TRAIL (CD253) Sensitizes Human Airway Epithelial Cells to Toxin-Induced Cell Death

Ricin toxin is a biothreat agent that is particularly damaging to lung tissue following inhalation. A hallmark of ricin exposure is widespread inflammation and concomitant destruction of the airway epithelium. In this study, we investigated the possible interaction between ricin and known proinflammatory cytokines associated with lung tissue. Using an established human airway epithelial cell line, we demonstrate that epithelial cell killing by ricin is significantly enhanced in the presence of the proinflammatory cytokine known as TRAIL (CD253). Moreover, epithelial cells that are simultaneously exposed to ricin and TRAIL produced large amounts of secondary proinflammatory signals, including IL-6, which in the context of the lung would be expected to exacerbate toxin-induced tissue damage. Our results suggest that therapies designed to neutralize proinflammatory cytokines such as TRAIL and IL-6 may limit the bystander damage associated with ricin exposure.

Ricin itself is a potent inducer of apoptosis (13). The toxin is derived from castor beans (Ricinus communis), where it accumulates in storage vesicles as a mature, 65-kDa glycosylated protein (14)(15)(16). Ricin's two subunits, RTA and RTB, are joined by a single disulfide bond. RTB is a galactose-and N-acetylgalactosamine (Gal/GalNAc)-specific lectin that promotes ricin attachment to cell surface glycoproteins and glycolipids and facilitates ricin's retrograde transport to the endoplasmic reticulum (ER). RTA is an RNA N-glycosidase (EC 3.2.2.22) that catalyzes the hydrolysis of a conserved adenine residue within the sarcin/ricin loop (SRL) of 28S rRNA (13,17,18). In the ER, RTA is liberated from RTB and is retrotranslocated via the Sec61 complex into the cytoplasm, where it inactivates ribosomes with great efficiency (17)(18)(19). Programmed cell death in alveolar macrophages and primary human bronchial epithelial cells occurs via caspase-3-dependent mechanisms, although the exact upstream signals (e.g., ribosome inactivation, ribotoxic stress, and mitogen-activated protein kinase [MAPK] and NFb signaling) are yet to be elucidated.
The goal of the current study was to better define the response of airway epithelial cells to the effects of ricin, especially in light of recent quantitative analyses of ribosomal depurination status that indicated that the pulmonary epithelial cells are disproportionately affected by ricin following intranasal challenge (11). Specifically, we reasoned that cells already compromised by ricin would be hypersensitive to secondary insults such as those represented by proinflammatory cytokines which are known to accumulate in the bronchoalveolar lavage (BAL) fluid of animals following a ricin inhalation, especially the early response cytokines interleukin-1 (IL-1) and tumor necrosis alpha (TNF-␣) (2,8,12,(20)(21)(22)(23)(24). TNF-␣ and related cytokines such as TRAIL (tumor necrosis factor-related apoptosis-inducing ligand; CD253) are suspected to have a role in driving lung epithelial cell death considering their established capacities to trigger extrinsic apoptotic cell death in different cell types experiencing intracellular stress from another insult (25).

TRAIL sensitizes Calu-3 cells to ricin toxin-induced death.
Ricin is a promiscuous toxin capable of killing virtually all mammalian cell types. In most cases, Ͼ50% cell death occurs within 12 to 48 h of exposure of cells to nanogram amounts of ricin (26). We chose Calu-3 cells as a model to understand the response of human airway epithelial cells to ricin toxin. Calu-3 cells are a lung adenocarcinoma line widely accepted as a model to study drug and nanomaterial interactions with pulmonary epithelium (27)(28)(29)(30)(31). Calu-3 cells have also been used as a model to assess transcytosis of botulinum toxin (32) and mucosal IgA (33), as well as infection by respiratory pathogens (34).
To assess the sensitivity of Calu-3 cells to ricin toxin, nearly confluent Calu-3 cells grown in microtiter plates were treated with a range of ricin toxin doses (0.08 to 10 g/ml) and then assessed for viability 24 to 72 h later. We found that Calu-3 cell viability was largely unaffected by ricin, as determined using an ATP-based cell viability assay ( Fig. 1A; see also Fig. S1 in the supplemental material). However, under these conditions, protein synthesis was inhibited to a level comparable to that achieved with cycloheximide treatment, indicating that the cells were indeed affected by exposure to ricin but that cell viability remained high (Fig. 2). Calu-3 cell monolayers grown on Transwell filters were similarly insensitive to killing by ricin, although transepithelial resistance (TER) declined significantly following ricin treatment, indicating that the toxin impacted epithelial integrity (Fig. S2). To determine whether lung epithelial cell lines are generally more resistant to ricin-induced cell death than other cell types, we performed parallel studies with A549 cells, another established human lung epithelial cell line. The viability profile of A549 cells exposed to ricin resembled that of Calu-3 cells (Fig. S3) (35). The issue was not related to the potency of ricin, as exactly the same lot of toxin was used in numerous Vero cell studies and mouse lethal challenge studies, as reported in separate manuscripts (36,37). Thus, we conclude that two well-established human lung epithelial cell lines are hyposensitive to ricin-induced cell death.
Lung epithelial cell death in vivo in response to infectious agents such as influenza virus is typically driven by local inflammatory responses. Therefore, we reasoned that proinflammatory cytokines such as TNF-␣, which is known to be released by alveolar macrophages in response to ricin, might sensitize Calu-3 cells to toxin-induced cell The sensitizing effect of TRAIL on ricin-induced cell death in Calu-3 cells. (A) TNF-ɑ or TRAIL (starting at 1 g/ml) in a 10-fold serial dilution was mixed with ricin (1 g/ml) and then administered to the cells seeded in 96-well plates for 24 h. The cells were then washed, and cell viability was measured 72 h later, as described in Materials and Methods. (B) In the dose experiment, cell viability was assessed at 72 h after exposure of the cells to the indicated concentrations of ricin and TRAIL. (C) In the time course experiments, cell viability was assessed 24 h, 48 h, or 72 h after the cells were exposed to ricin (0.25 g/ml) and TRAIL (0.1 g/ml). All treatments were performed in triplicate and repeated 3 times. Viability of 100% was defined as the average value obtained from wells in which cells had been treated with medium only. death (8). TRAIL (Apo2L; CD253) is another proinflammatory cytokine of interest, considering its role in accelerating lung epithelial cell death under conditions of ARDS (38,39). We therefore examined the viability of Calu-3 cells following treatment with a fixed amount of ricin (1 g/ml) plus recombinant TNF-␣ or TRAIL at doses ranging from 0.01 ng/ml to 1,000 ng/ml (Fig. 1A). Calu-3 cells treated with 1 g/ml of ricin alone displayed ϳ20% cell death at the 72-h time point. The addition of 10 ng/ml of TNF-␣ resulted in slight increase in cell killing (50% to 60% cell death), although further increasing the amount of the TNF-␣ did not further exacerbate ricin's cytotoxic activity. TRAIL was significantly more potent than TNF-␣ in promoting the enhancement of ricin-induced Calu-3 cell death. We observed virtually 100% cell death when cells were treated with ricin and Ն100 ng/ml TRAIL. Neither TNF-␣ (data not shown) nor TRAIL ( Fig. 2) by itself had an effect on protein synthesis in Calu-3 cells. To see if TRAIL had a similar effect on other cell epithelial airway cells, we treated A549 cells with ricin or TRAIL or the combination of ricin and TRAIL (Fig. S3). Cell viability was greatly reduced when the cells were treated with the combination of ricin and TRAIL (ϳ90% cell death), compared to control (untreated) cells. TRAIL alone did not impact cell viability, whereas ricin (10 ng/ml) induced ϳ25% cell death. Thus, TRAIL sensitizes lung epithelial cells to the effects of ricin.
To better define the observed synergy between ricin and TRAIL, we performed checkerboard analysis across a range of concentrations of ricin (0 to 1 g/ml) and TRAIL (0 to 1 g/ml) (Fig. 1B). This analysis identified the minimal doses of ricin (250 ng/ml) and TRAIL (100 ng/ml) required to achieve ϳ100% cell death within a 72-h period. A time course of Calu-3 cell viability in response to ricin (250 ng/ml) or TRAIL (100 ng/ml) or the combination of ricin and TRAIL is shown in Fig. 1C. The viability of ricin-treated Calu-3 cells declined only marginally within a 72-h period, while the viability of the TRAIL-treated cells was largely unchanged in that same time frame. In contrast, the cell viability of Calu-3 cells treated with the combination of ricin and TRAIL declined in a stepwise manner at 24 h, 48 h, and 72 h to Ͻ25%. The observed effects of ricin and TRAIL on Calu-3 cell death and protein synthesis arrest were annulled by anti-TRAIL antibodies (Abs) (Fig. 3A) or neutralizing monoclonal antibodies (MAbs) against ricin's FIG 2 Protein synthesis in Calu-3 cells following treatments with ricin and/or TRAIL. Calu-3 cells were plated at a density of 5 ϫ 10 4 cells/well in a 96-well clear-bottom black plate. After confluence was achieved, cells were treated with ricin (0.25 g/ml) or TRAIL (0.1 g/ml) or a mixture of ricin and TRAIL with or without anti-ricin MAb PB10 (15 g/ml) for 24 h. As a control, cells were treated with cycloheximide (250 g/ml). Cells were then incubated for 2 h in culture medium alone, culture medium containing OPP, or culture media containing OPP and the treatments as described above. The cells were then processed for detection of protein synthesis, as described in Materials and Methods. The results (means Ϯ standard deviations [SD]) represent a single experiment done in triplicate and repeated twice. ***, P Ͻ 0.001; ****, P Ͻ 0.0001 (versus only OPP-treated cells).
A subunit (Fig. 3B) or B subunit (Fig. S4). The 50% inhibitory concentrations (IC 50 s) for each of the RTA and RTB neutralizing MAbs in conjunction with the Calu-3 cells were similar to what has been reported for other cell types such as Vero cells and THP-1 cells (36,40). For example, MH3's IC 50 for Calu-3 cells was ϳ 1 g/ml, which is similar to what we reported for MH3 on Vero cells (1.6 g/ml) (36).
Transcriptional profiling of Calu-3 cells following ricin and TRAIL treatment. To better understand the interaction between ricin and TRAIL, we subjected Calu-3 cells to transcriptional profiling using a human immunology nCounter array encompassing ϳ600 target genes. RNA was isolated from Calu-3 cells treated with ricin (250 ng/ml) or TRAIL (100 ng/ml) or the combination of ricin and TRAIL for 3 h, 6 h and 18 h. At the 3-h and 6-h time points, there were no significant changes in RNA levels seen among the target genes represented on the human immunology array in comparisons of samples with ricin or TRAIL or ricin-plus-TRAIL treatments to medium control samples (data not shown). By 18 h, the picture was markedly different. Analysis of the RNA from Calu-3 cells treated with the combination of ricin and TRAIL indicated that there were ϳ80 genes whose expression was elevated Ͼ2-fold over the levels seen with the untreated controls, which corresponds to roughly 12% of all the genes on the human immunology nCounter array ( Fig. 6; see also Table S1 and Fig. S6 in the supplemental material). Most notable was an increase in the level of IL-6 (ϳ750-fold), followed by other proinflammatory cytokines such as beta interferon (IFN-␤) (ϳ120-fold), TNF-␣ (ϳ120fold), IL-8 (88-fold), IL-1␣ (60-fold), and CCL20 (90-fold). Virtually the same transcriptional profile was observed when Calu-3 cells were treated with just ricin, although the magnitude of the response was dampened compared to that seen with cells treated with ricin plus TRAIL ( Fig. 6; see also Table S1 and Fig. S6). TRAIL treatment alone did not significantly alter Calu-3 gene expression. These results are consistent with TRAIL enhancing the proinflammatory and apoptotic responses of Calu-3 to ricin rather than inducing parallel or convergent proinflammatory and apoptotic pathways. To validate the transcriptional profiling studies, Calu-3 cells were treated for 24 h with ricin or TRAIL or the ricin-plus-TRAIL combination, after which culture supernatants were assayed for levels of proinflammatory cytokines IL-6, IL-8, IL-1, IL-10, TNF-␣, and IL-12 by cytometric bead array (CBA). We found that the IL-8, IL-1, IL-10, TNF-␣, and IL-12 levels were unchanged, irrespective of whether Calu-3 cells were treated with ricin or TRAIL or the ricin-plus-TRAIL combination (Fig. 7). IL-6 levels, in contrast, were elevated Ͼ10-fold in supernatants from Calu-3 cells treated with ricin plus TRAIL, compared to controls (Fig. 7). Treatment of cells with ricin alone enhanced IL-6 levels, although not to a statistically significant degree. Thus, treatment of Calu-3 cells with the combination of ricin plus TRAIL results in the preferential secretion of IL-6.
We postulated that the absence of TNF-␣ in the Calu-3 cell supernatants following ricin and TRAIL treatments might have been due to autocrine signaling such that soluble cytokine was rapidly captured by the TNF-␣ receptor, which in turn might have contributed to ricin-induced cell death. To examine this possibility, Calu-3 cells were treated with ricin plus TRAIL in the presence of neutralizing anti-TNF-␣ antibody. We found that anti-TNF-␣ antibody treatment did not prevent or even reduce Calu-3 cell killing in response to ricin plus TRAIL, whereas treatment with an anti-TRAIL neutralizing antibody did rescue the cells (Fig. 8).

DISCUSSION
Widespread damage to the airway epithelium is a hallmark of inhalational ricin exposure, although the exact molecular events that culminate in epithelial cell destruction have not been fully elucidated (2,3,5,7,24,41). For this study, we utilized the well-characterized Calu-3 cell line as a prototype to better define the response of human airway epithelial cells to ricin (27)(28)(29)(30)(31). We found that Calu-3 cells, when grown to confluence on solid or permeable substrates, were largely impervious to the effects of ricin-induced cell death (as determined using an ATP-based cell viability assay), even though protein synthesis was effectively shut down (as determined using an O-propargyl-puromycin [OPP]-based assay) and epithelial integrity compromised (as manifested by a reduction in TER). Coadministration of soluble TRAIL (and, to a lesser degree, TNF-␣) rendered Calu-3 cells Ͼ1,000-fold more sensitive to toxin-induced apoptosis. Soluble TRAIL also magnified the proinflammatory transcriptional response of Calu-3 cells to ricin and the preferential secretion of IL-6. While the current investigation was limited to in vitro studies, the results are consistent with a model in which proinflammatory cytokines such as TRAIL amplify epithelial stress-induced signal transduction that promotes the recruitment of polymorphonuclear leukocytes (PMNs) to the lung and simultaneously lowers the threshold level of ricin required to induce epithelial apoptosis. We also present evidence to support the notion that ricin alters bronchial epithelial integrity and barrier function, thereby contributing to enhanced transudation into the alveolar space (42).
TRAIL has previously been implicated in driving respiratory pathology and airway epithelial cell death in mice and humans in response to pathogenic agents, notably,  Table S1. influenza virus, respiratory syncytial virus (RSV), and chlamydia (38,39,(43)(44)(45). In the case of influenza virus infection, alveolar macrophages are the primary source of TRAIL (38,39). Bronchial epithelial cells express a TRAIL receptor(s), which ultimately modulates TRAIL-dependent apoptosis (46). In rodents and nonhuman primates, alveolar macrophages are a primary target of ricin following intranasal and inhalational challenge, so it is plausible that these cells also serve as a source of TRAIL following toxin exposure (3,5,9,47). It remains unclear whether this scenario applies to ricin. In preliminary studies, we have not yet observed a measurable uptick in TRAIL levels in the BAL fluids from mice or nonhuman primates that had been challenged with ricin (Y. Rong and N. Mantis, unpublished results). However, we have observed that supernatants from toxin-treated monocyte/macrophage cell lines do sensitize human lung epithelial cells to the effects of ricin, and we are actively examining whether this factor is TRAIL or a member of the TNF-␣ family of cytokines (C. Kempen, Y. Rong, N. Mantis, and T. LaRocca, unpublished results).
The issue of how TRAIL sensitizes Calu-3 cells to ricin-induced cell death is of particular interest, considering that TRAIL activates cell death through an extrinsic apoptotic pathway, whereas ricin triggers intrinsic pathways induced as a result of the ribotoxic stress response (RSR), unfolded protein response (UPR), and/or increased levels of intracellular calcium (13,25). We postulate that caspase-3 serves as the central node through which ricin and TRAIL intersect. In human cells, activation of TRAIL receptor 1 (TRAIL-R1) and/or TRAIL-R2 stimulates caspase-8 activation, which in turn triggers caspase-3 activation (46). Ricin-induced programmed cell death is also dependent on caspase-3 activation, although which specific upstream signaling pathway(s) (e.g., RSR, UPR, ER stress) is most relevant in airway epithelial cell killing has not been completely elucidated (2,12). As demonstrated in this study, Calu-3 cell death following with ricin and TRAIL treatment coincided with an increase in caspase-3/7 activity and was partially inhibited by the addition of Z-DEVD-fmk but was not impacted by LEHD, an inhibitor of caspase-9. Calu-3 cell death may also be exacerbated by a concomitant decline in the levels of endogenous inhibitors of apoptosis such as cFLIP, due to ricin's capacity to arrest protein synthesis, as reported previously in a study of cells treated with Shiga toxin (48). In fact, the protein synthesis inhibitor cycloheximide is commonly used as a tool to sensitize cells to TRAIL-induced apoptosis (25,49). While the issue is somewhat a question of semantics, we would argue that TRAIL sensitizes Calu-3 cells to ricin-induced apoptosis rather than sensitizing ricin-sensitizing Calu-3 cells to TRAILinduced cell death. This claim is best supported by the results of the transcriptional profiling that we performed, which demonstrated that TRAIL amplifies the effects of ricin on the treatment of Calu-3 cells across the board. Treatment with TRAIL alone had no effect on Calu-3 gene expression, nor did TRAIL (by itself) negatively influence Calu-3 cell viability.
The relationship between protein synthesis arrest and apoptosis in response to ricin toxin is another interesting aspect of this study. Using the puromycin analogue OPP coupled with click chemistry to measure levels of protein synthesis (50), we found that ricin treatment induced protein synthesis arrest (i.e., a measurable decrease in ATP levels) in Calu-3 cells without triggering cell death. Only upon the addition of TRAIL did the cells undergo apoptosis. Thus, ricin-treated Calu-3 cells are primed to undergo programmed cell death upon exposure to a secondary insult, whether that insult is represented by TRAIL or possibly by other signals in vivo (25). As shown in Fig. 2 and 3 (see also Fig. S4 in the supplemental material), neutralizing antibodies against ricin or TRAIL protected Calu-3 cells from simultaneous ricin and TRAIL challenge. The relative kinetics of neutralizing activity were not investigated here. In other words, might anti-TRAIL antibodies have activity when administered after treatment of Calu-3 cells with ricin? Might the Calu-3 cell response to TRAIL be blunted with TRAIL receptor (e.g., DR2) antibodies? These are pertinent questions that are currently being addressed in a mouse model, even though the mouse differs considerably from humans with respect to TRAIL receptors (46).
We identified IL-6 as being markedly upregulated in Calu-3 cells at the transcriptional and protein levels following ricin and ricin-plus-TRAIL treatments. This finding may have important implications for understanding the pathology associated with pulmonary ricin exposure, especially in nonhuman primates, where elevated levels of IL-6 in bronchoalveolar lavage (BAL) fluids are associated with negative outcomes following toxin exposure (C. Roy, Y. Rong, D. Ehrbar, and N. Mantis, unpublished results). IL-6 also accumulates in BAL fluids and serum of mice following intranasal ricin challenge (Rong and Mantis, unpublished) (9,22,47,51). DaSilva and colleagues also noted a significant increase (albeit modest) in IL-6 mRNA levels when total lung tissues from ricin-challenged mice were analyzed by DNA microarray (20). Whether IL-6 is more than just a biomarker of ricin intoxication remains to be determined, but there is considerable evidence to implicate this cytokine in driving local and systemic pathologies (52,53). Surprisingly, Calu-3 cells did not secrete detectable levels of other "initiator" cytokines such as TNF-␣ and IL-1 or the chemokine IL-8, even though each of their respective mRNA transcripts was significantly unregulated by ricin or ricin plus TRAIL. Wong et al. noted that TNF-␣ secretion actually declined when primary human bronchial epithelial cells were exposed to ricin, presumably due to a global arrest in protein synthesis (12). The continued (and possibly preferential) synthesis of IL-6 in ricin-intoxicated cells may have to do with differential rates of mRNA stability, especially when mitogen-activated protein kinase (MAPK) signaling pathways are activated (54). It is worth noting that Wong and colleagues reported that TNF-␣ and IL-1␤ expression in primary human bronchial epithelial cells in response to ricin is dependent on NFB, while IL-6 expression is not (12). In summary, our report advances our understanding of the factors that may drive the severe pathology associated with aerosolized ricin exposure and offers support for therapeutic approaches aimed at neutralizing ricin plus immunomodulatory factors such as TRAIL (2,24).
Cell culture. The Calu-3 lung adenocarcinoma cell line was obtained from the American Type Culture Collection (ATCC; Manassas, VA) and cultured in Eagle's minimum essential medium (EMEM) supplemented with 10% fetal bovine serum, provided by the Wadsworth Center Media Services facility. Cells were grown in a humidified incubator containing 5% CO 2 and 95% air at 37°C. The cells were plated in 75-cm 2 cell culture flasks and subcultured at 70% to 90% confluence using a 0.25% trypsin-EDTA solution (Corning Life Sciences, Corning, NY). The culture medium was changed every 3 days. The cells were split 1:5 during each passage. The numbers of passages used for the following experiments were under 10. The A459 human lung epithelial cell line (ATCC CCL-185) was obtained from ATCC and maintained as recommended by ATCC.
Ricin cytotoxicity assay. Calu-3 and A549 cells were trypsinized, adjusted to 5 ϫ 10 5 cells per ml, seeded (100 l/well) into 96-well plates (Corning Life Sciences, Corning, NY), and incubated for 3 to 4 days until confluence. Calu-3 cells were then treated with ricin, TRAIL, a mixture of ricin and TRAIL, or medium alone (negative control) for 24 h. The cells were washed to remove noninternalized toxin or TRAIL and were then incubated for 24 to 72 h. Cell viability was assessed using CellTiter-GLO reagent (Promega, Madison, WI) and a SpectraMax L microplate reader (Molecular Devices, Sunnyvale, CA). All treatments were performed in triplicate and repeated at least 3 times. Viability of 100% was defined as the average value obtained from wells in which cells were treated with medium only.
On the basis of the cytotoxicity results from the ricin and TRAIL treatments described above, ricin (0.25 g/ml) and TRAIL (0.1 g/ml) were used in all the subsequent experiments. To measure the neutralizing activity of ricin-specific MAbs and neutralizing anti-TRAIL Ab, the anti-RTA MAbs (starting at 15 g/ml), anti-RTB MAbs (starting at 30 g/ml), or anti-TRAIL Abs (starting at 1 g/ml) in a 2-fold serial dilution were mixed with ricin and TRAIL and then administered to the cells seeded in 96-well plates for 24 h. After washing and then incubating for 3 days, cell viability was assessed using CellTiter-GLO reagent. Cell viability was normalized to cells treated with medium only. The protective effect of caspase inhibitors (Z-VAD-FMK, Z-LEHD-FMK, Z-DEVD-FMK, and Z-IETD-FMK), RIPK1 inhibitor (Nec-1), RIPK3 inhibitor (GSK=872), and MLKL inhibitor (NSA) was also evaluated using combinations with ricin and TRAIL and the concentrations and incubation times indicated for each experiment. Dimethyl sulfoxide (DMSO) was used as a control vehicle for all experiments.
Protein synthesis assay. Calu-3 cells were plated at a density of 5 ϫ 10 4 cells/well in a 96-well clear-bottom black plate (Corning Life Science, Corning, NY) and incubated for 3 to 4 days until confluence. Cayman's protein synthesis assay kit (Cayman Chemical, Ann Arbor, MI) was used to measure the protein synthesis. Cells were treated with ricin (0.25 g/ml), TRAIL (0.1 g/ml), the mixture of ricin and TRAIL combined with PB10 (15 g/ml), or cycloheximide (250 g/ml; positive control) for 24 h. Cells were then incubated for 2 h in culture medium alone, culture medium containing O-propargylpuromycin (OPP), or culture medium containing OPP plus the treatments described above. After fixing and washing, cells were incubated with 5 FAM-azide staining solution in the dark for 30 min for subsequent detection of OPP-labeled proteins. The cells were then examined immediately using a fluorescent plate reader (Becton, Dickinson, Franklin Lakes, NJ) set to excitation/emission at 485/535 nm.
Caspase-3/7 activity assay. For the quantification of caspase-3/7 activities after treatment, Calu-3 cells were labeled with 500 nM Cell Event caspase-3/7 green detection reagent (Invitrogen, Carlsbad, CA) for 30 min at 37°C in the dark. A total of 10,000 stained cells per sample were acquired and analyzed in a FACSCalibur flow cytometer by using CellQuest Pro software (Becton, Dickinson, Franklin Lakes, NJ). Data were expressed as a percentage of total cells.
Multiplex gene expression analysis using NanoString. Calu-3 cells were treated with ricin, TRAIL, the combination of ricin and TRAIL, or medium alone (negative control) for 24 h. RNA was extracted from treated or nontreated cells using an RNeasy Plus minikit with additional on-column DNase digestion with an RNase-Free DNase set (Qiagen Hilden, Germany). Protocols were followed according to the manufacturer's instructions. Extracted RNA samples were stored at Ϫ80°C until use. Upon assay, RNA integrity was verified by agarose gel electrophoresis. RNA quality and concentrations were measured using an Agilent 2100 Bioanalyzer (Life Technologies, Carlsbad, CA). RNA (100 ng) was hybridized with a predesigned nCounter human immunology panel that included 594 target genes and 15 internal reference genes. The geNorm algorithm was used to select the most stable of the reference genes (GAPDH [glyceraldehyde-3-phosphate dehydrogenase], PPIA, G6PD, EEF1G, GUSB, HPRT1, SDHA, and RPL19) used for normalization (56). The experimental procedures were carried out on a NanoString preparation station and digital analyzer according to the manufacturer's instructions. Two biological replicates were selected from each of the four groups for analysis.
Cell supernatant cytokine quantification by cytometric bead array (CBA). Cell supernatants were collected from treated Calu-3 cells. The BD CBA Human Inflammatory Cytokines kit (Becton, Dickinson, Franklin Lakes, NJ) was used to quantitatively measure levels of the following specific sets of cytokines: interleukin-8 (IL-8), interleukin-1␤ (IL-1␤), interleukin-6 (IL-6), interleukin-10 (IL-10), TNF-␣, and interleukin-12p70 (IL-12p70). Dilution series of human cytokine standards, included in the kit and prepared according to the manufacturer's instructions, were employed in each assay run to enable quantification. Assays were performed according to the manufacturer's instructions, and 50 l of assay beads, 50 l of the studied sample or standard, and 50 l of phycoerythrin (PE)labeled antibodies (Detection Reagent) were added consecutively to each sample tube and incubated at room temperature in the dark for 3 h. Next, the samples were washed and centrifuged, after which the pellet was resuspended in wash buffer and analyzed on the same day in a flow cytometer. Flow cytometry was performed using a four-laser BD FACSCalibur system utilizing BD CellQuest software for acquisition.
Statistical analyses. Statistical analyses were carried out using GraphPad Prism 7 (GraphPad Software, San Diego, CA), as well as the nCounter Advanced analysis module (v 1.1.4) of the nSolver Analysis software (v3). Differential expression of the genes examined was determined by multivariate linear regression, with group membership chosen as the predictor variable and binding density as a confounder. All P values derived from NanoString analysis were adjusted with Benjamini-Yekutieli correction for control of the false-discovery rate. Low-count genes were omitted using the default settings in the nCounter Advanced analysis software for all analyses except the linear regression of gene expression values.

ACKNOWLEDGMENTS
We gratefully acknowledge Renjie Song of the Wadsworth Center's Biochemistry and Immunology Core facility for assisting and for the use of the flow cytometer. We thank Navjot Singh and Susan L. McHale of the Wadsworth Center's Applied Genomics Technologies Core for assistance with the nCounter instrumentation. We also thank Amanda Poon (University at Albany) for helpful discussions.
This work was supported by contract no. HHSN272201400021C from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. The content is solely our responsibility and does not necessarily represent the official views of the National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The nCounter experiments were funded by the Director's Office of at the Wadsworth Center in support of collaborations between the Wadsworth Center and ACPHS.