Interleukins 4 and 13 increase intestinal epithelial permeability by a phosphatidylinositol 3-kinase pathway. Lack of evidence for STAT 6 involvement.

Interleukins 4 and 13 can affect their target cells by activation of signal transducer and activator of transcription 6 (STAT 6) or phosphatidylinositol 3-kinase (PI3K). We examined the signal transduction events involved in IL-4 and IL-13 regulation of epithelial paracellular permeability using T84 cells, a model human colonic epithelium. T84 cells treated with IL-4 or IL-13 displayed virtually identical dose- and time-dependent STAT 6 activation as assessed by electrophoretic mobility shift assay (EMSA) and decreases in transepithelial resistance (TER). STAT 6 DNA binding activity was maximal in nuclear extracts 30 min after exposure to IL-4 or IL-13, and TER was maximally reduced by 24 h post-treatment. Pretreatment of epithelia with transcription factor decoys (phosphorothioated DNA oligonucleotides containing the STAT 6 binding site) dramatically reduced STAT 6 activation as detected by EMSA, but did not attenuate the TER reduction by IL-4 or IL-13. In contrast, although the PI3K inhibitors wortmannin and LY294002 did not affect IL-4 or IL-13 STAT 6 activation, they significantly inhibited the ability of either cytokine to lower TER. Thus, we provide evidence for PI3K as the major proximal signaling event in IL-4 and IL-13 regulation of TER and speculate that pharmacological targeting of enterocytic PI3K activity may represent a means to manipulate epithelial permeability.

Interleukins 4 and 13 can affect their target cells by activation of signal transducer and activator of transcription 6 (STAT 6) or phosphatidylinositol 3-kinase (PI3K). We examined the signal transduction events involved in IL-4 and IL-13 regulation of epithelial paracellular permeability using T84 cells, a model human colonic epithelium. T84 cells treated with IL-4 or IL-13 displayed virtually identical dose-and time-dependent STAT 6 activation as assessed by electrophoretic mobility shift assay (EMSA) and decreases in transepithelial resistance (TER). STAT 6 DNA binding activity was maximal in nuclear extracts 30 min after exposure to IL-4 or IL-13, and TER was maximally reduced by 24 h posttreatment. Pretreatment of epithelia with transcription factor decoys (phosphorothioated DNA oligonucleotides containing the STAT 6 binding site) dramatically reduced STAT 6 activation as detected by EMSA, but did not attenuate the TER reduction by IL-4 or IL-13. In contrast, although the PI3K inhibitors wortmannin and LY294002 did not affect IL-4 or IL-13 STAT 6 activation, they significantly inhibited the ability of either cytokine to lower TER. Thus, we provide evidence for PI3K as the major proximal signaling event in IL-4 and IL-13 regulation of TER and speculate that pharmacological targeting of enterocytic PI3K activity may represent a means to manipulate epithelial permeability.
The intestinal epithelium is not a passive entity but rather is a dynamic tissue which, along with other functions, serves as a regulated barrier between the external and internal environments. Loss of epithelial barrier function has been reported in many enteropathies, including the inflammatory bowel diseases: Crohn's disease and ulcerative colitis (1,2). Whether this loss of barrier function is associated with the etiology of disease or acts as a potentiating factor is a point of debate. It has been shown that Crohn's disease is often accompanied by increased local, or circulating, levels of interferon-␥ (IFN␥) 1 and tumor necrosis factor-␣ and that IL-4 can be increased in cohorts of patients with ulcerative colitis or Crohn's disease (3,4). Furthermore, several in vitro studies have demonstrated that cytokines, including IFN␥, IL-4, and IL-13, can increase the permeability characteristics of intestinal epithelial monolayers (5)(6)(7).
Several cytokines exert their biological effects via the signal transducer and activator of transcription (STAT) system (8). Briefly, cytokine ligation of receptor causes receptor dimerization and phosphorylation by associated Janus kinase proteins. Subsequently, latent cytoplasmic STAT protein monomers bind to the receptor where they are phosphorylated, dimerize, and then translocate to the nucleus to modulate gene transcription. Studies with immune cells have illustrated that IL-4 and IL-13 specifically activate STAT 6 (8). Additional investigations showed that IL-4 and IL-13 also signal along the insulin receptor substrate-1/2, phosphatidylinositol 3-kinase (PI3K) pathway; for example, to support proliferation of B cells (9). Although IL-4 and IL-13 signal transduction events are being precisely characterized in immune cells considerably less information is available with specific reference to epithelial cells.
Given the ability of IL-4 and IL-13 to directly affect epithelial permeability, the present study was designed to: (a) define STAT 6 kinetics in model intestinal epithelia, primarily the human transformed T84 cell line, and (b) to use a pharmacological approach to explore the intracellular signal transduction pathway events that govern the IL-4-and IL-13-induced increase in epithelial permeability. The data presented here support the conclusions that IL-4 and IL-13 mobilize STAT 6 in similar dose-and time-dependent manners and activate the PI3K signal transduction pathway in T84 cells. Also, our in vitro observations suggest that the IL-4-and IL-13-induced increase in T84 epithelial permeability occur, at least in part, via a PI3K-dependent mechanism.

EXPERIMENTAL PROCEDURES
Cell Culture-The transformed human colonic epithelial cell line T84 was cultured in a 1:1 mixture of Dulbecco's modified eagle medium and Ham's F-12 medium supplemented with 10% (v/v) fetal calf serum, 2% (v/v) penicillin-streptomycin, and 1.5% (v/v) HEPES (all obtained from Life Technologies) (8,10). T84 cells were seeded onto 6-cm diameter Petri dishes (3 ϫ 10 6 cells) or onto 1-cm 2 semipermeable culture mediatreated filters (10 6 cells) (Costar, Corning Inc., Cornell, NY) for subsequent nuclear protein extraction or assessment of paracellular permeability, respectively. Additional experiments used the transformed human colonic epithelial cell line HT-29 or the transformed human lung epithelial cell line A549, for analysis of STAT 6 activation only.
Cytokine Stimulation and Inhibitor Studies-T84 monolayers were treated with IL-4 or IL-13 (Sigma) at doses of 1, 10, or 100 ng/ml for 30 min, or 100 ng/ml for 15 min, 30 min, 1 h, 4 h, 14 h, and 24 h to determine the dose-and time-dependent kinetics of STAT 6 activation. Time-matched controls were cultured in media only.
Transcription factor decoys (TFDs) were used to block IL-4 or IL-13 activation of STAT 6 (i.e. binding to nuclear DNA) (11,12). TFDs are double-stranded DNA oligonucleotides consisting of the same sequence as the STAT 6 probe (or the mutant probe) used for EMSA (see below) that have been phosphorothioated to increase their entrance into cells and render them less susceptible to degradation (12). Briefly, TFDs (0.2 M) were added to semiconfluent T84 cells in OPTI-MEM I-reduced serum medium (Life Technologies) as described (13). 24 h later the cells were rinsed in phosphate-buffered saline (3ϫ) to remove non-incorporated TFDs and received standard medium (10% fetal calf serum) for 1 h prior to cytokine stimulation. Subsequently, either a nuclear protein extraction was performed or physiology assessed. (Comparative pilot studies used Lipofectin (Life Technologies) to deliver the TFDs. This approach was no more effective than adding the TFDs alone.) Also, the effects of the specific PI3K inhibitors wortmannin (1 M) and LY294002 (20 M) (both Sigma) on IL-4-or IL-13-induced mobilization of STAT 6 were assessed (14 -16). Inhibitors were added to Petri dish-grown T84 cells for 10 min prior to cytokine stimulation, then concurrent with either 100 ng/ml IL-4 or IL-13. Nuclear protein extracts were collected 30 min later and snap frozen for subsequent EMSA analysis.
Electrophoretic Mobility Shift Assay-Nuclear protein extracts were obtained following the protocol of Andrews and Faller (17), with the addition of the enzyme inhibitors aprotinin (10 g/ml), pepstatin A (2 g/ml), leupeptin (2 g/ml), and phenylmethylsulfonyl fluoride (20 mg/ ml) (18) (all Sigma). Protein concentrations in the nuclear extracts were determined using the Bio-Rad/Bradford assay.
Epithelial Paracellular Permeability-Confluent T84 filter-grown monolayers with a transepithelial resistance (TER) Ͼ 1000 ⍀/cm 2 were exposed to 10 ng/ml of either IL-4 or IL-13 for 24-h Ϯ TFDs (0.2 M), wortmannin (0.1, 1 M), or LY294002 (10, 20 M) (15,16). The TFDs and inhibitors were added to the cells following the same procedure as that described for the EMSA analysis, with the addition that wortmannin was added every 8 h to the cells due to its reported instability in culture medium at 37°C (14). Cytokine and all reagents were added to the basolateral compartment of the culture well. TER was measured prior to and 24 h after cytokine administration using a voltmeter with chop-stick electrodes (Millipore Corp., Bedford, MA). Time-matched controls consisted of naïve T84 monolayers and those treated with drug or vehicle only.
Data Presentation and Statistical Analysis-TER data are presented as the percentage of pretreatment values Ϯ S.E., where n ϭ number of individual epithelial preparations. Data were compared by ANOVA followed by post-hoc Newman-Keuls statistical comparisons, where p Ͻ 0.05 was set as the level of statistically significant difference.

RESULTS
IL-4 and IL-13 Evoke Similar STAT 6 DNA Binding Activity in Epithelial Cells-When T84 cells were exposed to 1 ng/ml IL-4 or IL-13, a faint STAT 6 band was discernible on EMSA, whereas STAT 6 DNA binding activity was clearly evident in nuclear extracts from cells treated with 10 or 100 ng/ml of either cytokine (n ϭ 3) (Fig. 1). Time-dependent experiments with 100 ng/ml IL-4 or IL-13 showed STAT 6 was detectable in nuclear extracts 15 min, 30 min (ϭ maximum response), and 1 h post-treatment, was apparent but reduced by 4 and 14 h, and was undetectable at 24 h (n ϭ 2-3) (Fig. 2, A and B). STAT 6 identity was confirmed by the ability of a STAT 6-specific, but not an irrelevant isotype-matched antibody, to supershift the suspected band. Specificity of the DNA oligonucleotide probe for STAT 6 was confirmed by the ability of increasing concentrations (50 -200ϫ) of non-[␣-32 P]dCTP-labeled probe, but not the mutated probe sequence, to progressively compete out the suspect band (n ϭ 3) (Fig. 3). (Similarly, nuclear extracts from IL-4-or IL-13-treated cells were not reactive with a STAT 1-specific probe 2 ). In some, but not all, T84 nuclear protein extracts, a faint band on EMSA was detected that migrated slightly faster than the inducible STAT 6 signal (see Fig. 1). This band, when present, occurred in nuclear extractions from non-treated cells and those treated with IL-4 or IL-13. The exact nature of this constitutive STAT 6-like band is unclear (reactive with the anti-STAT 6 antibody and reduced in intensity by competitive binding (Fig. 1)). We speculate that it is a truncated form of STAT 6 analogous to isoforms of STAT 6 that have been identified in mast cells, fibroblasts, and T cells (20 -22). IL-4 and IL-13 (10 -100 ng/ml, 30-min exposure) were also found to activate STAT 6 in the HT-29 and A549 epithelial cell lines (data not shown).
TFDs Abrogate the IL-4-and IL-13-induced Nuclear STAT 6 Signal- Fig. 4 (A and B) illustrates that nuclear extracts from T84 cells pretreated with TFDs (0.2 M) and then exposed to IL-4 or IL-13 had a significantly reduced STAT 6 signal on EMSA (n ϭ 5 and 3). In contrast, epithelial monolayers treated with TFDs containing a mutation in the STAT 6 recognition sequence and then exposed to IL-4 displayed STAT 6 DNA binding activity on EMSA that was not different from nuclear proteins extracted from T84 cells treated only with cytokine (n ϭ 2; Fig. 4C).

Inhibitors of PI3K Do Not Alter IL-4-or IL-13-induced STAT
6 Activation-Nuclear extracts from T84 cells pretreated for 10 min with wortmannin or LY294002 and then exposed for 30 min to fresh inhibitor with or without IL-4 or IL-13 (100 ng/ml) were examined on EMSA. Consistently, neither wortmannin IL-4 and IL-13 Decreases in TER Are Reduced by Inhibitors of PI3K Activity-Both IL-4 and IL-13 (10 ng/ml) caused a significant decrease in the TER of filter-grown T84 monolayers, which was consistently reduced to 30 -60% of pretreatment values by 24 h post-treatment (n Ͼ 24; Figs. 4D, 5C, 6C), which is consistent with other investigations with these cytokines (5). Pretreatment with the STAT 6-specific TFDs (or the mutant TFDs as a negative control) did not preserve the TER of cytokine-treated T84 monolayers (n ϭ 8; Fig. 4D). However, treat-ment with the PI3K inhibitors wortmannin or LY294002 significantly preserved the TER of T84 monolayers treated with IL-4 or IL-13 24 h previously. Fig. 5C shows that both inhibitors of PI-3K activity dose dependently prevented the drop in showing that the inducible band detected following exposure to IL-4 (100 ng/ml, 30 min) is in fact STAT 6. The indicated STAT 6 band (arrowhead) is supershifted by inclusion of an anti-STAT 6 antibody (␣-STAT 6) but not an isotype-matched irrelevant antibody. Increasing concentrations of non-32 P-labeled probe as a cold competitor prevents detection of STAT 6, whereas a mutant cold competitor (100fold excess) does not. EMSA also shows that the transcription factor decoys act as a cold competitor, whereas the mutant TFDs do not significantly reduce the IL-4-induced STAT 6 band (n ϭ 3). In this experiment the faster migrating constitutive STAT 6-like band was detected and was reactive with the STAT 6 antibody and the cold competitor. (Identical results were obtained with IL-13 stimulation of T84 cells (data not shown).)

FIG. 4.
EMSAs showing that nuclear extracts from T84 cells pretreated for 20 h with transcription factor decoys (TFDs) and then exposed to (A) IL-4 or (B) IL-13 (30 min) have a significantly reduced STAT 6 DNA binding activity (n ϭ 5 and 3, respectively). C, pretreatment with mutant TFDs or vehicle only does not interfere with subsequent STAT 6 mobilization by IL-4 (n ϭ 2). D, pretreatment with neither TFDs nor the mutant TFDs affected the drop in T84 transepithelial resistance observed 24 h after exposure to IL-4 or IL-13 (both 10 ng/ml). n ϭ 8 epithelial monolayers from 4 experiments; mean Ϯ S.E.; *, p Ͻ 0.05 compared with control; no inhib., no inhibitor. T84 TER caused by 24-h exposure to IL-4 (n ϭ 6 -12). In general, the drug treatments did not completely prevent the drop in TER, which may be a reflection of the pharmacokinetics of the drugs in T84 cells or indicate a role for another intracellular signaling pathway (and we have not unequivocally ruled out STAT 6) in the regulation of epithelial paracellular permeability. Similarly, the IL-13-induced drop in TER was significantly reduced by wortmannin (0.1 M) and LY294002, which at the higher concentration completely preserved monolayer TER (n ϭ 9 -18; Fig. 6C). Exposure to TFDs, wortmannin or LY294002 only, or the vehicle for the inhibitors had no significant affect on T84 TER, and thus only media controls are shown on Figs. 4D, 5C, and 6C. DISCUSSION Cytokines directly affect a variety of epithelial functions, including electrolyte transport, barrier function, and mediator synthesis (23). Recent studies have begun to explore epithelial intracellular signaling cascades in response to cytokine exposure (24,25). Here we used a pharmacological approach to provide evidence for a PI3K-dependent, STAT 6-independent proximal signal transduction pathway in the IL-4 and IL-13 regulation of paracellular permeability in T84 cells, a model gut epithelium.
Cell-mediated immunity is typified by IFN␥ production (i.e. T helper 1 cell events), and humoral immunity is characterized by IL-4 and IL-10 synthesis (i.e. Th2 events). Investigations with patients with airways inflammation have shown elevated serum levels of Th2-type cytokines and, somewhat unexpectedly, concomitant increases in gut permeability (26). Increases in IL-4 mRNA in ileal resections from some patients with Crohn's disease and colonic tissue from cohorts of patients with ulcerative colitis have been shown (3,4). Also, the colitis that develops in mice lacking the ␣-chain of the T cell receptor is less severe in animals that also lack IL-4, although gut permeability has not been examined in this model (27). Moreover, in vitro studies have unequivocally shown that cytokines can directly alter epithelial permeability (23,28): IL-4 and IL-13 often exert similar biological activities and both decrease the barrier function of T84 monolayers (6,29). Consequently, excess IL-4 (or IL-13) in vivo may be important in the regulation of gut permeability and the development and/or progression of inflammatory disease.
The IL-4 and IL-13 receptors share a common chain, designated the IL-4R␣ chain, and studies with immune cells have revealed that following receptor ligation both cytokines acti- vate the STAT 6 and PI3K pathways (9). Although cell-specific events have been described, the bulk of the available data indicate that IL-4/STAT 6 events control transcriptional regulation of genes for CD23, major histocompatibility complex II, and IgE as well as facilitating the generation of Th2 cells (29,30). The PI3K pathway has been shown to regulate vesicle trafficking and, by inference, cytoskeletal rearrangements (31), proliferation, and rescue from apoptosis (32). The aims of our investigation were to assess if IL-4 and IL-13 activate STAT 6 to the same degree and to determine if IL-4 and IL-13 regulate epithelial permeability via a STAT 6 or a PI3K pathway.
Initial studies revealed that IL-4 and IL-13 activation of STAT 6 (as defined by DNA binding activity on EMSA) was time-and dose-dependent and virtually identical for both cytokines. The identity of STAT 6 on EMSA was confirmed by anti-STAT 6 antibody supershifts and competitive binding by a non-radiolabeled STAT 6 DNA oligonucleotide. Although these data fit with the predicted effect of IL-4 and IL-13, they are nevertheless the first demonstration of STAT 6 activation in enteric epithelium. Indeed, it is only recently that STAT signaling has piqued the interest of the epithelial biologist (33)(34)(35)(36). Additionally, a faster migrating constitutive STAT 6-like band was identified on EMSA. This may be a truncated STAT 6 isoform analogous to one of the STAT 6 isoforms (designated STAT 6, 6b, and 6c) that have been identified in a murine mast cell line (20), a human lung fibroblast cell line (21), and primary isolates of human bloodderived natural killer and T cells (23).
Currently there are no specific pharmacological inhibitors of STAT 6, with the possible exception of the immunosuppressive drug leflunomide, which was found to block STAT 6 activation in IL-4-treated B cells (37). Following the published leflunomide treatment protocol, we were unable to block either IL-4or IL-13-induced STAT 6 activation in T84 cells or the drop in TER. 3 These preliminary data may reflect cell-specific effects of leflunomide. Rather, we employed transcription factor decoys (TFDs) to block STAT 6 activation. With this strategy, activated STAT dimers will bind to the TFDs in the cytoplasm, effectively blocking their binding to host cell nuclear DNA (13,38,39). A recent report (12) suggests that TFDs may remain stable in the cytoplasm for up to 48 h. Importantly, TFDs bearing the particular binding element specific for STAT 6 used in our study (i.e. TTC-N 4 -GAA) have been used to functionally disrupt the IL-4-induced, STAT 6-driven proliferation of CD4 ϩ T cells (11). Pretreatment of T84 cells with the appropriate TFD, but not a scrambled form of the DNA sequence, significantly blocked the activation of STAT 6 by 100 ng/ml IL-4 or IL-13 as determined by subsequent EMSA. Similarly, the constitutive STAT 6-like band was also reduced by TFD exposure. However, TFD pretreatment did not interfere to any significant extent with the ability of a 10-fold less dose of IL-4 or IL-13 to reduce the TER across T84 monolayers. Thus, we present no FIG. 6. Neither the PI3K inhibitor (A) wortmannin nor (B) LY294002 alter the STAT 6 activity (arrowhead) of nuclear protein extracts from T84 cells treated with IL-13 (30 min) as judged by EMSA (n ϭ 3-4). C, using the same treatment regime, wortmannin (W) and LY294002 (LY) reduce the magnitude of the drop in transepithelial resistance caused by IL-13 over a 24-h period. n ϭ 9 -18 T84 monolayers from 3 experiments; mean Ϯ S.E.; *, p Ͻ 0.05 compared with control; #, p Ͻ 0.05 compared with control and monolayers treated only with IL-4; no inhib., no inhibitor. data in favor of the involvement of the inducible, or constitutive, form of STAT 6 in the regulation of epithelial paracellular permeability. The possibility remains that the EMSA was unable to detect low levels of STAT 6 that were capable of transducing a biological signal; that is, small amounts of STAT 6 below the EMSA detection limit, but enough to affect transcription may have escaped the TFDs. Unequivocal statements on the role of STAT 6 in regulating epithelial paracellular permeability await the advent of a stable and, preferably, inducible STAT 6 dominant-negative transfected enterocytic cell line suitable for electrophysiological studies.
Ligation of the IL-4 and IL-13 receptors results in sequential binding of the docking protein insulin receptor substrate-1/2 and then PI3K, followed by the production of lipid inositols that can activate a variety of signaling molecules, including PKC (40). Use of T84 cells and other epithelia have implicated PKC activity in the regulation of the epithelial tight junction function, the rate-limiting step governing paracellular permeability (41)(42)(43). Recently, biphasic PI3K activity with characteristic early and late functions has been described, suggesting that PI3K is capable of supporting immediate and sustained cellular responses (44). We found that inclusion of the inhibitors of PI3K activity, wortmannin, or the more specific compound LY294002, resulted in a significant ablation of the ability of IL-4 or IL-13 to decrease T84 monolayer TER. Indeed, these observations are in accordance with the recognized ability of PI3K to affect the cytoskeleton, which is a pivotal determinant in the control of epithelial tight junction activity (31). Additionally, rescue of the HT-29 epithelial cell line from apoptosis by exposure to IL-13 was shown to occur via a PI3K pathway (45).
A recent report (46) indicates that wortmannin and LY294002 can inhibit NFB activation in stimulated immune cells, raising the possibility that these drugs affect STAT 6 activation in response to IL-4 or IL-13. However, neither wortmannin nor LY294002, when used at the highest doses employed in the physiological studies, affected the STAT 6 activation by IL-4 or IL-13. Similarly, others have shown that wortmannin did not affect IL-4-evoked STAT 6 activation as assessed by Western blotting (47). Collectively, these data support the hypothesis that the proximal signal transduction events elicited by IL-4 and IL-13, namely STAT 6 and PI3K activation, are separate and distinct. The virtually indistinguishable kinetics of IL-4-and IL-13-induced STAT 6 mobilization, along with the pivotal role of PI3K in their modulation of epithelial paracellular permeability, adds support to the hypothesis that the common IL-4R␣ chain in the IL-4 and IL-13 receptors is crucial for signal transduction in enteric epithelia.
In conclusion, we have identified PI3K activity as a major proximal signaling event involved in the IL-4-and IL-13-induced increase in epithelial paracellular permeability. The data raise questions pertinent to downstream events from PI3K in the regulation of paracellular permeability such as the involvement of PKC isoforms in the modulation of epithelial tight junctions by cytokines. These issues are the focus of ongoing studies. Finally, our data lead to the tantalizing speculation that future therapies for enteropathies characterized by an epithelial permeability defect could specifically target PI3K while leaving STAT 6-driven events (e.g. major histocompatibility complex II expression) intact.