Results

Airway epithelial cells do not produce significant amounts of IL-1β in response to inflammasome stimulation
The inflammasomes and their respective activators examined in this study are listed in Table 1. Cells were stimulated in accordance with the schedule in Figure 1. In CF, airway epithelial cells have been shown to possess a hyper-inflammatory phenotype and produce an exaggerated pro-inflammatory cytokine response [4], [5]. To determine if airway epithelial cells contribute to the increased IL-1β production in patients with CF, CF and control bronchial epithelial cell lines were stimulated with the inflammasome inducers P. aeruginosa strain PAO1 (PAO1) and LPS followed by ATP. IL-1β levels in cell culture supernatants were not greatly increased in either the CF or control cell lines (Fig. 2a–d), although a small increase in IL-1β production was detected in NuLi-1 and CuFi-1 cells, but not in S9 and IB3-1 cells, by 24 hours. In contrast, these airway cells were highly responsive to other inflammatory stimuli, such as recombinant IL-1β, producing large quantities of IL-8 (Fig. 2a–d inserts).
10.1371/journal.pone.0037689.g001 Figure 1  Cell stimulation and inhibitor schedule.
Schedule outlines the timing of inhibitor addition and priming in relation to inflammasome stimulation (t = 0) for THP-1 reporter and PBMC cytokine quantification experiments. Inhibitor treatments and stimulations were carried out as described in the Materials and Methods section.
10.1371/journal.pone.0037689.g002 Figure 2  Airway epithelial cells do not significantly contribute to IL-1β production in response to inflammasome stimuli.
Control cell lines ((A) S9, (C) NuLi-1) and their corresponding CF cell lines ((B) IB3-1, and (D) CuFi-1) cells were stimulated with P. aeruginosa (MOI = 10), ATP (5 mM), or IL-1β (10 ng/ml), for the indicated times (n = 3 individual experiments). Cells were primed with LPS (100 ng/ml) for 4 hours where appropriate. Cell culture supernatants were assayed for IL-1β and IL-8 production by ELISA. Insert shows IL-8 secretion in response to stimulation with IL-1β (10 ng/ml).
10.1371/journal.pone.0037689.t001 Table 1  Inflammasome Activators.

Airway epithelial cells do not significantly upregulate caspase-1 activity in response to inflammasome stimulation
To examine if inflammasome activation occurs in these airway cells, caspase-1 activity was quantified by flow cytometry. There was no significant increase upon stimulation with live PAO1 or LPS+ATP at the times examined (Fig. 3a–b). Because previous studies have indicated a role for caspase-1 in the activation of NF-κB through Toll-like receptor (TLR) signaling [35], we examined whether chemical inhibition of caspase-1 altered NF-κB-dependent IL-6 production in response to P. aeruginosa. However, treatment with the caspase-1 inhibitor z-YVAD-fmk (YVAD) did not decrease IL-6 secretion by airway epithelial cells (Fig. 3c).
10.1371/journal.pone.0037689.g003 Figure 3  Airway cells do not strongly upregulate caspase-1 activity in response to inflammasome stimuli.
S9 and IB3-1 cells were examined for caspase-1 activation following inflammasome stimulation with P. aeruginosa (MOI = 10) and ATP (5 mM). Cells were primed with LPS for 5 hours where appropriate. A representative histogram of % caspase-1-active cells is shown in (A) and the averaged values are shown in (B) (n = 3 separate experiments). (C) IB3-1 cells (5×104 cells/well) were pre-treated for 1 hour with increasing concentrations of z-YVAD-fmk (10–30 µM) prior to stimulation with P. aeruginosa (MOI = 50). Cell culture supernatants were collected after 6 hours and assayed for IL-6 by ELISA (n = 3).

CD14 positive monocytes from CF patients and controls show similar increases in caspase-1 activity upon inflammasome stimulation
Monocytes were identified in PBMC populations using CD14 as a phenotyping marker. CD14 positive monocytes from CF patients and healthy controls showed a significant increase in caspase-1 activation upon stimulation with LPS+ATP, PAO1, and LPS+Poly(dA:dT) (Fig. 4a) but this activation was not different between CF and control subjects (Fig. 4b).
10.1371/journal.pone.0037689.g004 Figure 4  PBMCs from CF patients and controls show similar increases in caspase-1 activity upon inflammasome activation.
PBMCs from CF patients (n = 6) and healthy controls (n = 6) were primed with LPS (10 ng/ml) for 5 hours prior to stimulation with ATP (5 mM) for 1 hour or Poly(dA:dT) (1 µg/ml) for 3 hours. PBMCs were stimulated with P. aeruginosa strain PAO1 for 3 hours. A representative histogram of the % caspase-1-active cells is shown in (A) with the averaged values shown in (B).

PBMCs from CF patients do not produce increased amounts of IL-1β upon inflammasome stimulation
Previous studies have shown that the loss of CFTR results in increased NF-κB activity and pro-inflammatory cytokine secretion [4], [5], [32], [36], [37]. To further examine this relationship, PBMCs from CF patients and healthy adult controls were stimulated with PAO1, LPS+ATP, and LPS+Poly(dA:dT), to activate the NLRC4, NLRP3, and AIM2 inflammasomes, respectively. By 24 hours of stimulation, CF PBMCs did not produce increased amounts of IL-1β (Fig. 5a) or IL-8 (Fig. 5b) when compared to healthy controls. However, we did notice a transient decrease (P<0.001) in the amount of IL-1β produced by CF cells in response to LPS+ATP at 6 hours (data not shown). Stimulation of PBMCs with P. aeruginosa that lacks exsA (PAO1ΔexsA), a key regulator of type III secretion, produced three-fold less IL-1β compared to the parental PAO1 strain by 24 hours (Fig. 5a). Inflammasome stimulation without priming did not result in any IL-1β production in either CF or control PBMCs. Contrary to our hypothesis, these results indicate that PBMCs from CF patients do not display increased production of IL-1β or IL-8 with inflammasome activation nor do they suggest any increased basal or induced NF-κB activity. These results are consistent with our observation that caspase-1 activity is not different between CF and control PBMCs (Fig. 4).
10.1371/journal.pone.0037689.g005 Figure 5  PBMCs from CF patients do not produce increased amounts of IL-1β.
PBMCs from CF patients (n = 17–20) and healthy controls (n = 15–19) were primed with LPS (10 ng/ml) overnight and stimulated with P. aeruginosa PAO1 (MOI = 1), P. aeruginosa PAO1 lacking exsA (MOI = 1), ATP (5 mM), or Poly(dA:dT) (1 µg/ml) for 24 hours. P. aeruginosa lacking exsA was used as a type III secretion control in comparison with wild-type P. aeruginosa. Supernatants were assayed for (A) IL-1β and (B) IL-8.

NF-κB activation is required for IL-1β and IL-8 responses to P. aeruginosa 
We confirmed the dependence of PAO1-induced IL-1β and IL-8 production on NF-κB activation using THP-1 cells expressing a reporter driven by NF-κB and AP-1 response elements. We found that stimulation of primed THP-1 reporter cells with heat-killed PAO1 produced the highest levels of NF-κB/AP-1 activation (Fig. 6a) and this correlated with IL-8 secretion (Fig. 6b) but negligible amounts of IL-1β were secreted (Fig. 6c). Stimulation of primed THP-1 reporter cells with live PAO1 did not significantly increase NF-κB/AP-1 activity (Fig. 6a) or IL-8 secretion (Fig. 6b) over priming alone. However, IL-1β production was greatly augmented over primed cells stimulated with heat-killed PAO1 or unprimed cells stimulated with live PAO1 (Fig. 6c). This confirmed that NF-κB activation alone is not sufficient for maximal IL-1β secretion, but increased priming of NF-κB is capable of augmenting IL-1β production and secretion upon inflammasome stimulation. Dependency of these responses on NF-κB was confirmed by pharmacologic inhibition of NF-κB using the Bay11-7082 inhibitor of IκBα phosphorylation, which significantly reduced NF-κB/AP-1 activation (P<0.001) (Fig. 6d) and the subsequent production of IL-8 (P<0.01) (Fig. 6e) and IL-1β (P<0.001) (Fig. 6f) in response to both heat-killed and live PAO1. These results were also verified in CF and control PBMCs for each inflammasome examined (P<0.001) (Fig. 6g–h). Overall these results confirm that NF-κB is an important modulator of IL-1β production and that increased activation of NF-κB augments inflammasome-mediated production of IL-1β.
10.1371/journal.pone.0037689.g006 Figure 6  NF-κB activation potentiates the degree of IL-1β production and secretion upon inflammasome activation.
THP-1 reporter cells were primed overnight with heat-killed P. aeruginosa and stimulated the next day with live P. aeruginosa or additional heat-killed P. aeruginosa for the times indicated. Cell culture supernatants were assayed for (A) NF-κB/AP-1 activity, (B) IL-8, and (C) IL-1β secretion (n = 3–6 experiments). Using the same stimulation method, THP-1 reporter cells were treated with Bay11-7082 (20 µM) for 1 hour prior to priming with heat-killed PAO1 or live PAO1. Supernatants were assayed at 24 hours for (D) NF-κB/AP-1 activity, (E) IL-8, and (F) IL-1β secretion (n = 3–5). PBMCs from CF patients (n = 11–15) and controls (n = 10–13) were treated with z-YVAD-fmk (20 µM) or Bay11-7082 (10 µM) and stimulated with live PAO1 (MOI = 1), ATP (5 mM), or Poly(dA:dT) (1 µg/ml) according to the schedule in Figure 1. (G) IL-1β and (H) IL-8 levels were measured at 24 hours. Statistical analysis was performed using two way ANOVA with Bonferroni correction for multiple comparisons. *, **, and *** signify P<0.05, 0.01, and 0.001.

Bay11-7082 inhibits pro-IL-1β production in response to P. aeruginosa 
In addition to inhibition of NF-κB activity, Bay11-7082 can also directly inhibit the NLRP3 inflammasome [38]. To validate its use in this study as an NF-κB inhibitor, western blots for pro-IL-1β were performed alongside inhibition of CFTR activity by CFTRinh172 in response to PAO1 at 4 hours after stimulation (Fig. 7a). Our results indicate that Bay11-7082 prevents production of pro-IL-1β whereas CFTRinh172 does not seem to affect it. This was further corroborated by the ability Bay11-7082 to inhibit IκBα degradation at 0.5, 1, and 1.5 hours post PAO1 stimulation (Fig. 7b).
10.1371/journal.pone.0037689.g007 Figure 7  Bay11-7082 inhibits pro-IL-1β production in response to P. aeruginosa.
PMA-differentiated THP-1 cells were treated with 10 µM CFTRinh172 or 20 µM Bay11-7082 and harvested after (A) 4 hours (n = 3) or (B) 0.5, 1, and 1.5 hours (n = 3) stimulation with PAO1. One representative blot is shown with a graph of the averaged fluorescence intensity values over 3 experiments.

Disruption of CFTR activity does not increase IL-1β production in PBMCs and THP-1 cells
A previous study has indicated a role for chloride ion concentration in suppression of NLRP3 inflammasome activation [34]. To determine whether CFTR dysfunction alters IL-1β production, THP-1 cells and PBMCs from CF patients and healthy controls were treated with the CFTR inhibitor, CFTRinh172, prior to simulation with live P. aeruginosa. Treatment with CFTRinh172 did not alter IL-1β or IL-8 production in control subjects or CF patients (Fig. 8a–b). Similarly, IL-1β production was not different in monocyte-derived macrophages or THP-1 reporter cells treated with CFTRinh172 (Fig. 8c–d). IL-8 (Fig. 8e) and NF-κB activity (Fig. 8f) were also unchanged in CFTRinh172-treated THP-1 reporter cells.
10.1371/journal.pone.0037689.g008 Figure 8  Disruption of CFTR activity does not increase IL-1β production in PBMCs or macrophages.
PBMCs from CF patients (n = 15) and controls (n = 13) were treated with CFTRinh172 (10 µM) for 18 hours prior to stimulation with live PAO1 (MOI = 1). (A) IL-1β and (B) IL-8 production was measured at 24 hours. (C) Monocytes from controls (n = 3) were differentiated into macrophages. Macrophages were treated with CFTRinh172, stimulated as per monocytes, and measured for IL-1β production at 24 hours. THP-1 reporter cells were treated with CFTRinh172 24 hours prior to stimulation with PAO1 and measured for (D) IL-1β secretion, (E) IL-8, and (F) NF-κB/AP-1 activity at 24 hours (n = 4).

Di