Epithelial Barrier Defects in HT-29/B6 Colonic Cell Monolayers Induced by Tumor Necrosis Factor α

Epithelial Barrier Defects in HT-29/B6 Colonic Cell Monolayers Induced by Tumor Necrosis Factor α ALFRED H. GITTER, a,b KERSTIN BENDFELDT, a HEINZ SCHMITZ, c JÖRG-DIETER SCHULZKE, c CARL J. BENTZEL, d AND MICHAEL FROMM a a Institut für Klinische Physiologie and c Medizinische Klinik I, Gastroenterologie und Infektiologie, Universitätsklinikum Benjamin Franklin, Freie Universität Berlin, 12200 Berlin, Germany d Department of Medicine, East Carolina University, Greenville North Carolina 27858, USA ABSTRACT: The barrier function of intestinal epithelia relies upon the continuity of the enterocyte monolayer and intact tight junctions. After incubation with tumor necrosis factor TNF-, however, the number of strands that form the tight junctions decreases, and apoptosis is induced in intestinal epithelial cells. These morphological changes lead to a rise of transepithelial ion permeability, because the paracellular ion permeability increases and leaks associated with sites of apoptosis increase by number and magnitude. Thus apoptosis and degradation of tight junctions contribute to the increased permeability observed after exposure to TNF-. These mechanisms explain clinical manifestations in the inflamed intestinal wall containing cytokinesecreting macrophages for example, leak flux diarrhea and invasion of bacterial enterotoxins. The epithelial monolayer that lines the luminal surface of the intestine separates the mucosal and the serosal side by sealing the intercellular space with tight junctions. Impairment of the barrier function of intestinal epithelia may be the predominant mechanism in the pathogenesis of diarrhea in intestinal bowel disease (IBD). 1 Secretory mechanisms as a cause for IBD-related diarrhea are unlikely, because the inflamed mucosa shows decreased, rather than increased, rheogenic transport. 2 During the inflammation, cytokine-secreting macrophages are recruited, and thus the intestinal concentration of tumor necrosis factor α (TNF-α) rises. 3 5 Up to 10 ng/ml of bioactive TNF-α were measured in the circulation of patients with intestinal bowel disease. 6,7 Hence, 100 ng/ml or more may be reached in the relatively small volume of interstitial fluid space at the site of inflammation. Exposure to TNF-α leads to increased ion permeability of epithelia in vitro for example, in LLC-PK1 b Address for correspondence: Dr. Alfred H. Gitter, Institut für Klinische Physiologie, Universitätsklinikum Benjamin Franklin, 12200 Berlin, Germany. Voice: +49-30-8445-2789; fax: +49-30-8445-4239. gitter@medizin.fu-berlin.de 193

194 ANNALS NEW YORK ACADEMY OF SCIENCES kidney cells, 8 endothelial cells, 9 HT-29 clone 19A 10, and T84 intestinal cells. 11 In addition, however, there is evidence that macrophages induce apoptosis of enterocytes. 12 In order to investigate the putative role of TNF-α in epithelial barrier defects, we looked for morphological changes in cultured monolayers of human intestinal epithelial cells (HT-29/B6). 13 Employing the conductance scanning technique, 14 we determined the contribution of apoptosis to the overall conductivity of the epithelium. METHODS Cell Culture HT-29/B6, a subclone of the human colorectal cancer cell line HT-29, grows as highly differentiated, polarized epithelium with properties of Cl and mucussecreting crypt cells. In order to elicit apoptosis, the serosal side of HT-29/B6 cells was incubated in culture medium containing 100 ng/ml of TNF-α for 7 hours. Recombinant human TNF-α (10 7 units/mg) was provided by Schering (Berlin). Electrophysiology Confluent monolayers were mounted horizontally between the two halves of the conductance scanning chamber described previously. 14 During electrophysiological recordings, mucosal and serosal surface of the epithelium were superfused with (concentrations in mmol/l) 113.6 NaCl, 2.4 Na 2 HPO 4, 0.6 NaH 2 PO 4, 21 NaHCO 3, 5.4 KCl, 1.2 CaCl 2, 1.2 MgCl 2, 10 D(+)-glucose, 2.5 L-glutamine, D(+)-mannose, and β-hydroxybutyric acid. The solution was gassed with carbogen (95% O 2, 5% CO 2 ), which set the ph to 7.4. The temperature was maintained at 37 C. Ussing techniques and flux measurements were performed as described previously. Samples for determination of the serosal-to-mucosal flux of 22 Na + were taken at regular 30-minute intervals. Conductance scanning experiments were performed under bright-field light microscopical control. Two pattern types could be discriminated in the epithelial monolayer: nonapoptotic areas with approximately hexagonal symmetry and apoptotic spots that where surrounded by neighboring cells positioned in a distinctive rosettelike arrangement. The local conductivity (conductance referred to gross tissue area) was probed by measurement of the local electric field, generated by sinusoidal transepithelial current (AC, 0.3 ma/cm 2, 24 Hz) in the bath solution on the the mucosal side of the epithelium. By multiplication of the field strength measured locally and the specific resistivity of the bath solution, the current density was calculated. In nonapoptotic areas the distribution of transepithelial current was homogeneous, and the conductivity equaled the ratio of current density to transepithelial voltage. Near apoptotic rosettes the transepithelial current was inhomogeneously elevated. Here the current associated with a single rosette was computed by numerical integration of the current density exceeding the current density of nonapoptotic areas. From rosette current and transepithelial voltage, the conductance of a single rosette was derived.

GITTER et al.: BARRIER DEFECTS BY TNF- 195 All values are given as mean ± SEM for 20 apoptotic sites in 9 control monolayers or 21 apoptoses in 11 monolayers treated with TNF-α. Histochemistry In order to prove that the rosettes observed with intravital light microscopy indeed indicate apoptosis in the center of the rosette, we demonstrated apoptosis using different histological techniques. Monolayers were fixed in formaldehyde at room temperature and embedded in paraffin. Serial sections, made with a thickness of 3 µm, were stained with hematoxylin and eosin or dewaxed for immunofluorescence localization of apoptoses. Cellular DNA was either stained with 4,6-diamidino-2-phenylindole-2-HCl (DAPI) or inspected with a TUNEL (TdT-mediated X-dUTP nick end labeling) assay. In the latter, the blunt ends of double-stranded DNA exposed by strand breaks were, according to the description of the manufacturer (Boehringer Mannheim, Germany), visualized by means of enzymatic labeling (using terminal deoxynucleotidyl transferase) of the free 3 -OH termini with modified nucleotides (fluorescein-dutp) and fluorescence microscopy. Electron Microscopy Mounted in the in vitro setup, HT-29/B6 monolayers were fixed at room temperature with phosphate-buffered 2% glutaraldehyde. For scanning electron microscopy the tissue was frozen in Freon 22 and liquid nitrogen ( 100 C) and fractured with a double replica device. Freeze fractures were shadowed with platinum and carbon and examined in a Phillips 200 electron microscope. Morphometric analysis was performed using coded prints of freeze-fracture electron micrographs ( 60,000 magnification) on all tight junction regions with a delimited meshwork of strands in the tight junction. Vertical grid lines perpendicular to the most apical strand were drawn at intervals of 167 nm. Counting the intersections of strands with a grid line yielded its number of strands. The distance between the most apical and the most basal strands defined the depth of the tight junction. Aberrant strands were not observed, either in controls or after TNF-α exposure. In controls, 280 grid lines from 22 different tight junction regions were analyzed; in monolayers exposed to TNF-α, 320 grid lines from 22 tight junctions were analyzed. Ultrathin sections for transmission electron microscopy were prepared following standard procedures. RESULTS Overall transepithelial ion permeability of HT-29/B6 monolayers was evaluated by measurement of total epithelial conductivity (or its reciprocal, transepithelial resistance). After incubation with TNF-α (100 ng/ml), the transepithelial conductivity increased from 3.24 ± 0.07 ms cm 2 (309 Ω cm 2 ) to 14.65 ± 0.31 ms cm 2 (68 Ω cm 2 ). The short circuit current did not change under TNF-α; it was 0.1 ± 0.01 µmol h 1 cm 2 (n = 6) under control conditions and with the cytokine.

196 ANNALS NEW YORK ACADEMY OF SCIENCES FIGURE 1. The junctional region of HT-29/B6 monolayers comprised a delimited meshwork of strands in the tight junction. The freeze-fracture electron micrographs were made (a) under control conditions and (b) after incubation with 100 ng/ml TNF-α for 24 hours. The depth of the tight junction decreased from 265 ± 17 (control, n = 22) to 200 ± 14 nm (TNF-α, n = 22; p < 0.01).

GITTER et al.: BARRIER DEFECTS BY TNF- 197 FIGURE 2. Histograms showing the relative frequency of the number of horizontally oriented junctional strands for all tight junction sections analyzed. (a) Control monolayers (n = 280); (b) exposed to TNF-α (n = 320). Necrosis could be excluded, because the percentage of LDH released into the bath solution was equal in controls and monolayers with 24 h of exposure to TNF-α (4.0 ± 0.3% versus 4.2 ± 0.3%, n.s.). TNF-α was not effective after it was boiled (100 C, 30 min), excluding the possibility of artifacts due to endotoxin contamination. Degradation of Tight Junctions Freeze-fracture replicas of tight junctions in regular Ringer s showed the typical branching and anastomosing network of strands that appeared as ridges or furrows in the plane of the cell membrane (FIG. 1a). In the 280 tight junction sections inspected, the number of horizontally oriented strands was between 2 and 9, with a mean of 4.7 ± 0.2 (FIG. 2a). The mean depth of the tight junction was 267 ± 18 nm (22 cells). The network of strands was also observed after incubation with TNF-α (FIG. 1b); the mean number of strands had decreased to 3.3 ± 0.2 (p < 0.001 vs. control) (FIG. 2b), and the depth of tight junctions to 200 ± 14 nm (p < 0.01). In contrast to control cells, TNF-α treated cells showed a significant amount of tight junctional regions with only 1 or 2 strands (FIG. 2). With and without TNF-α, the histogram of strand numbers appeared Gaussian, but tight junctional regions without horizontal strands cannot be excluded because the method does not assess those. These findings suggest that the TNF-α induced decrease in the the number of strands affects the majority of the tight junctions in the same way, probably by interference with assembly/disassembly mechanisms. Paracellular Permeability In order to assess paracellular permeability, we measured the serosal-to-mucosal flux of 22 Na +. Under control conditions, it was 1.6 ± 0.1 µmol h 1 cm 2 (n = 6). After 4.5 hours of preincubation with TNF-α, the flux rose to 10.2 ± 1.0 µmol h 1 cm 2 (n = 6). Thus, the paracellular ion permeability increased under TNF-α.

198 ANNALS NEW YORK ACADEMY OF SCIENCES FIGURE 3. Monolayers of living HT-29/B6 cells in the conductance scanning chamber, viewed by light microscopy with bright-field illumination, show clusters suggestive of a rosette, composed of radially elongated cells around a central apoptotic cell ( arrow). Evidence of Apoptosis In the chamber used for conductance scanning experiments, the living monolayer was observed with an 40 water immersion object lens. Within the otherwise inconspicious cell sheet, we found rosette-like arrangements composed of radially elongated cells with a small cell (if any) in the center (FIG. 3). The frequency of these clusters, 3,700 ± 300 cm 2 under control conditions, increased to 11,100 ± 500 cm 2 after TNF-α treatment. The rosettes resembled the tissue remodeling during TNF-α induced apoptosis in LLC-PK1 renal epithelial cells. 16 Hence we performed several histological tests in order to prove apoptosis. Apoptotic fragments of the nucleus in the center of the rosette were visualized by staining with hematoxylin-eosin or the fluorochrome DAPI, or by fluorescent DNA end labeling (TUNEL). Transmission electron micrographs showed apoptotic cells and bodies in the center of rosettes and fragments engulfed by adjacent cells (FIG. 5). In some sections there was a hole in the center of the rosette, suggesting a superficial lacuna due to incomplete closure of the space that had been occupied by the apoptotic cell (FIG. 4). Conductance Scanning In order to determine the local conductance associated with single apoptoses, we scanned the monolayer with a conductance probe. 14 In nonapoptotic areas, a con-

GITTER et al.: BARRIER DEFECTS BY TNF- 199 FIGURE 4. Transmission electron micrograph demonstrating a central hole with apoptotic bodies (short arrow) and cytoplasma vacuoles (long arrow) in a section through the apical part of intestinal epithelial cells forming an apoptotic rosette. This finding suggests incomplete closure of the space that had been occupied by the apoptotic cell. The mean cell diameter was 6 µm. stant value reflecting the basic epithelial conductivity was recorded. Above apoptotic rosettes, however, the local conductivity increased along a line between the area with homogeneous conductivity and the rosette s center. Integration yielded the conductance associated with a single apoptotic rosette. The morphological appearance of a rosette correlated with its conductance. Three types were distinguished, as illustrated in the cartoon (FIG. 6): (a) a shrunk, sharply contoured cell with almost no conductance (except the basic epithelial conductivity);

200 ANNALS NEW YORK ACADEMY OF SCIENCES FIGURE 5. Transmission electron micrograph of a section through the basal part of intestinal epithelial cells forming an apoptotic rosette. In this plane, a single apoptic cell is surrounded by normal neighbors (rosette). Hence, there is no hole of cellular dimension in the epithelium, but the relatively wide intercellular space with few interdigitating membrane foldings suggests the possibility of disconnection between the apoptic cell and its neighbors. FIGURE 6. Three types of apoptotic rosettes observed with intravital light microscopy were correlated with apoptotic conductance: (a) a shrunk, sharply contoured cell, with no conductance (except the basic epithelial conductivity); (b) a transparent center and apoptotic fragments, with low or high conductance; (c) no central cell and the surrounding cells closing in, with no or low conductance.

GITTER et al.: BARRIER DEFECTS BY TNF- 201 FIGURE 7. Histogram showing epithelial conductivity and its components, basic epithelial conductivity and apoptotic conductivity, in HT-29/B6 monolayers under control conditions (left) and with TNF-α (right). After exposure to TNF-α, the epithelial conductivity rose dramatically, because both basic epithelial conductivity as well as frequency and conductance of apoptotic rosettes increased. (b) a transparent center and apoptotic fragments, with low or high conductance; and (c) no central cell and the surrounding cells closing in, with no or low conductance. Incubation with TNF-α increased the mean conductance of single apoptoses from 48 ns to 597 ns. Together with the three times higher frequency of apoptoses under TNFα, the contribution of apoptoses to the total epithelial conductivity increased from 0.18 ± 0.02 ms cm 2 (5.5%) under control conditions to 6.57 ± 0.30 ms cm 2 (45% of the total epithelial conductivity). However, the basic epithelial conductivity also increased under TNF-α, from 3.06 ± 0.07 ms cm 2 to 8.08 ± 0.05 ms cm 2 (FIG.7). DISCUSSION The present report documents two morphological alterations of the intestinal epithelial cells treated with TNF-α: degradation of tight junctions and apoptosis. With the conductance scanning technique the contribution of both effects to the increase in epithelial ion permeability was measured. TNF-α, applied at 100 ng/ml increased the basic epithelial conductivity and led to apoptosis associated with pronounced local barrier defects. Measurement of serosal-to-mucosal flux of 22 Na + indicated that the increase in basic epithelial conductivity was caused by an increase in paracellular conductivity. The latter can be attributed to a reduction in complexity of the network of strands that form the tight junction, as demonstrated by freeze-fracture analysis of TNF-α treated monolayers. Absence of changes in short-circuit current in monolayers treated with TNF-α indicate that transcellular ion permeability plays a minor role. Chloride secretion induced by TNF-α via prostaglandins is short-lived and not accompanied by a significant change in the transepithelial resistance. 17,18

202 ANNALS NEW YORK ACADEMY OF SCIENCES Our data demonstrate leaks in an epithelial monolayer caused by spontaneous apoptosis and, with more pronounced effect, by TNF-α induced apoptosis. The higher rate and leakiness of apoptoses, contrasting TNF-α treated with control epithelia, may be related to the degradation of tight junctions observed in freeze fractures, because cell cell contacts can be involved in the control of apoptosis. 16 Thus, epithelial integrity may be compromised during TNF-α mediated inflammatory processes, causing a clinical manifestation of leakiness in the intestinal wall diarrhea, for example. Invasion of bacterial enterotoxins may start a vicious cycle. ACKNOWLEDGMENTS We thank D. Sorgenfrei and S. Lüderitz for electrifying and cultivating collaboration. The work was supported by the Deutsche Forschungsgemeinschaft and by Freie Universität Medical Faculty funds. REFERENCES 1. SCHULZKE, J.D., M. FROMM, H. SCHMITZ, C. BARMEYER & E.O. RIECKEN. 1995. Barrier impairment in gastrointestinal diseases: ulcerative colitis and celiac sprue. Z. Gastroenterol. 33: 571 572. 2. SANDLE, G.I., N. HIGGS, P. CROWE et al. 1990. Cellular basis for defective transport in inflamed human colon. Gastroenterology 99: 97 105. 3. SARTOR, R.B. 1994. Cytokines in intestinal inflammation. Pathophysiological and clinical considerations. Gastroenterology 106: 533 539. 4. BREESE, E.J., C.A. MICHIE, S.W. NICHOLLS et al. 1994. Tumor-necrosis factor-α producing cells in the intestinal mucosa of children with inflammatory bowel disease. Gastroenterology 106: 1455 1466. 5. MCKAY, D.M., K. CROITORU & M.H. PERDUE. 1996. T cell-monocyte interactions regulate epithelial physiology in a coculture model of inflammation. Am. J. Physiol. 270: C418 C428. 6. SATEGNA-GUIDETTI, C., R. PULITANO, L. FENOGLIO et al. 1993. Tumor necrosis factor/ cachectin in Crohn s disease. Relation of serum concentration to disease activity. Rec. Prog. Med. 84: 93 99. 7. MARANO C.W., S.A. LEWIS, L.A. GARULACAN et al. 1998. Tumor necrosis factor-alpha increases sodium and chloride conductance across the tight junction of CACO-2 BBE, a human intestinal epithelial cell line. J. Membr. Biol. 161: 263 274. 8. MULLIN, J.M., K.V. LAUGHLIN, C.W. MARANO et al. 1992. Modulation of tumor necrosis factor-induced increase in renal (LLC-PK 1 ) transepithelial permeability. Am. J. Physiol. 263: F915 F924. 9. BURKE-GAFFNEY, A. & A.K. KEENAN. 1993. Modulation of human endothelial cell permeability by combinations of the cytokines interleukin-1 α/β, tumor necrosis factorα and interferon-γ. Immunopharmacology 25: 1 9. 10. RODRIGUEZ, P., M. HEYMAN, C. CANDALH et al. 1995. Tumor necrosis factor-α induces morphological and functional alterations of intestinal HT-29 cl.19a cell monolayers. Cytokine 7: 441 448. 11. MADARA, J.L. & J. STAFFORD. 1989. Interferon-γ directly affects barrier function of cultured intestinal epithelial monolayers. J. Clin. Invest. 83: 724 727. 12. IWANAGA, T., O. HOSHI, H. HAN et al. 1994. Lamina propria macrophages involved in cell death (apoptosis) of enterocytes in the small intestine of rats. Arch. Histol. Cytol. 57: 267 276. 13. KREUSEL, K.M., M. FROMM, J.D. SCHULZKE & U. HEGEL. 1991. Cl secretion in epithelial monolayers of mucus-forming human colon cells (HT-29/B6). Am. J. Physiol. 261: C574 582.

GITTER et al.: BARRIER DEFECTS BY TNF- 203 14. GITTER, A.H., M. BERTOG, J.D. SCHULZKE & M. FROMM. 1997. Measurement of paracellular epithelial conductivity by conductance scanning. Pflügers Arch. 434: 830 840. 15. SCHULZKE, J.D., M. FROMM, C.J. BENTZEL et al. 1992. Epithelial ion transport in the experimental short bowel syndrome of the rat. Gastroenterology 102: 497 504. 16. PERALTA SOLER, A., J.M. MULLIN, K.A. KNUDSEN & C.W. MARANO. 1996. Tissue remodeling during tumor necrosis factor-induced apoptosis in LLC-PK1 renal epithelial cells. Am. J. Physiol. 270: F869 879. 17. SCHMITZ, H., M. FROMM, C.J. BENTZEL et al. 1996. Tumor necrosis factor alpha induces Cl and K + secretion in human distal colon driven by prostaglandin E 2. Am. J. Physiol. 271: G669 G674. 18. KANDIL, H.M., H.M. BERSCHNEIDER & R.A. ARGENZIO. 1994. Tumour necrosis factor α changes porcine intestinal ion transport through a paracrine mechanism involving prostaglandins. Gut 35: 934 940. 19. HEYMAN M., N. DARMON, C. DUPONT et al. 1994. Mononuclear cells from infants allergic to cow s milk secrete tumor necrosis factor-α, altering intestinal function. Gastroenterology 106: 1514 1523. 20. RUSSELL, R.G., M. O DONNOGHUE, D.C. BLAKE JR. et al. 1993. Early colonic damage and invasion of Campylobacter jejuni in experimentally challenged infant Macaca mulatta. J. Infect. Dis. 168: 210 215. 21. MCKAY, D.M. & P.K. SINGH. 1997. Superantigen activation of immune cells evokes epithelial (T84) transport and barrier abnormalities via IFN-γ and TNFα: inhibition of increased permeability, but not diminished secretory responses by TGF-β 2. J. Immunol. 159: 2382 2390.

本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 (www.xuebalib.com) 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具