Interleukin-4 and interleukin-13 differentially regulate epithelial chloride secretion.

Intestinal epithelia are in intimate contact with subepithelial and intraepithelial lymphocytes. When stimulated, mucosal lymphocytes generate cytokines that act locally and influence functional aspects of many cell types. We have previously defined functional epithelial receptors for interferon-gamma, interleukin (IL)-4, and a recently described IL-4-like cytokine IL-13. In this study, we examine the ion transport properties of T84 cells, a crypt-like epithelial cell line, following exposure to IL-4 and IL-13. Basolateral exposure of epithelial monolayers to both IL-4 and IL-13 attenuated epithelial barrier function and increased paracellular flux of a dextran marker by greater than 65% in a dose- and time-dependent fashion. Stimulated Cl- secretion, as measured by epithelial short circuit current, however, was diminished only by IL-4 and not IL-13, demonstrating cytokine specificity in this epithelial function. Decreased Cl- secretion following IL-4 exposure was associated with diminished Cl- channel activity and IL-4 pretreatment of epithelia decreased expression of the cystic fibrosis transmembrane regulator. Finally, stimulated fluid transport across cultured epithelia was diminished following exposure to IL-4, but not IL-13. These results indicate that while post-receptor signaling events induced by IL-13 and IL-4 may be similar, end point function is cytokine-specific.

Intestinal epithelia are in intimate contact with subepithelial and intraepithelial lymphocytes. When stimulated, mucosal lymphocytes generate cytokines that act locally and influence functional aspects of many cell types. We have previously defined functional epithelial receptors for interferon-␥, interleukin (IL)-4, and a recently described IL-4-like cytokine IL-13. In this study, we examine the ion transport properties of T84 cells, a crypt-like epithelial cell line, following exposure to IL-4 and IL-13. Basolateral exposure of epithelial monolayers to both IL-4 and IL-13 attenuated epithelial barrier function and increased paracellular flux of a dextran marker by greater than 65% in a dose-and time-dependent fashion. Stimulated Cl ؊ secretion, as measured by epithelial short circuit current, however, was diminished only by IL-4 and not IL-13, demonstrating cytokine specificity in this epithelial function. Decreased Cl ؊ secretion following IL-4 exposure was associated with diminished Cl ؊ channel activity and IL-4 pretreatment of epithelia decreased expression of the cystic fibrosis transmembrane regulator. Finally, stimulated fluid transport across cultured epithelia was diminished following exposure to IL-4, but not IL-13. These results indicate that while post-receptor signaling events induced by IL-13 and IL-4 may be similar, end point function is cytokine-specific.
Epithelial cells of the intestine are uniquely positioned to serve as a direct line of communication between the immune system and the external environment. Intestinal epithelial surfaces are continuously exposed on the luminal side to foreign antigens and a myriad of microorganisms, while at the same time, intimately associated with the immune system via subepithelial lymphoid tissue. In the basal state, and especially during disease flares, lymphocyte-derived cytokines are readily detectable in the mucosa (1). For most cytokines, target cell function is pleiotropic, and individual cytokines may act singly, additively, antagonistically, or synergistically (2). Previous studies have demonstrated that intestinal epithelia bear functional receptors for a number of diverse cytokines, including but not limited to IFN-␥ 1 (3)(4)(5), IL-2 (6), IL-4 (7), IL-13 (8), tumor necrosis factor-␣ (9), transforming growth factor-␤ (10,11), and hepatocyte growth factor (12).
Interleukin-13 is a recently described cytokine produced by activated T lymphocytes (13). IL-13 and IL-4 share a number of biological responses when exposed to monocytes, macrophages, and B cells (14,15). We have recently demonstrated that IL-13, like IL-4, increases paracellular permeability of cultured intestinal epithelial monolayers (8). In addition, human intestinal epithelial cells posess IL-4 receptors (16), and while IL-13 does not bind to the IL-4 receptor, others have proposed overlapping signal transduction pathways (14,15). The genes for IL-13 and IL-4 are closely linked in both humans and mice, and these cytokines are approximately 30% homologous (17) (approximately the same degree of homology as IL-1␣ and IL-␤ (18)).
In this study, we demonstrate strict cytokine specificity of IL-4 and IL-13 actions on epithelial ion transport. While both IL-13 and IL-4 specifically diminish epithelial barrier function, only IL-4 attenuates epithelial electrogenic Cl Ϫ secretion (the transport event underlying hydration of mucosal surfaces), indicating cytokine specificity. Such decreases in Cl Ϫ secretion by IL-4 are attributable to diminished Cl Ϫ channel activity and associated with IL-4-mediated down-regulation of CFTR protein expression. These results suggest that, while the intracellular signaling events induced by IL-13 and IL-4 reportedly exhibit many similarities, the resulting end point phenotype of epithelial function demonstrates clear specificities between these cytokines. Such observations indicate that differences in IL-13 and IL-4 signal transduction pathways exist.

MATERIALS AND METHODS
Cell Culture-Confluent monolayers of the human intestinal cell line T84 were grown on collagen-coated permeable supports and maintained until steady-state transepithelial resistance was achieved. To measure agonist-stimulated short circuit currents, transepithelial potentials, and resistance, a commercially available voltage clamp (Iowa Dual Voltage Clamps, Bioengineering, University of Iowa) interfaced with an equilibrated pair of calomel electrodes and a pair of Ag-AgCl electrodes was utilized, as described in detail elsewhere (19). Using these values and Ohm's law (V ϭ IR), tissue resistance and transepithelial current were calculated. Fluid resistance within the system accounts for less than 5% of total transepithelial resistance. Hanks' balanced salt solution was used in both apical and basolateral baths during all experiments, unless otherwise noted. Assessment of paracellular flux of fluorescein isothiocyanate-labeled dextran (molecular mass: 10 kDa, Molecular Probes, Eugene, OR) was performed exactly as described before (8). All experiments were performed in a 37°C room to ensure that epithelial monolayers, solutions, plastic ware, etc. were maintained at uniform 37°C temperature. Tissue culture supplies were obtained from Life Technologies, Inc. and Costar (Cambridge, MA). Recombinant human IL-13 was a kind gift from Dr. Satish Menon at DNAX Research Institute (Palo Alto, CA), recombinant human IL-4 and was purchased from R&D Systems (Minneapolis, MN), and recombinant human IFN-␥ was a kind gift from Genentech (South San Francisco, CA). Other reagents were obtained from Sigma or Calbiochem.
T84 epithelial monolayers were exposed to cytokines at indicated concentrations for indicated periods of time. Unless otherwise noted, monolayers were exposed to cytokine on both apical and basolateral surfaces. All incubations included controls consisting of media alone.
Lactate Dehydrogenase Release Assay-T84 intestinal epithelial monolayers were grown on 96-well plates with and without addition of cytokine for 48 h. Cytotoxicity as determined by lactate dehydrogenase release was measured color metrically (Cytotox 96™ non-radioactive cytotoxicity assay, Promega, Madison, WI) according to the manufacturer's protocol.
Fluid Transport Assay-The methods for measuring transmonolayer fluid movement were adapted from those described by Smith and Welsh (20). T84 monolayers were exposed to indicated cytokines for 24 h. To measure net fluid movement, the apical solution was replaced with 100 l of media and layered with 1 ml of warm, sterile mineral oil to minimize evaporation. Others have shown that it is necessary to begin the experiment with a small amount of fluid (i.e. 100 l) so as to recover enough for analysis. Coating monolayers with mineral oil does not alter transport properties (20). In some monolayers, the cAMP agonists forskolin (50 M) and 3-isobutyl-1-methylxanthine (IBMX, 100 M) were added to the basolateral solution to promote fluid movement. After an additional 24 h (total exposure to cytokine, 48 h), the apical solution was collected, spun at high speed in an Eppendorf centrifuge, and the recovered fluid was weighed on a balance. 125 I Efflux Assays-Cl Ϫ channel activity was monitored using 125 I efflux assays with T84 cells grown on 1-cm 2 permeable supports, as described before (4,21). Rate constants of efflux were calculated as [ln(R 2 ) Ϫ ln(R 1 )]/(t 2 Ϫ t 1 ), where R x is the percent of radioactivity remaining monolayer-associated at time t x , as reported elsewhere (21).
CFTR Immunoprecipitation-T84 cells were grown on 5-cm 2 permeable supports in the presence or absence (media alone) of the indicated cytokines. Immunoprecipitation was performed using a mouse antihuman CFTR monoclonal antibody (Genzyme Corp., Cambridge, MA) as described previously (22). CFTR was in vitro phosphorylated by incubating immunoprecipitates with protein kinase A (20 ng/ml, Sigma) and 10 Ci [␥-32 P]ATP in 50 l of kinase buffer (50 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , and 100 g/ml bovine serum albumin) at 30°C for 60 min as described previously (22). Bands at ϳ165 kDa consistent with previously published CFTR from resulting autoradiograms were analyzed from scanned images using National Institutes of Health Image software (Bethesda, MD).
Statistical Analysis-Transepithelial resistance's were compared by Student's t test. Time course data and dose responses were compared by analysis of variance (ANOVA) for significance. All results are presented as the mean and S.E. of n experiments.

Functionally Defined IL-13 Receptors on T84
Intestinal Epithelial Cells-As we have shown previously, intestinal epithelial exposure to IL-4 (7,8) or IL-13 (8) exhibits a marked attenuation in transepithelial resistance and concomitant increases in paracellular permeability of labeled tracers. Thus, transepithelial resistance was used as a screening assay for cytokine bioactivity in this system. As shown in Fig. 1A, addition of IL-4 or IL-13 to confluent T84 cells for 48 h resulted in a dose-dependent decrease in epithelial barrier function (both p Ͻ 0.001 compared with media alone control, ANOVA). Efficacy was comparable between IL-4 and IL-13 (EC 50 : ϳ1.5 ng/ml for both). Predictably, and as we have shown before (8), paracellular flux of 10 kDa (Stokes radius: 23 Å) fluorescein isothiocyanate-dextran was increased for both IL-13 and IL-4 compared with media alone controls (flux of 3.7 Ϯ 0.81, 11.9 Ϯ 1.54, 10.9 Ϯ 2.01 pM/cm 2 ⅐h for media controls, IL-13-and IL-4-exposed monolayers (10 ng/ml, 48 h), respectively (p Ͻ 0.01 for IL-13 and IL-4 compared with control). Time course experiments (range 12-72 h, 10 ng/ml) and subsequent assessment of barrier function resulted in no distinguishable difference between IL-4 and IL-13 (t1 ⁄2 ϳ30 h for both IL-13 and IL-4, data not shown). Such results are not explained by IL-4 toxicity (7) or by IL-13 toxicity (spontaneous lactate dehydrogenase release measured over a 48-h period was equivalent for control monolayers and monolayers exposed to IL-13, data not shown). IL-13, like IL-4 (7), appeared to signal via a basolaterally restricted receptor. Using monolayers grown on permeable supports, basolateral pre-exposure to IL-13 (10 ng/ml, 48 h) elicited a 72 Ϯ 8.5% fall in transepithelial resistance (p Ͻ 0.01 compared with media alone), while apical exposure to this cytokine failed to significantly influence resistance (6 Ϯ 3.1% decrease compared with media alone, p Ͼ 0.05). Finally, exposure of T84 cells to a combination of IL-13 and IL-4 at approximate EC 50 (final concentration alone or in combination 2.5 ng/ml, 48 h) did not result in additive effects on barrier function (45 Ϯ 9.7, 47 Ϯ 6.5, and 53 Ϯ 7.2% fall in transepithelial resistance compared with media alone controls for IL-4 alone, IL-13 alone, and IL-4/IL-13 combined, respectively). These data indicate that IL-4 and IL-13 share many properties with regard to regulation Selective Down-regulation of Electrogenic Cl Ϫ Secretion by IL-4 and Not IL-13-As shown in Fig. 1B, and as we have demonstrated previously (7), IL-4 attenuates stimulated electrogenic Cl Ϫ secretion (measured as short circuit current, I sc ), the ion transport event responsible for mucosal hydration (23). However, a clear discrepancy between IL-4 and IL-13 was evident at this level of epithelial function. Indeed, while IL-4 dose-dependently attenuated forskolin-stimulated I sc (p Ͻ 0.001, ANOVA), no significant reduction was apparent at equivalent concentrations of IL-13 (p ϭ not significant, ANOVA). As shown in Fig. 2A, a significant shift in the dose response to stimulation by the cAMP agonist forskolin (range 0.01-10 M) was evident following epithelial exposure to IL-4 (10 ng/ml, 48 h, two-factor ANOVA, p Ͻ 0.025), but not to IL-13 (10 ng/ml, 48 h, two-factor ANOVA, p ϭ not significant). Baseline transepithelial resistances of these monolayers revealed an IL-13 effect (615 Ϯ 119 ohm-cm 2 ) as well as an IL-4 effect (559 Ϯ 97 ohm-cm 2 ) compared with no cytokine controls (1543 Ϯ 219 ohm-cm 2 ), indicating cytokine selectivity in cAMP-stimulated Cl Ϫ secretion. Similarly, Ca 2ϩ -mediated Cl Ϫ secretion (using carbachol, Fig. 2B) was attenuated by IL-4 (two-factor ANOVA, p Ͻ 0.01 compared with no cytokine controls), but no demonstrable decrease was apparent using IL-13 (two-factor ANOVA, p ϭ not significant compared with no cytokine controls). Last, the ED 50 for IL-4-mediated attenuation of Cl Ϫ secretion was comparable with that for IL-4 influences on barrier function (2.5 versus 1.5 ng/ml, respectively), indicating that these two functional consequences of IL-4 exposure were likely consequent to the same ligand-receptor binding event.
This IL-4-induced decrease in Cl Ϫ secretion cannot be competed by excess IL-13. Experiments were performed in which 10-and 100-fold excess concentrations of IL-13 (50 and 500 ng/ml, respectively, 48 h) were added to IL-4 treated (5 ng/ml, 48 h) monolayers and assayed for forskolin-stimulated (1 M final concentration) Cl Ϫ secretion. Indeed compared with no cytokine control, addition of IL-4 alone decreased Cl Ϫ secretion by 65 Ϯ 10% (I sc of 58 Ϯ 8.6 and 20 Ϯ 6.2 for control and IL-4 treated, respectively, p Ͻ 0.01), while no competitive effect was observed with addition of 50 ng/ml IL-13 (I sc of 22 Ϯ 7.1, p ϭ not significant compared with IL-4 alone) or 500 ng/ml IL-13 (I sc of 25 Ϯ 6.6, p ϭ not significant compared with IL-4 alone).
IL-4 Down-regulates Cl Ϫ Channel Activity-Attenuation of Cl Ϫ secretion, such as that shown above for IL-4 exposure, but not for IL-13, might relate to a variety of cytokine-elicited influences in cell function, the most direct of which would be down-regulated activity/expression of the Cl Ϫ channel itself. Thus, we next examined whether differences between IL-4 and IL-13 were apparent at the epithelial Cl Ϫ channel. The rate constant of 125 I efflux was used to estimate forskolin-stimulated activation of Cl Ϫ channel on permeable support grown monolayers of T84 epithelial cells with and without exposure to cytokine. This method has been validated as a reliable measurement of Cl Ϫ efflux from T84 epithelial cells (21). As shown in Fig. 3, pre-exposure of epithelial monolayers to IL-4 significantly decreased forskolin-stimulated 125 I efflux from T84 monolayers (two-factor ANOVA, p Ͻ 0.01 compared with media alone), while no significant effects were observed in response to IL-13 (two-factor ANOVA, p ϭ not significant compared with media alone). As a positive control for attenuated Cl Ϫ channel activity (4), monolayers were exposed to IFN-␥ (1000 units/ml, 48 h). Unstimulated efflux rate constants for 125 I were not different for monolayers exposed to IL-4, IL-13, or IFN-␥ (data not shown) compared with untreated controls (Fig. 3). These data demonstrate that IL-4, like IFN-␥ (4), diminishes stimulated Cl Ϫ channel activity, while IL-13 does not.
Influence of IL-13/IL-4 on Epithelial Expression of CFTR-T84 cells, as do native intestinal Cl Ϫ secreting cells, express the CFTR, a protein that serves as both a Cl Ϫ channel and a Cl Ϫ channel regulator (24). Thus, we determined whether decreased CFTR expression contributed to attenuated Cl Ϫ secretion and Cl Ϫ channel activity in T84 cells following exposure to IL-13 or IL-4. CFTR was detected in T84 cell lysates by adding [␥-32 P]ATP and protein kinase A to immunoprecipitates formed using monoclonal antibodies raised against CFTR (Fig. 4A). Pre-exposure of T84 cell monolayers to IL-4 resulted in decreased CFTR protein levels, similar to our positive control IFN-␥ (Fig. 4, A and B), as others have shown (25). Consistent with our results above, no differences were observed between IL-13 and media only controls at the level of CFTR. Impact of IL-4/IL-13 on Fluid Transport Intestinal Epithelia-The key end point of epithelial electrogenic Cl Ϫ secretion is net transepithelial secretion of isosmotic fluid. Thus, we examined whether IL-4 induced decreased fluid transport across epithelial monolayers and, as suggested by the above results, whether differences between IL-4 and IL-13 exist at this level. As shown in Fig. 5, base-line (i.e. unstimulated) fluid transport was negligible and did not differ between cells exposed to media, IL-4, IL-13, or IFN-␥. However, in the presence of the cAMP agonists forskolin (50 M) and IBMX (100 M) for 24 h, a marked difference between IL-4 and IL-13 was observed (Fig. 5). Indeed, epithelial exposure to IL-4 significantly attenuated fluid movement across epithelia (p Ͻ 0.01 compared with media only control) in a manner similar to IFN-␥ (p Ͻ 0.001 compared with media only). Monolayers exposed to IL-13, however, were not different than media only controls (p ϭ not significant), indicating that IL-4 and IL-13 also differentially regulate the net result of electrogenic Cl Ϫ secretion: isosmotic fluid secretion.

DISCUSSION
Soluble factors produced by lymphocytes adjacent to the epithelium modulate epithelial function. We and others have demonstrated that a diverse array of cytokines regulate the function of model intestinal epithelia (3,5,26,27). Here we use polarized model intestinal epithelia, which demonstrate readily detectable biophysical responses as a means to probe potential differences in IL-4-and IL-13-mediated signal transduction pathways. Our results indicate that both IL-4 and IL-13, which share many biological properties in a variety of systems (14,15), attenuate epithelial barrier function; IL-4 selectively down-regulates electrogenic Cl Ϫ secretion, indicating distinct cytokine specificity. In addition, these studies reveal that IL-4 attenuation of Cl Ϫ secretion, which we have observed previously (7), occurs through regulation of expression of the apical Cl Ϫ efflux channel and results in diminished epithelial fluid secretion.
IL-13 is a recently described cytokine that is secreted by activated T lymphocytes and shares many properties with IL-4 (14, 15). To date, functional IL-13 receptors have been demon- strated on monocyte/macrophages, B cells, NK cells, mast cells, fibroblasts, renal epithelia, and intestinal epithelia (8,14,15,28,29). In general, mucosal lymphocytes are subdivided into two populations, intraepithelial lymphocytes and lamina propria lymphocytes, each defined by their spatial proximity to the epithelia and each with characteristic surface markers and cytokine production patterns. The primary source of IL-13 and IL-4 appears to be the lamina propria lymphocyte population of CD4 ϩ , Th-0, Th-1, or Th-2 T-cells, although other cell types may also produce these cytokines (14,15). The present study demonstrates that in a time-and dose-dependent manner, exogenous exposure of intestinal epithelial monolayers to IL-13 diminishes barrier function (measured as transepithelial resistance and paracellular flux), an essential and primary role of mucosal epithelial cells (30). The functional IL-13 receptor is localized to the basolateral surface on T84 cells. This feature of diminished barrier following ligation of the IL-13 receptor on T84 cells resembles that of IL-4 (7) and IFN-␥ (4,5). At present, a commonality between these cytokines and modulation of epithelial barrier has not been identified.
The structure of the IL-13 receptor has remained elusive. While IL-13 does not bind to the IL-4 receptor, it has been shown that IL-13 can inhibit the binding of labeled IL-4 to cells that are responsive to both cytokines (31), and thus it has been proposed that the IL-13 receptor shares a component of the IL-4 receptor important in signal transduction. Of note on this accord, Reinecker and Podolsky (16) recently demonstrated that intestinal epithelia possess transcripts for and signal transduction capacity (tyrosine phosphorylation) for a number of cytokines that share the IL-2 receptor ␥c chain, including the IL-4 receptor. Alternatively, the IL-4 and IL-13 receptors may share redundant signal transduction pathways. We demonstrate that addition of IL-13 in combination with IL-4 at approximate ED 50 does not result in additive or synergistic effects on diminished transepithelial resistance. Moreover, addition of even 100-fold excess exogenous IL-13 failed to inhibit IL-4elicited diminution of Cl Ϫ secretion (see "Results"), indicating that it is unlikely IL-4 and IL-13 bind to the same receptor, or if this is the case, receptor affinities favor binding of IL-4 to a great extent. Finally, it is worthwhile to note that we have not examined whether IL-4 and/or IL-13 might liberate epithelialderived cytokines that could act through autocrine mechanisms to regulate such functions as barrier and ion transport. Further work on IL-13 receptor binding should reveal important information regarding the mechanism by which this cytokine regulates epithelial function.
While a number of studies have demonstrated that IL-4 and IL-13 share many biological properties, the present studies clearly demonstrate differences between IL-4 and IL-13 at the level of epithelial electrogenic Cl Ϫ secretion, the transport event responsible for mucosal hydration (23). Here we show that IL-4 selectively down-regulates Cl Ϫ transport. Such downregulation was evident in response to both cAMP (forskolin) and Ca 2ϩ (carbachol)-stimulated Cl Ϫ secretion, was observed at the level of apical Cl Ϫ channel activity, and resulted in decreased transmonolayer fluid movement. Additionally, diminished levels of the CFTR were associated with IL-4 preexposure. Whether such decreases in CFTR (ϳ65%) result in our observed decrease in Cl Ϫ secretion are not known at the present time. Another possible explanation for diminished Cl Ϫ secretion in response to IL-4 is attenuation of basal K ϩ efflux pathways. Current methods available to study K ϩ channel activity lack the sensitivity necessary to detect decrements that might be induced by IL-4, and thus, we have not directly examined basal K ϩ channel activity in T84 cells.
Such findings with IL-4 are remarkably similar to previous investigations using IFN-␥, including diminished epithelial barrier (3)(4)(5), decreased electrogenic Cl Ϫ secretion in response to a variety of agonists (4), decreased Cl Ϫ channel activity (4), and decreased CFTR protein (25). We have defined such IFN-␥-elicited changes as a "cell surface phenotype switch," in which epithelia exposed to such stimuli lose many epithelial characteristics and assume other more immune accessory celllike features (increased major histocompatibility class I and II expression, regulated neutrophil trafficking, etc.) (4). Such results suggest that a T-cell-derived cytokines specifically modulate the extent of electrolyte and water loss across the intestinal barrier, a hypothesis that has been proposed previously for IFN-␥ (26).
In summary, these data show that a human intestinal crypt epithelial model has diverse functional characteristics modulated by IL-4 and IL-13 exposure to the basolateral surface. Furthermore, end point functions elicited by IL-4 and IL-13 on intestinal epithelial biology are distinct. Such observations strengthen the hypothesis that local mucosal cytokine profiles will contribute significantly to the regulation of epithelial barrier function and ion transport characteristics.