Human Intestinal Epithelial Cells of Proinflammatory Cytokine

The Regulation and Functional Consequence of Proinflammatory Cytokine Binding on Human Intestinal Epithelial Cells

This information is current as Asit Panja, Stan Goldberg, Lars Eckmann, Priya Krishen and of September 27, 2021. Lloyd Mayer J Immunol 1998; 161:3675-3684; ; http://www.jimmunol.org/content/161/7/3675 Downloaded from

References This article cites 63 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/161/7/3675.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 27, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1998 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Regulation and Functional Consequence of Proinﬂammatory Cytokine Binding on Human Intestinal Epithelial Cells1

Asit Panja,2* Stan Goldberg,* Lars Eckmann,† Priya Krishen,* and Lloyd Mayer*

Products of an activated immune system may affect cells within the immune system as well as nonlymphoid cells in the local environment. Given the immunologically activated state of the intestinal tract, it is conceivable that locally produced cytokines could regulate epithelial cell function. To assess whether epithelial cells are targets for particular cytokines, we initiated studies on the binding of a panel of proinflammatory cytokines in freshly isolated epithelial cells from normal and inflammatory bowel disease (IBD) patients as well as in cell lines. Isolated intestinal epithelial cells (IEC) were stained with phycoerythrin-conjugated ␤ or biotinylated cytokines to determine the expression and density of receptors for IL-1 , IL-6, granulocyte-macrophage CSF Downloaded from (GM-CSF), and TNF-␣. Receptors for IL-1␤, IL-6, and GM-CSF were readily detectable in all epithelial cell preparations at levels equal to (GM-CSFR) or lower than those seen on monocytes. However TNF␣-R were not detectable on freshly isolated IECs. Receptor density was greater in surface vs crypt epithelial cells, but no significant differences were seen between normal and IBD epithelial cells. Expression of IL-1R and IL-6R was enhanced by LPS and IFN-␥. Functionally, IL-1␤ enhanced proliferation of the IEC cell line, DLD1, whereas GM-CSF treatment of de-differentiated crypt-like DLD1 and HT29 cells resulted in enhanced

expression of ICAM-1. Furthermore, TNF-␣ treatment enhanced the secretion of IL-8 and GRO-␣ in HT29 cells, but not in freshly http://www.jimmunol.org/ isolated IEC cultures. The differential binding and function of proinﬂammatory cytokines on IEC support the hypothesis that these cytokines may be involved in normal physiological processes as well as in regulating mucosal immune responses. The Journal of Immunology, 1998, 161: 3675–3684.

he pathophysiological mechanisms underlying the devel- propria lymphocytes (15–19) from inflammatory bowel disease opment of IBD are not fully understood. In ulcerative (IBD) patients. There is increasing evidence that intestinal epithe- T colitis (UC)3, the final target of the inflammatory process lial cells (IEC) may serve as targets for these locally produced appears to be the epithelium, and epithelial cell dysfunction has cytokines. In this scenario, local cytokines may result in alterations been demonstrated. In Crohn’s disease (CD), inflammation ex- of epithelial cell function or integrity. For example, the combina- by guest on September 27, 2021 tends transmurally, and the epithelium may be spared from de- tion of TNF-␣ and IFN-␥ is cytotoxic for the colonic epithelial cell struction. However, there is evidence that epithelial cell dysfunc- line, HT29 (20), and IL-1␤ up-regulates IL-6 synthesis (a proin- tion exists in this disease as well. Although separate cellular or flammatory cytokine) by epithelial cells (21). Alternatively or ad- molecular events may be involved in these inflammatory pro- ditionally, cytokines may play a role in the normal regulatory cesses, it has been widely accepted that locally produced cytokines mechanisms involved in epithelial cell growth and differentiation may play a critical role, comparable to their effects in a number of ensuring maintenance of homeostasis. An imbalance of the normal other chronic inflammatory disorders such as rheumatoid arthritis cytokine milieu seen in IBD might lead to the disruption of epi- (1–7), psoriasis (8), and glomerulonephritis (9, 10). Proinflamma- thelial function. There is evidence for selective effects of various tory cytokines such as IL-1, IL-6, granulocyte-macrophage CSF cytokines on IEC growth and differentiation, phenotype, and func- (GM-CSF), and TNF-␣ have been reported to be increased in se- tion through an autocrine/juxtacrine or paracrine pathway. One rum (11), cultures of biopsy tissues (12–14), or isolated lamina hypothetical model based on these scenarios might be that cytokine-mediated changes in the epithelial cell integrity might initiate pathologic events in the gut. The ability of cytokines to function in *Division of Clinical Immunology, Mount Sinai Medical Center, New York, NY this manner depends on the presence of specific receptors ex- 10029; and †Department of Medicine, University of California at San Diego, Lo Jolla, CA 92093 pressed on IECs. Therefore, we analyzed the binding of the proinflammatory cy- Received for publication October 23, 1997. Accepted for publication May 27, 1998. tokines IL-1␤, IL-6, GM-CSF, and TNF-␣ on both freshly isolated The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance intestinal epithelial cells and cell lines by flow cytometric analysis. with 18 U.S.C. Section 1734 solely to indicate this fact. Our results show that IECs from both IBD and normal control 1 This work was supported by a First Award Grant (to A.P.) and a Career Develop- samples express receptors for GM-CSF, IL-1␤, and IL-6 at com- ment Award (to L.E.) from the Crohn’s and Colitis Foundation of America and U.S. parable levels. TNF␣-R were not detected on freshly isolated ep- Public Health Service Grants AI-23504, AI-24671, and DK-44156 (to L.M.). ithelial cells, although they were present on colonic adenocarci- 2 Present address, correspondence and reprint request to Dr. Asit Panja, Division of Gastroenterology, Hepatology and Nutrition, Winthrop University Hospital, 222 Sta- noma cell lines. Functionally, IL-1 enhanced while GM-CSF tion Plaza North, Suite 511A, Mineola, Long Island, New York 11501. inhibited IEC growth. Furthermore, GM-CSF could induce or up- 3 Abbreviations used in this paper: UC, ulcerative colitis; CD, Crohn’s disease; GM- regulate ICAM-1 expression preferentially on crypt-like epithelial CSF, granulocyte-macrophage CSF; IBD, inflammatory bowel disease; IEC, intestinal cells. Taken together, these results support the potential for an epithelial cell; MCF, mean channel fluorescence; PE, phycoerythrin; ECAM-1, endothelial cell adhesion molecule-1; ELAM-1, endothelial leukocyte adhesion autocrine or paracrine effect of proinflammatory and prophago- molecule-1. cytic cytokines as potential modulators of epithelial cell integrity.

Materials and Methods Cell lines Isolation of IEC Cell lines (EL-4, L929, DLD1, HT-29, and Caco2) were obtained from the American Type Culture Collection (ATCC; Manassas, VA). The IL-6 re- Epithelial cells from resected bowel specimens were isolated by dispase or sponsive plasmacytoma cell line B9, was kindly provided by Dr. Lester EDTA (TNF␣-R) treatment as previously described (22, 23). Surgical May (New York Medical College, Valhalla, NY). Cell lines were either specimens included those from “normal” controls, i.e., patients with ade- used as positive controls for receptor assays or as sources of intestinal nocarcinoma (10 cm from tumor), colonic inertia, and familial polyposis, epithelium. Cells were thawed and grown in T75 tissue culture flasks as well as those from patients with active and inactive UC and CD. In the (Nunc, Naperville, IL). DLD1, HT29, and L-929 cell lines were grown in latter cases, tissues were obtained from both areas of active inflammation RPMI 1640 medium supplemented with 10% FCS (Life Technologies, and areas where no histologic inflammation was noted (where available). Gaithersburg, MD), Caco2 in DMEM with 10% FCS, EL-4 in DMEM with Briefly, after vigorous washing in PBS, the mucosal layer was dissected 10% horse serum, and B9 cells in RPMI 1640 with 10% FCS and 20 U/ml from the underlying submucosa. The dissected mucosal tissue was washed rIL-6 (kind gift of Dr. Edward Siden, Mount Sinai Medical Center, NY). several times in HBSS supplemented with 1% penicillin-streptomycin, am- All culture media contained 1% penicillin-streptomycin (ICN-Flow. photericin, and gentamicin (50 ␮g/ml), minced into small pieces, and fol- McLean, VA) and L-glutamine (ICN-Flow). Cells from suspension cultures lowed by treatment with the mucolytic agent DTT (1 mM) (Sigma, St. (EL-4, B9) or confluent monolayers (DLD1, HT29, Caco2), dissociated Louis, MO) in RPMI 1640 medium for 5 min at room temperature. Tissue from the plastic surface by treating with a nonenzymatic cell dissociation pieces were washed in PBS and subjected to dispase (1 mg/ml; Boehringer solution (Sigma), were washed and used for staining. In some experiments, Mannheim, Indianapolis, IN) or EDTA (0.75 mM; Sigma) treatment for 30 control cell lines or IEC lines were treated with dispase or EDTA to assess min at 37°C. During this 30-min incubation, tissue pieces were agitated the effects of the isolating procedure on cytokine binding. For receptor every 5 min by vortex mixing, resulting in the liberation of epithelial cells regulation studies secondary cultures were generated in the presence or and intraepithelial lymphocytes into the solution. This cell suspension was absence of modulatory agents for varying time periods. Receptor density collected and immediately diluted 10-fold in PBS to minimize any effect of was compared between stimulated and unstimulated cells by analysis of Downloaded from dispase on cytokine receptors (see below). Cells were spun down and mean fluorescence intensity or peak channel (see below). washed twice in PBS. Epithelial cells were further separated from the contaminating intraepithelial lymphocytes and RBCs by percoll density gra- In vitro models of crypt and surface epithelial cells dient centrifugation (23). The cells on top of the 30% layer contained 95% pure epithelial cells (Ͻ1% CD45ϩ cells or CD14ϩ cells) with a viability of In vitro models of fully differentiated or undifferentiated human colonic 95 to 100%. There was no difference in the viability of normal IECs vs epithelial cell lines were developed by using glucose free or high glucose those obtained from patients with IBD. containing conditioned medium, respectively, as previously described by For the purpose of obtaining crypt epithelial cells, minced tissue pieces Wice et al. (30). Briefly, HT29 (obtained from Dr. Burton M. Wice, Wash- http://www.jimmunol.org/ were incubated with dispase for up to five treatments (5 min for the first ington University School of Medicine, St, Louis, MO) and DLD1 cells treatment, 30 min for the second and third treatments, and then 40 min each (from ATCC) were cultured in glucose-free DMEM with 10% dialyzed for the next two treatments). At each interval, sample tissue (for histologic FCS (Summit Biotechnology, Fort Collins, CO) supplemented with 2.5 examination to detect the disappearance of surface epithelium and subse- mM inosine (Sigma) (for differentiated surface-like cells designated HT29ino or DLD1ino in the text) or 25 mM glucose (for de-differentiated quent crypt cell liberation) and cell suspensions were collected. Epithelial crypt-like cells designated HT29glu or DLD1glu in the text). Cells were cells were separated by percoll density gradient, and purity of IECs was grown for at least four passages and maintained in a postconfluent state for Ͼ95% (as evaluated by staining with anti-CD3, CD14, CD20, CD45 at least 10 days with daily replacement of fresh medium before their use in mAbs) and L12 (anti-epithelial cell) Abs. Hematoxylin and eosin staining the experiments. Although the features of full differentiation (in inosine of histologic sections of the treated tissue samples demonstrated that the cells) or de-differentiation (in glucose cells) of these cells have been char- surface epithelium was completely removed by the first three dispase treat- by guest on September 27, 2021 acterized previously by Wice et al. (30, 31) under these conditions, we ments with maintenance of the crypt epithelium. After the fifth treatment reconfirmed the differentiation states of these cells by demonstrating lower histologic sections revealed the absence of crypt epithelium. The IEC from levels of alkaline phosphatase and sucrase isomaltase (markers of a mature the fourth and fifth dispase treatments were used as a source of crypt cells enterocyte phenotype) in the “glucose-conditioned” cells compared with whereas the IEC from the first and second dispase treatments were used as the higher levels of these enzymes in the “inosine-conditioned” cells (data the source of surface epithelium. Purity of the surface or crypt cell popu- not shown). lations in the cell suspension from each treatment was determined by mor- phology (by phase contrast microscopy and histochemical staining), mea- Cytokines and chemicals surement of alkaline phosphatase (as a measure of surface epithelium) (24, 25), and [3H]thymidine uptake incorporation (as a measure of the crypt cell Human recombinant IL-1␤ was obtained from Regeneron (Boston, MA), population) for 4, 16, and 24 h. Consistent with the findings in the histo- IL-6 was a generous gift from Dr. Edward Siden, human GM-CSF was logic sections, alkaline phosphatase was detectable only in the lysates from purchased from R&D Systems (Minneapolis, MN), TNF-␣ from Cetus the cells isolated in the first three dispase treatments, suggesting that these (Emeryville, CA), IFN-␥ from Genentech (South San Francisco, CA), and fractions contained an enriched population of surface epithelial cells. Escherichia coli LPS and phorbol 12-myristate 13-acetate (PMA) from [3H]Thymidine uptake incorporation was minimal in these cells, whereas Sigma. the cells liberated in the fourth and fifth dispase treatments were able to Identification of cytokine binding proliferate with a 4- to 5-fold increase in thymidine uptake seen at 24 h. These data suggest that the latter fractions are enriched for crypt epithelial Cytokine receptors were identified using fluorochrome-labeled cytokines cells. No alkaline phosphatase activity was noted in lysates of cells from (Fluorokine; R&D Systems) and analyzed by flow cytometry. We utilized the fourth and fifth dispase treatments. In some experiments intact crypts phycoerythrin (PE) or biotin (followed by avidin-FITC) conjugated cyto- were liberated during the treatment, and these could be visualized by con- kines as the signal-to-noise ratio was greater than that seen with fluores- ventional light microscopy. Such crypts were teased apart by incubation cein. Briefly, purified epithelial cells were washed twice in RDF-1 buffer with 0.15 mol/liter NaCl3 mmol/liter tetraphenyl borate (26, 27) and sub- (wash buffer for fluorokine staining from R&D Systems) and resuspended jected to analysis. at 4 ϫ 106 cells/ml. A total of 25 ␮l of cell suspension (105 cells) was incubated with 10 ␮l (chosen by performing serial dilution analysis with control cells) of either PE-conjugated IL-1␤, IL-6, GM-CSF, TNF-␣,or Isolation of peripheral blood monocytes streptavidin for1honice. In the cases where staining was performed with biotinylated cytokines, 5 ␮l (chosen from serial dilution studies in normal Peripheral venous blood was obtained from the same individual from control cells) of biotin-conjugated cytokines or soybean trypsin inhibitor whom the epithelial cells were derived and monocytes were isolated by (as a negative control) was added to cells for1honicefollowed by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) density gradient centrifuga- incubation with 5 ␮l avidin-FITC for another hour. After this incubation, tion and plastic adherence, as described (28). The nonadherent cells were cells were washed twice with 2 ml of RDF-1 buffer. Cells were resus- removed, and the adherent cells were washed with PBS to remove all pended in 200 ␮l of PBS and analyzed by an Epics Profile-I flow cytometer loosely bound contaminating cells. Adherent monocytes were harvested by (Coulter, Hialeah, FL). At least 3000 cells were counted per analysis. In a rubber policeman. Purity and viability were assessed by staining with the preliminary studies, specificity of these conjugated cytokines was deter- anti-CD-14 mAb Leu M3 (Becton Dickinson, San Jose, CA) and pro- mined by the inhibition of specific staining on receptor bearing control pridium iodide (29), respectively. In all cases purity was Ͼ95% and via- cells by unlabeled cytokine. Cells were first incubated with unlabeled cy- bility was 100%. tokine at 100-fold excess molar concentration (or the fluorescent cytokines The Journal of Immunology 3677 were neutralized by incubation with specific mAbs against the cytokines Results for 10 min before their incubation with cells) and then incubated with IL-6, GM-CSF, and IL-1␤ receptors are constitutively expressed labeled cytokine for another hour on ice following the protocol described above. To assess purity of the cell preparations, in each set of experiments on IECs parallel immunofluorescence staining was performed as described previ- Freshly isolated epithelial cells and appropriate positive control ously (21) with Abs against HLA-class I (W6/32), CD14 (Leu-M3), CD20, ␤ CD45 (HLe-1), and an epithelial cell specific mAb L12 (32) followed by cells were stained with either PE-conjugated IL-1 , IL-6, GM- Ј CSF, TNF-␣, or streptavidin to detect the constitutive expression FITC-conjugated F(ab )2 goat anti-mouse IgG. An irrelevant mouse IgG1 mAb served as a negative control. This allowed us to gate on the L12 of receptors for these cytokines. The flow cytometric analysis for positive epithelial cells, thus avoiding any false positive cytokine binding one representative experiment is shown in Figure 1. A peak shift from contaminating nonepithelial cells. In another series of experiments, in MCF demonstrates the presence and density of receptors dual labeling analyses were performed. Cells were first stained with L-12/ goat anti-mouse FITC as described above and then incubated with PE- compared with the streptavidin-PE negative control. In each conjugated cytokines to discriminate between epithelial cells and other case binding of PE-labeled cytokines was specific since the contaminating cells in our IBD epithelial cell preparations. In each assay, binding could be inhibited by the corresponding unlabeled hu- positive control cells for IL-6R (B9), IL-1R (EL-4), GM-CSFR (mono- man recombinant cytokine. cytes), and TNFR-␣ (L929) binding were included to compare receptor ␤ density and reliability of the technique. IL-1 R was constitutively expressed on freshly isolated IECs from normal tissues (Fig. 1A, upper). The specificity of IL-1␤ PE ␤ Measurement of IL-8 and GRO-␣ binding to IL-1 R was demonstrated by inhibition of binding of the labeled cytokine to EL-4 cells in a blocking assay using 100 Freshly isolated IEC and HT29 colonic adenocarcinoma cell lines were ␤ fold excess unlabeled rIL-1 (Fig. 1A, middle) but not with IL-6 or Downloaded from cultured in the presence or absence of TNF-␣ (33) for 48 h, and culture supernatants were assayed for IL-8 and GRO-␣ production by ELISA as GM-CSF (data not shown). Blocking experiments were also per- described previously (34). Briefly, a polyclonal goat anti-human (GAH) formed in a series of experiments with freshly isolated IECs (data IL-8 or a monoclonal GAH GRO-␣ (R&D Systems) Ab were used as not shown) with similar results. Dispase treatment of EL4 cells had capturing Abs with a polyclonal rabbit anti-human IL-8 and monoclonal no effect on IL-1␤R expression (Fig. 1A, lower). mouse anti-human GRO-␣ as the detecting Ab. The IL-8 and GRO-␣ ELISAs were sensitive to 20 and 50 pg/ml, respectively. As seen in Figure 1B, IL-6R was also expressed on IECs (upper) but at a lower density compared with IL-1␤R. Because we previ- http://www.jimmunol.org/ Determination of the cytokine effect on epithelial cell growth ously demonstrated that IL-6 is secreted by IEC (21), the lower

5 density of IL-6R may relate to autocrine blockade of IL-6R by DLD1 cell lines were seeded (5 ϫ 10 cells per well) in 6-well plates. ␤ Cytokines were added at the time of adherence to the culture well. Cells IEC-secreted IL-6. Similar to IL-1 R, IL-6-PE binding was spe- were removed at different time points (24 and 48 h) by treatment with cell cific (Fig. 1B, middle) and dispase treatment had no effect on dissociation solution and counted in a Neiubauer chamber (Hausser Sci- IL-6R expression (Fig. 1B, lower) on the IL-6-dependent B9 cell entific, Horsham, PA). Three wells were used for each data point and line. counting was performed by three investigators independently. The results We next examined expression of receptors for GM-CSF, which from direct cell counts were validated in a parallel series of [3H]thymidine uptake assays. IEC cell lines were grown in the presence or absence of enhances phagocytosis in macrophages (36). Figure 1C, upper, varying concentrations of cytokines in 96-well (5000 cells/well) flat-bot- demonstrates that GM-CSFRs are expressed on epithelial cells at by guest on September 27, 2021 tom tissue culture plates in a humidified 5% CO2 incubator. After 48 h of high density. Since, in human, GM-CSF and IL-3 have been re- 3 culture, 1 ␮Ci [ H]thymidine was added, and after6hoffurther incubation ported to share a common ␤2 subunit, we performed blocking the culture medium was aspirated and thymidine incorporation was mea- sured in a 1450 Microbeta Plus counter (Wallac, Gaithersburg, MD). experiments with cold IL-3 as well. Binding of GM-CSF PE was partially inhibited by both cold GM-CSF (Fig. 1C, middle) and IL3 Detection of cytokine-induced ICAM-1 on in vitro models of (data not shown). However, the ability of GM-CSF but not IL-3 to surface and crypt epithelial cells regulate surface Ags on IECs (37, 38) confirms that the staining relates to GM-CSFR and not IL-3R. Dispase had no effect on GM- Conditionally differentiated (inosine medium) or de-differentiated (glucose CSFR expression (Fig. 1C, lower). medium) cell lines (DLD1 and HT29) were grown in T25 tissue culture ␣ flasks (Falcon, Franklin Lakes, NJ) as described above. Each type of cell Interestingly, receptors for TNF- were not detected in any ep- line or freshly isolated crypt and surface IEC were cultured either with ithelial cell preparation (Fig. 1D, upper), although TNF␣-R were recombinant human GM-CSF (20 ng/ml) or medium alone for 24 h at clearly demonstrated in the colonic adenocarcinoma cell lines (see 37°C. The cell lines were dissociated by treatment with nonenzymatic cell Table I below). In this case, TNF␣-R expression in control cell dissociation solution (freshly isolated IEC do not adhere to the plastic surface and did not require this cell dissociation procedure after stimula- lines was reduced by dispase treatment; therefore, EDTA was tion), washed three times in PBS/1% BSA/0.02% sodium azide, and pro- utilized to isolate IECs in this series of experiments. EDTA had cessed for immunofluorescence staining using an anti-ICAM-1 mAb (Bec- minimal effects on TNF␣-R expression. The variability in the ton Dickinson) or an irrelevant mouse IgG1 control mAb. staining of the L929 cells related to a reduced background in this experiment. Flow cytometric analysis Stained cell suspensions were analyzed on a Epics Profile flow cytometer Comparison of cytokine receptor expression on IECs and (Coulter) as previously described (35), gating on viable cells. Mean chan- conventional APCs nel fluorescence (MCF), which correlates with fluorescence intensity, was determined from the peak of positively stained cells and is recorded on a Since IECs possess some of the functional properties ascribed to log scale. APCs, it was of interest to compare cytokine receptor expression on IECs with that of monocytes. Therefore we performed studies Stimulation of epithelial cells using paired samples of epithelial cells (n ϭ 5) and isolated mono- Freshly isolated epithelial cells or cell lines were cultured in the presence cytes (n ϭ 5) from the same individuals. As can be seen in Figure or absence of IL-1␤ (100 ng/ml and 1 ␮g/ml), IFN-␥ (250 U/ml), LPS (1 2, with the exception of GM-CSF, the density of receptor expres- ␮g/ml), GM-CSF (20 ng/ml), or PMA (10 ng/ml) for 24 h in culture me- sion in IECs for all cytokines, was lower than that of monocytes. dium in a 5% CO2 humidified 37°C incubator. These concentrations were ␣ optimal for induction of specific receptors in conventional APCs. Cells As described earlier (Fig. 1), no TNF -R was detected in any of were washed and checked for viability, and receptor expression was ana- the epithelial cell preparations, whereas TNF␣-R were present in lyzed by the method described above. the monocytes. GM-CSFR expression was the highest of any of the 3678 CYTOKINE BINDING BY INTESTINAL EPITHELIAL CELLS

FIGURE 2. Comparison of cytokine receptor density in monocytes (professional APC) vs IEC (nonprofessional APC). Comparison of mean channel ﬂuorescence of cytokine receptors expressed on monocytes/macrophages (M) vs epithelium (E) in paired samples. With the exception of GM-CSF, all cytokine receptors are expressed at lower (or absent—TNF) density on IEC. Downloaded from

cytokine receptors tested and was comparable to monocytes. Cy- tokine receptor expression and density on IEC was similar regard- less of the site from which the IEC were isolated (e.g. ileum vs colon, right colon vs left colon). The variability seen in receptor expression between the different individuals likely relates to indi- http://www.jimmunol.org/ vidual variation rather than differences in isolation technique.

Cytokine receptor expression on IECs does not correlate with severe histologic inﬂammation in IBD Given the above data, it is clear that epithelial cells might be targets for proinﬂammatory cytokines released by the epithelial cells themselves as well as by other mucosal cells during the process of

inflammation. Receptor-cytokine interactions could alter the phe- by guest on September 27, 2021 notype of the epithelial cells resulting in abnormal cytokine receptor expression, alteration in function, growth, or induction of specific costimulatory signals for T cell activation. To determine whether such a phenomenon may be responsible for the epithelial cell dysfunction seen in IBD (39), we studied the presence and density of receptor expression for the four cytokines in epithelial cells isolated from inflamed (UC, n ϭ 11; CD, n ϭ 10) and uninflamed (UC, n ϭ 2; CD, n ϭ 8) tissues from patients with IBD and compared them with normal control epithelial cells (n ϭ 16). As shown in Table I, receptor expression on epithelial cells derived from inflamed tissues was not significantly different from normal control IECs, although there was a trend toward an increase in GM-CSFR in uninflamed UC and CD. However, when we directly compared the cytokine receptor expression and density on epithelial cells derived from inflamed and uninflamed areas in the bowel of the same IBD patient, we found that cytokine receptor expression did not correlate with inflammation in either UC or CD (data not shown).

FIGURE 1. Flow cytometric analysis of cytokine receptor expression Table I. Comparison of cytokine binding on IEC: normal vs IBD on freshly isolated intestinal epithelial cells. PE-conjugated cytokines (10 ␮l) or streptavidin (negative control) were added to 1 ϫ 105 freshly isolated IEC (upper) or positive control cells (middle). A, IL-1␤ PE, EL4 cells Cytokine Receptor (log MCF Ϯ SE) are the positive control. B, IL-6-PE, B9 cells are the positive control. C, Cell Source IL-1␤ IL-6 GM-CSF GM-CSF PE, monocytes are the positive control. D, TNF-␣ PE, L929 cells are the positive control. In each case, cold (unlabeled) cytokine was added Normal (n ϭ 16) 1.444 Ϯ 0.173 1.000 Ϯ 0.002 1.169 Ϯ 0.205 at 100-fold excess to document speciﬁcity (middle; termed “block”). The UC inv (n ϭ 11) 0.946 Ϯ 0.210 0.727 Ϯ 0.179 1.006 Ϯ 0.241 lower panels represent the lack of effect of dispase (for IL-1, IL-6, GM- UC uninv (n ϭ 2) 1.655 Ϯ 1.225 0.910 Ϯ 0.890 1.935 Ϯ 1.465 ϭ Ϯ Ϯ Ϯ CSF) or EDTA (for TNF) treatment of the control cells in terms of cytokine CD inv (n 10) 1.035 0.226 0.685 0.158 1.242 0.263 ϭ Ϯ Ϯ Ϯ binding. CD uninv (n 8) 1.144 0.281 0.948 0.300 2.163 0.622 The Journal of Immunology 3679

Table II. Cytokine receptor expression on intestinal epithelial cell lines Table III. TNF-␣ stimulation does not affect chemokine secretion by freshly isolated IECs Cell Lines (log MCF Ϯ SE) Chemokines Produced (ng/ml) Cytokine HT29 DLD1 CaCo-2 Added TNF-␣ Streptavidin PE 1.850 Ϯ 0.091 1.483 Ϯ 0.552 1.486 Ϯ 0.672 Cells (ng/ml) IL-8 GRO-␣ IL-6 PE 2.376 Ϯ 0.233 6.703 Ϯ 2.15 4.381 Ϯ 3.001 IL-1␤ PE 3.442 Ϯ 1.293 6.161 Ϯ 2.786 5.071 Ϯ 3.850 HT29 0 0.79 2.31 GM-CSF PE 5.863 Ϯ 2.370 7.742 Ϯ 5.321 3.754 Ϯ 1.00 10 26.89 13.28 TNF-␣ PE 6.063 Ϯ 2.139 1.429 Ϯ 0.653 2.250 Ϯ 1.163 1 6.09 7.49 IEC no. 1 0 0.58 1.32 10 0.43 1.79 1 0.48 1.63 IEC no. 2 0 5.55 0.85 10 5.75 0.99 Cell lines differ from freshly isolated epithelial cells in cytokine 1 5.32 0.62 receptor expression To develop a model for assessing the regulation of cytokine receptor expression in IECs, we examined receptor expression in the Comparison of cytokine binding (receptor expression) between epithelial cell lines DLD1, CaCO2, and HT29. Receptors for IL-

surface and crypt epithelium Downloaded from 1␤, IL-6, and GM-CSF were expressed on these cell lines at levels The differences in cytokine receptor expression between the dif- greater than those seen in freshly isolated IECs (Table II). Fur- ferent epithelial cell lines raised the possibility that cytokine re- thermore, unlike freshly isolated IECs, TNF␣-R were detected on ceptor expression may vary depending upon the differentiation all three cell lines. state of the cells. Previous studies have shown that the response to The presence of TNF␣-R in the cell lines but not in freshly various hormones and physiological modulators differs in crypt isolated IECs raised concerns about the possibility that TNF␣-R on and surface epithelial cells. Therefore, to determine whether the http://www.jimmunol.org/ the epithelial surface might be blocked by the binding of endog- two cell types (crypt and surface IEC) display any difference in the ␣ enous TNF- , resulting in negative staining by the PE-conjugated cytokine receptor expression, we isolated surface and crypt epi- cytokine. Such a possibility was rendered less likely by the finding thelial cells from 4 individual specimens. As can be seen in Figure ␣ of an absence of TNF- protein or mRNA in isolated IEC (21). To 3, the surface epithelial cells expressed cytokine receptors (IL-1␤ ensure that our isolation procedure was not responsible for strip- and IL-6) at a higher density than crypt epithelium. These findings ping off TNF␣-R, we subjected macrophage and control cell lines were reconfirmed in an in vitro cell line model system of crypt and to the same isolation procedure. As discussed earlier, dispase treat- surface epithelial cells (see Materials and Methods). HT29 and ment did not change receptor density on either macrophages or the DLD1 cell lines were conditionally differentiated (HT29ino and cell lines for any cytokine except for TNF␣-R. However we were DLD1ino cells, comparable to surface/villus IEC) or de-differen- by guest on September 27, 2021 still unable to detect TNF␣-R on IEC when the IEC were isolated tiated (HT29glu and DLD1glu, comparable to crypt cells) by using using EDTA (which had no inhibitory effect on TNFR expression glucose free or high glucose containing medium (see Materials and in fact had a modest stimulatory effect). These data suggest and Methods). After growth under such conditions, experiments that the absence of TNF␣-R in freshly isolated epithelial cells is were performed to detect cytokine receptor expression on these not due to the isolation procedure but that tumor bearing epi- cells. Consistent with our results with freshly isolated crypt and thelial cell lines differ from fresh IECs in their cytokine recep- surface epithelial cells, IL-1␤ and IL-6R expression was greater in tor expression. the differentiated (HT29ino and DLD1ino) cells than the de-differentiated (HT29glu and DLD1glu) crypt-like cells (data not shown). As seen with freshly isolated IEC, the GM-CSFR density Demonstration of a lack of functional response to TNF-␣ by was similar in both differentiated and dedifferentiated cell lines. freshly isolated IEC The finding of cytokine receptor heterogeneity in the crypt vs surface epithelial cells may relate to specific cytokine requirements Although our results of TNF-␣ binding (receptor expression) by malignant epithelial cell lines have been consistent with previously described functional effects of TNF-␣ on these cell lines, the lack of TNF-␣ binding on freshly isolated IEC raises the concern as to whether the nonbinding of TNF-␣ on isolated IEC is an artifact of either endogenous TNF-␣ production or the isolation procedure. To address this issue we compared the functional response of the HT29 cell line and freshly isolated IEC to TNF-␣ stimulation by examining chemokine (IL-8 and GRO-␣) production by these cells. As can be seen in Table III, consistent with previously reported results, TNF-␣ stimulation of HT29 cells resulted increased production of both IL-8 and GRO-␣ in a dose-dependent fashion, whereas the freshly isolated IECs from two different individuals showed no response to TNF-␣ stimulation even at the higher concentration of 10 ng/ml. Furthermore, it appeared less likely that FIGURE 3. Comparison of cytokine receptor expression on isolated ␣ ␣ endogenous TNF- was present since baseline IL-8 and GRO- crypt vs surface epithelium. Expression is denoted as mean channel fluo- was comparable to the HT29 cell line. These data support the lack rescence (log). This figure represents an average of three separate of TNF␣-R expression on freshly isolated IEC. experiments. 3680 CYTOKINE BINDING BY INTESTINAL EPITHELIAL CELLS Downloaded from FIGURE 5. Regulation of epithelial cell line growth by IL-1␤, IL-6, and GM-CSF. DLD1 cells (1 ϫ 106) were cultured with either IL-1␤ (100 ng/ml), IL-6 (60 U/ml), or GM-CSF (20 ng/ml) for 24 and 48 h. Total viable cells were counted by trypan blue dye-exclusion method. The data are averages of triplicate determinations. Consistent with the [3H]thymidine incorporation assay (data not shown), cell growth was enhanced by

addition of IL-1␤ and inhibited by GM-CSF and to a lesser extent by IL-6. http://www.jimmunol.org/ This ﬁgure is representative of ﬁve experiments.

Figure 4 (average of five experiments), IL-1␤R expression was enhanced by all of the stimuli used, although this did not achieve statistical significance (Fig. 4A). Unlike monocytes (40), IL-6Rs on IEC (Fig. 4B) were up-regulated by LPS and unchanged with PMA, suggesting a different regulatory pathway for IL-6R expression in both conventional and mucosal APCs. This lack of induction of IL-6R cannot relate to the production of (unlabeled) IL-6 in by guest on September 27, 2021 the cell line as we have previously seen that DLD1 cells do not secrete this cytokine (A. Panja and L. Mayer, unpublished data). GM-CSFR expression was unaffected by PMA, LPS, or IFN-␥ (Fig. 4C). Finally, TNF␣-R was neither altered in DLD1 cells nor induced in freshly isolated epithelial cells (data not shown) by any of these stimuli. FIGURE 4. Regulation of cytokine receptor expression in the DLD1 We next assessed the functional role of the cytokine receptors colonic adenocarcinoma epithelial cell line. DLD1 cells were cultured in by assessing the effect of recombinant human cytokines on IEC the presence or absence of LPS (1 ␮g/ml), IFN-␥ (250 U/ml), or PMA (10 growth and surface molecule (ICAM-1) expression. ng/ml) for 18 h and stained with PE conjugated IL-1␤ (A), IL-6 (B), and GM-CSF (C) as described in Figure 1. The results are expressed as fold IL-1␤, IL-6, and GM-CSF differentially regulate epithelial cell increase over baseline expression (stimulated MCF/baseline MCF). This growth figure is representative of three separate experiments. Similar results were obtained by using freshly isolated IECs. The effect of cytokines on epithelial cell growth was studied by culturing DLD1 cells in the continuous presence of IL-1␤ (10–100 ng/ml), IL-6 (60 u/ml), GM-CSF (20 ng/ml), or medium only. Cell for the modulation of growth, differentiation, function, or apopto- growth at various time points was quantitated by direct cell counts. sis of epithelial cells as they migrate from the crypt to the surface. As can be seen in Figure 5, IL-1␤ enhanced the cell growth by 3- to 4-fold, whereas GM-CSF and to a lesser extent IL-6 inhibited Regulation of cytokine binding (receptor expression) growth. In a second series of experiments cultures were grown in Several laboratories have recently shown that inflammatory cyto- 96-well flat-bottom plates (5000 cells/well) for 24 h in 1% FCS kine production is elevated in IBD mucosa and that these cytokines with 0.5% BSA. Cells were then treated with various concentra- can modulate expression of various cytokine receptors. Therefore tions of cytokines in the presence of 2% FCS for 72 h, and cell we assessed the regulation of cytokine receptor expression in IECs. proliferation was assessed by a thymidine incorporation assay. Based on the receptor profile of the epithelial cell lines, we chose Again, a stimulatory effect of IL-1␤ and an inhibitory effect of DLD1 cells for these studies since it most closely mimicked nor- GM-CSF was seen in a dose-dependent fashion (data not shown). mal IECs. A total of 106 DLD1 cells/well were cultured in the There was minimal or no effect of IL-6 in this system. presence or absence of LPS (1 ␮g/ml), IFN-␥ (250 U/ml), or PMA (10 ng/ml) for 24 h in 6-well tissue culture plates (Falcon), and GM-CSF treatment of crypt IEC induces ICAM-1 expression receptor expression and density were analyzed by flow cytometry To further evaluate the physiological role of the cytokine receptors and expressed as fold increase over baseline MCF. As shown in on IEC, we examined whether treatment of IECs with GM-CSF The Journal of Immunology 3681

(since GM-CSF production by IECs is increased in active inflammation (71) and GM-CSFR are present in both crypt and surface IEC) could regulate ICAM-1 expression. This integrin has been shown to be of importance in Ag-specific T cell responses and in lymphocyte homing to endothelium with subsequent migration of lymphocytes into inflammatory sites (41). As shown in Figure 6, de-differentiated crypt-like cell lines (DLD1glu and HT29glu) lack constitutive ICAM-1 expression. Treatment with GM-CSF induced ICAM-1 expression in both of these cell lines (significant level of expression in DLD1glu and a very low level of expression in HT29glu). However, the differentiated surface-like (DLD1ino and HT29ino) cell lines constitutively expressed ICAM-1 (low in HT29ino and high in DLD1ino) and this was not further enhanced by GM-CSF treatment in either HT29ino or DLD1ino cells. Sim- ilar results were seen in experiments (n ϭ 2) with freshly isolated crypt epithelial cells although the response in these cells was weaker than in the cell lines and there was no constitutive expression in either cell type (Fig. 6C). Downloaded from

Discussion The results of this study have implications for both normal and inﬂammatory states in the bowel. One the one hand, the expression of proinﬂammatory cytokine receptors by normal IEC may point

toward a role for these cytokines in normal growth and differen- http://www.jimmunol.org/ tiation processes, e.g., crypt to villus maturation. Alterations in either cytokine receptor expression or local cytokine concentrations would induce an imbalance in normal homeostatic mechanisms, potentially leading to epithelial barrier dysfunction. On the other hand, proinﬂammatory cytokine receptor expression renders the epithelium susceptible as a target for the negative effects of these cytokines as well, potentially promoting IEC destruction, ineffective repair, or altered APC function. In this case, the increase in cytokine production within the mucosa would adversely by guest on September 27, 2021 affect the IEC contributing to the clinical expression of disease. Thus the cytokine milieu within the intestinal epithelium has the potential to affect normal physiologic as well as immunoregulatory processes. The possibility that IECs could serve as a target for these proin- ﬂammatory cytokines is supported by similar studies with endothelial cells in vascular processes. Several groups have demon-

strated that IL-1 induces PGE2 secretion by fibroblasts, smooth muscle, and endothelial cells (42), which may have a role in promoting local cytoprotection or in some cases tissue destruction. In addition, IL-1, as well as other cytokines, has been shown to be able to induce the expression of adhesion molecules (ICAM, ELAM-1) on endothelial cells (43), which promotes cell-cell interactions and contributes to the inflammatory response. Proin- flammatory cytokine receptor expression by various epithelial cells has also been reported. IL-1R expression has been documented in the rat intestinal epithelial cell line, IEC-18 (44, 45), whereas Krueger and coworkers (46) have reported mitogenic effect of GM-CSF on basal keratinocytes, suggesting the presence of GM- CSFRs on these cells. This latter observation is of interest since FIGURE 6. Induction of ICAM-1 in IEC lines by GM-CSF. Condition- GM-CSF is a potent stimulator of phagocytosis and class II Ag ally de-differentiated (glu) or fully differentiated (ino) HT29 (A), DLD1 (B) expression in APCs (macrophages) (34). Specific examples of colonic epithelial cell lines and freshly isolated surface and crypt IECs (C) IECs serving as targets for such cytokines include studies by were cultured for 24 h with either medium alone (upper) or with medium Shanahan and coworkers (20), who demonstrated that the combi- containing 20 ng/ml GM-CSF (lower). ICAM-1 expression was detected nation of IFN-␥ and TNF-␣ is cytotoxic for the colonic epithelial by immunofluorescence staining using an anti-ICAM-1 mAb (dotted line) ␥ or an isotype control (solid line). Flow cytometric analysis demonstrates cell line HT-29; Madara et al. (47), who demonstrated that IFN- induction of ICAM-1 in HT29glu and freshly isolated crypt (although very alters the permeability of tight junctions in T84 cells allowing for weak) cells or enhanced expression of ICAM-1 in DLD1glu by GM-CSF neutrophil transmigration; Chang et al. (48), who reported that treatment. A and B are representative of five separate experiments. C is IL-1 induces chloride secretion and short circuit current by chicken representative of two separate experiments. enterocytes when added to the serosal side of chicken intestine 3682 CYTOKINE BINDING BY INTESTINAL EPITHELIAL CELLS mounted in an Ussing chamber (although these effects may have as well as epithelial cells might regulate the normal growth and been mediated through stromal cells); and Molmenti et al. (49), differentiation of IEC. It is of interest to note that many of the who documented acute phase response regulation by IL-1 and IL-6 cytokine knockout mice that develop intestinal inflammation dis- binding on Caco2 cells. Such data support the potential for auto- play histologic pictures characterized by marked crypt cell prolif- crine and paracrine effects of cytokines produced by epithelial cells eration. Thus alterations in cytokine production could alter normal and/or lamina propria mononuclear cells. Given the potential im- physiologic processes (i.e., growth and differentiation) and result plications of such possibilities, we studied the expression and den- in epithelial cell dysfunction. Such a scenario may be true in IBD sity of IL-1␤, IL-6, TNF-␣, and GM-CSF receptors on epithelial where both T cell-derived as well as proinflammatory cytokines cells from normal and IBD tissues. These cytokines were chosen are increased in inflamed tissues. The effects of these cytokines on because they have significant effects on the APC function of mono- IEC phenotype and function may account for some of the clinically cytes and have been reported to be increased in the mucosa of expressed components of these diseases. Since IBD IEC express patients with IBD. The findings reported here demonstrate that similar profiles of cytokine receptors to those seen on normal IEC, receptors for IL-1␤, IL-6, and GM-CSF were comparably ex- the epithelial cell dysfunction seen would relate to the higher con- pressed in both normal and IBD epithelial cells. However, consid- centrations of local cytokine. erable differences did exist in receptor density within differentiated Cytokines might be important for the growth, differentiation, and undifferentiated epithelial cells. TNF␣-R were not detected in and migration of epithelial cells during maturation from crypt cells any of the freshly isolated epithelial cell preparations. In contrast, as well. Alteration in any of these events may cause serious patho- conventional APCs from the same individuals as well as malignant logic consequences. Specific roles of cytokines in each of these epithelial cell lines expressed receptors for all four cytokines in- events require further exploration, but the identification of candi- Downloaded from cluding receptors for TNF-␣. Thus freshly isolated epithelial cells date cytokines by virtue of receptor expression is the first step. differ from professional APCs in their inability to express receptors That cytokines alter epithelial cell function is well described. IL-4 for TNF-␣. This latter finding is of importance as TNF has been enhances IEC-6 cell line proliferation (65) whereas the proinflam- shown to be an important factor in regulating cytokine production matory cytokines IL-1 and GM-CSF either enhance or inhibit by macrophages and as a potentially cytotoxic agent for IEC. The DLD1 cell line growth. IFN-␥, IL-1, and TNF-␣ have been shown absence of TNF␣-R on IEC suggests that these cells may not be up-regulate ICAM-1 expression on the surface of many cells (66, http://www.jimmunol.org/ targets for killing by TNF-␣ in diseases such as IBD where TNF-␣ 67) including intestinal epithelial cells (68). Although the role of levels are high. ICAM-1 expression in intestinal inflammation has been controver- Proinflammatory and prophagocytic cytokines have been known sial (69, 70), our in vitro data suggest that an increased GM-CSF to regulate APC functions as well. However, different APCs dis- concentration in the epithelium can induce ICAM-1 expression by play a distinct range of responses to these cytokines. For example crypt epithelial cells enhancing the chances of adhesion between IFN-␥ up-regulates class II MHC expression on murine macro- these cells and CD4ϩ lamina propria lymphocytes. This finding phages but fails to do so on B cells, although non-class II Ag may represent an important mechanism for the immune responses presentation is enhanced (50). Similar differences have been ob- seen in IBD. served with other cytokines (e.g. IL-4) (51, 52). Thus disparate In conclusion, our data support the concept that IEC can be by guest on September 27, 2021 effects of cytokines on different APCs are well described. Intestinal affected by cytokines in their local environment. Alterations in epithelial cells like conventional APCs express class II MHC mol- cytokine concentrations as a consequence of inflammation may ecules on their cell surface (53, 54) and are capable of taking up, therefore have profound effects on epithelial cell growth, pheno- processing, and presenting Ag to immunocompetent T cells (55– type, and function. Selective regulatory roles of proinflammatory 61). However, several differences exist as well. These include a cytokines on IEC growth and surface molecule expression may lower density of class II molecules on the IEC surface (62), a provide evidence for an unrecognized mechanism responsible for limited cytokine profile (lack of IL-1 production) (15, 21), prefer- inflammation, ulceration, or malignant transformation seen in IBD. ential stimulation of suppressor T cells in mixed cell culture sys- Further elucidation of the functional interactions between distinct tems (52), and a slower rate of Ag processing by IECs (57, 61). cytokines would be of importance in the understanding of cytokine These differences may relate to cytokines in the local environment mediated regulation of mucosal immune responses. and their effect on IEC function. Such correlations have not been addressed in in vivo systems. Acknowledgments It is well described that cytokines or bacterial products can mod- We thank Drs. Adrian Greenstein, Irwin Gelernt, Barry Salky, Joel Bauer, ulate cytokine receptor expression (62, 63). This may in turn in- and Tom Heiman for providing the surgical specimens, and Italas George fluence immune and inflammatory events by promoting or sup- for expert technical assistance with the flow cytometry. We also thank Dr. pressing activation, surface molecule expression, or costimulatory Burton Wice (Washington University, St. Louis, MO) for his generous signals for T cell activation. The regulatory mechanisms governing supply of HT29ino and HT29std cells as well as valuable advice for main- cytokine receptor expression in IECs are not known. Our data sug- tenance and growth of these cells. gest that the regulation of IL-6R is different from that of monocytes, i.e., LPS up-regulates IL-6R expression in DLD1, a colonic References epithelial cell line, whereas it down-regulates expression in mac- 1. Eastgate, J. A., J. A. Symons, N. C. Wood, F. M. Grisham, F. S. Di Giovini, and rophages. This differential regulation of cytokine receptors at a site G. W. Duff. 1988. Correlation of plasma interleukin-1 levels with disease activity where LPS is abundant may reflect a compensatory mechanism. in rheumatoid arthritis. Lancet ii:706. 2. Hirano, T., T. Matsuda, M. Turner, N. Miyasaka, G. Buchan, B. Tang, K. Sato, Still these cytokine receptors may play an important role in intes- M. Shimizu, R. Maini, M. Feldman, and T. Kishimoto. 1988. Excessive produc- tinal homeostasis. Reinecker and Podolsky (64) have recently tion of interleukin 6/B cell stimulatory factor in rheumatoid arthritis, Eur. J. Im- identified the presence of receptors for T cell-derived cytokines on munol. 18:1779. 3. Hossiau, F. A., J.-P. Devogelaer, J. Van Damme, C. N. De Deuxchaisnes, and IEC describing the expression of the common ␥-chain of the IL-2, J. Van Snick. 1988. Interleukin-6 in synovial fluid and serum of patients with IL-4, IL-7, IL-9, and IL-15 receptors in a rat IEC cell line. rheumatoid arthritis and other inflammatory arthritides, Arthritis Rheum. 31: 784. Taken together, one could develop a model where locally pro- 4. Hermann, E., B. Fleischer, W. J. Mayer, T. Poralla, and K. H. Meyer Zum Be- duced cytokines by resident T cells, macrophages, myofibroblasts, schernfeld. 1989. Correlation of synovial fluid interleukin-6 (IL-6) activities with The Journal of Immunology 3683

IgG concentrations in patients with inflammatory joint disease and osteoarthritis. 33. Eckmann, L., H. C. Jung, C. Schurer-Maly, A. Panja, E. Morzycka-Wroblewska, Clin. Exp. Rheumatol. 7:411. and M. F. Kagnoff. 1993. Differential cytokine expression by human intestinal 5. Shinmei, M., K. Masuda, T. Kikuchi, and Y. Shimomura. 1989. Interleukin 1, epithelial cell lines: regulated expression of interleukin-8. Gastroenterology 105: tumor necrosis factor, and interleukin-6 as mediators of cartilage destruction. 1689. Semin. Arthritis Rheum. 18:27. 34. Yang, S.-K., L. Eckmann, A. Panja, and M. F. Kagnoff. 1997. Differential and 6. Guerne. P.-A., B. L. Zuraw, J. H. Vaughan, D. A. Carson, and M. Lotz. 1989. regulated expression of C-X-C, C-C, and C-chemokines by human colon epithe- Synovium as a source of interleukin 6 in vitro. Contribution to local and systemic lial cells. Gastroenterology 113:1214. manifestations of arthritis. J. Clin. Invest. 83:585. 7. Hazenberg, B. P. C., M. H. Van Leeuwen, M. Van Rijswijk, A. C. Stern, and 35. Young, Y. T. 1977. Proof without prejudice: Use of the Kolmogorov-Smirnov E. Vellenga. 1989. Correction of granulocytopenia in Felty’s syndrome by gran- test for the analysis of histograms from flow systems and other sources. J. Histol. ulocyte macrophage colony stimulating factor: simultaneous induction of inter- Cytochem. 25:935. leukin-6 release and flare-up of the arthritis. Blood 74:2769. 36. Collins, H. L., and G. J. Bancroft. 1992. Cytokine enhancement of complement- 8. Grossman, R. M., J. Krueger, D. Yourish, A. GranelliPiperno, D. P. Murphy, dependent phagocytosis by macrophages: synergy of tumor necrosis factor-␣ and L. T. May, T. S. Kupper, P. B. Seghal, and A. B. Gottlieb. 1989. Interleukin-6 is granulocyte-macrophage colony-stimulating factor for phagocytosis of Crypto- expressed in high levels in psoriatic skin and stimulates proliferation of cultured coccus neoformans. Eur. J. Immunol. 22:1447. human keratinocytes. Proc. Natl. Acad. Sci. USA 86:6367. 9. Takamura, T., K. Yoshiaka, K. Murakami, N. Akano, M. Okada, N. Aya, and 37. Panja, A., and L. Mayer. 1995. Cytokine regulated co-stimulatory molecule ex- S. Maki. 1994. Cellular localization of inflammatory cytokines in human glo- pression on human intestinal epithelial cells. Gastroenterology 108:A890 (Ab- merulonephritis. Virchows Arch. 424:459. str.). 10. Abboud, H. E. 1994. Cytokine mediators of renal inflammation. Curr. Opin. 38. Panja, A., T. Zhigadlo, I. Jong, and L. Mayer. 1994. Differential induction of Nephrol. Hypertens. 3:329. MHC determinants in an intestinal epithelial cell line by proinflammatory cyto- 11. Gross, V., T. Andus, H. G. Leser, M. Ruth, and J. Scholmerich. 1991. Inflam- kines. Gastroenterology 106:A750 (Abstr.). matory mediators in chronic inflammatory bowel diseases (IBD) of unknown 39. Mayer, L., and D. Eisenhardt.1990. Lack of induction of suppressor T cells by etiology. Klin. Wochenschr. 69:981. intestinal epithelial cells from patients with inflammatory bowel disease. J. Clin.

12. Breese, E., C. P. Braegger, C. J. Corrigan, J. A. Walker-Smith, and Invest. 86:1255. Downloaded from T. T. MacDonald. 1993. Interleukin-2 and interferon-␥-secreting T cells in normal and diseased human intestinal mucosa. Immunology 78:127. 40. Bauer, J., T. M. Bauer, T. Kalb, T. Taga, G. Lengyel, T. Hirano, 13. Isaacs, K. L., R. B. Sartor, and S. Haskill. 1992. Cytokine mRNA profiles in T. Kishimoto, G. Acs, L. Mayer, and W. Gerok. 1989. Regulation of inter- inflammatory bowel disease mucosa detected by polymerase chain reaction am- leukin 6 receptor expression in human monocytes and monocyte derived mac- plification. Gastroenterology 103:1587. rophages. Comparison with expression in human hepatocytes. J. Exp. Med. 14. Stevens, C., G. Waiz, B. Zanker, C. Singaram, M. Lipman, and T. B. Strom. 170:1537. 1992. Tumor necrosis factor-␣, interleukin-1␤, and interleukin 6 expression in 41. Dustin, M. L., R. Rothlein, A. K. Bhan. C. A. Dinarello, and T. A. Spriger. 1986. inflammatory bowel disease (1BD). Dig. Dis. Sci. 37:818. Induction by IL-1 and interferon, tissue distribution, biochemistry and function of 15. Fiocchi, C. 1989. Lymphokine and intestinal immune response. Role in inflam- a natural adherence molecule (ICAM-1). J. Immunol. 137:245. http://www.jimmunol.org/ matory bowel disease. Immunol. Invest. 18:91. 42. Sartor, R. B., and D. W. Powell. 1991. Mechanisms of diarrhea in intestinal 16. Mahida, Y. R., K. Wu, and D. P. Jewell. 1989. Enhanced production of IL-1␤ by inflammation and hypersensitivity: immune system modulation of intestinal mononuclear cells isolated from mucosa with active ulcerative colitis or Crohn’s transport. In Diarrheal Diseases, M. Field, ed., Elsevier, New York, pp. 75– disease. Gut 30:835. 17. Simon, P. L., G. A. West, D. Rachmilewitch, and C. Fiocchi. 1989. Investigation 114. of interleukin-1 (IL-1) activity in the intestinal mucosa of Crohn’s disease and 43. Mantovani, A., and E. Dejana. 1989. Cytokines as communication signals be- ulcerative colitis patients. Gastroenterology 96:A473 (Abstr.). tween leukocytes and endothelial cells. Immunol. Today 10:370. 18. Satsangi, J., R. A. Wolstencroft, J. Cason, C. C. Ainley, D. C. Dumona, and 44. Sutherland, D. B., G. W. Varilek, and G. A. Neil. 1994. Identification and char- R. P. Thompson. 1989. Interleukin-1 in Crohn’s disease. Clin. Exp. Immunol. acterization of the rat intestinal epithelial cell (IEC-18) interleukin-1 receptor. 67:594. Am. J. Physiol. 266:C1198. 19. Pullman, W. E., S. Elsbury, M. Kobayashi, A. J. Hapel, and W. F. Doe. 1992. Enhanced mucosal cytokine production in inflammatory bowel disease. Gastro- 45. McGee, D. W., S. J. D. Vitkus, and P. Lee. 1996. The effect of cytokine stimu- by guest on September 27, 2021 enterology 102:529. lation on IL-1 receptor mRNA expression by intestinal epithelial cells. Cell. 20. Deem, R. L., F. Shanahan, and S. R. Targan. 1991. Triggered mucosal T cells Immunol. 168:276. release TNF␣ and IFN-␥ which kill human colonic epithelial cells. Clin. Exp. 46. Braunstein, S., G. Kaplan, A. B. Gottlieb, M. Schartz, G. Walsh, R. M. Abalos, Immunol. 83:79. T. T. Fajardo, L. S. Guido, and J. G. Krueger. 1994. GM-CSF activates regen- 21. Panja, A., E. Siden, and L. Mayer. 1995. Synthesis and regulation of accessory/ erative epidermal growth and stimulates keratinocyte proliferation in human skin proinflammatory cytokines by intestinal epithelial cells. J. Clin. Exp. Immunol. in vivo. J. Invest. Dermatol. 103:601. 100:298. 47. Madara, J. C., and J. Staffork. 1989. Interferon ␥ directly affects barrier function 22. Panja, A., R. S. Blumberg, S. P. Bulk, and L. Mayer. 1993. CD1d is involved in of cultured intestinal epithelial monolayers. J. Clin. Invest. 83:724. T cell-intestinal epithelial cell interactions. J. Exp. Med. 178:1115. 23. Panja, A., A. Barone, and L. Mayer. 1994. Stimulation of lamina propria lym- 48. Chang, E. B., M. Musch, and L. Mayer. 1990. Interleukin 1 and 3 stimulate anion phocytes by intestinal epithelial cells: evidence for recognition of nonclassical secretion in chicken enterocyte. Gastroenterology 98:1518. restriction element. J. Exp. Med. 179:943. 49. Molmenti, E. P., T. Ziambaras, and D. H. Perlmutter. 1993. Evidence for an 24. Weiser, M. M. 1973. Intestinal epithelial cell surface membrane glycoprotein acute phase response in human intestinal epithelial cells. J. Biol. Chem. 268: synthesis. I. An indicator of cellular differentiation. J. Biol. Chem. 248:2536. 14116. 25. Gibson, P. R., E. Van de Pol, L. E. Maxwell, A. Gabriel, and W. F. Doe. 1989. 50. Hawrylowicz, C. M., and E. R. Unanue. 1988. Regulation of antigen presen- Isolation of colonic crypt that maintain structural and metabolic viability in vitro. tation. I. IFN-␥ induces antigen presenting properties on B cells. J. Immunol. Gastroenterology 96:283. 141:4083. 26. Harrison, D. D., and H. L. Webstar. 1969. The preparation of isolated intestinal crypt cell. Exp. Cell Res. 55:257. 51. Cao, H., R. G. Wolff, M. S. Meltzer, and R. M. Crawford. 1989. Differential ␥ 27. Reinecker, H. C., E. Y. Loh, D. J. Ringler, A. Mehta, J. L. Rombeau, and regulation of class II MHC determinants of macrophages by IFN- and IL-4. R. P. MacDermott. 1995. Monocyte-chemoattractant protein-1 gene expression in J. Immunol. 143:3524. intestinal epithelial cells and inflammatory bowel disease mucosa. Gastroenter- 52. Te Velde, A. A., R. J. F. Huijbens, J. E. De Vries, and C. G. Frigdor. 1990. IL-4 ology 108:40. decreases Fc ␥ R membrane expression and Fc ␥ R mediated cytotoxic activity 28. Mayer, L., S. M. Fu, and H. G. Kunkel. 1982. Human T cell hybridomas secreting of human monocytes. J. Immunol. 144:3046. factor for IgA specific help, polyclonal B cell activation and B cell proliferation. 53. Mason, D. W., M. Dallman, and A. N. Barclay. 1981. Graft-versus-host disease J. Exp. Med. 156:1860. induces expression of Ia antigen in rat epidermal cells and gut eoithelium. Nature 29. Mayer, L., S. M. Fu, C. Cuningham-Rundles, and H. G. Kunkel. 1984. Poly- 293:150. clonal immunoglobulin secretion in patients with common variable immuno- deficiency using monoclonal B cell differentiation factors. J. Clin. Invest. 54. Wiman, K., B. Curman, U. Forsum, L. Klareskug, U. Malmnas-Tjernlund, 74:2115. L. Rask, L. Tragardh, and P. A. Peterson. 1978. Occurrence of a Ia antigens on 30. Wice, B. M., G. Trugman, M. Pinto, M. Rousset, G. Chevalier, E. Dussaulx, tissues of nonlymphoid origin. Nature 276:711. B. Lacroix, and A. Zweibaum. 1985. The intercellular accumulation of UDP-N- 55. Natali, P. G., C. DeMartino, V. Quaranta, M. R. Nicotra, F. Frezza, acetylhexosamines is concomitant with the inability of human colon cancer cells M. A. Pelligrino, and S. Ferrone. 1981. Expression of Ia-like antigens in normal to differentiate. J. Biol. Chem. 260:139. human nonlymphoid tissues. Transplantation 31:75. 31. Wice, B. M., and J. I. Gordon. 1992. A strategy for isolation of cDNAs encoding 56. Selby, W. S., L. W. Poulter, S. Hobbs, and D. P. Jewell. 1981. Expression of HLA proteins affecting human intestinal epithelial cell growth and differentiation: char- acterization of a novel gut-specific N-myristoylated annexin. J. Cell Biol. 116: DR antigens by colonic epithelium in inflammatory bowel disease. Clin. Exp. 405. Immunol. 44:453. 32. Mayer, L., E. Siden, S. Becker, and D. Eisenhardt. 1990. Antigen handling in the 57. Bland, P. W., and L. G. Warren. 1986. Antigen presentation by epithelial cells of intestine by normal enterocytes. In Advances in Mucosal Immunology. Mac- rat small intestine. I. Kinetics: antigen specificity and blocking by anti-Ia antisera. donald et al., eds. Academic, New York, 23–28. Immunology 58:1. 3684 CYTOKINE BINDING BY INTESTINAL EPITHELIAL CELLS

58. Mayer, L., and R. Shlien. 1987. Evidence for function of a Ia molecule on gut 65. McGee, D. W., and S. J. Vitkus. 1996. IL-4 enhances IEC-6 intestinal epithelial epithelial cells in man. J. Exp. Med. 166:1471. cell proliferation yet has no effect on IL-6 secretion. Exp. Immunol. Clin. 105: 59. Kaiserlian, D., K. Vidal, and J. P. Revillar. 1989. Murine enterocytes can present 274. soluble antigen to specific class II restricted CD4ϩ T cells. Eur. J. Immunol. 66. Rothlein, R., M. Czajkowski, M. M. O’Neill, S. D. Marlin, E. Mainolfi, and 19:1513. V. J. Merluzzi. 1988. Induction of intercellular adhesion molecule 1 on primary 60. Bland, P. W., and C. V. Whiting. 1989. Antigen processing by isolated rat in- and continuous cell lines by pro-inflammatory cytokines: regulation by pharma- testinal villus enterocytes. Immunology 68:497. cologic agents and neutralizing antibodies. J. Immunol. 141:1665. 61. Mayer, L., L. P. So., K. Pelton-Henrion, and G. Small. Antigen trafficking in the 67. Piela, T. H., and J. H. Korn. 1988. Lymphocyte-fibroblast adhesion induced by intestine. 1996. Ann. N.Y. Acad. Sci. 778:28. interferon-␥. Cell. Immunol. 114:149. 62. Raqib, R., A. A. Lindberg, L. Bjork, P. K. Bardhan, B. Wretlind, U. Andersson, 68. Kaiserlian, D., D. Rigal, J. Abello, and J.-P. Revillard. 1991. Expression, function and J. Andersson. 1995. Down-regulation of ␥ interferon, tumor necrosis factor and regulation of the intercellular adhesion molecule-1 (ICAM-1) on human in- type-1, interleukin-1 (IL-1) type-1, IL-3, IL-4, and transforming growth factor b testinal epithelial cell lines. Eur. J. Immunol. 21:2415. type 1 receptors at the local site during acute phase of Shigella infection. Infect. 69. Bloom, S., D. Simmons, and D. P. Jewell. 1995. Adhesion molecules intercellular Immun. 63:3079. adhesion molecule-1 (ICAM-1), ICAM-3 and B-7 are not expressed by epithe- 63. Huang, G. T., L. Eckmann, T. C. Savidge, and M. F. Kagnoff. 1996. Infection of lium in normal or inflamed colon. Clin. Exp. Immunol. 101:157. human intestinal epithelial cells with invasive bacteria upregulates apical inter- 70. Norton, J., J. P. Sloane, N. Al-Saffar, and D. O. Haskard. 1992. Expression of cellular adhesion molecule-1 (ICAM-1) expression and neutrophil adhesion. adhesion molecules in human intestinal graft-versus-host disease. Clin. Exp. Im- J. Clin. Invest. 98:572. munol. 87:231. 64. Reinecker, H. C., and D. K. Podolsky. 1995. Human intestinal epithelial cells 71. Panja, A., L. Eckmann, M.-F. Kagnoff, and L. Mayer. 1994. Increased production express functional cytokine receptors sharing the common gamma c chain of the of IL-8 and GM-CSF by intestinal epithelial cells in inflammatory bowel disease. interleukin 2 receptor. Proc. Natl. Acad. Sci. USA 92: 8353. Gastroenterology 160: A750 (Abstr.) Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021