MN) for 5–8 days. When mentioned, MDMs were primed with 1 μg/ml Salmonella typhimurium LPS (Sigma Aldrich, St. Louis, MO) for 30 min.

2.2. Collection of human serum

Blood (10 ml) was also collected from healthy volunteers to separate serum. After coagulation, sera were collected, centrifuged (10000g, 5min), filtered through 0.22 μm size sterile filter and stored at −80 °C for future use. Serum samples from at least 5 different donors were thawed, pooled and heat-inactivated (56 °C, 30 min). 10% heat-inactivated pooled human serum was used on MDM cultures.

2.3. Differentiation of murine BMDMs

Duox1-deficient mice (purchased from Lexicon Pharmaceuticals, Inc., The Woodlands, TX, USA) were previously characterized.[26] These mice were shown to contain a gene-trapping cassette between DUOX1 exons 20 and 21, resulting in the failure to produce detectable Duox1 protein in uroepithelial cells. Mice were maintained in a specific pathogen-free facility at the NIAID according to IUCAC animal protocols approved at the NIH. Bone marrow cell suspensions were isolated from tibias and femurs of 8- to 12-week-old C57Bl/6 mice by flushing bone marrows with RPMI1640 medium (+10% FCS, +1% Pen/Strep). Cell clumps were dislodged by pipetting, and debris was removed by passaging through a 70-μm cell strainer (Fisher Scientific, Pittsburg, PA, USA). Cells were washed twice with medium and seeded on 10 cm tissue culture dishes (for Ca²⁺ measurements) or on 24-well plates (for ELISA) (Corning Costar, Tewskbury, MA). Cells were grown in complete RPMI-1640 medium supplemented with 10 ng/mL recombinant murine M-CSF (R&D Systems, Minneapolis, MN) and cultured for 5–7 days in a humidified incubator at 37°C and 5% CO₂.

2.4. Manipulations of intra- and extracellular calcium

To scavenge extracellular Ca²⁺, human and murine macrophages were incubated in HBSS supplemented with 1 mM MgCl₂, 10 mM HEPES, 5 mM glucose and 100 μM EGTA, pH 7.4. Corresponding Ca²⁺-containing solutions contained 1 mM CaCl₂ and no EGTA (pH 7.4). To chelate intracellular Ca²⁺, human MDMs were incubated with 50 μM BAPTA-AM for 15 min at 37°C in HBSS and washed twice in PBS afterwards to remove excess dye.

2.5. Calcium measurements

Murine macrophages were primed with 1μg/ml Salmonella typhimurium LPS (Sigma Aldrich, St. Louis, MO) for 1hr. After two washes, cells were incubated in HBSS containing 2 μM FURA2-AM (Life Technologies, Grand Island, NY) for 30 min in the dark. Cells were washed, suspended in HBSS and were added to black 96-well plates (Corning Costar, Tewksbury, MA). After 10-minute incubation, cells were stimulated by different doses of ATP (0–3 mM) (Sigma Aldrich, St. Louis, MO) and changes in 510nm fluorescence emission while exciting at two wavelengths (340nm_ex/510nm_em, 390nm_ex/510nm_em) were followed for 60 min with shaking. The ratio of 340 nm/510nm and 390 nm/510 nm values were calculated and presented to reflect kinetic changes in cytosolic Ca²⁺ levels.

2.6. RNA isolation and quantitative real-time RT-PCR

RNA was isolated by Trizol/chloroform extraction and isopropanol precipitation from differentiated hMDMs. cDNA synthesis was carried out (Thermoscript cDNA synthesis kit, Life Technologies, Grand Island, NY). Human DUOX1, DUOX2, NOX5, NLRP3 and P2RX7 gene expression was measured by reverse transcription/quantitative real-time PCR using SYBR Green (Invitrogen) and the following primers:

Human DUOX1 (F: 5′-CACCTCCTGGAGACCTTTTTC -3′, R: 5′ GGCCTGGTTGATGTCCAG -3′, 60 bp product), human DUOX2 (F: 5′-GATGGTGACCGCTACTGGTT -3′, R: 5′– GCCACCACTCCAGAGAGAAG -3′, 323 bp product), human NOX5 (F: 5′-CAAGAATGAAGCCGCAGAC -3′, R: 5′-CCTGCAATGGTCTTAAACTGC -3′, 95 bp product), human NLRP3 (F: 5′-CTTCTCTGATGAGGCCCAAG -3′, R: 5′-GCAGCAAACTGGAAAGGAAG -3′, 200 bp product), human P2RX7 (F: 5′-GGGAACCAGAAGACCTGTGA -3′, R: 5′-AGTTTTCGGCACTGTTCAAGA -3′, 94 bp product) and β-actin (F: 5′-CCAACCGCGAGAAGATGA-3′, R: 5′-CCAGAGGCGTACAGGGATAG-3′, 97 bp product). Changes in fluorescence were followed in 7500 Real-time PCR system (Applied Biosystems, Grand Island, NY). Reaction parameters were 50°C for 2 minutes, 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. The relative abundance of mRNAs was obtained using the comparative cycle threshold method and was normalized to beta-actin as the endogenous control.

2.7. Measurement of human and murine IL-1β release by ELISA

Human MDMs or mouse BMDMs cultured on 24-well plates were primed with Salmonella typhimurium LPS (1 μg/ml, Sigma Aldrich, St. Louis, MO) for 60 min. Cells were washed twice with Ca²⁺-free HBSS before addition of inflammasome stimuli (ATP, 0–4 mM or 100 μg/ml silica crystals (Sigma Aldrich, St. Louis, MO)). At indicated time points, supernatants were collected, centrifuged (500g, 10 min) to remove cell debris, crystals etc. Supernatants were either frozen or immediately used for further analysis. Human and murine IL-1β release in macrophage supernatants was determined by commercial ELISA kits (DuoSet, R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions. Cytokine release was calibrated using standard curves.

2.8. Western Blotting

HEK-293 cells or human MDMs were washed three times with cold Ca²⁺- and magnesium-free PBS and lysed by NP-40 lysis buffer (Boston Biosciences, Ashland, MA) containing 150 μM PMSF (Fluka Biochemika-Sigma Aldrich, St. Louis, MO) and 1% protease inhibitor cocktail (dissolved in DMSO, Sigma Aldrich, St. Louis, MO). Lysates were centrifuged and protein concentrations in supernatants were determined using the BCA assay (Pierce, Rockford, IL). Equal amounts of protein were added to SDS sample buffer containing 2-mercaptoethanol (without boiling), and electrophoresed on SDS-polyacrylamide gels (8%; Tris-Glycine Gel, Life Technologies, Grand Island, NY). Gels were blotted on nitrocellulose membrane (Invitrogen) using the TE Series Transpher Electrophoresis Unit (Hoefer, Holliston, MA) wet blotting system. Blots were blocked overnight in TTBS (TBS-buffer containing 5% milk powder and 0.05% Tween-20). Blots were incubated with primary antibodies (RT, 1hr, TTBS), washed three times with TTBS and then probed with secondary HRP-linked antibodies (RT, 1hr, TTBS, GE Healthcare, Piscataway, NJ). After repeated washes, blots were developed by chemiluminescence using the Lumigen DS detection kit (GE Healthcare, Piscataway, NJ). The primary antibodies used were: anti-Duox1 (Clone T-14, 1:500, rabbit, Santa Cruz, Dallas, TX), anti-Hsp90 (1:1000, mouse, mouse, Santa Cruz, Dallas, TX), anti-caspase-1 p20 (rabbit, 1:500, Cell signaling, Danvers, MA), anti-actin (rabbit, 1:2000, Sigma Aldrich, St. Louis, MO), HA (mouse, 1:100, Covance, Princeton, NJ).

2.9. Caspase-1 enzymatic activity

Differentiated human MDMs were primed with 1 μg/ml Salmonella typhimurium LPS (Sigma Aldrich, St. Louis, MO, USA) for 30 min. To scavenge intracellular calcium, MDMs were treated with 50 μM BAPTA-AM (Dojindo laboratory, Rockville, Maryland USA) in a PBS with 5 mM glucose for 10 min. Equal volume of DMSO (Sigma Aldrich, St. Louis, MO,USA), solvent for BAPTA-AM, was used as a control. MDMs were then stimulated with 3mM ATP in HBSS containing 2mM glucose with or without 1mM Ca²⁺ for 10 min. Ca²⁺-free HBSS contained 1 mM MgCl₂ (Sigma Aldrich, St. Louis, MO,USA) and 100 μM EGTA (Sigma Aldrich St. Louis, MO, USA) was used to chelate any remaining Ca²⁺. MDMs were then scraped off to measure caspase-1 enzymatic activity using caspase-1 fluorometric assay kit (Enzo Life Sciences, Farmingdale, NY, USA). Following the manufacturer’s instruction, cold lysis buffer was added to macrophages. Equal volume of the reaction buffer containing 10 mM DDT was then added to the lysis buffer. Finally, 50 μM of caspase-1 YFAD-AFC substrate was used to detect fluorescence at excitation and emission wavelengths of 450 nm and 505 nm, respectively. Results were reported as relative fluorescence units (RFU) and kinetics of the caspase-1 signal development was obtained using Varioskan^™ Flash Multimode Reader (Thermo-scientific, USA) for 1.5 hours at 37 °C. Final endpoint fluorescent values were compared between 2 groups treated, one treated with media containing Ca²⁺ and once treated without Ca²⁺.

2.10. Statistical analysis

Data are represented as mean +/− S.E.M. or mean+/− S.D. Significance was calculated with ANOVA and Tukey’s pot-hoc test and was marked as * when p < 0.05, ** when p < 0.01 and *** when p < 0.001.

3.1. Calcium is critical for IL-1β secretion of human macrophages stimulated by ATP or silica

Recently, a critical role of Ca²⁺ in NLRP3 inflammasome activation has been shown.[1] IL-1β release from murine BMDMs and human monocytes is inhibited by intra- or extracellular chelation of Ca²⁺.[1, 11] Since the role of Ca²⁺ in IL-1β secretion by human macrophages has not been demonstrated so far and differences exist between human monocytes and macrophages in terms of regulation of IL-1β expression and inflammasome activation, we differentiated monocytes isolated from peripheral human blood into macrophages by treatment with 50 ng/ml human M-CSF and measured IL-1β release by ELISA.[27, 2] We stimulated inflammasome activation by ATP, a well-described stimulus to induce IL-1β release in monocytes/macrophages.[28] Extracellular ATP enhanced IL-1β secretion in a concentration-dependent manner (Fig. 1A). Chelation of extracellular Ca²⁺ by ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA) or intracellular Ca²⁺ by 1,2-bis(o-aminophenoxy)ethane- N,N,N′,N′-tetraacetic acid (BAPTA) decreased ATP- and silica-stimulated cytokine release (Fig. 1B). Combined effects of EGTA and BAPTA resulted in an 80.2+/−8.0% (mean+/−S.E.M., n=6) inhibition of silica- and 70.4+/−7.4% inhibition (mean+/−S.E.M., n=5) of ATP-activated IL-1β release (Fig. 2B). We observed similar robust inhibition of IL-1β secretion of human MDMs by using two additional known inflammasome activators, antimycin-A and calcium pyrophosphate dehydrate crystals (data not shown). Thus, Ca²⁺ plays an essential role in IL-1β release by human macrophages.

3.2. Calcium-inducing agonists trigger IL-1β release

To investigate whether the elevation in intracellular Ca²⁺ alone is sufficient to induce IL-1β secretion in human MDMs we used pharmacological manipulation to disrupt Ca²⁺ homeostasis. Thapsigargin (TG) is a widely used inhibitor of the sarco-/endoplasmic reticulum Ca²⁺ ATPase.[3] TG first empties intracellular calcium stores that activates the store-operated calcium entry pathway leading to subsequent calcium influx from the extracellular space.[3] TG-treatment of lipopolysaccharide (LPS)-primed human MDMs alone induced release of IL-1β (Fig. 2A). Ionomycin (iono), a calcium ionophore, enhances intracellular Ca²⁺ concentrations abruptly by releasing Ca²⁺ from both, extra- and intracellular stores.[6] Addition of ionomycin to human MDMs increased IL-1β secretion (Fig. 2B). ATP-induced IL-1β release did not differ in the presence or absence of ionomycin suggesting that ATP also releases Ca²⁺ from stores both inside and outside of the cell (Fig. 2B). Our data show that increases in cytosolic Ca²⁺ alone are enough to induce IL-1β secretion by human macrophages.

3.3. Macrophage inflammasome activation requires Ca²⁺

Next, we asked whether Ca²⁺ is required for IL-1β release either upstream or downstream of NLRP3 inflammasome activation. We stimulated human MDMs in the presence or absence of Ca²⁺ (BAPTA+EGTA) with 3 mM ATP, a well-characterized NLRP3 agonist, and measured the true indicator of inflammasome activation, procaspase-1 cleavage. Once assembled, the direct consequence of NLRP3 inflammasome activation is proteolytic activation of procaspase-1 into the bioactive, 20kD-size caspase-1.[20] While in presence of calcium the 20 kD caspase-1 cleavage product appeared upon stimulation by ATP, in absence of Ca²⁺, this band was missing (Fig. 3A). In addition, we measured caspase-1 enzymatic activity in human MDMs stimulated with 3mM ATP in the presence or absence of calcium (BAPTA+EGTA), and we experienced an inhibition in caspase-1 enzymatic activity upon calcium removal (Fig. 3B). These results show that Ca²⁺ is required for NLRP3 inflammasome activation in human macrophages (Fig. 3).

3.4. Duox1 is expressed in human macrophages

ROS were suggested to play a role in NLRP3 inflammasome regulation.[12] Mitochondria have been identified as the source of ROS but certain NADPH oxidases have been excluded.[16, 19, 20, 10] Studies on phagocytic NADPH oxidase-deficient monocytes showed that four members of the Nox/Duox family of NADPH oxidases (Nox1-4) that are dependent on p22^phox are not involved in inflammasome activation.[10, 19, 20] There are, however, three other members of this family which could be ROS sources required for inflammasome activation in macrophages: Nox5, Duox1 and Duox2.[17] Whether these oxidases are required for inflammasome activation is unknown. Nox5, Duox1 and Duox2 are all directly activated by increases in intracellular Ca²⁺ levels through their intracellular Ca²⁺-binding EF-hand motifs.[22, 21, 23] Innate immune functions of Duox1 have already been previously proposed but it is unknown if these enzymes are expressed in human macrophages.[29, 17, 30] Therefore, we measured gene expression levels of NOX5, DUOX1 and DUOX2 in human MDMs by real-time PCR. Nox5 mRNA was not detected at all, Duox2 mRNA was only detected in one of the four donors whereas Duox1 mRNA was consistently detected (Fig. 4A). DUOX1 expression was significantly lower than that of NLPR3 or P2RX7 (Fig. 4A). Probing human MDM cell lysates of two donors with a commercial Duox1 antibody, we detected a clear band at about 180 kDa molecular mass in both donors (Fig. 4B). Thus, we identified a new cell type in humans expressing Duox1: macrophages.

To test the isoform specificity of the Duox antibody, we used cell lysates prepared from our previously characterized HEK-293 cell model lines stably expressing bicistronic pcDNA5.1^™/FRT constructs of human Duox1+DuoxA1α or human Duox2+DuoxA2.[23] We used cell lines containing HA-tagged constructs to reliably determine the amounts of Duox by Western blotting.[10] The parental cell line (HEK-293) did not show detectable HA or Duox (Fig. 4C). The intensity of the HA-band was stronger in Duox2/DuoxA2-expressing cells compared to that of the Duox1/DuoxA1S sample, indicating higher protein/cell expression for Duox2 than Duox1 (Fig. 4C). The used Duox1 antibody detected Duox1 but not Duox2 (Fig. 4C), confirming that human MDMs express the Ca²⁺-activated NADPH oxidase Duox1.

3.5. Duox1 plays no role in NLRP3 inflammasome activation

To address the question whether Duox1 is required for inflammasome activation and IL-1β release, first we tested whether murine BMDMs produce cytosolic Ca²⁺ responses to ATP. Cytosolic Ca²⁺ levels did not rise above baseline unless cells were stimulated with 1 mM or higher doses of ATP (Fig. 5A). 1mM ATP triggered a short, temporary Ca²⁺ transient whereas 3mM ATP, a concentration triggering maximal inflammasome activation resulted in a robust, long-lasting Ca²⁺ signal (Fig. 5A). We next exposed wild-type and Duox1-deficient murine BMDMs to ATP or silica and measured IL-1β release.[26] Duox1-deficiency does not impair the ability of murine BMDMs to secrete IL-1β (Fig. 5B). Thus, we conclude that Duox1 does not play a role in inflammasome activation in macrophages.