Hank’s balanced salt solution (HBSS) used for live cell imaging and other experiments contained (in mM): 138 NaCl, 5.3 KCl, 0.34 Na₂HPO₄, 0.44 KH₂PO₄, 0.49 MgCl₂, 0.41 MgSO₄, 1.3 CaCl₂, 5.6 glucose, 20 HEPES free acid. HBSS was adjusted to pH 7.4 with NaOH. When 1x MEM amino acids was added to the HBSS, pH was adjusted after addition of the amino acids. Dulbecco’s phosphate-buffered saline (DPBS) contained (in mM): 140 NaCl, 10 mM NaH₂PO₄, 1.5 KH₂PO₄, 2.7 KCl, 1.8 CaCl₂, 1.5 MgCl₂, pH 7.4. Proper osmolarity of solutions was routinely verified using a vapor pressure osmometer (Wescore 5520). Apigenin (cat # 10010275), L-NAME (cat # 80210), D-NAME (cat #21687), SC79 (cat # 14972), N-3-oxo-dodecanoyl-L-Homoserine lactone (3oxoC12HSL; cat # 10007895), diphenhydramine (cat #11158), S-Nitroso-N-Acetyl-D,L-Penicillamine (SNAP; cat # 82250), quinine (cat # 23958), denatonium benzoate (cat #26350), phenylthiocarbamide (PTC; cat #30855), VX-770 (ivacaftor; cat #15145), VX-809 (lumacaftor; cat #22196), CFTR inhibitor 172 (CFTR_inh172; cat #15545), colistin sulfate (cat #17584), flufenamic acid (FFA; cat #21447), MK2206 (cat #11593), GSK690693 (cat #16891), and forskolin (cat# 11018) were from Cayman Chemical. DAF-FM diacetate (cat # D23844), fura-2 acetoxymethylester (fura-2 AM; cat #F1221), and SPQ (6-Methoxy-N-(3-Sulfopropyl)Quinolinium, Inner Salt; cat #M440) were from ThermoFisher Scientific. If not stated otherwise below, all other reagents were acquired from MilliporeSigma. Antibodies used are listed below in the immunofluorescence microscopy methods section.

Immortalized cells were grown in minimal essential media (MEM) with Earl’s salts (ThermoFisher Scientific) plus 1x cell culture penicillin/streptomycin (Gibco) and 10% FetalPlex serum substitute (Gemini Biosciences cat #100602/500) at 37°C with 5% CO₂. Parental CFBE41o- cells (F508del CFTR homozygous (64) and CFBE41o- cells stably expressing either Wt CFTR or F508del CFTR with puromycin resistance (65, 66) were a gift from Dr. R.C. Rubenstein (Division of Allergy, Immunology and Pulmonary Medicine, Washington University in St. Louis). Parental 16HBE cells were originally from Dr. D. C. Gruenert (University of California, San Francisco). CRISPR-modified 16HBE14o- cells homozygous for G542X or F508del CFTR (67) were obtained from the Cystic Fibrosis Foundation under a materials transfer agreement. Identity was confirmed by in-house sequencing (University of Pennsylvania Department of Genetics Next-Generation Sequencing Core). All cells were used within 12 passages of receipt. Submerged cells were grown on uncoated cell culture-treated plastic T75 flasks (Corning) and were split 1:5 using 0.25% trypsin + EDTA for 10 min when ~70% confluence was reached.

For air-liquid interface (ALI) cultures, 16HBE14o- or CFBE41o- cells were seeded onto collagen-coated 0.33 cm² transwell filters (24-well plate size) at ~50% confluency and grown to confluence for 5 days before apical air exposure, as performed previously (68, 69). Media from the apical side was removed by aspiration and cells were fed only from the basolateral side with 500 µL of MEM + Fetalplex + penicillin/streptomycin. At the time of air exposure, media was removed from the apical side by aspiration and basolateral media was changed from MEM + 10% Fetalplex to a differentiation media containing 1:1 Lonza bronchial epithelial cell basal media (BEBM):DMEM plus Lonza singlequot supplements (B-ALI™ Bronchial Air-Liquid Interface Medium BulletKit™ cat #00193514; 0.5 ng/ml hEGF, 5 ng/ml epinephrine, 0.13 mg/ml BPE, 0.5 ng/ml hydrocortisone, 5 ng/ml insulin, 6.5 ng/ml triiodothyronine, and 0.5 ng/ml transferrin, 0.1 nM retinoic acid) supplemented with 1x penicillin/streptomycin and 2% NuSerum (Corning cat #355500) as previously used to culture these cells at ALI (22). Cells were fed from the basolateral side only with the differentiation media for ~21 days before use. 16HBE and CFBE tight junction formation was confirmed by measurement of transepithelial resistance (Epithelial Volh-Ohm Meter, World Precision Instruments). Before imaging experiments, both apical and basolateral sides of the cultures were washed with HBSS to remove residual media containing phenol red, which can interfere with fluorescence measurements.

Primary human nasal epithelial cells were isolated from surgical specimens according to The University of Pennsylvania guidelines regarding use of residual clinical material. Tissue was obtained from patients undergoing sinonasal surgery at the Hospital of the University of Pennsylvania under institutional review board approval (#800614) with written informed consent in accordance with the U.S. Department of Health and Human Services code of federal regulation Title 45 CFR 46.116 and the Declaration of Helsinki. Inclusion criteria were patients ≥18 years of age requiring surgery for sinonasal disease or trans-nasal approaches to the skull base. Exclusion criteria included systemic inheritable disease (e.g., granulomatosis with polyangiitis, systemic immunodeficiences) or use of antibiotics, oral corticosteroids, or anti-biologics (e.g., Xolair) within one month of surgery. Vulnerable populations (patients ≤18 years of age, pregnant women, and cognitively impaired persons) were not included.

Tissue was transported to the lab on ice in saline. Mucosal tissue (1-2 mm strips) was washed with HBSS and then immediately removed for enzymatic dissociation with 1.4 mg/ml pronase, 0.1 mg/ml DNAse I, and 1x pen/strep in MEM for 1 hr at 37°C. Digestion was stopped by addition of 20% FBS for 3 min. Tissue was filtered through a 70 µm nylon mesh strainer (Fisher Scientific) followed by centrifugation at 500x g 5 min. Dissociated cells were resuspended in proliferation media (Bronchial Epithelial Cell Growth Media, BEGM™ Bronchial Epithelial Cell Growth Medium BulletKit cat #CC-3170; Lonza) and incubated in a T-25 culture flask for 1-2 hours at 37°C to remove any contaminating fibroblasts, macrophages, PMNs, lymphocytes, etc., which adhere before the epithelial cells. Supernatant (containing epithelial cells) was removed, centrifuged again, resuspended in proliferation media, and placed in fresh T25 flask. A confluence of ~80% was typically reached within one week. At this time, cells were dissociated, and seeded at high density (~90-100% confluence) on Corning transwells coated with type I bovine collage, fibronectin, and bovine serum albumin (22, 24, 69, 70). Culture medium was removed from the upper compartment the next day by aspiration, and basolateral media was changed to the differentiation medium as described above for 16HBE and CFBE cells. Cultures were genotyped for TAS2R38 PAV or AVI polymorphims (50, 71) as described (22, 24, 69) using a restriction digest protocol. We obtained cDNA from small tissue samples from cells that were being cultured; cDNA was digested with FNu4HI (New England Biolabs; cat # R0178S) for 1hr at 37°C, and run on an agarose gel. The presence of bands at 364 bp and 531 bp demonstrated presence of T2R38 AVI while bands at 337 and 465 bp revealed presence of T2R38 PAV, with both sets of bands occuring in heterozygotes. PAV/AVI heterozygote cells were primarily used in experiments with T2R14 agonists like quinine, apigenin, or diphenhydramine, as we have previously found these agonist responses to be independent of TAS2R38 genotype (69, 72, 73). Moreover, this preserved adequate numbers of PAV/PAV CF and non-CF cultures for PTC and 3oxoC12HSL experiments, where responses could be tested between PAV/PAV vs AVI/AVI non-CF cultures to ensure responses observed correlated with T2R38 functionality.

CF cells used here were identified as homozygous for phenyalanine 508 deletion (F508del/F508del) or to have one copy of F508del plus one copy of a minimally functional CFTR variant (e.g., G542X) based on their patient medical record. Because CF CRS patients typically undergo surgery at a younger age than the average CRS patient, and because age might be a factor in airway cell NO production, we used cells from a subset of younger CRS patients (mean age at surgery 35 ± 2 years) to match the age of our CF CRS patients (mean age at surgery 33 ± 3 years, p = 0.56 by Student’s t test). The two patient populations are shown in Supplemental Table 1 .

Primary human monocyte-derived macrophages were cultured as done previously (74, 75). De-identified human monocytes were isolated from healthy apheresis donors were obtained by the University of Pennsylvania Human Immunology core using RosetteSep™ immunodensity cell separation (StemCell Technologies). Monocytes were seeded onto glass 8-well chamber slides or 96 well plates (CellVis) and differentiated into M0 macrophages by adherence culture in high glucose RPMI 1640 + 10% human serum + 1x pen/strep for 12 days. Our prior studies suggest no differences in T2R responses among macrophages differentiated by adherence alone or by adherence plus M-CSF (43), and thus adherence alone was used for these studies.

HBSS was the primary buffer used in the in vitro experiments described below. SC79 was made as 10 mg/ml SC79 stock in DMSO. PTC was made as 1 M stock in DMSO, 3oxoC12HSL as 100 mM stock in DMSO, apigenin as 100 mM stock in DMSO. Control solutions were HBSS + 0.1% or 0.2% DMSO vehicle control as appropriate. No effects of DMSO alone were observed, as shown in figures below. Stock solutions of compounds in DMSO were stored at -20°C. Quinine, denatonium benzoate, sodium benzoate, and diphenhydramine were made fresh daily and dissolved directly in HBSS. T2Rs known to be activated by non-bacterial agonists used are as follows (taken from (76, 77)): denatonium benzoate (T2R4, T2R8, T2R10, T2R13, T2R39, T2R43, T2R46, T2R47); quinine (T2R10, T2R7, T2R10, T2R14, T2R39, T2R40, T2R43, T2R44, T2R46), PTC (T2R38), diphenhydramine (T2R14, T2R40), apigenin (T2R14, T2R39).

Fura-2 (Ca²⁺ indicator dye) and DAF-FM (NO indicator dye) were imaged as previously described (69, 74). Briefly, fura-2 was imaged using MetaFluor (Molecular Devices, Sunnyvale, CA USA) and dual excitation filter set on an IX-83 microscope (10x 0.4 NA PlanApo objective) equipped with a fluorescence xenon lamp (Sutter Lambda LS, Sutter Instruments, Novato, CA USA), excitation and emission filter wheels (Sutter Lambda 10-2), and Orca Flash 4.0 sCMOS camera (Hamamatsu, Tokyo, Japan). DAF-FM was imaged on a TS100 microscope (10x 0.3 NA PlanFluor objective; Nikon, Tokyo, Japan) with GFP filter set, XCite 110 LED (Excelitas Technologies, Waltham MA USA), and Retiga R1 Camera (Teledyne Qimaging, Surrey, BC, Canada). DAF-FM time course images were acquired using Micromanager variant of ImageJ (78). All experiments utilized background measurements of unloaded ALIs taken at the same microscope settings; background measurements were subtracted from experimental fluorescence values from each individual wavelength recorded for each experiment. Fura-2 340/380 ratio was taken after background subtraction of individual 340 and 380 image stacks.

Primary human ALIs were loaded for 90 min in the dark with 10 µM DAF-FM-diacetate or fura-2 acetoxymethyl ester (AM) in 20 mM glucose-free HEPES-buffered HBSS plus 0.1% pluronic F127 on the apical side and HBSS containing glucose supplemented with 1x MEM amino acids on the basolateral side, followed by washing three times with the same buffer (22, 24, 70). Compounds were added to the apical side in glucose-free HBSS. Macrophages were loaded with 5 µM fura-2-AM or DAF-FM DA for 45 min in glucose-containing HBSS as previously described (74, 75) and imaged with 20x 0.75 NA PlanApo objective.

Whole-field CBF was imaged at 120 frames per second using a Basler A602 camera and Nikon TS-100 microscope (40x long working distance objective) at ~26-28°C in a custom glass bottom chamber. Experiments utilized Dulbecco’s PBS (+ 1.8 mM Ca²⁺) on the apical side and 20 mM HEPES-buffered Hank’s Balanced Salt Solution supplemented with 1× MEM vitamins and amino acids on the basolateral side. Data were analyzed using the Sisson-Ammons Video Analysis system and normalized to baseline CBF as previously described (4, 22, 79–82).

IF was carried out as previously described (22, 24, 69), with ALI cultures fixed at room temperature in 4% formaldehyde for 20 min, followed by blocking and permeabilization for 1 hour at 4°C in Dulbecco’s phosphate buffered saline (DPBS) containing 5% normal donkey serum (NDS; Abcam cat # ab7475), 1% bovine serum albumin (BSA), 0.2% saponin, and 0.3% Triton X-100. After three washes in DPBS, primary antibody incubation (1:100 for anti-T2R antibodies, 1:250 for tubulin antibody) was performed overnight at 4°C in DPBS containing 5% NDS, 1% BSA, and 0.2% saponin. Subsequent incubation with AlexaFluor (AF)-conjugated donkey anti-mouse and anti-rabbit secondary antibodies (1:1000) was done for 2 hours at 4°C. Transwell filters were then removed from the plastic mounting ring and mounted with Fluoroshield with DAPI (Abcam; Cambridge, MA USA)). For co-staining of T2R14 and T2R38, primary antibodies were labeled directly using Zenon antibody labeling kits (Thermo Fisher Scientific) for AF546 or AF647 as described (22, 24, 69). Images of ALIs were taken on an Olympus Fluoview confocal system with IX-73 microscope and 60x (1.4 NA) objective and analyzed in FIJI (83). Images of submerged H441 cells were taken on an Olympus IX-83 microscope with 60x (1.4 NA) objective using Metamorph. Anti-T2R38 (ab130503; rabbit polyclonal) and anti-beta-tubulin IV (ab11315; mouse monoclonal) antibodies were from Abcam. Anti-T2R14 (PA5-39710; rabbit polyclonal) primary antibody, Anti-T2R46 (rabbit polyclonal; OSR00137W), and conjugated secondary antibodies (donkey anti-rabbit AlexaFluor 546 and donkey anti-mouse AlexaFluor 488) were from ThermoFisher Scientific. T2R38 antibody (Cat #AP59054) and C-terminal peptide (Cat #BP16929b) were from Abcepta. Blocking peptide was incubated with primary antibody at 10:1 molar ratio for 4 hrs at 4°C prior to use. T2R14 antibody (LS-C413403-100) and blocking peptide (LS-E44266-1) were from LSBio. Immunofluorescence images were analyzed in FIJI (83) using only linear adjustments (min and max), set equally between images obtained at identical microscope settings (exposure, objective, binning, etc.).

P. aeruginosa lab strains PAO-1 (ATCC 15692) and ATCC27853, as well as clinical CRS-isolates P11006, 2338, and L3847 (obtained from Drs. N. Cohen and L. Chandler, Philadelphia VA Medical Center) (73) were grown in LB media (Gibco). For anti-bacterial assays, P. aeruginosa were grown to OD 0.1 in Luria broth (LB) and resuspended in 50% saline containing 1 mM HEPES and 0.5 mM glucose, pH 6.5. Nasal ALIs were washed 24 hrs prior with antibiotic-free Ham’s F12K media (ThermoFisher Scientific) on the basolateral side. 30 uL of bacteria solution was placed on the apical side of the ALI for 10 min, followed by aspiration of bulk ASL fluid. After 2 hrs at 37°C, remaining bacteria were removed from the ALI culture by washing followed by life-dead staining with SYTO9 (live) and propidium iodide (dead) with BacLight Bacterial Viability Kit (ThermoFisher Scientific; cat # L7012). Control experiments were performed similarly using transwell filters with no cells and bacteria in solution ± 10 µg/ml colistin sulfate. Green (live)/red (dead) ratio was quantified in a Spark 10M (Tecan, Mannedorf, Switzerland) at 485 nm excitation and 530 nm and 620 nm emission. CFUs counting was done by taking aliquots of bacteria saline solution from similar experiments without the live/dead stain, diluting with saline as indicated, and spotting on LB agar plates.

Phagocytosis assays [as descried (74, 75)] were carried out by incubating macrophages on glass bottom 96 well plates with heat-killed FITC-labeled Escherichia coli strain K-12 bioparticles (ThermoFisher Scientific Vybrant phagocytosis assay kit; cat # V6694) at 250 µg/ml in phenol red-free, low glucose DMEM (15 min, 37°C) followed by immediate recording of fluorescence from living cells after quenching extracellular FITC with trypan blue per the manufacturer’s instructions. Fluorescence was recorded on a Spark 10M plate reader (Tecan) with 485 nm excitation and 535 nm emission. For representative images shown in the text, macrophages were fixed in 4% formaldehyde (Electron Microscopy Sciences; cat # 15714) for 10 min followed by addition of DAPI mounting media (Abcam Fluoroshield; cat #ab104139) and imaging with a 60x 1.4 NA objective.

Data were analyzed in Excel (Microsoft) and/or Prism (GraphPad software, La Jolla, CA). All data in bar graphs are shown as mean ± SEM. Multiple comparisons were made in Prism using one-way ANOVA with Bonferroni (pre-selected pairwise comparisons), Tukey-Kramer (comparing all values), or Dunnett’s (comparing to control value) post-tests; p <0.05 was considered statistically significant. Asterisks (* and **) indicate p <0.05 and p <0.01, respectively. Data points used to construct graphs are available upon request.

We examined primary nasal epithelial cells grown and differentiated at air-liquid interface (ALI), which form motile cilia ( Figure 1A ) that also express bitter receptors like T238 ( Figures 1B, C ) (22), T2R4, T2R14, and T2R16 ( Figure 1D ). Over the course of ALI differentiation, there was no significant difference between TAS2R expression in F508del/F508del CF vs non-CF cells by qPCR ( Figure 1D ). Co-staining for T2R38 and T2R14 suggested that both T2R isoforms were cilia-localized in both CF and non-CF cells (Supplemental Figures 1A–C ), and co-localization was similar when measured by Pearson’s correlation coefficient ( Supplemental Figure 1D ) or antibody-based FRET efficiency calculations ( Supplemental Figures 2A-C ). Thus, both T2R expression and localization appear to be similar in CF and non-CF cells.

T2R activation results in Ca²⁺ elevation ( Figure 2A ). We measured Ca²⁺ responses downstream of T2R38 agonist PTC (50) using Ca²⁺ indicator fura-2. Responses to PTC were observed in non-CF cells genotyped for homozygous functional T2R38 (PAV variant) but not in non-CF cells genotyped for homozygous non-functional T2R38 (AVI variant) ( Figure 2B ). CF cells genotyped as PAV/PAV responded similarly to PAV/PAV non-CF cells ( Figure 2B ). We previously determined that Ca²⁺ responses to P. aeruginosa AHL 3-oxo-dodecanoyl-homoserine lactone (3oxoC12HSL (84)) in nasal ALIs is also dependent on T2R38. Like PTC, non-CF PAV/PAV (functional T2R38) cells responded to 3oxoC12HSL while non-CF AVI/AVI (non-functional T2R38) cells did not ( Figure 2C ). Also, like PTC, CF PAV/PAV and non-CF PAV/PAV cells responded similarly ( Figure 2C ). We also saw similar Ca²⁺ responses to T2R14 agonist apigenin (69) between CF and non-CF cells ( Figure 2D ). Responses in non-CF cells were blocked by antagonist 6-methoxyflavanone (85), supporting activation of T2R14 ( Figure 2D ). Finally, we noted similar Ca²⁺ responses to T2R14 agonist diphenhydramine (DPD) between CF and non-CF cells ( Figure 2E ).

Downstream of T2R Ca²⁺ signaling is NO production ( Figure 2A ). This occurs via endothelial nitric oxide synthase (eNOS) localized to the base of the cilia. Confirming this, we used a published siRNA protocol for primary ALIs (86) and found that eNOS siRNA but not nNOS siRNA reduced T2R Ca²⁺ responses ( Figure 3A ). We also noted that, despite nearly identical Ca²⁺ responses, F508del/F508del CF ALI cultures produced less NO in response to T2R agonists PTC, 3oxoC12HSL, quinine, apigenin, and DPD ( Figures 3B–F ). Individual CF ciliated cells isolated directly from turbinate brushings also exhibited less NO production compared with non-CF cells in response to multi-T2R agonist quinine (individual cell traces shown in Figure 3G and average of independent experiments shown in Figure 3H ), suggesting (along with qPCR data above) that this is not due to reduced numbers of ciliated cells resulting in reduced T2R expression.

We questioned if the defect in NO production was a general defect in NO production, and tested this using small molecule Akt activator SC79, which results in phosphorylation and activation of eNOS in nasal epithelial cells [(87) and Supplemental Figure 3A ]. SC79 stimulated NO production that was reduced by Akt inhibitors MK2206 and GSK690693 ( Supplemental Figure 3B ) as well as NOS inhibitor L-NAME ( Supplemental Figure 3B ). There was no difference in SC79-activated NO production with T2R38 genotype ( Supplemental Figure 3C ), and Akt inhibitors did not decrease NO during stimulation with T2R agonist 3oxoC12HSL ( Supplemental Figure 3D ). SC79 and 3oxoC12HSL activated NO production with different kinetics, with SC79 inducing a more sustained NO production and 3oxoC12HSL inducing more of a “burst” of higher level NO production ( Supplemental Figure 3E ), supporting that these two compounds activate eNOS through different mechanisms. We saw no difference in SC79 stimulated NO production between non-CF and F508del/F508del CF cells ( Supplemental Figure 3F ). This supports that there is not a general defect in eNOS function in CF cells, but rather there is a more specific defect in T2R signaling to eNOS downstream of T2R activation.

To confirm that this was really due to CFTR mutations and not another cause like a genetic linkage artifact, we tested NO production in 16HBE cells that were CRISPR modified to contain CFTR mutations, either F508del or premature stop codon G542X (67). 16HBE cells are an SV40 immortalized non-CF bronchial cell line (88) that produces NO in response to bitter taste agonist denatonium benzoate (89). CFTR has two polymorphisms at residue 470, M or V, that can affect protein stability and disease severity. The F508del mutation is almost exclusively paired with the M470 polymorphism (67). However, the Wt parental 16HBE cells contain V470 CFTR. We saw that both F508del M470 and F508del V470 CFTR 16HBE cells grown at ALI had reduced NO during stimulation with denatonium benzoate compared with the parent non-CF cells ( Supplemental Figures 4A, B ). Because of this, we used the more clinically relevant F508del M470 cell line for the rest of the studies here. As with F508del, G542X cells exhibited less NO production during denatonium stimulation compared with the parent non-CF cells ( Figure 4A ). Sodium benzoate was used as an osmotic/pH control. We ensured that DAF-FM traces reflected NOS activity using NOS inhibitor L-NAME and inactive control D-NAME (1 hr. pretreatment; 10 µM; Figure 4A ). These data suggest that CFTR mutations alone can reduce NO signaling during T2R stimulation, even within an isogenic background.

To further test if CFTR function is required for T2R to NO signaling, we treated F508del CFTR 16HBE ALI cultures for 48 hrs with corrector/potentiator combination VX-770 (ivacaftor; 1 µM) and VX-809 (lumacaftor; 3 µM) or vehicle (DMSO) alone ( Figure 4B ). VX-770 + VX-809 pre-treatment increased NO production during denatonium stimulation ( Figure 4B ). Wt and G542X cells were unaffected. Corrector/potentiator therapies cannot restore CFTR activity with premature stop codon mutations, so the lack of effect on G542X cells suggests that the effects of VX-770 + VX-809 is dependent on their ability to restore CFTR function. Using a Cl- substitution SPQ assay (70) ( Supplemental Figure 5A ), we confirmed that VX-770 + VX-809 treatment restored substantial apical membrane cAMP-stimulated ion permeability in F508del but not G542X cells ( Supplemental Figure 5B ). We also found that VX-770 + VX-809 did not affect T2R Ca²⁺ responses in F508del 16HBE cells ( Supplemental Figure 6 ), ruling out increased Ca²⁺ as a mechanism for the effects of VX-770 + VX-809 on NO production.

To further confirm effects of corrector potentiator therapies, we tested CFBE41o- cells, also an SV40 immortalized bronchial cells but from a F508del/F508del CF patient lung (64). CFBE41o- cells produce NO in response to T2R14 agonist DPD (73). This NO production was increased after VX-770 + VX-809 treatment ( Figure 4C ). Even further supporting the role of CFTR in T2R NO production, we tested CFBE41o- cells stably over-expressing Wt or F508del CFTR. Wt CFTR expression increased DPD-stimulated NO production while F508del was without effect ( Figure 4D ). Lastly, F508del/F508del primary CF cells pretreated with VX-770 + VX-809 exhibited increased NO production to T2R agonists 3oxoC12HSL, quinine, apigenin, and DPD ( Figure 4E ). Interestingly, co-application (48 hrs) of VX-770 + VX-809 and CFTR_inh172 (5 µM) abrogated the effect of VX-770 + VX-809 ( Figure 4F ).

How does this affect innate immunity in CF cells? To answer this, we first tested ciliary beat frequency (CBF), which drives mucociliary transport and clearance of inhaled/inspired pathogens. T2R activation and stimulation of NO production activates guanylyl cyclase to increase CBF via protein kinase G (24, 75, 90). CF ALIs exhibited reduced CBF increases with 3oxoC12HSL ( Figure 5A ), apigenin ( Figure 5B ), and DPD ( Figure 5C ) compared with non-CF ALIs. Because the whole field analysis of CBF takes into account only actively beating areas as determined by the software (80), this further confirms that effects observed here are not due to differences in numbers of ciliated cells, but rather defects intrinsic to the ciliated cells themselves.

Next, we examined the ability of CF and non-CF nasal ALIs to kill both lab (strain PAO1) and clinical (strain P11006) P. aeruginosa. In this assay, incubation of P. aeruginosa with nasal ALIs results in bacterial killing over 2 hours that requires T2R38 and NO production, as previously demonstrated (22, 75). In this assay, we have found that PAV/PAV ALIs kill P. aeruginosa while AVI/AVI ALIs do not, demonstrating a requirement for T2R38 function in bacterial killing (22, 75). Bacterial viability is measured by the fluorescence ratio of stains for live (Syto9) and dead (propidium iodide [PI]). PAV/PAV (functional T2R38) non-CF cultures killed both strains of P. aeruginosa to levels comparable with antibiotic colistin ( Figure 5D ). AVI/AVI non-CF cultures and PAV/PAV CF cultures both exhibited reduced bacterial killing ( Figure 5D ). NOS inhibitor L-NAME, but not inactive D-NAME, inhibited non-CF nasal epithelial cell bacterial killing ( Figure 5D ), supporting the role for NO in this assay.

We verified the live-dead staining assay using colony forming unit (CFU) counting (shown for PAO-1 in Figure 5E ), as previously done (75). CF PAV/PAV cells killed less bacteria than non-CF PAV/PAV cells across multiple P. aeruginosa strains ( Figure 5F ). Together with CBF data above, these data suggest that CF nasal epithelial cells have a reduced capacity to both clear and kill bacteria during T2R stimulation.

We next tested if this was a general effect of CFTR function on T2R signaling or if this was specific to nasal epithelial cells. We previously reported that macrophages exhibit T2R-induced Ca²⁺ signaling and NO production downstream of eNOS (43, 75). Instead of regulating ciliary beating, this pathway instead regulates phagocytosis. We confirmed that macrophages express T2R38 ( Figures 6A, B ), T2R14 ( Figures 6C, D ), and T2R46 ( Figure 6E ) with localization reminiscent of plasma membrane marker Na⁺/K⁺ ATPase ( Figure 6F ). We examined NO production by DAF-FM imaging and found that pre-treatment (24 hrs) with CFTR_inh172 (10 µM) resulted in reduced NO production when stimulated with multi-T2R agonists denatonium benzoate or quinine as well as T2R14-specific agonist flufenamic acid (FFA; Figures 6G, H ). This was accompanied by reduced phagocytosis of FITC-labeled E. coli ( Figures 6I, J ), suggesting that loss of CFTR function can reduce T2R signaling in multiple cell types.