ORIGINAL ARTICLE. Potential Role of Abnormal Ion Transport in the Pathogenesis of Chronic Sinusitis

Transcription

1 ORIGINAL ARTICLE Potential Role of Abnormal Ion Transport in the Pathogenesis of Chronic Sinusitis Kenji Dejima, MD; Scott H. Randell, PhD; M. Jackson Stutts, PhD; Brent A. Senior, MD; Richard C. Boucher, MD Objectives: To create well-differentiated cultures of normal and chronic sinusitis paranasal sinus epithelial cells and to compare their electrophysiologic properties. Design: In vitro investigation using primary sinus epithelial cells, initially cultured on plastic tissue culture dishes. Cells were characterized by means of immunocytochemical analysis and then passaged to air-liquid interface culture conditions. The morphologic features of air-liquid interface cultures were assessed using light and electron microscopy. Epithelial Na channel, Na -K -2Cl cotransporter, cystic fibrosis transmembrane conductance regulator, and Ca 2 -activated Cl channel function were investigated in Ussing chambers. Subjects: Specimens were obtained from 15 patients undergoing transsphenoidal pituitary procedures, tumor removal, or trauma repair and from 9 patients with chronic sinusitis. Results: After culture at an air-liquid interface for 21 days, the epithelium was pseudostratified and contained basal, mucous secretory, and ciliated cells. There were no detectable morphologic differences between normal and chronic sinusitis cells. In cultures of normal cells, median basal short circuit current was 4.7 µa/cm 2, and Na transport, defined as the amiloride hydrochloride sensitive component, was approximately 2% of the total. Basal and amiloride-sensitive short circuit currents were greater in cultures of chronic sinusitis cells. Basal short circuit currents in both types of cultures were insensitive to the Cl transport inhibitor bumetanide, but all responded to forskolin or uridine triphosphate. After amiloride pretreatment, forskolin and uridine triphosphate responses were greater in chronic sinusitis cells. Conclusions: We established methods for welldifferentiated sinus epithelial cultures. The cells exhibited Na absorption and Cl secretion, and elevated rates of ion transport may be pathophysiologically relevant in chronic sinusitis. Arch Otolaryngol Head Neck Surg. 26;132: Author Affiliations: Department of Otorhinolaryngology, Kyoto Prefectural University of Medicine, Kyoto, Japan (Dr Dejima); and Department of Cell and Molecular Physiology (Dr Randell), Department of Medicine, Cystic Fibrosis/Pulmonary Research and Treatment Center (Drs Dejima, Randell, Stutts, and Boucher), and Department of Otolaryngology/Head and Neck Surgery (Dr Senior), The University of North Carolina at Chapel Hill. Dr Dejima is now with the Department of Otorhinolaryngology, Kyoto Second Red Cross Hospital, Kyoto. BILATERALLY SYMMETRICAL PAranasal sinuses are highly developed in humans compared with many other mammals. These sinuses are thought to enhance vocal resonance, warm and humidify inhaled air, reduce skull weight, protect intracranial and orbital structures from trauma, and perhaps generate nitric oxide to regulate airway or vascular tone. 1 However, their principal physiologic function has not been established. 2 Diseases of the sinuses, including sinusitis, are common globally, consuming substantial health care resources and resulting in large economic losses. Chronic sinusitis is a frequent occurrence and is resistant to therapy. Despite many basic and clinical studies, the pathogenesis of sinusitis is not well understood. Epithelial ion transport regulates hydration of respiratory mucosal surfaces, which, in turn, promotes effective mucociliary clearance. Thus, altered ion transport, and hence failure to clear bacteria or toxins from sinus surfaces, could play a role in the pathogenesis of sinus disease. However, the electrophysiologic properties of sinus epithelium in the normal or diseased state have not been elucidated. Air-liquid interface (ALI) cultures of airway epithelial cells reproduce key aspects of in vivo airway epithelial structure and function and have proved to be invaluable for understanding airway diseases such as cystic fibrosis. 3 Air-liquid interface cultures are routinely generated with epithelial cells derived from excised nasal turbinate, nasal polyp, and tracheobronchial tissue. However, to our knowledge, ALI cultures of normal and pathologic human paranasal sinus epithelia have not been reported. One reason for the absence of data may be the scarcity of adequate human paranasal sinus tissue specimens. Since the mid-198s, endoscopic sinus surgery has been the method of choice for the surgical treatment of sinus disease, except cancer. 4 Typically, endoscopically excised, pathologic mucosal tissues are small macerated fragments thought to be inadequate for culture, es- 1352

2 pecially compared with large nasal turbinate and polyp specimens. Furthermore, normal sinus mucosa was rarely available. However, in recent years, endoscopic transsphenoidal sinus surgery has been applied to the treatment of pituitary disease, increasing access to normal sinus mucosa. Furthermore, advances in cell culture techniques permit differentiation of cells from very small specimens, such as cytologic brushes 5 and murine tracheal epithelium. 6 Improved passaging of airway epithelial cells has also allowed expansion of cell numbers. 7 In view of the few studies of sinus epithelial cell cultures, the greater tissue availability, and the improved methods, our goal was to produce and characterize ALI cultures of human paranasal sinus epithelia. First we established conditions that produce mucociliary differentiation of normal paranasal sinus epithelial (NPSE) cells, and then we measured their ion transport properties, including basal rates of Na absorption and Cl secretion and pharmacologic regulation of Cl transport. These data were compared with those from epithelial cell cultures derived from patients with chronic sinusitis (CSE cells) to test the hypothesis that altered epithelial ion transport may contribute to the pathogenesis of sinusitis. Protease XIV Passage 1 Cells on Collagen-Coated Porous Support Primary Cells on Plastic METHODS PRIMARY CULTURE Normal paranasal sinus tissue specimens (n=15) were obtained from patients treated with endonasal surgery for pituitary tumors (n=1), benign sinonasal tumors (n=3), or facial trauma (n=2). Specimens were also obtained from patients with chronic sinusitis (n=9), excluding cystic fibrosis, bronchial asthma, allergic fungal sinusitis, and primary ciliary dyskinesia. The University of North Carolina Committee for the Rights of Human Subjects approved the use of excess surgical pathology tissue. Discrete tissue samples were obtained under endoscopic guidance in the operating room with the patient under general anesthesia. Cutting forceps were used to remove the tissue, and freshly removed tissue was placed in isotonic sodium chloride solution before transfer to the laboratory. Procedures for sinus epithelial cell culture were adapted from methods previously described for nasal turbinate and polyp tissues and tracheobronchial specimens 8 (Figure 1). Freshly excised sinus tissues were collected as small fragments and were incubated at 4 C for 3 to 8 hours in Joklik minimum essential medium containing.1% protease type XIV (Sigma-Aldrich Corp, St Louis, Mo),.1 mg/ml of deoxyribonuclease (Sigma-Aldrich Corp), and penicillin G sodium, streptomycin sulfate, gentamicin sulfate, and amphotericin B in the concentrations given later in this paragraph. Ten percent fetal bovine serum was added to neutralize the protease, and, because the epithelial surface was difficult to identify, cells were detached by gentle scraping of all tissue surfaces. The cells were resuspended, counted, and then plated at a density of.5 to cells/35-mm-, 1 to cells/6-mm-, and 3 to cells/1-mm-diameter collagencoated tissue culture dish in modified LHC9 (Laboratory of Human Carcinogenesis) medium, 9 termed bronchial epithelial growth medium. The modifications included increasing the epidermal growth factor concentration to 25 ng/ml, adjusting the retinoic acid and gentamicin concentrations to M and 5 µg/ml, respectively, and adding.5 mg/ml of bovine serum albumin,.8% bovine pituitary extract, 1 U/mL of penicillin, 1 µg/ml of streptomycin, and. 25 µg/ml of amphotericin B. 8 Collagencoated dishes were prepared by incubating dishes for 2 hours at Figure 1. A schematic illustration of the tissue harvest and cell culture process. Removal of mucosa using a cutting forceps instrument is shown at the upper left. Primary epithelial cells are dissociated from excised tissue fragments via protease action and gentle scraping and are grown on tissue culture plastic before subpassage to porous supports for culture at an air-liquid interface. 37 C with 4 µg/ml of a type I and type III collagen mixture (Inamed Biomaterials, Fremont, Calif) in distilled water in a volume to yield 1 µg of collagen per 1 cm 2 of surface area. The collagen solution was then aspirated, and the dishes were dried and sterilized for 1 minutes under UV light. Cultures were maintained at 37 C in an atmosphere of 5% carbon dioxide in air. Culture medium was changed on day 1 after seeding and every other day thereafter until the cultures reached approximately 75% confluence, at which time they were dissociated with trypsin-edta solution. Passage 1 cells were subcultured as described in the following subsection. ALI CULTURE Passage 1 cells (.1 to ) were plated in.2 ml of ALI culture medium on 12-mm,.4-µm-pore culture inserts (Transwell-Col or Snapwell; Costar Corp, Cambridge, Mass). The Snapwells were coated with human placental collagen type IV (product No. C7521; Sigma-Aldrich Corp) as described previously 8 and were used for Ussing chamber experiments. The ALI medium is similar to bronchial epithelial growth medium except that a 5:5 mixture of LHC basal and Dulbecco modified Eagle medium-h was used as the base, amphotericin B and gentamicin were omitted, and the epidermal growth factor concentration was reduced to.5 ng/ml. 8 On reaching confluence, the apical surface was rinsed with phosphate-buffered saline, and ALI medium was replaced only in the bottom compartment of the culture. Cultures were maintained at 37 C in an atmosphere of 5% carbon dioxide in air, and the medium was changed every other day. IMMUNOHISTOCHEMICAL ANALYSIS Cytospin slides were made with 5 to cells per slide from primary dissociations and passage 1 NPSE cells and were 1353

3 stored without fixation at 8 C until staining. Slides were stained with 1 of 4 primary antibodies, followed by biotinylated secondary antibodies, streptavidin horseradish peroxidase, diaminobenzidine color development, and light green counterstaining. We used the following antibodies: mouse monoclonal anti human pancytokeratin (1:2 dilution), mouse monoclonal anti human epithelial membrane antigen (1:1 dilution), mouse monoclonal anti human vimentin (1:16 dilution), and mouse monoclonal anti-cd45 (1:1 dilution) (all from Sigma-Aldrich Corp) and appropriate isotype controls. The number of dark brown cells in approximately 3 cells on each slide was scored to determine the mean percentage of positively stained cells. HISTOLOGIC ANALYSIS After 1 or 21 days, ALI cultures were fixed with 1% neutral buffered formalin and embedded in paraffin. Five-micrometerthick sections were cut and stained with hematoxylin-eosin or Alcian blue periodic acid Schiff (AB-PAS). These sections were used to measure epithelial thickness, the number of cell layers, the number of ciliated cells, and the number of cells containing AB-PAS positive granules (defined as mucous secretory cells). Five to 1 consecutive high-power fields per section were counted. TRANSMISSION ELECTRON MICROSCOPY Transmission electron microscopy was performed on cultured NPSE cells using a special fixation procedure designed to preserve the structure of airway surface liquid. 3 Briefly, cultures were fixed with perfluorocarbon osmium tetroxide and embedded in epon-araldite resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a transmission electron microscope. USSING CHAMBER STUDIES The electrophysiologic properties of ALI cultures of NPSE and CSE cells were measured as previously described. 1 Briefly, Snapwell cultures were mounted in Ussing chambers (Physiologic Instruments, La Jolla, Calif). Each bath contained warmed (37 C) Krebs bicarbonate Ringer solution (containing 14mM Na, 12mM Cl, 5.2mM K, 1.2mM Ca 2, 1.2mM Mg 2, 2.4mM HPO 4 2,.4mM H2PO 4, 25mM HCO 3, and 5mM glucose), circulated by gas lift with 95% oxygen and 5% carbon dioxide. Bioelectric properties were digitally recorded from the output of voltage clamps (Physiologic Instruments) using a software program (ACQUIRE; Physiologic Instruments). The transepithelial voltage was clamped to mv, and the short circuit current (I sc ) was continuously recorded. To measure basal bioelectric properties and rates of Na absorption and Cl secretion, tissues were mounted and I sc and potential difference were recorded after I sc stabilized (mean±sd, 14.8±1. minutes). Transepithelial resistance (R t ) was calculated using Ohm s law. Low-resistance tissues (R 3 cm 2 ) were excluded from the study. In one protocol, an epithelial Na channel inhibitor, amiloride hydrochloride (1µM) (Sigma-Aldrich Corp), was added to the apical bath to measure basal Na transport rates. In a second protocol, an inhibitor of the Na -K -2Cl channel, bumetanide (1µM) (Sigma-Aldrich Corp), was added to the basolateral bath to measure the Na -K -2Cl cotransport system mediated current, that is, basal Cl transport. In this protocol, after 2 minutes, amiloride was added to measure Na transport, and a purinergic receptor agonist, uridine triphosphate (UTP), 1µM (Sigma-Aldrich Corp), which ultimately raises intracellular Ca 2 levels, was added to the apical bath to assess Ca 2 -activated Cl secretion. Two additional protocols were designed to also explore regulation of Cl transport. First, the cyclic adenosine monophosphate stimulating agent forskolin (1µM) (Sigma-Aldrich Corp) was added bilaterally to activate cystic fibrosis transmembrane conductance regulator, followed by 1µM UTP and, finally, 1µM bumetanide. Second, after pretreatment with 1µM amiloride, 1µM UTP was added to the apical bath, and then 1µM forskolin was added bilaterally. Separate culture preparations for each patient were included in each protocol. STATISTICAL ANALYSIS The data were subjected to initial testing for normal distribution and equal variance using Kolmogorov-Smirnov and Bartletts tests, respectively (SigmaStat; Systat Software Inc, Richmond, Calif). Several of the data sets did not meet assumptions for parametric testing, and nonparametric testing was consistently applied when statistical comparisons were sought. Medians and 25th and 75th percentiles are presented in the tables, and the actual values are given in the figures. The Wilcoxon test for paired data, the Kruskal-Wallis test, and Dunn s multiple comparison test or the Mann-Whitney rank sum test were applied as indicated. Because nonparametric testing has less power to detect differences when they actually exist, we illustrate when a 1-tailed P=.6 is present to indicate a trend, but a 1-tailed P.5 was considered statistically significant. RESULTS CHARACTERIZATION OF PRIMARY AND PASSAGE 1 CELLS Endosinus tissues obtained from patients without chronic sinusitis exhibited no obvious inflammatory changes, whereas tissues obtained from patients with chronic sinusitis appeared grossly erythematous and edematous, with variable amounts and character of associated mucus. A small portion of every normal sinus tissue specimen and a representative sample of chronic sinusitis tissues (6 specimens) were examined histologically (Figure 2). All normal sinus epithelia consisted of ciliated, basal, and scattered goblet cells, and scattered inflammatory cells were observed in some specimens. Edema, severe inflammatory cell infiltration, fibrosis, goblet cell hyperplasia, and squamous cell metaplasia were not observed in normal tissues, consistent with normal paranasal sinus mucosa. In contrast, chronic sinusitis specimens displayed variable abnormalities, including general epithelial hyperplasia, goblet cell hyperplasia, a thickened basement membrane, intraepithelial and submucosal mixed cell infiltrates, and subjacent fibrotic and granulation tissue with reactive glands. Mean±SD NPSE and CSE primary cell yields after protease exposure were 3.2±.6 (n=15) and 6.7± (n=9) cells, respectively. Aliquots of the NPSE cells were used to make cytospin slides, and NPSE and CSE cells were plated on plastic dishes in bronchial epithelial growth medium. In primary culture, a variable proportion of the cells attached and increased in number, gradually reaching confluence in 5 to 14 days. Mean±SD NPSE and CSE passage 1 cell yields after primary culture were 2.3±.5 (n=15) and 1.7± (n=9) cells, respectively. Cy- 1354

4 tospin slides of freshly dissociated primary cells and passage 1 NPSE cells were immunostained to confirm the nature of seeded cells. Photomicrographs are shown in Figure 3, and mean data are given in Figure 4. Most cells in both populations were positive for cytokeratin, and passage 1 cells were 96.4% cytokeratin positive. Approximately half of all cells were positive for epithelial membrane antigen. The vimentin-positive fraction was lower in passage 1 cells. Eighteen percent of the primary NPSE cells were positive for the leukocyte marker CD45, whereas this value was reduced to less than 1% in passage 1 cells. These results are consistent with inflammatory and, potentially, fibroblastic cell contamination of the primary cell harvests. However, the serumfree culture conditions do not support the survival of nonepithelial cells, which were eliminated by the time of passage. Thus, the passage 1 cells that were seeded on porous supports for ALI culture were almost exclusively epithelial cells. A B MORPHOLOGIC FEATURES OF ALI CULTURES FROM PASSAGE 1 NPSE AND CSE CELLS When grown on plastic, primary sinus epithelial cells appeared flat and cobblestone (Figure 5A). These were dissociated into passage 1 cells and were seeded on Transwell-Col or Snapwell culture inserts and grown submerged in ALI culture medium until confluence, 2 to 7 days after seeding. At this point, the apical surface was rinsed with phosphate-buffered saline, and media was replaced only in the basolateral compartment. The cells opposed a hydrostatic pressure gradient, maintaining a dry apical surface, and cellular differentiation continued under ALI conditions. Representative wells of NPSE cells were fixed on day 1 and examined histologically. The epithelial cells were flat or cuboidal, and no cilia were observed at the apical surface on day 1, but a few cells contained AB-PAS positive mucous granules (Figure 5B and C). On day 21, the epithelium from NPSE (Figure 5D and E) and CSE (Figure 5F and G) cells was well differentiated, stratified with mature ciliated and AB-PAS positive goblet cells visible across the membrane. The quantitative morphologic properties of the cells are summarized in Table 1. There was a trend for the number of cell layers, epithelial thickness, and the number of goblet cells to increase in NPSE cells between days 1 and 21, but they were not statistically significantly increased. However, the number of ciliated cells increased significantly during this interval. There were no statistically significant morphologic differences between NPSE and CSE cells on day 21. Day 21 ALI NPSE cell cultures were fixed with perfluorocarbon osmium tetroxide to preserve airway surface liquid and were examined by means of electron microscopy (Figure 6). Mature cilia and microvilli were present in an electron-lucent layer close to the cell surface. A thin, more electron-dense layer, likely mucous, was observed above the clear periciliary layer (Figure 6A). In addition, images were captured presumably representing mucous granule filled membrane extensions from mucous secretory cells into the airway surface liquid (Figure 6B). Figure 2. Representative histologic appearance of the resected sinus tissue specimens. Hematoxylin-eosin stained normal (A) and chronic sinusitis (B) specimens are shown (original magnification 4). Note the general epithelial hyperplasia, goblet cell hyperplasia, basal lamina thickening, and predominantly mononuclear cell infiltrate in the chronic sinusitis specimen. Scale bar=2 µm. ELECTROPHYSIOLOGIC PROPERTIES OF ALI CULTURES To investigate transepithelial ion transport by NPSE and CSE cells, we mounted day 21 ALI Snapwell culture inserts in Ussing chambers and recorded I sc continuously. Under baseline conditions, sinus epithelium exhibited relatively high R t and relatively low I sc (Table 2) compared with reported properties of primary nasal surface epithelial cells seeded at high density and studied shortly after confluence The R t was similar to that of passage 2 tracheobronchial epithelial cells grown under identical conditions for 21 days, but again, I sc and potential difference were somewhat lower. 14 The I sc and potential difference were significantly higher in CSE cells than in NPSE cells, and there was no difference in basal R t (Table 2). We next tested the effects of the epithelial Na channel blocker amiloride on NPSE and CSE cells (Figure 7). Amiloride inhibited approximately 2% of the basal I sc in NPSE cells (n=6; P=.3) and induced an even larger decrease in I sc in CSE cells, an approximately 6% reduction (n=7; P=.2). We found little evidence of basal bumetanideinhibitable, Na -K -2Cl channel mediated Cl secretion in NPSE and CSE cells (Figure 8A and B). The effect of bumetanide on basal I sc was variable and inconsistent. In ALI cultures of 1 tissue donor, basal I sc was decreased by 9.36 µa/cm 2 (approximately 7%), but in cul- 1355

5 A B C D Figure 3. Immunohistochemical characterization of isolated normal paranasal sinus epithelial cells. Representative images of cytospin slides of passage 1 cells stained (brown reaction product) for cytokeratin (A), epithelial membrane antigen (B), vimentin (C), and CD45 (D) (methyl green counterstain, original magnification 1). Scale bar=1 µm. Cells, % P1 1 P1 1 P1 1 P1 Cytokeratin EMA Vimentin CD45 Figure 4. Mean percentage of dissociated primary (1 ) and passage 1 (P1) normal paranasal sinus epithelial cells positively stained for cytokeratin, epithelial membrane antigen (EMA), vimentin, and CD45 (n=4). Error bars represent SEM. tures from other patients it was increased or not changed after the addition of bumetanide. Amiloride addition, more than 2 minutes after bumetanide exposure, showed a trend toward decreased I sc in NPSE and CSE cells (Figure 8C and D). A large residual I sc was observed after bumetanide and amiloride exposure, approximately 6% and 45% of basal I sc in NPSE and CSE cells, respectively. A comparison of basal Na transport rates between NPSE and CSE cells is shown in Figure 9. The CSE cells exhibited relatively large I sc responses to amiloride, whether amiloride was added alone or after bumetanide, suggesting increased Na transport rates in CSE cells. Regarding regulation of Cl transport, forskolin, which stimulates cyclic adenosine monophosphate formation and activates cystic fibrosis transmembrane conductance regulator, induced a trend toward increased I sc in NPSE and CSE cells. Uridine triphosphate, which raises intracellular Ca 2 levels and activates Ca 2 -activated Cl channel, when added after forskolin, significantly increased I sc in NPSE cells and induced a trend toward increased I sc in CSE cells (Figure 1A). In the stable state after forskolin and UTP administration, bumetanide showed a trend toward inhibiting I sc in NPSE and CSE cells. There were no differences in response to drug addition between NPSE and CSE cells in this series of experiments. We also investigated the regulation of Cl secretion by UTP and forskolin in the presence of amiloride (Figure 1B). Amiloride in other respiratory epithelia increases the electrochemical driving force for Cl secre- 1356

6 A B C D E F G Figure 5. Morphologic features of cultured paranasal sinus epithelial cells. A, Typical appearance of confluent primary normal paranasal sinus epithelial cells on plastic as viewed by means of phase microscopy (original magnification 2). Scale bar=5 µm. B-G, Histologic features of air-liquid interface cultures of passage 1 cells (original magnification 4). B, C and D, E are representative normal paranasal sinus epithelial cultures 1 and 21 days after plating, respectively. F, G are chronic sinusitis paranasal sinus epithelial cells 21 days after plating. The porous support is visible below the cells. Hematoxylin-eosin staining was used in B, D, and F and Alcian blue periodic acid Schiff staining was used in C, E, and G (scale bar=2 µm). tion and increases the magnitude of the response to Cl secretagogues. Under these conditions, UTP increased I sc in NPSE cells and induced an even larger increase in CSE cells that was significantly greater than that in NPSE cells. Subsequent addition of forskolin also increased I sc in NPSE cells and again produced a significantly larger response in CSE cells. COMMENT To our knowledge, this is the first study reporting ALI culture of human paranasal sinus epithelial cells. The methods are modified from the standard ALI culture procedures for human nasal turbinate and polyp and tracheobronchial specimens. 8 The initial incubation with.1% protease for cell detachment was reduced from the 24 to 48 hours typically used with other tissues to much shorter intervals. This reduction is consistent with studies of other very small resected specimens. For example, murine tracheal epithelium was incubated in.1% protease for only 6 minutes, 6 and Black et al 5 did not use protease to isolate cells from nasal brush samples. Thus, the duration of protease exposure is probably directly related to specimen size. We found that 3 to 8 hours was optimal for the typical small and thin tissue fragments obtained from sinus mucosa. 1357

7 Table 1. Morphologic Properties of ALI Cultures* Group Day No. No. of Cell Layers Thickness, µm Ciliated No. of Cells/3 µm AB-PAS Normal (4-6) 44 (34-81) 1 (-2) Normal (5-7) 61 (49-67) 7 (2-16) 4 (2-6) Sinusitis (5-9) 86 (7-115) 12 (1-32) 4 (3-8) Abbreviations: AB-PAS, Alcian blue periodic acid Schiff; ALI, air-liquid interface. *Data are given as median (25th-75th percentile). P.5 compared with normal at day 1 (Kruskal-Wallis test followed by Dunn s multiple comparison test). There were no significant differences between normal and sinusitis at day 21. Because identification of the mucosal side of the specimen to selectively remove epithelial cells by scraping was difficult, we used immunohistochemical analysis of cytospin slides to determine the phenotype of proteolytically released primary and passage 1 NPSE cells. Before primary culture, approximately 2% of the freshly harvested cells were pancytokeratin negative and CD45 or vimentin positive and were presumably hematopoietic cells or fibroblasts. However, the contaminating CD45- positive cells did not persist in culture and were absent in suspensions of passage 1 cells. Vimentin-positive cells were observed in the passage 1 population, but these were not likely fibroblasts and instead were respiratory epithelial cells exhibiting pancytokeratin and vimentin double staining as previously reported. 15 Epithelial membrane antigen staining was similar (approximately 43%) in primary and passage 1 cells. Although this marker is variably expressed, we speculate that this result indicates minimal terminal squamous differentiation during culture on plastic. Thus, primary culture of sinus cells on plastic in serum-free media eliminated inflammatory cells and fibroblasts and resulted in an enriched population of passage 1 respiratory epithelial cells able to undergo mucociliary differentiation typical of the sinus epithelium in vivo. The morphologic features of ALI cultures of NPSE cells were examined on days 1 and 21. The cells progressed from flat and poorly differentiated cells on day 1 to polarized, columnar ciliated and AB-PAS positive mucous secretory cells on day 21. Previous studies 16 of nasal epithelial cells in ALI culture on day 21 showed approximately 3 ciliated cells per 1 mm and few secretory cells. We found a similar degree of ciliogenesis (Figure 5). Usui and colleagues 17 found that approximately half of the cells in ALI nasal surface cultures were AB-PAS positive. Although we found fewer AB-PAS positive cells, it is generally accepted that the paranasal sinus contains fewer goblet cells than the nasal surface, 2 and the present results may reflect this in vivo difference. Thus, ALI cultures on day 21 resembled, but did not precisely recapitulate, the morphologic features of the native epithelium. The cultured cells were generally more stratified than pseudostratified, ciliated cells were not as abundant, and cilia seemed to be shorter than in vivo. Nevertheless, the overall morphologic similarity to the paranasal sinus respiratory epithelium suggests that this system will be useful for modeling the function of this tissue. We found no histologic differences between NPSE and CSE cell cultures on day 21, suggesting that the in vivo effect of injury and inflammation had become minimized, at least at the morphologic level. We next used perfluorocarbon osmium tetroxide fixation and electron microscopy to examine the detailed morphologic features of the airway surface (Figure 6). We detected cilia in a clear, serous periciliary layer and an overlying, apparently mucus-rich layer. The ALI cultures of sinus epithelium may be useful for studying factors that affect the relative depth of the periciliary and mucus layers, which are presumed to strongly affect mucociliary clearance rates. Researchers have extensively evaluated transepithelial ion transport of cultured airway epithelium in Ussing chambers. 11,14 The present studies of cultured paranasal sinus epithelial cells revealed low basal I sc and high R t compared with previous studies 12,13 of primary nasal turbinate and polyp epithelial cells seeded at high density and studied after several days in culture. In human nasal surface epithelia, amiloride-sensitive Na absorption seems to be the dominant active ion transport system under basal physiologic conditions. 12,18 In the present cultured passage 1 NPSE cells, amiloride-sensitive I sc was surprisingly small, 2% of basal I sc (Figure 7). These results may be due to the effects of the culture system or may indicate that Na transport plays a less substantial role in basal transepithelial ion transport in normal human paranasal epithelium. The reduction in basal I sc induced by the Na -K -2Cl cotransporter specific inhibitor bumetanide was not statistically significant (Figure 8), consistent with little basal Cl secretion. However, the expression and contribution of the Na -K -2Cl cotransporter to Cl secretory function was indicated by a trend toward inhibition of I sc by bumetanide after forskolin and UTP treatment (Figure 1). These data suggest that there is a basal current that we have not characterized. Possibilities include amino acid or glucose-dependent Na transport or anion secretion reflecting HCO 3 transport or Cl transport mediated by a basolateral Cl entry path different than the Na -K -2Cl cotransporter, such as parallel Cl -HCO 3 and Na -H exchangers. One clue that the residual I sc may have reflected Cl secretion came from the observation that Cl secretion tended to be increased by forskolin and UTP without amiloride treatment. This response differs from nasal turbinate and polyp derived cells and suggests that a driving force for 1358

8 A Apical Mucus Layer Periciliary Layer B Figure 6. Ultrastructural morphologic features of cultured paranasal sinus epithelial cells and airway surface liquid (ASL). Transmission electron micrographs of air-liquid interface cultures of normal paranasal sinus epithelial cells after perfluorocarbon osmium tetroxide fixation to preserve the ASL (uranyl acetate lead citrate, original magnification 3). A, Cilia in a relatively clear periciliary layer and overlying mucus are visible. B, Protrusion of a mucous granule filled cell into the ASL. Scale bar=2.5 µm. Table 2. Basal Electrophysiologic Properties of Day 21 ALI Cultures* Group No. I sc, µa/cm 2 Resistance, cm 2 PD, mv NPSE cells ( ) 699 (47-138) 2.4 (1.8-6.) CSE cells ( ) 787 (522-18) 7.2 ( ) Abbreviations: ALI, air-liquid interface; CSE, chronic sinusitis epithelial; I sc, short circuit current; NPSE, normal paranasal sinus epithelial; PD, potential difference. *Data are given as median (25th-75th percentile). The number of individual Ussing chambers, which were derived from 7 different NPSE and 7 different CSE cell preparations. P.5 compared with NPSE cells (Mann-Whitney rank sum test). Cl secretion exists but that the apical Cl channels are not fully active. In contrast, in the nasal surface epithelium, a driving force for Cl secretion does not generally exist without pharmacologic blockade of epithelial Na channel with amiloride. The observed responses to forskolin and UTP exposure suggest the existence of cystic fibrosis transmembrane conductance regulator and Ca 2 -activated Cl channel in the apical membrane of NPSE cells. The increase in I sc stimulated by forskolin and UTP leads us to speculate that normal paranasal sinus epithelium has a large and regulatable capacity for Cl secretion. The low rate of Na absorption and the high potential for Cl secretion are similar to the physiologic features of nasal glands, 19 which function to secrete fluid. Thus, we suggest that the balance of ion transport in NSPE cells may favor fluid secretion. Amiloride-sensitive I sc was statistically significantly greater in CSE cells than in NPSE cells. In contrast, Cl secretion inhibited by bumetanide exposure was not greater in CSE cells than in NPSE cells. However, after amiloride pretreatment to block Na transport and establish driving forces for Cl secretion, the combination of UTP (Ca 2 ) and forskolin (cyclic adenosine monophosphate) stimulated Cl transport to a greater extent in CSE cells than in NPSE cells. These results suggest that basal Na absorption and the capacity to secrete Cl were up-regulated in CSE cells. This could be due to at least 2 alternative mechanisms. First, it is possible that CSE cells 1359

9 BIsc, µa/cm2 AIsc, µa/cm Basal Normal Amiloride BIsc, µa/cm Basal Sinusitis Amiloride Figure 7. Short circuit current (I sc ) before and after inhibition of Na transport with 1µM amiloride in air-liquid interface cultures of normal paranasal sinus epithelial (NPSE) (A) and chronic sinusitis epithelial (CSE) (B) cells. Each paired specimen is connected by a line. *P.5 compared with basal I sc, Wilcoxon test for paired data. AIsc, µa/cm2 16 Normal 6 Isc, µa/cm N S Amiloride N S Bumetanide N S Amiloride Figure 9. Amiloride and bumetanide-amiloride inhibition of short circuit current (I sc ) in cultured paranasal sinus epithelial cells. The change in basal I sc after inhibition of Na transport with 1µM amiloride in normal paranasal sinus epithelial (N) and chronic sinusitis epithelial (S) cells is illustrated on the left and after sequential bumetanide and amiloride treatment is shown on the right. *A significant difference between NPSE and CSE cells, P.5, Mann-Whitney rank sum test Basal I sc, µa/cm Bumetanide Sinusitis 1 Bumetanide Amiloride Isc, µa/cm Isc, µa/cm I sc, µa/cm N S Forskolin N UTP S ( ) Amiloride N S 3 N S N Bumetanide UTP Forskolin () Amiloride S Basal Bumetanide in culture are generally more fit than NPSE cells. Specifically, chronic inflammation may result in more replication-competent progenitor cells in the CSE cell population that are poised to regenerate a more functionally active epithelium in vitro. However, the morphologic features of NPSE and CSE cells in ALI cultures were quantitatively similar (Table 1), and a functional difference Bumetanide Amiloride Figure 8. Effects of sequential bumetanide and amiloride administration on basal short circuit current (I sc ) of cultured paranasal sinus epithelial cells. The I sc is shown before and after exposure to an Na -K -2Cl cotransport inhibitor (1µM bumetanide) in air-liquid interface cultures of normal paranasal sinus epithelial (NPSE) (A) and chronic sinusitis epithelial (CSE) (B) cells. Also shown is the effect of inhibiting Na transport with 1µM amiloride in NPSE (C) and CSE (D) cells 2 minutes after bumetanide pretreatment. Each paired specimen is connected by a line. P=.6 compared with basal I sc, Wilcoxon test for paired data. Figure 1. Changes in short circuit current (I sc ) after forskolin, uridine triphosphate (UTP), and bumetanide treatment (in the absence of amiloride) (A) and after stimulation with UTP and forskolin (in the presence of amiloride) (B) in normal paranasal sinus epithelial (N) and chronic sinusitis epithelial (S) cells. A trend toward inhibition of I sc (P=.6). *P.5 compared with baseline or a significant difference between N and S cells (P.5), Mann-Whitney rank sum test. would have to be independent of histologic appearance. Alternatively, the inflammatory milieu in chronic sinusitis may have altered the ion transport phenotype of the epithelium in situ, which persisted in vitro. Inflammatory mediators such as histamine 2 and bradykinin 21 and cytokines stimulate ion transport in the human airway epithelium. Although the effects of chronic inflammation found in cystic fibrosis airways in vivo persist for at least several days in primary cultures, they seem to be lost on extended culture or in passaged cells. 26,27 Further studies are necessary to resolve whether the observed differences in the electrophysiologic properties of 136

10 K K H 2 O Na CI H 2 O Na CI ENaC CFTR CaCC ENaC CFTR CaCC K K Na Na K 2CI Na Na K 2CI Normal Chronic Sinusitus Figure 11. A depiction of ion transport in endosinus epithelial cells and observed differences between normal and chronic sinusitis epithelial cells. The studies suggest the presence of amiloride-sensitive, epithelial Na channel (ENaC) mediated Na transport; forskolin-stimulated, cystic fibrosis transmembrane conductance regulator (CFTR) mediated Cl transport; and uridine triphosphate (UTP) stimulated, Ca 2 -activated Cl channel (CaCC) mediated Cl transport in the apical plasma membrane of endosinus epithelial cells. A role for the bumetanide-sensitive, Na -K -2Cl mediated Cl transport in the basolateral membrane was suggested by a trend toward reduced short circuit current after sequential treatment with forskolin, UTP, and bumetanide. Inwardly rectifying K channels and Na K exchanger activities are assumed to be present in the basolateral membrane. Basal short circuit current, amiloride-sensitive Na transport, and forskolin-stimulated CFTR activity (after UTP and amiloride exposure) were greater in chronic sinusitis cells. The overall greater levels of ion transport are predicted to enhance water and ion transport across the epithelium, which may contribute to the development of chronic sinusitis by depletion of airway surface liquid that, in turn, is predicted to inhibit mucus clearance. Enhanced fluid removal may also promote the development of submucosal edema, which may contribute to ostial blockage. NPSE vs CSE cells were intrinsic, acquired, or somehow related to the in vitro culture system, for example, due to differences in growth and differentiation of the excised NPSE vs CSE cells. These uncertainties highlight the limitations of the in vitro approach and point to the need to ultimately verify the findings in humans in vivo. Abnormal ion transport in CSE cells, as illustrated in Figure 11, could contribute to the pathogenesis of chronic sinusitis. With respect to the increased capacity for Cl secretion, Passali and colleagues 28,29 reported that furosemide (a bumetanide-like drug) could prevent postsurgical recurrence of sinonasal polyposis. Although bumetanide did not affect basal I sc, it tended to inhibit I sc after treatment with forskolin and UTP. If inhibition of Cl secretion under activated conditions is important to prevent the recurrence of sinusitis, this study may partially explain the outcome of the clinical studies, 28,29 although furosemide may also be active against inflammatory cells. Ostial obstruction 2 and subsequent bacterial infection are likely important in the pathogenesis of chronic sinusitis. Na absorption may become up-regulated in the paranasal sinus epithelium owing to the effects of chronic inflammation or after ostial block as a compensatory mechanism to clear residual fluid. Inflammationinduced Na hyperabsorption may deplete sinus surface liquid, impairing mucus transport and resulting in ostial occlusion by thickened mucus and inflammatory debris. After ostial blockage, it is possible that submucosal edema induced by up-regulated Na absorption, in an effort to remove accumulated materials, will further exacerbate ostial block. Thus, a vicious cycle of increased Na absorption may be important in the pathogenesis of chronic sinusitis. In conclusion, we established methods to produce welldifferentiated ALI cultures of normal and chronic sinusitis epithelia from small surgical specimens. The use of specific inhibitors and agonists in Ussing chambers revealed epithelial Na channel, Na -K -2Cl cotransporter, cystic fibrosis transmembrane conductance regulator, and Ca 2 -activated Cl channel activity. Sodium absorption in normal human paranasal sinus cells was relatively low, and bumetanide-dependent chloride secretion was undetectable under basal conditions. Cultured cells derived from chronic sinusitis specimens exhibited larger Na absorption and a greater capacity for Cl secretion compared with normal sinus mucosa. Each abnormal ion transport property found in vitro suggests potentially important mechanisms in the pathogenesis of chronic sinusitis, and studies in humans are needed to verify whether similar alterations are present in vivo. Submitted for Publication: February 15, 26; final revision received July 7, 26; accepted August 2, 26. Correspondence: Scott H. Randell, PhD, Cystic Fibrosis/ Pulmonary Research and Treatment Center, University of North Carolina, CB 7248, Thurston-Bowles Building, Room 411, Chapel Hill, NC (randell@med.unc.edu). Author Contributions: Dr Randell had full access to all 1361

11 the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Dejima, Randell, and Boucher. Acquisition of data: Dejima, Randell, Stutts, and Senior. Analysis and interpretation of data: Dejima, Randell, Stutts, and Boucher. Drafting of the manuscript: Dejima, Randell, Senior, and Boucher. Critical revision of the manuscript for important intellectual content: Dejima, Randell, Stutts, Senior, and Boucher. Statistical analysis: Dejima and Randell. Obtained funding: Randell and Boucher. Administrative, technical, and material support: Randell and Senior. Study supervision: Randell, Stutts, and Boucher. Financial Disclosure: None reported. Funding/Support: This work was supported by grants from the National Institutes of Health (Drs Randell, Stutts, and Boucher). Previous Presentation: This study was presented in abstract form at the 41st Meeting of the Japan Rhinology Society; September 27, 22; Hiroshima City, Japan. Acknowledgment: We thank Kim Burns, BS, HTL, and Tracy Bartolotta, BS, HTL, of the Cystic Fibrosis/ Pulmonary Research and Treatment Center, University of North Carolina Histology Core for technical assistance with histology and electron microscopy; Elizabeth Bazemore, BS, for help with the Ussing chambers; and Lisa Brown, MA, for editorial and graphics assistance. REFERENCES 1. Weitzberg E, Lundberg JO. Humming greatly increases nasal nitric oxide. Am J Respir Crit Care Med. 22;166: Wagenmann M, Naclerio RM. Anatomic and physiologic considerations in sinusitis. J Allergy Clin Immunol. 1992;9: Matsui H, Grubb BR, Tarran R, et al. Evidence for periciliary liquid layer depletion, not abnormal ion composition, in the pathogenesis of cystic fibrosis airways disease. Cell. 1998;95: Kennedy DW, Zinreich SJ, Rosenbaum AE, Johns ME. Functional endoscopic sinus surgery: theory and diagnostic evaluation. Arch Otolaryngol. 1985;111: Black HR, Yankaskas JR, Johnson LG, Noah TL. Interleukin-8 production by cystic fibrosis nasal epithelial cells after tumor necrosis factor- and respiratory syncytial virus stimulation. Am J Respir Cell Mol Biol. 1998;19: Davidson DJ, Kilanowski FM, Randell SH, Sheppard DN, Dorin JR. A primary culture model of differentiated murine tracheal epithelium. Am J Physiol Lung Cell Mol Physiol. 2;279:L766-L Gray TE, Guzman K, Davis CW, Abdullah LH, Nettesheim P. Mucociliary differentiation of serially passaged normal human tracheobronchial epithelial cells. Am J Respir Cell Mol Biol. 1996;14: Fulcher ML, Gabriel S, Burns KA, Yankaskas JR, Randell SH. Well-differentiated human airway epithelial cell cultures. Methods Mol Med. 25;17: Lechner JF, LaVeck MA. A serum-free method for culturing normal human bronchial epithelial cells at clonal density. J Tissue Cult Methods. 1985;9: Stutts MJ, Lazarowski ER, Paradiso AM, Boucher RC. Activation of CFTR Cl conductance in polarized T 84 cells by luminal extracellular ATP. Am J Physiol. 1995; 268:C425-C Yankaskas JR, Cotton CU, Knowles MR, Gatzy JT, Boucher RC. Culture of human nasal epithelial cells on collagen matrix supports: a comparison of bioelectric properties of normal and cystic fibrosis epithelia. Am Rev Respir Dis. 1985; 132: Willumsen NJ, Boucher RC. Sodium transport and intracellular sodium activity in cultured human nasal epithelium. Am J Physiol. 1991;261:C319-C Bernstein JM, Yankaskas JR. Increased ion transport in cultured nasal polyp epithelial cells. Arch Otolaryngol Head Neck Surg. 1994;12: Becker MN, Sauer MS, Muhlebach MS, et al. Cytokine secretion by cystic fibrosis airway epithelial cells. Am J Respir Crit Care Med. 24;169: Kasper M, Stosiek P. The expression of vimentin in epithelial cells from human nasal mucosa. Eur Arch Otorhinolaryngol. 199;248: Rhee CS, Min YG, Lee CH, et al. Ciliary beat frequency in cultured human nasal epithelial cells. Ann Otol Rhinol Laryngol. 21;11: Usui S, Shimizu T, Kishioka C, Fujita K, Sakakura Y. Secretory cell differentiation and mucus secretion in cultures of human nasal epithelial cells: use of a monoclonal antibody to study human nasal mucin. Ann Otol Rhinol Laryngol. 2; 19: Boucher RC. Human airway ion transport (part 1). Am J Respir Crit Care Med. 1994;15: Rochelle LG, Li DC, Ye H, Talbot CR, Boucher RC. Distribution of ion transport mrnas throughout murine nose and lung. Am J Physiol Lung Cell Mol Physiol. 2;279:L14-L Clarke LL, Paradiso AM, Boucher RC. Histamine-induced Cl - secretion in human nasal epithelium: responses of apical and basolateral membranes. Am J Physiol Lung Cell Mol Physiol. 1992;263: Clarke LL, Paradiso AM, Mason SJ, Boucher RC. Effects of bradykinin on Na and Cl transport in human nasal epithelium. Am J Physiol. 1992;262:C644- C Danahay H, Atherton H, Jones G, Bridges RJ, Poll CT. Interleukin-13 induces a hypersecretory ion transport phenotype in human bronchial epithelial cells. Am J Physiol Lung Cell Mol Physiol. 22;282:L226-L Hauber HP, Manoukian JJ, Nguyen LH, et al. Increased expression of interleukin-9, interleukin-9 receptor, and the calcium-activated chloride channel hclca1 in the upper airways of patients with cystic fibrosis. Laryngoscope. 23;113: Chu S, Blaisdell CJ, Bamford P, Ferro TJ. Interferon-gamma regulates ClC-2 chloride channel in lung epithelial cells. Biochem Biophys Res Commun. 24; 324: Gray T, Coakley R, Hirsh A, et al. Regulation of MUC5AC mucin secretion and airway surface liquid metabolism by IL-1 in human bronchial epithelia. Am J Physiol Lung Cell Mol Physiol. 24;286:L32-L Ribeiro CMP, Paradiso AM, Schwab U, et al. Chronic airway infection/ inflammation induces a Ca 2 i-dependent hyperinflammatory response in human cystic fibrosis airway epithelia. J Biol Chem. 25;28: Ribeiro CMP, Paradiso AM, Carew MA, Shears SB, Boucher RC. Cystic fibrosis airway epithelial Ca 2 i signaling: the mechanism for the larger agonist-mediated Ca 2 i signals in human cystic fibrosis airway epithelia. J Biol Chem. 25;28: Passali D, Mezzedimi C, Passali GC, Bellussi L. Efficacy of inhalation form of furosemide to prevent postsurgical relapses of rhinosinusal polyposis. ORL J Otorhinolaryngol Relat Spec. 2;62: Passali D, Bernstein JM, Passali FM, Damiani V, Passali GC, Bellussi L. Treatment of recurrent chronic hyperplastic sinusitis with nasal polyposis. Arch Otolaryngol Head Neck Surg. 23;129: