Flavones modulate respiratory epithelial innate immunity: Anti-inflammatory effects and activation of the T2R14 receptor

Chronic rhinosinusitis has a significant impact on patient quality of life, creates billions of dollars of annual healthcare costs, and accounts for ∼20% of adult antibiotic prescriptions in the United States. Because of the rise of resistant microorganisms, there is a critical need to better understand how to stimulate and/or enhance innate immune responses as a therapeutic modality to treat respiratory infections. We recently identified bitter taste receptors (taste family type 2 receptors, or T2Rs) as important regulators of sinonasal immune responses and potentially important therapeutic targets. Here, we examined the immunomodulatory potential of flavones, a class of flavonoids previously demonstrated to have antibacterial and anti-inflammatory effects. Some flavones are also T2R agonists. We found that several flavones inhibit Muc5AC and inducible NOS up-regulation as well as cytokine release in primary and cultured airway cells in response to several inflammatory stimuli. This occurs at least partly through inhibition of protein kinase C and receptor tyrosine kinase activity. We also demonstrate that sinonasal ciliated epithelial cells express T2R14, which closely co-localizes (<7 nm) with the T2R38 isoform. Heterologously expressed T2R14 responds to multiple flavones. These flavones also activate T2R14-driven calcium signals in primary cells that activate nitric oxide production to increase ciliary beating and mucociliary clearance. TAS2R38 polymorphisms encode functional (PAV: proline, alanine, and valine at positions 49, 262, and 296, respectively) or non-functional (AVI: alanine, valine, isoleucine at positions 49, 262, and 296, respectively) T2R38. Our data demonstrate that T2R14 in sinonasal cilia is a potential therapeutic target for upper respiratory infections and that flavones may have clinical potential as topical therapeutics, particularly in T2R38 AVI/AVI individuals.

Chronic rhinosinusitis (CRS) 2 is a syndrome of chronic inflammation and/or infection of the upper respiratory tract, which leads to substantial decreases in patient quality of life, creates Ͼ$8 billion in direct healthcare costs in the United States alone, and can seed lower respiratory infections and exacerbate lung diseases (1)(2)(3). CRS is also an important public health concern, as it accounts for ϳ20% of antibiotic prescriptions in adults in the United States (1, 4 -8), making it a significant driver for the emergence of antibiotic-resistant organisms (9 -15). An attractive therapeutic strategy to avoid the selective pressures for antibiotic resistance is to stimulate endogenous innate host defenses. However, this requires a better understanding of the receptors involved in activating upper airway innate immunity.
It is believed that multiple etiologies contribute to the pathogenesis of CRS, but a common phenotype of the disease is defective mucociliary clearance, the major physical defense of the conducting airways (1,3). A thin layer of mucus secreted by airway goblet cells and submucosal glands traps inhaled pathogens and particulates. Coordinated beating of motile cilia drives transport of this mucus to the oropharynx, where it is cleared by expectoration or swallowing (16). In CRS, chronic infection and persistent inflammation is associated with stasis of sinonasal secretions, possibly through alterations of basal (17,18) or stimulated (19) ciliary beat frequency (CBF), mucus viscosity (17,20), or ion/fluid transport (21). Enhancing mucociliary clearance may have potential therapeutic benefit for CRS patients (1,3).
We previously showed that a bitter "taste" G-protein-coupled receptor, T2R38, is expressed in the motile cilia of cells of the upper airway (nose and sinuses) and modulates mucociliary clearance through calcium (Ca 2ϩ )-dependent production of nitric oxide (NO) (2,(22)(23)(24)(25)(26)(27). Our data suggest that T2R38 is activated by acyl-homoserine lactone (AHL) quorum-sensing molecules secreted by Gram-negative bacteria (27), including the common airway pathogen Pseudomonas aeruginosa. When activated, T2R38-activated NO production drives an increase in CBF through protein kinase G (PKG), which phosphorylates ciliary proteins (28). This NO also directly diffuses into the airway surface liquid with antibacterial effects (27). Because NO damages the cell walls and DNA of bacteria (29,30), NO production by the sinonasal epithelium is thought to be important for preventing infection. Supporting this, nitric-oxide synthase (NOS) polymorphisms have been associated with severe CRS (31). We found that patients homozygous for a polymorphism in the TAS2R38 gene resulting in non-functional T2R38 protein (the AVI (alanine, valine, isoleucine at positions 49, 262, and 296, respectively) polymorphism (32)) are more susceptible to Gram-negative bacterial infection (27), have a higher incidence of biofilm-forming bacteria (33), are at higher risk for CRS requiring functional endoscopic sinus surgery (FESS) (34,35), and may have worse outcomes after FESS for CRS without nasal polyps (36) compared with patients homozygous for the functional (PAV: proline, alanine, and valine at positions 49, 262, and 296, respectively) allele of TAS2R38. There are 25 different T2R bitter receptors in humans (22,24,37,38), and other researchers have confirmed that polymorphisms in at least two TAS2R genes, TAS2R38 and TAS2R13, correlate with CRS by genome-wide association (39).
The full picture of T2R38's role in sinonasal immunity in different patient populations is still being elucidated. A subsequent study found no correlation of TAS2R38 genotype with CRS in Italian patients (40), but this population had different clinical characteristics (e.g. more recalcitrant disease and stronger Th2 inflammatory phenotype) than our own studies. A more recent study supported a correlation between TAS2R38 genotype and CRS severity in a Polish population (41), and an Australian study concluded that AVI/AVI TAS2R38 is predictive of the presence of culturable bacteria in CRS patient sinuses (42). Moreover, a recent study reported that T2R10 and T2R14, but not T2R38, respond to some AHLs when expressed in HEK293 cells (43). The reason for this discrepancy is not yet clear, although other studies have suggested that T2R38 is indeed important for immune cell detection of AHLs (44,45). Understanding the full range of T2Rs endogenously expressed by primary differentiated cell types as well as their oligomerization properties will help clarify the best therapeutic compounds to target these receptors to stimulate innate immune responses.
Plants produce thousands of polyphenolic flavonoids (46 -49), which are of great biomedical interest, because they have effects on both eukaryotic and prokaryotic cells (47)(48)(49)(50)(51)(52)(53)(54)(55)(56)(57)(58). Flavones are a sub-group of flavonoids that have been demonstrated to have antibacterial, antioxidant, and anti-inflammatory effects in various in vitro models (48,51,53,55,57,59,60). Moreover, the flavones chrysin and apigenin activate T2R14 and T2R39 (61,62). Because of their potential antibacterial properties, and because multiple TAS2R polymorphisms have been linked to CRS (39), we sought to investigate the effects of flavones on airway epithelial cells. We previously examined a separate flavonoid class, anthocyanidins, and found that they did not activate sinonasal cilia T2Rs (58). However, anthocyanidins have a Ն1-log higher effective concentration (EC) for T2R activation than flavones (e.g. 250 M EC for the anthocyanidin cyanidin versus 8 M EC for the flavone apigenin for T2R14) (61,62). The goal of this study was to begin to investigate the potential role of flavones as a treatment for upper respiratory infections, with a special focus on their anti-inflammatory properties as well as potential reactivity with upper airway T2Rs.

Immunomodulatory activities of flavones
We sought to determine whether flavones exhibited antiinflammatory effects in airway epithelial cells, as reported previously in other cell types (51). The protein kinase C (PKC) pathway is utilized by many inflammatory mediators, and activation of PKC with phorbol 12-myristate 13-acetate (PMA) is frequently used to induce inflammatory effects in vitro. PKC inhibitors are utilized as anti-inflammatory compounds in a number of diseases of chronic inflammation (66 -68). Apigenin and other flavones may have PKC inhibitory activity (69,70), suggesting they may be useful for attenuating PKC-dependent inflammatory effects. When 16HBE14oϪ (16HBE) airway epithelial cells were stimulated with PMA (10 nM, 18 h), we observed a marked up-regulation of Muc5AC expression via immunofluorescence microscopy (Fig. 2a). Muc5AC is a gelforming mucin in the airway, and Muc5AC up-regulation is a common hallmark of chronic obstructive pulmonary disease and other inflammatory airway diseases (71,72). Other studies have also demonstrated that PKC up-regulates Muc5AC (73). Individual flavones (10 M each) significantly repressed PMAinduced Muc5AC up-regulation (Fig. 2a), suggesting they may have an anti-inflammatory role in airway cells.
We tested another important marker of inflammation, inducible nitric-oxide synthase (iNOS) (74 -77). Whereas acute NO production can be antibacterial, chronic excessive NO production through transcriptional up-regulation of iNOS may exacerbate and prolong inflammation. PMA induced an upregulation of iNOS in A549 alveolar-derived airway cells (Fig.  2b) that was likewise reduced with individual flavone compounds (10 M each; Fig. 2b). As a control for the specificity of iNOS as an inflammatory marker, we examined changes in endothelial NOS (eNOS), typically not up-regulated during inflammation. Immunofluorescence for eNOS exhibited a perinuclear Golgi-like localization, as reported previously (78,79). PMA had no effect on eNOS, either alone or in the presence of flavones (Fig. 2c).
We also directly quantified cytokines released from air-liquid interface (ALI) cultures of primary human sinonasal cells. ALIs from primary cells differentiate into ciliated and goblet cells mimicking the in vivo epithelium (24,27,80,81). Primary sinonasal ALIs were stimulated as above with PMA on the apical side in the presence or absence of chrysin or apigenin (Fig. 3a). Levels of secreted interleukin-8 (IL-8), granulocyte colonystimulating factor (GCSF), and granulocyte macrophage colony-stimulating factor (GMCSF) were quantified from basolateral media by ELISA. PMA induced robust secretion of all three cytokines (Fig. 3a), and exposure to apigenin or chrysin reduced cytokine release (Fig. 3a). To test whether flavones have direct inhibitory activity against PKC, we expressed a PKC FRET-reporter construct, CKAR (82,83), in 16HBE cells. CKAR includes the FHA2 domain of RAD53p and a PKC phosphorylation sequence sensitive to all PKC isoforms, flanked by an N-terminal enhanced CFP and C-terminal citrine YFP. When phosphorylated, a reversible conformational change moves the CFP and YFP apart, decreasing FRET emission. A decrease in FRET emission thus correlates with increased PKC activity and vice versa. Single transfected cells were imaged, and the ratio of FRET to CFP emission at CFP excitation is reported, as described previously (80). We noted that cells pretreated for 1 h with 50 M apigenin exhibited a reduced PKC activation (FRET decrease) in response to 1 M PMA (Fig. 3b).
We likewise examined IL-8 secretion in both 16HBE ALIs (Fig. 3c) and primary sinonasal ALIs (Fig. 3d) in response to tumor necrosis factor ␣ (TNF␣) stimulation. The flavonol quercetin was reported to reduce TNF␣-induced IL-8 secretion in 16HBE cells through a phosphatidylinositol 3-kinase (PI3K)dependent pathway (84). We found that TNF␣ induced IL-8 secretion ϳ10-fold in both 16HBE (Fig. 3c) and primary sino- under control conditions or pre-treated with apigenin for 1 h prior to experiment. Cells were stimulated with PKC-activating PMA (10 M) followed by the global PKC inhibitor Gö6983 (10 M). c and d, similar experiments performed with 16HBE ALIs (c) and primary sinonasal ALIs (d) stimulated with TNF␣ (10 ng/ml; 24 h) in the presence or absence of apigenin or chrysin. e, IL-8 release (pg/ml) in response to pyocyanin in 16HBE, Calu-3, and primary sinonasal ALI cultures (n ϭ 5 ALIs for each condition). Significance was determined by one-way ANOVA, Tukey-Kramer, or Bonferonni post-test; *, p Ͻ 0.05; **, p Ͻ 0.01 for all graphs, and for D, ##, p Ͻ 0.01 versus pyocyanin alone. Data values from bar graphs are reported in supplemental Material.

Flavones activate T2R14 expressed in sinonasal epithelial cell cilia
The above data suggest that flavones modify sinonasal epithelial inflammation-induced up-regulation of Muc5AC, iNOS, and cytokines. As mentioned above, apigenin and chrysin may also activate T2R14 and T2R39 (61), whereas tangeritin may activate T2R14 and T2R46 but not T2R39 (99). We hypothesized that flavones may activate one or more of these T2Rs located in the upper respiratory epithelium (22,25). The molecular structures of flavones were examined using BitterX, an open-access platform for predicting whether a molecule is bitter as well as predicting which T2R receptors are likely to respond (100). BitterX identified T2R14 and T2R39 as having the highest probabilities of being activated by all flavone molecules examined here (supplemental Table 1).
We examined endogenous T2R14 and T2R39 expression in primary sinonasal ALIs with anti-C-terminal antibodies validated against cross-reactivity by heterologous expression of T2R14 and T2R39 in A549 cells (supplemental Fig. 1). Using these antibodies, we detected endogenous expression of T2R14 but not T2R39 in primary sinonasal ALIs (Fig. 4a). T2R14 colocalized with the motile cilia marker ␤Tubulin IV and T2R38 ( Fig. 4a), previously demonstrated to be localized to sinonasal cilia (27). Surface localization of T2R14 was confirmed by staining un-permeabilized cells with an antibody raised against an epitope (amino acids 213-262) containing an extracellular loop between transmembrane helix 6 and 7 (Fig. 4b). The co-localization observed between T2R38 and T2R14 was highly distinct from the pattern observed with T2R38 or T2R14 and T2R4. T2R4 localized only to the distal tips of cilia (supplemental Fig.  2, a and b). This distal localization of T2R4 was previously observed in bronchial cilia (101), and it may be related to previous HEK293T heterologous expression data suggesting that recombinant T2R4 does not oligomerize with T2R14 or T2R38, whereas T2R14 and T2R38 can oligomerize with each other (102,103). T2R14 and T2R38 co-localization was similar to that observed with T2R16 (supplemental Fig. 2, c and d).
Because endogenous T2R14 and T2R38 appeared to strongly co-localize, and because the primary antibodies used here were both raised against the C termini (Fig. 5a), we examined whether endogenously expressed T2R38 and T2R14 interacted closely enough to observe Förster (also known as fluorescence) resonance energy transfer (FRET; schematic shown in Fig. 5b) in primary sinonasal ALIs. FRET is a highly quantitative measure of co-localization. Although T2Rs have been previously proposed to oligomerize, presumably as dimers (102,103), these studies have been done in heterologous overexpression HEK293 cell systems, and thus results must be interpreted with . Immunofluorescence co-localization of endogenous T2R38 and T2R14 in sinonasal cilia. a, upper panels show a representative confocal immunofluorescence image of a primary sinonasal ALI culture and co-localization of the motile cilia marker ␤-tubulin IV (␤-TubIV, green) with T2R38 (cyan) and T2R14 (magenta). Detection of T2R39 was not observed (lower panels). Orthogonal slice of T2R14 and ␤-tubulin IV is shown at top right. b, T2R14 immunofluorescence was observed in non-permeabilized cells using an antibody recognizing an extracellular epitope, supporting surface localization; ␤-tubulin IV staining was conducted after subsequent second fixation and permeabilization. Images are representative of results observed with ALI cultures from at least three individual patients.
caution. Overexpression can force molecules to interact in vitro that may not necessarily interact in vivo at endogenous expression levels. These results must be interpreted in light of findings from studies of endogenous expression.
We used AlexaFluor (AF) 555 and AF647 as our FRET pair (Förster radius [R o ] ϭ 51 Å). Primary antibodies were labeled with AF-conjugated Fab fragments. We noted a significant T2R14 FRET signal in the AF647 emission channel when T2R38 AF555 was excited using a 559-nm laser (Fig. 5c). Similar T2R38 FRET signals were observed when T2R14 was labeled with AF555 and T2R38 was labeled with AF647. When the acceptor fluorophore was bleached using a 647 laser, donor fluorescence increased (Fig. 5, d-f). Quantification of the relative increase in donor fluorescence after acceptor bleaching allows calculation of the FRET efficiency (E). E was measured to be 15 Ϯ 3% when T2R38 was the donor and T2R14 was the acceptor, 17 Ϯ 1% when T2R14 was the donor and T2R38 was the acceptor, and 16 Ϯ 2% when the two conditions were pooled. R o is the distance at which the E is 50% (51 Å for AF555 and AF647); FRET efficiency decreases to the 6th power as distance (R) increases: R ϭ R o ⅐((1 Ϫ E)/E) 1/6 . Thus, with 0.16 E, R ϭ 51 Å⅐((1 Ϫ 0.16)/0.16) 1/6 ϭ 67 Å. Although we cannot conclude that the two proteins form oligomers per se from our data due to the sizes of the antibodies involved in our measurements, similar antibody-based FRET measurements have been used to suggest a "close association" of endogenously expressed proteins (105). These quantitative FRET measurements demonstrate a close co-localization of T2Rs to a higher resolution (ϳ7 nm) than yet achievable by most current super-resolution light microscopy techniques (106).
With the knowledge that T2R14 is endogenously expressed in primary sinonasal ciliated epithelial cells and the knowledge that this receptor is activated by flavones, we sought to determine whether flavones activate T2R14-dependent Ca 2ϩ -driven nitric oxide (NO) responses in these cells, as demonstrated previously with T2R38 (27). Apigenin and chrysin activated lowlevel Ca 2ϩ responses (Fig. 6, a-c) identical to what was previously observed with stimulation of sinonasal cilia-localized T2R38 (27). Although these [Ca 2ϩ ] i changes are small, it is important to note that these are measurements of global [Ca 2ϩ ] i ; localized responses at or near the apical membrane and/or within cilia could be much higher (107). Importantly, the ECs observed here fit with ECs previously determined for these compounds for heterologously expressed T2R14 (8 M for apigenin and 63 M for chrysin (61,62)). These responses were also blocked in the presence of 4Ј-fluoro-6-methoxyflavanone (Fig. 6, d and e), a previously identified T2R14 inhibitor (64,65). We found that 4Ј-fluoro-6-methoxyflavanone inhibited apigenin-induced T2R14 [Ca 2ϩ ] i responses at 50 M but had activating effects at higher concentrations, suggesting that 4Ј-fluoro-6-methoxyflavanone may act as a partial agonist rather than a competitive antagonist (supplemental Fig. 5). Nonetheless, this inhibition in primary cells supports that these responses are mediated by T2R14. No difference in [Ca 2ϩ ] i responses was observed when cultures from T2R38 PAV/PAV (homozygous taster) or AVI/AVI (homozygous non-taster) individuals were compared, suggesting that, despite close co-localization and possible heterodimerization, the non-functional AVI T2R38 does not adversely affect T2R14 Ca 2ϩ signaling in cilia (supplemental Fig. 6, a, b, and e). Supporting this, we found no inhibition of apigenin-induced Ca 2ϩ responses with the T2R38 inhibitor probenecid (1 mM) (108) and no inhibition of phenylthiocarbamide-induced Ca 2ϩ responses in PAV/PAV cultures with 4Ј-fluoro-6-methoxyflavanone (supplemental Fig. 6, c-e).

Activation of T2R14 in sinonasal cilia results in NO production and increases ciliary beat frequency
As observed with T2R38, flavone/T2R14-induced Ca 2ϩ responses were associated with increases in fluorescence of DAF-FM, which reacts with intracellular reactive nitrogen species (Fig. 7a). DAF-FM fluorescence increases reflected NO production, as they were blocked by the nitric-oxide synthase (NOS) inhibitor L-N G -nitroarginine methyl ester (L-NAME) but not by its inactive analogue D-NAME (Fig. 7b). These responses were blocked by the phospholipase C inhibitor U73122 but not by its inactive analogue U73343 (Fig. 7c). Like flavone-induced Ca 2ϩ signals, flavone induced NO production was also inhibited by 4Ј-fluoro-6-methoxyflavanone (Fig. 7d). These results are summarized in Fig. 7e.
The constant and coordinated beating of airway motile cilia drives mucociliary clearance (1,3,110,111). Because NO is known to increase ciliary beating through activation of guanylyl cyclase and production of cGMP (27,28), we measured CBF during stimulation with apigenin or chrysin. Both compounds increased CBF (Fig. 8a; n ϭ 6 -8 ALIs from at least three patients per condition). The increases in CBF were dependent on NO production, as they were blocked by L-NAME but not by D-NAME (Fig. 8b).

Discussion
This study first demonstrates that flavones have anti-inflammatory effects on Muc5AC and iNOS up-regulation as well cytokine secretion in airway epithelial cells. This supports prior studies that apigenin, chrysin, and wogonin inhibit Muc5AC up-regulation in NCI-H292 bronchial cancer cells in response to epidermal growth factor (91,112,113) or TNF␣ (114,115). Our data suggest this occurs partly through direct inhibition of kinase C rather than upstream or downstream mechanisms, as they were directly able to block PKC activity in response to PMA. We also demonstrate relevance of prior findings from cell lines to primary well differentiated airway epithelial cells. We postulate that flavones may be useful potential anti-inflammatory mediators in airway diseases, as has also been suggested for the related, but structurally different, flavonol quercetin (84).
Second, we demonstrate that flavones activate T2R14 in primary sinonasal epithelial cell cilia. T2R14 is also expressed in bronchial cilia (101) and airway smooth muscle cells (116). This is the first demonstration of T2R14 in sinonasal cilia, and the finding that T2R14 closely co-localizes (Ͻ7 nm) with T2R38 is important and novel. These studies were facilitated by the commercial availability of valid and specific antibodies directed at moieties on the same side of the plasma membrane. Although we cannot directly determine whether these T2Rs oligomerize by the methods used here due to the size of the labeled antibodies, this study is one of the first to demonstrate such a close interaction with endogenously expressed T2Rs. Previous studies of T2R interactions have been carried out in heterologous expression HEK293 models (102,103). Results from such models, where overexpression may induce interactions that do not normally occur at physiological protein levels, must be interpreted with caution and must be verified in more differentiated in vitro models and/or tissues. Future studies in airway epithelial cells may help to clarify the oligomerization properties of the subset of T2Rs endogenously expressed in motile cilia, shedding important light into the cell biology of these receptors. Although T2R38 is required for AHL detection in sinonasal epithelial cells (27) and immune cells (44,45), another study has suggested T2R10 responds to AHLs in a heterologous expression system (43). Oligomerization has substantial functional consequences for many G-protein-coupled receptors  (117,118). A better understanding of the oligomerization of endogenous T2Rs, both with other T2Rs as well as with other receptor classes (116), may help explain some such discrepancies that exist in the literature.
Activation of T2R14 results in similar Ca 2ϩ -dependent NO production and ciliary beat increases to those observed during T2R38 stimulation (27). These CBF increases were smaller than those observed with ATP but were sustained and were similar to elevations previously observed with T2R38 stimulation (27) that may have clinical relevance (27, 34 -36, 39, 41, 42, 119).
Quercetin was previously shown to increase CBF in primary human sinonasal cells (120), and our data suggest this activation may at least partly occur through activation of T2R14. Interestingly, T2R14 is activated by several other compounds that have been used clinically, including NFA, the antihistamine diphenhydramine, the non-steroidal anti-inflammatory drug flufenamic acid, and quercetin (62). NFA is commonly used experimentally at high micromolar-to-low millimolar concentrations as a Ca 2ϩ -activated Cl Ϫ channel blocker (121)(122)(123)(124)(125)(126). NFA was shown to activate Ca 2ϩ release from intracellular stores in rat pulmonary artery smooth muscle cells (127); as bronchial smooth muscles express T2R14 (116, 128), arterial  smooth muscle expression of T2R14 may explain this effect. Off-target T2R effects of NFA should be taken into account when using this compound with cells that express T2R14. It is also likely that there is either an endogenous host ligand or a pathogen ligand for physiological activation of sinonasal T2R14, which remains to be identified. Screening of bacterial or fungal compounds against this protein may reveal the identity of such a compound and help clarify the physiological role of T2R14 in sinonasal innate immunity.
We previously postulated that T2R38 may be a useful therapeutic target to activate NO-driven endogenous innate immune responses in patients to treat airway infections without the use of antibiotics (2,22,(25)(26)(27)129). However, there is a high frequency of non-functional (non-taster) TAS2R38 polymorphisms. In Caucasians, ϳ25% of the population is homozygous for the AVI non-functional TA2R38 (27, 32, 130 -132). These patients would be expected to be completely non-responsive to T2R38-directed therapeutics, and we have previously confirmed that ALIs grown from AVI/AVI patients do not respond to T2R38 agonists (27). The observation that T2R14 activates a highly similar set of responses from both T2R38 PAV/PAV and AVI/AVI sinonasal cells suggests that T2R14 may be a viable therapeutic target to activate innate immune pathways in T2R38 AVI/AVI patients. However, polymorphisms in TAS2R14 also exist (133), although their effects are not as well studied. Future work is needed to understand how known TAS2R14 polymorphisms affect T2R14 function. Additionally, identifying the full gamut of sinonasal cilia taste receptors will allow elucidation of agonist mixtures that activate multiple sinonasal T2Rs to achieve maximum possible benefit in the largest patient population possible. T2R agonists such as flavones that have additional beneficial non-T2R-dependent effects (e.g. PKC inhibition) may increase efficacy of this therapeutic strategy.
For ALI cultures, cells were seeded onto collagen-coated ThinCert cell culture inserts (0.4-m pore size, transparent PET membrane, Greiner Bio-One, Germany) in 12-well plates (1.1-cm 2 surface area). After ϳ3-5 days of culture and confirmation of confluence and establishment of transepithelial electrical barrier (Ͼ300 ohms⅐cm 2 ), cultures were exposed to air on the apical side by removal of apical media and washing three times with PBS. Cultures were fed from the basolateral side for at least 2 weeks before use to allow for full differentiation. For cytokine release studies, ALI cultures were stimulated on the apical side only (to simulate mucosal exposure) with PMA, TNF␣, or pyocyanin dissolved in sterile PBS. Cytokines were quantified from the basolateral medium.
HEK293T cells (ATCC) were cultured in high-glucose DMEM (Gibco). Transfection of A549 and HEK293T cells for heterologous expression experiments was carried out with Lipofectamine 2000 according to the manufacturer's instructions. Sequences of human TAS2R4, TAS2R5, TAS2R14, TAS2R16, TAS2R38 (PAV functional isoform), and TAS2R39 were cloned into pcDNA3.1 vectors containing the first 45 amino acids of the rat type 3 somatostatin receptor at the N terminus to enhance membrane expression, as described previously (103,137). Each T2R was co-transfected with pcDNA3.1 containing G␣ 16 -gust44, a chimeric G␣ protein containing the last 44 amino acids of gustducin, as described previously (137). Cells were used 24 -48 h after transfection. Transfection of 16HBE cells with the protein kinase C (PKC) FRET-reporter construct CKAR (82,83) was performed with Lipofectamine 3000; cells were used 24 h. after transfection.

Generation of primary sinonasal ALI cultures
All experimental protocols were carried out in accordance with the University of Pennsylvania School of Medicine guidelines regarding use of residual clinical material in research. Patients undergoing sinonasal surgery were recruited from the Department of Otorhinolaryngology at the University of Pennsylvania with full IRB approval (#800614), and written informed consent was obtained for all participating patients in accordance with the United States 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. Wegener's granulomatosis, sarcoidosis, and cystic fibrosis), immunodeficiencies, or use of antibiotics, oral corticosteroids, or anti-biologics (e.g. Xolair) within 1 month of surgery. Human sinonasal epithelial cells were enzymatically dissociated and grown to confluence in proliferation medium (DMEM/Ham's F-12 plus BEBM; Clonetics, Cambrex, East, NJ) for 7 days as described previously (27,138). 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 were changed to differentiation medium (1:1 DMEM/BEBM) containing human epidermal growth factor (0.5 ng/ml), epinephrine (5 ng/ml), BPE (0.13 mg/ml), hydrocortisone (0.5 ng/ml), insulin (5 ng/ml), triiodothyronine (6.5 ng/ml), and transferrin (0.5 ng/ml), supplemented with 100 units/ml penicillin, 100 g/ml streptomycin, 0.1 nM retinoic acid, and NuSerum (BD Biosciences) as described previously (27,138).

Live cell imaging of intracellular Ca 2ϩ , reactive nitrogen species production, and PKC activation
[Ca 2ϩ ] i and reactive nitrogen species were imaged in primary sinonasal ALIs using the Fura-2 and DAF-FM, respectively, as described previously (23,24,27,139). ALI cultures were loaded with Fura-2 AM (5 M applied apically) for 90 min followed by washing and 20 min of incubation in the dark. Cultures were similarly loaded with 10 M DAF-FM diacetate for 90 min in the presence of 5 M cPTIO, followed by washing to remove unloaded DAF-FM and cPTIO and incubation for 15 min prior to imaging. Imaging was performed using an Olympus IX-83 microscope (10 ϫ 0.4 NA PlanApo objective, Olympus Life Sciences, Tokyo, Japan) equipped with a fluorescence xenon lamp (Sutter Lambda LS, Sutter Instruments, Novato, CA), excitation and emission filter wheels (Sutter Instruments), and a 16-bit Hamamatsu Orca Flash 4.0 sCMOS camera. Images were acquired using MetaFluor (Molecular Devices, Sunnyvale, CA). Dual excitation of fura-2 was carried out with 340/11-nm and 380/11-nm bandpass excitation filters, 510 long pass dichroic mirror, and 560/80 nm emission filter (set 79002-ET, Chroma Technologies, Rockingham, VT). Excitation of DAF-FM was carried out with 470/40-nm excitation filter, 495 long pass dichroic, and 525/40-nm emission filter (Chroma 49002-ET). For fura-2 [Ca 2ϩ ] i measurement in submerged HEK293T cells, cells grown on coverslips were loaded with 2 M fura-2 AM for 30 min in Hanks' balanced salt solution plus 10 mM HEPES, followed by washing and imaging with a ϫ30 (1.05 NA) UPLSAPO silicone immersion objective (Olympus) using MetaFluor.
Calibration of fura-2 fluorescence from primary nasal cells to [Ca 2ϩ ] i was carried out as described previously (123,124,126) using the method of Grynkiewicz et al. (140) and R min , R max , and ␤ values empirically determined for this specific imaging system using live cells with 0-Ca 2ϩ (1 mM EGTA) and high Ca 2ϩ (10 mM) solutions containing 10 g/ml ionomycin and thapsigargin. DAF-FM measurements utilized raw fluorescence values to compare experiments performed under identical conditions with identical microscope settings.

Measurement of CBF
Whole-field CBF was measured using the Sisson-Ammons Video Analysis system (141) as described previously (27,81,139) at ϳ28 -30°C. Cultures were imaged using at 100 frames/s using a Leica DM-IL microscope (ϫ20/0.8 NA objective) with Hoffman modulation contrast (Leica IMC). Experiments utilized Dulbecco's PBS (1.8 mM Ca 2ϩ ) on the apical side and HEPES-buffered Hanks' balanced salt solution supplemented with 1ϫ MEM vitamins and amino acids on the basolateral side.

Immunofluorescence microscopy
Immunofluorescence was carried out as described previously (27), with modifications outlined below. ALI cultures were fixed in 4% formaldehyde for 20 min at room temperature, followed by blocking and permeabilization in phosphate-buffered saline (PBS) containing 1% bovine serum albumin (BSA), 5% normal donkey serum, 0.2% saponin, and 0.3% Triton X-100 for 1 h at 4°C. A549 or 16HBE cells were fixed in 4% formaldehyde for 20 min at room temperature, followed by blocking and permeabilization in PBS containing 1% BSA, 5% normal donkey serum, 0.2% saponin, and 0.1% Triton X-100 for 30 min at 4°C. Primary antibody incubation (1:100 for anti-T2R antibodies and 1:250 for tubulin antibodies) were carried out at 4°C overnight. AlexaFluor-labeled donkey anti-mouse or rabbit secondary antibody incubation (1:1000) was carried out for 2 h at 4°C. Transwell filters were removed from the plastic mounting ring and mounted with Fluoroshield with DAPI (Abcam). For direct labeling of primary antibodies to co-localize T2R38 and T2R14, Zenon antibody labeling kits (Invitrogen/Molecular Probes/ Thermo Fisher Scientific) were used according to the manufacturer's instructions. Images of ALIs were taken on an Olympus Fluoview confocal system with IX-81 microscope and ϫ60 (1.4 NA) objective. Images of A549 and 16HBE cells were taken in wide field with a ϫ60 (1.4 NA oil) objective on an Olympus IX-83 inverted microscope running Metamorph. Images were analyzed using Metamorph, Fluoview software, and/or FIJI (142)/ImageJ (W. Rasband, Research Services Branch, National Institute of Mental Health, Bethesda).

Data analysis and statistics
One-way ANOVA was performed in GraphPad Prism with appropriate post-tests as indicated; p Ͻ 0.05 was considered statistically significant. For comparisons of all samples within a data set, Tukey-Kramer post-test was used. For preselected pairwise comparisons, Bonferroni post-test was performed. For comparisons to a single control group, one-way ANOVA with Dunnett's post-test was used. All other data analysis was performed in Excel. For all figures, one asterisk (*) indicates p Ͻ 0.05 and two asterisks or pound signs (** or ##) indicate p Ͻ 0.01, respectively; "n.s." indicates no statistical significance. All data are presented as mean Ϯ S.E.