Fungal Aflatoxins Reduce Respiratory Mucosal Ciliary Function

Aflatoxins are mycotoxins secreted by Aspergillus flavus, which can colonize the respiratory tract and cause fungal rhinosinusitis or bronchopulmonary aspergillosis. A. flavus is the second leading cause of invasive aspergillosis worldwide. Because many respiratory pathogens secrete toxins to impair mucociliary immunity, we examined the effects of acute exposure to aflatoxins on airway cell physiology. Using air-liquid interface cultures of primary human sinonasal and bronchial cells, we imaged ciliary beat frequency (CBF), intracellular calcium, and nitric oxide (NO). Exposure to aflatoxins (0.1 to 10 μM; 5 to 10 minutes) reduced baseline (~6–12%) and agonist-stimulated CBF. Conditioned media (CM) from A. fumigatus, A. niger, and A. flavus cultures also reduced CBF by ~10% after 60 min exposure, but effects were blocked by an anti-aflatoxin antibody only with A. flavus CM. CBF reduction required protein kinase C but was not associated with changes in calcium or NO. However, AFB2 reduced NO production by ~50% during stimulation of the ciliary-localized T2R38 receptor. Using a fluorescent reporter construct expressed in A549 cells, we directly observed activation of PKC activity by AFB2. Aflatoxins secreted by respiratory A. flavus may impair motile and chemosensory functions of airway cilia, contributing to pathogenesis of fungal airway diseases.

. Acute exposure to AFB 2 and AFB 1 decreased basal sinonasal CBF in a PKC-dependent manner.
(a) Structures of aflatoxins B 1 and B 2 (AFB 1 and AFB 2 ). (b-e) Mean traces of CBF normalized to baseline (n = 3-6 cultures from separate patients each) during stimulation with vehicle (DMSO) alone (b), AFB 2 (5 min exposure for each concentration; c), AFB 2 + Gö6983 (d), AFB 1 (e; purple trace), and AFB 1 + Gö6983 (e; green trace). Normalized CBF was 0.99 ± 0.01, 1.0 ± 0.01, and 1.0 ± 0.01 after 5 min application of 0.001%, 0.01%, and 0.1% DMSO, respectively. After 5 min application of AFB 2 , CBF decreased to 0.94 ± 0.01 (0.1 μ M AFB 2 ; n.s. pathogenic Aspergillus, A. flavus is the second-leading cause of invasive aspergillosis 20,21 . A. flavus infection is rare in the US and Europe. However, bronchiopulmonary and sinonasal apergillosis from A. flavus is common in India, Africa, South East Asia, and the Middle East, possibly due to an increased ability of A. flavus to thrive in arid conditions 20 . A. flavus in the upper respiratory tract is often associated with chronic granulomatous sinusitis. A. flavus is of importance because it produces aflatoxins, which are among the most potent naturally-occuring hepatic carcinogens known 19 . Ingestion of contaminated foods results in metabolism ("activation") of aflatoxins in the liver into reactive DNA-damaging epoxides that cause hepatic necrosis, cirrhosis, and/or carcinoma 22,23 .
Inhalation of aflatoxins has been associated with occupations involving exposure to environmental molds 24 , such as grain processing. However, the effects of inhaled aflatoxins or aflatoxin-producing fungi on the airway epithelium are not well characterized. There is some evidence that airway cells can activate aflatoxins in vitro 4,25 and in vivo [26][27][28] , though the link between aflatoxin exposure and human lung cancer is unclear. However, aflatoxins can increase protein kinase C (PKC) activity in some cell lines in vitro [29][30][31] . Because PKC can decrease CBF 32,33 through phosphorylation of ciliary proteins 32,33 , we hypothesized that aflatoxins may have acute effects on MCC that contribute to A. flavus pathogenesis.

Aflatoxin B 2 Decreases CBF in a PKC-Dependent Manner.
We examined epithelial responses to a common aflatoxin, aflatoxin B 2 (AFB 2 ). AFB 2 (Fig. 1a) was chosen as the model aflatoxin for testing in this study because it has less carcinogenicity than AFB 1 34,35 and thus should have less DNA-damaging nonspecific toxic effects. We utilized air-liquid interface cultures (ALIs) derived from human sinonasal and bronchial epithelial cells 36 . ALIs mimic the polarized respiratory epithelium with well differentiated ciliated and goblet cells 37 . Highspeed imaging was used to track changes in CBF. Acute mucosal exposure (apical side only) of sinonasal ALIs to AFB 2 (1 μ M and 10 μ M) significantly decreased basal CBF after only 5 minutes, while vehicle (DMSO) had no effect (Fig. 1b,c). The protein kinase C (PKC) inhibitor Gö6983 38 (10 μ M; 5 min apical pre-treatment before experiment) significantly blunted the AFB 2 -mediated inhibition (Fig. 1d). AFB 1 had nearly identical effects (Fig. 1e). Results are summarized in Fig. 1f,g. No additive effects of AFB 2 were observed when sinonasal ALIs were pretreated with the phorbol ester phorbol-12-myristate-13-acetate (PMA), nor was CBF further reduced when PMA was added to ALIs pre-treated with AFB 2 (Fig. 1h), supporting the hypothesis that AFB 2 reduces CBF through a PKC-dependent pathway.
We noted that short-term exposures to AFB 2 and AFB 1 also impaired activation of CBF in response to the purinergic agonist ATP (Fig. 1b,d), an important signaling molecule in the airway 33 . We thus carried out a detailed examination of the effects of AFB 2 on stimulated CBF using several physiologically important agonists after 10 min exposure to 0.5 μ M AFB 2 . AFB 2 inhibited CBF during stimulation with 1 μ M ATP (added apically; Fig. 2a), 10 μ M isoproterenol (added apically; Fig. 2b), and 10 μ M VIP (added basolaterally; Fig. 2c). CBF reductions (summarized in Fig. 2d) were blocked by the PKC inhibitors Gö6983 and calphostin C 38 . AFB 2 exposure also reduced CBF increases in response to a mechanically-simulated "sneeze" (Fig. 2e), which stimulates CBF through apical ATP release and downstream calcium signaling 36 .
When we examined ALIs grown from human bronchial epithelial (HBE) cells, we found that AFB 2 similarly reduced both basal and ATP-stimulated CBF via a PKC-dependent mechanism (Fig. 3). Interestingly, when we examined ALI cultures derived from mouse nasal septum, we found that AFB 2 inhibited basal CBF but not ATP-stimulated CBF ( Supplementary Fig. S1), reflecting a species-specific difference. AFB 2 Acts Independently of Calcium. Calcium is a master regulator in airway cells, controlling both ion transport 39 as well as CBF 32 . Data above show that AFB 2 reduces CBF in human ALIs in response to both ATP and the sneeze puff, which both require intracellular calcium, as well as VIP, which acts independently of calcium through cyclic AMP (cAMP) in these cells 40 . Thus, we hypothesized that AFB 2 likely has direct effects on cilia function, possibly through PKC phosphorylation of cilia proteins, as previously described 33 , rather than by indirectly affecting calcium levels. However, because many isoforms of PKC are regulated by calcium 33 , we examined if AFB 2 affects baseline or stimulated calcium signaling. We examined changes in intracellular calcium concentration in sinonasal ALIs loaded with the calcium-sensitive indicator fluo-4 during exposure to 10 μ M AFB 2 . AFB 2 had no detectible effect on intracellular calcium, nor did it affect the magnitude or kinetics of ATP-induced calcium signaling ( Supplementary Fig. S2), supporting the hypothesis that AFB 2 activates PKC independently of calcium. compared with vehicle), 0.90 ± 0.02 (1 μ M AFB 2 ; P < 0.05 compared with vehicle) and 0.87 ± 0.02 (10 μ M AFB 2 ; P < 0.01 vs vehicle). In the presence of Gö6983, CBF with AFB 2 was 0.99 ± 0.01 (0.1 μ M AFB 2 ), 0.96 ± 0.02 (1 μ M AFB 2 ), and 0.96 ± 0.03 (10 μ M AFB 2 ; P < 0.05 vs 10 μ M AFB 2 alone; n.s. vs vehicle) after 5 minutes. With AFB 1 , CBF decreased to 0.96 ± 0.01 (0.1 μ M AFB 1 ; n.s. compared with vehicle), 0.92 ± 0.01 (1 μ M AFB 1 ; P < 0.05 vs. vehicle), and 0.89 ± 0.01 (10 μ M AFB 1 ; P < 0.05 vs. vehicle). With AFB 1 in the presence of Gö6983, CBF was 0.99 ± 0.01 (0.1 μ M AFB 1 ; n.s. compared with vehicle), 0.98 ± 0.01 (1 μ M AFB 1 ; n.s. compared with vehicle) and 0.98 ± 0.02 (10 μ M AFB 1 ; n.s. compared with vehicle). (f,g) Plot of Normalized CBF after 5 min vs log AFB 2 (f) and AFB 1 (g) Data points for 0.01 μ M AFB 1 and AFB 2 were from separate experiments (not shown). Asterisks denote significance vs. DMSO alone (vehicle control). All significances determined by 1-way ANOVA with Bonferroni post-test; *P < 0.05, **P < 0.01. (h) Additive effects on CBF were not observed between the PKC-activator PMA (1 μ M) and AFB 2 .
Scientific RepoRts | 6:33221 | DOI: 10.1038/srep33221 AFB 2 Acts Independently of the Y 2 Neuropeptide Y Receptor. Neuropeptide Y (NPY) is one of the few neurotransmitters known to reduce CBF through Y 2 receptor activation of PKC in primary human tracheal and bronchial ciliated cells 41 . In sinonasal ALIs, NPY decreased basal CBF by ~10% through a mechanism blocked by both the Y 2 antagonist BIIE-0246 and Gö6983 (Fig. 4a,c). CBF was also reduced by the Y 2 agonist NPY-(16-36) but not the Y 1 agonist [Leu 31 ,Pro 34 ]-NPY (Fig. 4b,c). No additive effects were observed when NPY was added after AFB 2 (Fig. 4d), suggesting they partially share the same pathway. However AFB 2 reduction of CBF was not blocked by BIIE-0246, the broad spectrum neuropeptide receptor inhibitor antagonist G 42 , or the phospholipase C inhibitor U73122 (Fig. 4e). These data demonstrate that AFB 2 functions independently of the Y 2 receptor and likely other neurotransmitter receptors.
Scientific RepoRts | 6:33221 | DOI: 10.1038/srep33221 25% ( Fig. 5a,b) and confirmed in bronchial ALIs at 25% (Fig. 5c,d). A. flavus CM significantly reduced baseline CBF after 60-75 min and significantly blunted ATP-induced CBF increase (Fig. 5b,d). While the kinetics of the CM-induced reduction in CBF was slower than observed with purified aflatoxin, these effects were nonetheless blocked by Gö6983 as well as when the A. flavus CM was pre-treated with anti-aflatoxin antibodies (recognizing both B and G group aflatoxins). These data strongly suggest that cultured A. flavus secretes aflatoxins at low concentrations that are nonetheless high enough to reduce airway CBF. We observed that CM from A. fumigatus and A. niger, which cannot secrete aflatoxins, was still observed to reduce CBF ( Supplementary Fig. S3). The effects of A. fumigatus and A. niger CM were blocked by Gö6983 but not by anti-aflatoxin antibodies ( Supplementary  Fig. S3), suggesting that these species secrete other mycotoxins that can target PKC, perhaps including gliotoxin, fumagillin, and/or helvoilic acid. The identities and mechanisms of action of A. fumigatus and A. niger ciliotoxins remain to be determined in future studies.

AFB 2 impairs sinonasal epithelial nitric oxide (NO) innate immune responses. Nitric oxide (NO)
is an important mediator of host airway defense because it directly kills pathogens as well as increases CBF 43,44 . We recently showed that a bitter taste receptor, T2R38, is expressed in sinonasal epithelial cilia and drives NO production in response to bacterial acyl-homoserine lactone (AHL) quorum sensing molecules [43][44][45] . Because PKC can phosphorylate nitric oxide synthase (NOS) and prevent its activation 46,47 , we tested the effects of AFB 2 on sinonasal NO production in response to the T2R38 agonist and Pseudomonas quorum sensing molecule N-3-oxo-dodecanoyl-L-homoserine lactone (C12HSL) 43 . Reactive nitrogen species (RNS) production was measured using the fluorescent indicator DAF-FM. RNS production was reduced by approximately one half in the presence of AFB 2 , and this effect was blocked by Gö6983 (Fig. 6a,b). To test if AB 2 -induced PKC activity had a general effect on NOS function or a specific effect on T2R38 function, we measured RNS production during global calcium elevation in cells exposed to the calcium ionophore ionomycin (10 μ M) and the sarco/endoplasmic reticulum calcium ATPase (SERCA) inhibitor thapsigargin (10 μ M). AFB 2 also significantly reduced NO production under these conditions through a Gö6983-sensitive pathway (Fig. 6c,d), suggesting AFB 2 has a direct effect on NOS activation rather than on T2R38 function.

AFB 2 activates PKC in A549 cells in vitro.
To further test the hypothesis that aflatoxins can activate PKC activity, we utilized a Förster resonance energy transfer (FRET)-based PKC construct, CKAR 48,49 . Because AFs activate PKC in a variety of cell types 29-31,50 , we hypothesized that the mechanism of activation was not cell-type dependent. As primary sinonasal ALIs are very difficult to transfect, even with viral systems, CKAR was transfected into A549 cells, a commonly used lung epithelial cell line. CKAR contains the FHA2 domain of RAD53p as well as a PKC phosphorylation sequence designed to be phosphorylated by all PKC isoforms. These sequences are flanked by an eCFP the N-terminus and a citrine YFP variant on the C-terminus. When phosphorylated, the substrate sequence binds the FHA2 phospho-peptide-binding domain, resulting in a conformational change that keeps the CFP and YFP further apart, reducing FRET emission. Thus, a decrease in FRET emission correlates with an increase in PKC activity, and vice verse. This conformational change is reversible by phosphatases. Single transfected cells were imaged by conventional wide-field low-light-level microscopy, collecting light at three wavelengths: 1) CFP excitation, CFP emission, 2) CFP excitation, YFP emission, and 3) YFP excitation, YFP emission. Data are reported as the signal of wavelength 2 divided by wavelength 1 (i.e. the yellow/cyan emission ratio at cyan excitation, or FRET/CFP ratio) as previously described 48 .
Application of 10 μ M AFB 2 caused a decrease in CKAR FRET/CFP ratio that was reversible by addition of Gö6983 in the continued presence of AFB 2 (Fig. 7a-c). As a control, A549 CKAR-transfected cells treated with PMA exhibited a fast decrease in CKR FRET that was reversed with application of Gö6983 (1 μ M) in the continued presence of PMA (Fig. 7c). Application of forskolin, an activator of adenylate cyclase, had no effect on CKAR fluorescence, as previously described 48 (Fig. 7c). These results strongly support the hypothesis that AFB 2 exposure increases PKC activity.

Discussion
The average person inhales hundreds to thousands of airborne Aspergillus spores daily 10 . In immune-competent individuals, these fungi are typically cleared without consequence. However, in individuals with impaired respiratory defenses (e.g., patients with CRS, diabetes, CF or otherwise immunocompromised), fungal infection can be a significant or even a fatal complication 9 . Understanding the effects of mycotoxins on the respiratory epithelium is important for understanding the pathogenesis of respiratory (upper and lower) aspergillosis. Here, we show that a class of Aspergillus mycotoxins, aflatoxins, can slow basal and stimulated respiratory CBF, potentially enhancing  4,52,53,56,57 . However, it must also be determined how environmental aflatoxin exposure, often measured in ppm of aflatoxin-contaminated dust, actually translates to concentrations seen by the airway epithelial cells. As exposure often occurs through contaminated dust, airway deposition will be affected by particle size and sinonasal airflow patterns 63 . This would be further confounded by the fact that the most commonly used animal models, such as mice, have significantly different paranasal sinus anatomy than humans 64,65 . Sampling of airway surface liquid and mucus from patients with respiratory A. flavus infections may shed light on concentrations of aflatoxins generated during active A. flavus infection. Moreover, since a significant amount of aflatoxin contamination occurs in grain-based livestock and pet foods [66][67][68] , inhalation of aflatoxin-contaminated dust may also be a contributor to respiratory infection in non-human animals as well. Antibiotic use in animals is a major driver for the emergence of resistant pathogenic microorganisms 69 . Further studies of aflatoxin exposure levels in at-risk (c) Average traces (mean ± SEM) of CKAR FRET/CFP ratio during stimulation with forskolin and PMA (left) and AFB 2 . Traces are the average of 4 (forskolin/PMA) and 9 (AFB 2 ) experiments. Bar graph to the right shows % change in CKAR FRET/CFP ratio, which was 4.3 ± 0.2% with PMA, 3.4 ± 0.1% with AFB2, and 0.2 ± 0.09% with forskolin. Significances determined by 1-way ANOVA with Bonferroni post test; *P < 0.05 vs control.
humans and animal models are critically needed to help complete our understanding of the consequences of both acute and chronic aflatoxin respiratory exposure in humans an animals.
Coupled with previous data that Aspergillus gliotoxin, fumagillin, and helvoilic acid 17,18 slow CBF, the current data emphasize that Aspergillus have evolved an armament of mycotoxins to impair MCC and reduce host innate defense. Our data also show that A. niger and A. fumigatus, which cannot secrete aflatoxins, nevertheless secrete mycotoxins that also activate PKC. Certain PKC isoforms have also been linked to inflammation 70 and apoptosis 71 , and thus chronic exposure to aflatoxins and other Aspergillus mycoctoxins may stimulate these processes, exacerbating epithelial damage. PKC inhibitors have been proposed as therapeutics for inflammatory diseases 70 . Our study suggests that PKC inhibitors may also have potential for fungal airway diseases by relieving mycotoxin-induced repression of ciliary beating. Moreover, the ability of AFB 2 to impair sinonasal NO production in response to bacterial AHL-stimulation of T2R receptors suggests that aflatoxins may play an important role in the generation of mixed fungal and bacterial biofilms sometimes observed in airway diseases 72 . Because we have shown that reduced T2R38 function correlates with gram-negative bacterial infection 43 , risk of chronic rhinosinusitis 73,74 , and surgical outcomes in non-polypoid chronic rhinosinusitis 75 , exposure to aflatoxins and resulting reduction in downstream components of the T2R38 pathway may have important implications for all of these clinical parameters. Moreover, the ability of aflatoxins to impair ciliary activity may have likewise profound clinical consequences during pulmonary aspergillosis caused by A. flavus 76 .
In conclusion, exposure of ciliated respiratory epithelial cells to AFB 2 resulted in a decrease in both baseline and stimulated CBF through calcium-independent activation of PKC. AFB 2 also impaired sinonasal epithelial cell bitter taste receptor-driven NO innate immune responses to gram-negative bacterial quorum sensing molecules. These results suggest that aflatoxins may impair MCC and other innate defense pathways, enhancing the pathogenicity of A. flavus and possibly other co-infecting pathogens as well. In addition to their anti-inflammatory effects, PKC inhibitors may be potential therapeutics for fungal respiratory diseases due to their ability to counteract mycotoxin-induced decreases in ciliary beating and MCC.

Materials and Methods
All experimental protocols were reviewed and approved by the Research and Development Committee at the Philadelphia Veterans Affairs Medical Center and were carried out in accordance with both The University of Pennsylvania and The Philadelphia VA Medical Center guidelines regarding use of residual clinical material in research.

Reagents and solutions.
Unless indicated, all reagents and solutions were as previously described 40,43,77,78 .

Generation of sinonasal ALI cultures.
Patients undergoing sinonasal surgery were recruited from the Department of Otorhinolaryngology at the University of Pennsylvania and the Philadelphia Veterans Affairs Medical Center with full approval of both Institutional Review Boards (Penn#800614, PVAMC#00781) and written informed consent was obtained for all participating patients in accordance with the U.S. Department of Health and Human Services code of federal regulation Title 45 CFR 46.116. Exclusion criteria included a history of systemic diseases (e.g., Wegner's, Sarcoid, CF), immunodeficiences, or use of antibiotics, oral corticosteroids, or anti-biologics (e.g. Xolair) within one month of surgery. Human sinonasal epithelial cells were enzymatically dissociated grown to confluence in proliferation medium (DMEM/Ham's F-12 plus BEBM; Clonetics, Cambrex, East, NJ, USA) for 7 days as previously described 37,43 . Confluent cells were dissociated and seeded on porous polyester membranes coated with BSA, type I bovine collagen, and fibronectin in cell culture inserts in LHC basal medium (Invitrogen). Culture medium was removed from the upper compartment and basolateral media was changed to differentiation medium (1:1 DMEM:BEBM) containing hEGF (0.5 ng/ ml), epinephrine (5 g/ml), BPE (0.13 mg/ml), hydrocortisone (0.5 g/ml), insulin (5 g/ml), triiodothyronine (6.5 g/ml), and transferrin (0.5 g/ml), supplemented with 100 U/ml penicillin, 100 g/ml streptomycin, 0.1 nM retinoic acid, and NuSerum (BD Biosciences, San Jose, CA) as previously described 37,43 .

Measurement of ciliary beat frequency (CBF).
Whole-field CBF was measured using the Sisson-Ammons Video Analysis system 79 as previously described 36,40,43 at ~28-30 °C. Cultures were imaged using at 100 frames/second using a Leica Microscope (20x/0.8NA objective) with Hoffman modulation contrast. Experiments utilized Dulbecco's PBS (1.8 mM calcium) on the apical side and HEPES-buffered Hank's Balanced Salt Solution supplemented with 1× MEM vitamins and amino acids on the basolateral side.
Calcium and nitric oxide (NO) imaging. Calcium and NO were imaged using the Fluo-4 and DAF-FM, respectively, as previously described 36,43,77,78 . Cultures were loaded with Fluo-4 AM (10 μ M applied apically) for 2 hrs followed by washing and 20 min incubation in the dark. Cultures were similarly loaded with 10 μ M DAF-FM diacetate for 90 min in the presence of 5 μ M carboxy-PTIO, followed by washing to remove unloaded DAF-FM and cPTIO and incubation for 15 minutes prior to imaging. Imaging was performed using an Olympus Fluoview confocal system with IX-81 microscope and 10x (0.3 NA UPlanFLN) objective. Images were analyzed using Fluoview software as previously described 36 A549 cell culture, transfection, and CKAR FRET imaging. A549 cells were obtained from American Type Culture Collection (ATCC, Manassas, VA) and cultured in Kaighn's modification of Ham's F12 media (F12K) with 10% fetal bovine serum and 1x penicillin-streptomycin mix (Gibco/Thermo Fisher Scientific, Waltham, MA). Cells were used at passage 15-20. Cells were transfected with CKAR 48,49 (Alexandra Newton, University of California San Diego, Addgene, Cambridge, MA, plasmid#14860) by standard calcium phosphate transfection in a 150 mm dish at ~75% confluency. The day after transfection, cells were trypsinized and re-plated into chambered coverglass wells (CellVis, Mountain View, CA) at 50% confluency. Cells were used at 48 hrs after transfection, and cells media was replaced with 10 mM HEPES-buffered Hank's balanced salt solution (HBSS) and imaging was performed at room temperature on the stage of an Olympus IX-83 inverted microscope (60x PlanApo 1.4 NA oil-immersion objective; Olympus Life Sciences, Tokyo, Japan) equipped with excitation and emission filter wheels (Sutter Instruments, Novato, CA) and a CFP-YFP FRET filter set (89003-ET, Chroma Technologies, Rockingham, VT). Images were acquired (12 sec intervals) and analyzed using MetaFluor (Molecular Devices, Sunnyvale, CA) and ratio images were constructed using ImageJ (W. Rasband, National Institutes of Mental Health, Research Services Branch, Bethesda, MD). Both background (estimated using an off-cell area) and baseline drift were subtracted as described 49 before averaging of traces.
Data analysis and statistics. One-way analysis of variance (ANOVA) was performed in GraphPad Prism with appropriate post-tests as indicated; P < 0.05 was considered statistically significant. All other data analysis was performed in Excel. For all figures, one (*) and two (**) asterisks indicate P < 0.05 and P < 0.01 respectively; "n.s." indicates no statistical significance. All data are mean ± SEM.