Acute Regulation of Tight Junction Ion Selectivity in Human Airway Epithelia

Acute regulation of tight junction ion selectivity in human airway epithelia

Andrea N. Flynn, Omar A. Itani, Thomas O. Moninger, and Michael J. Welsh1

Howard Hughes Medical Institute, Departments of Internal Medicine and Molecular Physiology and Biophysics, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, IA 52242

Contributed by Michael J. Welsh, December 31, 2008 (sent for review December 17, 2008) Electrolyte transport through and between airway epithelial cells erties of individual claudins (5, 8, 9). However, because most controls the quantity and composition of the overlying liquid. epithelia express multiple different claudins, the properties Many studies have shown acute regulation of transcellular ion conferred by individual claudins remain uncertain and appear to transport in airway epithelia. However, whether ion transport depend on context, including the cell type and presence of other through tight junctions can also be acutely regulated is poorly claudins. Moreover, claudins form homooligomers and heterooli- understood both in airway and other epithelia. To investigate the gomers, both within the same cell and across adjacent cells (8, 10). paracellular pathway, we used primary cultures of differentiated Regulation of tight junction function is also poorly under- human airway epithelia and assessed expression of claudins, the stood. Previous work on individual claudins showed that several primary determinants of paracellular permeability, and measured contain a C terminus subject to phosphorylation, which may .(transepithelial electrical properties, ion fluxes, and La3؉ move- modify claudin localization or paracellular permeability (11–14 ment. Like many other tissues, airway epithelia expressed multiple However, for the reasons outlined above, ascribing causal mech- claudins. Moreover, different cell types in the epithelium expressed anisms to specific modifications has remained difficult. Acute the same pattern of claudins. To evaluate tight junction regulation, regulation of tight junction function is even less well understood. ϩ we examined the response to histamine, an acute regulator of One example is apical Na -coupled glucose entry into small airway function. Histamine stimulated a rapid and transient in- intestinal epithelial cells, which triggered an acute increase in -crease in the paracellular Na؉ conductance, with a smaller increase paracellular permeability (15, 16). The increase involved disrup ؊ in Cl conductance. The increase was mediated by histamine H1 tion of tight junction integrity with contraction of the actomyosin .receptors and depended on an increase in intracellular Ca2؉ con- ring and dilations within the tight junctions centration. These results suggest that ion flow through the para- Previous work in airway epithelia showed that several hor- cellular pathway can be acutely regulated. Such regulation could mones, peptides, neurotransmitters, and other agents regulate facilitate coupling of the passive flow of counter ions to active electrolyte transport through the cellular pathway (1–3, 17). In transcellular transport, thereby controlling net transepithelial salt many cases, ligands alter transcellular ion transport within and water transport. seconds. Because both the cellular and paracellular pathways determine net transport, we hypothesized that the paracellular claudin ͉ histamine ͉ paracellular pathway pathway might be regulated in parallel with the cellular pathway. To test this hypothesis, we chose histamine as an agonist. First, irway epithelia determine the quantity and composition of it has a rapid and transient local effect after its release by mast Athe overlying liquid (1–3). Two parallel pathways control cells (18). Second, it can acutely regulate active transcellular ion transepithelial electrolyte transport, the cellular pathway and the transport (19, 20). Third, it is a physiological regulator of epithelial function and may be involved in airway disease (21). paracellular pathway. The cellular pathway comprises the epi- Fourth, previous studies indicated that histamine acutely in- thelial cells with their distinct apical and basolateral membranes creases transepithelial electrical conductance (G ) in airway PHYSIOLOGY containing channels, transporters, and pumps. Together, they t epithelia (19, 22). In endothelia and cell lines, activation of generate active (and sometimes passive) transepithelial ion histamine receptors disrupts the interaction of vascular endo- transport. The paracellular pathway provides a route for passive thelial-cadherin with the vimentin cytoskeleton to increase transepithelial transport, with ions moving in either direction paracellular conductance (G ) (23). Those studies focused on driven solely by their electrochemical gradient. The permeability p histamine’s effect on adherens junctions, but whether tight and ion selectivity of the paracellular pathway is critical for junction function is also altered was not addressed. Therefore, in establishing or dissipating ion concentration gradients and hence this study we tested the hypothesis that histamine might acutely for determining the ionic composition of the apical compart- regulate tight junction permeability in airway epithelia. ment and net volume flow. Thus, cellular and paracellular pathways must work in concert, being functionally matched to Results meet the transport requirements of a specific tissue. Differentiated Human Airway Epithelia Express Multiple Claudins. We The paracellular pathway is formed by tight junctions, located studied primary cultures of differentiated human airway epithe- near the apical side of the cells, and the lateral intercellular lia (24). After Ϸ2-weeks culture at the air–liquid interface, spaces, where adjacent epithelial cells interdigitate and are held epithelia contain basal cells, ciliated cells, goblet cells, and together by scattered adherens junctions (4–6). Tight junctions form the functional and structural boundary that separates apical and basolateral compartments. Importantly, they also Author contributions: A.N.F., O.A.I., T.O.M., and M.J.W. designed research; A.N.F., O.A.I., determine the ion transport properties of the paracellular path- and T.O.M. performed research; A.N.F., O.A.I., T.O.M., and M.J.W. analyzed data; and A.N.F. way. Tight junctions are comprised of a complex of proteins that and M.J.W. wrote the paper. includes claudins, which determine the ion selectivity and con- The authors declare no conflict of interest. ductance of the paracellular pathway. Freely available online through the PNAS open access option. Identification of claudin mutations associated with human 1To whom correspondence should be addressed. E-mail: [email protected]. disease have emphasized their importance in maintaining barrier This article contains supporting information online at www.pnas.org/cgi/content/full/ function and ion selectivity (7). Overexpression, knockout, 0813393106/DCSupplemental. knockdown, and mutagenesis studies have suggested the prop- © 2009 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0813393106 PNAS ͉ March 3, 2009 ͉ vol. 106 ͉ no. 9 ͉ 3591–3596 Downloaded by guest on October 2, 2021 +RT A -RT 40

30 Ct 20

B i ii iii

Fig. 2. Claudins 1, 3, and 7 localize to tight junctions between ciliated cells (A–C) and goblet cells (D–F). Data are en face projections of Alexa Fluor 488 bound to rabbit anti-claudins 1, 3, and 7 (green) and Alexa Fluor 568 bound to antiacetylated ␣-tubulin, a ciliated cell marker (A–C; red), or to Muc5AC, a iv v vi goblet cell marker (D–F; red). Arrows point to one example of tight junctions between neighboring ciliated or goblet cells that express claudins 1, 3, or 7. Data are stacks of images, and therefore the claudin immunostaining appears slightly blurry in some panels. 60ϫ magniﬁcation. Experiments were performed twice with epithelia from 2 different donors.

examined junctions between 2 identical cell types; if a claudin appeared at this location, we concluded that the specific cell type expressed it. We identified ciliated cells with antibodies to Fig. 1. Differentiated human airway epithelia express multiple different acetylated ␣-tubulin (27) and goblet cells with antibodies to claudins. (A) Real-time RT-PCR of primary cultures of differentiated airway epithelia. Ct indicates cycle threshold. The black and light gray bars indicate mucin Muc5AC (28). Claudins 1, 3, and 7 localized to tight samples with and without reverse transcriptase (RT), respectively. (B) Immu- junctions between neighboring ciliated cells and between neigh- nocytochemistry revealed staining of claudins 1, 3, 4, and 7 in a pattern boring goblet cells and to junctions between nonidentical cell consistent with tight junction localization (i–iv). Data are en face (Upper) and types (Fig. 2). Thus, both ciliated and goblet cells express X-Z projections (Lower) of Alexa Fluor 488 bound to rabbit anti-claudin claudins 1, 3, and 7. antibodies (green). ToPro-3 stained nuclei are in blue. Dotted line indicates position of ﬁlter. (v and vi) Immunostaining of claudin 2 was not detected in The Airway Epithelial Paracellular Pathway Is Cation-Selective. To airway epithelia (v) even though it could be detected in MDCKII cells (vi). Data investigate the paracellular pathway, we sought to eliminate the in v and vi are Alexa Fluor 488 bound to mouse anti-claudin 2 antibodies (green), and ethidium bromide-stained nuclei are in red. 40ϫ magniﬁcation. contribution of ion flow through the transcellular pathway. Experiments were repeated 3 times with epithelia from 3 different donors. Therefore, we used cystic fibrosis (CF) epithelia that lack CF transmembrane conductance regulator (CFTR) anion channels, and we added amiloride to inhibit epithelial Naϩ channels nonciliated, nongoblet columnar cells. To assess the claudin expression profile, we used RT-PCR and detected transcripts for (ENaC). Thus, unless otherwise noted, all functional studies occludin and claudins 1, 3, 4, 7, and 16 and variant 2 of claudin used CF epithelia; CF epithelia express the same pattern of 10, but not claudin 2 or variant 1 of claudin 10 (Fig. 1A). As an claudins by immunostaining and have a similar paracellular ion selectivity. We measured the relative selectivity of the paracel- additional test for airway epithelial claudins, we used immuno- ϩ Ϫ cytochemistry (Fig. 1B). When viewed en face, claudins 1, 3, 4, lular pathway for Na vs. Cl (PNa/PCl) by reducing the apical and 7 appeared in a cobblestone pattern typical of tight junction bathing solution NaCl concentration from 135 to 60 mM (Fig. 3A proteins (5, 6, 8). In vertical sections, they localized at the apical and B). We measured the changes in Vt under open-circuit side of tight junction cell–cell contacts and along the lateral cell conditions and used them to calculate PNa/PCl. Under baseline membrane, as reported for other epithelia (5, 6, 8). There was conditions, the paracellular pathway was cation-selective (Fig. 3 little, if any, claudin expression in basal cells. We did not detect C–E), which is consistent with an earlier report (25). claudin-2 expression in airway epithelia, although we detected it Na Cl in Madin-Darby canine kidney (MDCK) II cells used as a Histamine Transiently Increases GP and P /P . When we added positive control. Thus, as with many other epithelia (6, 8), airway histamine, Gt increased to a peak within 45 s and then returned epithelia express multiple claudins. Our data agree with earlier to baseline values (Fig. 3 A–C and Fig. S1). In these epithelia studies reporting that primary cultures of airway epithelia ex- lacking CFTR and with ENaC inhibited, histamine-induced press claudins 1, 3, 4, and 7, but not claudins 2, 5, 6, 9, 10, 11 or changes in G reflect predominantly changes in G . Histamine 15 (25, 26). One difference is that we detected claudin-10 variant t p 2 (but not variant 1) and claudin 16, whereas they were not had little effect on Vt. Although there was substantial variability detected in another study (25). The reason for this minor between cultures from different donors, histamine transiently discrepancy is uncertain, but might relate to the low level of increased Gt in all epithelia. claudin expression and limitations of antibody detection. Histamine also increased PNa/PCl (Fig. 3D). The increase was rapid and transient and paralleled the increase in estimated Gp. Na Cl Distinct Airway Epithelial Cell Types Express the Same Claudins. We used the changes in Gt and P /P to calculate the estimated ϩ Ϫ Na Cl Expression of multiple claudins raised the question of whether paracellular Na and Cl conductances (G p and G p) and Na Cl distinct cell types express the same claudins. As a test, we found that G p nearly doubled, whereas G p increased by Ϸ1/3

3592 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0813393106 Flynn et al. Downloaded by guest on October 2, 2021 Na Cl Fig. 4. Stimulation of the H1 receptor mediates the increase in Gt and P /P . Epithelia were treated with the histamine H1 receptor antagonist pyrilamine (100 ␮M basolateral), the H2 receptor antagonist cimetidine (100 ␮M basolateral), or neither as indicated. Histamine (100 ␮M) was added basolaterally as indicated by gray bars. Data are means Ϯ SEM (n ϭ 3). *, P Ͻ 0.05 compared with control.

Previous studies showed that activation of the histamine H1 receptor causes a rapid transient increase in the intracellular 2ϩ 2ϩ Ca concentration ([Ca ]I) in airway epithelia (30). In contrast, activation of the histamine H2 receptor primarily increased intracellular cAMP (31). Those observations plus the rapid time Na 2ϩ course of the G p changes suggested that a [Ca ]i elevation was Na Cl responsible for the increase in estimated Gp and P /P .Totest this hypothesis, we treated epithelia with the Ca2ϩ chelator ϩ Fig. 3. Histamine increases paracellular Na permeability. (A and B) Tracings 1,2-bis-(o-aminophenoxy)-ethane-N,N,NЈ,NЈ-tetraacetic acid of Vt in response to repeated reductions of apical solution NaCl concentration tetraacetoxymethyl ester (BAPTA-AM) and found that it elim- from 135 to 60 mM. Epithelia were treated with basolateral vehicle control (A) inated the effect of histamine on G and PNa/PCl (Fig. 5 A and B). or 100 ␮M histamine (B) during time indicated by bars. (C–E) Experiments like t those shown in B were used to calculate G (C), PNa/PCl (D), and GNa and GCl (E). Likewise, treating epithelia with thapsigargin to deplete intra- t 2ϩ In C and D, each set of symbols and lines is from a different experiment. cellular Ca stores abolished the response to histamine. These

(Fig. 3E). We found no relationship between either basal Gt or Na Cl Na Cl P /P and the histamine-induced changes in Gt,P /P ,or individual conductances. Although airway epithelia can express Ca2ϩ-activated ClϪ channels, the small transient Vt depolarization after histamine application was in the opposite direction expected for activation ofaClϪ channel. For example, activating the CFTR ClϪ channel in non-CF epithelia hyperpolarized Vt (Fig. S2). As a control, we repeated the experiments in the presence of 4,4Ј-diisothiocyano- Ј

stilbene-2,2 -disulfonic acid (DIDS), which inhibits these chan- PHYSIOLOGY nels, and found that the effect of histamine on estimated Gp was unaltered (Fig. S3). If there were any component of Ca2ϩ- activated ClϪ conductance, these calculations would overesti- Cl mate the increase in G p. We also considered that if histamine simply disrupted tight junction integrity, then PNa/PCl would fall rather than increase. Consistent with this prediction, we treated epithelia with EGTA (5 mM) to chelate Ca2ϩ, disrupt E- cadherin-based cell–cell contacts, and increase Gp (29). We applied EGTA only briefly so that it increased Gt to approxi- mately the same extent as occurred with histamine (Fig. S4). As predicted, PNa/PCl fell. In addition, the EGTA-induced changes did not rapidly reverse. These data indicate that the increased PNa/PCl is not caused by a nonspecific disruption of the paracellular pathway, but rather are caused by a specific effect on tight junctions.

؉ Histamine Interacts with the H1 Receptor to Increase Paracellular Na 2ϩ Conductance. To determine which histamine receptor increases Fig. 5. An increase in [Ca ]i is required and sufﬁcient to increase estimated Na Cl ␮ G and PNa/PCl, we used the histamine H antagonist pyrilamine Gp and P /P .(A and B) Epithelia were treated with BAPTA-AM (25 M, apical p 1 ␮ and H antagonist cimetidine. Neither antagonist alone had an and basolateral, 1 h), thapsigargin (100 M, basolateral, 15 min), or vehicle 2 control for 1 h before adding histamine (100 ␮M, basolateral). Data are effect on the paracellular pathway (Fig. S5). Although cimeti- means Ϯ SEM (n ϭ 3). (C and D) Ionomycin (100 ␮M, apical), forskolin (100 ␮M, dine had a partial effect, pyrilamine abrogated the effect of apical) plus IBMX (100 ␮M, apical), or histamine (100 ␮M, basolateral) were Na Cl Na Cl histamine on estimated Gp and P /P (Fig. 4). These results are added to epithelia, and Gt and P /P were determined as indicated above. (E) consistent with an earlier study suggesting that histamine H1 GNa and GCl were calculated from results in C and D for epithelia treated with receptors are responsible for changes in Gt (19, 22). ionomycin.

Flynn et al. PNAS ͉ March 3, 2009 ͉ vol. 106 ͉ no. 9 ͉ 3593 Downloaded by guest on October 2, 2021 Fig. 6. Histamine increases paracellular Na flux. (A) Apical to basolateral and basolateral to apical unidirectional fluxes of 22Na were measured in CF epithelia in the presence of 100 ␮M amiloride. Fluxes are from a 25-min period before histamine addition, a 1-min period immediately after histamine (100 ␮M), and the subsequent 25-min period. Data are means Ϯ SEM. n ϭ 10 basolateral to apical and 12 apical to basolateral. *, P Ͻ 0.05 compared with the pretreatment flux. (B)Naϩ and ClϪ fluxes in non-CF epithelia treated with 100 ␮M amiloride and 10 ␮M GlyH-101. Timing of fluxes was the same as in A. n ϭ 8 epithelia; the basolateral to apical and apical to basolateral unidirectional fluxes for Naϩ and for ClϪ did not differ and were combined.

2ϩ results suggest that a histamine-stimulated increase in [Ca ]i Na Cl increased estimated Gp and P /P . We tested this hypothesis directly by applying the Ca2ϩ ionophore ionomycin. Ionomycin Na Cl ϩ increased estimated Gp,P /P , and Na conductance (Fig. 5 C–E). In contrast, adding forskolin and 3-isobutyl-1-methylxan- thine (IBMX) to increase cellular levels of cAMP failed to alter Na Cl Gp or P /P . Thus, interventions that prevent an increase in 2ϩ [Ca ]i inhibited the response to histamine and increasing 2ϩ Na Cl [Ca ]i increased estimated Gp and P /P .

Histamine Increases Passive Transepithelial Na؉ Flux. Finding that histamine augmented paracellular cation conductance predicted that it would increase transepithelial Naϩ fluxes. In CF epithelia treated with amiloride, the Naϩ fluxes from the apical to basolateral surface and from the basolateral to apical surface were not significantly different, and thus net Naϩ flux was not different from 0 (Fig. 6A). Histamine induced an abrupt and Fig. 7. Histamine increases tight junction permeability to La3ϩ.(A–D) Epi- ϩ transient increase in unidirectional Na flux without inducing a thelia were treated with vehicle control (A), 100 ␮M histamine for 45 s (B), 100 net Naϩ flux. ␮M histamine for 10 min (C), or 5 mM EGTA (D). Three examples of each We also repeated the flux experiments in non-CF epithelia condition are shown. When La3ϩ penetrates the tight junction, the lateral treated with GlyH-101 to inhibit CFTR ClϪ channels (32); we intercellular space shows electron dense material. (Scale bars: 1.4 ␮m.) (E) 3ϩ ϭ used non-CF epithelia because of limited availability of CF Percentage of tight junctions penetrated by La . n 59–67 for each of the various conditions. , P Ͻ 0.05. epithelia. There was no net Naϩ or ClϪ flux and the 2 unidi- * rectional fluxes were combined. Histamine transiently increased ϩ Ϫ transepithelial Na but not Cl flux (Fig. 6B). These flux fluxes, by 10 min after histamine treatment the percentage of experiments support the conclusion that histamine transiently La3ϩ-labeled paracellular pathways had decreased to 19%. As a ϩ increases tight junction permeability to Na . They are also control, we treated epithelia with EGTA to open tight junctions consistent with an earlier study of canine tracheal epithelium (29) and found that La3ϩ labeled 54% of lateral intercellular under short-circuit conditions, which showed that histamine spaces. ϩ increased the passive (basolateral to apical) Na flux without Examination by TEM revealed no detectable alterations in the Ϫ altering the passive (apical to basolateral) Cl flux (19). morphology of the tight junctions or lateral intercellular spaces (Fig. 7). As an independent assessment for altered structure, we Histamine Increases Tight Junction Permeability to Lanthanum With- immunostained epithelia for ZO-1, claudin 1, and claudin 7, and out Inducing Morphological Alterations. As an independent method we examined rhodamine–phalloidin staining of F-actin. Hista- of testing the effect of histamine on paracellular permeability, we mine did not alter their appearance or distribution (Fig. S6). ϩ examined the movement of lanthanum (La3 ) through tight Thus, histamine increased tight junction permeability to a junctions by using transmission electron microscopy (TEM) (33). polyvalent cation without inducing morphological alterations La3ϩ is an electron dense element with a hydrated radius (0.4 that were detectable by TEM or a redistribution of several tight nm) similar to that of Naϩ (0.3 nm). Under basal conditions, junction-associated proteins. La3ϩ crossed tight junctions and moved into the lateral intercellular space in 13% of junctions (Fig. 7). However, 45 s after Discussion adding histamine, we found that La3ϩ crossed 32% of tight Our data indicate that tight junctions can rapidly change their junctions to label the lateral intercellular spaces. In keeping with selectivity for cations vs. anions with histamine stimulation. The ϩ ϩ ϩ the transient nature of the increase in estimated Gp and Na increased paracellular Na conductance, transepithelial Na

3594 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0813393106 Flynn et al. Downloaded by guest on October 2, 2021 flux, and La3ϩ permeability all indicate a substantial transient It is interesting that different cell types in the airway epithe- elevation of paracellular Naϩ permeability without morpholog- lium expressed the same claudins. Because all cells touching the ical changes. This enhanced selectivity points to a tight junction apical surface are exposed to the same ionic composition of function different from previous studies exploring interventions airway surface liquid, a similar claudin expression profile would that disrupt tight junction integrity. ensure that the paracellular pathway makes a uniform contri- Several observations argue that alterations in tight junctions bution to maintaining transepithelial ion and liquid transport. At were responsible for the histamine-induced changes. First, we present, knowledge of claudin function is insufficient to predict used several methods (transepithelial dilution potentials, transepithelial permeability properties. However, of the claudins we epithelial passive ion fluxes, and tight junction penetration of detected in airway epithelia, claudin 7 (41) and 16 (42) increase La3ϩ), and all suggested that histamine increased paracellular cation selectivity, whereas claudin 4 (43) reduces cation selec- permeability. Second, the tight junction is the epithelial struc- tivity. Hopefully, future studies will allow predictions of selec- ture that determines paracellular ion selectivity, and thus in- tivity and discover how direct or indirect claudin modifications creased ion selectivity implicates changes in tight junction func- acutely alter claudin function. tion. In contrast, when we disrupted tight junctions and Airway epithelia exhibit 2 main transcellular processes that 2ϩ Na Cl ϩ increased Gp by chelating extracellular Ca ,P /P fell. Third, drive active ion transport, ENaC-dependent Na absorption and an increase in transcellular conductance does not explain the CFTR-dependent anion secretion (1–3, 29). Under open-circuit Ϫ changes. CFTR Cl channels were lacking in these CF epithelia, conditions, net cationic and anionic fluxes across the epithelium ϩ and amiloride inhibited ENaC Na channels. An increase in must be equal, and the paracellular pathway therefore plays an 2ϩ 2ϩ Ϫ [Ca ]i can activate apical membrane Ca -activated Cl chan- important role in determining net salt movement. When ENaC- nels, however, that would have reduced, rather than increased dependent Naϩ absorption dominates cellular transport, increas- Na Cl Na ϩ P /P , and we obtained similar responses in the presence of ing G p would minimize net NaCl absorption by shunting Na DIDS, which inhibits these channels. Fourth, the effect of back across the tight junctions toward the lumen, driven by the histamine would not be solely explained by changes in the lateral Ϫ Ϫ Vt. Conversely, when CFTR-dependent Cl or HCO3 secretion Na intercellular space, which lies in a series arrangement with the dominates cellular transport, increasing G p would maximize tight junctions. The width of the lateral intercellular space can net salt secretion by providing a pathway for Naϩ to accompany Ϫ influence GP (34). For example, histamine has been reported to Cl . Thus, through its regulation of tight junctions, histamine loosen adherens junctions in endothelia (35). However, the released from mast cells might transiently and locally increase lateral intercellular space is a watery pathway without ion the amount of NaCl, and hence liquid, on the apical surface, selectivity (36). Thus, increased ion flow through this space thereby facilitating mucociliary clearance. should not increase ion selectivity. Moreover, we observed no Finding acute regulation of tight junction permeability and morphological changes in the lateral intercellular space (or tight selectivity suggests the potential for involvement in disease. For junctions) like those reported to occur when cadherin interac- ϩ example, histamine-dependent changes might contribute to the tions are disrupted by Na -coupled transport in small intestine pathophysiology of asthma. In cystic fibrosis, we speculate that 2ϩ (15) or Ca chelation in MDCK epithelia (29). variations in the genes encoding tight junction proteins might Although other studies have examined tight junction regula- affect the clinical phenotype. Finally, manipulations of tight tion in airway epithelia, the changes and mechanisms differ from junction function could potentially be a therapeutic strategy in our findings. Inflammatory cytokines increased Gp over 24–72 airway disease, including cystic fibrosis. h, but they decreased tight junction selectivity and altered tight junction morphology (37). In contrast to our data, azithromycin Materials and Methods and shear stress decreased Gp. The former occurred over the Detailed materials and methods are provided in SI Text. course of days in association with altered processing of tight junction proteins (26), and the latter required serial activation of Primary Cultures of Differentiated Human Airway Epithelia. We used previously

the transient receptor potential vanilloid-4 channel and L-type described methods (24) to culture differentiated human airway epithelia at PHYSIOLOGY voltage-gated calcium channel (38). Studies in other epithelia the air–liquid interface. We used CF airway epithelia for all functional studies have also revealed regulation of tight junction permeability. As except where specifically noted. See SI Text. indicated above, Naϩ-coupled solute entry into intestinal epi- RT-PCR. We used standard methods for RT-PCR. See SI Text. thelia increases Gp by disrupting tight junctions, which would be expected to reduce any ion selectivity (15). Another example is the aldosterone-induced increase in claudin 4 phosphorylation Immunofluorescence. We used standard procedures for immunocytochemis- and paracellular ClϪ permeability (14, 39). An example of a large try. See SI Text. (9-fold) increase in G comes from studies of the Malpighian p Transepithelial Electrophysiology. Epithelia were studied in Ussing chambers as tubule of the mosquito Aedes aegypti where leukokinin stimu- described (24) under open-circuit conditions. G was calculated from the Cl t lates a rapid and reversible increase in G p (40). change in V induced by constant current pulses. Amiloride (100 ␮M) was 2ϩ t Our results indicate that changes in [Ca ]i were required and added apically, and histamine (0.1 mM) was added basolaterally. To assess ϩ Na 2 Na Cl sufficient to increase G p. The end target of the elevated [Ca ]i P /P , the NaCl concentration of the apical solution was reduced to 60 mM, is almost certainly airway epithelial claudins because they es- and mannitol was added to maintain osmolarity. Changes in Vt induced by tablish the ion selectivity and conductance of the tight junctions. solution changes were used to calculate PNa/PCl by using the Goldman- 2ϩ Na Cl However, the processes that lie between [Ca ]i and the claudins Hodgkin-Katz equation. We calculated paracellular G and G from the 2ϩ Kimizuka–Koketsu equation (44). See SI Text. remain uncertain. It could be that an elevated [Ca ]i leads to phosphorylation of claudins or modification of one or more of the many proteins associated with claudins and tight junctions. Transepithelial Ion Fluxes. We used standard techniques to measure ion fluxes. Deciphering the steps involved may prove challenging based on See SI Text. the number of claudins expressed by airway epithelia and the TEM. Epithelia were treated with vehicle, histamine, or EGTA, fixed, and then difficulty associated with attempts to understand how and which incubated with apical La3ϩ. After washing they were processed for TEM. The claudins contribute to tight junction function in other epithelia investigator was blinded for quantification of labeled junctions. See SI Text. (5, 6, 8). Nevertheless, our data obtained with histamine suggest 2ϩ that other agonists that elevate [Ca ]i might also increase ACKNOWLEDGMENTS. We thank Philip Karp, Tamara Nesselhauf, Pamela paracellular pathway conductance. Hughes, and Theresa Mayhew for excellent assistance; Joseph Zabner, Michael

Flynn et al. PNAS ͉ March 3, 2009 ͉ vol. 106 ͉ no. 9 ͉ 3595 Downloaded by guest on October 2, 2021 Winter, and D. Michael Shasby for helpful discussions; John Widness and Fibrosis Foundation Grants R458-CR02 and ENGLH9850, National Institute of Robert Schmidt for use of their gamma counter; and the In Vitro Models and Diabetes and Digestive and Kidney Disease Grant DK54759, and National Cell Culture Core (supported in part by National Heart, Lung, and Blood Institutes of Health Grant HL61234. M.J.W. is an Investigator of the Howard Institute Grant HL51670) for cell culture. This work was supported by Cystic Hughes Medical Institute.

1. Welsh MJ (1987) Electrolyte transport by airway epithelia. Physiol Rev 67:1143–1184. 25. Coyne CB, Gambling TM, Boucher RC, Carson JL, Johnson LG (2003) Role of claudin 2. Pilewski JM, Frizzell RA (1999) Role of CFTR in airway disease. Physiol Rev 79(Suppl interactions in airway tight junctional permeability. Am J Physiol 285:L1166– 1):S215–S255. L1178. 3. Verkman AS, Song Y, Thiagarajah JR (2003) Role of airway surface liquid and submu- 26. Asgrimsson V, Gudjonsson T, Gudmundsson GH, Baldursson O (2006) Novel effects of cosal glands in cystic fibrosis lung disease. Am J Physiol 284:C2–C15. azithromycin on tight junction proteins in human airway epithelia. Antimicrob Agents 4. Farquhar MG, Palade GE (1963) Junctional complexes in various epithelia. J Cell Biol Chemother 50:1805–1812. 17:375–412. 27. Piperno G, Fuller MT (1985) Monoclonal antibodies specific for an acetylated form of 5. Tsukita S, Furuse M (2002) Claudin-based barrier in simple and stratified cellular sheets. ␣-tubulin recognize the antigen in cilia and flagella from a variety of organisms. J Cell Curr Opin Cell Biol 14:531–536. Biol 101:2085–2094. 6. Van Itallie CM, Anderson JM (2006) Claudins and epithelial paracellular transport. 28. Hovenberg HW, Davies JR, Herrmann A, Linden CJ, Carlstedt I (1996) MUC5AC, but not Annu Rev Physiol 68:403–429. 7. Van Itallie CM, Anderson JM (2004) The molecular physiology of tight junction pores. MUC2, is a prominent mucin in respiratory secretions. Glycoconj J 13:839–847. Physiology 19:331–338. 29. Martinez-Palomo A, Meza I, Beaty G, Cereijido M (1980) Experimental modulation of 8. Angelow S, Ahlstrom R, Yu AS (2008) Biology of claudins. Am J Physiol 295:F867–F876. occluding junctions in a cultured transporting epithelium. J Cell Biol 87:736–745. 9. Heiskala M, Peterson PA, Yang Y (2001) The roles of claudin superfamily proteins in 30. Harris RA, Hanrahan JW (1994) Effects of EGTA on calcium signaling in airway epithelial paracellular transport. Traffic 2:93–98. cells. Am J Physiol 267:C1426–C1434. 10. Daugherty BL, Ward C, Smith T, Ritzenthaler JD, Koval M (2007) Regulation of hetero- 31. Leurs R, Smit MJ, Timmerman H (1995) Molecular pharmacological aspects of histamine typic claudin compatibility. J Biol Chem 282:30005–30013. receptors. Pharmacol Ther 66:413–463. 11. D’Souza T, Agarwal R, Morin PJ (2005) Phosphorylation of claudin-3 at threonine 192 32. Muanprasat C, et al. (2004) Discovery of glycine hydrazide pore-occluding CFTR by cAMP-dependent protein kinase regulates tight junction barrier function in ovarian inhibitors: Mechanism, structure-activity analysis, and in vivo efficacy. J Gen Physiol cancer cells. J Biol Chem 280:26233–26240. 124:125–137. 12. Tanaka M, Kamata R, Sakai R (2005) EphA2 phosphorylates the cytoplasmic tail of 33. Friend DS, Gilula NB (1972) Variations in tight and gap junctions in mammalian tissues. claudin-4 and mediates paracellular permeability. J Biol Chem 280:42375–42382. J Cell Biol 53:758–776. 13. Kahle KT, et al. (2004) Paracellular ClϪ permeability is regulated by WNK4 kinase: Insight 34. Smulders AP, Tormey JM, Wright EM (1972) The effect of osmotically induced water into normal physiology and hypertension. Proc Natl Acad Sci USA 101:14877–14882. flows on the permeability and ultrastructure of the rabbit gallbladder. J Membr Biol 14. Yamauchi K, et al. (2004) Disease-causing mutant WNK4 increases paracellular chloride 7:164–197. permeability and phosphorylates claudins. Proc Natl Acad Sci USA 101:4690–4694. 35. Winter MC, et al. (1999) Histamine alters cadherin-mediated sites of endothelial 15. Pappenheimer JR (1990) Paracellular intestinal absorption of glucose, creatinine, and mannitol in normal animals: Relation to body size. Am J Physiol 259:G290–G299. adhesion. Am J Physiol 277:L988–L995. 16. Madara JL, Pappenheimer JR (1987) Structural basis for physiological regulation of 36. Kottra G, Fromter E (1983) Functional properties of the paracellular pathway in some paracellular pathways in intestinal epithelia. J Membr Biol 100:149–164. leaky epithelia. J Exp Biol 106:217–229. 17. Boucher RC (2003) Regulation of airway surface liquid volume by human airway 37. Coyne CB, et al. (2002) Regulation of airway tight junctions by proinflammatory epithelia. Pfluuërs Arch 445:495–498. cytokines. Mol Biol Cell 13:3218–3234. 18. Franzen, Norrby K (1980) Local mitogenic effect of tissue mast cell secretion. Cell Tissue 38. Sidhaye VK, Schweitzer KS, Caterina MJ, Shimoda L, King LS (2008) Shear stress Kinet 13:635–642. regulates aquaporin-5 and airway epithelial barrier function. Proc Natl Acad Sci USA 19. Marin MG, Davis B, Nadel JA (1977) Effect of histamine on electrical and ion transport 105:3345–3350. properties of tracheal epithelium. J Appl Physiol 42:735–738. 39. Le Moellic C, et al. (2005) Aldosterone and tight junctions: Modulation of claudin-4 20. Shen BQ, Finkbeiner WE, Wine JJ, Mrsny RJ, Widdicombe JH (1994) Calu-3: A human phosphorylation in renal collecting duct cells. Am J Physiol 289:C1513–C1521. Ϫ airway epithelial cell line that shows cAMP-dependent Cl secretion. Am J Physiol 40. Beyenbach KW (2003) Regulation of tight junction permeability with switch-like 266:L493–L501. speed. Curr Opin Nephrol Hypertens 12:543–550. 21. Thurmond RL, Gelfand EW, Dunford PJ (2008) The role of histamine H1 and H4 41. Alexandre MD, Lu Q, Chen YH (2005) Overexpression of claudin-7 decreases the receptors in allergic inflammation: The search for new antihistamines. Nat Rev Drug paracellular ClϪ conductance and increases the paracellular Naϩ conductance in Discov 7:41–53. LLC-PK1 cells. J Cell Sci 118:2683–2693. 22. Zabner J, et al. (2003) Histamine alters E-cadherin cell adhesion to increase human 42. Hou J, et al. (2008) Claudin-16 and claudin-19 interact and form a cation-selective tight airway epithelial permeability. J Appl Physiol 95:394–401. 23. Shasby DM, Ries DR, Shasby SS, Winter MC (2002) Histamine stimulates phosphoryla- junction complex. J Clin Invest 118:619–628. tion of adherens junction proteins and alters their link to vimentin. Am J Physiol 43. Van Itallie C, Rahner C, Anderson JM (2001) Regulated expression of claudin-4 de- 282:L1330–L1338. creases paracellular conductance through a selective decrease in sodium permeability. 24. Karp PH, et al. (2002) An in vitro model of differentiated human airway epithelia: J Clin Invest 107:1319–1327. Methods and evaluation of primary cultures. Epithelial Cell Culture Protocols,edWise 44. Kimizuka H, Koketsu K (1964) Ion transport through cell membrane. J Theor Biol C (Humana, Totowa, NJ), Vol 188, pp 115–137. 6:290–305.

3596 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0813393106 Flynn et al. Downloaded by guest on October 2, 2021