Interleukin-2 Receptor β Subunit-dependent and -independent Regulation of Intestinal Epithelial Tight Junctions*

Interleukin (IL)-15 is able to regulate tight junction formation in intestinal epithelial cells. However, the mechanisms that regulate the intestinal barrier function in response to IL-15 and the involved subunits of the IL-15 ligand-receptor system are unknown. We determined the IL-2Rβ subunit and IL-15-dependent regulation of tight junction-associated proteins in the human intestinal epithelial cell line T-84. The IL-2Rβ subunit was expressed and induced signal transduction in caveolin enriched rafts in intestinal epithelial cells. IL-15-mediated tightening of intestinal epithelial monolayers correlated with the enhanced recruitment of tight junction proteins into Triton X-100-insoluble protein fractions. IL-15-mediated up-regulation of ZO-1 and ZO-2 expression was independent of the IL-2Rβ subunit, whereas the phosphorylation of occludin and enhanced membrane association of claudin-1 and claudin-2 by IL-15 required the presence of the IL-2Rβ subunit. Recruitment of claudins and hyperphosphorylated occludin into tight junctions resulted in a more marked induction of tight junction formation in intestinal epithelial cells than the up-regulation of ZO-1 and ZO-2 by itself. The regulation of the intestinal epithelial barrier function by IL-15 involves IL-2Rβ-dependent and -independent signaling pathways leading to the recruitment of claudins, hyperphosphorylated occludin, ZO-1, and ZO-2 into the tight junctional protein complex.


Interleukin (IL)-15 is able to regulate tight junction formation in intestinal epithelial cells. However, the mechanisms that regulate the intestinal barrier function in response to IL-15 and the involved subunits of the IL-15 ligand-receptor system are unknown. We determined the IL-2R␤ subunit and IL-15-dependent regulation of tight junction-associated proteins in the human intestinal epithelial cell line T-84. The IL-2R␤ subunit was expressed and induced signal transduction in caveolin enriched rafts in intestinal epithelial cells. IL-15-mediated tightening of intestinal epithelial monolayers correlated with the enhanced recruitment of tight junction proteins into Triton X-100-insoluble protein fractions. IL-15-mediated up-regulation of ZO-1 and ZO-2 expression was independent of the IL-2R␤ subunit, whereas the phosphorylation of occludin and enhanced membrane association of claudin-1 and claudin-2 by
The regulation of the intestinal barrier is determined by the assembly of tight junctions (1). Tight junctions not only create a primary barrier to prevent paracellular transport of solutes, but they also restrict the lateral diffusion of membrane lipids and proteins to maintain cellular polarity (1)(2)(3)(4). Tight junctions are macromolecular assemblies that form a specialized membrane domain at the most apical region of polarized epithelial cells. Tight junctions have a highly dynamic structure whose permeability, assembly, and/or disassembly can be regulated by a variety of cellular and metabolic mediators including cytokines, which have major functions in the immune system (5)(6)(7)(8)(9)(10). Immune modulators therefore may control tight junction-dependent intestinal barrier function during development, wound healing, and pathological processes such as cancer or chronic inflammation. To date, the claudins (11) and occludin (12,13) have been identified as tight junction-specific integral membrane proteins. Occludin and claudins can interact with the PDZ domains of zonula occludens proteins ZO-1 and ZO-2 (14,15). ZO-1 and ZO-2 are membrane-associated guanylate kinase-like homologues (MAGUKs) 1 that may play a general role in creating and maintaining specialized membrane domains by cross-linking multiple integral membrane proteins at the cytoplasmic surface of plasma membranes in various cells types (16 -18). Furthermore, ZO-1 and ZO-2 bind directly to actin filaments at their COOH-terminal regions, suggesting that these molecules function as cross-linkers between tight junction strands and actin filaments (19 -22). In addition, ZO-2 has been reported to associate with ZO-1 directly (20,21). Therefore zonula occludens proteins may establish a framework to combine the functional components of tight junctions during the generation of the intestinal barrier.
Disruption of the intestinal epithelial cell barrier function may play an important role in the pathogenesis of inflammatory bowel disease (23)(24)(25). Although it is not clear whether intestinal barrier dysfunction is involved in the initiation or is the result of intestinal inflammation, it is likely that dysfunction of the intestinal barrier would perpetuate intestinal inflammatory responses. Cytokines, which regulate intestinal immune response, may perturb the intestinal barrier or could contribute to mechanisms that may have evolved to rapidly seal the intestinal monolayer to limit the influx of highly antigenic substances into the intestinal lamina propria.
The IL-15 cytokine receptor complex consists of the IL-15R␣, the IL-2R␤, and the ␥c receptor subunit, which is common to several cytokine receptor complexes (26 -29). Because the IL-15 receptor complex shares the IL-2R␤ and the ␥c receptor subunits with IL-2, much attention has focused on the overlapping functional effects of IL-15 and IL-2 in the regulation of T-and B-lymphocytes (26 -29). However, the expression of the IL-15R␣ subunit in a wide variety of different tissues suggests that the functional role of IL-15 is not limited to the regulation of lymphoid cells (30). Recently IL-15 has been recognized as a cytokine up-regulated during inflammatory bowel disease as well as during the colitis in IL-2-deficient mice (31)(32)(33). Nevertheless, it is not clear whether IL-15 contributes to the perpetuation of inflammation or the regulation of intestinal tissue repair. Furthermore, IL-15 may be an important regulator of intestinal intraepithelial lymphocytes. IL-15 is a potent activator of intestinal intraepithelial lymphocytes (34 -36), and the number of intestinal intraepithelial lymphocytes is reduced in mice lacking the IL-15R␣ subunit (37).
T-84 cells provide a well established model for the assembly of intercellular junctions and the development of apical-basolateral polarity (5). T-84 cells express the IL-15R␣ subunit (9) and the common ␥c receptor subunits (9,38). However, T-84 cells lack, in contrast to primary intestinal epithelial cells, the expression of the IL-2R␤ subunit (9,38). Nevertheless, T-84 cells respond to stimulation with IL-15 with an increase in transepithelial electrical resistance (TER) (9). However, the mechanisms responsible for the regulation of the intestinal barrier function by IL-15 have not been determined.
To study the regulation of tight junction formation mediated through the IL-15 ligand receptor system in intestinal epithelial cells, we stable transfected T-84 cells with IL-2R␤ and characterized the subsequent basal and IL-15-mediated regulation of tight junction-associated proteins.
Expression Vector Construction-To generate an expression vector encoding the IL-2R␤ subunit, cDNAs encoding extracellular and intracellular regions were subcloned separately. Reverse transcriptase-PCR to generate the cDNA for the NH 2 -terminal region of the IL-2R␤ subunit was carried out with reverse transcribed RNA isolated from peripheral blood mononuclear cells with the primer pair 5Ј-CAG CAC CGG GGA GGA CTG GA-3Ј and 5Ј-CGC CAG GGC TGA AGG ACG AT-3Ј for 40 cycles at 94°C for 1 min, 70°C for 1 min, and 72°C for 2 min. The PCR product was cloned into pBluescript KSϩ (Stratagene, La Jolla, CA), and the BamHI/SacI fragment was released for ligation. To generate the COOH-terminal region of the IL-2R␤ cDNA, nested PCR was carried out for 20 cycles at 94°C for 1 min, 70°C for 1 min, and 72°C for 2 min with first sets of primers of 5Ј-CTG CAA GGC GAG TTC ACG AC-3Ј and 5Ј-AGC AGC AGT GGA GGT TTG GA-3Ј. A second PCR was carried out using a 1:100 dilution of the first PCR product as a template for 30 cycles with the second set of primers (5Ј-GGC TTT TGG CTT CAT CT-3Ј and 5Ј-AGC TGC AAC TGG ACA CTG AG-3Ј) at 94°C for 1 min, 70°C for 1 min, and 72°C for 2 min. The PCR product was cloned into pBluescript KSϩ, and the SacI/XhoI fragment was released. Extracellular and intracellular parts of IL-2R␤ cDNA were ligated at the SacI site and ligated into a BamHI/XhoI-digested pcDNA3.1ϩ vector (Invitrogen, San Diego, CA). Sequence was confirmed in both directions using the dideoxy termination method.
Cell Proliferation Assay-T-84␤ cells and parental T-84 cells were seeded into 96-well plates at a density of 5 ϫ 10 3 cells/well and were cultured in serum-free Dulbecco's modified Eagle's medium in the presence of various concentration of recombinant human IL-15 at 37°C for 24 -72 h. Proliferation was assessed by MTS assay using CellTiter 96 Aqueous kit (Promega, Madison, WI) according to the manufacturer's instructions. Each assay was performed in triplicate. 20 l of a fresh mixture of MTS tetrazolium compound and phenazine methosulfate was added to each well and incubated at 37°C for 1 h. The amount of formazan corresponds to the number of viable cells and was measured at an absorbance of 490 nm by an enzyme-linked immunosorbent assay reader. Wells containing medium, but no cells, were subtracted as background from the raw absorbance values.
Measurement of TER-For sequential TER measurements, T-84 monolayers were maintained on 24-well Transwell collagen-treated permeable supports (Corning Coster, Cambridge, MA). We determined the IL-2R␤-dependent regulation of TER during the assembly of tight junctions after replating T-84 cells to model the reconstitution of the intestinal epithelial barrier function during in intestinal wound healing. T-84 cells were plated at a concentration of 1.5 ϫ 10 6 /well. The cells were stimulated with 100 ng/ml IL-15 24 h later in the presence or absence of anti-IL-2R␤ or control antibodies, and the TER measurements were carried out at the indicated time points in triplicate samples. Millicell-ERS epithelial volt-ohmmeter (World Precision Instruments, New Haven, CT) was utilized under temperature-controlled conditions at 37°C with electrodes reproducibly placed. TER values were calculated by subtracting the contribution of the bare filter and medium. Statistical analysis was performed by Student's t test.
Preparation of Triton X-100-soluble and -insoluble Protein Fractions-To isolate Triton X-100-soluble and -insoluble protein fractions, confluent T-84 cell monolayers grown on 10-cm dishes were washed three times with ice-cold PBS, lysed in Triton X-100 buffer (1% Triton X-100, 100 mM NaCl, 10 mM HEPES, pH 7.6, 2 mM EDTA, 1 mM phenylmethylsulfonylfluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 g/ml pepstatin, 4 mM sodium orthovanadate, 40 mM sodium fluoride), and then passed through a 21-gauge needle ten times. The lysates were then centrifuged at 15,000 ϫ g for 30 min at 4°C. The resulting supernatant was considered the Triton X-100-soluble fraction. The pellet was solubilized in Triton X-100 buffer containing 1% SDS using an ultrasonic disintegrator, cleared by centrifugation at 15,000 ϫ g for 5 min at 4°C, and referred to as the Triton X-100-insoluble fraction. The protein concentration of each sample was quantified by the Bradford method.
Isolation of Detergent-insoluble Glycolipid-enriched Membrane Microdomains-Glycolipid-enriched membrane microdomains, or detergent-insoluble glycolipid rafts, were isolated as described before with minor modifications (39 -41). Four dishes of T-84 or T-84␤96 cells (diameter, 10 cm) were washed three times with ice-cold PBS, and the cells were solubilized in 1 ml of lysis buffer (25 mM MES, pH 6.8, 150 mM NaCl, 1% Triton X-100 supplemented with protease/phosphatase inhibitor mixture). After incubation for 30 min on ice, the lysate was gently processed in a tight fitting Dounce homogenizer 10 times and cleared by centrifugation for 5 min at 1000 ϫ g. The supernatant was adjusted to 40% sucrose with an equal volume of 80% sucrose in MBS (25 mM MES, pH 6.8, 150 mM NaCl supplemented with the protease/ phosphatase inhibitor mixture) and transferred to the bottom of a centrifugation tube. A sucrose step gradient with 30, 25, 20, 15, and 5% in MBS (2 ml each) was layered on top. After centrifugation in a SW 41 rotor (Beckman) at 39,000 rpm for 14 -16 h at 4°C, fractions of 1 ml each were taken, starting from the top and going to the bottom. Protein measurement of the cellular fractions was performed with BCA protein assay reagent, according to the manufacturer's instructions (Pierce).
Gel Electrophoresis and Western Blotting Analysis-The samples were electrophoresed through a 4 -20% gradient SDS-polyacrylamide gel and transferred onto polyvinylidene difluoride membranes (Millipore, Bedford, MA). After 1 h of blocking (Tris-buffered saline, 0.1% Tween 20, 1% bovine serum albumin), the blots were incubated overnight at 4°C with primary antibodies diluted in blocking buffer. After washing in Tris-buffered saline, 0.1% Tween 20, the membrane was incubated with appropriate secondary antibody diluted in blocking buffer for 60 min at room temperature. The hybridized band was detected by an ECL kit (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Immunoblots were stripped with 62.5 mM Tris, pH 6.8, 2% SDS containing 10 mM ␤-mercaptoethanol at 50°C for 30 min.
RNA Extraction and Northern Blot Analysis-Total RNA was isolated from cultured cells using Trizol reagent (Life Technologies, Inc.). Thirty micrograms of total RNA were electrophoresed in a 1% agarose formaldehyde gel and then transferred onto a nylon membrane (Magna NT, MicroSeparations Inc., Westbrough, MA) by capillary blotting. The probes were labeled with [␣-32 P]dCTP using the Rediprime Random Primer Labeling Kit (Amersham Pharmacia Biotech). Membranes were hybridized with radiolabeled cDNA probes in Quickhyb solution (Stratagene, La Jolla, CA) at 65°C for 1 h. The membranes were washed with 1% SDS, 2ϫ sodium chloride sodium citrate buffer. The blots were analyzed by autoradiography. The blots were sequentially hybridized with IL-15 and glyceraldehyde-3-phosphate dehydrogenase probes as described previously (42).
Confocal Microscopy-T-84 and T-84␤96 cells were grown for 3 days with or without 100 ng/ml IL-15 on chamber slide culture chambers (Nunc Inc., Naperville, IL). T-84 cells or T-84␤96 cells were washed three times with Dulbecco's PBS (Life Technologies, Inc.) and were fixed for 10 min in 2% paraformaldehyde for immunostaining with antibodies recognizing occludin or for 30 s in aceton at Ϫ20°C for the detection of claudin-1, claudin-2, ZO-1, and ZO-2. Thereafter, the cells were blocked with 1:200 PBS-diluted normal donkey serum for 1 h at 20°C and incubated with 1:200 in PBS-diluted primary antibodies overnight at 4°C. After three washes with PBS, the monolayers were incubated at room temperature with anti-mouse or anti-rabbit IgG fluorescein isothiocyanate-conjugated antibody (1:500 in PBS) for 1 h in the dark at room temperature and analyzed with a confocal immunofluorescent microscope (Bio-Rad).

Stable Transfection with the IL-2R␤ Subunit Initiates the Concentration-dependent Autocrine Activation of Phosphotyrosine Kinases in T-84
Cells-An initial screen of 98 IL-2R␤transfected T-84 clones yielded two clones with a high expression of functional IL-2R␤ subunits. The low number of stable transfected clones reflects the difficulties in stably transfecting T-84 cells. Both IL-2R␤ subunit expressing T-84 cell clones (T-84␤29 and T-84␤96) exhibited a similar phenotype and were characterized by an increased resistance to separation by a 1 mM EDTA, 0.25% trypsin solution. T-84␤29 and T-84␤96 monolayers required 15-20 min of incubation to obtain singlecell suspensions instead of the 5-7 min of incubation time required for the parental T-84 monolayers. Both T-84␤29 and T-84␤96 cells form uniform, tightly packed monolayers within 3 days after separation at a 1:4 split ratio.
Expression levels and membrane targeting of the IL-2R␤ subunit in stable transfected T-84 cells was analyzed in Triton X-100-soluble and -insoluble protein fractions (Fig. 1A). As demonstrated in Fig. 1A, T-84␤29 and T-84␤96 cells expressed the IL-2R␤ subunit, whereas the IL-2R␤ chain was not detectable in parental T-84 cells. When analyzed densitometrically, the expression of the IL-2R␤ was 3-fold higher in T-84␤29 cells than in T-84␤96 cells (Fig. 1A). In both T-84␤29 and T-84␤96 cells, the IL-2R␤ subunit was highly enriched in Triton X-100insoluble protein fractions, demonstrating that the stable expressed cytokine receptor was integrated into T-84 cell membranes (Fig. 1A).
Overexpression of cytokine receptors may lead to autonomous activation of signal transduction events. Therefore, we determined the basal activation of tyrosine phosphorylation in parental T-84 cells, T-84␤29, and T-84␤96 cells. As demonstrated in Fig. 1B, T-84␤29 cells demonstrated an enhanced basal level of phosphotyrosine kinase activity compared with the parental T-84 cells and T-84␤96 cells when membraneassociated Triton X-100-insoluble protein fractions were analyzed (Fig. 1B, lane 4 compared with lanes 2 and 6). The elevation of tyrosine phosphorylation in the T-84␤29 cells correlated with the high level of IL-2R␤ subunit expression in Triton X-100-insoluble membrane fraction in this T-84 cell clone.
The IL-2R␤ Subunit Transduces IL-15-mediated Signals in Caveolin-1-enriched Detergent-insoluble Glycolipid-enriched Membrane Microdomains in T-84 Cells-To further characterize the membrane compartment containing the IL-2R␤ subunit, we separated postnuclear membrane protein fractions by sucrose density gradient centrifugation. This method separates detergent-insoluble membrane fractions characterized by their specific lipid composition (40,43,44). As demonstrated in Fig.  2A, the IL-2R␤ subunit was present in low sucrose gradient fractions of T-84␤96 cells. IL-2R␤ was present in membrane fractions corresponding to sucrose concentrations of 20 -24%, with the highest concentration in the 20% sucrose gradient fraction ( Fig. 2A). These low sucrose gradient fractions have been associated with detergent-insoluble glycolipid-enriched membrane fractions or rafts (45). In epithelial cells, these membrane fractions are characterized by the presence of the scaffolding protein caveolin (46). As demonstrated in Fig. 2A, Western blot analysis of these cellular compartments revealed that the expression of caveolin-1 was restricted to the same sucrose gradients as that of the IL-2R␤.
In the next set of experiments, we determined whether the expression of the IL-2R␤ subunit enhanced the responsiveness of T-84 cells to IL-15. In these experiments, we used T-84␤96 cells, which had a base-line activation of phosphotyrosine kinase comparable with T-84 cells (Fig. 1B). T-84 and T-84␤96 cells were cultured for 15 min with or without IL-15. As demonstrated in Fig. 2B, IL-15 stimulation was able to induce a strong up-regulation of phosphotyrosine kinase activity in T-84␤96 cells after 15 min (Fig. 2B, lane 10). IL-15 induced tyrosine phosphorylation of proteins with a molecular mass of 100 kDa, as well as a number of proteins with molecular masses between 42 and 54 kDa in T-84␤96 cells (Fig. 2B, lane  10). In contrast, IL-15 induced only a weak tyrosine phosphorylation of proteins in the parental T-84 cell line, which had a molecular mass of 160 kDa in the 20% sucrose gradient fraction and of 150 kDa in the 32-38% sucrose gradient fractions (Fig.  2B, lanes 10 and 14 -16). The majority of the proteins that were tyrosine-phosphorylated in response to IL-15 migrated with membrane proteins of the 20% sucrose gradient (Fig. 2B, lanes  2 and 10), which corresponds to the same membrane fraction in which the IL-2R␤ subunit was enriched in T-84␤96 ( Fig. 2A). Collectively, these data indicate that the IL-2R␤ subunit stable expressed in T-84 cells constitutes a functional IL-15 receptor. The IL-2R␤ subunit is largely sequestered and initiates signal transduction in a microdomain of T-84 cell membranes that display the biophysical characteristics of caveolae (45).

Expression of the IL-2R␤ Subunit Enhances the Development of Transepithelial Electrical Resistance and Renders T-84 Cells
Responsive to IL-15-The assembly of tight junctions is correlated with the development of TER across T-84 intestinal epithelial monolayers. Increased cell-cell adhesion of the T-84␤29 and T-84␤96 cells suggested enhanced formation of adherens and tight junctions. Furthermore, IL-15 has been identified as a regulator of TER in T-84 cells (9). We therefore compared the development of TER in T-84␤96 and parental T-84 cells. As demonstrated in Fig. 3A, T-84␤96 cells developed a higher steady state TER than the parental T-84 cells. The TER of T-84␤96 cells increased 26 Ϯ 4% within 2 days after reaching confluence. In contrast, the TER of the parental T-84 cells increased only 4 Ϯ 3% in the same time period (*, p Ͻ 0.001; Fig. 3A). Overall, the TER of the parental T-84 cells increased 9 Ϯ 4% over the 4-day observation period, significantly below the increase in TER of 29ϩ3% observed in the T-84␤96 cells (^, p Ͻ 0.005; Fig. 3A). In the presence of anti-IL-2R␤ antibody, TER of T-84␤96 cells increased only 12 Ϯ 3% over 2 days, whereas the control IgG-treated T-84␤96 cells reached steady state TER with an increase of 26 Ϯ 4% over the same time period (ϩ, p Ͻ 0.01; Fig. 3A). The assembly of tight junctions with the subsequent development of TER in T-84␤96 cells was enhanced by exogenous IL-15. In the presence of IL-15, T-84␤96 cells reached steady state TER levels 24 h earlier (30 Ϯ 4% increase) than unstimulated T-84␤96 cells, which increased their TER 12 Ϯ 4% within 24 h after reaching confluence ( ¶, p Ͻ 0.001; Fig. 3A). The addition of IL-15 to the parental T-84 cells increased the TER 15 Ϯ 3%, compared with an increase of 8 Ϯ 4% of unstimulated T-84 cells within the observed time period (Fig. 3A). As demonstrated in Fig. 3B, T-84 and T-84␤96 cells did not differ in their proliferation rate in the presence or absence of IL-15.
Furthermore, the regulation of tight junction formation in T-84 and T-84␤96 cells was not due to the expression of IL-15 by the intestinal monolayers themselves. As demonstrated in Fig. 3C, parental T-84 cells express very small amounts of IL-15 mRNA during the observed time period in comparison with Caco-2 cells. Furthermore, only the larger of the two IL-15 isoform transcripts was detectable in T-84 cells (Fig. 3C). IL-15 mRNA expression was not detectable in the T-84␤96 cells (Fig.  3C). Most importantly, IL-15 stimulation of either T-84 cell line did not induce IL-15 mRNA expression (Fig. 3C). These data suggest that the basal regulation of TER in T-84␤96 cells may be mediated by low level of autonomous signaling of the over- expressed IL-2R␤. In addition, IL-2R␤ subunit expression enhanced the biological response to the exogenous IL-15 stimulation in T-84␤96 cells.

Expression of the IL-2R␤ Subunit in T-84 Cells Regulates Expression and Membrane Association of Tight Junction
Proteins-Assembly of tight junctions is associated with the development of resistance of tight junction protein complexes to solubility in detergent-salt extractions. Conversely, disassembly of tight junction proteins has been shown to correlate with internalization or diffuse cytoplasmic distribution of tight junction proteins, making them more extractable with detergentsalt solutions. Therefore, the solubility of tight junction proteins can be used as an indicator of membrane and cytoskeletal association (47)(48)(49). Moreover, an increased insoluble:soluble ratio of tight junction proteins can be correlated with an increase in transepithelial resistance (50). To determine the expression of tight junction-associated proteins in T-84, T-84␤29, and T-84␤96 cells, we carried out Western blot analysis of transmembrane proteins claudin-1, claudin-2, and occludin and the MAGUK family members ZO-1 and ZO-2 in Triton X-100-insoluble and -soluble protein fractions.
The anti-occludin antibodies detected a low molecular mass species (65-71 kDa) in soluble protein fractions and a high molecular mass species (72-79 kDa) in insoluble protein fractions in T-84 cells that have been recognized as hyperphosphorylated forms enriched in tight junctions (Fig. 4A) (49,51,52). Nonphosphorylated and phosphorylated basal occludin levels were comparable in T-84, T-84␤29, and T-84␤96 cells (Fig. 4, A  and D). As demonstrated in Fig. 4 (A and D), phosphorylated occludin was almost exclusively detected in the Triton X-100- insoluble protein fraction in T-84, T-84␤29, and T-84␤96 cells. Therefore, the preferential expression of claudin-1 and ZO-2 in the Triton X-100-soluble protein fraction suggests an expression of both proteins in membrane or cellular compartments distinct from tight junctions. Enhanced expression of claudins and redistribution of ZO-2 from Triton X-100-soluble into -insoluble protein fractions correlated with the higher expression of the IL-2R␤ in the T-84␤29 cell line.

IL-15 Regulates the Expression and Membrane Association of Tight Junctional Proteins by IL-2R␤-dependent and -independent Mechanisms-
The regulation of claudin-2 and ZO-2 expression in T-84␤29, which are characterized by a high expression of IL-2R␤, suggested that signals mediated through the IL-2R␤ may be able to regulate the intestinal barrier through modulation of the expression and the cellular distribution of tight junction-associated proteins. We therefore assessed the effect of IL-15 on expression and membrane association of tight junctional proteins in T-84 and T-84␤96 cells. T-84 cells or T-84␤96 cells were stimulated with IL-15 (100 ng/ml) 1 day after replating and harvested after 24, 48, and 72 h. Triton X-100-insoluble and -soluble protein fractions were prepared to determine the expression and membrane association of tight junction-associated proteins in stimulated or unstimulated T-84 and T-84␤96 cells. In T-84␤96 cells, IL-15 enhanced expression of claudin-1 and claudin-2 in Triton X-100-insoluble membrane fractions within 3 days 1.8 Ϯ 0.17-and 1.9 Ϯ 0.15-fold, respectively, compared with unstimulated cells (Fig. 5, A-C). In contrast, recruitment of claudin-1 and claudin-2 into Triton X-100-insoluble membrane protein fraction was not altered by IL-15 within the same time period in parental T-84 cells (Fig. 5, A-C).
In the absence of IL-15, claudin-1 and claudin-2 remained almost constant over the observed time period in both cell lines. As demonstrated in Fig. 5 (A, D, and E), IL-15 induced a 1.9 Ϯ 0.13-fold increase in nonphosphorylated occludin in Triton X-100-insoluble fractions of T-84␤96 cells.
Furthermore, IL-15 induced a 1.5 Ϯ 0.1-fold increase after 1 day and a 2.1 Ϯ 0.16-fold increase of phosphorylated occludin species after 3 days in the Triton X-100-insoluble fractions in T-84␤96 cells. In T-84 cells the expression of nonphosphorylated occludin increased 1.3 Ϯ 0.12-fold over the 3-day observation period. IL-15 did not further increase occludin expression in the parental T-84 cell line (Fig. 5, A and D). In contrast to the regulation of transmembrane tight junctional proteins, recruitment of ZO-1 and ZO-2 into the Triton X-100-insoluble protein fraction in T-84 cells was enhanced by IL-15 independent of the expression of the IL-2R␤ chain. As demonstrated in Fig. 5 (A, F,  and G), after replating without IL-15 stimulation, ZO-1 and ZO-2 levels in the Triton X-100-insoluble fractions decreased in T-84 cells 0.77 Ϯ 0.1-and 0.8 Ϯ 0.08-fold, respectively. In T-84␤96 cells, ZO-2 levels in the Triton X-100-insoluble protein fraction decreased 0.63 Ϯ 0.09-fold, whereas ZO-1 levels appeared to be unaltered over the 3-day observation period (Fig.  5, A, F, and G). In both T-84 and T-84␤96 cells, IL-15 was able to induce the expression and membrane association of ZO-1 and ZO-2. In T-84 cells, IL-15 increased the ZO-1 and ZO-2 levels in Triton X-100-insoluble protein fractions 1.4 Ϯ 0.11fold. In T-84␤96 cells, ZO-1 levels increased 1.6 Ϯ 0.17-fold, and ZO-2 levels increased 1.5 Ϯ 0.13-fold in response to IL-15. Whereas ZO-2 membrane association was regulated by IL-15, the available pool of ZO-2 in the cytoplasm was not altered in T-84 and T-84␤96 cells (Fig. 5A).
In the next set of experiments, we determined the expression and membrane association of claudin-1 and claudin-2, occludin, ZO-1, and ZO-2 by Z-section confocal microscopy in T-84 and T-84␤96 cells. In these experiments T-84 and T-84␤96 cells were stimulated with IL-15 for 3 days, and fixed, and stained for the presence of tight junction-associated proteins. As demonstrated in Fig. 6, IL-15 increased the recruitment of claudin-1 and claudin-2 into tight junctional complexes in T-84␤96 cells (B compared with A, and F compared with E) but not in the parental T-84 cells (C, D, G, and H). IL-15-induced signals also regulated the recruitment of occludin into the area of tight junctions in T-84␤96 cells (Fig. 6, compare J with I) but not in the parental T-84 cells (Fig. 6, K and L). In contrast IL-15 induced the recruitment and integration of ZO-1 and ZO-2 into tight junctions in T-84␤96 as well as the parental T-84 cells (Fig. 6, M-T). Claudin-2, occludin, ZO-1, and ZO-2 immunostaining reached further down the lateral cell contact areas after stimulation with IL-15, suggesting the expansion of tight junctional contact areas. In contrast, claudin-1 immunostaining appeared to be recruited into the apical portion of cell contacts, suggesting a distinct role of claudin-1 and claudin-2 in the structuring of tight junctional protein complexes. In T-84␤96 cells, IL-15 also increased the cytoplasmic staining for tight junction-associated proteins. In agreement with the Western blot analysis, ZO-2 was present constitutively in the cytoplasm of T-84␤96 and T-84 cells (Fig. 6, Q-T). DISCUSSION Characterization of the IL-2R␤ subunit overexpressing T-84 cell lines revealed that the assembly of tight junctions by IL-15 is mediated by the differential regulation of the expression and membrane association of tight junction-associated proteins. In our experiments, IL-15 was able to regulate the expression and membrane association of claudins, hyperphosphorylated occlu- membrane microdomains or rafts in T-84 cells. In these rafts, the IL-2R␤ subunit co-isolated with caveolin-1, a scaffolding protein abundant in detergent-insoluble glycolipid rafts. Similar isolated protein membrane fractions were recently shown to contain phosphorylated occludin and ZO-1 (49). These findings led to the hypothesis that tight junctions themselves have a raftlike membrane compartment that is characterized by the presence of caveolin-1. However, evidence is mounting that detergent-insoluble glycolipid rafts in intestinal epithelial cells are heterogenous and may include membrane microdomains, which do not contain caveolin-1 (53). Therefore, it is currently not clear whether signal transduction mediated through the IL-2R␤ subunit can lead to the direct activation of tight junction-associated proteins in the same raftlike membrane compartments.
The development of intestinal barrier function correlated with the induction of tyrosine phosphorylation in T-84␤96 cells. IL-15 induced tyrosine phosphorylation in T-84 as well as T-84␤96 cells. Changes in phosphotyrosine kinase activity have been linked to the regulation of tight junctions (54). Tyrosine phosphorylation of proteins may be involved in both the establishment and the disassembly of tight junctions. Tyrosine phosphorylation has been demonstrated to play an important role in the reassembly of occludin and other tight junction proteins during ATP repletion (55). Tyrosine phosphorylation of ZO-1 and other proteins occurred during the formation of podocyte junctions in glomeruli (56). Furthermore, epidermal growth factor stimulation caused a transient increase in tyrosine phosphorylation of both ZO-1 and ZO-2 in A431 cells (54). Recently, the tyrosine phosphorylation of occludin and ZO-1 has been linked to the regulation of tight junction formation in Rastransformed MDCK cells (50). However, tyrosine phosphorylation of both ZO-1 and ZO-2 induced by v-Src did not change either tight junction structure or TER (57). In contrast, the protein-tyrosine phosphatase inhibitors vanadate and H 2 O 2 resulted in a rapid increase in paracellular permeability and the redistribution of E-cadherin and ZO-1 in MDCK cells (58). Similarly, inhibition of tyrosine phosphatases by sodium pervanadate resulted in a decrease in TER in both MDCK cells and brain endothelial cells (59).
The regulation of tight junction-associated proteins by cytokines is not well understood. We recently demonstrated that immune modulators regulate tight junction formation in human intestinal epithelial cells by regulating the expression of the claudin family of proteins (10). Furthermore, recent observations demonstrated that claudin-1 and ZO-1 isolate in Triton X-100-insoluble membrane compartments during tight junction formation in Ras-transformed MDCK cells (50).
Our data indicate that claudin-1 and claudin-2 may localize into different, partially overlapping, cellular compartments within T-84 cells and, therefore, may have distinct functions in the regulation of cellular junctions. Whereas claudin-1 was mostly detected in Triton X-100-soluble protein fractions, claudin-2 was highly enriched in tight junctional protein fractions characterized by the expression of hyperphosphorylated occludin in T-84 cells. Both proteins were recruited into Triton X-100-insoluble protein fractions upon induction of tight junction formation in T-84␤96 cells by IL-15. However, claudin-1 was recruited to the apical pole of cell-cell contacts, whereas claudin-2 positive staining was observed in a larger lateral apical membrane compartment.
The enhanced expression of claudins in tight junctions by IL-15 was dependent on the presence of the IL-2R␤ subunit. Claudins are a multigene family consisting of at least 24 members (60). Claudin-1 and claudin-2 have the ability to induce the formation of networks of strands and grooves at cell-cell contact sites when introduced into fibroblasts lacking tight junctions (11), and their increased expression correlates with the tightening of the paracellular barrier in human intestinal epithelial cells (10). Increasing evidence suggests that claudin expression is tissue type-specifically regulated (61). Expression of claudin proteins in primary tissues has been demonstrated for claudin-1, -2, -3, -4, and -8 in kidney and liver tissues (11,62). Claudin-5 has been demonstrated to be a component of tight junctions in endothelial cells (63), and claudin-11 expression has been demonstrated in tight junctions of sertoli cells and brain myelin sheets (61,63). Furthermore, the kidney epithelial cell line MDCK preferably expresses claudin-1 and claudin-4 but little claudin-2 and claudin-3 (64), whereas the human intestinal epithelial cell line T-84 is able to express both claudin-1 and claudin-2 in tight junctional complexes (10).
The high TER induced by IL-15 in T-84␤96 cells correlated with up-regulation of membrane-associated hyperphosphorylated occludin by IL-15. Occludin is a transmembrane protein within tight junctions (12), and it plays an essential role in the formation of the tight junction paracellular permeability barrier (65,66). Formation of tight junctions in MDCK cells has been shown to correlate with serine/threonine as well as tyrosine phosphorylation of occludin, resulting in a number of occludin proteins migrating as 70 -82-kDa bands on SDS-polyacrylamide gel electrophoresis (50 -52). Furthermore, highly phosphorylated occludin was found to be selectively concentrated at tight junctions, whereas less or nonphosphorylated occludin localized within the cytoplasm and to the basolateral membrane of MDCK cells and T-84 cells (49,51,52). How phosphorylation of occludin is regulated and which kinases are involved in the activation process leading to the assembly of tight junctions is not known. Among the analyzed tight junction-associated proteins, only phosphorylated occludin was significantly up-regulated after 1 day of IL-15 stimulation in T-84␤96 cells. At this time point T-84␤96 cells demonstrated the highest difference in TER in comparison with the parental T-84 cells, suggesting that phosphorylation of occludin may tighten the paracellular barrier. Alternatively, additional unknown components of tight junctions may be regulated by IL-15 in T-84␤96 cells. Together, our data suggest that IL-15 may be able to regulate intestinal epithelial cell barrier function through the combined regulation of occludin expression and phosphorylation.
Whereas the regulation of claudins and occludin by IL-15 was only observed in the T-84␤96 cell line, the regulation of the membrane association of ZO-1 and ZO-2 was independent of the IL-2R␤ subunit. ZO-1 and ZO-2 interact with occludin and claudins through their PDZ domains. These interactions are predicted to stabilize a network of tight junctional protein complexes and the cytoskeleton (19 -22). Our data demonstrate that ZO-1 and ZO-2 have a different cellular distribution in T-84 cells. ZO-1 was enriched together with claudin-2 and hyperphosphorylated occludin in Triton X-100-insoluble protein fractions, whereas ZO-2 was detected at high levels in the soluble protein fraction, which contained most of claudin-1. It remains to be determined whether the differential expression of claudin-1 together with ZO-2 and claudin-2 with ZO-1 is due to direct interaction of these proteins.
After replating of T-84 cells, expression of the initial high protein levels of ZO-1 and ZO-2 decreased over a 3-day period, whereas in the presence of IL-15, the recruitment of both proteins into the Triton X-100-insoluble membrane fraction was enhanced. The recruitment and phosphorylation of ZO-1 in Triton X-100-insoluble protein fractions during the formation of tight junctions has been recently demonstrated in MDCK cells (50). In contrast, the role of ZO-2 in tight junction formation is not well characterized. Our data indicate that T-84 may express ZO-2 in cellular compartments distinct from tight junc-tions. In these compartments, ZO-2 may have additional functions, possibly in the regulation of adherens junction in intestinal epithelial cells, because ZO-1 as well as ZO-2 can localize to adherence junctions in a number of different cell types (21). The regulation of ZO-1 and ZO-2 by IL-15 was observed in both parental T-84 cells and the IL-2R␤ subunit expressing T-84 cell lines and therefore may be mediated by the IL-15R␣ and ␥c alone as proposed for the IL-15 receptor complex on mast cells (67). Alternatively, IL-15 or its receptor complex may be able to interact with additional receptors expressed in intestinal epithelial cells. One alternative receptor for IL-15 may include the epithelial cell kinase (Eck) receptor, which can be activated not only by its specific ligand B61 but also by different additional mediators, which include IL-15 (68). Activation of the Eck correlates with tightening of the paracellular barrier in the human intestinal epithelial cell line Caco-2 (68). The IL-15mediated regulation of ZO-1 and ZO-2 was independent of the IL-2R␤ subunit, whereas the phosphorylation of occludin and enhanced membrane association of claudins by IL-15 required the presence of the IL-2R␤ subunit. Therefore, the recruitment of intracellular zonula occludin proteins into tight junctions may be able to regulate the epithelial barrier function independently from the induction of the expression of transmembrane tight junctional proteins. Moderate increases in TER may be achieved by the enhanced recruitment of ZO-1 and ZO-2 into tight junctional complexes, but strong induction of intestinal barrier function may require the parallel induction of occludin and claudin protein expression. The IL-2R␤ subunitindependent regulation of recruitment of ZO-1 and ZO-2 into Triton X-100-insoluble membranes by IL-15 may explain why untransfected T-84 cells are able to up-regulate their TER (9).
IL-15 and IL-15R␣ mRNA expression has been detected in numerous tissues, many of which are not sites of immune responses, indicating the potential for additional nonimmune functions (26). In inflammatory bowel disease, elevated IL-15 expression derived from peripheral blood mononuclear cells, intestinal macrophages, or intestinal epithelial cells has been associated with disease stages (31,32,69). Furthermore, IL-15 mRNA is expressed at high levels in normal and diseased mouse intestine (70) and up-regulated during intestinal inflammation in IL-2-deficient mice. Our data indicate that IL-15, in addition to its ability to regulate intestinal NK-cells, dendritic cells, and intraepithelial lymphocyte populations (37), is also a cytokine able to regulate the intestinal barrier function.
The broad pleiotropic effects of IL-15 on multiple tissues and cell types outside of the classical immune system are unusual for a cytokine, and further understanding of how IL-15 may affect nonimmune cells could provide information relevant to the understanding of the role of IL-15 in the regulation of inflammation and wound healing.
Together, our experiments support a model in which IL-15 functions to regulate the expression and membrane association of tight junction-associated proteins during the formation of the intestinal barrier. Distinct IL-15-initiated signaling pathways can mediate either the recruitment of MAGUK family members into tight junctional complexes or the enhanced expression and membrane association of transmembrane tight junctional proteins.