CHAC1 Is Differentially Expressed in Normal and Cystic Fibrosis Bronchial Epithelial Cells and Regulates the Inflammatory Response Induced by Pseudomonas aeruginosa

In cystic fibrosis (CF), Pseudomonas aeruginosa (Pa) colonizes the lungs, leading to chronic inflammation of the bronchial epithelium. ChaC glutathione-specific γ-glutamylcyclotransferase 1 (CHAC1) mRNA is differentially expressed in primary human airway epithelial cells from bronchi (hAECBs) from patients with CF and healthy patients at baseline and upon infection with Pa. CHAC1 degrades glutathione and is associated with ER stress and apoptosis pathways. In this study, we examined the roles of CHAC1 in the inflammatory response and apoptosis in lung epithelial cells. First, we confirmed by reverse transcription quantitative polymerase chain reaction that CHAC1 mRNA was overexpressed in hAECBs from patients without CF compared with the expression in hAECBs from patients with CF upon Pa (PAK strain) infection. Moreover, the Pa virulence factors LPS and flagellin were shown to induce CHAC1 expression in cells from patients without CF. Using NCI-H292 lung epithelial cells, we found that LPS-induced CHAC1 mRNA expression was PERK-independent and involved ATF4. Additionally, using CHAC1 small interfering RNA, we showed that reduced CHAC1 expression in the context of LPS and flagellin stimulation was associated with modulation of inflammatory markers and alteration of NF-κB signaling. Finally, we showed that Pa was not able to induce apoptosis in NCI-H292 cells. Our results suggest that CHAC1 is involved in the regulation of inflammation in bronchial cells during Pa infection and may explain the excessive inflammation present in the respiratory tracts of patients with CF.

In cystic fibrosis (CF), Pseudomonas aeruginosa (Pa) colonizes the lungs, leading to chronic inflammation of the bronchial epithelium. ChaC glutathione-specific γglutamylcyclotransferase 1 (CHAC1) mRNA is differentially expressed in primary human airway epithelial cells from bronchi (hAECBs) from patients with CF and healthy patients at baseline and upon infection with Pa. CHAC1 degrades glutathione and is associated with ER stress and apoptosis pathways. In this study, we examined the roles of CHAC1 in the inflammatory response and apoptosis in lung epithelial cells. First, we confirmed by reverse transcription quantitative polymerase chain reaction that CHAC1 mRNA was overexpressed in hAECBs from patients without CF compared with the expression in hAECBs from patients with CF upon Pa (PAK strain) infection. Moreover, the Pa virulence factors LPS and flagellin were shown to induce CHAC1 expression in cells from patients without CF. Using NCI-H292 lung epithelial cells, we found that LPS-induced CHAC1 mRNA expression was PERK-independent and involved ATF4. Additionally, using CHAC1 small interfering RNA, we showed that reduced CHAC1 expression in the context of LPS and flagellin stimulation was associated with modulation of inflammatory markers and alteration of NF-κB signaling. Finally, we showed that Pa was not able to induce apoptosis in NCI-H292 cells. Our results suggest that CHAC1 is involved in the regulation of inflammation in bronchial cells during Pa infection and may explain the excessive inflammation present in the respiratory tracts of patients with CF.
Keywords: cystic fibrosis, bronchial epithelium, Pseudomonas aeruginosa, inflammation, ChaC glutathionespecific γ-glutamylcyclotransferase 1, ER stress INTRODUCTION Cystic fibrosis (CF), the most common severe autosomal recessive genetic disease in the Caucasian population, is caused by mutations in the CF transmembrane conductance regulator (CFTR) gene (1)(2)(3). Lung injury is predominant and is the leading cause of morbidity and mortality in CF patients. In airways, CFTR dysfunction leads to alterations in mucociliary clearance and impairment of innate immune host defenses. This results in chronic bacterial infection and inflammation, leading to degradation of the lung epithelium and progressive lung damage in CF patients (4). Pseudomonas aeruginosa (Pa) is the leading cause of chronic pulmonary infection in the airways of adults with CF (5). Pa colonization induces a strong inflammatory response in CF patients, particularly neutrophil accumulation and hypersecretion of inflammatory cytokines, such as IL-8 (6,7).
The lung epithelium plays an essential role in the fight against infections by orchestrating innate immunity (8). Lung epithelial cells are sufficient to generate an innate protective immune response in mice infected with Pa (9). In CF, establishment of chronic pulmonary infection with Pa is linked to alterations in epithelial cell responses, which do not allow establishment of an adequate defense system (4). Previously, we showed that this response was indeed different by comparative analysis of the transcriptomic signatures of primary CF and non-CF bronchial epithelial cells during Pa infection (10). We identified several up-and downregulated mRNAs in CF cells compared with those in non-CF cells, suggesting that these genes may be involved in the exacerbated inflammatory response to Pa found in CF patients. Among these dysregulated mRNAs, ChaC glutathione-specific γ-glutamylcyclotransferase 1 (CHAC1) showed dramatically differential expression between non-CF and CF cells both at baseline and after Pa infection.
CHAC1 was identified in mammalian cells for the first time in 2009 as a new component of the unfolded protein response (UPR) pathway (11). This UPR pathway is induced in response to ER stress due to the presence of misfolded proteins in the ER lumen. The UPR consists of three pathways initiated by inositol-requiring enzyme 1α, eukaryotic translation initiation factor 2-α kinase 3 (PERK), and activation transcription factor (ATF) 6 (12). These three pathways alleviate ER stress by decreasing protein synthesis, facilitating protein folding, and increasing protein degradation. CHAC1 is induced by the PERK pathway downstream of ATF4 and the pro-apoptotic factor DNA damage-inducible transcript 3 (DDIT3 also called CHOP) and has been described as necessary and sufficient to induce markers of apoptosis, particularly the cleavage of the apoptotic factor poly(ADP-ribose) polymerase 1 (PARP) (11). Moreover, CHAC1 can degrade glutathione in the cytosol of mammalian cells through its γ-glutamylcyclotransferase activity (13). Various studies have shown a link between ER stress and inflammation (12,14,15). In the context of CF, atypical activation of the UPR pathway has been observed; the resulting ER stress can potentiate the inflammatory response (12). Moreover, the UPR is implicated in the regulation of cytokine production for bacterial infections (16). The role of CHAC1 in this interplay between ER stress and inflammation in the context of Pa infection has never been investigated.
Accordingly, in this study, we hypothesized that CHAC1 may play a role in regulating the inflammatory process induced by Pa infection and apoptosis.

Bacterial Strains, Growth Conditions, and Infection Protocols
The Pa WT PAK strain, the mutant FliC with deletion of flagellin and the double mutant pscF xcpQ with deletion of type II and III secretion systems were used and grown as previously (20,21). The OD measured at 600 nm was adjusted to give the desired MOI. MOI were verified by serial 10fold dilutions of the bacterial suspensions and plating on LB agar (Supplementary Figure 3). Cells seeded in 6-well (hAECBs) or 12-well plates (NCI-H292) were infected with Pa in their culture medium without antibiotics. After each stimulation, supernatants were collected and centrifuged at 3,000 × g for 15 min to remove the bacteria. Clinical strains of Pa were isolated according to the standards procedures of the microbiological department of St-Antoine Hospital, Paris, France. Sputum from patients was diluted (v/v) with prediluted Digest-Eur (Eurobio, Courtaboeuf, France), vortexed, and incubated for 15 min at room temperature. A 10-µL aliquot was plated on different selective media for 5 days at 37 • C under aerobic conditions. Pa was isolated from Drigalski's agar plates, and identification was confirmed by mass spectrometry (Maldi-Biotyper, Brucker, France). Stenotrophomonas maltophilia was freshly isolated from a CF patient, and its identification was confirmed by mass spectrometry. S. maltophilia and Staphyloccocus aureus [Newman strain (22)] were cultured using the same protocol for Pa as described above.

RT qPCR
RNA was isolated using a NucleoSpin RNA/Protein kit (Macherey Nagel, Duren, Germany). RT was performed using a high-capacity cDNA kit (Applied Biosystems, Foster City, CA, USA). Real-time qPCR was performed with an ABI StepOnePlus, using TaqMan Fast Universal PCR Master Mix (Applied Biosystems), TaqMan probes for CHAC1 (Hs_00225520), IL-8 (Hs_174103), IL-6 (Hs_985639), CCL2 (Hs_00234140), GAPDH (Hs_2786624) and cDNA as a template. For relative quantification, the amount of target genes was normalized to the expression of GAPDH relative to control cells used as a calibrator and was calculated using the 2 − Ct method.

Fluorescence Microscopy
CHAC1-GFP expression was monitored in live NCI-H292 cells using an EVOS cell imaging system (Thermo Fisher Scientific).

NF-κB Luciferase Activity
NF-κB luciferase activity in A549 cells was assayed with a Luciferase kit (Promega, Madison, WI, USA) according to the manufacturer's protocol with a FLUOstar OPTIMA luminometer (BMG LABTECH, Champigny sur Marne, France).

Statistical Analysis
Differences among groups were assessed for statistical significance using Prism 7.00 software (GraphPad Software, La Jolla, CA, USA) as indicated in the Figure Legends. Differences with p < 0.05 were considered statistically significant.

Defective Regulation of CHAC1 in CF Cells
In a previous transcriptomic analysis, we identified CHAC1 as one of the differentially expressed mRNAs between human airway epithelial cells from bronchi (hAECBs) of patients with and without CF during Pa infection (10). Indeed, CHAC1 mRNA was only induced in hAECBs from non-CF patients ( Figure 1A, Figure 1F) in CF cells.
We further validated the differential expression of CHAC1 by qPCR in the same samples used for RNA-seq. Indeed, at baseline, we observed significantly higher expression of CHAC1 in non-CF cells compared with that in CF cells. As expected from the transcriptomic data, we observed upregulation of CHAC1 mRNA expression after 6 h of Pa infection only in non-CF cells ( Figure 1B).

LPS and Flagellin Induced CHAC1 Overexpression in Primary Bronchial Epithelial Cells and NCI-H292 Cells
Because Pa induced CHAC1 expression in non-CF cells, we investigated which Pa virulence factors were involved. We studied the effects of two very well-known virulence factors, LPS and flagellin from Pa, on hAECBs from CF and non-CF patients and on NCI-H292 cells, which expressed CHAC1 mRNA at sufficient levels for further knockdown experiments in comparison to BEAS-2B cells (Supplementary Figure 2A). Notably, 6 h of stimulation with LPS and flagellin induced CHAC1 expression in hAECBs from non-CF patients, similar to whole bacteria at a MOI of 0.25 for 6 h. This induction by LPS and flagellin and by living bacteria was not found in hAECBs from CF patients (Figure 1C). We observed similar effects of LPS and flagellin on CHAC1 expression in NCI-H292 cells (Figure 2A). To validate the role of flagellin in CHAC1 expression, we used a mutant of the Pa PAK strain, FliC, in which flagellin was deleted. We observed a significantly lower increase in CHAC1 mRNA expression when NCI-H292 cells were stimulated with the FliC strain than with the WT PAK strain (Figure 2B). We also tested a double mutant of the Pa PAK strain with both type II and III secretion systems deleted ( pscF xcpQ). This mutant induced CHAC1 mRNA expression, similar to the WT PAK strain (Figure 2B). The difference in CHAC1 expression observed with the FliC strain was not due to different MOI since equal numbers of Pa colonies were found on LB agar (Supplementary Figure 3A). Subsequently, in order to determine whether the observed CHAC1 upregulation was restricted to the laboratory PAK strain, we used Pa strains freshly isolated from different patients with CF. We observed that clinical isolates of Pa were also able to induce CHAC1 mRNA expression in NCI-H292 cells (Figure 2C) with an effect similar to PAK. This induction is highly variable between the clinical strains used and may reflect difference of virulence of each strain, since the number of the bacteria used for stimulation is similar (Supplementary Figure 3B). CHAC1 induction by PAK strain is also highly variable (Figure 2C), and is likely due to the variability in the inoculum used (Supplementary Figure 3B). We also observed that the induction of CHAC1 mRNA expression by PAK was higher in NCI-H292 cells than in hAECBs (Figures 2B,C vs. Figure 1C). Finally, we tested two other bacterial species known to colonize the airways of patients with CF, i.e., S. aureus and S. maltophilia. Using a similar MOI, we observed reduced CHAC1 mRNA induction in NCI-H292 cells infected with S. aureus and S. maltophilia compared with that in the cells infected with Pa. Concomitantly, induction of IL-8 mRNA expression was reduced ( Figure 2D). The differences in CHAC1 and IL-8 mRNA expression observed with S. aureus and S. maltophilia were not caused by different MOI because equal numbers of colonies were found on LB agar (Supplementary Figure 3C).
Next, we examined CHAC1 protein expression in response to Pa stimulation. Consistent with previous reports indicating that CHAC1 protein is very unstable due to its rapid degradation by the proteasome in Neuro2a cells (23), we were unable to detect endogenous CHAC1 protein in NCI-H292 at baseline (Supplementary Figure 2B) or after Pa infection (data not shown). Additionally, we were unable to detect CHAC1 protein expression after treating the cells with tunicamycin, a known inducer of CHAC1 mRNA expression (24) (data not shown). CHAC1 was only detected at the protein level when it was either overexpressed with a plasmid in NCI-H292 cells (Figure 2E) or when the cells were treated with at least 1 µM of the proteasome inhibitor MG-132 ( Figure 2F).

CHAC1 Upregulation Was ATF4 Dependent and PERK Independent
Because CHAC1 has been shown to be associated with the PERK pathway (11), we hypothesized that induction of CHAC1 by LPS from Pa may be mediated by PERK. Therefore, we used a PERK inhibitor and tunicamycin (TM) as a positive control for  Figure 2E). activation of the CHAC1 and PERK pathway (11). As expected, compared with untreated NCI-H292 cells, LPS and TM induced CHAC1 mRNA expression. When cells were treated with TM and PERK inhibitor, there was a significant 90.3% decrease in CHAC1 mRNA expression compared with that in cells treated with TM alone (Figure 3A). However, when cells were treated with LPS and PERK inhibitor, CHAC1 mRNA expression was not significantly different from that in cells treated with LPS only (Figure 3A). We then hypothesized that this induction of CHAC1 by LPS would directly pass through ATF4. Indeed, we observed that LPS was able to activate ATF4 and that siRNA against ATF4 inhibited this induction ( Figure 3B). The inhibition of ATF4 expression by the siRNA ATF4 (∼60%) was associated with an inhibition of CHAC1 induction by LPS (Figures 3C,D).

CHAC1 Overexpression Induced by Pa Did Not Stimulate Apoptosis
CHAC1 has been shown to play a role in apoptosis (11); therefore, we investigated the effects of Pa stimulation on apoptosis. We infected NCI-H292 cells with two different MOIs (0.5 and 1) for 5 or 8 h and examined the levels of cleaved PARP and cleaved caspase-3. As expected, we observed that the positive control staurosporine induced cleavage of PARP and caspase-3. In contrast, Pa did not induce cleavage of PARP and caspase-3 ( Figure 4A). PARP and caspase-3 cleavages were observed when CHAC1 was overexpressed with a plasmid in NCI-H292 cells (Figure 4B).

CHAC1 Regulated the Inflammatory Response of Bronchial Epithelial Cells Induced by LPS or Flagellin
CHAC1 is a component of the ER stress pathway, and a link between ER stress and the inflammatory response has already been described in the context of CF (12). Therefore, we investigated the role of CHAC1 in the inflammatory response induced by LPS or flagellin. Transfection with CHAC1 siRNA induced around 50% reduction in CHAC1 mRNA expression (Figure 5A). We then examined different inflammatory parameters, particularly the secretion of IL-8, IL-6, and PGE2. As expected, we observed that 1 µg/mL LPS, 50 ng/mL flagellin, and the combination of both (LPS+flagellin) induced significant increases in both IL-8 and IL-6 secretion and tended to increase PGE 2 secretion (Figures 5B-D). Under these stimulation conditions, when CHAC1 mRNA expression was decreased using siRNA, greater significant increases in IL-8 secretion ( Figure 5B) and PGE 2 secretion ( Figure 5D) were observed. In contrast, in the presence of CHAC1 siRNA, we observed significant inhibition of IL-6 secretion ( Figure 5C) in cells stimulated with LPS (80% inhibition), flagellin (30% inhibition), or the combination of both virulence factors (80% inhibition). We obtained similar results for the mRNA expression of IL-8 and IL-6 (Supplementary Figures 4A,C). CCL2 mRNA expression was also modulated similar to IL-8 expression under these experimental conditions (not detectable by ELISA; Supplementary Figure 4B).

CHAC1 Regulated NF-κB Activation
To determine whether regulation of these inflammatory parameters involved NF-κB, we studied its phosphorylation by western blot in NCI-H292 cells. We observed that LPS induced a significant increase in NF-κB p65 phosphorylation in CHAC1-knockdown cells after 1 h of stimulation ( Figure 6A). To confirm these data, we used A549-NF-κB luciferase cells. As in NCI-H292 cells, 60% inhibition of CHAC1 mRNA expression was observed after siRNA transfection (data not shown). Similar to NCI-H292 cells, we observed that 1 µg/mL LPS, 50 ng/mL flagellin, and the combination of both (LPS+flagellin) induced a significant increase in IL-8 secretion when A549 cells were transfected with CHAC1 siRNA compared with that in cells transfected with siRNA control (Figure 6B). Additionally, as expected, we observed that 1 µg/mL LPS, 50 ng/mL flagellin, and the combination of both (LPS+flagellin) induced significant increases in NF-κB activity (Figure 6C). Under these experimental conditions, when we decreased CHAC1 mRNA expression, we observed an even greater increase in NF-κB activity when cells were stimulated with LPS and/or flagellin ( Figure 6C). IL-6 and CCL-2 were not detectable by ELISA (data not shown).
(25) pathways. In our previous study, we identified the CHAC1 mRNA as one of the most highly differentially expressed transcripts between primary bronchial cells from CF and non-CF patients, infected with Pa. We also found that CF cells exhibited an expected excessive inflammatory response, including higher IL-8 production (10). Because CHAC1 was recently described as a new component of the UPR in the PERK pathway (12), we aimed to evaluate whether CHAC1 was involved in the modulation of the inflammatory response induced by Pa in lung epithelial cells.
We confirmed our previous transcriptomic data showing that CHAC1 mRNA was overexpressed in primary non-CF epithelial cells compared with its expression in CF cells at baseline and upon Pa infection. Additionally, Pa and two of its virulence factors, LPS and flagellin, the respective ligands of Toll-Like Receptor (TLR) 4 and TLR5, were able to induce CHAC1 at the transcriptional level in bronchial epithelial NCI-H292 cells. As previously shown in Neuro2a (23) or HEK cells (26), we were able to detect CHAC1 protein expression only after MG-132 treatment, suggesting the involvement of the proteasome pathway in the stability of the protein (26). This induction of CHAC1 mRNA expression was observed not only with the laboratory strain PAK but also with freshly isolated clinical strains. Interestingly, S. maltophilia and S. aureus, both known to chronically infect CF patients, were not able to induce CHAC1 mRNA expression to similar levels. Using cellular infection models, other transcriptomic studies have consistently shown CHAC1 upregulation with microorganisms, including human coronavirus (27), human cytomegalovirus (28), tick-borne flaviviruses (29), Zika virus (30), and the bacterium Mycoplasma hominis (31). In these studies, the specific role of CHAC1 was not investigated. However, CHAC1 is essential as embryonic lethality was observed in mice with deletion of the Chac1 gene (32).
In our study, CHAC1 mRNA induction by LPS and Pa was not related to PERK activation, as is TM, but was dependent on ATF4. In contrast, LPS from E. coli has previously been shown to activate both PERK and ATF4 in A549 cells (33). Notably, ATF4 has been shown to be directly recruited by TLR4 independent of ER stress in human monocytes (34). Consistent with our results, another ER stress-associated molecule, GADD34, was recently shown to be upregulated by several virulence factors from Pa, including pyocyanin and AprA, independent of PERK but dependent on ATF4 (35). Taken together, these results were consistent with emerging evidence showing that induction of the UPR pathway by microorganisms did not reflect ER stress but was an integral part of the immune response and TLR pathways (36).
The fine tuning of immune responses by the UPR has been extensively studied (37). However, its contribution to the regulation of cytokine production in bacterial infection is not well understood (16). Here, we observed that inhibition of CHAC1 mRNA expression by siRNA promoted IL-8 and CCL2 expression and PGE2 secretion, suggesting that CHAC1 prevents excessive inflammatory responses dependent on the NF-κB p65 pathway. These results were consistent with a recent study reporting that bronchial epithelial cells undergoing ER stress showed reduced IL-8 and CXCL1 production in the presence of inflammatory stimuli (38). Indeed, pretreatment of bronchial epithelial cells with TM [shown here and by others (11) to be able to increase CHAC1 expression] can prevent IL-8 and CXCL1 production. In contrast, inhibition of CHAC1 mRNA expression by siRNA inhibits IL-6 production, suggesting that CHAC1 may promote IL-6 expression. Our results were consistent with a previous study showing that TM promotes IL-6 secretion in bronchial IB3-1 (CF deficient) and C38 cells stimulated with flagellin (12). Previous works have demonstrated that IL-6 production in A375 skin cells is dependent on CHOP (39). However, CHOP is expected to be upstream in the CHAC1 pathway and does not explain our results; future studies are needed to determine how the IL-6 pathway is affected specifically in lung epithelial cells.
In addition to its role in UPR, CHAC1, as a member of the γglutamylcyclotransferase family, is able to hydrolyze glutathione only in its reduced form (GSH) (40). Glutathione has been widely studied owing to the function of CFTR as a glutathione transporter (41). GSH homeostasis is defective in CF, and GSH promotes oxidative stress and related inflammation (42). The roles of CHAC1 activity in relation to cellular glutathione homeostasis and the observed cytokine regulation remain to be determined, although previous studies have demonstrated that the levels of glutathione-related enzyme can regulate cytokine production (43).
In our experiments, PAK did not induce apoptosis in NCI-H292 cells. However, these cells are not completely resistant to apoptosis since staurosporine was able to induce caspase-3 and PARP cleavage. Thus, although CHAC1 was induced by Pa, it was not sufficient to induce cell apoptosis. It is conceivable that the level of induction of CHAC1 may be too low to promote apoptosis. In contrast, high induction of CHAC1 using an overexpression plasmid led to induction of PARP cleavage, indicative of cell apoptosis. Apoptosis dysfunction is still highly controversial in CF (44), and our results are partly consistent with some previous studies. Indeed, even if Pa has been previously shown to induce apoptosis in lung epithelial cells, variable susceptibility between lung epithelial cells has been observed. Moreover, the PAO1 strain of Pa has been shown to induce apoptosis in human tracheal 9HTEo cells but not in primary nasal epithelial cells from polyps or in 16HBE14o bronchial cells, which are both resistant (45). In contrast, other bronchial cell lines are permissive to apoptosis induced by Pa, such as IB3-1 (F508del/W1282X), C38, S9 (both IB3-1 corrected cells), CFT1-LCFSN, and CFT1-LC3 (F508del) cells. More recently, using Calu-3 polarized lung epithelial cells, researchers showed that PAO1 was able to induce apoptosis markers, such as active caspase-3, within 6-8 h. Additionally, the effects of the CFTR genotype on apoptosis of lung epithelial cells at baseline (44) or after Pa infection are still debated and may be dependent on which specific marker of apoptosis is being evaluated (45,46).
In conclusion, we showed that inhibition of CHAC1 in bronchial epithelial cells resulted in the regulation of inflammatory mediators. Moreover, CHAC1 was induced preferentially in normal hAECBs in comparison to CF hAECBs in response to Pa and its virulence factors, LPS and flagellin. Thus, these results suggest that low CHAC1 expression may contribute to the excessive inflammation observed in CF bronchial epithelial cells after Pa infection and thus may be associated with the exacerbation of inflammation characterizing the lungs from patients with CF. Inflammation is complex and thought to have a critical impact during the course of CF lung disease (47). A better understanding of the link between infection and inflammation is thus required for the development of new therapeutics.

AUTHOR CONTRIBUTIONS
LG, MC, and VB designed the study.