Proinflammatory cytokines tumor necrosis factor-α and interferon-γ alter tight junction structure and function in the rat parotid gland Par-C10 cell line

Olga J. Baker, Jean M. Camden, Robert S. Redman, Jonathan E. Jones, Cheikh I. Seye, Laurie Erb, Gary A. Weisman

Abstract

Sjögren's syndrome (SS) is an autoimmune disorder characterized by inflammation and dysfunction of salivary glands, resulting in impaired secretory function. The production of the proinflammatory cytokines tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) is elevated in exocrine glands of patients with SS, although little is known about the effects of these cytokines on salivary epithelial cell functions necessary for saliva secretion, including tight junction (TJ) integrity and the establishment of transepithelial ion gradients. The present study demonstrates that chronic exposure of polarized rat parotid gland (Par-C10) epithelial cell monolayers to TNF-α and IFN-γ decreases transepithelial resistance (TER) and anion secretion, as measured by changes in short-circuit current (Isc) induced by carbachol, a muscarinic cholinergic receptor agonist, or UTP, a P2Y2 nucleotide receptor agonist. In contrast, TNF-α and IFN-γ had no effect on agonist-induced increases in the intracellular calcium concentration [Ca2+]i in Par-C10 cells. Furthermore, treatment of Par-C10 cell monolayers with TNF-α and IFN-γ increased paracellular permeability to normally impermeant proteins, altered cell and TJ morphology, and downregulated the expression of the TJ protein, claudin-1, but not other TJ proteins expressed in Par-C10 cells. The decreases in TER, agonist-induced transepithelial anion secretion, and claudin-1 expression caused by TNF-α, but not IFN-γ, were reversible by incubation of Par-C10 cell monolayers with cytokine-free medium for 24 h, indicating that IFN-γ causes irreversible inhibition of cellular activities associated with fluid secretion in salivary glands. Our results suggest that cytokine production is an important contributor to secretory dysfunction in SS by disrupting TJ integrity of salivary epithelium.

  • salivary epithelium
  • Sjögren's Syndrome
  • claudin-1
  • ion secretion

sjögren's syndrome (SS) is an autoimmune disorder characterized by chronic inflammation and dysfunction of salivary and lacrimal glands, with progressively decreasing secretion of saliva and tears leading to xerostomia and xeropthalmia (2). The diminished function of exocrine glands in SS is related to the extent of lymphocytic infiltration, destruction of acini (2, 15, 17), and the local production of proinflammatory cytokines, such as tumor necrosis factor-α (TNF-α), interferon-γ (IFN-γ), interleukin (IL)-1β, IL-6, and IL-10 (7, 16, 28, 49, 51). Despite extensive molecular, histological, and clinical studies, the underlying cause and pathogenesis of SS remain largely unknown (24).

Previous studies showed that proinflammatory cytokines such as TNF-α and IFN-γ decrease transepithelial electrical resistance (TER) and increase paracellular permeability to ions and normally impermeant molecules in intestinal (1, 8, 14, 18, 35, 38, 59, 77) and airway epithelia (12). Decreases in the integrity of intercellular tight junctions (TJs) of intestinal epithelium have been associated with Crohn's disease (31, 67) and celiac sprue (65), and cystic fibrosis in airway epithelium (12). Electrophysiological studies of polarized epithelial cell lines show that TNF-α and IFN-γ decrease short-circuit current (Isc) in response to muscarinic receptor activation by carbachol or increase intracellular cAMP levels induced by forskolin (14, 80).

Epithelial cell TJs consist of a narrow belt-like structure in the apical region of the lateral plasma membrane that circumferentially binds each cell to its neighbor (48). TJs not only separate the apical from the lateral plasma membranes (“fence function”), but are also the rate-limiting barriers that restrict flow through the paracellular pathway (“gate function”) (26, 27). Occludin, claudins, and junctional adhesion molecules are transmembrane TJ proteins that participate in sealing the paracellular space (20, 21, 41). Rodent salivary glands express claudin-1, -3, -4, -5 (52), and -7 (32) similar to human salivary glands (36). TNF-α and IFN-γ induce internalization of the TJ proteins occludin, claudin-1, claudin-4, and junctional adhesion molecule-1 (JAM-1) in intestinal T84 epithelial cells (8). Additionally, TNF-α and IFN-γ decrease occludin gene expression in human colonic HT-29/B6 cells (39, 40) and reduce the expression of zonula occludens-1 (ZO-1) and JAM-1 in primary human airway cells (12). Furthermore, TNF-α decreases claudin-1 expression and causes ZO-1 redistribution in Madin-Darby canine kidney epithelial cells (55). These studies suggest that changes in TJ protein expression levels and protein redistribution induced by cytokines in epithelia lead to decreases in the TER and increases in the paracellular permeability of ions and normally impermeant molecules.

In currently accepted models, the transepithelial movement of Cl is the primary driving force for fluid and electrolyte secretion by salivary acinar cells (43). Agonist-stimulated secretion is initiated by concomitant activation of Ca2+-dependent Cl channels on the apical surface and K+ channels on the basolateral surface (53). The stimulated efflux of K+ and Cl down their electrochemical gradients produces a transepithelial potential difference that is followed by Na+ and water diffusion across the epithelial TJ, creating an isotonic primary secretion in the rat parotid gland (43). Polarized rat parotid gland (Par-C10) epithelial cell monolayers exhibit many features of the acinar cells from which the cell line was derived, including the presence of TJs that maintain cell polarization and TER, and Isc induced by carbachol or UTP due to apical transepithelial anion secretion (73). In the present study, we used Par-C10 cell monolayers to investigate the effects of proinflammatory cytokines on epithelial TJ integrity that is necessary for establishing transepithelial ion gradients that drive saliva secretion. Our findings indicate that TNF-α and IFN-γ alter TJ structure and function in Par-C10 cell monolayers leading to salivary epithelial dysfunction.

MATERIALS AND METHODS

Par-C10 cell culture.

The polarized rat parotid cell line (Par-C10) was derived from freshly isolated rat parotid gland acinar cells by transformation with simian virus 40 and exhibits morphological, biochemical, and functional characteristics of freshly isolated acinar cells (57). Par-C10 cells (5 × 105; passages 4060) were plated on Falcon permeable supports (diameter 1.2 cm, pore size 0.4 μm; Becton Dickinson, Franklin Lakes, NJ). The cultures were grown to confluence in DMEM-Ham's F12 (1:1) containing 2.5% (vol/vol) fetal bovine serum (GIBCO BRL, Gaithersburg, MD) and the following supplements: 0.1 μM retinoic acid, 80 ng/ml epidermal growth factor, 2 nM triiodothyronine, 5 mM glutamine, 0.4 μg/ml hydrocortisone, 5 μg/ml insulin, 5 μg/ml transferrin, 5 ng/ml sodium selenite, 50 μg/ml gentamicin, and 8.4 ng/ml cholera toxin (all from Sigma, St. Louis, MO). Cells were cultured at 37°C in a humidified atmosphere of 95% air-5% CO2 and used at confluence (a time when monolayer cultures exhibited maximum TER), typically 4 days after plating.

Cytokine treatment of Par-C10 cell monolayers.

TNF-α (0.01–100 ng/ml) and/or IFN-γ (0.01–100 ng/ml) were added to the basolateral side of Par-C10 cell monolayers grown on permeable supports and incubated for 18–48 h. For recovery studies, cytokine-treated monolayers were cultured for 24 h in cytokine-free medium.

Measurement of TER in Par-C10 cell monolayers.

Changes in TER in response to TNF-α and IFN-γ in polarized Par-C10 cell monolayers grown on permeable supports were measured as a function of time using an epithelial volt-ohmmeter (EVOM; World Precision Instruments, New Haven, CT) with miniature dual chopstick electrodes. After subtraction of bare filter resistance (120 Ω), tissue resistance values in Ω were multiplied by effective membrane area (π) (d2)/4 = (3.14) (1.20 cm)2/4 = 1.13 cm2. Therefore, TER is expressed as Ω·cm2.

Measurement of Isc in Par-C10 cell monolayers.

Par-C10 cell monolayers treated with or without TNF-α and/or IFN-γ (10 ng/ml each) for 48 h were mounted in modified Ussing chambers equipped with a recirculating water jacket for measurement of changes in Isc induced by carbachol (100 μM) or UTP (100 μM), as described previously (73). Standard medium (5 ml; Krebs-Ringer-HCO3 buffer, pH 7.5, containing 118 mM NaCl, 3 mM KCl, 1.2 mM MgSO4, 25 mM NaHCO3, 1.0 mM CaCl2, and 10 mM glucose) was added to the apical and basolateral reservoirs at 37°C. The medium in both reservoirs was mixed and oxygenated by bubbling with 95% O2-5% CO2. Isc was measured continuously and transepithelial potential difference was monitored intermittently using a VCC-600 automatic voltage-clamp apparatus (Physiologic Instruments, San Diego, CA) and an electrode set (EasyMount; Physiologic Instruments) connected to the chamber system with 3% (wt/vol) agar-KCl bridges. Isc and automatic fluid resistance compensation current were applied through Ag-AgCl electrodes connected to the chamber system with 4% (wt/vol) agar-KCl bridges. Changes in Isc (μA/cm2 membrane) are expressed as the peak change obtained in response to agonist minus the basal value.

Intracellular free Ca2+ concentration measurements.

The intracellular free Ca2+ concentration ([Ca2+]i) was quantified in single cells within polarized monolayers of TNF-α- and/or IFN-γ-treated or -untreated Par-C10 cells grown on permeable supports and preloaded with Fura-2, a Ca2+-sensitive fluorescent dye, using an InCyt Dual-Wavelength Fluorescence Imaging System (Intracellular Imaging, Cincinnati, OH). For Fura-2 preloading, Par-C10 cell monolayers on permeable supports were incubated in assay buffer [120 mM NaCl, 4 mM KCl, 1.2 mM KH2PO4, 1.2 mM MgSO4, 1 mM CaCl2, 10 mM glucose, 15 mM HEPES, pH 7.4, and 0.1% (wt/vol) BSA] containing 2.0 μM Fura-2-AM (Molecular Probes, Eugene, OR) for 45 min at 37°C, washed, and further incubated for 20 min at 37°C. The cell monolayers on permeable supports were positioned on the stage of a fluorescence microscope and stimulated with agonists at 37°C, as described in the figure legends. The cell monolayers were exposed sequentially to 340/380 nm light and fluorescence emission was detected at 505 nm and converted to [Ca2+]i using a standard curve created with solutions containing known concentrations of Ca2+. Increases in [Ca2+]i are expressed as the peak response obtained under the indicated conditions (see figure legends) minus the basal [Ca2+]i.

Measurement of paracellular permeability of normally impermeant proteins in Par-C10 cell monolayers.

Microperoxidase MP-11 (1.9 kDa, 10 μM; Sigma), horseradish peroxidase type VI (44 kDa, 10 μM; Sigma), or bovine milk lactoperoxidase (80 kDa, 10 μM; Sigma) was added to the basolateral compartment of untreated or TNF-α- and/or IFN-γ-treated polarized Par-C10 cell monolayers grown on permeable supports. After 2 h, supernatants were collected from the apical and basolateral compartments and peroxidase activity was measured using a liquid substrate system for ELISA [2,2′-Azino-bis (3-ethylbenzothiazoline)-6-sulfonic acid; Sigma], according to the manufacturer's instructions. Paracellular peroxidase flux was determined in three wells per experiment, and experiments were repeated at least three times.

Electron microscopy.

Par-C10 cell monolayers were fixed in 2% (vol/vol) glutaraldehyde and 2% (vol/vol) formaldehyde (prepared from paraformaldehyde) in 0.1 M cacodylate buffer, pH 7.2, overnight at 4°C. Fixed cells were rinsed four times for 10 min each with 30 mM HEPES, pH 7.2, 70 mM NaCl, and 6% (wt/vol) sucrose, rinsed three times with 20 mM Tris, pH 7.2, containing 120 mM NaCl and 5 mM CaCl2, postfixed with osmium tetroxide 1% (wt/vol) OsO4, 70 mM NaCl, 5 mM CaCl2, 30 mM HEPES buffer, pH 7.4, for 10 min, and rinsed three times for 10 min each with distilled water. Fixed cells were stained overnight in aqueous 0.5% (wt/vol) uranyl acetate (pH 6.0) at room temperature and infiltrated with Epon-Araldite epoxy resin (Electron Microscopy Sciences, Hatfield, PA). The infiltrate was placed in fresh resin in ballistic-electron-emission microscopy (BEEM) embedding capsules and polymerized at 60°C. Sections of 70-nm thickness were cut on a Leica ultra cut microtome (UCT) and stained with 5% (wt/vol) uranyl acetate and Sato's triple lead salt stain consisting of 1% (wt/vol) lead citrate, 1% (wt/vol) lead acetate, 1% (wt/vol) lead nitrate, and 2% (wt/vol) sodium citrate (61). Samples were viewed and photographed in a Japan Electro Optics Laboratories (JEOL 1400) transmission electron microscope (TEM).

Freeze-fracture analysis.

Par-C10 cell monolayers were fixed at 4°C overnight in 2% (vol/vol) glutaraldehyde and 2% (vol/vol) formaldehyde in 0.1 M phosphate buffer, pH 7.2. Fixed cells were cryoprotected by incubation for 1 h with a gradient of increasing concentrations of sucrose (0.25 to 1.8 M) in 0.1 M cacodylate buffer, pH 7.2. Cells on permeable supports in 1.8 M glucose were shipped overnight to the University of Iowa where cells were scraped from supports, loaded on holders, and plunged into liquid ethane. Frozen specimens were transferred to a Balzers 301 freeze-fracture apparatus for fracturing at −110°C for 1 min causing water sublimation. Etched fracture surfaces were replicated with platinum at 45° supported by carbon deposited from 90°. Replicas were cleaned for 1 h using commercial household bleach followed by rinsing with double distilled water. Formvar-coated grids were used to pick up the cleaned replicas. For freeze- fracture quantification, eight replicas per monolayer were obtained, replicas were examined in a JEOL 1400 TEM equipped with a 4k × 4k Gatan model 890 on-axis charge couple device (CCD) camera, and images were generated at the University of Missouri Electron Microscopy Core. The TJs of Par-C10 monolayers were located, photographed, and printed at a final magnification of ×60,000 (3).

Morphometry.

Strands of TJs run parallel to the apical membrane at the basal side of the microvilli of Par-C10 cells. However, the freeze-fracture technique could not always manifest TJs as full-length or integral strands, and the TJs in each cell had indefinite intervals. Therefore, as previously described for quantitative analysis of TJ morphometry (3, 25), main grid lines (see Fig. 6F) perpendicular to the parallel integral strands were drawn at intervals of 3 cm (equivalent to a real interval of 1 μm) on electron micrographs at a magnification of ×60,000. A total of 130 compartments (of each grid area) were analyzed for maximum and minimum depth of TJ strands (see Fig. 6F; red arrows) using a ruler. The number of strands was analyzed using three additional grid lines (see Fig. 6F green lines) drawn parallel to the main grids at intervals of 1 cm and the total number of strands crossing these additional grid lines was counted. To calculate the number of strands for each compartment, the total number of strands was divided by three and the number of strands was expressed as a mean value.

Western blot analysis.

Par-C10 cell monolayers were lysed in 200 μl of 2× Laemmli buffer and lysates were sonicated for 5 s with a Branson Sonifier 250 (microtip; output level 5; duty cycle 50%) and boiled for 3 min. The lysates were subjected to 7.5% (wt/vol) SDS-PAGE on mini-gels and transferred to nitrocellulose membranes. Membranes were blocked for 1 h with 5% (wt/vol) nonfat dry milk in Tris-buffered saline [0.137 M NaCl, 0.025 M Tris (hydroxymethyl)-aminomethane, pH 7.4] containing 0.1% (vol/vol) Tween-20 (TBST) and immunoblotted overnight with primary antibodies at 4°C in TBST containing 3% (wt/vol) BSA and 0.02% (wt/vol) sodium azide. TJ proteins were detected with the following antibodies: rabbit anti-claudin-1 antibody (1:1,000 dilution; Invitrogen, Carlsbad, CA) that recognizes an epitope of a synthetic peptide corresponding to the COOH terminus of human claudin-1, rabbit anti-claudin-3 antibody (1:1,000 dilution; Invitrogen) that recognizes a synthetic peptide derived from the COOH-terminal region of mouse claudin-3, mouse anti-claudin-4 antibody (1:1,000 dilution; Invitrogen) that recognizes a synthetic peptide derived from the COOH-terminal region of human claudin-4, rabbit anti-claudin-10 antibody (1:1,000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA) that recognizes amino acids 171–228 mapping at the COOH terminus of human claudin-10, mouse anti-ZO-1 antibody (1:1,000 dilution; Invitrogen) that recognizes amino acids 334–634 of human ZO-1, mouse anti-human occludin (1:1,000 dilution; Invitrogen) that recognizes the COOH-terminal region of human occludin, and goat anti-mouse JAM-1 antibody (1:1,000 dilution; R&D Systems, Minneapolis, MN) that recognizes the extracellular domain of rat and mouse JAM-1. After incubation with the primary antibody, membranes were washed three times for 15 min each with TBST and incubated with peroxidase-linked goat anti-rabbit IgG antibody (1:2,000 dilution; Santa Cruz Biotechnology), goat anti-mouse IgG antibody (1:2,000 dilution; Santa Cruz Biotechnology), or donkey anti-goat IgG antibody (1:2,000 dilution; Santa Cruz Biotechnology), as appropriate, at room temperature for 1 h. The membranes were washed three times for 15 min each with TBST and then treated with chemiluminescence detection reagent containing 20 mM Tris buffer, pH 8.5, 250 mM luminol, and 90 mM coumaric acid (Sigma) and protein bands were visualized on X-ray film and their densities were quantified using a computer-driven scanner and Quantity One software (Bio-Rad, Hercules, CA). For signal normalization, membranes were treated with stripping buffer [0.1 M glycine, pH 2.9, and 0.02% (wt/vol) sodium azide] and reprobed with goat anti-rabbit (total) extracellular signal-regulated kinase (ERK) antibody (1:1,000 dilution; Santa Cruz Biotechnology). Claudin-1 is expressed as a ratio of normalized values of the band intensities of claudin-1 to total ERK. All experiments were performed in duplicate and repeated at least three times.

Statistical analysis.

Data are means ± SE of results from three or more determinations. P values <0.05 calculated from a two-tailed t-test represent significant differences.

RESULTS

TNF-α and/or IFN-γ cause a time- and dose-dependent decrease in TER in Par-C10 cell monolayers.

To determine whether treatment with specific cytokines associated with SS causes a time-dependent decrease in TER in salivary epithelium, Par-C10 cell monolayers were treated with TNF-α and/or IFN-γ and TER was measured over 48 h. A significant decrease in TER was detected in TNF-α- and/or IFN-γ-treated Par-C10 cell monolayers within 24 h, compared with cells treated in the absence of cytokines (Fig. 1A). This decrease in TER was significant at >1 ng/ml TNF-α and/or IFN-γ (Fig. 1B). These results suggest that prolonged exposure of salivary gland epithelium to cytokines can affect TJ permeability properties, similar to intestinal (1, 8, 14, 18, 35, 38, 59, 77) and airway epithelia (12).

Fig. 1.

Tumor necrosis factor-α (TNF-α) and/or interferon-γ (IFN-γ) alone or in combination cause a time- (A) and dose-dependent (B) decrease in transepithelial resistance (TER) in Par-C10 cell monolayers. Par-C10 cells were cultured on permeable supports, as described in materials and methods. After reaching maximal TER, cells were exposed to TNF-α (10 ng/ml) and/or IFN-γ (10 ng/ml) for 18–48 h (A) or TNF-α and/or IFN-γ (0–100 ng/ml each) for 48 h (B). TER values were obtained with an epithelial volt-ohmmeter (EVOM) equipped with miniature dual chopstick electrodes. Following subtraction of medium resistance (120 Ω), tissue resistance was multiplied by the effective membrane area (1.13 cm2) and expressed as means ± SE of results from 10 independent experiments, where *P < 0.05 indicates a significant difference from control cells.

TNF-α and/or IFN-γ decrease agonist-induced transepithelial anion secretion in Par-C10 cell monolayers.

Previous studies indicate that salivary gland-relevant receptor agonists, including the basolateral muscarinic cholinergic receptor agonist carbachol (6, 19) and the apical P2Y2 nucleotide receptor (P2Y2R) agonist UTP (56, 58), induce increases in [Ca2+]i in Par-C10 cells (57). In Par-C10 cell monolayers, changes in Isc induced by UTP or carbachol are due to an increase in transepithelial anion secretion across the apical membrane (73). To determine whether TNF-α and/or IFN-γ affect agonist-induced transepithelial anion secretion, Par-C10 cell monolayers were treated with 10 ng/ml TNF-α and/or IFN-γ for 48 h, and then cell monolayers were mounted in Ussing chambers to measure agonist-induced changes in Isc. UTP was applied to the apical side of Par-C10 cell monolayers or carbachol was applied to the basolateral side and changes in Isc were monitored immediately after the addition of the agonist. TNF-α and/or IFN-γ significantly decreased UTP- (Fig. 2A) or carbachol-induced (Fig. 2B) changes in Isc in Par-C10 cell monolayers. These results suggest that decreases in agonist-induced transepithelial anion secretion are related to the reduction in TER caused by TNF-α and/or IFN-γ treatment (Fig. 1, A and B).

Fig. 2.

TNF-α and/or IFN-γ decrease agonist-induced transepithelial anion secretion in Par-C10 cell monolayers. Par-C10 cells were cultured on permeable supports, as described in materials and methods. After reaching maximal TER, cells were exposed to TNF-α (10 ng/ml) and/or IFN-γ (10 ng/ml) for 48 h. Permeable supports were placed in Ussing chambers and changes in short-circuit current (Isc) were monitored in response to UTP (100 μM) applied to the apical compartment (A) or carbachol (100 μM) applied to the basolateral compartment (B). Values represent the maximum Isc obtained and are expressed as means ± SE of results from 3 or more experiments, where *P < 0.05 indicates significant differences from control cells.

TNF-α and/or IFN-γ do not affect agonist-induced increases in [Ca2+]i in Par-C10 cell monolayers.

Activation of G protein-coupled P2Y2 or muscarinic receptors in Par-C10 cells stimulates phospholipase C to mediate inositol 1,4,5-trisphosphate-dependent increases in [Ca2+]i due to calcium release from intracellular stores and calcium entry via activation of store-operated calcium channels (5, 6, 73). Therefore, we determined whether the decrease in transepithelial anion secretion in response to UTP or carbachol in TNF-α- and/or IFN-γ-treated Par-C10 cell monolayers was associated with alterations in the calcium signaling pathway. However, UTP (Fig. 3A) or carbachol (Fig. 3B) induced increases in [Ca2+]i that were unaffected by a 48-h treatment with TNF-α and/or IFN-γ. These results indicate that decreases in agonist-induced transepithelial anion secretion in response to TNF-α and/or IFN-γ in Par-C10 cell monolayers are not due to changes in receptor-mediated calcium signaling.

Fig. 3.

TNF-α and/or IFN-γ do not alter agonist-induced calcium signaling in Par-C10 cell monolayers. Par-C10 cells were cultured on permeable supports, as described in materials and methods. After reaching maximal TER, cells were exposed to TNF-α (10 ng/ml) and/or IFN-γ (10 ng/ml) for 48 h. UTP (100 μM) was applied to the apical compartment (A) or carbachol (100 μM) was applied to the basolateral compartment (B) and changes in [Ca2+]i were monitored in Par-C10 cell monolayers, as described in materials and methods. Changes (Δ) in [Ca2+]i were expressed by subtracting the basal [Ca2+]i (before agonist addition) from the peak agonist-induced increase in [Ca2+]i. Data are expressed as means ± SE of results from 3 experiments.

TNF-α and/or IFN-γ increase paracellular permeability to normally impermeant proteins in Par-C10 cell monolayers.

To investigate whether the decrease in TER and agonist-induced transepithelial anion secretion caused by TNF-α and/or IFN-γ treatment correlated with alterations in the barrier function of TJs in Par-C10 cell monolayers, the paracellular permeability of different size proteins was measured. As shown in Fig. 4, paracellular permeability of microperoxidase (2 kDa) and peroxidase (40 kDa) was increased by TNF-α and/or IFN-γ treatment of Par-C10 cell monolayers, as compared with untreated controls. In contrast, the paracellular permeability of lactoperoxidase (80 kDa) was not affected by TNF-α and/or IFN-γ treatment (Fig. 4). These results indicate that TNF-α and/or IFN-γ lower TER by decreasing the barrier function of TJs, thereby diminishing the magnitude of agonist-induced transepithelial anion secretion in Par-C10 cell monolayers.

Fig. 4.

TNF-α and/or IFN-γ increase the paracellular permeability of normally impermeant proteins in Par-C10 cell monolayers. Par-C10 cells were cultured on permeable supports, as described in materials and methods. After reaching maximal TER, cells were exposed to TNF-α (10 ng/ml) and/or IFN-γ (10 ng/ml) for 48 h. Then, microperoxidase, peroxidase, or lactoperoxidase (42 μg/ml) was added to the basolateral compartment. After 2 h, peroxidase activity in 100 μl of medium from the apical compartment was detected with 100 μl of the liquid peroxidase substrate [2′2-Azino-bis (3-ethylbenzothiazoline)-6-sulfonic acid] system for ELISA (Sigma). After 5 min, the reaction was stopped with 100 μl of 1% (wt/vol) SDS and the absorbance was measured at 405 nm. Data are expressed as means ± SE of results from 3 independent experiments, where *P < 0.05 indicates a significant difference from control cells.

TNF-α and IFN-γ alter Par-C10 cell morphology.

A 48-h treatment with TNF-α and/or IFN-γ (10 ng/ml each) alters Par-C10 cell morphology, as shown by TEM (Fig. 5). Par-C10 cell monolayers in the absence of TNF-α and/or IFN-γ were closely packed and columnar (Fig. 5A), as reported previously (73). The nuclei were located near the basolateral side of cells and were uniform in size and shape (Fig. 5A). Cells treated with TNF-α (Fig. 5B) or IFN-γ (Fig. 5C) looked similar to control cells, although the intercellular space appears to be increased. In contrast, cell monolayers treated with TNF-α and IFN-γ developed a flatter morphology and were squamous-like with flat nuclei that oriented in parallel to the bottom of the culture dish (Fig. 5D). Thus, a combination of TNF-α and IFN-γ dramatically alters Par-C10 cell morphology.

Fig. 5.

TNF-α and IFN-γ alter Par-C10 cell morphology. Shown are transmission electron micrographs (TEMs) of Par-C10 cells grown on permeable supports for 48 h in the absence (A) or presence of TNF-α (10 ng/ml), IFN-γ (10 ng/ml; B and C), or TNF-α (10 ng/ml) and IFN-γ (10 ng/ml; D). Cells were processed for morphological analysis, as described in materials and methods. Magnification ×4,000.

TNF-α and/or IFN-γ alter TJ morphology in Par-C10 cell monolayers.

Since TNF-α and/or IFN-γ increased the paracellular permeability of normally impermeant proteins in Par-C10 cell monolayers, we assessed whether TNF-α and/or IFN-γ could cause morphological changes in TJ structure. As summarized in Fig. 6E, electron micrographs of freeze-fractured membrane replicas from Par-C10 cell monolayers treated with TNF-α and IFN-γ (Fig. 6D) or TNF-α alone (Fig. 6B) showed decreased TJ depth, as compared with cells treated with IFN-γ alone (Fig. 6C) or untreated cells (Fig. 6A). However, TNF-α and/or IFN-γ significantly decreased the number of TJ strands, as compared with untreated controls (Fig. 6E). In addition, discontinuous TJ strands were observed in TNF-α-treated cells (Fig. 6, B and D; yellow arrows), but not in cells treated with IFN-γ alone (Fig. 6C) or in the absence of cytokines (Fig. 6A). These results demonstrate that alterations in TER (Fig. 1) and the paracellular permeability of proteins (Fig. 4) induced by TNF-α and/or IFN-γ in Par-C10 cell monolayers are associated with changes in TJ morphology.

Fig. 6.

TNF-α and IFN-γ alter tight junction (TJ) morphology in Par-C10 cell monolayers. TEM of freeze-fractured membrane replicas from untreated Par-C10 cells (A) or cells treated with TNF-α (10 ng/ml; B), IFN-γ (10 ng/ml; C), or TNF-α (10 ng/ml) and IFN-γ (10 ng/ml; D) was performed, as described in materials and methods. TJ strands are shown between black arrows (AD). E: quantification of TJ depth and strand number. F: methods for quantifying TJ morphometry, as described previously (3). Main grid lines (black) perpendicular to the parallel strands were drawn at 3-cm intervals over the TJ area. The maximum depth (red arrow a) and the minimum depth (red arrow b) were measured for each compartment separated by the grid lines. Three additional grid lines (green) were drawn parallel to the main grid lines at 1-cm intervals in each compartment. The total number of strands was determined as the number of points where strands crossed the 3 grid lines (green). #1 cm. Data are expressed as means ± SE of results from 3 experiments, where *P < 0.05 indicates significant differences from control cells.

TNF-α and/or IFN-γ downregulate claudin-1 expression in Par-C10 cell monolayers.

We investigated whether effects of cytokines on Par-C10 cell monolayers and TJ morphology correlated with changes in the expression levels of TJ proteins in Par-C10 cell monolayers. Exposure of Par-C10 cell monolayers to TNF-α and/or IFN-γ caused a decrease in the expression of claudin-1, compared with control cells (Fig. 7). The expression levels of other TJ proteins including claudins-3, -4, and -10, occludin, ZO-1, and JAM-1 were not significantly altered by TNF-α and/or IFN-γ treatment (Fig. 7B). Claudins-5 and -7 were not expressed in Par-C10 cells (data not shown). These results indicate that claudin-1 is selectively downregulated among TJ proteins by TNF-α and/or IFN-γ treatment of Par-C10 cell monolayers.

Fig. 7.

TNF-α and IFN-γ downregulate expression of claudin-1, but not other TJ proteins, in Par-C10 cell monolayers. A: lysates were prepared from Par-C10 cell monolayers treated with or without 10 ng/ml TNF-α and/or IFN-γ and claudin-1 and total ERK expression were detected by Western analysis, as described in materials and methods. Data are expressed as means ± SE of results from 3 or more experiments, where *P < 0.05 indicates significant differences from control cells. Results from a representative experiment are shown at the top of the figure. B: lysates were prepared from Par-C10 cell monolayers treated with or without 10 ng/ml TNF-α and/or IFN-γ and expression of ZO-1, occludin, JAM-1, and claudins-3, -4, and -10 was detected by Western blot analysis. Results from a representative experiment (n = 3) are shown.

Decreases in TER, agonist-induced transepithelial anion secretion, and claudin-1 expression are reversible in TNF-α- but not IFN-γ-treated Par-C10 cell monolayers.

Since TNF-α and IFN-γ are known to cause apoptosis of human salivary gland cell lines (34, 42) and salivary gland cell apoptosis has been associated with SS (23, 33, 69), we investigated whether Par-C10 cells treated with TNF-α and/or IFN-γ could recover normal function after removal of the cytokines. As shown in Figs. 8 and 9, the decreases in TER and agonist-induced transepithelial anion secretion caused by a 48-h TNF-α treatment of Par-C10 cell monolayers were reversible after a 24-h incubation in cytokine-free medium. In contrast, TER and agonist-induced transepithelial anion secretion in Par-C10 cell monolayers treated with IFN-γ with or without TNF-α did not recover to normal values after a 24- (Figs. 8 and 9) or 72-h (data not shown) incubation in cytokine-free medium. Claudin-1 expression also was restored to basal levels in TNF-α-treated Par-C10 cell monolayers after incubation in cytokine-free medium for 24 h, but cells treated with IFN-γ in the presence or absence of TNF-α maintained low levels of claudin-1 expression after cytokine removal (Fig. 10). Furthermore, TNF-α and/or IFN-γ treatment of Par-C10 cell monolayers did not increase cell loss over 48 h, as compared with untreated cells (data not shown). Taken together, these results indicate that the loss of TJ integrity associated with decreased claudin-1 expression in Par-C10 cell monolayers treated with TNF-α and/or IFN-γ is not due to cell apoptosis.

Fig. 8.

TNF-α-mediated decreases in TER in Par-C10 cell monolayers are reversible. Par-C10 cell monolayers were treated with or without 10 ng/ml TNF-α and/or IFN-γ for 48 h and cells were washed and incubated in cytokine-free medium for 24 h. Then, TER was measured, as described for Fig. 1. Data are expressed as means ± SE of results from 3 or more experiments, where *P < 0.05 indicates a significant difference from cells treated without TNF-α removal.

Fig. 9.

TNF-α-mediated decreases in agonist-induced transepithelial anion secretion in Par-C10 cell monolayers are reversible. Par-C10 cell monolayers were treated with or without 10 ng/ml TNF-α and/or IFN-γ for 48 h and cells were washed and incubated in cytokine-free medium for 24 h. Then, changes in Isc were determined in response to UTP (100 μM; A) or carbachol (100 μM; B), as described for Fig. 2. Data are expressed as means ± SE of results from 3 or more experiments, where *P < 0.05 indicates significant differences from cells treated without TNF-α removal.

Fig. 10.

TNF-α-mediated decreases in claudin-1 expression in Par-C10 cell monolayers are reversible. Par-C10 cell monolayers were treated with or without 10 ng/ml TNF-α and/or IFN-γ for 48 h and cells were washed and incubated in cytokine-free medium for 24 h. Claudin-1 expression in cell lysates was normalized to total ERK, as described for Fig. 7. Data are expressed as means ± SE of results from 3 or more experiments, where *P < 0.05 indicates a significant difference from cells treated without TNF-α removal. Results from a representative experiment are shown at the top.

DISCUSSION

Previous studies showed that levels of TNF-α and IFN-γ are elevated not only in plasma but also in the minor salivary glands of patients with SS (7, 16). Inflammatory cytokines in SS are synthesized and released by lymphocytes (16, 29) and salivary gland epithelium (68, 76). Cytokines have been shown to alter intestinal and airway epithelial TJ structure and function (1, 8, 12, 14, 18, 35, 38, 54, 59, 77). Moreover, alterations in TJ structure have been associated with other inflammatory autoimmune diseases, such as Crohn's disease (31), celiac sprue (65), and cystic fibrosis (12). The present study indicates that the proinflammatory cytokines TNF-α and IFN-γ alone or in combination alter parotid (Par-C10) cell and TJ morphology (Figs. 5 and 6) associated with decreases in TER (Fig. 1) and the expression of the TJ protein claudin-1 (Fig. 7) and increases in the paracellular permeability of normally impermeant proteins <80 kDa (Fig. 4). Accordingly, these results suggest that TNF-α and/or IFN-γ promote salivary gland dysfunction in SS by altering epithelial TJ structure and function.

Exposure of Par-C10 cell monolayers to TNF-α and/or IFN-γ also decreases muscarinic and nucleotide receptor-mediated changes in Isc (Fig. 2), indicative of a decrease in agonist-stimulated transepithelial anion secretion. These effects were most pronounced when monolayers were treated with a combination of TNF-α and IFN-γ, and are likely related to effects of these cytokines on TJs that serve to lower the transmembrane electrical potential that drives agonist-induced anion secretion. The results presented here are consistent with previous studies indicating that both TNF-α and IFN-γ alter barrier function of intestinal and airway epithelia (12, 14, 38, 54, 59), although this is the first report of effects of SS-associated cytokines on salivary gland epithelial TJ functions.

Previous studies demonstrated secretory dysfunction in salivary epithelia of humans with SS (13) and in animal models of SS (10, 30). Here, Par-C10 cell monolayers were employed as an in vitro model to demonstrate that SS-associated cytokines can affect anion secretion induced by the muscarinic cholinergic receptor agonist carbachol or the P2Y2R agonist UTP (Fig. 2). Saliva secretion is a process that depends on activation of second messenger systems that regulate increases in the [Ca2+]i and activation of calcium-dependent chloride channels (43). Since TNF-α and/or IFN-γ decrease Isc in response to carbachol or UTP (Fig. 2), and these cytokines do not alter agonist-induced calcium responses (Fig. 3), we conclude that secretory dysfunction induced by TNF-α and/or IFN-γ is not associated with alterations in calcium signaling, but rather with disruption of TJ structure and function.

Our findings are consistent with previous studies suggesting that cytokine exposure alters TJ ultrastructure in intestinal (63) and airway epithelia (12). Moreover, TJ morphology is altered in jejunum of patients with acute celiac sprue (64), in colonic epithelium from patients with ulcerative colitis (62), and in airway epithelium from patients with cystic fibrosis (12). Therefore, salivary gland dysfunction in SS is likely associated with cytokine-induced alterations in TJ structure and function, although our results with Par-C10 cell monolayers should be extended to studies of TJ morphology in salivary epithelium of patients with SS.

The claudin family of transmembrane proteins plays a critical role in maintaining TJ integrity and may regulate the selective passage of ions and molecules through the paracellular space (for review, see Refs. 71, 72). Immunoreplica electron microscopy revealed that claudins are exclusively localized on TJ strands (4446). The number of TJ strands is an important factor in determining the barrier properties of TJs, but the molecular mechanism underlying the regulation of strand number remains unknown (72). When Madin-Darby canine kidney cells, which express claudin-1 and claudin-4, were specifically depleted of claudin-4, a marked decrease was observed in the number of TJ strands and in paracellular barrier function (66). Furthermore, when claudins were overexpressed in L fibroblasts, a large network of TJ strands was formed (22). These findings indicate that the number of TJ strands might be determined by the level of claudin expression in individual cells.

Our results show that TNF-α and/or IFN-γ selectively decrease claudin-1 expression in Par-C10 cell monolayers (Fig. 7), since expression levels of other TJ proteins including claudins-3, -4, -10, and occludin, ZO-1 and JAM-1 are unaffected by TNF-α and/or IFN-γ (Fig. 7B). Further studies are needed to determine whether TJ protein distribution or activity at the junctional complex is altered by TNF-α and/or IFN-γ in Par-C10 cell monolayers, although we did not detect redistribution of ZO-1 or occludin in Par-C10 cell monolayers treated with these cytokines (data not shown). Previous studies showed that IFN-γ downregulates claudin-1 expression and impairs the barrier function of primary thyrocytes (70), and the downregulation of claudin-1 expression has been associated with chronic plaque psoriasis (75). Moreover, increased breast cancer recurrence correlates with decreased claudin-1 expression (47). Thus, it seems plausible that decreases in claudin-1 expression caused by the SS-associated cytokines TNF-α and/or IFN-γ play a role in salivary gland dysfunction in SS patients.

TNF-α and IFN-γ are known to induce apoptosis in a variety of epithelia (4, 9, 11, 50), including human salivary gland cells (34, 42). However, apoptosis does not correlate with loss of barrier function (8, 37, 60, 79), and we did not detect any decrease in cell number in Par-C10 cell monolayers treated with TNF-α and/or IFN-γ, compared with untreated cells (data not shown). Previous studies using high-resolution epithelial surface analysis demonstrated that a continuous epithelial cell layer is maintained during apoptotic cell extrusion (78). Furthermore, incubation of the intestinal epithelial cell line, T84, with TNF-α and IFN-γ induced cell flattening from their normal tall columnar phenotype that was accompanied by extrusion of apoptotic cells (8). In airway (12) and intestinal (74) epithelia, TNF-α- and/or IFN-γ-induced decreases in TER are reversible after treatment with cytokine-free medium (12). This is consistent with our results on the reversibility of TNF-α-induced decreases in TER, agonist-induced transepithelial anion secretion, and claudin-1 expression in Par-C10 cell monolayers (Figs. 810), although the mechanisms underlying the irreversible effects of IFN-γ warrant further investigation.

In summary, the present study demonstrates that TNF-α and/or IFN-γ cause disruption of barrier function in Par-C10 cell monolayers associated with changes in cell and TJ morphology and the decreased expression of the TJ protein claudin-1, alterations that correlate with decreases in TER and agonist-induced anion secretion and increases in paracellular permeability of normally impermeant proteins. Thus, Par-C10 cell monolayers provide an in vitro model for studying cytokine-dependent alterations in salivary gland epithelial functions related to saliva secretion and suggest that TNF-α and IFN-γ contribute to secretory dysfunction in SS by disrupting TJ integrity.

GRANTS

This work was supported by the National Institutes of Health National Institute of Dental and Craniofacial Research Grants R01-DE-017591-01, R01-DE-07389-19, and K08-DE-017633-01, and a Sjögren's Syndrome Foundation Research Grant.

Acknowledgments

The authors acknowledge R. Tindall from the University of Missouri Electron Microscopy Core Facility and R. Nessler from the University of Iowa Central Microscopy Research Facility for assistance in the preparation and imaging of specimens for this study.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract