Aspergillus fumigatus Supernatants Disrupt Bronchial Epithelial Monolayers: Potential Role for Enhanced Invasion in Cystic Fibrosis

Aspergillus fumigatus is the most commonly isolated fungus in chronic lung diseases, with a prevalence of up to 60% in cystic fibrosis patients. Despite this, the impact of A. fumigatus colonisation on lung epithelia has not been thoroughly explored. We investigated the influence of A. fumigatus supernatants and the secondary metabolite, gliotoxin, on human bronchial epithelial cells (HBE) and CF bronchial epithelial (CFBE) cells. CFBE (F508del CFBE41o−) and HBE (16HBE14o−) trans-epithelial electrical resistance (TEER) was measured following exposure to A. fumigatus reference and clinical isolates, a gliotoxin-deficient mutant (ΔgliG) and pure gliotoxin. The impact on tight junction (TJ) proteins, zonula occludens-1 (ZO-1) and junctional adhesion molecule-A (JAM-A) were determined by western blot analysis and confocal microscopy. A. fumigatus conidia and supernatants caused significant disruption to CFBE and HBE TJs within 24 h. Supernatants from later cultures (72 h) caused the greatest disruption while ΔgliG mutant supernatants caused no disruption to TJ integrity. The ZO-1 and JAM-A distribution in epithelial monolayers were altered by A. fumigatus supernatants but not by ΔgliG supernatants, suggesting that gliotoxin is involved in this process. The fact that ΔgliG conidia were still capable of disrupting epithelial monolayers indicates that direct cell–cell contact also plays a role, independently of gliotoxin production. Gliotoxin is capable of disrupting TJ integrity which has the potential to contribute to airway damage, and enhance microbial invasion and sensitisation in CF.


Introduction
Aspergillus fumigatus is the most commonly isolated fungal pathogen in chronic lung diseases and particularly in those individuals with cystic fibrosis (CF) [1][2][3][4]. CF is a lifethreatening hereditary disorder that affects multiple organs of the body, with the respiratory system most severely affected. Mutations in the CF transmembrane conductance regulator (CFTR) gene result in the absence of functioning CFTR proteins causing decreased chloride secretions and increased sodium absorption at the apical membrane of the airway epithelial cells leading to abnormally viscous mucous [5]. Coupled with impaired mucociliary clearance, this makes the CF airways an ideal environment for persistent microbial colonisation. Airway disease in CF is characterised by a continuous cycle of infection and inflammation, playing a significant role in exacerbations and disease progression [6,7]. The most common manifestation of A. fumigatus in CF is allergic bronchopulmonary aspergillosis (ABPA) occurring in 1-15% of CF patients and causing reduced pulmonary function [8]. Up to 60% of CF patients are colonised with Aspergillus [9][10][11][12] and we have shown that CF patients

Microbiological Strain Maintenance, Storage and Culture Conditions
A. fumigatus strain Af293 (ATCC ® MYA-4609 TM ), a gliotoxin deficient mutant ∆gliG derived from the Af293 wild type [31] and two clinical A. fumigatus strains isolated from CF sputum samples from two different patients, namely AF1 and AF2, were all maintained on Malt Extract Agar (MEA) (Fannin, Dublin, Ireland). Stocks of all A. fumigatus isolates were kept in 15% glycerol (v/v) at −80 • C. The ∆gliG mutant was generated using the bipartite marker technique and is deficient in producing gliotoxin [31,32]. The AF1 strain was isolated from 6 consecutive sputum samples from a single CF patient taken over the course of 6 months and confirmed using the STRAf assay [33] to be an indistinguishable genotype ( Figure S1). The AF2 strain was isolated from a single sputum sample taken from a CF patient and not subsequently isolated from the same patient ( Figure S1).

Infection of CFBE and HBE Monolayers with A. fumigatus Isolates
For tight junction assays, CFBE or HBE cell lines were seeded onto each 0.4 µm polyester transwell filter (Merck) at a dilution of 1 × 10 5 /0.5 mL with 1.5 mL of media applied to the basolateral side of the transwell support. A blank well containing only S-MEM [26] and a control with cells only were also included. Following 18 to 24 h incubation, the apical medium was removed creating an air-liquid interface [26]. S-MEM was replaced on alternate days and cells were cultured over a 5-day period to enable the formation of TJs. On day 6, Trans-epithelial electrical eesistance (TEER) measurements were taken in triplicate for all wells using an EVOM meter (World Precision Instruments) and only those exceeding TEER measurement of 150 Ω·cm 2 (above blank) were used for the experiment. CFBE and HBE monolayers were infected with Af293, ∆gliG, AF1 and AF2 conidia at a multiplicity of infection (MOI) of 2:1 and incubated at 37 • C with 5% CO 2 . Sterile latex beads (Merck) 3 µm in diameter (MOI 2:1) were also seeded onto CFBE and HBE monolayers as controls. TEER values were recorded at 0, 2, 4, 6, 8, 10, 12 and 24 h post infection. Experiments were performed on three independent occasions. Triplicate ten-fold dilutions of conidia concentrations were made for plating to confirm conidia concentration of MOI for each experiment.
TEER values were calculated using the following formula: Resistance of cells on filter-resistance of blank filter (with no cells) 1.12 cm 2 (area of filter) = TEER Ω·cm 2 Results were recorded as a percentage of the control TEER for each time point.

Treatment of CFBE and HBE Monolayers with Supernatants from A. fumigatus Isolates
Seven cultures containing 2 × 10 5 conidia/3 mL minimal essential medium (MEM) (Merck) for each of the four isolates were incubated at 37 • C for 0, 4,8,12,24,48 or 72 h. Cultures were passed through a sterile cell sieve and then sterile filtered using a 0.2 µm syringe filter (Sarstedt, Lower Saxony, Germany) to remove all A. fumigatus conidia and hyphae. Supernatants were confirmed as containing no viable conidia or hyphae by plating 20 µL of the filtered supernatants onto an MEA plate in triplicate and incubated at 37 • C for 7 days. No growth confirmed sterile filtration of supernatants. CFBE and HBE monolayers with TEER measurements exceeding 150 Ω·cm 2 were inoculated with 0, 4,8,12,24,48 or 72 h supernatants of the four A. fumigatus strains. TEERs were monitored every 2 h from 0 to 12 h and again at 24 h post infection. Experiments were performed on three independent occasions.

Exposure of CFBE and HBE Monolayers to Gliotoxin
Gliotoxin (Merck) was suspended in DMSO at a stock concentration of 0.1 M and was diluted to working solutions in MEM. All final gliotoxin concentrations contained no more than 0.08% DMSO for tissue culture assays. Gliotoxin was applied to CFBE and HBE monolayers at a concentration of 0.8, 8 or 80 µM. Controls of untreated monolayers and 0.08% DMSO-treated monolayers were included. TEERs were monitored every two hours from 0 to 12 h and again at 24 h post infection. Experiments were performed on three independent occasions.

Western Blot Analysis of TJ Proteins JAM-A and ZO-1
CFBE and HBE monolayers were prepared as previously described. Supernatants of Af293 and ∆gliG grown for 4, 8 or 72 h were prepared as previously described and were applied to transwell inserts containing monolayers. For all experiments, a cell-only control was included. Following 24 h exposure, epithelial cell monolayers were washed three times with ice cold PBS and protein was harvested by adding 100 µL radioimmunoprecipitation assay buffer (RIPA) (Merck) plus protease inhibitors (Roche, Basel, Switzerland) to each transwell filter. Cells were gently scraped off using a sterile cell scraper and lysed with repeated pipetting. The suspension was transferred to a fresh tube on ice and incubated for at least 15 min. Samples were then sonicated for 1 min and transferred immediately to −80 • C. Proteins were quantified using the Qubit fluorometer and protein quantifi- Membranes were washed and then incubated with either horseradish peroxidase (HRP)conjugated anti-mouse (1:50,000 for GAPDH and JAM-A) or HRP-conjugated anti-rabbit (1:50,000 for ZO-1) (Thermoscientific, Waltham, MA, USA) for 1 h at room temperature. Following immunoblotting, proteins were detected by chemiluminescence and visualised using a Fuji film ImageQuant LAS 4000 analyser. The density of each band was compared with the corresponding control band and normalised against GAPDH by densitometry using Quantity One version 4.6.6 software. Results are presented as percentage of the untreated control.
2.7. Immuno-Fluorescence and Confocal Microscopy of CFBE and HBE Monolayers Exposed to Supernatants of Af293 or ∆gliG CFBE and HBE monolayers were grown on transwell inserts as previously described. After 24 h exposure to medium alone or supernatants of Af293 or ∆gliG, monolayers were prepared for immuno-fluorescent imaging. Monolayers were washed with PBS, then permeablised with ice cold methanol (Merck) for 30 min and blocked with PBS containing 1% BSA (w/v) for 10 min. Cells were then incubated with 10 µg/mL primary antibody (rabbit anti-ZO-1 (Life Technologies) or mouse anti-JAM-A (Santa Cruz Biotechnology)) for 1 h at room temperature. Cells were then washed four times with PBS containing 1% BSA (w/v) and subsequently incubated with FITC-conjugated anti-rabbit (ZO-1) or anti-mouse (JAM-A) (Jackson Immunoresearch, West Grove, PA, USA) antibodies for 1 h at room temperature protected from the light. Monolayers were washed five times with PBS containing 1% BSA (w/v) and then post-fixed in PBS containing 4% paraformaldehyde (w/v) for 10 min. Following fixation, filters were removed from the inserts and mounted on slides with Vectashield ® containing DAPI (Vector Laboratories) and examined by confocal microscopy. Confocal images were captured at ×600 magnification using the Olympus FLOUVIEW FV1000 microscope utilising the FITC and DAPI channels and FV10-ASW software version 03.01. Images are representative of 3 fields of view taken per treatment.

Statistical Analysis
All statistical analysis was performed on GraphPad Prism version 5 and 9. Statistical analysis using the two-way ANOVA and Bonferroni post-test was carried out on each experiment investigating the ability of conidia or supernatants or gliotoxin to affect the TEER of CFBE and HBE monolayers at 2 h intervals over a 24 h period. Comparison of densitometry results from western blots was performed using one-way ANOVA and Bonferroni post-test. p values of <0.05 were considered significant.

A. fumigatus Isolates and Their Supernatants Opened CFBE and HBE TJs
A. fumigatus conidia of the Af293 reference strain and CF clinical isolates (AF1 and AF2) disrupted CFBE TJs within 12 h ( Figure 1A) and HBE TJs within 24 h ( Figure 1B). A statistically significant decrease in the TEER values was observed between infected versus noninfected (control) CFBE monolayers from 8 h post-infection for Af293 (p < 0.01; two-way ANOVA) and from 12 h post-infection for both clinical isolates (p < 0.01; two-way ANOVA) ( Figure 1A). In contrast, a significant decrease in the TEER values of HBE monolayers was only reached following 24 h infection (p < 0.01; two-way ANOVA) ( Figure 1B). There was no significant difference in the ability of AF1 and AF2 conidia to impair CFBE or HBE TJs ( Figure 1A). Untreated CFBE and HBE monolayers maintained TJ integrity, with TEER values ranging from 90% to 114% and 83% to 105% of the initial TEER for the duration of the experiments, respectively.
Supernatants of Af293, AF1 and AF2 were capable of opening the TJs of CFBE cells ( Figure 1C). The 72 h supernatants of Af293 caused the greatest degree of disruption to TJ integrity. The 12 h and 24 h supernatants of the two clinical isolates were more disruptive to CFBE integrity than the corresponding supernatants of Af293 (p < 0.01; two-way ANOVA). In contrast, the later supernatants (48 h and 72 h) of Af293 were more disruptive to CFBE TJs than the later supernatants of the clinical isolates (p < 0.05; two-way ANOVA). The 72 h supernatant of AF2 was significantly more disruptive to TJ integrity than the 72 h supernatant of AF1 (p < 0.001). All isolate supernatants from the 4 h and 8 h cultures were capable of significantly disrupting TJ integrity within 24 h (p < 0.0001; two-way ANOVA) ( Figure 2A). The disruption to TJ integrity by the 72 h supernatants of A. fumigatus was comparable in CFBE and HBE monolayers with a significant decrease in TEER occurring at 2 h post-exposure for both cell types (p < 0.0001; two-way ANOVA) ( Figure 1D).
Overall, the 72 h supernatant of Af293 caused the greatest disruption to CFBE and HBE TJ integrity and this effect was induced as early as 2 h post-exposure to the supernatant ( Figure 1D).

Supernatants from the Gliotoxin Deficient A. fumigatus Mutant, ∆gliG, Do Not Maintain the Ability of the Af293 Wild Type to Disrupt CFBE and HBE Monolayers
Conidia of the wild type Af293 and the gliotoxin mutant, ∆gliG, caused a significant decrease in the CFBE monolayer TEER compared with the control 24 h post-infection (p < 0.0001; two-way ANOVA) (Figure 2A). A similar trend was observed in HBE cells with a significant decrease in TEER occurring following 24 h exposure to conidia (p < 0.0001) ( Figure 2B). As in CFBE cells, the conidia of Af293 caused significant disruption to the HBE monolayers from 10 h post-infection (p < 0.01) (Figure 2A,B) whereas conidia of ∆gliG took 24 h to cause significant disruption to monolayers.
Supernatants from Af293 culture time points as early as 4 h and 8 h were shown to disrupt CFBE TJs after 24h ( Figure 1C); therefore, the effect of 4 h, 8 h and 72 h supernatants of Af293 and ∆gliG were compared. The 72 h supernatant of Af293 caused the greatest decrease in the CFBE TEER at 24 h (p < 0.0001; two-way ANOVA), while the 72 h supernatant of ∆gliG caused no significant decrease in TEER ( Figure 2C). The 4 h and 8 h supernatants of Af293 and ∆gliG both caused comparable decreases in TEER at 24 h (p < 0.05; two-way ANOVA). These results were mirrored in HBE cells ( Figure S2). To further investigate whether gliotoxin is capable of opening CFBE and HBE TJs, monolayers were treated with 0.8 µM, 8 µM and 80 µM pure gliotoxin (Sigma). All concentrations of gliotoxin caused significant reductions in CFBE and HBE monolayer TEER values at 24 h (p < 0.0001; one-way ANOVA) ( Figure 2D). The 0.08% DMSO control (gliotoxin diluent) caused no significant reduction in TEERs. Conidia of the wild type Af293 and the gliotoxin mutant, ΔgliG, caused a significant decrease in the CFBE monolayer TEER compared with the control 24 h post-infection (p < 0.0001; two-way ANOVA) (Figure 2A). A similar trend was observed in HBE cells with a significant decrease in TEER occurring following 24 h exposure to conidia (p < 0.0001) ( Figure 2B). As in CFBE cells, the conidia of Af293 caused significant disruption to the HBE monolayers from 10 h post-infection (p < 0.01) (Figure 2A,B) whereas conidia of ΔgliG took 24 h to cause significant disruption to monolayers.
Supernatants from Af293 culture time points as early as 4 h and 8 h were shown to disrupt CFBE TJs after 24h ( Figure 1C); therefore, the effect of 4 h, 8 h and 72 h supernatants of Af293 and ΔgliG were compared. The 72 h supernatant of Af293 caused the greatest decrease in the CFBE TEER at 24 h (p < 0.0001; two-way ANOVA), while the 72 h supernatant of ΔgliG caused no significant decrease in TEER ( Figure 2C). The 4 h and 8 h supernatants of Af293 and ΔgliG both caused comparable decreases in TEER at 24 h (p < 0.05; two-way ANOVA). These results were mirrored in HBE cells ( Figure S2). To further investigate whether gliotoxin is capable of opening CFBE and HBE TJs, monolayers were TJs are specialised multi-protein complexes that define epithelial organisation. Two proteins with key roles in maintaining barrier function are ZO-1 and JAM-A. Thus, to further investigate the abilities of A. fumigatus supernatants to impair CFBE and HBE epithelial integrity, western blot analysis of two TJ proteins, ZO-1 and JAM-A, was performed. Although 4 h and 8 h supernatants of Af293 and ∆gliG were capable of opening TJs, no reduction in either ZO-1 or JAM-A in CFBE ( Figure 3A) or HBE cells ( Figure 3B) was observed in response to these supernatants. In contrast CFBE cells ( Figure 3A) and HBE cells ( Figure 3B) exposed to 72 h supernatants of Af293 showed reduced levels of ZO-1 and JAM-A TJ proteins. This was not observed in CFBE or HBE cells exposed to the 72 h supernatant of the gliotoxin deficient mutant, ∆gliG. The levels of ZO-1 and JAM-A detected in CFBE cells exposed to the 72 h supernatant of Af293 were significantly lower than in CFBE cells exposed to the 72 h supernatant of ∆gliG (p < 0.01; two-way ANOVA) ( Figure 3A). This result was mirrored in HBE cells ( Figure 3B). GAPDH levels in CFBE and HBE cells exposed to the 72 h supernatant of Af293 were also slightly reduced; therefore, the concentrations of JAM-A and ZO-1 levels were normalised to GAPDH in all the sample treatment scenarios and normalised change ZO-1 and JAM-A highlights the specificity of the effect on TJ proteins.  Figure 2D). The 0.08% DMSO control (gliotoxin diluent) caused no significant reduction in TEERs.

TJ Proteins, ZO-1 and JAM-A, Are Disrupted by 72 h Supernatant of Af293 but Not the ΔgliG
TJs are specialised multi-protein complexes that define epithelial organisation. Two proteins with key roles in maintaining barrier function are ZO-1 and JAM-A. Thus, to further investigate the abilities of A. fumigatus supernatants to impair CFBE and HBE epithelial integrity, western blot analysis of two TJ proteins, ZO-1 and JAM-A, was performed. Although 4 h and 8 h supernatants of Af293 and ΔgliG were capable of opening TJs, no reduction in either ZO-1 or JAM-A in CFBE ( Figure 3A) or HBE cells ( Figure 3B) was observed in response to these supernatants. In contrast CFBE cells ( Figure 3A) and HBE cells ( Figure 3B) exposed to 72 h supernatants of Af293 showed reduced levels of ZO-1 and JAM-A TJ proteins. This was not observed in CFBE or HBE cells exposed to the 72 h supernatant of the gliotoxin deficient mutant, ΔgliG. The levels of ZO-1 and JAM-A detected in CFBE cells exposed to the 72 h supernatant of Af293 were significantly lower Error bars represent standard error of at least three independent replicates. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

The 72 h Supernatant of Af293, but Not ∆gliG, Disrupts CFBE and HBE Monolayer TJ Physiology
FITC staining of ZO-1 and JAM-A TJ proteins in untreated CFBE cells (control) showed that ZO-1 ( Figure 4A) and JAM-A ( Figure 4B) are localised primarily in the plasma membranes of CFBE cells. As expected, JAM-A is localised mainly in the plasma membranes of cells but also localises in the cytoplasm of the CFBE cells. The staining of these proteins highlights the network of CFBE cells forming a monolayer and the honeycomb arrangement of TJ proteins, the hallmark of an intact epithelial monolayer. CFBE monolayers exposed to 72 h supernatants of Af293 show a complete delocalisation of ZO-1 ( Figure 4A) and JAM-A ( Figure 4B). Instead, ZO-1 and JAM-A staining is present in the cytoplasm of some cells (arrow) and/ or in vesicles within the cell as demonstrated by punctate staining (arrow) with greatly reduced plasma membrane staining. In contrast, CFBE monolayers exposed to 72 h supernatants of the gliotoxin deficient mutant, ∆gliG, appeared similar to the untreated CFBE monolayers with evident networks of TJs. These results were again mirrored in the HBE monolayers ( Figure 4C,D). These results strongly suggest the role of gliotoxin in TJ disruption and the destruction of the epithelial barrier.
x FOR PEER REVIEW 8 of 15 than in CFBE cells exposed to the 72 h supernatant of ΔgliG (p < 0.01; two-way ANOVA) ( Figure 3A). This result was mirrored in HBE cells ( Figure 3B). GAPDH levels in CFBE and HBE cells exposed to the 72 h supernatant of Af293 were also slightly reduced; therefore, the concentrations of JAM-A and ZO-1 levels were normalised to GAPDH in all the sample treatment scenarios and normalised change ZO-1 and JAM-A highlights the specificity of the effect on TJ proteins.    Bar charts present densitometric analysis of ZO-1 (225 kDa) and JAM-A (36 kDa) bands from western blots to determine relative protein levels of monolayers exposed to Af293 supernatants (black bars) and ∆gliG supernatants (white bars). Band intensities from three independent experiments were normalised against GAPDH values and are expressed as the mean percentage of the control. Error bars represent standard error of at least three independent replicates. ** p < 0.01, **** p < 0.0001.
FITC staining of ZO-1 and JAM-A TJ proteins in untreated CFBE cells (control) showed that ZO-1 ( Figure 4A) and JAM-A ( Figure 4B) are localised primarily in the plasma membranes of CFBE cells. As expected, JAM-A is localised mainly in the plasma membranes of cells but also localises in the cytoplasm of the CFBE cells. The staining of these proteins highlights the network of CFBE cells forming a monolayer and the honeycomb arrangement of TJ proteins, the hallmark of an intact epithelial monolayer.

Discussion
All A. fumigatus isolates were capable of disrupting CFBE TJ integrity by 12 h postinfection, with the reference strain opening the TJs as early as 8 h post-infection. Both clinical A. fumigatus isolates opened the CFBE TJs at a similar rate and all A. fumigatus isolates caused the same level of disruption to TJ integrity by 24 h. Amitani et al. identified conidia within the TJs of cells 6 h post-infection and hyphae penetrating epithelial cells and through the TJs 12 h post-infection of human bronchial mucosal tissue [35]. We observed initial disruptions of TJs of CFBE cells between 8 h and 12 h post-infection, corroborating previous findings in normal bronchial epithelial cells and giving a quantifiable measure of the disruption of CFBE TJs by A. fumigatus using TEER. Interestingly, conidia of the gliotoxin deficient mutant, ∆gliG, were equally capable of disrupting TJ integrity; however, this disruption was delayed. Both the wild type and ∆gliG mutant have been shown to cause similar levels of mortality in an insect infection model [31]. Davis et al. noted that there was no significant difference in the growth rates of the ∆gliG mutant and the wild type [31], ruling this out as a possible cause of the delayed TJ disruption by ∆gliG observed here.
Supernatants of all A. fumigatus isolates in this study were capable of disrupting CFBE TJ integrity. A. fumigatus supernatants as early as 4 h and 8 h were capable of significantly disrupting TJ integrity. The effects of these early supernatants are unlikely to be due to gliotoxin as the production of gliotoxin by A. fumigatus is known to be minimal before 24 h growth [36]. Additionally, we observed that even the 4 h and 8 h supernatants of the gliotoxin mutant were capable of disrupting TJs to a degree. Other groups have shown spore diffusates of A. fumigatus to inhibit phagocytosis by alveolar macrophages [37] and neutrophils [38]. These findings raise questions regarding the damage that early and shortterm A. fumigatus colonisation may exert on the CF airways. The component/s of early A. fumigatus supernatants capable of disrupting TJs were not identified here and this warrants further investigation.
Overall, the 72 h supernatants of A. fumigatus isolates caused the greatest disruption to CFBE TJ integrity, and this effect was induced as early as 2 h post-treatment. A. fumigatus allergens and proteases have been reported to cause damage to epithelial barriers [35,39]. Amitani et al. investigated the effect of 7-day A. fumigatus supernatants on human nasal epithelial cells (HNCEs) from 14 different A. fumigatus clinical isolates [27]. They reported that A. fumigatus supernatants caused significant damage to HNCEs and slowed ciliary beat frequency [27]. However, the secretions of proteases are known to be minimal in the ATCC Af293 strain prior to day 4 of culture in MEM [40]. Considering that the 72 h supernatants in this study caused the most rapid and significant reduction in TJ integrity, we investigated whether gliotoxin, a secondary metabolite of A. fumigatus produced in highest quantities at 72 h growth [32,41], was capable of disrupting TJ integrity in CFBE monolayers. We found that commercially available gliotoxin at concentrations similar to those detected from BAL fluid of CF patients colonised with A. fumigatus [4] was capable of significantly reducing CFBE TER. Gliotoxin is known to have numerous immunoevasive effects including slowing ciliary beat frequency [27], inhibiting B and T lymphocyte proliferation [42] and hindering the ability of macrophages and neutrophils to phagocytose and kill microorganisms [43]. Gliotoxin has also been shown to significantly down-regulate the expression of the Vitamin D receptor in CFBE and HBE cell lines [4]. Khoufache et al. stated that gliotoxin isolated from A. fumigatus supernatants did not cause a decrease in the TEER of nasal epithelium [44]; however, we have shown that gliotoxin did significantly decrease the TEER of HBE and CFBE monolayers. To confirm the involvement of gliotoxin in disrupting CFBE monolayers, we employed a gliotoxin mutant strain of A. fumigatus which has been shown to be deficient in gliotoxin production by liquid chromatography-mass spectrometry (LC-MS) and reversed phase-high performance liquid chromatography (RP-HPLC) analysis of supernatants [31]. The 72 h supernatants of ∆gliG did not impair CFBE TJs, strongly suggesting that gliotoxin contributes to the breach of CFBE TJs caused by the 72 h supernatant of A. fumigatus. This has not been reported before. In fact, apart from one recent study showing that gliotoxin modulates actin cytoskeleton rearrangement to facilitate A. fumigatus internalisation into lung epithelial cells [45], there is a dearth of studies investigating the impact of gliotoxin on airway epithelia. Interestingly, we found that CFBE monolayers were more susceptible to gliotoxin than HBE monolayers. CF respiratory epithelia have been shown to respond differently than normal respiratory epithelium to microbial exposure [35] and it has been reported that the normal organisation and function of TJs is disturbed by the localisation of delF508-CFTR to the cytoplasm and retention within the endoplasmic reticulum [46]. Furthermore, CFTR has been shown to colocalise with ZO-1 at the TJ to regulate TJ assembly [47].
ZO-1 is a peripheral membrane protein found on the cytoplasmic surfaces of epithelial cell membranes and has been suggested to have both structural and signaling roles in the TJ [48]. JAM-A is a member of the immunoglobulin super family expressed in the TJs of epithelial cells and has been reported to play a role in leukocyte migration, cell-cell adhesion, homophilic interactions [25] and TJ barrier regulation [49]. Both TJ proteins are integral to the maintenance of epithelial barriers and their delocalisation is associated with TJ disruption. The western blot and confocal evidence presented here revealed that CFBE cells exposed to 72 h supernatants of Af293 had markedly delocalised ZO-1 and JAM-A. In contrast, CFBE cells exposed to the 72 h supernatants of the ∆gliG mutant exhibited no changes in ZO-1 and JAM-A levels and their localisation was similar to that of the control untreated monolayers. Given the previously reported abilities of gliotoxin to interfere with protein activity [50], and translocation [29,30], it is not surprising that gliotoxin can interfere with TJ proteins and epithelial barrier integrity. It is likely that the disruption of airway epithelial barriers by A. fumigatus is multifactorial with host factors playing a significant role in the ability of A. fumigatus to exert these effects in vivo. The breakdown of TJ integrity in the CF airway has obvious implications. Wan et al. demonstrated that a breakdown in TJ integrity in response to the house dust mite allowed migration of Der p allergens outside of the respiratory epithelium which could initiate an inflammatory response [51]. It has been hypothesised that disruption of airway epithelial TJ integrity could permit allergens access to the underlying capillaries [52,53].
The CF airway is host to a diverse microbiome [18] and interactions within the airway microbiome influence virulence, antimicrobial resistance and disease progression [54,55]. A. fumigatus and P. aeruginosa co-infections contribute to increased pathogenicity in a Galleria mellonella model and these co-infecting microbes contribute to an altered inflammatory response in CFBE cells [56]. Gliotoxin has been shown to cause apoptosis in human lung epithelial cells [57] and here we have shown that it can impair the integrity of the epithelial barrier. Therefore, damage of the epithelial barrier by A. fumigatus secretions could allow the translocation of A. fumigatus conidia (and other species in the microbiome), penetration of hyphae and the potential release of allergens into the blood stream which may exacerbate or initiate an allergic reaction such as ABPA. The evidence reported here, alongside the building body of evidence in the literature, reaffirms that A. fumigatus is not an innocent bystander in the chronic airway and that colonisation of the airways in the absence of ABPA has the potential to cause silent damage to the epithelium.
No studies have investigated the effects of A. fumigatus supernatants or gliotoxin on the integrity of bronchial epithelium with the delF508 mutation associated with CF. We report here that both reference and clinical strains of A. fumigatus can disrupt the TJs of epithelial monolayers. We show for the first time that A. fumigatus supernatants disrupt CFBE TJ integrity and we have provided evidence that gliotoxin in later supernatants is contributing to this effect. A. fumigatus can persistently colonise the CF airways ( Figure S1) and gliotoxin has been detected in the airways of CF patients asymptomatically colonised with A. fumigatus [4]. Given the findings presented here, the exposure of CF airway epithelium to A. fumigatus secretions and gliotoxin can compromise the integrity of the epithelial barrier and potentially enhance the invasion of microorganisms in the lungs of people with CF.