Skin-resident T cells sense ultraviolet radiation-induced injury and contribute to DNA repair

Abstract

Skin-resident T cells have been shown to play important roles in tissue homeostasis and wound repair, however, their role in ultraviolet radiation (UVR)-mediated skin injury and subsequent tissue regeneration is less clear. Here, we demonstrate that acute UVR rapidly activates skin-resident T cells in humans and dendritic epidermalγδ T cells (DETC) in mice through mechanisms involving the release of ATP from keratinocytes. Following UVR, extracellular ATP leads to an increase in CD69 expression, proliferation, and IL-17 production, and to changes in DETC morphology. Furthermore, we find that the purinergic receptor P2X7 and caspase-1 are necessary for UVR-induced IL-1 production in keratinocytes, which increases IL-17 secretion by DETC. IL-17, in turn, induces epidermal TNF related weak inducer of apoptosis (TWEAK) and Growth arrest and DNA damage associated gene 45 (GADD45), two molecules linked to the DNA repair response. Finally, we demonstrate that DETC and human skin-resident T cells limit DNA damage in keratinocytes. Together, our findings establish a novel role for skin-resident T cells in the UVR-associated DNA repair response and underscore the importance of skin-resident T cells to overall skin regeneration.

Introduction

Excessive exposure to ultraviolet radiation (UVR), in particular in the mid-wave length (UVB, 290-320nm), leads to inflammatory reactions of the skin, including sunburn and skin aging. It is the major risk factor for the development of skin cancers such as squamous cell carcinomas (SCC) and their precursor lesions actinic keratoses. Each year, there are more new cases of skin cancer than the combined incidence of cancers of the breast, prostate, lung, and colon (1).

Human skin is populated by 1-2×10¹⁰ resident T cells, a number which exceeds that of circulating T cells (2). This population of T cells resides within the (supra-)basal epidermis and upper dermis and is comprised of αβ and γδ T cells (3). The relevance of an intact T cell immune response for skin cancer surveillance is supported by observations that SCC are particularly numerous in patients taking T cell immunosuppressants (4) and that SCC are characterized by a numeric reduction of skin-resident T cells (4,5). Dendritic epidermal T cells (DETC) reside in the epidermis of mouse skin in immediate contact with neighboring keratinocytes. DETC express an invariant TCR containing the Vγ3 and Vδ1 chains and recognize a yet unidentified antigen expressed by damaged or diseased keratinocytes as well as other immune receptors (6-9). Murine dermis harbors both αβ andγδ T cells (10,11). The in vivo relevance of DETC and human skin-resident T cells to cutaneous repair and immunity is supported by previous findings demonstrating the essential role of these sentinel cells to the wound healing response, antimicrobial barrier function, and tissue surveillance (2,3,6,12,13). However, studies have not been performed to elucidate early immunological mechanisms exerted by skin-resident T cells in acute UVR-induced skin injury. Despite differences in T cell compositions in humans and mice, the importance of skin-resident T cells for protective skin surveillance function is highly comparable. Therefore, immunological studies utilizing DETC are not only crucial to investigate the role of murine epithelial γδ T cell biology, but are also likely to uncover mechanisms of immune cell interactions and inflammatory mediators that operate to control UVR-induced damage in human skin.

Increase of extracellular ATP (eATP) acts as an early and sensitive signal of cellular stress and dying cells. Changes in eATP levels control functional responses of excitatory and non-excitatory cells through activation of purinergic receptors, including the ionotropic P2X and metabotropic G-protein-coupled P2Y receptors (14). Notably, eATP regulates not only innate immune responses, but has recently been linked to adaptive immunity as well (15,16). Keratinocytes are sensitive to UVR and rapidly release ATP following UVR (17). However, the role of eATP in cutaneous immune function is not well understood. Based on their sentinel role, we hypothesized that skin-resident T cells sense UVR-induced ATP release and provide protective surveillance and repair functions in the context of keratinocyte UVR damage, early before carcinogenesis evolves.

In the present study, we demonstrate that UVR-induced ATP release leads to human skin-resident T cell and DETC activation. UVR increases CD69 expression and IL-17 production by skin-resident T cells and DETC in an eATP-dependent manner. IL-17, in turn, upregulates epidermal TNF related weak inducer of apoptosis (TWEAK) and Growth arrest and DNA damage associated gene 45 (GADD45), two genes with known functions in DNA repair (18,19). We furthermore demonstrate that human skin-resident T cells and DETC play a critical role in limiting UVR-induced DNA damage-associated γH2AX and CPD formation in keratinocytes. Together, this study identifies a previously unknown role of skin-resident T cells in sensing solar injury and potentiating the keratinocyte DNA repair response. Our findings indicate that the eATP pathway could be therapeutically targeted to alter susceptibility or treat UV-induced skin cancer and may offer an alternative to phototherapy.

Materials and Methods

Human skin samples, cell preparation and stimulation

This study was approved by the Scripps Investigational Review Board. Normal skin samples were obtained from otherwise discarded tissue from plastic surgery procedures performed at Scripps Green Hospital, La Jolla, CA and Scripps Clinic Ambulatory Surgical Center Carmel Valley, San Diego CA. Tissue samples were used to perform skin organ cultures or to obtain skin-resident T cells as previously described (2,3) with the exception that cells were kept in complete RPMI 1640 medium with 10% FCS without cytokines or, for experiments where T cell supernatants were used, were stimulated in EpiLife medium (Cascade Biologics). Anti-CD3 (OKT3, Biolegend) antibodies were diluted in ELISA coating buffer (50mM Tris, 150mM NaCl; pH 8.0) and immobilized to individual wells of 96-well flat-bottom microtiter ELISA plates (Immulon). Approximately 0.5×10⁶ ml⁻¹ skin-resident T cells were stimulated with plate-bound anti-CD3 antibody (OKT-3), ATP (2 mM, Sigma-Aldrich), apyrase (15U ml⁻¹, Sigma-Aldrich), keratinocyte conditioned medium (described below) or combinations thereof, as stated in figure legends. Skin organ cultures from healthy donors were stimulated for 24 hours with rhIL-17 (200 ng ml⁻¹, R&D), skin-resident T cell supernatants, or were irradiated with an EB-280C/12 UVB lamp (200mJ cm⁻², Spectroline, predominant emission 312nm, 270-390nm emission range) before immunofluorescence staining was performed. Normal human keratinocytes were purchased from Cascade Biologics and were grown in serum-free EpiLife cell culture medium containing 0.06 mM Ca²⁺, EpiLife Defined Growth Supplement, 50 U ml⁻¹ penicillin and 50mg ml⁻¹ streptomycin. Keratinocyte cultures were maintained for up to five passages. Keratinocytes were used at approximately 75-80% confluence and were stimulated in 6-12 well-plates (Corning) with rhIL-17, rhIL-4, or rhIFNγ(R&D) for subsequent RNA isolation or were stimulated in 2-4 well chamber slides (Lab-Tek) with cell-free supernatants from anti-CD3ε-activated skin-resident T cells (T cell supernatants, TC sups) and were irradiated with 15mJ cm⁻² UVB before immunohistochemical analyses. For some experiments, cell culture medium was collected from human keratinocytes, called keratinocyte conditioned medium, from either UVR-treated or non-treated keratinocytes, which were used for stimulation experiments of human skin-resident T cells.

Mice and in vivo UV irradiation

C57BL/6, Rag^−/−, P2×7^−/−, and Tcrd^−/− mice on the C57BL/6 background were purchased from The Jackson Laboratory. Il17a^−/− mice were kindly provided by Dr. Y. Iwakura, University of Tokyo, Japan and Dr. K. Ley, La Jolla Institute for Allergy and Immunology, CA. We received P2×7^−/− mice as a kind gift by Dr. K. Mowen from The Scripps Research Institute, CA who purchased these mice from Jackson Laboratory, which originally received this strain from Pfizer Pharmaceuticals. Casp1^−/− mice were a kind gift of Drs. R. Flavell from Yale and R. Ulevitch from The Scripps Research Institute, CA. Rag^−/− mice were kindly provided by Dr. C. Surh from The Scripps Research Institute, CA. All mice were bred at The Scripps Research Institute and all animals were housed in specific pathogen-free conditions according to The Scripps Research Institute IACUC guidelines. For some experiments, mice were irradiated with UVB (100 mJ cm⁻²) after removing their back hair and were sacrificed after 6-8 hours (CD69 staining) or 18 hours (mRNA analyses). All studies were reviewed and approved by IACUC at The Scripps Research Institute, La Jolla, CA.

Mouse skin preparation and cell stimulation

DETC were freshly isolated from mouse skin as previously described (7) and were cultured in complete RPMI 1640 containing 10% FCS. For generation of DETC short term cell lines, epidermal cells were cultured in complete RPMI 1640 containing 10% FCS in the presence of 5 U ml⁻¹ IL-2. Approximately 0.5-1×10⁶ ml⁻¹ cells were stimulated in complete DMEM or complete RPMI 1640 with 10% FCS with various concentrations of plate-bound anti-CD3ε antibody (500A2), ATP (2 mM), apyrase (15 U ml⁻¹), anakinra (150 μg ml⁻¹, kind gift from Dr. H. Hoffman, University of California San Diego, CA) or combinations thereof, as stated in figure legends. For some experiments, skin organ cultures were irradiated with UVB (100 mJ cm⁻²) in the presence or absence of apyrase (15 U ml⁻¹) for 8 hours before isolating DETC. For analyses of DETC rounding, ears were excised, separated into dorsal and ventral ear halves, and were floated epidermis side-up in complete DMEM (Dulbeccos) with 10% FCS and were irradiated with 100 mJ cm⁻² UVB (24 hours) or were treated with ATP (2 mM, 3 hours). Primary murine keratinocytes were isolated and maintained in defined K-FSM medium (Gibco). Cell culture supernatants, designated as ‘keratinocyte conditioned medium’, were collected from keratinocytes within 30-60 minutes following irradiation with UVB (15 mJ cm⁻²) or without. This keratinocyte conditioned medium or control medium was used in some experiments to stimulate DETC.

Flow cytometry and fluorescence-activated cell sorting

Antibodies and appropriate IgG controls were conjugated to FITC, PE, PerCP-Cy5.5, Pe-Cy7, Pacific blue or APC. Antibodies to Vγ3 (536), TCRβH57-597), CD4 (GK1.5) and pan-γδTCR (GL-3) were purchased from BD Bioscience. Thy1.2 (53-2.2), CD3 (HIT3a, 17A2 and 145-2C11), CD45 (HI30) antibodies were purchased from Biolegend. Antibodies to IL-17 (ebio64DEC17), CD8 (53-5.8) and CD69 (H1.2F3), were purchased from eBioscience, antibody to P2X1 receptor was purchased from Alomone lab. For detection of intracellular IL-17, human skin-resident T cells were stimulated with keratinocyte conditioned medium for 18 hours in the presence or absence of apyrase (15 U ml⁻¹) and were restimulated with PMA (2ng ml⁻¹) and ionomycin (1μM) for additional 3 hours in the presence of brefeldin A (5μl ml⁻¹) and monensin (eBioscience, 1:1000), before cells were fixed and permeabilized with cytofix and cytoperm reagents from BD Bioscience. Cells were acquired with DiVa 5.0 software on a Digital LSRII and analyzed with FlowJo software (Tree Star, Inc.). For some experiments, DETC were purified by FACS-sorting based on Vγ3⁺Thy1.2⁺ or CD3⁺CD45⁺ expression, keratinocytes were sorted from epidermal cell suspensions being CD45⁻CD3⁻; the purity of sorted cell populations was between 91-99%.

Proliferation, ELISA, and bioluminescence assays

For proliferation assays and ELISA, approximately 0.5-1×10⁶/ml cells were stimulated as stated in figure legends. For proliferation assays, DETC or human skin-resident T cells were stimulated for 18-24 hours before addition of 0.5 μCi/well [³H]thymidine (MP Biomedicals). Samples were harvested 14-18 hours later and [³H]thymidine incorporation was measured using a Beckman LS181 scintillation counter (Beckman Coulter). For analyses of secreted cytokines, supernatants were removed at various time points and immediately stored at −20°C until use. Supernatants from DETC were analyzed for presence of IL-17A by ELISA (eBioscience). Bioluminescence-based ATP measurements (Roche) were performed on supernatants collected from UVR-treated or non-treated skin organ cultures.

RNA isolation and RT-PCR

Total RNA was isolated from cells using RNeasy Micro Kit (Qiagen) or from tissue using the TRIZOL reagent (Invitrogen). RNA was reverse transcribed using the iScript cDNA Synthesis Kit (BioRad) and resulting cDNA was amplified using FastStart Universal SYBR Green Master Mix (Roche). Primers for amplification were used as previously described (13). Fold induction of gene expression was normalized to β-actin and calculated using the 2^(−ΔΔCt) method.

Epidermal sheet and skin immunofluorescence

Epidermal ear sheets from mice were stained with PE-conjugated antibodies to Vγ3 and DAPI (Sigma-Aldrich) as previously described (13). Quantification of DETC numbers was performed using Image J software. For histological analysis of γH2AX (Active Motif), CPD (Kamiya Biomedical Company), and TWEAK (CARL-1, AbD serotec), tissue was methanol- or paraformaldehyde-fixed and incubated in a blocking solution of PBS containing 2.5% normal goat serum (Jackson Immuno Research), 2.5% normal donkey serum (Jackson Immuno Research), 1% bovine serum albumin (Calbiochem), 2% fish gelatin (Sigma), and 0.1% Triton-X100 for 1 hour at room temperature before incubation with primary antibodies or appropriate IgG controls overnight at 4°C. Secondary antibodies were FITC, PE- or Cy3-conjugated (Jackson ImmunoResearch Laboratories). After subsequent washing, sections were mounted with Prolong Gold antifade containing DAPI (Invitrogen). Sections were visualized using a Nikon Eclipse E800 microscope and digital images were acquired with a Zeiss AxioCam HRc camera.

Statistical analyses

Data are presented as means, error bars are SEM. Data shown are representative of at least three independent experiments unless otherwise indicated. Statistical significance was measured using two-tailed Student t-test. A P value of <0.05 was considered significant. *p<0.05, **p<0.01, ***p<0.001.

Results

Soluble factors released from UV-irradiated keratinocytes activate DETC in mice

UVR is a common environmental hazard of the skin, however its role in skin-resident T cell immunology is not well understood. First, we investigated whether exposure to UVR activates DETC in WT mouse skin. Like other tissue-resident T cells, DETC are present in a semi-activated state and constitutively express CD69 (2,8,11, 20). Once activated, DETC upregulate CD69 expression and change their cellular morphology from a dendritic to a circular cell shape (7,21). Following in vivo UVR treatment, we observed an increase in CD69 expression on DETC (Fig. 1a). DETC rounding began within 30 minutes following UV radiation and by 3 hours, the majority of DETC had a rounded phenotype that was retained for at least 18 hours post UVR (Fig. 1b, c). These findings suggested that acute UVR leads to DETC activation.

Figure 1. UV radiation rapidly activates DETC.

(a) Representative flow cytometric analysis of CD69 expression on DETC 8 hours following in vivo UV radiation (UVR) treatment of the skin (gated on live Thy1.2⁺Vγ3⁺). Gray shaded graph indicates IgG control for non-UVR-treated skin and was similar to UVR-treated control IgG (data not shown). (b) Detection of DETC morphology changes by immunofluorescence staining of epidermal ear sheets following in vitro UVR-treatment with anti-GL-3, specific for the γδTCR. (c) Quantification of DETC from (b) with indicated morphology. Data are presented as mean from at least n=350 DETC, error bars=SEM. h, hours

Soluble factors released from epithelial cells have been shown to modulate T cell responses (22-25), therefore we next assessed the capacity of cell-free supernatants from UVR-treated keratinocytes to induce DETC proliferation (Fig. 2a). Supernatants collected from UVR-treated or mock-treated keratinocytes 30 minutes post UVR exposure increased DETC proliferation in the presence, but not in the absence of TCR-stimulation (Fig. 2a and data not shown), suggesting that a soluble factor from keratinocytes is rapidly released upon UVR and enhances TCR-mediated DETC activation. Expanding the previous finding of increased ATP release from keratinocytes upon UVR (17), we detected increased levels of eATP in supernatants from UVR-treated skin organ cultures (Fig. 2b), indicating a possible role for eATP in the acute UV response. This idea was supported by three observations: First, enzymatic hydrolyzation of ATP by apyrase inhibited the capacity of supernatants from UVR-treated keratinocytes to increase DETC proliferation (Fig. 2a). Second, apyrase treatment of UVR-treated skin organ cultures blocked the UVR-induced increase in CD69 expression on DETC (Fig. 2c). Third, exogenous ATP was sufficient to mediate DETC activation in situ, as eATP alone stimulated DETC rounding, visualized by immunofluorescent staining of epidermal ear sheets with antibodies recognizing the γδTCR (Fig. 2d). Together, these findings strongly suggested a previously unrecognized role for UV-induced eATP in regulating DETC activation.

Figure 2. UV-irradiated keratinocytes activate murine DETC in an ATP-dependent manner.

(a) UVR-induced ATP release from keratinocytes promotes DETC proliferation in the presence of anti-CD3 stimulation. Proliferation was measured in triplicates by ³H-thymidine incorporation in DETC 48 hours following stimulation with conditioned medium collected from keratinocytes that were either UVR-treated (Kerat. cond. med., UV) or untreated (Kerat. cond. med., no UV) in the presence or absence of apyrase. Culture medium alone and apyrase alone were included as controls and did not stimulate DETC proliferation (data not shown). Data are presented as means from triplicates, error bars=SEM. (b) Measurement of ATP in the supernatants from mock- and UVR-treated skin organ cultures by bioluminescence. Data are presented as means, n=5, error bars=SEM (c) UVR-induced CD69 expression on DETC is mediated via extracellular ATP. Representative flow cytometric analyses of CD69 on DETC 6 hours following UVR-treatment of skin organ cultures in the presence or absence of apyrase (gated on live Thy1.2⁺Vγ3⁺). Gray shaded graph indicates IgG control for non-UVR-treated skin and was similar to UVR-treated control IgG or apyrase-treated control IgGs (data not shown). (d) DETC change their cellular morphology following ATP stimulation. Ear sheets were treated with 2mM ATP for 3 hours and DETC morphology was evaluated as described in Fig. 1b. *p<0.05 by two-tailed student’s t-test.

Murine epidermal cell populations express distinct ATP-sensing P2 receptors

Extracellular ATP signaling is mediated through membrane-bound purinoreceptors and is implicated in the regulation of both innate and adaptive immune responses (15,26,27). Among the studied P2X receptors, P2X1, P2X4, and P2X7 receptors have been previously described to be expressed by peripheral αβ and γδ T cells (28). However, the presence of purinoreceptors on DETC and the functional role of eATP signaling for DETC biology is unknown. Analyses of epidermal cell populations revealed that DETC and keratinocytes express distinct P2 receptors (Fig. 3), suggesting that eATP serves multiple functions in the skin, in line with previous reports (27). The mRNAs encoding the ionotropic P2X1, P2X2, P2X3, P2X5 and the metabotropic P2Y6 and P2Y12 receptors are expressed at higher levels by DETC compared to keratinocytes (Fig. 3). In contrast, P2X4, P2X7, and P2Y1 receptors are expressed at higher levels by keratinocytes compared to DETC. These results suggested that epidermal cells express several distinct purinoreceptors.

Figure 3. Distinct expression of purinoreceptors in murine keratinocytes and DETC.

Representative RT-qPCR analysis of various P2X and P2Y receptors in FACS-sorted populations of DETC and epidermal keratinocytes from mouse skin.

Dual functions of extracellular ATP for murine DETC activation

While under steady state conditions, eATP levels are low, transient eATP increases facilitate desensitization of distinct purinoreceptors, whereas upon cell stress or cell burst, eATP concentrations rise rapidly and can activate purinoreceptors with even low binding constants to eATP, such as the P2X7 receptor. Our laboratory has recently shown that a subset of DETC produces IL-17A following acute skin injury to promote wound repair (13). As eATP was found to be increased in UV-irradiated mouse skin (Fig. 2b), a form of skin injury, and DETC are shown to readily express multiple P2X and P2Y receptors (Fig. 3), we hypothesized that eATP may modulate IL-17 production by DETC. When pure populations of Vγ3⁺ DETC were stimulated with eATP in the presence of low concentrations of immobilized anti-CD3, DETC produced significantly more IL-17 than when stimulated with low concentrations of immobilized anti-CD3 alone. The response of DETC to eATP could be blocked by the preincubation of ATP with apyrase (Fig. 4a), whereas addition of apyrase alone or eATP alone did not alter IL-17 production (Fig. 4a and data not shown).

Figure 4. ATP increases IL-17 production by murine DETC via direct and indirect mechanisms.

(a) IL-17 production by FACS-sorted populations of anti-CD3-stimulated WT DETC 24 hours after addition of exogenous ATP in the presence or absence of apyrase. Data are presented as means from triplicates, error bars=SEM (b) IL-17 production by WT DETC 24 hours following stimulation with conditioned medium collected from UVR-treated (Kerat. cond. medium, UV) or non-treated (Kerat. cond. medium, no UV) keratinocytes in the presence or absence of apyrase or the IL-1R antagonist anakinra. (c) Conditioned medium from UVR-treated WT keratinocytes induces significantly more IL-17 secretion by anti-CD3-activated WT DETC 24 hours following stimulation than conditioned medium from P2×7^−/− and Casp1^−/− UVR-treated keratinocytes. Data are presented as means from triplicates, error bars=SEM, *p<0.05, **p<0.01 by two-tailed student’s t-test.

We then sought to test whether conditioned culture medium from UVR-treated keratinocytes was sufficient to increase IL-17 secretion by DETC. Indeed, stimulation of WT DETC with conditioned medium from UVR-treated, but not mock-treated WT keratinocytes, increased IL-17 production by WT DETC in the presence of low concentrations of immobilized anti-CD3 (Fig. 4b). Apyrase inhibited the capacity of keratinocyte conditioned medium from UVR-treated WT keratinocytes to induce IL-17 secretion by WT DETC (Fig. 4b). These results suggested that UVR-treated keratinocytes can stimulate WT DETC cytokine production through an ATP-dependent mechanism.

Keratinocytes are a rich source of IL-1α, and to a lesser extent IL-1β. Both, IL-1α and IL-1β have been shown to play roles in modulating IL-17-T cell responses (13,29). Previous studies from our laboratory have demonstrated that recombinant IL-1β together with IL-23 increases IL-17 production by WT DETC in the presence of TCR stimulation (13). We therefore tested the possibility that UVR-induced IL-1 release from keratinocytes affects WT DETC IL-17 production. The biological activity of IL-1α and IL-1β can be simultaneously blocked by antagonizing the IL-1 receptor (IL-1R). When anakinra, a clinically used recombinant, non-glycosylated form of the human IL-1R antagonist (IL-1Ra), was added to WT DETC cultures during stimulation with supernatants from UVR-treated keratinocytes, IL-17 secretion by DETC was significantly decreased (Fig. 4b).

IL-1β secretion is mediated by a pathway involving NOD-like receptor family, pyrin domain containing 3 (Nlrp3)-inflammsome and caspase-1 activation, following binding of eATP to the P2X7 receptor, whereas IL-1α secretion is at least partially dependent on IL-1β secretion (29-32). Compared to other members of the P2 receptor family, P2X7 receptor has only a low binding constant to eATP, acting especially in response to high eATP levels and is therefore considered a danger signaling receptor (33,34). In contrast to DETC, keratinocytes express the P2X7 receptor (Fig. 3 and ref. (33,34)), as well as all components for Nlrp3-inflammasome assembly (31, 33, 34). To test the possibility that activation of P2X7 receptors by UVR-induced ATP release from keratinocytes induces IL-1 to stimulate IL-17 production by WT DETC, conditioned medium from UVR-treated WT, P2×7^−/−, or Casp1^−/− keratinocytes were added to WT DETC cultures. Increased IL-17 production by WT DETC was observed only when conditioned medium from WT keratinocytes was added to DETC cultures, whereas conditioned medium from UVR-treated P2x7^−/− or Casp1^−/− keratinocytes had minor effects on IL-17 production by WT DETC (Fig. 4c). In concert with the finding that secretion of IL-1β and to a far lesser extent also IL-1α, were diminished in P2×7^−/− keratinocytes upon UV treatment (Suppl. Figure 1), our findings suggest that UVR induces ATP-mediated IL-1 release from keratinocytes which increases IL-17 production by WT DETC. Together, these results highlight that both direct and indirect mechanisms may account for eATP-mediated DETC activation and IL-17 production.

WT but not Tcrδ^−/− mice are protected from UVR-induced DNA damage-associated γH2AX and CPD formation

Keratinocytes are sensitive targets of UVR which causes DNA damage characterized by DNA double-strand breaks (DSB), phosphorylated histone 2A variant H2AX (γH2AX) and cyclobutane pyrimidine dimer (CPD) formation. If proper DNA repair fails, cells may undergo cellular proliferation and oncogenic development. Tissue-resident T cells, such as DETC, provide local surveillance functions (35-39). Therefore, we hypothesized that following UVR-induced activation, DETC may initiate a protective response to limit UV damage to keratinocytes early before skin carcinogenesis can evolve. To test this hypothesis, we assessed the DNA damage response in keratinocytes 3, 24, and 48 hours following UVR treatment. Cells positive for γH2AX, a marker of DSB damage ³⁹, as well as CPD⁺ cells were observed with similar frequency in skin of WT and Tcrδ^−/− mice at 3 hours following UVR (Fig. 5a-c), indicating that equal DNA damage occurred. In contrast, at 24 or 48 hours following UVR, a higher frequency of γH2AX⁺ and CPD⁺ cells was observed in UV-exposed skin from Tcrδ^−/− compared to WT mice (Fig. 5a, b and c). Together, these results suggested that DNA repair of UV-induced lesions is reduced in the absence of γδ T cells in the skin of mice. To further demonstrate that UV-induced eATP could to initiate an epidermal response aimed at limiting DNA damage, we also tested whether eATP reduces the frequency of CPD formation in UVR-treated skin. Pretreatment of epidermal sheets with eATP prior to UVR, resulted in only a minor reduction of CPD formation measured at 24 hours, whereas pretreatment with apyrase enhanced the frequency of CPD⁺ cells following UV compared to eATP treated epidermal ear sheets but the increase was not significant compared to non-treated ear sheets (Suppl. Figure 2).

Figure 5. Tcrδ^−/− mice have more DNA-damaged keratinocytes upon UVR than WT mice.

(a) γH2AX expression (green) in epidermal ear sheets 3h and 24h following UVR-treatment of WT and Tcrδ^−/− skin. Cell nuclei are stained with DAPI (blue). (b) Quantification of γH2AX⁺ cells from Fig. 5a. Data are presented as means from at least 96 cells per genotype, error bars=SEM (c) Quantification of CPD⁺ cells in epidermal ear sheets in WT and Tcrδ^−/− skin similar to (a, b). (d, e) RT-qPCR analysis of epidermal Gadd45 and Tweak expression 18 hours following in vivo UVR-treatment. Data are presented as means from triplicates, error bars=SEM, *p<0.05,**p<0.01 by two-tailed student’s t-test.

GADD45 is a major participant in genomic stability and DNA repair (19,36,38,40). It exhibits low constitutive expression, is predominantly intra-nuclear, and becomes transcriptionally activated by UVR, hyperoxia, and endotoxin (37,41,42). The relevance of GADD45 to the UV protection response is supported by previous observations that GADD45 is essential for protection against UV-induced skin cancerogenesis (19). TWEAK is a soluble protein known to bind to the TWEAK receptor, the fibroblast growth factor-inducible 14 receptor (FN14), and has been previously shown to induce Gadd45 (18). Here, we show that Gadd45 and Tweak and are upregulated upon in vivo UVR treatment in the epidermis from WT mice (Fig. 5d,e). In contrast, Tcrδ^−/−, Il17a^−/− and Rag^−/− mice showed impaired upregulation of Gadd45 and Tweak following UVR (Fig. 5d,e), demonstrating that DETC and IL-17 play critical roles in the regulation of Gadd45 and Tweak expression in the skin. Thus, the lack of GADD45 and TWEAK in Tcrδ^−/− skin following UVR may contribute to defective keratinocyte DNA repair, as demonstrated by the higher frequency of γH2AX⁺ and CPD⁺ keratinocytes in Tcrδ^−/− mice.

Human skin-resident T cells contribute to the UVR response

Identification of the protective function of DETC in the UVR response in mice and the critical role of human skin-resident T cells in skin immunity, raised the possibility that human skin-resident T cells also sense solar injury. To assess whether the eATP-skin-resident T cell-axis is active in human skin, we first examined whether human skin-resident T cells respond to eATP. Increased proliferation was observed in human skin-resident T cell cultures following eATP stimulation, independent of the presence of anti-CD3 stimulation (Fig. 6a). Several mRNAs of purinergic receptors, including those encoding P2X1, P2X4, and P2X7 receptor have been recently identified in human T cells ²⁸. Since P2X1 receptor expression, but not P2X4 or P2X7 receptor expression was high in murine DETC (Fig. 3), we focused on the human P2X1 receptor, and demonstrate that it is detected on activated human skin-resident T cells (Suppl. Fig. 3a,b). Stimulation of human skin-resident T cells with conditioned medium from normal human keratinocytes that were subjected to UVR, allowed further investigation of the involvement of keratinocyte-derived ATP in skin-resident T cell activation. IL-17 production in skin-resident T cells was observed following stimulation with conditioned medium from UVR-treated keratinocytes (Fig. 6b). The absolute frequency of IL-17 producing skin-resident T cells did not dramatically change between stimulation with conditioned medium from UVR-treated keratinocytes and conditioned medium from non-UVR treated keratinocytes, however, UVR-treated keratinocyte supernatants increased the proportion of T cells producing high amounts of IL-17 (Fig. 6b). This subset of IL-17^hi producers was sensitive to ATPase treatment (Fig. 6b). Furthermore, recombinant human IL-17 (rhIL-17) upregulated GADD45A and TWEAK mRNA levels in human cultured keratinocytes, whereas rhIL-4 exerted an effect only on GADD45; and IFNγ had minor effects on GADD45 and TWEAK (Fig. 6c). Increased TWEAK immunoreactivity was also observed in human skin organ cultures treated with rhIL-17 and was localized to the epidermal layer, suggesting that keratinocytes are the major cell type in the skin to upregulate TWEAK upon IL-17 stimulation (Fig. 6d). Finally, supernatants from activated skin-resident T cells reduced the frequency of γH2AX⁺ cells in human UVR-treated skin (Suppl. Fig. 3c). To confirm that it was UVR-induced DNA damage present in keratinocytes that was blocked by skin-resident T cell supernatants, cultured keratinocyte monolayers were analyzed for γH2AX immunoreactivity (43) following UVR treatment (Fig. 6e). UVR-treated keratinocytes showed a higher frequency of γH2AX⁺ keratinocytes than did those incubated with skin-resident T cell supernatants (Fig. 6e, f). Together, these results show a novel function of human skin-resident T cells in providing protection for keratinocytes against UVR-induced DNA damage.

Figure 6. Human skin-resident T cells promote DNA repair of UV-damaged keratinocytes.

(a) Proliferation of human skin-resident T cell explant cultures was measured in duplicate following treatment with exogenous ATP in the presence or absence of anti-CD3-stimulation. Data are presented as means, error bars=SEM. (b) Intracellular flow cytometry for IL-17 in skin-resident T cells upon stimulation for 18 hours with conditioned medium from UVR-treated (Kerat. cond. med., UV) or non-treated keratinocytes (Kerat. cond. med., no UV) or control medium in the presence or absence of apyrase. The percentage of IL-17A⁺ cells is shown in the right corner, numbers of IL-17^high and IL-17^low cells are shown inside marked boxes. Gated on live CD45⁺CD3⁺cells. (c) RT-qPCR analysis of GADD45 and TWEAK in human keratinocytes upon stimulation for 18 hours with rhIL-17, rhIL-4, and rhIFNγor vehicle control. (d) TWEAK protein expression (red) in human skin organ cultures following treatment with rhIL-17A for 24 hours. Cell nuclei were stained with DAPI (blue). Dashed line represents the epidermal-dermal border. (e) γH2AX expression (red) in cultured human keratinocytes which were treated for 24 hours with supernatants from activated human skin-resident T cells following UV-irradiation. DAPI staining (blue) visualizes nuclei. Dashed line represents the epidermal-dermal border. (f) Quantification of γH2AX⁺ keratinocytes treated as indicated in (e). Data are presented as means from at least 700 keratinocytes per condition, error bars=SEM. **p<0.01,***p<0.001 by two-tailed student’s t-test.

Discussion

UVR-induced DNA damage has been causatively linked to many skin cancers, including SCC (1, 16, 44). An intact T cell immune system is essential to maintain tissue surveillance and prevent skin carcinogenesis (5,16). In fact, T-cell immunosuppression bears a high risk for cutaneous SCC development and skin-resident T cell numbers are critically reduced in human cutaneous SCC lesions, supporting the idea that impaired function or loss of this protective T cell population in the skin may be associated with SCC development. However, the role of skin-resident T cells in the acute response to cutaneous UVR has not been well studied. Since skin-resident T cells play critical roles in cutaneous immunity, we examined the status of these sentinel cells in UV-irradiated skin of humans and mice and found that these cells become activated upon UV exposure and play a novel and yet unrecognized role in the epidermal DNA repair response.

UVR leads to skin-resident T cell activation through a mechanism involving eATP, a danger signaling molecule. Our study demonstrates that release of ATP from UVR-treated keratinocytes results in autocrine and paracrine immune responses, ultimately promoting skin-resident T cell activation. The mechanisms by which eATP affects DETC activation may occur on multiple levels, including cell morphological changes, proliferation, Ca²⁺ influx, and effects on neighboring keratinocytes (summarized in Suppl. Fig. 4). Interestingly, activation of the P2X1 receptor has been previously linked to changes in cell shapes and Ca²⁺ influx (45), whereas P2X7 receptor activation has been previously linked to inflammasome-mediated IL-1 secretion (46). In line with the latter observation, we found that IL-17 production in skin-resident T cells is enhanced by UVR-induced IL-1 release from keratinocytes. Furthermore, we demonstrate that eATP enhances TCR-mediated DETC activation, but is ineffective in the absence of TCR stimulation. This result is in line with previous findings from our laboratory showing that DETC require TCR stimulation for optimal activation (13). Hence, knowledge about the nature of the yet unidentified DETC TCR ligand will help to further delineate DETC biology.

ATPase treatment or IL-1R-blockade diminished the production of IL-17 by WT DETC when stimulated with UVR-treated WT keratinocyte supernatants. This important finding is consistent with observations from this work and others that UVR induces IL-1 secretion, and that WT keratinocytes have the cellular machinery to orchestrate P2X7 receptor signaling, inflammasome assembly, and subsequent IL-1 maturation and processing (29,46). Our findings expand these observations by directly showing a link between UVR, ATP release, keratinocyte activation, IL-1 release, and DETC IL-17 production (summarized in Suppl. Fig. 4). We cannot exclude the possibility that eATP may also activate other epidermal cells, such as Langerhans cells in vivo, which express multiple purinoreceptors as well (unpublished observation and ref. (47)) to provide additional IL-1, (48) but may also contribute through other ATP-dependent biological effects to the cutaneous UVR response in vivo (49). Interestingly, treatment of mouse skin with a P2X7 receptor ligand has been shown to inhibit formation of skin papillomas and carcinomas in the murine two-stage-carcinogenesis model (50), suggesting a protective role for P2X7 receptor activation. In addition, our results on IL-1R blockade further support clinical precaution advisories stating that, while taking the IL-1R antagonist anakinra, patients should practice enhanced skin cancer prevention, i.e. sun protection, as approximately 9% of patients under anakinra therapy have been reported to develop skin cancers (51). Together, our data strongly support that eATP signaling in the skin is an important pathway for DETC activation following UVR exposure and may be important to alarm the immune system about solar skin damage.

We demonstrate that human skin-resident T cells were activated by eATP, however, in contrast to DETC, they did not require concomitant stimulation through the TCR. For our studies, human skin-resident T cell explant cultures were used. It is possible, that explant cultures may contain other skin cells that contribute to human skin-resident T cell activation and which may explain this difference. Nonetheless, our results suggest that eATP-mediated signals are critical for both DETC and human skin-resident T cell activation.

Our findings may have multiple implications for human health. SCC comprise epithelial-derived cancers of the skin, lung, esophagus, urinary bladder, prostate, lung, cervix and often develop over many years through a multi-step process, including initiation, proliferation, and progression; and are accompanied by local or systemic immunosuppression (52,53). Cutaneous SCC are frequently caused by excessive UV irradiation and UVR-induced DNA-damaged keratinocytes bear the risk to eventually develop into carcinomas. We find that skin-resident T cells contribute to the DNA repair response in keratinocytes upon acute UVR exposure, supporting their critical role in skin homeostasis and surveillance function. Thus, our findings raise the possibility that skin-resident T cells may be involved in the very early control of DNA damage in keratinocytes to protect from skin cancer development. This newly discovered role for skin-resident T cells may explain previous clinical reports that T cell-immunosuppressive drugs, such as cyclosporine, are associated with higher risk of SCC development, as often observed in organ transplant patients (16). Therefore, this study increases the overall understanding of epidermal repair responses upon acute solar injury and provides novel and previously unrecognized insight into the role of skin-resident T cells in skin immunity. Future studies are necessary to elucidate the susceptibility to photo-carcinogenesis in WT and Tcrδ^−/− mice, which will be critical in further defining the roles and requirements of DETC in epidermal surveillance function and long-term maintenance of genomic integrity.

DNA is the major target of direct or indirect UV-induced cellular damage. The DNA damage response is complex, and comprises cell-cycle checkpoint control and DNA repair, which collectively are aimed to prevent photo-carcinogenesis. Consistent with the idea that skin-resident T cells serve host protective functions, we find that the presence of DETC increased Gadd45a and Tweak while decreasing the frequency of DNA damage associated-γH2AX⁺ and CPD⁺ keratinocytes upon UVR. This critical finding was confirmed in human skin cells as well, corroborating the idea that skin-resident T cells serve critical immune functions in the keratinocyte DNA repair response. Previous studies demonstrated that GADD45 associates with mononucleosomes that have been altered by UVR, allowing GADD45 to recognize an altered chromatin state and modulate DNA accessibility to cellular repair proteins (54,55). Furthermore, the lack of G2/M arrest coupled with reduced DNA repair leads to higher UVR sensitivity of Gadd45-deficient keratinocytes (56). Together, these findings may explain the underlying mechanisms for the essential role of GADD45 in protection against UV-induced skin carcinogenesis (19). TWEAK is known to regulate GADD45 and CDC2-phosphorylation (18). However, TWEAK expression is either reduced or increased in cutaneous malignancies suggesting a more complex role in carcinogenesis and inflammatory responses (57). TWEAK may exert functions in tissue repair as well as destructive pathways (58). However, our findings show that DETC support keratinocyte DNA repair and suggest that lack of induction of Gadd45 and/or Tweak in Tcrδ^−/− skin may contribute to the decreased DNA repair response following acute UV exposure (summarized in Suppl. Fig. 4). However, we cannot exclude the possibility that DETC-derived cytokines may modulate additional repair mechanisms, such as nucleotide excision repair enzymes, to enhance DNA repair, similar to previous observations made for IL-12 and other cytokines (59).

Knowledge of the eATP/skin-resident T cell axis may have major implications for the development of therapeutic targets to improve UVR-induced skin damage, enhance our understanding of the use of phototherapy to ameliorate disease states, such as psoriasis, eczema, or mycosis fungoides, and shed light on ATP-mediated immunity in T cell and epithelial cell biology.

Abbreviations used in this article

UVR

ultraviolet radiation

DETC

dendritic epidermalγδ T cells

CPD

cyclobutane pyrimidine dimer

SCC

squamous cell carcinoma(s)

eATP

extracellular ATP

Nlrp3

NOD-like receptor family, pyrin domain containing 3

TWEAK

TNF related weak inducer of apoptosis

GADD45

Growth arrest and DNA damage associated gene 45

γH2AX

phosphorylated from of histone 2A