Silver Nanoparticles Induce Degradation of the Endoplasmic Reticulum Stress Sensor Activating Transcription Factor-6 Leading to Activation of the NLRP-3 Inflammasome

Background: Some nanoparticles are known to induce endoplasmic reticulum (ER) stress and lead to cell death. Results: Silver nanoparticles induce ATF-6 degradation, leading to activation of the NLRP-3 inflammasome and pyroptosis. Conclusion: ATF-6 is an important target to silver nanoparticles. Significance: Our results provide a new link between ER stress and activation of the NLRP-3 inflammasome. In the past decade, the increasing amount of nanoparticles (NP) and nanomaterials used in multiple applications led the scientific community to investigate the potential toxicity of NP. Many studies highlighted the cytotoxic effects of various NP, including titanium dioxide, zinc oxide, and silver nanoparticles (AgNP). In a few studies, endoplasmic reticulum (ER) stress was found to be associated with NP cytotoxicity leading to apoptosis in different cell types. In this study, we report for the first time that silver nanoparticles of 15 nm (AgNP15), depending on the concentration, induced different signature ER stress markers in human THP-1 monocytes leading to a rapid ER stress response with degradation of the ATF-6 sensor. Also, AgNP15 induced pyroptosis and activation of the NLRP-3 inflammasome as demonstrated by the processing and increased activity of caspase-1 and secretion of IL-1β and ASC (apoptosis-associated speck-like protein containing a CARD domain) pyroptosome formation. Transfection of THP-1 cells with siRNA targeting NLRP-3 decreased the AgNP15-induced IL-1β production. The absence of caspase-4 expression resulted in a significant reduction of pro-IL-1β. However, caspase-1 activity was significantly higher in caspase-4-deficient cells when compared with WT cells. Inhibition of AgNP15-induced ATF-6 degradation with Site-2 protease inhibitors completely blocked the effect of AgNP15 on pyroptosis and secretion of IL-1β, indicating that ATF-6 is crucial for the induction of this type of cell death. We conclude that AgNP15 induce degradation of the ER stress sensor ATF-6, leading to activation of the NLRP-3 inflammasome regulated by caspase-4 in human monocytes.

Nanoparticles (NP) 3 are used in many applications and in a variety of sectors, including textile, aerospace, electronics, and medical healthcare. Silver nanoparticles (AgNP) are among the most commonly used NP in nanomedicine, mainly because of their potent antimicrobial properties, increasing the interest to use them for drug delivery (1). Indeed, silver ions and nanosilver were shown to be highly toxic for various types of microorganisms, including Pseudomonas spp. and Escherichia spp. (2,3). Even if potential exposure of humans to AgNP is already high, it will certainly increase in the becoming years. Because the toxicity of AgNP in humans is not fully understood, it is highly relevant to investigate their mode of action at the cellular and molecular level in humans.
Endoplasmic reticulum (ER) stress leads to unfolded protein response, a major hallmark of cytotoxicity. To date, three ER stress sensors have been documented: protein kinase RNA-like endoplasmic reticulum kinase (PERK), inositol-requiring enzyme 1 (IRE-1), and activating transcription factor 6 (ATF-6). IRE-1 and PERK both contain cytoplasmic kinase domains known to be activated by homodimerization and autophosphorylation in the presence of ER stressors (4 -6). In the case of ATF-6, accumulation of unfolded proteins induces ATF-6 transition to the Golgi, where it is cleaved by two transmembrane proteins, Site-1 and Site-2 proteases (7). ATF-6 cleavage yields a cytoplasmic protein acting as an active transcription factor. Although short-term ER stress events lead to pro-survival transcriptional activities, prolonged ER stress activates the major apoptotic pathways (8,9). Moreover, ER stress-related events were recently proposed as an early biomarker for nanotoxicological evaluation (10). A few studies have reported ER stressrelated events induced by NP in human cell lines and in zebrafish (10 -12).
Pyroptosis, a type of programmed cell death sharing common features with apoptosis and necrosis, leads to the assembly of the inflammasomes and the formation of large structures called pyroptosomes characterized by aggregation of apoptosis-associated speck-like protein containing a CARD domain (ASC) (13). Formation of pyroptosomes allows recruitment and processing of caspase-1 into two active fragments, p10 and p20 (14). Caspase-1 controls processing and secretion of IL-1␤, one of the most potent endogenous pyrogenic molecules. IL-1␤ is responsible for inflammatory cell infiltration and is known to induce cyclooxygenase and increase expression of adhesion molecules, production of reactive oxygen species, and other inflammatory soluble mediators (15). Secretion of high concentrations of IL-1␤ is also associated with chronic inflammatory conditions, including rheumatoid arthritis and inflammatory bowel diseases (16). Interestingly, treatment of some auto-immune diseases with anti-IL-1␤ antibodies results in significant reduction of disease severity and symptoms. Pyroptosis also leads to the release of cytosolic content via formation of pore in the cellular membrane, thereby increasing the inflammatory process (17). Some NP were shown to induce pyroptosis in human cells, namely carbon nanotubes, carbon black NP, and AgNP (18 -20). Therefore, studying the impact of several distinct NP in the regulation of the inflammasome has become highly relevant for investigating their toxicity.
In this study, we show that low concentrations of silver nanoparticles of 15 nm (AgNP 15 ) induced ER stress response but did not led to cell death, whereas higher concentrations resulted in atypical ER stress response associated with ATF-6 degradation and pyroptotic cell death through NLRP-3 inflammasome activation. Our data suggest a link between these two processes.
Characterization of AgNP 15 -The AgNP 15 suspension obtained from the manufacturer was examined by transmission electronic microscopy using a Hitachi H-7100 transmission electron microscope (21). The size distribution and electric charge ( potential) of AgNP 15 were determined by dynamic light scattering (DLS) using a Malvern Zetasizer Nano-ZS (model ZEN3600; Malvern Instruments Inc., Westborough, MA). Measurements were performed at 1, 5, 10, and 25 g/ml AgNP 15 in RPMI 1640 with HEPES and 100 units/ml penicillin, 100 g/ml streptomycin (further referred to as RPMI) ϩ 10% heat-inactivated FBS, at 37°C.
Cell Culture-WT human monocytic THP-1 cells were purchased from ATCC. Stable caspase-4-deficient (clone TB) THP-1 cells and the corresponding cell line containing a scrambled sequence (vector control) were kindly given by Dr. Alan G. Porter (22). These cells line were cultured in RPMI. Cell density never exceeded 1 ϫ 10 6 cells/ml, and viability was evaluated before each experiment and was always above 95%. For primary culture of monocytes and macrophages, PBMCs were first isolated from healthy blood donors after centrifugation over Ficoll-Hypaque gradient, as previously published (23). For monocyte isolation, 25 ϫ 10 6 PBMCs were incubated at 37°C in a 5% CO 2 atmosphere for 2 h in RPMI containing 10% autologous heat-inactivated serum in a Petri dish. Monocytes obtained by removing the non-adherent PBMCs were further incubated in RPMI ϩ 10% heat-inactivated FBS for another 12 h. The monocytes were washed twice with Hanks' balanced salt solution without Ca 2ϩ and Mg 2ϩ with 2 mM EDTA and harvested with a cell scraper. Human monocytederived macrophages were generated by incubating 2 ϫ 10 6 PBMCs at 37°C in a 5% CO 2 atmosphere for 2 h in RPMI with 10% heat-inactivated autologous serum in 48-well plates. Cells were then washed twice with warm Hanks' balanced salt solution and further incubated in RPMI 1640 supplemented with 2 ng/ml GM-CSF for 7 days, with medium renewal every 3 days, to obtain macrophages.
Cell Death Assessment-For assessment of cell death, THP-1 cells were treated for 1 or 24 h with the indicated agents. After different periods of time, cells were harvested and washed twice in PBS and then stained with FITC-annexin-V and propidium iodine, according to the manufacturer's protocol (Life Technologies). Cells were analyzed by flow cytometry using a BD FACScan.
Western Blot Analysis-Cells were stimulated at 1 ϫ 10 6 cells/ml with the indicated agonists for various periods of time, as specified. At the end of the incubation periods, cells were lysed in Laemmli sample buffer (0.25 M Tris-HCl (pH 6.8), 8% SDS, 40% glycerol, and 20% 2-mercaptoethanol), and aliquots corresponding to 5 ϫ 10 5 cells were loaded onto 10% SDS-PAGE gels and transferred to nitrocellulose membranes for the detection of specific proteins. Membranes were blocked for 1 h at room temperature in 5% nonfat dry milk. After washing, the primary antibodies were added at a final dilution of 1:1000 in TBS-Tween 0.15%. The membranes were kept overnight at 4°C, and then washed with TBS-Tween and incubated for 1 h at room temperature with the appropriate secondary HRP antibody 1:25,000 in TBS-Tween followed by several washes. Protein expression was revealed using Luminata Forte Western HRP substrate (Millipore). Membranes were stripped with ReBlot Plus Strong (Millipore) and reprobed to confirm equal loading of proteins. Chemiluminescence was revealed with a chemiDocTM MP imaging system from Bio-Rad.
Assembly of Pyroptosome by Confocal Microscopy-THP-1 cells were incubated on poly-L-lysine coverslips for 30 min at 37°C and then incubated with the indicated agonist for 30 min.
In some experiments, cells were primed with LPS (100 ng/ml) for 4 h. Coverslips were then washed three times with PBS and fixed in 3.7% paraformaldehyde. After three washes in PBS, cells were permeabilized in PBS containing 0.1% saponin and 0.05% Tween 20. Cells were blocked with a blocking solution containing 5% goat serum, 1% BSA, and 1% dry fat milk and stained with mouse anti-ASC (2 g/ml) primary antibody. After three washes with PBS, cells were incubated with Alexa Fluor-488 goat anti-mouse IgG (1:500) secondary antibody. Cells were washed three times with PBS, and coverslips were mounted using ProLong Gold antifade reagent containing DAPI. Cells were then visualized with a Zeiss LSM780 on Axio Observer Z1 confocal microscope using a Plan-Apochromat 63 ϫ 1.4 NA oil differential interference contrast objective. Five fields with at least 50 cells/field were evaluated for the presence of pyroptosome.
Caspase-1 Assay-Caspase-1 activity was measured with the caspase-1 colorimetric assay kit (R&D Systems) as described previously (24). Briefly, 4 ϫ 10 6 cells were stimulated with agonists for the indicated periods of time and lysed with ice-cold lysis buffer after incubation for 10 min on ice. Cells were centrifuged at 10,000 ϫ g for 1 min, and supernatants were used for enzymatic reaction. Caspase-1 colorimetric substrate (WEHD-pNA) was added to each reaction mixture with the corresponding volume of cell lysate, reaction buffer, and DTT. The plate was incubated for 2 h at 37°C. After incubation, results were read with a microplate reader using a wavelength of 405 nm. Results are expressed as -fold increase of caspase-1 activity.
Caspase-3 and Caspase-4 Activity Assays-Caspase-3 and caspase-4 activity assays were performed as previously published with a few modifications (25). Briefly, THP-1 cells (1 ϫ 10 6 cells/ml) were stimulated as indicated in figure legends. Cells were washed twice in PBS and disrupted using a lysis buffer containing 25 mM Hepes, pH 7.5, 5 mM EDTA, 5 mM MgCl 2 , 10 mM DTT, 0.5% Triton X-100, and protease inhibitor mixture (Pierce, Thermo Fisher). Protein concentration was determined using the Bradford assay. An amount of 10 g of proteins was mixed in 200 l of reaction buffer containing 50 mM Hepes, 10% sucrose, 0.1% CHAPS, 10 mM DTT with 100 M of the caspase-3 (DEVD-pNA) or caspase-4 (LEVD-pNA) substrate for 2 h at 37°C. Cleavage of the caspase substrate was monitored using a spectrophotometer at 405 nm.
Transmission Electronic Microscopy-Cells were treated with 25 g/ml AgNP 15 or buffer for 1 h and then fixed with glutaraldehyde (2.5%) and examined by transmission electron microscopy, as above.
IL-1␤ Production-The measurement of IL-1␤ was determined with a commercially available ELISA kit (Life Technologies). THP-1 cells, primary monocytes, or macrophages were incubated with the indicated agonists for 1 h in a 24-well plate in RPMI-10% FBS or autologous serum. In some experiments, cells were pre-incubated with LPS (100 ng/ml) for 4 h. Supernatants were harvested after centrifugation and stored at Ϫ80°C before determining the concentration of IL-1␤.
Statistical Analysis-Experimental data are expressed as means Ϯ S.E. One-way analysis of variance (Dunnett's multiplecomparison test) and two-way analysis of variance (Bonferroni's post test) were performed using GraphPad Prism (version 5.01). 15 -We first characterized AgNP 15 from the manufacturer stock solution by transmission electron microscopy, confirming a diameter close to 15 nm, but also indicating that the suspension contains NP with different ratios between smaller and larger diameters (Fig. 1A). Because RPMI 1640 supplemented with 10% FBS was the medium used for all experiments throughout this study, the size distribution and potential of AgNP 15 were determined by DLS for all tested concentrations in this medium at 37°C, the temperature at which cells were incubated. As shown in Fig. 1B and Table 1, the DLS analysis revealed a tri-modal size distribution for a concentration of 1, 5, and 25 g/ml and a bi-modal size distribution at 10 g/ml with a polydispersion index of 0.4 Ϯ 0.3, 0.4 Ϯ 0.3, 0.2 Ϯ 0.1, and 0.3 Ϯ 0.1 (n ϭ 6) for 1, 5, 10, and 25 g/ml, respectively. The potential was stable, ranging from Ϫ8.4 Ϯ 0.4 to Ϫ9.5 Ϯ 1.0 mV.

Characterization of AgNP
AgNP 15 Induce Morphological Changes in Human THP-1 Monocyte Cells- Fig. 2 illustrates that, after 24 h of treatment, the cell morphology remained unchanged at a concentration of 1 or 5 g/ml AgNP 15 . However, morphological changes became apparent at 10 g/ml AgNP 15 where numerous vacuoles were observed in the cytosol (panel e). Cell morphology was markedly altered at 25 g/ml AgNP 15 (panel f), an effect that is clearly different from that of the apoptosis-inducing agent staurosporine, suggesting a different type of cell death.
AgNP 15 Are Internalized in THP-1 Cells-We next investigated potential internalization of AgNP 15 because we previously reported that silver nanoparticles of 20 nm (AgNP 20 ) could be found inside human neutrophils, another cell type of myeloid origin (21). As illustrated in Fig. 3, incubation of AgNP 15 with THP-1 cells leads to internalization of nanoparticles in the cytosol and nucleus as determined by transmission electron microscopy (Fig. 3). Larger aggregates of AgNP 15 were visibly taken up by a mechanism requiring the formation of phagosomes/vacuoles, whereas well dispersed or small aggregates were mainly found inside the nucleus. Well dispersed nanoparticles were also found in the cytosol but were not internalized inside vacuoles.
AgNP 15 Induce ER Stress-related Events-A 24-h treatment of cells with AgNP 15 resulted in a dose-dependent dephosphorylation of IRE-1, suggesting that this pathway was not activated (Fig. 4A). Although the PERK pathway was strongly activated with 1-10 g/ml AgNP 15 , it remained similar to basal level with a treatment with 25 g/ml. In addition, low concentrations of AgNP 15 induced synthesis of HSP-70, whereas the higher concentration induced degradation of HSP-90. GRP-78 expression levels remained unchanged in all tested conditions except thapsigargin, the positive control. Although the expression level of ATF-6 remained similar in cells treated with 0 -10 g/ml AgNP 15 , this protein was degraded at 25 g/ml, despite the fact that an equivalent of proteins was loaded (see GAPDH lane). Interestingly, degradation of ATF-6 occurred rapidly after 1 h (Fig. 4B). Degradation of other components such as HSP-70 was also observed, whereas the expression level of total IRE-1 and HSP-90 remained relatively stable after 1 h of stimulation. These results suggest that AgNP 15 induce ER stress-related events in human THP-1 cells.
AgNP 15 Induce a Rapid Cell Death Distinct from Apoptosis-As prolonged ER stress response is usually associated with apoptosis, we next investigated the role of AgNP 15 in THP-1 cell death. After only 1 h, cells treated with 25 g/ml AgNP 15 were annexin-V-and PI-positive, whereas cells treated with the proapoptotic agent staurosporine were negative for both markers (Fig. 5, A and B). However, after 24 h, cells were annexin-V-and PI-positive whether they were treated with AgNP 15 or staurosporine. This suggests that AgNP 15 , at 25 g/ml, induce a cell death that is different from staurosporine-induced apoptosis. To confirm this hypothesis, we next monitored the processing of several important caspases, including the executioner caspase-3 and caspase-7, and the inflammatory caspase-4, A, a sample of the solution stock from the manufacturer was used for characterization by transmission electronic microscopy illustrating that the suspension was relatively homogeneous with the majority of AgNP with a diameter close to 15 nm, but also that other particles had different ratios between smaller and larger diameters. B, representative data obtained by DLS analysis performed at 37°C with a concentration of 1, 5, 10, or 25 g/ml AgNP 15 in RPMI 1640 supplemented with 10% FBS, the experimental conditions used in this study. Data are from one representative experiment out of three.

Silver Nanoparticles in ER Stress and Inflammasome Activation
known to be associated with ER stress-induced apoptosis (27). As illustrated in Fig. 5C, staurosporine induced processing of caspase-3, caspase-4, and caspase-7, but not AgNP 15 at 1, 5, or 10 g/ml. At 25 g/ml, AgNP 15 were found to induce the processing of caspase-4 and caspase-7 but not caspase-3. We also observed a variation in the expression of level of GAPDH in staurosporine-treated cells, a phenomenon previously observed at the mRNA level in neutrophils (28). However, membrane coloration revealed an equivalent amount of proteins (data not shown). The activity of caspase-3 and caspase-4 was also monitored using a caspase assay. Only the positive control staurosporine induced activation of caspase-3, whereas AgNP 15 strongly  activated caspase-4 at 25 g/ml (Fig. 5, D and E). These results indicate that AgNP-induced cell death is distinct from apoptosis. AgNP 15 Induce Pyroptosis and Activation of the Inflammasome-Knowing that activation of caspase-4 and caspase-7 is associated with activation of the inflammasome (13,29,30), in addition to the loss of membrane integrity observed after PI staining (this study), we next investigated the role of AgNP 15 in the activation of inflammasomes. Because pyroptosis, another type of cell death, is dependent on caspase-1, we also followed its processing as well as its activa-  tion and other components of the NLRP-3 inflammasome. AgNP 15 treatment at 25 g/ml resulted in the presence of the p20 caspase-1 fragment. (Fig. 6A). Such caspase-1 processing was also observed at 25 g/ml AgNP 15 when cells were pretreated with LPS, suggesting that the effect is due to AgNP alone independently of the LPS pre-treatment. Of note, in the absence of LPS, pro-IL-1␤ was undetectable, but when cells were pretreated with LPS, a strong expression of pro-IL-1␤ was observed. ASC expression remained relatively constant whether or not cells were pre-treated with LPS. Because AgNP 15 induced processing of caspase-1, suggesting that it induces its activity, we next verified whether AgNP 15 could activate caspase-1, using a specific activity assay. As shown in Fig. 6B, the caspase-1 assay revealed that 25 g/ml AgNP 15 significantly increased caspase-1 activity. Also, the activity of caspase-1 was found to correlate with the secretion of IL-1␤ in the extracellular milieu where 25 g/ml AgNP 15 significantly increased the IL-␤ secretion from 9.5 Ϯ 2.8 pg/ml to 298.2 Ϯ 52.7 pg/ml (Fig. 6C). Of note, secretion of mature IL-1␤ required priming with LPS because unstimulated cells failed to secrete this cytokine (see Fig. 11B). We next investigated the importance of AgNP 15 internalization for its effect on IL-1␤ secretion. Therefore, cells were pre-treated or not with the endocytosis dynamin inhibitor (Dynasore) or with the actin depolarizing agent (cytochalasin D) before incubation with AgNP 15 . Cytochalasin D reduced slightly the secretion of IL-1␤ induced by AgNP 15 , whereas dynasore has no effect on its secretion (Fig. 6D). These results suggested that internalization of AgNP 15 might be dispensable for their effect on secretion of IL-1␤.
Formation of pyroptosome is a well documented feature of pyroptotic cell death, characterized by aggregation of ASC proteins into normally one aggregate as observed in murine macrophages or in human monocyte THP-1 cells (13,14). Therefore, we have followed the formation of this structure by confocal microscopy using an ASC specific antibody. As illustrated in Fig. 7A, AgNP 15 induced formation of a pyroptosome structure (Fig. 7A, arrows) similar to that of ATP (positive control) (31), whereas the localization of ASC protein remained diffuse in control cells. The number of cells expressing pyroptosome structures was also quantified in response to 25 g/ml AgNP 15 where close to 20% of cells showed the presence of ASC pyroptosome in the experimental conditions tested (Fig. 7B).
AgNP 15 Induce Activation of the NRLP-3 Inflammasome-Several inflammasomes are found in monocytes including NLRP-3, the most common and best characterized inflammasome (32). To determine whether AgNP 15 induce specific activation of the NLRP-3 inflammasome, siRNA targeting NLRP-3 were transfected in THP-1 cells. NLRP-3 expression levels were then followed by Western blot to determine the efficiency of transfection. A marked decreased of NLRP-3 protein expression was observed after a transfection period of 48 h, but not 24 h (Fig. 8, A and B). Therefore, we next selected this 48-h time point for further experiments. As illustrated in Fig.  8C, a significant decrease in IL-1␤ secretion was observed in cells transfected with silencing NLRP-3 siRNA, confirming that AgNP 15 activate the NLRP-3 inflammasome. Activation of the NLRP-3 inflammasome is dependent on caspase-1 (33). We have therefore evaluated the importance of caspase-1 in the activation of the NLRP-3 inflammasome by AgNP 15 . Caspase-1 was found to be crucial for activation of this inflammasome because its inhibition with YVAD-CHO completely abolished the secretion of IL-1␤, thereby confirming the importance of caspase-1 in this process (Fig. 8D).
Caspase-4 Regulates the Activation of the NRLP-3 Inflammasome-Caspase-4 is known to be required for the activation of the inflammasome through interaction with caspase-1 (30). Because AgNP 15 induced activation of the NLRP-3 inflammasome and also processing of caspase-4, we tested whether caspase-4 was involved in this process. To do so, we used caspase-4-deficient THP-1 cells (22,34) to characterize the role of this caspase in AgNP 15 -induced pyroptosis. The absence of caspase-4 protein expression was confirmed in the caspase-4deficient THP-1 cell clone TB in contrast to the clone vector control where caspase-4 was strongly expressed (Fig. 9A). Western blot experiments indicate that NLRP-3 and ASC protein levels were relatively stable in all tested conditions (Fig.  9B). However, the protein expression of pro-IL-1␤ was undetectable in caspase-4-deficient cells even if primed with LPS, a known NLRP-3 inflammasome primer (35). AgNP 15 were found to induce cleavage of caspase-1 and appearance of the p20 caspase-1 fragment in both control and caspase-4-deficient cells primed with LPS. Despite the fact that AgNP 15 induced an increase caspase-1 activity in caspase-4-deficient cells (Fig. 9C), the IL-1␤ secretion was decreased in caspase-4-deficient cells versus the corresponding control (Fig. 9D). In addition, a significantly higher proportion of PI-positive cells was observed in caspase-4-deficient cells (Fig. 9E), classically used to identify pyroptotic cells (reviewed in Ref. 33). Finally, caspase-4-deficient cells had an increased amount of pyroptosome when compared with control cells (Fig. 9F). These results suggested that caspase-4 is essential for priming of the NLRP-3 inflammasome component pro-IL-1␤, but would negatively regulate its activation by AgNP 15 .
ATF-6 Degradation Leads to Inflammasome Activation and Pyroptotic Cell Death-Because AgNP 15 induce ER stress events accompanied by a rapid degradation of the ATF-6 sen-sor, we next determined whether or not this pathway could regulate the activation of the NLRP-3 inflammasome. To do so, we used specific inhibitors of S1P and S2P, two proteases involved in ATF-6 cleavage (7). We first evaluated the efficiency of the two inhibitors of PF-429242 (S1P inhibitor) and 1,10phenanthroline (S2P inhibitor) on processing of one of their main substrates, SREBP-1. 1,10-Phenanthroline strongly inhibited the processing of SREBP-1, whereas PF-429242 did not appear to block its degradation. Of note, AgNP 15 alone reduced its processing. Pre-treatment of the cells with PF-429242 reduced moderately AgNP 15 -induced ATF-6 processing into its active fragment, but the use of 1,10-phenanthroline was found to drastically prevent its processing and appearance of its cleaved fragment (Fig. 10A). The effect of AgNP 15 was similar to the ER stressor tunicamycin, which alone induced processing of ATF-6, but after 6 h of stimulation (Fig. 10B). In addition, ATF-6 processing correlated with increased caspase-1 processing and activity (Fig. 10C) as well as with IL-1␤ secretion (Fig. 10D) and the percentage of PI-positive cells (Fig. 10E), suggesting an important role for ATF-6 in the control of pyroptosis. Importantly, treatment with the S2P inhibitor led to a reduction in caspase-1 activity, IL-1␤ secretion, and the percentage of PI-positive cells. These results suggest a possible link between ATF-6 degradation and induction of pyroptotic cell death.
AgNP 15 Induce Activation of the Inflammasome in Primary Human Monocytes and Macrophages-We then wanted to compare the effect of AgNP 15 on activation of the NLRP-3 inflammasome in primary human monocytes and macrophage cells. Expression of pro-IL-1␤ and NLRP-3 was clearly different in primary monocytes and macrophages when compared with THP-1 cells. Indeed, expression of pro-IL-1␤ was undetectable in THP-1 cells and macrophages, whereas we observed a basal level in monocytes. Moreover, NLRP-3 expression in THP-1 cells remained similar even after 6 h of priming, whereas its expression in monocytes and macrophages increased over time after 3-6 h of stimulation with LPS (Fig. 11A). Secretion of mature IL-1␤ was also quantified in these three cell types in response to AgNP 15 . Priming with LPS alone was sufficient to induce the secretion of IL-1␤ in THP-1 and macrophages (Fig.  11B). Conversely, treatment with AgNP 15 alone, without LPS FIGURE 7. AgNP 15 induce pyroptosome assembly in human monocyte THP-1 cells. Cells were stimulated with 25 g/ml AgNP 15 , 2 mM ATP, or buffer (Ctrl) for 1 h. A, cells were then fixed, and formation of pyroptosome was followed by confocal microscopy, as described under "Experimental Procedures." DIC, differential interference contrast. B, quantification of percentage of cells with pyroptosome assembly. Data are from one representative experiment out of three (A) or are expressed as means Ϯ S.E. of three independent experiments (B). Differences were considered statistically significant as follows: ***, p Յ 0.005 versus control or appropriate diluent. FEBRUARY 27, 2015 • VOLUME 290 • NUMBER 9

Silver Nanoparticles in ER Stress and Inflammasome Activation
priming, led to a significant production of IL-1␤ in primary monocytes, suggesting that these cells respond differently than THP-1 and macrophages. However in these three cell types, priming with LPS followed by stimulation with AgNP 15 induced the strongest response.

DISCUSSION
AgNP are gaining increasing attention as the utilization of these NP is becoming more and more important in a plethora of medical products (36) and, therefore, increasing potential human exposure, raising important toxicological considerations. In this study, we provide new evidences for a potential cytotoxic role of AgNP 15 , as demonstrated by activation of ER stress events accompanied by the activation of the NLRP-3 inflammasome and pyroptosis. Different types of nanoparticles were shown to induce ER stress response, including zinc oxide (10). However, only a few studies reported that AgNP can induce ER stress events in zebrafish and in human Chang liver cells (11,12).
Accumulation of unfolded proteins in the ER leads to transition of ATF-6, where it is processed by S1P and S2P (7). The results of the present study indicate that 25 g/ml AgNP 15 induce rapid processing of ATF-6 (within 1 h) in human monocytes, an effect not observed at lower concentrations. Interestingly, activation of the NLRP-3 inflammasome was also induced only in response to stimulation with 25 g/ml AgNP 15 , suggesting that ATF-6 could act as a molecular switch to induce the activation of this inflammasome. Indeed, inhibition of S2P (and S1P, to a lesser extent) blocked not only ATF-6 processing, but also activation of the inflammasome. Our results are in agreement with others reporting that inhibition of S2P was more efficient than S1P to block processing of ATF-6 (38,39). Interestingly, a recent study demonstrated that the classical ER stressor, tunicamycin, is sufficient for the induction of IL-1␤, suggesting a possible link between ER stress and the inflammasome (40). Moreover, tunicamycin was previously shown to induce ATF-6 degradation (7). Another study also demonstrates that inhibition of ER stress by sodium 4-phenylbutyrate results in a significant reduction in IL-1␤ levels, supporting the role of ER stress in IL-1␤ production (41). However, to the best of our knowledge, the role of ATF-6 in the activation of the NLRP-3 inflammasome had never been investigated before.
The inflammasomes are multiprotein complexes activated upon cytosolic perturbations such as cellular infection (17,32). This complex mediates the activation of the inflammatory caspase-1, which is critical for the secretion of mature IL-1␤ and formation of pores at the cell membrane (42). Multiple inflammasomes have been characterized including NLRC-4, AIM-2, and NLRP-3. Among them, NLRP-3 is the one that has been more studied and is known to be activated upon various types of stimuli, including asbestos, silica, uric acid, and ATP (43)(44)(45). Conversely to the NLRC-4 or the AIM-2, the NLRP-3 inflammasome is controlled by a priming step that requires de novo protein translation (35). Although carbon nanotubes and carbon black have also been reported to activate this inflam- masome in THP-1 and murine bone marrow-derived cells (20,46), the effect of AgNP on the inflammasome has never been reported. Herein, we show a cytotoxic and inflammatory role of AgNP 15 as evidenced by induction of pyroptotic cell death and IL-1␤ secretion. Indeed, AgNP 15 -treated cells harbored the characteristics of pyroptotic cell death: increased processing and activity of caspase-1, pyroptosome formation, loss of membrane integrity, and secretion of IL-1␤. The experiments on THP-1, primary monocytes, and macrophages also provide new mechanisms of action of the regulation of the NLRP-3 inflammasome. For instance, the basal expression level of the different components of this pathway were found to be slightly different in THP-1, primary monocytes, and macrophages. This resulted in a significant variation in the amount of mature IL-1␤ secreted within the extracellular space. Indeed, AgNP 15treated primary monocytes secreted significant amount of IL-1␤ without the priming of LPS. Of note, we could detect a significant expression level of NLRP-3 and pro-IL-1␤ in resting monocytes, possibly explaining the secretion of IL-1␤ in AgNP 15 -treated cells.
Secretion of IL-1␤ is associated with acute inflammation and pathogenesis of multiple chronic inflammatory diseases (16), fitting well with previous studies indicating that nanosilver particles possess some toxic and proinflammatory effects (47)(48)(49)(50). However, in a subacute murine inhalation model, some authors reported that AgNP induce minimal lung toxicity or inflammation (51). These observations explain why some authors are interested in trying to answer whether AgNP are allies or adversaries (52). Therefore, the results of the present study are in line with the need to further study the mechanisms of the action of AgNP on mammalian cells at the molecular level, to assure a safe application of AgNP.
Although we provide new evidences in the toxicity of AgNP, the exact mechanisms underlying the effects of AgNP 15 still remain to be clarified. Even if we clearly observed AgNP 15 inside cells, particularly close to or in the nucleus, we cannot definitely conclude that AgNP have to penetrate cells to exert biological effects. Based on the internalization inhibition experiments, we may speculate that internalization of AgNP 15 might be dispensable for its effects on IL-1␤ secretion. Moreover, the effect of AgNP 15 was clearly dependent on caspase-1 as inhibition of this protein resulted in abolition of IL-1␤ secretion. Furthermore, although we observed ATF-6 degradation in response to AgNP 15 , it cannot be ruled out that silver ions are somewhat involved. However, it is important to specify that we have used a commercial source of AgNP, and we have used them as is, as they are probably used in different applications. Our results also indicate that the processing of ATF-6 and the appearance of the cleaved fragment correlate with the increased expression of caspase-1 p20, caspase-1 activity, IL-1␤ secretion, and the percentage of PI-positive cells, supporting inflammasome activation. Through inhibition of ATF-6 processing, we clearly see a reduction in caspase-1 activity, IL-1␤ secretion, and the percentage of PI-positive cells, suggesting that ATF-6 degradation is an event related to activation of the NLRP-3 inflammasome and pyroptotic cell death. It is noteworthy that this effect was clearly different from apoptosis cell death.
Caspase-4 belongs to the human inflammatory caspase family along with caspase-1, caspase-5, and caspase-12 (53). The involvement of caspase-4 in ER stress has been previously established (25,27) but has also been recently challenged (34,  B) for Western blot experiments, lysed in caspase-1 (Casp-1) lysis buffer for caspase-1 assay (C), or stained with PI for flow cytometry analyses (E). For IL-1␤ quantification, cells were primed for 4 h with 100 ng/ml LPS before incubation with inhibitors and stimulation with the indicated agonists. D, supernatants were harvested to perform IL-1␤ ELISA. Data are from one representative experiment out of three (A and B) or are expressed as means Ϯ S.E. of three independent experiments(C-E). Differences were considered statistically significant as follows: *, p Յ 0.05 and ***, p Յ 0.005 versus control or appropriate diluent. 54). Nevertheless, to date, only a few studies reported the potential involvement of caspase-4 in the activation of the inflammasomes (30,55). Our results indicate that caspase-4 is indeed required for the priming of the inflammasome, as the absence of this caspase resulted in a severe impairment of pro-IL-1␤ synthesis. However, the absence of caspase-4 also resulted in a significant increase in the activity of active caspase-1 and the proportion of PI-positive and pyroptosomepositive cells. These results suggest that caspase-4 would therefore regulate negatively the NLRP-3 inflammasome activation but would play an important role in its priming step. It is noteworthy that, in caspase-4-deficient THP-1 cells, the NF-B pathway is markedly affected because caspase-4 was previously found to interact with TRAF-6, agreeing with the lack of pro-IL-1␤ in these cells (22). In another study, IL-1␤ production was found to be dependent on NF-B in THP-1 cells (37). Together with the fact that priming of the inflammasomes is dependent of NF-B activation and that caspase-4-deficient cells have reduced NF-B transduction, this information is in agreement with our observation on the impairment of pro-IL-1␤ in caspase-4-deficient cells. Because we made our observations at a concentration of 25 g/ml AgNP 15 , and not at 1-10 g/ml, it would be of interest in the future to establish whether the same would be true for AgNP with a different starting diameter because this variable is highly important in the induction of different biological effects previously reported for cell toxicity of AgNP in bacteria, yeast, algae, crustaceans, and mammalian cells in vitro (26). Because of the great potential and interest that AgNP represent for developing future therapies based on drug delivery, further studies need to be conducted to limit undesired potential effects of AgNP.