Developmental and Functional Defects of Thymic and Epidermal Vγ3 Cells in IL-15-Deficient and IFN Regulatory Factor-1-Deficient Mice¹

Thymocyte development in the murine thymus starts at days 11–12 of fetal development with the influx of precursor cells. The first wave of cells in thymic ontogeny is unique in that they express an invariant γδ TCR characterized by a lack of junctional diversity. Vγ3 cells are the first TCR-positive cells that can be found in the fetal thymus around gestation days 14–16 (1, 2, 3). At a later time point, other TCR γδ and TCR αβ thymocytes appear. In adult mice, TCR Vγ3 cells are detected only in the epidermis (2). Due to their dendritic morphology, these cells are called dendritic epidermal T cells (DETCs)³ and represent the main T cell population in the epidermis of mice (4). It has been shown that mature fetal Vγ3 thymocytes are the precursors of DETCs in the skin (5, 6).

Mice deficient in the common γ-chain (γ_c), which is shared by the IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21 receptors, have a marked impairment of B, T, and NK cell development (7). The defect in T cell development seen in γ_c^−/− mice mainly reflects the absence of IL-7R signaling, while the defect in NK cell development seen in these mice is due to the absence of IL-15R signaling. IL-7^−/− and IL-7R^−/− mice have major defects in T cell development, but NK cell development is not compromised (8, 9). IL-7 and IL-7R play a critical role in lymphopoiesis by inducing survival and proliferation of progenitor T lymphocytes (10). Studies in IL-15^−/− mice and IL-15Rα^−/− mice have confirmed a critical role for IL-15 in regulating the development and/or expansion of NK cells, NK-T cells, and distinct intestinal intraepithelial lymphocyte populations (11, 12). In addition, these studies have revealed a role for IL-15 in the maintenance of the memory CD8⁺ T cell population in the periphery (11). The finding that IL-15^−/− and IL-15Rα^−/− mice are lymphopenic suggests that IL-15 may also support adaptive CD4⁺ and CD8⁺ T cells. Whether IL-15 regulates adaptive lymphocyte differentiation remains to be elucidated (11, 12).

Some cytokines have also been demonstrated to affect the growth and differentiation of γδ T cells. γδ T cells derived from the fetal thymus and from adult skin, spleen, or peritoneal cavity can proliferate in vitro in response to IL-2, IL-7, or IL-15 (13, 14, 15, 16, 17, 18, 19). γ_c^−/− mice have confirmed these findings, as these mice have defects in γδ T cell development. γ_c^−/− mice have severely reduced numbers of immature fetal Vγ3 cells and lack mature fetal thymic Vγ3 cells. Vγ3 DETCs are absent from the skin epidermis of γ_c^−/− mice (20). In IL-7R^−/− and IL-7^−/− mice, maturation of Vγ3 cells in the fetal thymus is inhibited and no Vγ3 DETCs are detected in the skin, showing the importance of IL-7 in the development and/or survival of Vγ3 cells or their precursors (8, 21). In addition, several other studies have suggested a role for IL-15 and IL-2 in the development of Vγ3 cells (18, 22, 23). Both cytokines interact with receptor complexes that contain the γ_c, the IL-2Rβ chain, and a specific IL-2R or IL-15R α-chain (22, 23). Mature fetal Vγ3 thymocytes and Vγ3 DETCs are known to express the IL-2Rβ chain (14). IL-2Rβ-deficient mice show a moderate reduction of mature Vγ3 cells in the fetal thymus. Small numbers of Vγ3 DETCs are detected in the fetal skin, but they are absent in adult mice (17). Because Vγ3 cells are present in normal numbers in the fetal thymus and in the adult skin of IL-2^−/− mice (24), IL-15 rather than IL-2 signaling through the IL-2Rβ chain seems to be important for the development and/or the expansion of Vγ3 cells and the maintenance of these cells in the skin.

Our aim was to determine whether intrathymic IL-15 is required for the generation of fetal thymic Vγ3 cells and whether peripheral expression of IL-15 in the skin is necessary for the development and/or survival of Vγ3 DETCs. In addition, we wanted to determine whether the transcription factor IFN regulatory factor-1 (IRF-1) is required for the development of Vγ3 cells in the fetal thymus and in the skin epidermis. IRF-1 binds regions within the promoter of type I IFNs and several IFN-inducible genes and is responsible for the induction of IL-15 but not the constitutive expression of this gene. Mice that do not express the transcription factor IRF-1 have been shown to exhibit a severe NK, NK-T, and intestinal intraepithelial lymphocyte deficiency (25). To clarify the role of IL-15 and IRF-1 in Vγ3 T cell development, we studied the development and function of these cells in IL-15^−/− mice and IRF-1^−/− mice. Our results suggest a redundant role for IL-15 expression during Vγ3 T cell development in the fetal thymus and a nonredundant function for localization of Vγ3 cells in the skin. In addition, we show that the transcription factor IRF-1 is important for the morphological maturation of DETCs, probably by regulating IL-15 expression in the skin epidermis during ontogeny. Furthermore, our data imply an important role for IRF-1 in regulating Vγ3 T cell-mediated cytotoxicity.

C57BL/6J wild-type (WT) mice were provided by Proefdierencentrum (Catholic University Leuven, Leuven, Belgium). IRF-1^−/− mice (C57BL/6 background) (26) were kindly provided by Dr. P. Matthys (Catholic University Leuven). IL-15^−/− mice (C57BL/6 background) were kindly provided by Dr. J. Peschon (Immunex, Seattle, WA) (11). Recombination-activating gene (RAG)-1^−/− mice (C57BL/6 background) were purchased from Kankerinstituut (Amsterdam, The Netherlands) and C57BL/6J-Rag^tm1/Mom mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were bred in our facility and housed in a specific pathogen-free environment. Mice were treated and used in agreement with institutional guidelines.

Monoclonal Abs used for staining were anti-FcγRII/III (unconjugated, rat IgG2b; kindly provided by Dr. J. Unkeless, Mount Sinai School of Medicine, New York, NY), anti-heat-stable Ag (HSA; biotin-conjugated, rat IgG2b; BD PharMingen, San Diego, CA), anti-NK1.1 (PE-conjugated, mouse IgG2a; BD PharMingen), anti-Thy1.2 (PE- and FITC-conjugated, rat IgG2b; BD PharMingen), anti-IL-2Rβ (FITC-conjugated, rat IgG2b; kindly provided by Dr. T. Tanaka, Tokyo, Japan), and anti-TCR Vγ3 (FITC-conjugated, hamster hybridoma F536 (kindly provided by Dr. J. P. Allison, University of California, Berkeley, CA) and PE-conjugated (BD PharMingen)).

Epidermal sheets were prepared as described previously (27). Epidermal sheets were labeled with FITC-conjugated anti-Thy1.2 mAb or with FITC-conjugated anti-Vγ3 mAb at 4°C for 18 h. DETCs were counted with a fluorescence microscope in a field that equaled 0.2 mm². Ten mice were used for each strain and for each specimen five random fields were counted. Data are expressed as the mean (± SD) number of positive cells per square millimeter.

Skin samples were freed of fatty tissue and were floated dermal side down in a petri dish containing 0.3% trypsin-PBS solution (Difco, Detroit, MI) at 4°C for 18 h. Epidermal sheets were peeled from the underlying dermis. Epidermal skin samples were then pooled in DMEM (Life Technologies, Paisley, U.K.) containing 0.25% DNase (Boehringer Mannheim, Mannheim, Germany). Single cell suspensions were prepared as described before (27). Cells were counted with trypan blue to exclude dead cells.

Mice were mated overnight. Thymuses from FD17 (plug date = day 0) WT IRF-1^−/− mice and IL-15^−/− mice were removed and disrupted using a small potter homogenizer. Cells were counted with trypan blue to exclude dead cells. Thymocytes were suspended in RPMI 1640 medium supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 0.03% glutamine, and 5 × 10⁻⁵ M 2-ME (all from Life Technologies). This medium will be further referred to as RPMI 1640 medium.

TRIzol LS Reagent (Life Technologies) was added to the sorted cells or total cell suspensions and RNA was extracted according to the manufacturer’s instructions. Before reverse transcription, digestion of DNA was performed with DNase I (Life Technologies). cDNA was synthesized with oligo(dT) as primer using the Superscript kit (Life Technologies). Primers used for RT-PCR for murine hypoxanthine phosphoribosyltransferase (HPRT), a housekeeping gene, were GTAATGATCAGTCAACGGGGGAC (sense) and CCAGCAAGCTTGCAACCTTAACCA (antisense). For IL-12Rβ2, primers used were AAAGCCAACTGGAAAGCATTCG (sense) and AGTTTTGAGTCAGGGTCTCTGC (antisense). Semiquantitative RT-PCR amplification was performed using a PTC-200 Peltier Thermal Cycler (MJ Research, Biozym, Landgraaf, The Netherlands) for 35 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min (HPRT) or with 35 cycles at 94°C for 30 s, 60°C for 30 s, and 72°C for 1 min (IL-12R). For semiquantitative RT-PCR, three 3-fold dilutions of each cDNA were amplified. H₂O and genomic DNA were used as negative controls (data not shown).

Amplification reactions for IL-15, IL-7, and HPRT mRNA were performed with the SYBR Green assay which contained 1× SYBR PCR buffer, 3 mM MgCl₂, 0.2 mM dATP, 0.2 mM dCTP, 0.4 mM dUTP, 1.25 U AmpliTaq Gold, and 0.5 U AmpErase UNG (all from PE Applied Biosystems, Foster City, CA). Primers for murine IL-15 (AAAGCTTTATACGCATTGTCCAAA T (sense) and CATGCAGTCAGGACGTGTTGAT (antisense)), murine IL-7 (GGAATTCCTCCACTGATCCTTG (sense) and TTCCTGTCATTTTGTCCAATTCA (antisense)), and HPRT (AATACGAGGAGTCCTGTTGATGTTG (sense) and CATTCATAGAAGGTTCATGCAAAAAG (antisense)) were designed with Primer Express 1.0 software (PE Applied Biosystems) and used at 50 nM (IL-15) and 200 nM (IL-7, HPRT) concentrations. The PCR conditions were 95°C for 10 min followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. Melting curves were generated after amplification. Amplification reactions for GAPDH, a housekeeping gene, were performed with the TaqMan assay kit for GAPDH amplification from PE Applied Biosystems. Amplification was performed in 1× TaqMan Universal PCR Master Mix (PE Applied Biosystems), using 100 nM of each primer and 200 nM probe for rodent GAPDH. Data were collected using the 5700 SDS thermal cycler (PE Applied Biosystems). Each sample was tested in triplicate and all PCR runs were performed three times.

FD17 thymic cell suspensions were prepared and cultured in 24-well plates (Falcon; BD Biosciences, Mountain View, CA) at 2 × 10⁶ cells/well in 2 ml RPMI 1640 medium with a final concentration of 50 ng/ml recombinant human (rhu)IL-15 (R&D Systems, Abingdon, U.K.). After culture for 4 days in 5% CO₂ at 37°C, cells were harvested, washed, and counted with trypan blue. Cells were sorted into Vγ3⁺ and NK1.1⁺Vγ3⁻ populations. Sorted cells were cultured in 96-well plates (Falcon; BD Biosciences) at 1 × 10⁵ cells/well in 200 μl RPMI 1640 medium for an additional 2 days, with a final concentration of 50 ng/ml rhuIL-15 with or without 2 ng/ml rIL-12 (PeproTech, Rocky Hill, NJ). After these additional 2 days of culture, the purity of the Vγ3 and NK cells was >99%.

To avoid aspecific binding, the FcγR was blocked by preincubating the cells with saturating amounts of anti-FcγRII/III mAb. Freshly isolated FD17 thymocytes were incubated with anti-HSA (biotin-conjugated), anti-Vγ3 (PE-conjugated), and anti-IL2Rβ (FITC-conjugated) at 4°C for 45 min. After washing, cells were incubated with streptavidin-allophycocyanin (BD Biosciences) at 4°C for 20 min. ECs were incubated with anti-Thy1.2 (PE-conjugated) mAb at 4°C for 45 min. Cells were analyzed for fluorescence using a FACSCalibur (BD Biosciences) equipped with an argon (488 nm) and helium (325 nm) laser with the CellQuest software program (BD Biosciences) for data acquisition and analysis. Propidium iodide was added to the cells (2 μg/ml) just before flow cytometric analysis (FCA). Gating was done on propidium iodide-negative cells to exclude dead cells.

FD17 thymocytes from WT mice and IRF-1^−/− mice cultured for 4 days in the presence of rhuIL-15 were incubated with anti-NK1.1 (PE-conjugated) and anti-Vγ3 (FITC-conjugated) mAbs at 4°C for 45 min. Vγ3⁺ cells and Vγ3⁻NK1.1⁺ cells were sorted to a purity of >99% using a FACSVantage flow cytometer (BD Biosciences) equipped with an argon laser.

Freshly isolated FD18 thymocytes (10⁷) from WT and IL-15^−/− mice were i.v. injected into syngeneic RAG-1^−/− and/or IL-15^−/− mice. Six weeks after injection, mice were analyzed for the presence of Vγ3⁺ DETCs by immunolabeling of epidermal sheets. Experiments were repeated three times with three mice of each genotype per experiment.

The tumor target used was the YAC-1 cell line (kindly provided by Dr. M. Joniau, K. V. Leuven, Kortrijk, Belgium). Target cells (10⁶) were labeled with 100 μCi ⁵¹Cr (Amersham International, Little Chalfont, U.K.) for 60 min at 37°C. Cells were washed three times. Effector cells used were sorted FD17 thymic Vγ3⁺ cells or NK cells derived from WT, IL-15^−/−, or IRF-1^−/− mice, cultured in the presence of IL-15 with or without IL-12. Graded effector cell numbers were cocultured in triplicate with 10^3 51Cr-labeled YAC-1 cells in a total volume of 100 μl RPMI 1640 medium in 96-well V-bottom plates (Nunc, Roskilde, Denmark). Alternatively, to determine the spontaneous and maximal ⁵¹Cr release, medium and 1% Triton X-100 solution, respectively, was added to the target cells instead of effector cells. After incubation for 6 h at 37°C, 70 μl supernatant was removed from each well. Then, 225 μl Optiphase Supermix (Wallac, Turku, Finland) was added to the supernatants, and radioactivity was measured using a 96-well scintillation counter (Microbeta; Wallac). Data are expressed as the mean percentage of specific ⁵¹Cr release. Percentage of specific release was calculated as follows: 100 × ((experimental − spontaneous release)/(maximal − spontaneous release)).

To examine the role of IL-15 in fetal Vγ3 T cell development, we analyzed FD17 thymocytes from IL-15^−/− mice. Fetal Vγ3 thymocytes consist of immature HSA^high and mature HSA^low cells (28). Maturation of Vγ3 thymocytes is also associated with the expression of the IL-2/IL-15Rβ chain (14). The total cell number of the fetal thymus of WT and IL-15^−/− mice was comparable (data not shown). As shown in Fig. 1, a selective reduction of mature HSA^low Vγ3⁺ thymocytes in FD17 IL-15^−/− mice could be observed, but all mature HSA^lowVγ3⁺ thymocytes from IL-15^−/− mice expressed the IL-2Rβ chain at normal levels. To examine whether regulation of IL-15 expression determines fetal Vγ3 T cell development, we examined FD17 thymocytes from IRF-1^−/− mice. Others have shown that mice deficient for IRF-1 fail to up-regulate IL-15 expression after stimulation, but low basal amounts of IL-15 can be detected (25). We found no reduction of mature Vγ3⁺ thymocytes in the fetal thymus of IRF-1^−/− mice (Fig. 1). The total cell number of the fetal thymus of WT and IRF-1^−/− mice was comparable (data not shown).

Vγ3 thymocytes migrate from the fetal thymus to the skin (5, 6). After birth, Vγ3 cells can only be detected in the epidermis (2). IL-2Rβ^−/− mice have no Vγ3 cells in the skin epidermis (17), whereas IL-2^−/− mice have normal DETC numbers (24). To determine whether the absence of epidermal Vγ3 cells in IL-2Rβ^−/− mice is due to the lack of IL-15 signaling, epidermal sheets from IL-15^−/− mice were examined. No Vγ3⁺ DETCs could be detected in the skin of adult (12 wk) IL-15^−/− mice by in situ immunofluorescent staining of epidermal sheets (Fig. 2,B). To determine whether the transcription factor IRF-1 is important during development of Vγ3⁺ DETCs, we also examined epidermal sheets of IRF-1^−/− mice. First, we determined whether IL-15 mRNA was expressed in ECs of IRF-1^−/− mice. As expected, unstimulated ECs expressed mRNA for IL-15 but, compared with WT mice, no up-regulation was found after stimulation with LPS and IFN-γ (data not shown). In contrast to IL-15^−/− mice, Vγ3 cells could be detected in normal numbers in epidermal sheets of adult IRF-1^−/− mice (Fig. 2,C). However, IRF-1^−/− Vγ3 DETCs differed from WT cells in their morphology, as Vγ3⁺ DETCs from IRF-1^−/− mice had less extensive dendrites (Fig. 2).

Because no Vγ3⁺ DETCs could be detected in the skin of adult IL-15^−/− mice and DETCs of adult IRF-1^−/− mice had a less mature dendritic morphology, we studied the role of IL-15 for the emergence kinetics of Vγ3⁺ and Thy-1⁺ DETCs. Epidermal sheets from WT, IL-15^−/−, and IRF-1^−/− mice were examined at different time points after birth. Epidermal sheets 2 days after birth contained round-shaped Thy1⁺ ECs and Vγ3⁺ ECs in both WT and IRF-1^−/− mice, although lower numbers were detected in IRF-1^−/− compared with WT mice (Figs. 3 and 4). During the next few days, WT mice showed a gradual increase of Vγ3⁺ DETCs with peak numbers 2 wk after birth (Fig. 4). After 2 wk of age, WT Vγ3⁺ DETC numbers decreased until steady cell numbers were reached. At 12 wk after birth, all WT Vγ3⁺ DETCs had a dendritic morphology (Fig. 3,A). In IRF-1^−/− mice we also found an increase of Vγ3⁺ DETCs until 2 wk after birth, but cell numbers at 1 and 2 wk were significantly lower compared with WT mice. No decrease was observed at later time points. The end result was that at 12 wk there was no difference in the cell number of Vγ3⁺ DETCs in IRF-1^−/− vs WT mice. In contrast, higher numbers of Thy1⁺ DETCs were reached in IRF-1^−/− mice compared with WT mice at 2 and 4 wk of age; at 12 wk there was again no difference detectable (Fig. 4). DETCs from IRF-1^−/− mice had fewer dendrites at all time points examined (Fig. 3). No Vγ3⁺ DETCs could be detected at any time point after birth in the epidermis of IL-15^−/− mice, and only marginal numbers of Thy1⁺ cells were present (Figs. 3 and 4). FCA of EC suspensions showed that these Thy1⁺ cells stained positive for CD3, although expression levels were lower compared with WT Thy1⁺ cells (data not shown). To exclude the possibility that we could not detect very small numbers of Vγ3⁺ DETCs in the epidermis of IL-15^−/− mice by immunolabeling in situ, we prepared EC suspensions at different time points after birth, cultured them with IL-15 for 24 h, and examined them by flow cytometry for the presence of Vγ3⁺ DETCs and/or other Thy1⁺ cells. Compared with adult WT mice, only small numbers of Thy1⁺ cells were present in EC suspensions of IL-15^−/− mice, but no Vγ3 cells could be detected (data not shown).

Because IL-15 seems to be essential for the proliferation and/or survival of Vγ3⁺ DETCs, we determined whether the difference observed in emergence kinetics and dendritic morphology between WT Vγ3⁺ DETCs and IRF-1^−/− Vγ3⁺ DETCs correlated with a different IL-15 mRNA expression in the skin during ontogeny. We prepared EC suspensions at different time points after birth. Real-time PCR for IL-15 showed lower mRNA levels in the skin of IRF-1^−/− mice compared with WT mice at all time points examined (Fig. 5).

Previously it has been shown that circulating CD3⁺TCR Vγ3⁺ fetal murine thymocytes home to the skin and give rise to proliferating DETCs (5, 6). In addition, an important role for IL-7 during the survival and/or proliferation of Vγ3⁺ DETCs in vitro has been suggested (15). To determine further the role of IL-15 and IL-7 during Vγ3 cell development in the skin, we injected WT FD18 thymocytes i.v. into syngeneic IL-15^−/− mice. We also injected WT FD18 thymocytes i.v. into RAG-1^−/− mice as a positive control, and IL-15^−/− FD18 thymocytes were adoptively transferred into RAG-1^−/− mice to see whether IL-15^−/− fetal Vγ3 T cells developed normally within the thymus. Six weeks after injection, epidermal sheets were prepared and Vγ3⁺ DETCs were detected by in situ immunofluorescent staining. Vγ3⁺ DETCs could be detected in the skin of RAG^−/− mice after adoptive transfer of WT and IL-15^−/− FD18 thymocytes. In contrast, Vγ3⁺ DETCs could not be detected in the skin of IL-15^−/− mice after adoptive transfer of WT FD18 thymocytes (Fig. 6). We compared the expression of IL-7 mRNA by real-time PCR in the skin of WT and IL-15^−/− mice. Levels of mRNA for IL-7 in the skin were comparable between WT and IL-15^−/− mice (Fig. 7).

To determine whether Vγ3 cells from IL-15^−/− mice and IRF-1^−/− mice are functional we tested their cytolytic activity against YAC-1 cells. FD17 thymocytes from WT mice and from both knockout mice were cultured in the presence of IL-15. After 4 days, Vγ3 cells and NK cells, as a control, were sorted and cultured for an additional 2 days in the presence of IL-15 to remove the anti-Vγ3 and anti-NK1.1 mAbs from the cell surface. Compared with WT Vγ3 cells, Vγ3 cells from IL-15^−/− mice exhibited reduced killing activity, but cytotoxicity could clearly be detected. The same results were found for IL-15^−/− NK cells (Fig. 8). In contrast, IRF-1^−/− Vγ3 cells exhibited very low killing activity, while IRF-1^−/− NK cells cultured in the presence of IL-15 exhibited killing activity, but weaker when compared with WT NK cells (Fig. 8).

Because the addition of IL-12 augments the killing capacity of NK cells, NK-T cells, and human γδ T cells (29, 30, 31, 32, 33, 34), we determined whether addition of IL-12 could restore the killing activity of IRF-1^−/− Vγ3 cells. IL-15-cultured Vγ3 cells from IRF-1^−/− mice were sorted and cultured for an additional 2 days in the presence of IL-15 plus IL-12. IL-12 did not increase the killing activity of IRF-1^−/− Vγ3 cells, while the killing activity of WT Vγ3 cells and IL-15^−/− Vγ3 cells was significantly enhanced (Fig. 8). Next, we determined whether the lack of cytolytic activity seen in Vγ3 cells from IRF-1^−/− mice upon IL-12 triggering was due to a diminished IL-12R expression. RT-PCR confirmed the presence of mRNA of IL-12Rβ2 chain in Vγ3 cells from IRF-1^−/− mice cultured in the presence of IL-15, and no difference was found as compared with WT Vγ3 cells (Fig. 9).

In this report, we determined the role of IL-15 in the development and maturation of Vγ3 cells. We demonstrate that IL-15, but not the transcription factor IRF-1, is important for the phenotypic maturation of Vγ3 thymocytes within the fetal thymus and is essential for the presence of Vγ3⁺ DETCs in the epidermis of adult mice. The results found in IRF-1^−/− mice suggest that the level of IL-15 expression during ontogeny might determine the normal morphologic maturation of Vγ3⁺ DETCs within the epidermis. The in vivo experiments of adoptive transfer of fetal thymocytes and the data obtained by real-time PCR for IL-7 confirmed the nonredundant role of IL-15 during Vγ3 T cell development in the skin. Our data also indicate that IRF-1 is essential for the induction of Vγ3 T cell-mediated cytotoxicity.

Our results show that mature Vγ3 thymocytes were decreased but not completely eliminated in IL-15^−/− mice, whereas normal numbers were found in IRF-1^−/− mice. These observations indicate that 1) other cytokines can support the development and/or survival of mature fetal Vγ3 thymocytes, and 2) minimal concentrations of IL-15 are sufficient to support the normal survival and/or differentiation of mature fetal Vγ3 thymocytes. In this context, it has been reported that IL-7 is involved in the maturation of Vγ3 cells in the fetal thymus (14, 35). IL-7 signaling is important during thymic differentiation of Vγ3 cells, because maturation of fetal Vγ3 thymocytes is almost completely blocked in IL-7^−/− and IL-7Rα^−/− mice (8, 9). In addition, IL-7R signaling induces germline transcription in the TCR γ locus and supports the proliferation and survival of lymphocyte precursors (10, 36). The absence of mature Vγ3 cells in IL-7^−/− mice shows that IL-15 alone is not sufficient for maturation of Vγ3 thymocytes to occur.

γ_c^−/− mice have severely reduced numbers of immature fetal Vγ3 thymocytes, no mature Vγ3 thymocytes, and an absence of epidermal Vγ3⁺ DETCs (20). IL-7 and IL-15 are known to share the γ_c. Vγ3⁺ DETCs are absent in IL-7^−/− and IL-7Rα^−/− mice, indicating a role for IL-7 in Vγ3⁺ DETC development (8, 9, 21). However, our data also show that IL-15 is nonredundant for the development and/or survival of Vγ3⁺ DETCs in vivo, because these cells were completely absent in the epidermis of IL-15^−/− mice. Adoptive transfer of WT FD18 thymocytes into IL-15^−/− mice clearly indicated that IL-15 plays a nonredundant role in Vγ3⁺ DETC development. In addition, we found comparable levels of IL-7 mRNA in the skin of IL-15^−/− and WT mice, suggesting that although IL-7 is present in the skin of IL-15^−/− mice it is not sufficient to support the development of Vγ3⁺ DETCs. These data indicate that, whereas other cytokines like IL-7 can promote the development of fetal Vγ3 thymocytes and support the proliferation and survival of DETCs in vitro (8, 9, 15), they cannot compensate for IL-15 during development of Vγ3⁺ DETCs in the epidermis. Results obtained by others have already shown a reduction of mature fetal Vγ3 thymocytes and the absence of DETCs in IL-2Rβ^−/− mice. Furthermore, in contrast to the IL-7R, an essential role for the IL-2Rβ chain in the proliferation and survival of DETCs in the skin has been shown (37). Our data are in agreement with these results and, in addition, confirm the previous assumption that signaling through the IL-2Rβ chain by the cytokine IL-15, rather than IL-2, is a key factor in DETC development.

The kinetic studies we performed ( Figs. 3–5) support the assumption that IL-15 is a key factor for normal development of Vγ3 cells in the skin. Newly arriving DETC precursors in the skin, derived from fetal Vγ3⁺ thymocytes (5, 38), are known to undergo proliferation in the epidermis (6, 39, 40). Colonization of the epidermis occurs in the perinatal period. The results we found for the kinetics in appearance of mature Thy⁺ and Vγ3⁺ DETCs in the epidermis of WT mice are in agreement with previously reported data (39). Different kinetics in appearance of Thy1⁺ and Vγ3⁺ DETCs were found in IRF-1^−/− mice. In addition, morphologically, most Thy1⁺ and Vγ3⁺ DETCs from IRF-1^−/− mice had less extensive dendrites compared with WT DETCs at each time point investigated. No Vγ3⁺ DETCs could be found at any time point after birth in the skin of IL-15^−/− mice. Because the transcription factor IRF-1 regulates IL-15 expression (25), we postulate that the amount of IL-15 expression during ontogeny could determine the maturation of DETC precursors in the epidermis. IL-15 mRNA levels in the epidermis of WT mice were indeed higher compared with the levels found in IRF-1^−/− mice at each time point examined. Although this might indicate that also higher levels of IL-15 protein are present, one has to keep in mind that IL-15 synthesis and secretion can be negatively regulated at multiple levels, i.e., at the levels of transcription, translation, and intracellular trafficking (41, 42). We were not able to measure IL-15 protein levels, as no reliable IL-15 ELISA method is available at the moment. It has already been postulated that skin epithelium has the capacity to induce DETC maturation (40) and that cytokines expressed in the epidermis might determine the localization and maturation of Vγ3⁺ DETCs in the epidermis (8, 17, 18, 43, 44). Our results obtained in IL-15^−/− and IRF-1^−/− mice suggest that the presence of threshold amounts of IL-15 in the skin epithelium is essential during the development and maturation of DETCs in the epidermis.

Vγ3 cells, cultured in the presence of IL-2 or IL-15, proliferate and acquire lymphokine-activated killing capacities (45). IL-15^−/− Vγ3 cells, cultured in the presence of IL-15, acquired reduced but significant killing capacity when compared with WT Vγ3 cells, whereas IRF-1^−/− Vγ3 cells exhibited a drastically decreased killing activity. In contrast, and in agreement with published data (25), IRF-1^−/− NK cells acquired significant killing activity in response to IL-15, but it was still somewhat lower when compared with WT NK cells. Our in vitro studies on Vγ3 and NK killing capacity from IL-15^−/− mice and IRF-1^−/− mice, and previous data showing the importance of IL-15 in the development of functional cytotoxic cells (11), suggest that IRF-1 controls the expression of other, non-IL-15, target gene(s) that contributes to the acquisition of cytotoxicity in Vγ3 cells. In this context, IL-12 has been shown to enhance γδ T cell cytotoxicity activity (29, 33). Because IRF-1 is known to regulate IL-12 expression (46), IL-12 could be the missing factor to induce lytic activity in IRF-1^−/− Vγ3 cells. But, in contrast to WT Vγ3 cells and IL-15^−/− Vγ3 cells, addition of IL-12 did not enhance the killing capacity of IRF-1^−/− Vγ3 cells. As this might be due to inadequate expression of the IL-12R, we measured IL-12R mRNA levels by semiquantitative RT-PCR. However, Vγ3 cells from IRF-1^−/− mice expressed the same levels of mRNA for IL-12R as compared with WT Vγ3⁺ thymocytes. This is in contrast with a previous report which shows a reduced expression of IL-12R mRNA in IRF-1^−/− mice. But mRNA levels were determined only for hepatic and pulmonary tissues and macrophages, not for γδ T cells (47). IL-12 induces cytotoxicity, presumably through the induction of genes involved in target cell lysis, such as perforin or granzyme B (48, 49). The induction of lytic granules containing perforin and granzymes might be defective in IRF-1^−/− mice, explaining the lack of lytic activity. But Vγ3 cells from IRF-1^−/− mice showed normal expression of perforin and contained lytic granules (data not shown). Because the presence of perforin is sufficient for the lysis of YAC-1 target cells (50), we did not examine whether granzymes were present. Furthermore, IRF-1^−/− Vγ3 cells showed normal adhesion to the target cells and expressed normal levels of the 2B4 NK receptor (data not shown), which has been shown to augment the killing capacity of Vγ3 cells (51, 52).

The difference observed in Vγ3 T cell morphology and function between IRF-1^−/− and WT mice can be due to differences in basal and inducible expression levels, respectively, of IL-15 (25). In addition, because IRF-1 is known to regulate several other genes in addition to IL-12, such as type I IFNs, IL-18, inducible NO synthase, and Fas ligand (25, 46, 53, 54, 55, 56, 57, 58), and is located downstream from IFN-γ, IFN-αβ, IL-6, IL-1, and TNF (59, 60, 61, 62), we cannot rule out the possibility that one of these factors is involved in the morphological maturation and function of Vγ3 T cells.

IL-15^−/− mice specifically lack NK cells, NK-T cells, intestinal intraepithelial lymphocytes, and memory CD8⁺ T cells (11). The loss of these cells demonstrates that IL-15 is mainly critical for the development and/or maintenance of lymphoid cells of the innate immune system. Vγ3 cells share common features with both NK and NK-T innate immune cells, as Vγ3 cells also express NK cell markers including 2B4 (52), IL-2Rβ/IL-15Rβ (30), Ly49E, and CD94/NKG2 (63), and as they express a canonical TCR Vγ3/Vδ1 (14, 52, 64). IL-7 is necessary for normal development of lymphoid cells of the adaptive immune system (10). Others have shown that IL-7 is also essential for thymic differentiation of Vγ3 cells (9, 21). These data, in combination with our finding that Vγ3 cell development is impaired in the skin IL-15^−/− mice, indicate that the cytokine requirements during fetal thymic differentiation of Vγ3 cells are characteristic of adaptive immune cells, whereas cytokine requirements for peripheral survival are characteristic of innate immune cells. This resemblance of Vγ3 cells with both cells from the innate immune system and the adaptive immune system confirms the assumption that Vγ3 cells are at the transition between these two systems.

We thank Dr. J. Peschon and Dr. P. Matthys for providing us with IL-15^−/− and IRF-1^−/− mice, respectively. We also thank M. De Smedt for purification of Abs and C. Collier and G. De Smet for animal care. We thank T. Van Belle for optimizing and performing the real-time PCR for IL-15 and T. Taghon and T. Kerre for excellent discussions.