IL-15 is crucial for the development of intestinal intraepithelial lymphocytes (IEL) and delivery is mediated by a unique mechanism known as trans-presentation. Parenchymal cells have a major role in the trans-presentation of IL-15 to IELs, but the specific identity of this cell type is unknown. To investigate whether the intestinal epithelial cells (IEC) are the parenchymal cell type involved, a mouse model that expresses IL-15Rα exclusively by the IECs (Villin/IL-15Rα Tg) was generated. Exclusive expression of IL-15Rα by the IECs restored all the deficiencies in the CD8αα+TCRαβ+and CD8αα+TCRγδ+ subsets that exist in the absence of IL-15Rα. Interestingly, most of the IEL recovery was due to the preferential increase in Thy1low IELs, which compose a majority of the IEL population. The differentiation of Thy1highCD4CD8 thymocytes into Thy1CD8αα IELs was found to require IL-15Rα expression specifically by IECs and thus, provides evidence that differentiation of Thy1low IELs is one function of trans-presentation of IL-15 in the intestines. In addition to effects in IEL differentiation, trans-presentation of IL-15 by IECs also resulted in an increase in IEL numbers that was accompanied by increases in Bcl-2, but not proliferation. Collectively, this study demonstrates that trans-presentation of IL-15 by IECs alone is completely sufficient to direct the IL-15-mediated development of CD8αα+ T cell populations within the IEL compartment, which now includes a newly identified role of IL-15 in the differentiation of Thy1low IELs.

Intestinal intraepithelial lymphocytes (IEL)3 are a unique population of T cells residing within the gut epithelium as these cells is mostly CD8+ T cells, many of which express the CD8αα homodimer (1, 2). Additionally, a large proportion of IELs express the γδ form of the TCR. In general, three main populations of CD8+ T cells can be found within the IEL compartment of the small intestine. One population is comprised of conventional CD8αβ+TCRαβ+ cells that develop in the thymus and home to the gut epithelium in response to activation in the periphery (3, 4); these memory CD8 T cells likely provide protection against microorganisms. The other two populations are made up of TCRαβ+ and TCRγδ+ IELs, which are considered unconventional because they express the CD8αα homodimer (5) and their origin and functions have been unclear. Previous studies have provided evidence that CD8αα IELs function in preserving the integrity of the epithelium and maintaining intestinal homeostasis (6, 7, 8). CD8αα IELs are also unique among T cells as a majority of these cells express low levels of Thy1 (9). Although the relevance of low Thy1 expression is not known, it likely relates to their unconventional development and distinctive functions.

The development of unconventional CD8αα+TCRαβ+ IELs has long been controversial, with a theory that these cells are generated extrathymically; however, more recent evidence has demonstrated the existence of CD8αα+TCRαβ+ IEL precursors in the thymus (10, 11). Similar to CD8αα+TCRαβ+ IELs, the development of CD8αα+TCRγδ+ IELs is not well defined and evidence exists for both thymic dependent and independent pathways (12). Despite an incomplete understanding in the development and function of the CD8αα+TCRαβ+ and CD8αα+TCRγδ+ IEL subsets, it is evident that the pleiotropic cytokine, IL-15, is crucial for their development and maintenance. IL-15 belongs to the four α-helix bundle cytokine family that uses three receptor subunits: IL-15Rα, IL-2Rβ, and γC. In the absence of IL-15, IL-15Rα, or IL-2Rβ, CD8αα+TCRαβ+ and CD8αα+TCRγδ+ IEL numbers are severely reduced (13, 14, 15). In vitro, IL-15 has been shown to enhance survival and proliferation of CD8αα+TCRαβ+ and CD8αα+TCRγδ+ IELs (16, 17). Additionally, it has been suggested that IL-15 can promote TCR Vγ5 rearrangement in TCRγδ+ precursors in the absence of IL-7 (18). So, whereas IL-15 can act directly on IELs to modulate survival, proliferation, and differentiation events, the main function of IL-15 in IEL development in vivo is uncertain.

Because of the unidentified roles of IL-15 in IEL development and the dual location of IEL development, the point in IEL development in which IL-15 functions remains unclear. This question may now be better addressed if one considers that IL-15 acts locally through a mechanism called trans-presentation. As first illustrated by Dubois et al. (19), IL-15 is expressed on the cell surface bound to IL-15Rα which can stimulate neighboring cells expressing the IL-2Rβ/γC subunits. Evidence that trans-presentation is used in vivo is supported by work demonstrating that IL-15Rα expression is required by cells in the environment rather than by IELs, memory CD8 T cells, or NK cells (20, 21, 22, 23). For development of both CD8αα+TCRαβ+ and CD8αα+TCRγδ+ IELs, parenchymal cell expression of IL-15Rα and IL-15, but not IL-2Rβ, is specifically required (21). In contrast, these IELs need to express IL-2Rβ, but not IL-15Rα, for their development (21), indicating that IL-15 acts directly on the IELs. In studies addressing the possibility that thymic parenchymal cells may trans-present IL-15 to IEL precursors, Lai et al. (24) used thymectomies along with thymic transplants to demonstrate that expression of IL-15 in the thymus was not required for development of CD8αα+ IELs. Altogether, these studies clearly show that parenchymal cells outside of the thymus trans-present IL-15 for IEL development, but did not precisely identify the cells providing IL-15 signals to IELs.

Because trans-presentation requires cell-cell contact and intestinal epithelial cells (IEC) are adjacent to IELs and express IL-15 (25), we hypothesized that IECs are the parenchymal cell type trans-presenting IL-15 to the IELs. To this end, we developed a mouse model that expresses IL-15Rα exclusively by the IECs to address the role of IL-15 trans-presented by IECs. By using this model, we demonstrate that sole expression of IL-15Rα by IECs restores the deficiencies in the CD8αα+TCRαβ+ and CD8αα+TCRγδ+ subsets present in the intestinal epithelium of IL-15Rα−/− mice. Furthermore, we identified that the main effect of transgenic (Tg) IL-15Rα expression was in the appearance of Thy1lowCD8α+TCR+ IELs.

C57BL/6J mice were purchased from The Jackson Laboratory and C57BL/6J Ly5.2 mice were purchased from National Cancer Institute. IL-15Rα−/− mice were generously provided by Averil Ma and backcrossed to C57BL/6 mice 15 generations (14). The full length murine IL-15Rα cDNA (from nucleotide 49–932 of precursor IL-15Rα sequence U22339) was cloned by PCR using the primers 5′-gtcactgctggggacaattg-3′ and the 5′-ggatccctaactgcccttgtatcttc-3′ from cDNA generated from C57BL/6J spleen RNA using TA Topo cloning kit (Invitrogen). A BamHI site was added to the 3′ end of the IL-15Rα cDNA during PCR cloning. The IL-15Rα cDNA was then subcloned upstream of an internal ribosome entry site-enhanced GFP (IRES-EGFP) reporter by ligating into the pIRES2-EGFP vector (BD Clonetech) using EcoRI and BamHI. Cloning of the IL-15Rα-IRES-EGFP insert downstream of the Villin promoter required the introduction of a BsiWi site on the 5′ end of the IL-15Rα cDNA and a MluI site on the 3′ end of the EGFP. This was achieved by mutagenesis using QuickChange II XL Site-Directed Mutagenesis Kit (Stratagene). Following mutagenesis, the entire IL-15Rα-IRES-EGFP region was sequenced and then subcloned into the pBS KS Villin MES SV40 poly vector (provided by Sylvie Robine; Refs. 26, 27) using BsiWi and MluI. The final 11.2 kb Villin-IL-15Rα-IRES-EGFP fragment was removed from the vector backbone by SalI digestion and used for microinjection into IL-15Rα+/− C57BL/6J pronuclei for generation of transgenic mice. Microinjection was performed by the Genetically Engineered Mouse Facility at the University of Texas M.D. Anderson Cancer Center. Tg-positive mice were identified by PCR of tail DNA and confirmed by IL-15Rα staining and GFP expression on IECs. Tg-positive mice were bred to IL-15Rα−/− background. All mice were maintained under specific pathogen-free conditions at the University of Texas M.D. Anderson Cancer Center in accordance with the Institutional Animal Care and Use Committee guidelines.

IELs and IECs were isolated from mice between 6- and 8-wk-old as previously described (28, 29). In brief, small intestines were cut longitudinally and then into 5-mm pieces. The pieces were washed in Ca2+ and Mg2+-free (HBSS with 1 mM HEPES and 2.5 mM NaHCO3) and stirred at 37°C in buffer containing 10% FBS and 1 mM dithioerythritol (Calbiochem) for IELs or 0.5 mM DTT (Sigma-Aldrich) for IECs. The cells were then centrifuged on a 44–67% or 25–44% Percoll (Amersham Biosciences) gradient to enrich for IELs or IECs, respectively. IELs from the colon were isolated in the same manner as described above for small intestines. For analysis of peripheral lymphocytes, single-cell suspensions were made by homogenizing spleens, peripheral and mesenteric lymph nodes, and thymuses using frosted glass slides in HBSS (containing HEPES, l-glutamine, penicillin, and gentamicin sulfate) and filtered through Nitex. RBC were lysed with Tris-ammonium chloride.

Isolated lymphocytes were resuspended in PBS, 0.2% BSA, 0.1% NaN3 (FACS buffer) at a concentration of 1–2 × 107 cells/ml followed by incubation with 100 μl of properly diluted mAb at 4°C for 20 min. The following Abs were used for flow cytometric analysis and were purchased from BD Biosciences, unless otherwise noted: CD45 (30-F11)-Pacific Blue (Biolegend), CD8α (Ly-2)-PerCp-Cy5.5, γδTCR (GL3)-PE or -FITC, Thy1.2 (53–2.1)-allophycocyanin or –PE-Cy7 (eBioscience), αβTCR (H57)-APC-Alexa750 (eBioscience) or -APC, CD8β (Ly3.2)-PE or –FITC, CD44 (IM7)-allophycocyanin or Pacific Blue, CD4 (L3T4)-PE, and CD122 (TM-β1)-PE, IL-15Rα was detected with goat anti-IL-15Rα-biotin (R&D Systems) followed by streptavidin-allophycocyanin (Jackson ImmunoResearch Laboratories). Background staining was determined by staining analogous populations from IL-15Rα−/− mice and with a biotinylated Ig control (Jackson ImmunoResearch Laboratories). Biotin anti-Vγ5 mAb was previously described (30) and visualized with streptavidin-allophycocyanin-Alexa Fluor 750 (Caltag Laboratories). After staining, cells were washed and fixed in 3% paraformaldehyde buffer. For intracellular staining of Bcl-2, a FITC conjugated Bcl-2 Ab reagent set (BD Biosciences) was used according to manufacturer’s instructions. To determine BrdU incorporation, BrdU (2 mg/mouse) was injected i.p. 12 h before analysis; isolated cells were stained with conjugated BrdU Ab set (BD Biosciences) according to manufacturer’s instructions and analyzed immediately by flow cytometry. All cells were acquired using a LSRII (Becton Dickinson) and analyzed using FlowJo software (Tree Star). Lymphocyte percentages and total cell numbers were calculated and evaluated by using the Student’s t test. Values of p < 0.05 were considered statistically significant.

Small intestines were fixed in 4% paraformaldehyde and then in 30% sucrose before being embedded in OCT. Six-micrometer sections of small intestine were blocked for endogenous biotin by treating with 0.001% avidin followed by 0.001% biotin in PBS. Sections were then stained with anti-mouse IL-15Rα biotin (R&D Systems) or anti-mouse goat Ig biotin (Jackson ImmunoResearch Laboratories). Staining was visualized using strepavidin-Cy5 (Jackson ImmunoResearch Laboratories) and analyzed using a Leica SP2 SE confocal microscope. For H&E staining, pieces from all three regions of the intestine (duodenum, jejenum, and ileum) were embedded in OCT, sectioned at 6 μm, stained, and analyzed using an Olympus BX41 microscope.

Intestinal IECs were lysed in RIPA lysis buffer (Santa Cruz Biotechnology) and protein (150,000 cells/lane) was separated on a 4–20% Tris-glycine gel (Invitrogen), transferred to a polyvinylidene fluoride membrane (Immobilon-P; Millipore), and analyzed by Western blot using anti-IL-15Rα (AZ-12; Santa Cruz Biotechnology) as the primary Ab. The secondary Ab used was anti-rat biotin (Santa Cruz Biotechnology), followed by streptavidin-HRP (Pierce). Signals were detected using the Supersignal West Pico Chemiluminescent Substrate System (Pierce).

Thymocytes from CD45.1+ C57BL/6J mice were purified by incubating in anti-CD4 (GK1.5) and anti-CD8α (2.43) coated Dynal beads (Dynal Mouse T cell Negative Isolation Kit; Invitrogen) before separation in a magnetic field. The resulting double negative thymocytes were then incubated with CD8α (Ly-2)-PerCp-Cy5.5, Thy1.2 (53–2.1)-APC, and CD4 (L3T4)-PE (all from BD Pharmingen) and sorted for CD4CD8αThy1high thymocytes using a FACSAria (BD Biosciences). IELs from CD45.1+C57BL/6J mice were stained with CD8α (Ly-2)-PerCp-Cy5.5, Thy1.2 (53–2.1)-APC, CD8β (Ly3.2)-PE, and CD4 (L3T4)-FITC (all from BD Pharmingen) and sorted for CD8αα+Thy1high and CD8αα+Thy1low IELs. The donor CD4CD8αThy1high thymocytes (∼750,000 cells/mouse), CD8αα+Thy1high IELs (∼100,000 cells/mouse), or CD8αα+Thy1low IELs (∼460,000 cells/mouse) were then injected i.v. into sublethally irradiated (750 cGy) Villin/IL-15Rα Tg, wild-type (Wt), and IL-15Rα−/− mice (all CD45.2+). Among the CD4CD8 Thy1high thymocytes, ∼25% were TCRαβ+ and 7.5% were TCRγδ+; IELs and spleens from recipient mice were analyzed 2–3 wk later.

To develop a model where IL-15Rα is exclusively expressed by the IECs, murine IL-15Rα cDNA and an IRES-EGFP reporter sequence was cloned downstream of the villin promoter and used to generate four Tg founder mice; these Villin/IL-15Rα-GFP Tg mice were then backcrossed to IL-15Rα−/− mice and will be referred from here on as Villin/IL-15Rα Tg mice. Each of the four founder lines, before and after being backcrossed to the IL-15Rα−/− background were found to be generally healthy and did not display any defects in growth, development, or reproduction. In histological analysis, staining with IL-15Rα Ab was observed specifically in the IECs of the small intestine from Villin/IL-15Rα Tg mice, but was not observed using an Ig control (Fig. 1,A). Interestingly, IL-15Rα staining appeared to be more concentrated along the basolateral cell surface. Green fluorescence, due to the GFP reporter, was also observed intracellularly and appeared evenly distributed within the epithelial cells (Fig. 1,A). GFP could not be detected in thymus or spleen sections (data not shown), confirming previous studies demonstrating specificity of the villin promoter to the intestinal epithelium (31, 32, 33, 34, 35). Expression of both GFP and IL-15Rα was also detected in CD45-negative cells in Villin/IL-15Rα Tg mice by flow cytometric analysis (Fig. 1,B). In side-by-side comparisons, the four founder lines expressed similar levels of IL-15Rα and GFP by IECs (data not shown). Detection of IL-15Rα by flow cytometry was insufficient in discerning differences in IL-15Rα expression in Wt and IL-15Rα−/− mice, which was likely due to the high level of background fluorescence of IECs. As such, IL-15Rα expression of IECs was examined by Western blot analysis and was indeed expressed by IECs in Wt and Tg mice but not IL-15Rα−/− mice (Fig. 1 C). Overall, Villin/IL-15Rα Tg mice specifically express IL-15Rα by IECs, which correlates with GFP expression.

FIGURE 1.

IL-15Rα and GFP are expressed in the intestinal epithelium of Villin/IL-15Rα Tg mice. A, IL-15Rα (red) and GFP (green) was detected in 6-μm sections of small intestine of Villin/IL-15Rα Tg mice. Top row, IL-15Rα expression as detected using anti-IL-15Rα Ab-biotin followed by visualization with streptavidin-Cy5. Bottom row, Control staining using goat Ig biotin. B, Histograms in left panel show GFP expression by IECs isolated from Wt and Villin/IL-15Rα Tg mice. Histograms in right panel show cell surface IL-15Rα expression by IECs isolated from the indicated mice. Both panels show fluorescence after gating on CD45-negative cells (IECs). C, Western blot analysis. Protein from intestinal IECs from Wt, IL-15Rα−/−, and Villin/IL-15Rα Tg mice were analyzed by Western blot using anti-IL-15Rα Ab. D, Phenotype of T cells isolated from spleen, mesenteric lymph node (MLN), and peripheral lymph nodes (PLN). Cells were isolated from the indicated mice and stained for cell surface markers to identify proportions in T cell populations. Left panel, Representative dot plots of lymphocytes stained with CD4 and CD8 from the indicated tissues. Right panel, Staining of CD44 after gating on CD8+ cells; marker designates the proportion of memory-phenotype (CD44high) cells. E, Proportion of peripheral NK cells. Lymphocytes from PLN, MLN, and spleen from the indicated mice were stained with NK1.1 and CD3. Box denotes the NK cell population as defined by CD3NK1.1+ staining and numbers represent the average percent of NK cells present.

FIGURE 1.

IL-15Rα and GFP are expressed in the intestinal epithelium of Villin/IL-15Rα Tg mice. A, IL-15Rα (red) and GFP (green) was detected in 6-μm sections of small intestine of Villin/IL-15Rα Tg mice. Top row, IL-15Rα expression as detected using anti-IL-15Rα Ab-biotin followed by visualization with streptavidin-Cy5. Bottom row, Control staining using goat Ig biotin. B, Histograms in left panel show GFP expression by IECs isolated from Wt and Villin/IL-15Rα Tg mice. Histograms in right panel show cell surface IL-15Rα expression by IECs isolated from the indicated mice. Both panels show fluorescence after gating on CD45-negative cells (IECs). C, Western blot analysis. Protein from intestinal IECs from Wt, IL-15Rα−/−, and Villin/IL-15Rα Tg mice were analyzed by Western blot using anti-IL-15Rα Ab. D, Phenotype of T cells isolated from spleen, mesenteric lymph node (MLN), and peripheral lymph nodes (PLN). Cells were isolated from the indicated mice and stained for cell surface markers to identify proportions in T cell populations. Left panel, Representative dot plots of lymphocytes stained with CD4 and CD8 from the indicated tissues. Right panel, Staining of CD44 after gating on CD8+ cells; marker designates the proportion of memory-phenotype (CD44high) cells. E, Proportion of peripheral NK cells. Lymphocytes from PLN, MLN, and spleen from the indicated mice were stained with NK1.1 and CD3. Box denotes the NK cell population as defined by CD3NK1.1+ staining and numbers represent the average percent of NK cells present.

Close modal

To provide evidence that functional activities of IL-15 are not present in the periphery of Villin/IL-15Rα Tg mice, lymphocytes were analyzed for deficiencies in CD8 T cells, memory-phenotype CD8 T cells (CD44high), and NK cells, which are characteristic of an IL-15Rα deficiency (14). Lymphocytes isolated from the spleen and peripheral and mesenteric lymph nodes of Villin/IL-15Rα Tg mice had a similar phenotype to those of IL-15Rα−/− mice, where total CD8 T cells and NK cells were decreased and displayed a deficiency in memory phenotype (CD44high) CD8 T cells (Fig. 1, D and E). The composition of lymphocytes in the spleens of Villin/IL-15Rα Tg mice crossed to the Wt background (Villin/IL-15Rα Tg (Wt)) did not appear different from normal Wt mice (supplemental data).4 The phenotype of the lymphocytes in the secondary lymphoid tissues of Villin/IL-15Rα Tg mice is biological evidence that IL-15Rα is not expressed outside the intestine.

Upon isolation of cells from the IEL compartment, the total number of CD45+ and CD45-negative cells isolated from the intestinal epithelium was compared among the different groups of mice. Among multiple experiments, the total number of CD45+ cells was increased 3.3-fold in the Villin/IL-15Rα Tg mice compared with the Wt mice (Fig. 2,A). Conversely, there was only a slight increase in the number of CD45-negative cells in the Villin/IL-15Rα Tg mice compared with Wt (Fig. 2,A). This was reflected by an increased proportion of CD45+ cells from 28% ± 2.4 present in Wt mice to 48% ± 0.8 in Villin/IL-15Rα Tg mice (Fig. 2,B). As increased cell numbers could be due to an increase in surface area, the general structure of the intestinal villi and crypts was examined in all three regions of the small intestines (duodenum, jejenum, and ileum). Between the three groups of mice, the length of villi and depths of crypts were similar (Fig. 2 C). Therefore, IL-15Rα expression by IECs enhances the overall number of lymphocytes in the intestinal epithelium.

FIGURE 2.

IEC expression of IL-15Rα increases lymphocyte numbers in the intestinal epithelium. A, Total cell numbers (IELs and IECs) were counted using a hemocytometer and trypan blue exclusion. Total IEL and IEC numbers were determined from the percentage of CD45+ or CD45neg, respectively. Average of IEL and IEC cell numbers from four independent experiments ± SE. ∗, p < 0.05 compared with Wt and Villin/IL-15Rα Tg mice; ∗∗, p < 0.001 compared with Wt and IL-15Rα−/− mice. B, Percent of CD45+ cells among all cells isolated from intestinal epithelium (average of four independent experiments ± SE). C, Six-micrometer sections of ileum stained for H&E.

FIGURE 2.

IEC expression of IL-15Rα increases lymphocyte numbers in the intestinal epithelium. A, Total cell numbers (IELs and IECs) were counted using a hemocytometer and trypan blue exclusion. Total IEL and IEC numbers were determined from the percentage of CD45+ or CD45neg, respectively. Average of IEL and IEC cell numbers from four independent experiments ± SE. ∗, p < 0.05 compared with Wt and Villin/IL-15Rα Tg mice; ∗∗, p < 0.001 compared with Wt and IL-15Rα−/− mice. B, Percent of CD45+ cells among all cells isolated from intestinal epithelium (average of four independent experiments ± SE). C, Six-micrometer sections of ileum stained for H&E.

Close modal

To determine whether exclusive IL-15Rα expression by IECs affects the composition of IEL populations, IELs were isolated from Villin/IL-15Rα Tg mice and compared with that of Wt and IL-15Rα−/− mice. Among CD45+ IELs, the frequency of CD8α+TCRγδ+ IELs was reduced in IL-15Rα−/− mice compared with Wt mice by ∼75% (Fig. 3,A) as previously described (13). In contrast, the IELs from Villin/IL-15Rα Tg mice contained a similar or higher proportion of CD8α+TCRγδ+ cells than Wt mice (Fig. 3,A). As CD8α+TCRγδ+ IELs contain subsets with different Vγ usage and varying levels of Thy1 expression, the proportion of these subsets were assessed in the different groups of mice. Whereas CD8α+TCRγδ+ IELs in Wt mice predominately use Vγ5 and have a high frequency of Thy1low subset, the Vγ5+ and Thy1low TCRγδ+ IELs were virtually absent in IL-15Rα−/− mice (Fig. 3,A). In the Villin/IL-15Rα Tg mice, the proportion of Vγ5+ and Thy1low TCRγδ+ IELs was restored to normal (Fig. 3,A). The differences in frequency of TCRγδ+ IEL subsets correlated to changes in absolute numbers (Fig. 3,B). Although not previously described, the total number of both Vγ5+ and Vγ5Thy1lowCD8α+TCRγδ+ IELs was severely reduced by ∼95% in IL-15Rα−/− mice compared with Wt mice (p < 0.01 and p < 0.05, respectively, Fig. 3,B). In contrast, the number of Thy1high CD8α+TCRγδ+ IEL subsets was not significantly different between Wt and IL-15Rα−/− mice and was independent of Vγ5 usage (Fig. 3,B). In the Villin/IL-15Rα Tg mice, the total numbers of all CD8α+TCRγδ+ IEL subsets were dramatically increased (p < 0.05; Fig. 3 B), which was largely due to the increased number of total lymphocytes that exist in the intestinal epithelium of Tg mice. Overall, exclusive expression of IL-15Rα by IECs restored the deficiency in Thy1lowCD8α+TCRγδ+ IELs that occurs in the absence of IL-15Rα expression as well as enhanced the numbers of all CD8α+TCRγδ+subsets.

FIGURE 3.

IL-15-mediated IEL development is recovered in Villin/IL-15Rα Tg mice. A, Flow cytometric analysis of TCRγδ+ IEL subsets. Top panel, Proportion of CD8α+TCRγδ+ cells among CD45+ IELs. Bottom panel, Staining for Thy1 and Vγ5 expression after gating on CD8α+TCRγδ+ cells. B, Absolute numbers of the subpopulations of TCRγδ+ IELs isolated from each group of mice (average of four independent experiments, n = 10–12 mice/group). Circles in top row indicate gating of analysis in bottom row. Error bars represent SE. C, Flow cytometric analysis of TCRαβ+ IEL subsets. Top panel, Proportion of CD8α+TCRαβ+ cells among CD45+ IELs. Bottom panel, Staining for Thy1 and CD8β expression after gating on CD8α+TCRαβ+ cells. D, Absolute numbers of the subpopulations of TCRαβ+ IELs isolated from each group of mice (average of four independent experiments, n = 10–12 mice/group). Error bars represent SE. E, Flow cytometric analysis of IELs isolated from the colon. First row shows proportion of TCRαβ+ and TCRγδ+ IELs among CD45+ IELs. Remaining rows have gated populations indicated on the left. F and G, Absolute numbers of TCRγδ+ (F) and TCRαβ+ (G) IELs from the colon. Error bars represent S.D. ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001 compared with Wt and IL-15Rα−/− mice.

FIGURE 3.

IL-15-mediated IEL development is recovered in Villin/IL-15Rα Tg mice. A, Flow cytometric analysis of TCRγδ+ IEL subsets. Top panel, Proportion of CD8α+TCRγδ+ cells among CD45+ IELs. Bottom panel, Staining for Thy1 and Vγ5 expression after gating on CD8α+TCRγδ+ cells. B, Absolute numbers of the subpopulations of TCRγδ+ IELs isolated from each group of mice (average of four independent experiments, n = 10–12 mice/group). Circles in top row indicate gating of analysis in bottom row. Error bars represent SE. C, Flow cytometric analysis of TCRαβ+ IEL subsets. Top panel, Proportion of CD8α+TCRαβ+ cells among CD45+ IELs. Bottom panel, Staining for Thy1 and CD8β expression after gating on CD8α+TCRαβ+ cells. D, Absolute numbers of the subpopulations of TCRαβ+ IELs isolated from each group of mice (average of four independent experiments, n = 10–12 mice/group). Error bars represent SE. E, Flow cytometric analysis of IELs isolated from the colon. First row shows proportion of TCRαβ+ and TCRγδ+ IELs among CD45+ IELs. Remaining rows have gated populations indicated on the left. F and G, Absolute numbers of TCRγδ+ (F) and TCRαβ+ (G) IELs from the colon. Error bars represent S.D. ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001 compared with Wt and IL-15Rα−/− mice.

Close modal

In the analysis of TCRαβ IELs, similar percentages of CD8α+TCRαβ+ IELs were present in each group of mice (i.e., Villin/IL-15Rα Tg mice, Wt, and IL-15Rα−/− mice) (Fig. 3,C); however, within this population, the CD8αα+TCRαβ+ IELs are the most heavily dependent on IL-15 (13, 14). Therefore, because CD8α+TCRαβ+ IELs contain CD8αβ+ as well as Thy1high and Thy1lowCD8αα+ cells, the effect of IL-15Rα on each of these CD8α+TCRαβ+ IELs subsets were examined (Fig. 3,C). In Wt mice, the CD8β IELs (CD8αα) make up a majority of the CD8α+TCRαβ+ IELs, with most of those cells expressing low levels of Thy1 (Fig. 3,C). The frequency of CD8αα+TCRαβ+ IELs was affected in IL-15Rα−/− mice with a preferential deficiency in the Thy1low IELs and an enhancement in the CD8αβ+ IELs (Fig. 3,C). Most significantly, the Villin/IL-15Rα Tg mice had normal percentages of all CD8α+TCRαβ+ IEL subsets (Fig. 3,C). Comparison of absolute numbers indicated that the Thy1low subset of CD8αα+ IELs was dramatically lost in the IL-15Rα−/− mice (97% decrease, p < 0.001), but the Thy1high CD8αα+ IELs were not deficient, similar to that observed for the CD8+TCRγδ+ IELs (Fig. 3,D). In Villin/IL-15Rα Tg mice, cell numbers of both Thy1low and Thy1high CD8αα+TCRαβ+ IELs were restored to levels beyond that found in Wt mice due to the increase in total lymphocyte numbers (p < 0.05 for Thy1low; Fig. 3,D). Surprisingly, the proportion and total numbers of CD8αβ+Thy1high cells was higher in IL-15Rα−/− and Villin/IL-15Rα Tg mice compared with Wt mice (p < 0.01, Fig. 3, C and D). Because an increase in CD8αβ+TCRαβ+ IELs in IL-15Rα−/− mice has not been previously described and cannot be presently explained, further investigation is required. Regardless, these findings show that IL-15Rα expression by IECs similarly recovers the defective development of Thy1lowCD8αα+TCRαβ+ IELs while enhancing the Thy1highCD8α+TCRαβ+ IELs.

The composition of lymphocytes in the IEL compartment was also examined in Villin/IL-15Rα Tg mice crossed to the Wt background. Overall, the proportion of CD8αα+TCRγδ+ subsets based on Vγ5 and Thy1 expression did not differ between Wt mice and Villin/IL-15Rα Tg (Wt) mice (supplemental data). In addition, no differences in the proportions of CD8αα+TCRαβ+ IELs were observed among the various groups of mice (supplemental data). Furthermore, in Villin/IL-15Rα Tg (Wt) mice, the lymphocytes in the spleen displayed a normal phenotype with no evidence that IELs migrated out of the intestines (supplemental data). This analysis demonstrates that transgenic expression of IL-15Rα by the IECs, in combination with normal endogenous expression of IL-15Rα, does not have an additional impact on IEL development.

Villin promoter expression, in addition to being active in the small intestine, is also active in the large intestine. As such, phenotypic analysis of IELs from the large intestine was also examined in this study. Similar to that observed in the small intestine, TCRγδ+ IELs were deficient in IL-15Rα−/− mice but were restored in theVillin/IL-15Rα Tg mice beyond that of Wt levels (Fig. 3, E and F). For the TCRαβ+ IELs, a deficiency was also present in the IL-15Rα−/− mice but to a greater degree in the colon than in the small intestine. In contrast to the TCRγδ+ IELs, transgenic expression of IL-15Rα was less effective in restoring the defect in CD8αα+TCRαβ+ IELs (Fig. 3, E and G). Upon further analysis of the TCRαβ+ and TCRγδ+ IELs, Thy1low subsets were preferentially affected by IL-15Rα expression as the Thy1low cells were absent in IL-15Rα−/− mice and regained when IL-15Rα was expressed by IECs. Therefore, IL-15Rα expression by IECs in the colon has similar effects on IEL development as in the small intestine by preferentially affecting Thy1low CD8αα+TCRαβ+ and TCRγδ+ IELs.

In an attempt to identify a mechanism by which IL-15Rα increases overall IEL numbers and preferentially restores Thy1low IELs, cell survival and proliferation were examined among IEL subsets. Because IL-15 mediates survival by regulating Bcl-2 (36), Bcl-2 expression in IELs was analyzed in each group of mice. Bcl-2 expression in Wt mice was higher than that of IL-15Rα−/− mice in all IEL subsets indicating that IL-15 normally regulates Bcl-2 levels in IELs (Fig. 4 A). In Villin/IL-15Rα Tg mice, expression of Bcl-2 was comparable to that of Wt mice in all IEL subsets, except the CD8αα+Thy1high subset, which had lower Bcl-2 levels than Wt mice but still higher than that observed in IL-15Rα−/− mice. The restored deficiencies in Bcl-2 expression in most IEL subsets suggest that trans-presentation of IL-15 by IECs helps maintain overall IEL survival. This up-regulation of Bcl-2 may result in a decrease in IEL turnover, thus causing the overall increase in IEL cell numbers observed in the Villin/IL-15Rα Tg mice.

FIGURE 4.

Effects of IL-15Rα expression on Bcl-2 expression and basal proliferation in IEL subsets. A, Bcl-2 expression measured by intracellular staining after gating on the indicated CD8α+TCRγδ+ IEL subsets (top row) or CD8α+TCRαβ+ IEL subsets (bottom row) isolated from indicated mice. Data are representative of two independent experiments (n = 2 mice/group/experiment). B and C, Mice (6 wk old) were given 2 mg of BrdU i.p. followed by analysis of BrdU incorporation 12 h later. B, Average percent of BrdU incorporated by TCRγδ IEL subsets. C, Average percent BrdU incorporated by TCRαβ IEL subsets. D, The expression of IL-2/15Rβ-chain was examined by staining with anti-CD122 Ab on IELs isolated from Wt (bold outline), Villin/IL-15Rα Tg mice (shaded histogram), or IL-15Rα −/− mice (thin outline). Data are representative of two independent experiments (n = 2 mice/group/experiment).

FIGURE 4.

Effects of IL-15Rα expression on Bcl-2 expression and basal proliferation in IEL subsets. A, Bcl-2 expression measured by intracellular staining after gating on the indicated CD8α+TCRγδ+ IEL subsets (top row) or CD8α+TCRαβ+ IEL subsets (bottom row) isolated from indicated mice. Data are representative of two independent experiments (n = 2 mice/group/experiment). B and C, Mice (6 wk old) were given 2 mg of BrdU i.p. followed by analysis of BrdU incorporation 12 h later. B, Average percent of BrdU incorporated by TCRγδ IEL subsets. C, Average percent BrdU incorporated by TCRαβ IEL subsets. D, The expression of IL-2/15Rβ-chain was examined by staining with anti-CD122 Ab on IELs isolated from Wt (bold outline), Villin/IL-15Rα Tg mice (shaded histogram), or IL-15Rα −/− mice (thin outline). Data are representative of two independent experiments (n = 2 mice/group/experiment).

Close modal

To determine whether effects of IEC expression of IL-15Rα were due to IL-15-mediated cell expansion, BrdU incorporation of IELs was measured after 12 h of BrdU treatment. In general, each of the CD8αα+TCRγδ+ and the CD8αα+TCRαβ+ IEL subsets from Villin/IL-15Rα Tg mice had a similar level of BrdU incorporation as Wt mice, with the exception of CD8αβ+TCRαβ+ cells, which actually had less BrdU incorporation than the Wt mice (Fig. 4, B and C). Surprisingly, proliferation was increased among Thy1low subsets regardless of TCR expression in IL-15Rα−/− mice compared with Wt mice (Fig. 4, B and C). Increased cell division in IL-15Rα−/− mice could be due to a compensatory mechanism that occurs when lymphocyte numbers are deficient or a result of defective development. In general, these findings suggest that trans-presentation of IL-15 does not specifically enhance the basal rate of IEL proliferation and thus, does not provide a mechanism for the increased number of lymphocytes observed in Villin/IL-15Rα Tg mice. In contrast to the predicted role of IL-15 on IEL proliferation, our data provides evidence that the absence of IL-15Rα results in a dysregulation of proliferation in Thy1low IELs.

Previous reports have demonstrated that expression of the IL-2/15Rβ-chain (CD122) correlates with the level of IL-15 responsiveness (37). As such, we determined whether the expression of IL-15Rα affects the expression level of CD122 among CD8α+ IEL subsets. In Wt mice, expression of CD122 was higher in Thy1low IELs than Thy1high IEL for both CD8αα+TCRαβ+ and TCRγδ+ subsets, suggesting that Thy1low IELs are more sensitive to IL-15 than Thy1high IELs (Fig. 4,D). Among CD8αβ+TCRαβ+ IELs, the CD122 expression was very low (data not shown) as previously described (38). In comparing the different groups of mice, CD122 expression was similar within the TCRγδ+ (regardless of Thy1 expression levels) and Thy1highCD8αα+ TCRαβ+ IELs (Fig. 4,D). Conversely, CD122 expression in Thy1lowCD8αα+TCRαβ+ IELs was decreased in the absence of IL-15Rα−/− but restored to normal levels in Villin/IL-15Rα Tg mice (Fig. 4 D). This finding suggests that Thy1lowCD8αα+TCRαβ+ IELs in IL-15Rα−/− mice failed to become adequately responsive to IL-15, which might explain the deficiency in this population.

Previous studies identified Thy1highCD4CD8TCRαβ+ and CD4CD8TCRγδ+ thymocytes as precursors for CD8αα+TCRαβ+ and CD8αα+TCRγδ+ IELs, respectively (11, 39, 40). To determine whether trans-presentation of IL-15 by IECs drives the differentiation of Thy1high CD4CD8 thymic IEL precursors into Thy1lowCD8αα+IELs, CD4CD8 Thy1high thymocytes were sorted from congenic Wt mice to a purity of >99% and injected into irradiated Villin/IL-15Rα Tg, Wt, and IL-15Rα−/− mice. Two to three weeks after transfer, the presence of donor cells in the IEL and spleen were compared among the three different recipients. In Wt hosts, donor-derived cells in the IEL were mostly TCRγδ+ (∼90–95%), suggesting that this thymocyte population has a greater potential for development into TCRγδ+ IELs (Fig. 5,A). This developmental potential or expansion of thymocytes into TCRγδ+ IELs was not observed in IL-15Rα−/− mice but was restored in Villin/IL-15Rα Tg mice (Fig. 5,A). Upon gating on the donor-derived CD8α+TCRγδ+ IELs (>95%), almost all the cells expressed low levels of Thy1 in Wt and Villin/IL-15Rα Tg mice while this population was predominately Thy1high in IL-15Rα−/− mice (Fig. 5,A). Among donor-derived CD8α+TCRαβ+ IELs, Thy1 expression was low in Wt mice and high in IL-15Rα−/−, similar to that observed for TCRγδ+ IELs. In contrast, only a small portion of CD8α+TCRαβ+ IELs were Thy1low in the Villin/IL-15Rα Tg mice, suggesting the TCRαβ+ thymic progenitors may not receive an adequate IL-15 signal from IECs. Regardless of the host, a small percent of donor-derived cells were detected in the spleen, which were a mixture of TCRαβ+ and TCRγδ+ cells and were all Thy1high (Fig. 5,B). In addition, all the donor-derived CD8+TCRαβ+ splenocytes expressed the CD8αβ heterodimer (Fig. 5 B). These data demonstrate that Thy1high CD4CD8 thymocytes, upon migration to the intestinal epithelium, preferentially differentiate into Thy1lowCD8+TCRγδ+ IELs through the signals delivered by intestinal epithelium expressing IL-15Rα. This does not exclude the possibility that intermediary stages of development may occur in the thymus prior to these cells migrating into the intestinal epithelium. In contrast, while Thy1high thymocytes can differentiate into Thy1lowTCRαβ+ IELs in an IL-15Rα-dependent fashion, this differentiation is only minimally influenced by trans-presentation of IL-15Rα by IECs.

FIGURE 5.

In vivo differentiation of thymic IEL precursors. A and B, CD4CD8αThy1high thymocytes were isolated from Wt mice (CD45.1+) and injected i.v. into sublethally irradiated (750 cGy) Villin/IL-15Rα Tg, Wt, and IL-15Rα−/− mice (CD45.2+). Recipient mice were analyzed 2–3 wk later. Flow cytometric plots show phenotype of donor-derived cells among IELs (A) and in the spleen (B) from the indicated mice. Gated populations are indicated on the left. CD8αα+Thy1high (C) or CD8αα+Thy1low (D) cells were sorted from the IELs of Wt mice (CD45.1+) and transferred into sublethally irradiated Villin/IL-15Rα Tg, Wt, and IL-15Rα−/− mice (all CD45.2+). Donor-derived cells present in the intestinal epithelium were analyzed for Thy1 expression 2–3 wk after transfer. Data are representative of three experiments.

FIGURE 5.

In vivo differentiation of thymic IEL precursors. A and B, CD4CD8αThy1high thymocytes were isolated from Wt mice (CD45.1+) and injected i.v. into sublethally irradiated (750 cGy) Villin/IL-15Rα Tg, Wt, and IL-15Rα−/− mice (CD45.2+). Recipient mice were analyzed 2–3 wk later. Flow cytometric plots show phenotype of donor-derived cells among IELs (A) and in the spleen (B) from the indicated mice. Gated populations are indicated on the left. CD8αα+Thy1high (C) or CD8αα+Thy1low (D) cells were sorted from the IELs of Wt mice (CD45.1+) and transferred into sublethally irradiated Villin/IL-15Rα Tg, Wt, and IL-15Rα−/− mice (all CD45.2+). Donor-derived cells present in the intestinal epithelium were analyzed for Thy1 expression 2–3 wk after transfer. Data are representative of three experiments.

Close modal

To examine differentiation of CD8αα+ IELs at a later stage of development, Thy1high CD8αα+ IELs were sorted from congenic mice; transferred into irradiated Villin/IL-15Rα Tg, Wt, and IL-15Rα−/− mice; and examined 2 wk later. In Villin/IL-15Rα Tg mice, donor-derived TCRαβ+ and TCRγδ+ IELs were both mostly Thy1low (Fig. 5,C). Down-regulation of Thy1 also occurred in TCRαβ+ IELs from Wt recipients, but was less dramatic in TCRγδ+ IELs (Fig. 5,C). In IL-15Rα−/− mice, both donor-derived TCRαβ+ and TCRγδ+ IELs remained Thy1high (Fig. 5,C). When Thy1lowCD8αα+ IELs were transferred, the donor-derived TCRαβ+ and TCRγδ+ IELs remained Thy1low in both Villin/IL-15Rα Tg and Wt recipients (Fig. 5,D). Interestingly, the transfer of Thy1low CD8αα+ IELs into IL-15Rα−/− recipients resulted in the up-regulation of Thy1 expression in these mice (Fig. 5 D). After all IEL transfers, no donor-derived cells were detected in the spleen (data not shown). These experiments suggest that IL-15 trans-presentation by IECs induces the down-regulation of Thy1 and maintains this phenotype in established CD8αα+ IELs.

Trans-presentation of IL-15 by parenchymal cells has been shown to drive IL-15-mediated development of IELs (21); however, the identity of the parenchymal cell type involved had yet to be determined. Because IECs have long been recognized as a source of IL-15 in the intestines and are adjacent to IELs, IECs were speculated to be a likely cell type trans-presenting IL-15 for IEL development. As studies have demonstrated that some IEL precursors are thymic-derived, have increased levels of CD122 expression, and are responsive to IL-15 (11), it is possible that this early stage of development may require IL-15 regulation by thymic parenchymal cells. In this current study, we demonstrate that expression of IL-15Rα solely by IECs was completely sufficient to restore the deficiencies in both the CD8αα+TCRαβ+ and CD8αα+TCRγδ+ subsets observed in the absence of IL-15Rα. These findings indicate that the development of IELs does not require the expression of IL-15Rα by the parenchymal cells of the thymus or in any other site outside of the intestine and are in agreement with the study conducted by Lai et al. (24). Whereas Lai et al. (24) demonstrated that regulation of CD8αα+ IELs does not require IL-15 in the thymus, this study did not precisely identify the cells providing IL-15 signals to IELs. Up until now, the point at which IL-15 becomes important for IEL development was not known, but our study provides evidence that IL-15 is not essential until the IEL precursors arrive in the intestinal epithelium.

In addition to demonstrating that IECs trans-present IL-15 for IEL development, our study also elucidates a novel role of IL-15 in IEL development. Previous studies have reported that IL-15 or IL-15Rα deficiencies preferentially affect the Vγ5 subset (21) and this correlated to the finding that rearrangements of Vγ5 were deficient in the absence of IL-15 (18). Vγ5 IELs are unusual T cells in that they are predominantly Thy1low, whereas the other Vγ1 and Vγ2 IELs are predominantly Thy1high. Upon a more thorough analysis, we found that Thy1highVγ5+ cells are present at normal levels in IL-15Rα−/− mice while the Thy1lowVγ5+ cells are clearly deficient. The same is true for Vγ5 (Vγ1+ and Vγ2+) IELs, where the small portion of Thy1lowVγ5 IELs are deficient in the absence of IL-15 signals but the Thy1highVγ5 IELs are not. Because this deficiency was also observed in the Thy1lowCD8αα+ subpopulation of the TCRαβ+CD8α+ IELs in IL-15Rα−/− mice, it suggests that the major effect of IL-15 on development is in the differentiation and expansion of Thy1low IELs and is independent of TCR rearrangement.

Although a number of models for IEL development have been proposed, most do not include alterations in Thy1 expression. As such, the relationship between the Thy1high and Thy1low cells is not clear. Two possibilities exist, where these cells could be two independently derived subsets or could be developmental intermediates of each other. Our results showing that Thy1highCD4CD8 thymocytes develop into Thy1lowTCR+CD8αα+ IELs indicates that the transition from Thy1high to Thy1low is along the same pathway of development. This hypothesis was further strengthened upon demonstrating that resident Thy1highCD8αα+ IELs also converted to Thy1lowCD8αα+ IELs. Most importantly, these conversion events were dependent on IEC expression of IL-15Rα, which provide evidence that IL-15 trans-presentation mediates this differentiation. Interestingly, the transition from Thy1highCD4CD8 thymocytes to Thy1lowTCRαβ+CD8αα+ IELs was only partially driven by IEC expression of IL-15Rα and could be an indication that additional IL-15 signals outside the intestinal epithelial compartment are required for this particular event. In contrast, once TCRαβ+CD8αα+ IELs are present in the IEL compartment, IL-15 trans-presentation was efficient in down-regulating Thy1 expression. Lastly, while we initially suspected that conversion of Thy1high to Thy1low IELs may be unidirectional, Thy1low IELs re-expressed Thy1 in the absence of IL-15Rα. Therefore, IL-15 signals delivered by IECs appear to down-regulate Thy1 expression and maintain this expression. One model of IEL differentiation did include the variable expression of Thy1; however, Thy1 was suggested to be down-regulated before TCR expression (12). Altogether, our findings support a model of IEL development in which Thy1high IEL precursors migrate to the intestinal epithelium, acquire expression of CD8αα, and is followed by the IL-15-mediated down-regulation of Thy1 expression.

Our model of restricted IL-15 trans-presentation also supports past findings that IEC-derived IL-15 is important for IEL survival by regulating Bcl-2 levels in vivo. In IL-15Rα−/− mice, all CD8α+ IEL subsets had lower levels of Bcl-2 expression than those found in Wt mice, indicating that IL-15Rα is important for maintaining Bcl-2 levels. Concurrently, Bcl-2 levels in IELs from Villin/IL-15Rα Tg mice were comparable to those found in Wt mice, demonstrating that IL-15 trans-presentation specifically by IECs was delivering a signal that regulated Bcl-2 levels. Unfortunately, we were unable to correlate the levels of Bcl-2 with indicators of IEL apoptosis, such as Annexin V staining. This inability to detect apoptotic cells may be due to the close proximity to the gut lumen, as apoptotic cells could be easily expulsed. Our results follow those observed by Nakazato et al. (36), which demonstrated that the transgenic expression of Bcl-2 in IL-15Rα−/− mice could restore the numbers of TCRγδ+ IELs; however, that study did not examine the contribution of transgenic Bcl-2 expression on Thy1low and Thy1high TCRγδ+ or TCRαβ+ IELs. Although our data demonstrate that IL-15 regulates Bcl-2 levels in all IEL subsets, it is possible that Thy1low IELs may be more sensitive to effects of IL-15-mediated survival than Thy1high IELs, thus contributing to the increase in the Thy1low population. This idea is supported by our observation that Thy1low IELs express higher levels of CD122.

In addition to its effects on IEL survival, IL-15 is also known for the ability to enhance proliferation of isolated IELs in vitro (16). Therefore, it was surprising that BrdU incorporation of IELs was not defective in IL-15Rα−/− mice. Moreover, the absence of IL-15Rα expression actually enhanced BrdU incorporation in the Thy1low IELs. Presently, this cannot be explained but could be a result of the severely reduced cell numbers and an over-compensation mediated by an alternative mechanism, such as IL-7. Indeed, TCRγδ IELs in IL-7−/− mice also have enhanced BrdU incorporation compared with Wt mice (41). Regardless of the high proliferation rate in the absence of IL-15Rα, these IELs are present at very low levels.

In summary, we have generated a novel model to study the specific effects of IL-15 trans-presented by the intestinal epithelium for IEL development. Using this model, we have provided further evidence that IL-15 trans-presentation is the main mechanism by which IL-15 is delivered to IELs. We have also given conclusive evidence that cells of the intestinal epithelium are the only cell type that needs to trans-present IL-15 to IELs. In addition to identifying the cell source and localization of IL-15, this study identifies a new function of IL-15 in the preferential differentiation of Thy1lowCD8αα+ IELs. Because the specific effect on Thy1low IELs applied to both TCRαβ+ and TCRγδ+ IELs, a common link likely exists in the differentiation between these two separate lineages.

We thank Eliseo Castillo for irradiating mice, Bhavin Shah for technical assistance, Pam Grant for tissue sectioning, Anna Zal for assisting with confocal microscopy, the Genetically Engineered Mouse Facility at the University of Texas M.D. Anderson Cancer Center in part supported by CCSG Grant NCI CA016672, and Sylvie Robine for providing the Villin promoter construct.

The authors have no financial conflict of interest.

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 with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Research is supported by National Institutes of Health Grant AI070910 and the M.D. Anderson Trust Fellowship (to K.S.).

L.J.M. performed a majority of the experiments, analyzed, interpreted data, and cowrote the manuscript. L.A. performed Western blots; T.L. analyzed confocal images; and K.S.S. designed, analyzed, interpreted data, and cowrote the manuscript.

3

Abbreviations used in this paper: IEL, intraepithelial lymphocyte; IEC, intestinal epithelial cell; Wt, wild type; TG, transgenic; IRES, internal ribosome entry site; EGFP, enhanced GFP.

4

The online version of this article contains supplemental material.

1
Cheroutre, H., F. Lambolez.
2008
. The thymus chapter in the life of gut-specific intra epithelial lymphocytes.
Curr. Opin. Immunol.
20
:
185
-191.
2
Lefrancois, L..
1991
. Phenotypic complexity of intraepithelial lymphocytes of the small intestine.
J. Immunol.
147
:
1746
-1751.
3
Kim, S. K., K. S. Schluns, L. Lefrancois.
1999
. Induction and visualization of mucosal memory CD8 T cells following systemic virus infection.
J. Immunol.
163
:
4125
-4132.
4
Lefrancois, L., C. M. Parker, S. Olson, W. Muller, N. Wagner, M. P. Schon, L. Puddington.
1999
. The role of β7 integrins in CD8 T cell trafficking during an antiviral immune response.
J. Exp. Med.
189
:
1631
-1638.
5
Jarry, A., N. Cerf-Bensussan, N. Brousse, F. Selz, D. Guy-Grand.
1990
. Subsets of CD3+ (T cell receptor α/β or γ/δ) and CD3 lymphocytes isolated from normal human gut epithelium display phenotypical features different from their counterparts in peripheral blood.
Eur. J. Immunol.
20
:
1097
-1103.
6
Chen, Y., K. Chou, E. Fuchs, W. L. Havran, R. Boismenu.
2002
. Protection of the intestinal mucosa by intraepithelial γδ T cells.
Proc. Natl. Acad. Sci. USA
99
:
14338
-14343.
7
Poussier, P., T. Ning, D. Banerjee, M. Julius.
2002
. A unique subset of self-specific intraintestinal T cells maintains gut integrity.
J. Exp. Med.
195
:
1491
-1497.
8
Leishman, A. J., L. Gapin, M. Capone, E. Palmer, H. R. MacDonald, M. Kronenberg, H. Cheroutre.
2002
. Precursors of functional MHC class I- or class II-restricted CD8αα+ T cells are positively selected in the thymus by agonist self-peptides.
Immunity
16
:
355
-364.
9
Parrott, D. M., C. Tait, S. MacKenzie, A. M. Mowat, M. D. Davies, H. S. Micklem.
1983
. Analysis of the effector functions of different populations of mucosal lymphocytes.
Ann. NY Acad. Sci.
409
:
307
-320.
10
Eberl, G., D. R. Littman.
2004
. Thymic origin of intestinal αβ T cells revealed by fate mapping of RORγt+ cells.
Science
305
:
248
-251.
11
Gangadharan, D., F. Lambolez, A. Attinger, Y. Wang-Zhu, B. A. Sullivan, H. Cheroutre.
2006
. Identification of pre- and postselection TCRαβ+ intraepithelial lymphocyte precursors in the thymus.
Immunity
25
:
631
-641.
12
Lambolez, F., O. Azogui, A. M. Joret, C. Garcia, H. von Boehmer, J. Di Santo, S. Ezine, B. Rocha.
2002
. Characterization of T cell differentiation in the murine gut.
J. Exp. Med.
195
:
437
-449.
13
Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers, N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al
2000
. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice.
J. Exp. Med.
191
:
771
-780.
14
Lodolce, J. P., D. L. Boone, S. Chai, R. E. Swain, T. Dassopoulos, S. Trettin, A. Ma.
1998
. IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation.
Immunity
9
:
669
-676.
15
Suzuki, H., G. S. Duncan, H. Takimoto, T. W. Mak.
1997
. Abnormal development of intestinal intraepithelial lymphocytes and peripheral natural killer cells in mice lacking the IL-2 receptor β chain.
J. Exp. Med.
185
:
499
-505.
16
Inagaki-Ohara, K., H. Nishimura, A. Mitani, Y. Yoshikai.
1997
. Interleukin-15 preferentially promotes the growth of intestinal intraepithelial lymphocytes bearing γδ T cell receptor in mice.
Eur. J. Immunol.
27
:
2885
-2891.
17
Lai, Y. G., V. Gelfanov, V. Gelfanova, L. Kulik, C. L. Chu, S. W. Jeng, N. S. Liao.
1999
. IL-15 promotes survival but not effector function differentiation of CD8+ TCRαβ+ intestinal intraepithelial lymphocytes.
J. Immunol.
163
:
5843
-5850.
18
Zhao, H., H. Nguyen, J. Kang.
2005
. Interleukin 15 controls the generation of the restricted T cell receptor repertoire of γδ intestinal intraepithelial lymphocytes.
Nat. Immunol.
6
:
1263
-1271.
19
Dubois, S., J. Mariner, T. A. Waldmann, Y. Tagaya.
2002
. IL-15Rα recycles and presents IL-15 In trans to neighboring cells.
Immunity
17
:
537
-547.
20
Schluns, K. S., K. D. Klonowski, L. Lefrancois.
2004
. Transregulation of memory CD8 T-cell proliferation by IL-15R α+ bone marrow-derived cells.
Blood
103
:
988
-994.
21
Schluns, K. S., E. C. Nowak, A. Cabrera-Hernandez, L. Puddington, L. Lefrancois, H. L. Aguila.
2004
. Distinct cell types control lymphoid subset development by means of IL-15 and IL-15 receptor α expression.
Proc. Natl. Acad. Sci. USA
101
:
5616
-5621.
22
Burkett, P. R., R. Koka, M. Chien, S. Chai, F. Chan, A. Ma, D. L. Boone.
2003
. IL-15R α expression on CD8+ T cells is dispensable for T cell memory.
Proc. Natl. Acad. Sci. USA
100
:
4724
-4729.
23
Prlic, M., B. R. Blazar, M. A. Farrar, S. C. Jameson.
2003
. In vivo survival and homeostatic proliferation of natural killer cells.
J. Exp. Med.
197
:
967
-976.
24
Lai, Y. G., M. S. Hou, Y. W. Hsu, C. L. Chang, Y. H. Liou, M. H. Tsai, F. Lee, N. S. Liao.
2008
. IL-15 does not affect IEL development in the thymus but regulates homeostasis of putative precursors and mature CD8αα+ IELs in the intestine.
J. Immunol.
180
:
3757
-3765.
25
Reinecker, H. C., R. P. MacDermott, S. Mirau, A. Dignass, D. K. Podolsky.
1996
. Intestinal epithelial cells both express and respond to interleukin 15.
Gastroenterology
111
:
1706
-1713.
26
Pinto, D., S. Robine, F. Jaisser, F. E. El Marjou, D. Louvard.
1999
. Regulatory sequences of the mouse villin gene that efficiently drive transgenic expression in immature and differentiated epithelial cells of small and large intestines.
J. Biol. Chem.
274
:
6476
-6482.
27
Robine, S., F. Jaisser, D. Louvard.
1997
. Epithelial cell growth and differentiation, IV: controlled spatiotemporal expression of transgenes: new tools to study normal and pathological states.
Am. J. Physiol.
273
:
G759
-G762.
28
Goodman, T., L. Lefrancois.
1988
. Expression of the γ-δ T-cell receptor on intestinal CD8+ intraepithelial lymphocytes.
Nature
333
:
855
-858.
29
Flint, N., F. L. Cove, G. S. Evans.
1991
. A low-temperature method for the isolation of small-intestinal epithelium along the crypt-villus axis.
Biochem. J.
280
: (Pt 2):
331
-334.
30
Goodman, T., L. Lefrancois.
1989
. Intraepithelial lymphocytes. Anatomical site, not T cell receptor form, dictates phenotype and function.
J. Exp. Med.
170
:
1569
-1581.
31
Robine, S., C. Huet, R. Moll, C. Sahuquillo-Merino, E. Coudrier, A. Zweibaum, D. Louvard.
1985
. Can villin be used to identify malignant and undifferentiated normal digestive epithelial cells?.
Proc. Natl. Acad. Sci. USA
82
:
8488
-8492.
32
Bretscher, A., K. Weber.
1979
. Villin: the major microfilament-associated protein of the intestinal microvillus.
Proc. Natl. Acad. Sci. USA
76
:
2321
-2325.
33
Osborn, M., G. Mazzoleni, D. Santini, D. Marrano, G. Martinelli, K. Weber.
1988
. Villin, intestinal brush border hydrolases and keratin polypeptides in intestinal metaplasia and gastric cancer; an immunohistologic study emphasizing the different degrees of intestinal and gastric differentiation in signet ring cell carcinomas.
Virchows Arch. A Pathol. Anat. Histopathol.
413
:
303
-312.
34
Elsasser, H. P., G. Kloppel, H. G. Mannherz, K. Flocke, H. F. Kern.
1991
. Immunohistochemical demonstration of villin in the normal human pancreas and in chronic pancreatitis.
Histochemistry
95
:
383
-390.
35
Grone, H. J., K. Weber, U. Helmchen, M. Osborn.
1986
. Villin: a marker of brush border differentiation and cellular origin in human renal cell carcinoma.
Am. J. Pathol.
124
:
294
-302.
36
Nakazato, K., H. Yamada, T. Yajima, Y. Kagimoto, H. Kuwano, Y. Yoshikai.
2007
. Enforced expression of Bcl-2 partially restores cell numbers but not functions of TCRγδ intestinal intraepithelial T lymphocytes in IL-15-deficient mice.
J. Immunol.
178
:
757
-764.
37
Zhang, X., S. Sun, I. Hwang, D. F. Tough, J. Sprent.
1998
. Potent and selective stimulation of memory-phenotype CD8+ T cells in vivo by IL-15.
Immunity
8
:
591
-599.
38
Masopust, D., V. Vezys, E. J. Wherry, D. L. Barber, R. Ahmed.
2006
. Cutting edge: gut microenvironment promotes differentiation of a unique memory CD8 T cell population.
J. Immunol.
176
:
2079
-2083.
39
Lambolez, F., M. L. Arcangeli, A. M. Joret, V. Pasqualetto, C. Cordier, J. P. Di Santo, B. Rocha, S. Ezine.
2006
. The thymus exports long-lived fully committed T cell precursors that can colonize primary lymphoid organs.
Nat. Immunol.
7
:
76
-82.
40
Smith, X. G., E. M. Bolton, H. Ruchatz, X. Wei, F. Y. Liew, J. A. Bradley.
2000
. Selective blockade of IL-15 by soluble IL-15 receptor α-chain enhances cardiac allograft survival.
J. Immunol.
165
:
3444
-3450.
41
Laky, K., J. M. Lewis, R. E. Tigelaar, L. Puddington.
2003
. Distinct requirements for IL-7 in development of TCR γ δ cells during fetal and adult life.
J. Immunol.
170
:
4087
-4094.