Differential Roles of IL-2–Inducible T Cell Kinase-Mediated TCR Signals in Tissue-Specific Localization and Maintenance of Skin Intraepithelial T Cells Free

Unlike conventional αβ T cells that primarily reside in secondary lymphoid organs for adaptive immune responses, various subsets of γδ T cells preferentially reside in epithelial tissues, such as the skin, reproductive tract, respiratory tracts, and intestines, where they function as the first line of defense (1). The different tissue-specific γδ T cells preferentially use different subsets of TCRs. In mice, skin-specific intraepithelial γδT lymphocytes (sIELs, also called dendritic epidermal T cells), a prototype of the tissue-specific T cells, almost exclusively express canonical γδ TCRs composed of Vγ3–Jγ1Cγ1 and Vδ1–Dδ2–Jδ2Cδ chains, whereas vaginal epithelial γδ T cells express Vγ4/Vδ1⁺ TCRs. By comparison, γδ T cells in secondary lymphoid organs express more diverse TCRs, predominantly of Vγ2 and Vγ1.1 associated with several Vδ chains. The preferential usage of specific TCRs by the different tissue-specific γδ T cells is suggested to be important for their tissue-specific functions. sIEL-specific Vγ3⁺ γδ TCRs react with Ags upregulated on diseased skin cells and play an important role in tumor surveillance and wound healing among others to maintain the integrity of the skin.

Precursors for the different tissue-specific γδ T cells are generated in the thymus at different stages of ontogeny. The Vγ3⁺ sIEL precursors are generated exclusively in the early fetal thymus, where they are the first T cell population to arise during ontogeny (around day 15 of embryonic gestation, E15). Once out of the thymus, they take residence in the skin epithelium, where they expand and sustain locally for the life span of mice (2–4). In contrast, fetal thymic Vγ4⁺ γδ T cells localize to peripheral destinations, such as the reproductive tract. In adults, the generation of Vγ3⁺ and Vγ4⁺ γδ T cells is completely suppressed, whereas Vγ2⁺ and Vγ1.1 γδ⁺ T cells are predominately generated and preferentially emigrate to secondary lymphoid organs, among other tissues. However, mechanisms regulating tissue-specific development of the various γδ T cell subsets are not well understood.

It has become clear recently that selection is involved in the development of tissue-specific γδ T cells, at least in the case of sIELs (5–7). We reported that fetal thymic γδ T cell populations that display activated or memory phenotypes correlated with their development into sIELs (7). Compared to other γδ T cells, fetal thymic Vγ3⁺ γδ T cells express a unique set of chemokine and cytokine receptors, including high levels of sphingosine-1-phosphate (S1P) receptor 1 (S1PR1) and CCR10 (7), which are potentially important for their thymic egress and skin localization (8–11), and the cytokine receptor CD122 (IL-15Rβ), which is critical for their survival/expansion in the skin (12, 13). In absence of positive selection, as observed in a substrain of FVB mice (FVB/Taconic) that express mutated Skint1, a selecting molecule for the Vγ3⁺ sIEL precursors, these cells could not develop into sIELs (14, 15). However, if transgenic (Tg) fetal thymic γδ T cells are positively selected to express the proper chemokine and cytokine receptors, then they could develop into sIELs (7, 16). These findings suggest that the TCR-dependent positive selection of fetal thymic γδ T cells is critical for their development into sIELs by promoting the expression of proper homing and cytokine receptors for epidermal localization and expansion.

Previous studies using various knockout mice found that multiple TCR signaling molecules, including Lck, Syk, and ZAP-70, are important for sIEL development (17–20). Although these molecules are involved in TCR signaling in general, they may differentially affect the development of sIELs and other T cell populations. For example, mice deficient in Syk, a kinase downstream of the TCR signaling, have normal development of conventional αβ T cells and splenic γδ T cells but impaired sIEL development, suggesting that there is a unique molecular signaling requirement for sIEL development (17, 21). However, mechanisms by which TCR signaling molecules affect tissue-specific sIEL development are poorly understood.

IL-2–inducible T-cell kinase (ITK) is a Tec family nonreceptor tyrosine kinase that plays multiple roles downstream of TCR signaling. During TCR signaling, ITK forms a complex with the adaptor molecule Slp-76 and is involved in the phosphorylation of phospholipase Cγ and intracellular Ca²⁺ mobilization. The ITK-regulated signal is also involved in the activation of the ERK/MAPK pathway and the activation of transcription factors AP-1 and NFAT. In addition, ITK-mediated signals modulate the TCR-induced reorganization of the actin cytoskeleton by interfacing with the guanine nucleotide exchange factor Vav1, another important TCR signaling molecule (22–24). Besides its role in the TCR signaling, ITK could transduce integrin and chemokine receptor-initiated signaling (22, 25–29).

ITK deficiency affects the development of various T cell populations differentially (30–34). Although the absence of ITK impairs the development of conventional CD4⁺ and CD8⁺ T cells, the development of nonconventional or “innate memory phenotype” CD4⁺ and CD8⁺ T cells remains intact. The nonconventional CD8⁺ T cells in ITK-deficient mice exhibit activated/memory phenotypes, including the expression of memory markers CD44, CD122, and NK1.1, rapid production of cytokines, and dependency on IL-15, features shared by innate lymphocytes (35–37). ITK is also required for the development and function of invariant NKT cells (38, 39). More recently, we and others have reported that ITK deficiency increased the generation of an IL-4–producing Vγ1.1⁺ T cell population (40, 41). These studies suggest that ITK is a key signal component that differentially regulates the development of various T cell populations.

In this study, we identified ITK as a critical signaling molecule specifically involved in the positive selection-associated acquirement of the unique skin-homing property by the fetal thymic sIEL precursors for their specific peripheral location, but dispensable for their maintenance, which provides molecular insights into differential requirements for TCR signaling in peripheral localization and survival/expansion of the tissue-specific γδT cells.

ITK^−/−, TCRδ^−/−, Vav1^−/−, and β2M^−/− and KN6 Tg mice were previously described (32, 42, 43). CCR10 knockout/EGFP knockin (CCR10^EGFP/EGFP) mice were generated in the Xiong laboratory at Pennsylvania State University. ITK^−/−KN6 and ITK^−/−CCR10^+/EGPP mice were generated by crossing ITK^−/− mice with KN6 and CCR10^EGFP/EGFP mice, respectively. All of the mice were kept in specific pathogen-free conditions and used for experiments at ages of 6–8 wk unless indicated otherwise in the text or figure legends. Experiments were approved by the Institutional Animal Care and Use Committee at Pennsylvania State University.

Epidermal cells were prepared as previously described (44). Briefly, hair was removed from the skin with Nair (Church & Dwight, Princeton, NJ). The treated skin was excised and trimmed of s.c. fat. Skin strips were digested with 0.3% trypsin/glucose/NaCl/KCl solution for 45 min at 37°C. Epidermal layers of skin strips were gently removed and incubated with the 0.3% trypsin/glucose/NaCl/KCl solution containing 0.0001% DNase for 10 min at 37°C. The epidermal cells were washed with medium and purified with Ficoll (GE Healthcare, Piscataway, NJ). Cells were cultured overnight in media containing IL-2 (20 U/ml) and used for analyses.

FITC-conjugated anti-Vγ2 (or Vγ3), R-phycoerythrin-Cy5 (PECy5)-conjugated anti-CD3, and PE-conjugated anti-CD24 Abs were purchased from BD Biosciences (San Jose, CA). FITC-conjugated anti-CD122 and biotin-conjugated anti-BrdU, anti-CD122, and anti-γδTCR Abs were purchased from eBioscience (San Diego, CA). PE- or FITC-conjugated streptavidin was purchased from Invitrogen. S1P was purchased from VWR (San Diego, CA), and CCL27 was purchased from PeproTech (Rocky Hill, NJ).

Cells were incubated with fluorescent Abs for 30 min at 4°C. For biotin-labeled Abs, streptavidin–PE was added in the second step and incubated for 20 min at 4°C. All of the samples were analyzed using the flow cytometer FC500 (Beckman Coulter, Miami, FL).

The epidermal sheets were prepared as described (45). Briefly, 6-wk-old mice were sacrificed; the ears were cut off, mechanically split into dorsal and ventral sides, and then placed in EDTA solution. After incubation, the epidermis was peeled off as a single sheet and stained with FITC-conjugated anti-Vγ2 or Vγ3 TCR Abs and analyzed on a fluorescence microscope (BX61, Olympus, Center Valley, PA, or Eclipse TE300, Nikon, Melville, NY).

Total RNA was extracted from sorted cells or skin tissue with TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. The first-strand cDNA was synthesized from the RNA using SuperScript III Reverse Transcriptase (Invitrogen). For semiquantitative PCR, serial 5-fold dilutions of cDNA were subjected to PCR with primer sets for rearranged Vγ3, CCL27, or β-actin. Quantitative real-time PCR was performed using SYBR Green Master Mix (Invitrogen). Primer sets for individual genes are as follows: CCR6F, 5′-AGAACTCCAAGAGGCACAGAGCAA-3′; CCR6R, 5′-TGTTGTGAGGGATCTGACAAGCCA-3′; CCR10F, 5′-TTCCTAGCCTGTATCAGCG; CCR10R, 5′-TAGAGCCAGAAACAGCGAC-3′; S1PR1F, 5′-GTGTAGACCCAGAGTCCTGCG-3′; S1PR1R, 5′-AGCTTTTCCTTGGCTGGAGAG-3′; KLF2F, 5′-TGTGAGAAATGCCTTTGAGTTTACTG-3′; KLF2R, 5′-CCCTTATAGAAATACAATCGGTCATAGTC-3′; β-actinF, 5′-CCCATCTACGAGGGCTAT-3′; β-actinR, 5′-TGTCACGCACGATTTCC-3′; L3 and J1 primers were previously described (46).

Cells were loaded with CFSE by incubation at a concentration of 1 × 10⁷ cells per milliliter in PBS for 10 min with 2.5 μM CFSE (Molecular Probes, Eugene, OR) at 37°C, then washed with cold complete medium twice, and resuspended in culture medium. The labeled cells (5 × 10⁶ cells per well, 2 ml) were cultured for 3 d in 12-well tissue culture plates (Becton Dickinson Labware, Franklin Lakes, NJ) in the presence of IL-2 (10 U/ml) only, IL-15 (50 ng/ml) and IL-2 (10 U/ml), or anti-γδ TCR Ab (GL4, 1 μg/ml) and IL-2 (10 U/ml). After the culture, the cells were stained with PE-conjugated anti-γδ TCR (GL3) and PECy5-conjugated anti-CD3 Abs and analyzed by flow cytometry.

Mice were injected i.p. with 100 μl BrdU dissolved in PBS (10 mg/ml) at the onset of experiments. At the same time, mice were fed water that contained 0.8 mg/ml BrdU and 5% glucose. For the incorporation of BrdU into 2-wk-old mice, the mice were injected i.p. with 50 μl of 10 mg/ml BrdU every other day. Nine days later, sIELs were isolated and assessed for BrdU incorporation by staining with biotin-conjugated anti-BrdU Ab/PE–streptavidin and FITC-conjugated anti-Vγ3 and PECy5-conjugated anti-CD3 Abs and flow cytometric analysis.

The assay was performed using 24-well chemotaxis chambers (Corning Costar, Lowell, MA). E16 fetal thymocytes (2 × 10⁵ cells per well, 100 μl) were added to the upper chamber, and 100 nM S1P, 100 nM CCL27, or fetal skin culture medium was placed in the bottom chamber. Cells were then incubated for 4 h at 37°C and 5% CO₂, and the cells in the upper and lower chambers were collected and analyzed. The percentage of migration was determined from the original cell input.

These experiments were performed using the TMR red In Situ Cell Death Detection Kit (Roche Applied Science, Indianapolis, IN). Briefly, freshly isolated ear epidermal sheets were fixed with 4% paraformaldehyde for 20 min and permeablized with 0.1% Triton X-100 for 2 min on ice, followed by culture with the TUNEL reaction mixture for 30 min and FITC-conjugated anti-Vγ2 Ab overnight. The stained epidermal sheets were analyzed by Olympus FluoView TM FV300 laser scanning microscope.

KN6 Vγ2⁺ Tg γδ T cells were purified from E16 fetal thymus of ITK-sufficient or knockout KN6 mice by a cell sorter and injected i.p. into 1-wk-old β2M^−/−TCRδ^−/− recipients (5 × 10⁵ cells per mouse). Eight weeks after the transfer, ear epidermal sheets of the recipients were analyzed for donor-derived sIELs by in situ immunofluorescent staining.

All of the data are expressed as mean ± SD. Statistical significance was determined by two-tailed Student t tests. p < 0.05 is considered significant.

To evaluate the role of ITK-mediated signaling in sIEL development, we first assessed the sIEL populations in ITK^−/− and wild-type (WT) mice by flow cytometry (Fig. 1A). Compared to the WT controls of the same ages, 6- to 8-wk-old ITK^−/− mice had reduced percentages of Vγ3⁺ sIELs (14.73 ± 0.41% versus 3.56 ± 0.32%, p < 0.0001). In situ examination of sIELs on skin epidermal sheets by immunofluorescent microscopy confirmed the impaired development of sIELs in ITK^−/− mice (181.8 ± 19.5 [WT] versus 66.3 ± 4.3 [ITK^−/−] cells per field, p < 0.05), demonstrating that ITK-mediated signaling is important in sIEL development (Fig. 1B, 1C). In addition, although the remaining sIELs in ITK^−/− mice displayed the normal dendritic morphology (Fig. 1B), they produced less IFN-γ when stimulated in vitro with anti-γδ TCR Ab (Supplemental Fig. 1), suggesting that the ITK-transduced TCR signaling is also important for the function of sIELs.

Vav1 lies downstream of ITK and has been suggested to act directly with ITK to transduce TCR-activated signals (24). In addition, Vav1 has been found previously to be critical for γδTCR-mediated T cell proliferation (47), raising the possibility that it is also involved in sIEL development. To test this, we analyzed the development of sIELs in Vav1 knockout mice. Surprisingly, unlike ITK^−/− mice, 6- to 8-wk-old Vav1^−/− mice had similar numbers of sIELs in the skin as WT controls (60.5 ± 6.3 [WT] versus 54 ± 7.1 [Vav1^−/−] cells per field) (Fig. 1D–F), suggesting that Vav1-regulated signals are not involved in the ITK-mediated sIEL development.

Signals regulated by ITK could be potentially involved in multiple stages of sIEL development, from the TCR-mediated positive selection of the fetal thymic sIEL precursors to their peripheral expansion in the skin. We therefore determined whether the seeding of fetal skin by fetal thymic Vγ3⁺ sIEL precursors was impaired in ITK^−/− mice. Compared to the WT controls, transcripts of rearranged Vγ3 TCR in the fetal skin were dramatically decreased in ITK knockout mice (>10-fold reduction) (Fig. 1G), suggesting that the defective sIEL development originates at the fetal stage, likely due to impaired generation, selection, or both of the fetal thymic sIEL precursors. As a control, the reduction of transcripts of rearranged Vγ3 TCR in the skin of 6- to 8-wk-old ITK^−/− mice was ~3-fold and correlated with the reduction of the Vγ3⁺ sIEL numbers (Fig. 1A–C, 1G).

To address how ITK deficiency affects the development of sIEL precursors, we characterized the fetal thymic Vγ3⁺ γδ T cells of ITK^−/− mice. First, we determined whether ITK deficiency affects their positive selection and maturation. As previously reported, the positive selection and maturation of the fetal thymic Vγ3⁺ cells are associated with upregulation of CD122 and downregulation of CD24 in WT mice (7, 48). Surprisingly, this was also the case in fetal thymic Vγ3⁺ cells of ITK^−/− mice (Fig. 2A), indicating that ITK deficiency does not affect their general selection and maturation processes.

Generation of the fetal thymic Vγ3⁺ sIEL precursors was not impaired significantly in ITK^−/− mice either. Although there was a slight delay in the appearance of Vγ3⁺ γδ T cells in E15 fetal thymi of ITK^−/− mice, this was no longer the case by E16 (Fig. 2B). In fact, as the fetuses aged, there was gradual accumulation of Vγ3⁺ cells in the ITK^−/− fetal thymi. By E18 when the number of Vγ3⁺ cells was decreased in WT fetal thymi due to the egress of mature Vγ3⁺ γδ T cells and the reduced generation of new Vγ3⁺ cells, the number of Vγ3⁺ cells continued to increase in the ITK^−/− fetal thymi, resulting in significantly more of these cells in ITK^−/− mice than in WT mice (Fig. 2B). These findings raise a possibility that the ITK deficiency might affect the proper migration of the sIEL precursors, which would correlate with their impaired seeding in the fetal skin (Fig. 1G).

TCR-mediated positive selection of the fetal thymic Vγ3⁺ sIEL precursors promotes a coordinated switch in expression of multiple “migration” molecules, including the upregulation of S1PR1 and CCR10 and downregulation of CCR6, which are important for their thymic egress and skin-homing (7). To determine whether ITK-mediated signals regulate their expression, we sorted CD122⁻ and CD122⁺ Vγ3⁺ γδ T cells from ITK^−/− and WT fetal thymocytes and analyzed the expression of these migration molecules by real-time RT-PCR. Compared to the WT controls, the upregulation of CCR10 and S1PR1 expression in the ITK^−/− fetal thymic CD122⁺ Vγ3⁺ γδ T cells was significantly impaired (Fig. 3A). Therefore, although its deficiency did not affect the maturation of fetal thymic sIEL precursors, ITK is required for promoting the proper expression of the migration molecules in the mature fetal thymic sIEL precursors. The mature fetal thymic CD122⁺ Vγ3⁺ T cells of ITK^−/− mice also expressed a lower level of KLF2 (Fig. 3A), a transcription factor critical in the regulation of chemokine receptor expression in positively selected αβ T cells (49, 50), suggesting that KLF2 might be involved in the ITK-mediated chemokine receptor expression. In contrast to the defect in upregulation of the migration molecules in the mature ITK^−/− CD122⁺ Vγ3⁺ T cells, there was little difference in their expression on the immature CD122⁻ Vγ3⁺ T cells of the ITK^−/− and WT mice (Fig. 3A), suggesting that the ITK-mediated signaling is mainly involved in regulation of the positive selection-associated acquisition of the skin-homing property.

To further confirm that the ITK deficiency impaired the CCR10 upregulation in the positively selected sIEL precursors, we crossed ITK^−/− mice with CCR10 knockout/EGFP knockin mice that use EGFP as a reporter for CCR10 expression. The resultant ITK^−/−CCR10^+/EGFP mice had a significantly reduced percentage of CCR10(EGFP)⁺ Vγ3⁺ γδ T cells in the early fetal thymus (Fig. 3B, top panels). However, as the fetus aged, this reduction disappeared (Fig. 3B, bottom panels). Considering that the ITK^−/− mature Vγ3⁺ γδ T cells are defective in the upregulation of S1PR1, one plausible explanation for this is that although the ITK^−/− Vγ3⁺ T cells have impaired upregulation of CCR10 after positive selection, the smaller numbers of CCR10⁺ Vγ3⁺ cells are unable to emigrate and accumulate in the thymus. By contrast, WT mature CCR10⁺ Vγ3⁺ sIEL precursors constantly migrate out of the thymus.

To further dissect this, we assessed the in vitro migration capabilities of ITK^−/− fetal thymic Vγ3⁺ γδ T cells toward S1P, a ligand for S1PR1 involved in mature thymic T cell emigration (9, 50–52). As shown in Fig. 3C, ITK^−/− fetal thymic Vγ3⁺ γδ T cells migrated much less efficiently than those in WT controls toward the S1P attraction. The ITK^−/− Vγ3⁺ γδ T cells also had defects in migration toward CCL27 and culture media of the fetal skin (Fig. 3C), suggesting that they have impaired ability in CCL27-mediated skin homing. Together, these results demonstrate that ITK-mediated signaling is important for positive selection-associated acquisition of the unique homing property in the fetal thymic Vγ3⁺ sIEL precursors for their egress from the thymus and localization in the skin.

We noted that even though the ITK^−/− fetal thymic sIEL precursors exhibited severe defects in skin-seeding at the fetal stage, the reduction in the number of sIELs in adult ITK^−/− mice is not as severe (Fig. 1A–C, 1G). This suggests that there might be a homeostatic compensation by which the few ITK-deficient sIEL precursors that make it to the skin are capable of extensive expansion, which would result in the reduced difference in numbers of sIELs between ITK^−/− and WT mice when they grow older. Consistent with this, compared with WT controls of same ages, 3- to 4-wk-old ITK^−/− mice had on average 6-fold reduction in the number of sIELs (8.5 ± 1.5% [WT] versus 1.5 ± 0.7% [ITK^−/−], p < 0.01, n = 5 each), whereas the reduction was 3- to 4-fold in the 6- to 8-wk-old adults (Fig. 1A–C).

To directly assess whether the ITK deficiency affected the proliferation of sIELs, we performed in vivo BrdU labeling experiments of WT and ITK^−/− mice and found that the ITK-deficient sIELs incorporated BrdU at a level similar as, if not higher than, WT sIELs (Fig. 4A). Therefore, although ITK is critical for the regulated expression of multiple migratory molecules on fetal thymic sIEL precursors for their epidermal localization, it is not required for their peripheral proliferation.

Correlating with their normal in vivo proliferation, there was no significant difference in the in vitro proliferation of the ITK^−/− and WT fetal thymic sIEL precursors in response to the anti-γδ TCR Ab stimulation (Fig. 4B), which was in contrast with the requirement of Vav1 in the TCR signaling-mediated proliferation of γδ T cells (Fig. 4C) (47). Together with the fact that the Vav1 knockouts did not have any defects in sIEL development (Fig. 1D–F), these results support a notion that TCR-mediated signaling is not required for the peripheral expansion and maintenance of sIELs. Likely, the expansion of sIELs in the skin is driven by IL-15/receptor signaling (12, 13). Consistent with this, both ITK^−/− and Vav1^−/− fetal thymic γδ T cells proliferated normally in response to IL-15 (Fig. 4D).

To directly characterize the role of ITK-regulated TCR signaling in the development of sIELs, we crossed ITK^−/− mice with KN6 γδTCR Tg mice (42). The Vγ2⁺ KN6 γδTCR recognizes ligands T10/T22, two nonclassical MHC class I molecules whose expression is high in C57BL/6 (B6), low in BALB/c, but absent in β2M^−/− mice (53, 54).

As observed in ITK^−/− Vγ3⁺ sIEL precursors, the fetal thymic Tg Vγ2⁺ T cells of WT and ITK^−/− KN6 mice on the high-ligand B6 background could be positively selected to undergo the Vγ3-like maturation process, but the ITK^−/− Tg Vγ2⁺ cells had defects in seeding the fetal skin (Supplemental Fig. 2). Surprisingly, the absence of ITK had opposite effects on the development of the Tg Vγ2⁺ and natural Vγ3⁺ sIELs in adult mice. Compared to the ITK-sufficient KN6 mice, ITK^−/− KN6 mice had significantly increased numbers of Tg Vγ2⁺ sIELs (Fig. 5A–C), whereas ITK deficiency impaired Vγ3⁺ sIEL development (Fig. 1A–C).

Closer microscopic examination found a morphological difference in the Tg sIELs of ITK^−/− and WT background. Although the Tg sIELs in ITK^−/− KN6 mice displayed the normal dendritic morphology, those of ITK-sufficient KN6 mice looked more rounded (Fig. 5C, ×400 magnification), resembling activated sIELs (19). Because ligands for the KN6 γδTCR, unlike those for the natural sIEL-specific Vγ3⁺ γδTCR, are highly expressed in the skin of B6 mice (53), this suggests that the continuous TCR/ligand-mediated signaling in Tg sIELs, transduced via ITK, may lead to their reduction, likely due to persistent activation-induced apoptosis, whereas a reduction of such signaling in the absence of ITK reverses the effect. Supporting this idea, the in situ TUNEL analysis of epidermal sheets found significantly lower percentages of apoptotic Tg sIELs in ITK^−/− KN6 than those in ITK-sufficient KN6 mice (Fig. 5D, Supplemental Fig. 3).

These data suggest the increased number of Tg sIELs in adult ITK^−/− KN6 mice to be a result of improved peripheral maintenance, which overcomes the reduced initial skin-seeding by the ITK^−/− fetal thymic Tg γδ T cells. To further test this, we eliminated the effect of peripheral TCR/ligand interaction-induced signaling by transferring ITK^−/− or ITK^+/+ fetal thymic KN6 Tg γδ T cells into ligand-negative β2M^−/−TCRδ^−/− recipients, which lack endogenous γδ T cells. Eight weeks after the transfer, the recipients were analyzed for donor-derived sIELs. As shown in Fig. 5E, the adoptively transferred ITK^−/− fetal thymic KN6 Tg γδ T cells gave rise to fewer sIELs in the ligand-negative recipients than the ITK^+/+ donor cells, a difference that is a reversal from that seen in ITK^−/− KN6 mice. In addition, the sIELs that developed in the recipients also displayed normal dendritic morphology (Fig. 5E). Together, these results demonstrate that the continuous peripheral TCR/ligand interaction, signaling through ITK, impairs the maintenance of Tg sIELs by promoting their apoptosis, which could be corrected by removing ITK-mediated signals.

Although it is increasingly realized that the various subsets of tissue-specific γδ T cells are important components of the immune system critical for the first line of defense, mechanisms regulating their development are poorly understood. Our recent studies found that thymic “educational” processes of different γδ T cell subsets promote their acquirement of unique homing properties (7, 48), suggesting a critical role of the TCR-mediated selection signaling in programming thymic γδ T cells for their tissue-specific development. In this study, we investigated molecular mechanisms underlying the involvement of TCR signaling in tissue-specific development of the skin-specific γδ sIELs and identified ITK as a critical signaling molecule that specifically controls the skin-homing property of fetal thymic sIEL precursors and their seeding into the skin. The fetal thymic sIEL precursors from ITK^−/− mice could not undergo the coordinate switch in expression of S1PR1 and CCR10 after positive selection and had impaired migration abilities toward their ligand attraction, suggesting that ITK-transduced selection signaling is critical to upregulate the expression of these migration molecules for their exit from thymus and migration into the skin (9, 50–52). In addition, considering previous studies that ITK is also involved in chemokine receptor-mediated signaling (22, 26–29), ITK deficiency might potentially impair the migration of the sIEL precursors by directly affecting chemokine receptor signaling.

TCR-mediated selection signaling in the fetal thymic sIEL precursors not only promotes their acquirement of the unique skin-homing property but also endows them with capacities to survive and expand in the skin, such as upregulation of CD122 (IL15Rβ) that is critical for the survival/expansion of sIELs (12, 13). Interestingly, the absence of ITK does not affect the CD122 upregulation and normal maturation of the fetal thymic sIEL precursors. In addition, although there were fewer sIELs in the ITK knockout mice, their in vivo proliferation rates were the same as, if not higher than, those of WT mice, indicating that the ITK-mediating signaling is not involved in controlling the survival/proliferation capacity of the fetal thymic sIELs in the skin. These suggest that different TCR signaling molecules are responsible for promoting the skin-homing and survival/proliferation properties of the selected fetal thymic sIEL precursors. In this regard, it is likely that although other TCR downstream signaling molecules, such as Lck, Syk, and Zap-70, are all involved in the sIEL development (17–20), they could affect different aspects of the development. Consistent with this, sIELs in ZAP-70^−/− mice displayed significant morphological changes, whereas sIELs in ITK^−/− mice maintain the normal dendritic shape (17–19). This difference might reflect the fact that ZAP-70, which is located at the upstream of the ITK signal pathway, may regulate a larger subset of TCR signals than ITK, with correspondingly greater effect (55).

ITK^−/− fetal thymic sIELs proliferate normally in response to TCR stimulation in vitro, consistent with the normal peripheral survival/expansion of sIELs in vivo. In addition, even in Vav1^−/− mice whose Vγ3⁺ fetal thymic sIELs are defective in TCR-mediated proliferation, they still had normal sIEL development, suggesting that TCR signaling-mediated proliferation is not required for the maintenance of sIELs in the skin. This agrees with the notion that TCR-specific ligand(s) of sIELs are not expressed in normal skin but are upregulated on “stressed” or “diseased” keratinocytes that would activate the sIELs for proper functions (56, 57). Therefore, there is a close interplay between the establishment of sIELs and their subsequent function. Considering that positively selected fetal thymic sIEL precursors display an activated “memory”-like phenotype and are independent of TCR signaling for survival and proliferation in the periphery, this suggests that their development is intrathymically programmed through TCR signaling molecule-mediated selection for their specific function.

Not only is the sIEL development independent of the peripheral TCR signaling, but also the continuous stimulation of TCRs in the sIELs impairs their development by promoting apoptosis. Such enhanced cell death was reduced by ITK deficiency, suggesting the involvement of ITK in the TCR-induced activation of sIELs that would result in the apoptosis if persisting. Consistent with this, ITK^−/− sIELs are defective in producing IFN-γ in response to TCR stimulation. Therefore, although the ITK-transduced TCR signaling in the peripheral sIELs is not required for their normal maintenance, it is important for their activation, suggesting that the ITK-transduced thymic and peripheral TCR signals are differentially involved in the development and function of sIELs. However, to fully understand these, how the ITK signaling is involved in the in vivo functions of sIELs needs to be addressed.

There are increasing types of tissue-specific lymphocytes that function in various roles to provide the first lines of defense (1, 58–60). In humans, the preferential distribution of specific T cell subsets in the skin was also reported (61–63). Although the human and murine skin T cells use different TCR compositions, they seem to perform the similar functions. It was reported that like the murine sIELs, the human skin γδ T cells could lyse skin tumor cells and produce similar cytokines in response to stimulation in vitro (61). In addition, human epidermal γδ T cells, as well as epidermal αβ T cells, were shown to contribute to wound healing (63), suggesting that even though they have different TCR usages, the human and murine epidermal T cells share similar functional properties and might develop similarly. Our findings with the murine skin-specific sIELs would aid in understanding how ITK and other TCR-associated signaling is involved in the development of the human skin T cells as well as other different tissue-specific lymphocytes. In addition, in light of the role of ITK in regulating CCR10 expression in the sIEL precursors, whether ITK is involved in its expression in other cells of the skin, such as melanocytes and melanoma tumor cells, under physiological and pathological conditions would also be interesting questions.

We thank Joonsoo Kang for critical comments and Christina Saylor for technical support. The experiments with Vav1^−/− mice were initiated in the laboratory of David Raulet (University of California at Berkeley, Berkeley, CA).

Disclosures The authors have no financial conflicts of interest.