Dermatitis Lesions T Cells Initiate Atopic + Skin-Infiltrating

Home , House dust mite

Skin-Infiltrating CD8+ T Cells Initiate Atopic Dermatitis Lesions Ana Hennino, Marc Vocanson, Yann Toussaint, Karen Rodet, Josette Benetière, Anne-Marie Schmitt, This information is current as Marie-Françoise Aries, Frédéric Bérard, Aurore Rozières of September 28, 2021. and Jean-François Nicolas J Immunol 2007; 178:5571-5577; ; doi: 10.4049/jimmunol.178.9.5571

http://www.jimmunol.org/content/178/9/5571 Downloaded from

References This article cites 30 articles, 7 of which you can access for free at: http://www.jimmunol.org/content/178/9/5571.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 28, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Skin-Inﬁltrating CD8؉ T Cells Initiate Atopic Dermatitis Lesions1

Ana Hennino,*† Marc Vocanson,*† Yann Toussaint,*† Karen Rodet,*‡ Josette Benetie`re,*† Anne-Marie Schmitt,§ Marie-Franc¸oise Aries,§ Fre´de´ric Be´rard,*†‡ Aurore Rozie`res,*†‡ and Jean-Franc¸ois Nicolas2*†‡

Skin lesions in the allergic form of atopic dermatitis (AD) are induced by allergen-specific T cells that infiltrate the skin at the site of allergen exposure. Although Th2-type CD4؉ T cells appear to be crucial in AD pathophysiology, little is known about the contribution of CD8؉ T cells in the development of the allergic skin inflammation. In the present study, we have analyzed the respective role of CD8؉ and CD4؉ T cells in the development of AD skin lesions in a mouse model of allergen-induced AD. In sensitized mice, CD8؉ T cells are rapidly and transiently recruited to the allergen-exposed site and initiate the inflammatory process leading to skin infiltration with eosinophils and Th1/Th2-producing cells. CD8؉ T cell-depleted mice show no inflamma- Downloaded from tion, demonstrating that these cells are mandatory for the development of AD. In contrast, CD4؉ T cell-depleted mice develop a severe form of eczema. Furthermore, adoptive transfer of CD8؉ T cells from sensitized mice into naive recipient mice leads to skin -inflammation soon after allergen exposure. These data indicate that allergen-primed CD8؉ T cells are required for the develop ment of AD-like lesions in mice. The Journal of Immunology, 2007, 178: 5571–5577.

topic diseases including atopic dermatitis (AD),3 Der p supporting the concept that immune responses in AD skin http://www.jimmunol.org/ asthma, and allergic rhinitis have a prevalence of up to are mediated by allergen-specific T cells and can be elicited by A 30% in developed countries and a high overall morbidity environmental aeroallergens (3, 5). However, the nature of T cells (1, 2). AD is a T cell-mediated chronic inflammatory skin disease responsible for the skin inflammation and the mechanisms by associated with cutaneous hyperreactivity to environmental Ags which they induce the AD skin lesions remain unknown. that are innocuous to normal nonatopic individuals. Patients with CD4ϩ T cells are considered pivotal in the development of ec- AD develop specific IgE and T cells to aeroallergens (3). In this zema and skin eosinophilic inflammation by producing a mixed respect, the house dust mites Dermatophagoides farinae (Der f) pattern of Th1 and Th2 cytokines including IL-4 and IL-13, which

and Dermatophagoides pteronyssinus (Der p) are major aeroaller- are known to induce isotype switching to IgE synthesis, as well as by guest on September 28, 2021 gens in AD pathophysiology, as up to 75% of patients with AD IL-5, which plays an important role in eosinophil development and have skin prick test reactivity and/or specific IgE. More impor- survival (6, 7). Such hypothesis relies on indirect observations in tantly, epicutaneous application of house dust mite extract by human AD, mostly using peripheral blood CD4ϩ T cells and T cell atopy patch test on uninvolved skin of patients with AD elicits clones and on histological findings showing that the majority of eczema in 30–50% of these patients (3, 4). T cells isolated from skin-infiltrating T cells in active AD lesions were CD4ϩ T cells AD skin lesions and allergen patch test sites selectively respond to (8). Whereas in acute AD lesions a predominance of Th2-producing cells has been observed, a shift to a Th1 cytokine pattern has been described in chronic lesions (9). Although IFN-␥ has been *Institut National de la Sante´ et de la Recherche Me´dicale, Unite´ 503, L’Institut shown to be produced in both acute and chronic skin lesions in † Fe´de´ratif de Recherche 128 BioSciences Lyon-Gerland, and UniversiteĆlaude Ber- humans (10, 11), high numbers of CD4ϩ and CD8ϩ T cells nard Lyon 1, Lyon, France; ‡Hospices Civils de Lyon, Department of Clinical Im- munology and Allergy, Centre Hospitalier Universitaire de Lyon Sud, Pierre Beńite, expressing IFN-␥ infiltrate acute AD lesions (12). Moreover, a France; and §Institut de Recherche Pierre Fabre, Hoˆtel Dieu St-Jacques, Toulouse, significant proportion of house dust mite-specific T cell clones France isolated from AD lesional skin were of a CD8 phenotype (13). Received for publication September 6, 2006. Accepted for publication February 6, 2007. Increased frequencies of both cutaneous lymphocyte-associated Ag (CLA)-positive CD4ϩ and CD8ϩ T cells producing type 2 The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance cytokines have been detected in the blood of patients with AD with 18 U.S.C. Section 1734 solely to indicate this fact. (14). These circulating CLA-positive CD8ϩ T cells are able to 1 This work was supported by institutional grants from Institut National de la Sante´ induce IgE production by B cells and enhance eosinophils survival et de la Recherche Me´dicale, UniversiteĆlaude Bernard Lyon 1, and Hospices Civils ´ (15). Seneviratne et al. (16) recently reported the presence of Der de Lyon and by a grant from the Centre de Recherche sur la Peau et les Epithe´lium ϩ de Reveˆtement, Pierre Fabre Laboratories (Castres, France). p-specific CD8 T cells able to produce type 1 and type 2 cyto- 2 Address correspondence and reprint requests to Dr. Jean-François Nicolas, Institut kines in the blood of patients with AD, but not in the blood of National de la Santeét de la Recherche Me´dicale Unite´503, L’Institut Fe´de´ratif nonatopic individuals. More importantly these authors observed a de Recherche 128, 21 av Tony Garnier, 69375 Lyon, France. E-mail address: ϩ [email protected] or [email protected] strong correlation between the level of specific CD8 T cells and 3 the severity of the disease. Together these studies point out to a Abbreviations used in this paper: AD, atopic dermatitis; FasL, Fas ligand; CLA, ϩ cutaneous lymphocyte-associated Ag; Der f, Dermatophagoides farinae; Der p, potential role for CD8 T cells in the development of AD lesions. Dermatophagoides pteronyssinus; LN, lymph node; SFC, spot-forming cell; To examine the respective contribution of CD4ϩ and CD8ϩ T TARC, thymus- and activation-regulated chemokine. cells in AD pathophysiology, we investigated the allergic response ϩ Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 of wild-type C57BL/6 mice as compared with that of CD4 and www.jimmunol.org 5572 CD8ϩ T CELLS IN AD

ϩ CD8 T cell-deficient mice, using a mouse model of allergen (Der proliferation assay. Irradiated splenocytes from naive C57BL/6 mice were f)-induced skin inflammation reproducing most of the features of used as APCs. For cytokine production (IFN-␥ and IL-4) the cells were seeded ϫ 5 ␥ human AD (17). We show that effector CD8ϩ T cells mediate at a concentration of 3 10 /well. IFN- and IL-4 production was titrated ␥ ϩ by ELISA using the R&D Systems DuoSet kit following the manufacturer inflammation and AD-like lesions. IFN- -producing CD8 T cells protocol. For ELISPOT assay (IFN-␥ and IL-5), 96-well nitrocellulose infiltrate the challenged skin a few hours after Der f exposure of plates (MAIP N4510; Millipore) were coated overnight at 4°C with anti- sensitized mice and initiate a mixed Th1/Th2-type skin inflamma- IFN-␥ Ab or anti-IL-5 Ab (R46A2; BD Pharmingen) and blocked with tion leading to the recruitment of mononuclear cells and eosino- RPMI 1640 for1hat37°C. The plates were washed five times with 0.05% PBS/Tween 20 before use. Auricular LNs of allergen- or vehicle-treated phils and to the development of epidermal changes typical of AD. mice were harvested at day 3, and cell suspensions were prepared. LN cells were dispensed in the plates (105/well) and incubated with 1 mg/ml Der f

Materials and Methods proteins or Der f peptide for 36 h at 37°C, 5% CO2 in the presence of Animals irradiated splenocytes as APCs. Der f class I peptide was also used for restimulation at a final concentration of 2 ␮M. The peptide 113–122 sequence Female C57BL/6 mice were purchased from Charles River Breeding Lab- (YGISNYCQI) was determined by prediction using the BioInformatics and ␤ oratories. C57BL/6 mice with a mutation in the A gene (MHC class II °/°) Molecular Analysis Section and SYFPEITHI epitope prediction databases. Af- and in the CD3␧ cluster (CD3␧°/°) were provided by C. Benoist and D. ter washing, plates were incubated2hat37°C, 5% CO2 with a biotinylated Mathis (Harvard Medical School, Boston, MA) and M. Malissen (Centre anti-IFN-␥ or anti-IL-5 mAb (BD Pharmingen). Spot-forming cells (SFCs) d’Immunologie, Marseille-Lumigny, Marseille, France), respectively. All were developed using avidin HRP (CliniSciences) incubated for 20 min at mice were used between 2 and 3 mo of age. Groups consisted of 10 mice room temperature, and extensively washed before adding substrate (amino- each. All experimental procedures were in accordance with the Comite´ ethyl carbazole; Sigma-Aldrich). re´gional d’e´thique pour l’expe´rimentation animale guidelines on animal For the cellular analysis of the AD-like lesions, ears were split into welfare. dorsal and ventral halves and incubated for 60 min at 37°C with 1000 U/ml Downloaded from Mouse model of AD collagenase IA (Sigma-Aldrich), 1000 U/ml hyaluronidase (Sigma- Aldrich), and 200 U/ml DNase I (Roche Diagnostics). Cells were incubated Ear mouse eczema was induced by repeated topical application on the left with the 2.4G2 hybridoma supernatant to block Fc␥R and then stained with ear of a protein allergen solution (250 ␮g of protein per ear) of Der f house various combinations of the following anti-mouse Abs: FITC anti-CD45 dust mite extract provided by P. Moingeon (Stallergenes, Antony, France) (leukocyte common Ag Ly5, clone 30-F11; BD Pharmingen), PE anti- once a week for 6–10 wk. Control mice were painted onto the left ear with CD8, and FITC anti-CD4. Immunofluorescence levels were analyzed on a the vehicle (70% DMSO in water) alone. Evaluation of ear thickness was FACSCalibur flow cytometer with CellQuest software (BD Biosciences). http://www.jimmunol.org/ done for each mouse ear before allergen application and each day after allergen application with a spring-loaded micrometer (J15; Blet). Ear swelling was calculated by subtracting the initial value from the value Histology and immunohistochemistry recorded on the corresponding day (for both allergen- and vehicle- Ears were fixed in a 3% formalin solution for 24 h and processed through treated mice). a routine 15-h cycle to paraffin wax embedding. The 4-␮m sections were Ab treatment cut using a microtome and mounted on Superfrost Plus slides. Sections were dried overnight at 37°C. The slides were dewaxed in Ottix baths and The rat anti-mouse CD4 mAb GK 1.5 was obtained from American Type stained with H&E. For immunohistochemistry, ears were deep frozen in Culture Collection (ATCC) and the rat anti-mouse CD8 mAb H35.17.2 isopentane precooled in liquid nitrogen, embedded in OCT compound (Tissue- was provided by G. Milon (Institut Pasteur, Paris, France). A ␤-galactosi- Tek). Cryostat sections (5 ␮m) were mounted on Superfrost Plus slides by guest on September 28, 2021 dase-specific rat IgG (clone GL113; ATCC) was used in some experiments and dried overnight at room temperature, before being fixed in ice-cold as isotype control for the anti-CD4 and anti-CD8 mAbs. Hybridomas were acetone for 10 min. Slides were rehydrated in TBS/Tween 20, and incu- grown as ascites in pristane-primed Swiss nude mice. Mice were given i.p. bated with primary Abs, anti-CD4 (L3T4 clone; BD Pharmingen) and anti- injections of 20 ␮g of anti-CD4 mAbs or 16 ␮g of anti-CD8 mAbs in 200 CD8␣ (Lyt-2, clone KT15; Immunotech) for1hatroom temperature in a ␮l of saline on days Ϫ1 and 0 of ear painting. Cell depletion was assessed humid chamber. Specific binding was revealed after 30 min incubation in each mouse by staining for CD4 and CD8 molecules on whole draining with biotinylated goat anti-rat IgG Ab (Vector Laboratories) before strepta- lymph nodes (LNs) as well as spleens, using FITC- or PE-labeled anti-CD4 vidin biotin complex and HRP incubation (DakoCytomation). or anti-CD8 mAbs (Tebu-Bio).

RNA extraction and RT-PCR analysis Adoptive transfer At different intervals (24 and 48 h) after allergen exposure, ear samples Axillary and inguinal LNs from allergen- or vehicle-treated mice were were collected from vehicle- and allergen-treated mice and frozen in liquid harvested at day 28 after sensitization and used to prepare single LN cell nitrogen. The detection of RNA was conducted as described in detail else- suspensions and purified CD8ϩ and CD4ϩ T cells. A total of 12 ϫ 106 where (18). In brief, total RNA was extracted using the RNAXEL kit cells/200 ␮l of saline were i.v. transferred through the orbital vein into (Eurobio Laboratories). After DNase I treatment, 1 mg of total RNA was naive syngenic CD3␧ °/° mice. Recipient mice were challenged 1 day after reverse transcribed using poly(dT)15 primers and Subscript II reverse tran- cell transfer and ear swelling was measured the following days. scriptase (90 min, 37°C; Invitrogen Life Technologies). The amount of RNA to be used for detection was normalized using the housekeeping gene HPRT as reference. The cDNA obtained was amplified using different sets of primers. Results For HPRT, CD4, CD8, IFN-␥ the primers have been described else- Epicutaneous exposure to house dust mite allergen induces where (19). For IL-4, the sense primer is 5Ј-TCGGCATTTTGAACGAG AD-like skin inflammation in normal mice GTC-3Ј and antisense 5Ј-GAAAAGCCCGAAAGAGTCTC-3Ј. For thymus- and activation-regulated chemokine (TARC) the sense primer is 5Ј- C57BL/6 mice were sensitized once a week for six consecutive CAG GAA GTT GGT GAG CTG GTA ATA-3Ј and the antisense primer weeks by skin application of a solution of Der f on the left ear and is 5Ј-TTG TGT TCG CCT GTA GTG CAT A-3Ј. of vehicle alone on the right ear, as previously described (17). The amplifications were conducted with 29 cycles for HPRT and 33 cycles for the other genes (1 min at 94°C, 1 min 30 s at 60°C, 2 min at Control mice were sensitized with the vehicle alone. Twenty-four 72°C). The PCR products were analyzed on 1.5% agarose gel. hours after the fourth allergen exposure, i.e., at day 22, Der f- treated mice developed an ear skin inflammation as manifested by Flow cytometry, T cell isolation, and cell culture increased ear thickness (Fig. 1A). Ear swelling was detectable at Draining LNs from vehicle- or allergen-treated mice were harvested on day 22 and increased after each subsequent allergen exposure lead- ϩ ϩ days 21 or 28 to prepare single-cell suspensions. CD8 or CD4 T cells ing to a very strong skin inflammation at day 36, i.e., 24 h after the were purified using anti-CD8 or anti-CD4 mAb-coated microbeads and positive selection columns (Miltenyi Biotec). Purity was routinely Ͼ95%. sixth allergen exposure. Histopathological analysis of the ear skin The cell cultures were performed in 96-well plates. The whole draining of Der f-treated mice at day 36 revealed skin inflammatory LNs as well as isolated CD8ϩ or CD4ϩ T cells were seeded at 105/well for changes characteristic of AD-like lesions (Fig. 1B), including The Journal of Immunology 5573

homing of T lymphocytes in AD lesions in both humans and mice (20, 21). PCR analysis demonstrated that mRNA encoding for these molecules were either not found (CD8, IFN-␥, IL-4, and CCL17) or expressed at very low levels (CD4) in the skin of vehicle-treated mice. In contrast, mRNA for CD8, IFN-␥, IL-4, and CCL17 were induced in the skin of Der f-treated mice, and in CD4 transcripts only moderately increased compared with vehicle- treated skin (Fig. 1C). We next analyzed the allergen-specific immune response gen- erated at day 35 in auricular LNs, draining the allergen-challenged ear. As shown in Fig. 1D, LN cells from Der f-treated mice vig- orously proliferated in response to in vitro restimulation with allergen-pulsed syngeneic cells. Purified CD8ϩ T cells gave higher proliferation rates than CD4ϩ T cells, suggesting that both allergen-specific CD8ϩ and CD4ϩ T cells had been primed. Moreover, the Der f-specific immune response in total draining LN cells was associated with the production of high levels of IFN-␥ and IL-5, whereas no IL-4 could be detected (Fig. 1E). Finally, Der f-treated

mice exhibited high serum IgE levels compared with vehicle- Downloaded from treated mice (Fig. 1F). Although not shown in this experiment, similar histological, molecular, and immunological data were obtained in the skin and LNs of mice at days 22 and 29, i.e., after the fourth and ﬁfth allergen exposure, respectively. Thus, Der f skin sensitization in-

duced an AD-like disease associated with a mixed Th1/Th2-type http://www.jimmunol.org/ dermatitis, specific CD4ϩ and CD8ϩ T cells and high serum IgE levels. FIGURE 1. Repeated Der f skin exposure induces AD-like skin inflammation in C57BL/6 mice. A, Allergen-induced skin inflammation was an- ϩ alyzed in C57BL/6 mice painted on the left ear with a Der f solution (f) CD8 T cells mediate AD-like skin lesions or the vehicle (Ⅺ) once a week for seven consecutive weeks. Results are These results showed that development of AD-like lesions in expressed as the mean ear swelling (in micrometers) at 24 h after each C57BL/6 mice was associated with prominent up-regulation of sensitization. Results are representative of three independent experiments, ϩ using five mice per group. B, Histological analysis of ear sections at 24 h CD8 mRNA suggestive of an infiltration of the skin by CD8 T

ϩ ϩ by guest on September 28, 2021 after the sixth sensitization, i.e., at day 36, showing normal histology in cells. To study the respective role of CD4 and CD8 T cells in vehicle-treated mice (panel 1) and major skin inflammation in Der f-treated our model, we next analyzed the development of AD-like inflam- ϩ mice, including ear swelling (panel 2), epidermal thickening and spongi- mation in C57BL/6 mice and in mice acutely depleted in CD8 or osis (panel 4, asterisk), presence of epidermal apoptotic cells (panel 5, CD4ϩ T cells by in vivo treatment with mAbs specific for CD8 and arrow), and infiltration of the dermis by eosinophils (panel 6, arrowhead). CD4, respectively. Ab treatment was achieved twice a week, be- Eosinophils were counted in 15 high-power fields and results are expressed fore and on the same day as Der f painting, and yielded to a Ͼ95% as cell number per high-power field, using HES staining. Original magni- depletion of the respective T cell subset (data not shown). As fication ϫ50 (panels 1 and 2), ϫ400 (panel 4), and ϫ1000 (panels 5 and shown in Fig. 2A, normal undepleted mice, as well as mice treated 6). C, RT-PCR analysis of CD8, IFN-␥, CD4, IL-4, and CCL17 transcripts in situ in whole ear mRNA extract from vehicle-treated (veh) or Der f- by a rat IgG isotype control (data not shown), developed a skin inflammation 24 h after the fourth topical allergen exposure (day treated (Der f) mice at 24 h after the sixth sensitization (day 36). D, Der ϩ f-specific T cell proliferation in LNs of day-35 vehicle-treated (u)orDer 22). CD4 T cell-depleted mice developed a quicker and stronger f-treated (f) mice. Total T cells, purified CD4ϩ or CD8ϩ T cells from the response than undepleted mice because skin inflammation was al- auricular LNs were restimulated in vitro for 3 days with Der f. Thymidine ready significant after the second allergen painting (day 8) and incorporation was conducted during the last 16 h of culture. Results are increased to a severe dermatitis after the fourth Der f application Ϫ expressed as total cpm (ϫ 10 3). E, Production of IFN-␥, IL-4, and IL-5 (day 22). This result suggests that CD4ϩ T cells do not behave as by LN cells. Total LN cells were recovered from the auricular LNs of Der effector cells but rather contain cells endowed with down-regula- f u f-sensitized ( ) and vehicle-sensitized ( ) mice at day 35 and restimulated tory properties. More importantly, no skin inflammation was in- in vitro for 3 days with Der f. IFN-␥ and IL-4 produced in day-3 culture ϩ ϩ duced in CD8 T cell-deficient mice, demonstrating that CD8 supernatant was tested by ELISA. IL-5-producing cells were enumerated by ELISPOT assay. F, ELISA titration of serum IgE (in nanograms per T cells are necessary for the development of AD-like skin .p Ͻ 0.01. inflammation ,ءء .milliliter) in either vehicle- or Der f-treated mice at day 36 Analysis of the allergen-specific immune response in draining LNs at day 22 was done by IFN-␥ ELISPOT assay, using in vitro thickening of the epidermis (Fig. 1B, panels 2 and 4) with apo- restimulation of T cells by a H-2b-restricted, class I Der f peptide. ptosis of keratinocytes (Fig. 1B, panel 5), blood vessel enlarge- Fig. 2B shows that sensitized mice displayed a mean of 80 SFCs/ ment, infiltration of mononuclear cells, and more importantly, 106 LN cells. Der f-specific T cells were found in the CD8ϩ T prominent recruitment of eosinophils (Fig. 1B, panels 3 and 6). cell subset because depletion of CD8ϩ T cells, but not of CD4ϩ Because human AD lesions are infiltrated by Th1/Th2 cells, we T cells, abrogated the IFN-␥ response. That CD8ϩ T cells were examined the presence of CD4ϩ and CD8ϩ T cells and the ex- the major IFN-␥-producing cells in our model was confirmed by pression of IL-4, IL-5, and IFN-␥ in the skin of allergen-induced IFN-␥ ELISPOT experiments using purified CD8ϩ and CD4ϩ T AD-like lesions in normal C57BL/6 at day 36. We also tested the cells recovered from draining LNs of day-22 sensitized mice expression of CCL17 (TARC), a chemokine associated with the (Fig. 2C). 5574 CD8ϩ T CELLS IN AD

FIGURE 2. AD-like skin inflammation is mediated by CD8ϩ T cells. A, Effect of in vivo depletion of CD4ϩ and CD8ϩ T cells on allergen-induced skin inflammation. Anti-CD4 mAb-treated (E), anti-CD8 mAb-treated (f)or untreated (Ⅺ) mice were ear sensitized with Der f once a week for 4 consecutive weeks. Results are expressed as the mean ear swelling (in micrometers) the days following each sensitization. Results are representative of three independent experiments. B, Der f-specific IFN-␥-producing cells were determined by ELISPOT assay in LN cells recovered on day 21 from Der f- or vehicle-sensitized C57BL/6 mice, untreated or treated with anti-CD4- or anti- CD8-depleting mAbs, as shown under the chart. Results are expressed as the number of SFCs per 106 LN cells. C, Der f-specific IFN-␥-producing cells were determined by ELISPOT assay using purified CD8ϩ and CD4ϩ LN T cells from day 21 Der f-sensitized C57BL/6 mice. Results are expressed as the number of SFCs per 106 cells. D, FACS analysis of skin cells at day 22, i.e., 24 h after the fourth sensitization. Cells extracted from the ears of either Downloaded from vehicle-treated unsensitized mice (panel 1), wild-type Der f-sensitized mice (panel 2), CD4ϩ T cell-depleted Der f- sensitized mice (panel 3)orCD8ϩ T cell-depleted Der f-sensitized mice (panel 4) were stained using anti-CD45, anti-CD4, and anti-CD8 mAbs. Dot plot analysis of double staining for CD4ϩ and CD8ϩ cells gated on CD45ϩ cells

is shown. Data are representative of three independent ex- http://www.jimmunol.org/ periments. E, FACS analysis of the recruitment of CD8ϩ and CD4ϩ T cells in the ear skin of Der f-sensitized undepleted mice at days 22 and 23, i.e., 24 and 48 h after the fourth sensitization. Results are expressed as the mean Ϯ SD of the percentage of CD8ϩ and CD4ϩ T cells among total CD45ϩ cells from the ears of vehicle-treated (Ⅺ)or Der f-treated mice (f) obtained in three mice per group. F, Immunohistochemical staining of CD8ϩ T cells in cryostat sections of ears of CD4ϩ T cell-depleted mice at days 22 and 23. Cartilage (C) and dermal-epidermal junction (dot- by guest on September 28, 2021 ted line) are noted. Inset is magniﬁed view (ϫ400) of epidermal CD8ϩ T cells. Original magniﬁcation is ϫ200. Data are representative of three independent experiments.

Analysis of CD8ϩ and CD4ϩ T cell recruitment in allergen- cells, which occurs however later than the CD8ϩ T cell recruit- treated ear skin was done by FACS analysis of skin cell suspen- ment, and is observed in both Der-f and vehicle-challenged mice. sions at day 22 (i.e., 24 h after the fourth Der f application) (Fig. Indeed, CD4ϩ T cells represent up to 6% of CD45ϩ skin cells at 2, D and E). In undepleted mice, treated with the vehicle alone, 48 h but only 2% at 24 h postchallenge (Fig. 2E). Immunohisto- most of the T cells infiltrating the skin were CD4ϩ T cells (5% of chemical analysis of CD8ϩ T cells in the skin of CD4ϩ T cell- total CD45ϩ cells), whereas CD8ϩ T cells represented only depleted mice confirmed FACS data showing that Der f exposure 0.5–1% of CD45ϩ cells (Fig. 2D, upper left panel). In undepleted, at day 21 resulted in a marked infiltration of the skin by CD8ϩ T Der f-treated mice, the proportion of CD4ϩ T cells in the skin was cells, which peaked at 24 h (day 22) (Fig. 2F). At that time, CD8ϩ similar to that found in vehicle-treated mice. In contrast, CD8ϩ T T cells were observed in the dermis and more importantly in the cells represented up to 5% of infiltrating cells, indicating that al- epidermis, whereas at 48 h only dermal CD8ϩ T cells could be found. lergen exposure was responsible for the preferential recruitment of Taken together these data indicate that the magnitude of the Der CD8ϩ T cells (Fig. 2D, upper right panel). f-induced eczema is positively correlated to the number of CD8ϩ In CD4ϩ T cell-depleted and Der f-treated mice, which devel- T cells infiltrating the skin. ϩ oped severe skin inflammation, CD8 T cells accounted for up to ϩ Ͻ ϩ Mice deficient in CD4 T cells develop severe AD-like skin 10–20% of infiltrating cells, whereas 0.5% of CD4 T cells ϩ were found (Fig. 2D, lower left panel). Finally, in CD8ϩ T cell- lesions and enhanced priming of allergen-specific CD8 T cells depleted and Der f-treated mice, in which no skin inflammation To confirm that allergen-induced CD8ϩ T cells can induce AD- could be obtained, no CD8ϩ T cell infiltration was observed and like inflammation in the absence of CD4ϩ T cells, we tested CD4ϩ T cells represented 5% of skin CD45ϩ cells (Fig. 2D, lower whether AD-like disease could be induced by repeated Der f sen- right panel). Fig. 2D shows that infiltration of CD8ϩ T cells in the sitization in A␤ gene °/° (MHC class II °/°) mice, genetically de- challenged skin is a transient phenomenon because an increased ficient in MHC class II-restricted CD4ϩ T cells. Similarly to anti- number of CD8ϩ T cells are observed at 24 h but not at 48 h after CD4 mAb-treated B6 mice (Fig. 2A), MHC class II °/° mice Der f painting. At that time the percentage of CD8ϩ T cells rep- developed a quicker and more intense skin inflammation as com- resent only 1.5% of CD45ϩ skin cells. Of note, development of the pared with normal C57BL/6 mice (Fig. 3A). Skin inflammation skin inflammation is also associated with infiltration of CD4ϩ T was already significantly up-regulated after the second allergen The Journal of Immunology 5575

FIGURE 3. MHC class II-deficient (II °/°) mice develop severe AD-like inflammation. A, Allergen- induced skin inflammation was analyzed in wild- type C57BL/6 (Ⅺ) and MHC class II-deficient mice (f) sensitized with a Der f solution on the left ear, once a week for five consecutive weeks. MHC class II-deficient mice treated with the vehicle alone (E) were included as controls. Results are expressed as the mean ear swelling (in micrometers) at 24 h after each sensitization, using five mice per group Ϯ SD. Results are representative of three independent experiments. B, Histological analysis of the ear skin of Der f-sensitized C57BL/6 mice and MHC class II- deficient mice (II °/°) at day 22 (i.e., 24 h after the fourth sensitization). HES staining at a magnifica- tion of ϫ200 is shown. C, RT-PCR analysis of CD8, IFN-␥, IL-4, and CCL17 transcripts in ears of Der f- treated C57BL/6 (BL/6) and MHC class II-deficient mice (II °/°) at days 22 and 23. Control includes Downloaded from mRNA expression in the ears of vehicle-treated (veh) MHC class II-deficient mice at day 22. D, Der f-specific proliferative responses in total auricular LN cells from Der f-sensitized (f) and vehicle-sensitized (u) C57BL/6 and MHC class II-deficient mice at day 21 after in vitro restimulation with Der f. Thymidine was http://www.jimmunol.org/ added for the last 16 h of culture. Results are expressed p Ͻ 0.01. E, Der ,ءء .(as total cpm supress (ϫ10Ϫ3 f-specific IFN-␥-producing cells were determined by ELISPOT assay in LN cells from day 21 Der f-sensitized (f) or vehicle-sensitized (u) C57BL/6 and MHC class II-deficient (II °/°) mice, after overnight restimulation with Der f. Results are expressed as the number of SFCs per 106 LN cells. Results in D and E are presented as the mean Ϯ SD in culture triplicate. F, Titration of serum IgE (nanograms per by guest on September 28, 2021 milliliter) in either Der f-sensitized (f) or vehicle- sensitized (u) C57BL/6 and MHC class II-deficient mice at day 36 by ELISA. Results are presented as the mean (three animals Ϯ SD).

exposure and further augmented to a severe dermatitis after the Adoptive transfer of CD8ϩ but not CD4ϩ T cells from Der fourth Der f application (day 21). Histological analysis of the ears f-sensitized mice can induce AD-like skin lesions of MHC class II °/° mice at day 22 showed enhanced thickening of To confirm the dominant role of CD8ϩ T cells in the pathophys- the epidermis and increased skin infiltration by mononuclear cells, iology of Der f-induced eczema, CD8ϩ T cells (or CD4ϩ T cells) compared with C57BL/6 mice (Fig. 3B). Increased skin inflam- purified from the LNs of vehicle- or Der f-treated mice at day 21 mation in MHC class II °/° mice correlated with a rapid up-regu- were i.v. transferred into naive T cell-deficient CD3␧°/° recipients, lation of CD8, IFN-␥, and IL-4 transcripts in the allergen-exposed ear as well as appearance of CCL17 mRNA, at 24 h after Der f exposure (Fig. 3C). Analysis of the Der f-specific immune responses at day 21 in draining auricular LNs revealed that MHC class II °/° mice exhibited increased Der f-specific T cells as assessed by proliferative responses (Fig. 3D) and frequency of IFN-␥-producing cells (Fig. 3E). However, the number of IL-5-producing LN cells was similar in MHC class II °/° mice to that observed in wild-type C57BL/6 mice and ranged from 30 to 80 IL-5 SFCs/106 LN cells (data not shown). IgE was undetectable in the serum of Der f-sensitized MHC class II °/° mice, a finding compatible with the need for ϩ ϩ FIGURE 4. CD8 T cells are effector cells of AD-like skin inflamma- CD4 T cell help for the differentiation of B cells in IgE-production. CD8ϩ T cells and CD4ϩ T cells were purified from the LNs of Der ing plasma cells (Fig. 3F). These data show that mice deficient in f-sensitized (f) or vehicle-sensitized (u) C57BL/6 at day 21 and i.v. trans- ϩ CD4 T cells develop a severe AD-like skin inflammation, which ferred (12 ϫ 106)toCD3␧°/° recipient mice, which were ear challenged ϩ is mediated by CD8 T cells primed in draining LNs and recruited with Der f 24 h later. Results are expressed as ear swelling at day 6 in the skin early after allergen exposure. postchallenge. 5576 CD8ϩ T CELLS IN AD that were challenged with Der f 1 day after transfer. As shown in in wild-type and MHC class II °/° mice) and are no longer detect- the Fig. 4, mice reconstituted with CD8ϩ T cells from Der f-treated able at 48 h. Conversely, CD4ϩ T cells are recruited only at 48 h mice developed a severe skin inflammation starting 24 h after Der of challenge, i.e., subsequent to skin infiltration by CD8ϩ T cells. f challenge and peaking at day 6 postchallenge. In contrast, no skin These observations suggest that, once recruited into skin, allergen- inflammation was observed upon Der f challenge in CD3␧°/° re- primed CD8ϩ T cells require a narrow window of time for exerting ϩ cipients adoptively transferred with purified CD4 T cells. These their effector function that initiates the subsequent skin inflamma- ϩ data provide evidence that CD8 T cells are the pathogenic effec- tory process. Persistence of these CD8ϩ effector T cells in skin is tor T cells responsible for the onset of Der f-induced mouse AD. not necessary for the skin inflammation to proceed probably because other cells involved in AD pathophysiology, especially eo- Discussion sinophils (24) have been recruited in the meantime. Our data are ϩ In this study, we demonstrate that CD8 T cells represent the reminiscent to previous studies in human AD showing that, al- dominant effector cell responsible for allergen-induced skin in- though CD4ϩ T cells predominated in AD skin lesions, T cells flammation. The AD-like dermatitis that developed after repeated infiltrating the epidermis were almost exclusively CD8ϩ T cells exposure to the aeroallergen Der f displays all the major histolog- (12). Along these lines, studies in mouse models of AD reported ical, molecular, and immunological features of human allergic AD. that allergen challenge is associated with epidermal recruitment of Histological changes include dermal thickening, vasodilatation, in- CD8ϩ T cells (25), which is dramatically enhanced by coexposure filtration by eosinophils and mononuclear cells including CD4ϩ ϩ to the superantigen staphylococcal enterotoxin B (26). However, and CD8 T cells, and epidermal thickening with presence of sev- only a few CD8ϩ T cells, compared with CD4ϩ T cells, were eral apoptotic cells. Molecular changes in the skin consist of up- detected in the skin in these studies probably because of the time Downloaded from ␥ regulation of transcripts for Th1-type (IFN- ) and Th2-type (IL-4) at which T cell infiltration was analyzed. It may be proposed that this cytokines and for the chemokine CCL17, involved in the recruit- peculiar transient kinetics of CD8ϩ T cell recruitment and activation ment of type 2-producing cells (22). In lymphoid organs, Der f- ϩ ϩ in the allergen-exposed skin may explain why previous studies have primed T cells are present within both the CD4 and CD8 T cell underestimated the effector role of CD8ϩ T cells in AD. ␥ subsets and produce high levels of IFN- and IL-5. Finally, Der The peak of IFN-␥ production in the challenged skin is maximal f-primed mice have high levels of circulating IgE. Therefore, the at 24 h and decreases rapidly from 48 h. This is in contrast with http://www.jimmunol.org/ Der f-induced AD mouse model is suitable for the study of AD human studies on atopy patch test that showed an early and tran- pathophysiology. sient up-regulation of IL-4 at 24 h followed by a switch toward Our data clearly demonstrate that CD8ϩ T cells represent the IFN-␥ production, which appeared at 48 h only (27). Although this effector cells that mediate skin inflammation. Our finding is sup- discrepancy may reflect subtle differences between mouse and hu- ported by several lines of evidence. First, repeated in vivo skin man disease models, it could be hypothesized that the previous sensitization with the allergen Der f is able to prime a vigorous ϩ ϩ studies have missed the first wave of allergen-specific T cells in- CD8 T cell response in skin draining LNs and to recruit CD8 T ϩ ϩ filtrating the challenged skin and transiently producing IFN-␥. On- cells in skin. Second, in vivo Ab depletion of CD8 , but not CD4

going studies testing for the presence of type 1 and type 2 cyto- by guest on September 28, 2021 T cells, in wild-type mice completely abrogates the skin inflam- ϩ kines in atopy patch tests after 6 and 12 h postchallenge may help mation, indicating that CD8 T cells are required for the skin ϩ to better understand the early events involved in the induction of inflammatory process. Third, CD8 T cell-deficient MHC class II ϩ the AD skin inflammation. °/° mice develop severe AD lesions. Finally, primed CD8 T cells The fate of skin-infiltrating CD8ϩ T cells is not known precisely from LNs of allergen-sensitized mice can adoptively transfer AD ϩ lesions upon allergen challenge in naive recipients. yet. It is possible that the recruited CD8 T cells rapidly leave the ϩ skin through the lymphatic vessels shortly after having delivered The observation that CD4 T cell-depleted mice develop an ϩ early and exacerbated inflammation indicates that CD8ϩ T cells their effector function. Alternatively, CD8 T cells may die by can be primed and develop into effector cells without the need for activation-induced cell death after allergen-induced T cell activa- CD4ϩ T cell help. Moreover, several observations indicate that tion. In this respect, Akdis et al. (8) demonstrated that apoptosis of CD4ϩ T cells may comprise a subset of regulatory cells controlling circulating memory/effector T cells is a constant feature of AD the skin inflammatory process by regulating the intensity of the skin inflammation that involves Fas-Fas ligand (FasL) interaction. CD8ϩ T cell response. Indeed, mice with a CD4ϩ T cell defect The fact that activation-induced cell death is mostly confined to (including both anti-CD4 mAb-treated C57BL/6 and MHC class II Th1 cells, whereas Th2 cells are resistant to apoptosis, could ex- °/° mice) exhibit a more rapid appearance of AD lesions. This plain why a Th2-skewed immune response is consistently de- response is associated with a higher CD8ϩ T cell response in scribed in patients with AD (8). Finally it is tempting to speculate ϩ draining LNs and a higher number of CD8ϩ T cells recruited into that CD4 T cells endowed with down-regulatory functions are ϩ the skin. Although the subset of CD4ϩ T cells endowed with reg- responsible for the rapid clearance of CD8 T cells, inasmuch as ϩ ulatory properties has not been tested, our findings are reminiscent recruitment of CD4 T cell observed at 48 h, but not at 24 h, ϩ to previous studies in AD patients, indicating that IL-10-producing paralleled the disappearance of CD8 T cells from the allergen- ϩ CD4ϩ regulatory T cells control the magnitude of AD skin in- exposed skin. The mechanisms by which CD4 T cells may con- ϩ flammation (23). tribute to the clearance of Der f-primed CD8 T cells remain, CD8ϩ T cell recruitment and activation in AD-like lesions is a however, unknown. ϩ rapid but transient phenomenon occurring in the hours following How CD8 T cells induce the AD skin pathology? Epidermal allergen exposure. This outcome was demonstrated by kinetic abnormalities including epidermal spongiosis and vesicle forma- studies of the phenotype of skin-infiltrating T cells using flow cy- tion are pathological hallmarks of AD (22). Apoptosis of keratin- tometry, immunohistochemistry, and RT-PCR analysis. Because ocytes caused by skin-infiltrating, FasL-expressing T cells plays a CD8ϩ T cells are the major IFN-␥ producers in our model, the key role in spongiosis formation. Studies of Akdis et al. (8) in up-regulation of IFN-␥ mRNA, which paralleled CD8ϩ T cell in- human AD indicate that T cell-derived IFN-␥ is responsible for filtration, is thought to reflect CD8ϩ T cell activation. CD8ϩ T up-regulation of Fas by keratinocytes, which become susceptible cells are found in the epidermis at 24 h postallergen exposure (both to apoptosis through Fas-FasL interaction. Moreover, perforin and The Journal of Immunology 5577 granzyme B are cytotoxic molecules produced by activated cyto- diseases is due to preferential apoptosis of circulating memory/effector Th1 cells. toxic CD8ϩ T cells, which are strongly expressed in AD lesional FASEB J. 17: 1026–1035. 9. Hamid, Q., M. Boguniewicz, and D. Y. Leung. 1994. Differential in situ cytokine skin (28). Our data showing a correlation between the infiltration gene expression in acute versus chronic atopic dermatitis. J. Clin. Invest. 94: of IFN-␥-producing CD8ϩ T cells in the epidermis, the appearance 870–876. of apoptotic cells, and the development of the skin inflammation 10. Leung, D. Y., and T. Bieber. 2003. Atopic dermatitis. Lancet 361: 151–160. 11. Thepen, T., E. G. Langeveld-Wildschut, I. C. Bihari, D. F. van Wichen, are in line with our reported findings and suggest that, similarly to F. C. van Reijsen, G. C. Mudde, and C. A. Bruijnzeel-Koomen. 1996. Biphasic their mode of action in allergic contact dermatitis to haptens (29), response against aeroallergen in atopic dermatitis showing a switch from an ini- ϩ tial TH2 response to a TH1 response in situ: an immunocytochemical study. allergen-specific CD8 T cells could exert their pathogenic func- J. Allergy Clin. Immunol. 97: 828–837. tion via cytotoxicity against allergen-presenting keratinocytes. 12. Simon, D., R. L. Lindberg, E. Kozlowski, L. R. Braathen, and H. U. Simon. 2006. Thus, the following scheme could be proposed to explain the Epidermal caspase-3 cleavage associated with interferon-␥-expressing lympho- ϩ cytes in acute atopic dermatitis lesions. Exp. Dermatol. 15: 441–446. pathogenic role of CD8 T cells in induction of AD-like skin 13. Werfel, T., A. Morita, M. Grewe, H. Renz, U. Wahn, J. Krutmann, and A. Kapp. inflammation in allergen-exposed mice. First, CLA-positive aller- 1996. Allergen specificity of skin-infiltrating T cells is not restricted to a type-2 gen-specific CD8ϩ T cells, primed in skin draining LNs, are re- cytokine pattern in chronic skin lesions of atopic dermatitis. J. Invest. Dermatol. 107: 871–876. cruited into the epidermis through interactions with E-selectins and ϩ 14. Teraki, Y., T. Hotta, and T. Shiohara. 2000. Increased circulating skin-homing P-selectins expressed on endothelial cells. Second, specific CD8 cutaneous lymphocyte-associated antigen (CLA)ϩ type 2 cytokine-producing ϩ T cells are activated by allergen-presenting epidermal cells result- cells, and decreased CLA type 1 cytokine-producing cells in atopic dermatitis. Br. J. Dermatol. 143: 373–378. ing in both cytotoxic activity and activation of skin-resident cells. 15. Akdis, M., H. U. Simon, L. Weigl, O. Kreyden, K. Blaser, and C. A. Akdis. 1999. Thirdly, amplification of the inflammatory reaction occurs through Skin homing (cutaneous lymphocyte-associated antigen-positive) CD8ϩ T cells production of cytokines and chemokines by skin cells, such as respond to superantigen and contribute to eosinophilia and IgE production in Downloaded from atopic dermatitis. J. Immunol. 163: 466–475. CCL17/TARC and CCL22/macrophage-derived chemokine, pro- 16. Seneviratne, S. L., L. Jones, A. S. King, A. Black, S. Powell, A. J. McMichael, ϩ moting the subsequent recruitment of leukocytes, including CD4 and G. S. Ogg. 2002. Allergen-specific CD8ϩ T cells and atopic disease. J. Clin. T cells and eosinophils (22). Further studies will have to define Invest. 110: 1283–1291. ϩ 17. Gao, X. K., N. Nakamura, K. Fuseda, H. Tanaka, N. Inagaki, and H. Nagai. 2004. more precisely the mechanisms responsible for the initial CD8 T Establishment of allergic dermatitis in NC/Nga mice as a model for severe atopic cell recruitment in the skin. dermatitis. Biol. Pharm. Bull. 27: 1376–1381. ϩ 18. Delassus, S., G. C. Coutinho, C. Saucier, S. Darche, and P. Kourilsky. 1994. In summary, we have identified a critical role for CD8 T cells http://www.jimmunol.org/ ϩ Differential cytokine expression in maternal blood and placenta during murine in allergen-induced mouse AD and have shown that CD8 T cells gestation. J. Immunol. 152: 2411–2420. are required for the full development of the allergic skin inflam- 19. Akiba, H., J. Kehren, M. T. Ducluzeau, M. Krasteva, F. Horand, D. Kaiserlian, F. Kaneko, and J. F. Nicolas. 2002. Skin inflammation during contact hypersen- mation. Our data are in line with recent studies reporting on the ϩ ϩ sitivity is mediated by early recruitment of CD8 T cytotoxic 1 cells inducing effector role of CD8 T cells in allergic asthma (30, 31) and keratinocyte apoptosis. J. Immunol. 168: 3079–3087. strongly support the idea that the CD4ϩ Th2 dogma in allergic 20. Vestergaard, C., H. Yoneyama, M. Murai, K. Nakamura, K. Tamaki, Y. Terashima, T. Imai, O. Yoshie, T. Irimura, H. Mizutani, and K. Matsushima. diseases should be revisited (7). However, although the results 1999. Overproduction of Th2-specific chemokines in NC/Nga mice exhibiting presented come from a representative mouse model, studies in hu- atopic dermatitis-like lesions. J. Clin. Invest. 104: 1097–1105. mans are required to confirm the hypothesis that CD8ϩ T cells are 21. Jahnz-Rozyk, K., T. Targowski, E. Paluchowska, W. Owczarek, and A. Kucharczyk.

2005. Serum thymus and activation-regulated chemokine, macrophage-derived che- by guest on September 28, 2021 involved in the development of AD skin lesions. mokine and eotaxin as markers of severity of atopic dermatitis. Allergy 60: 685–688. 22. Pastore, S., F. Mascia, and G. Girolomoni. 2006. The contribution of keratino- Acknowledgments cytes to the pathogenesis of atopic dermatitis. Eur. J. Dermatol. 16: 125–131. We thank the staff of the Plateau de Biologie Experimentale de la Souris. 23. Akdis, M., J. Verhagen, A. Taylor, F. Karamloo, C. Karagiannidis, R. Crameri, S. Thunberg, G. Deniz, R. Valenta, H. Fiebig, et al. 2004. Immune responses in We particularly thank Dominique Kaiserlian for helpful scientific discus- healthy and allergic individuals are characterized by a fine balance between al- sions. We also thank P. Moingeon for providing Der f reagent and helpful lergen-specific T regulatory 1 and T helper 2 cells. J. Exp. Med. 199: 1567–1575. scientific discussion. 24. Simon, D., L. R. Braathen, and H. U. Simon. 2004. Eosinophils and atopic dermatitis. Allergy 59: 561–570. Disclosures 25. Huang, C. H., I. C. Kuo, H. Xu, Y. S. Lee, and K. Y. Chua. 2003. Mite allergen induces allergic dermatitis with concomitant neurogenic inflammation in mouse. The authors have no financial conflict of interest. J. Invest. Dermatol. 121: 289–293. 26. Savinko, T., A. Lauerma, S. Lehtima¨ki, M. Gombert, M. L. Majuri, N. Fyhrquist- Vanni, M. C. Dieu-Nosjean, L. Kemeny, H. Wolff, B. Homey, and H. Alenius. 2005. References ϩ 1. Herd, R. M., M. J. Tidman, R. J. Prescott, and J. A. Hunter. 1996. Prevalence of Topical superantigen exposure induces epidermal accumulation of CD8 T cells: a atopic eczema in the community: the Lothian Atopic Dermatitis Study. mixed Th1/Th2-type dermatitis and vigorous production of IgE antibodies in the Br. J. Dermatol. 135: 18–19. murine model of atopic dermatitis. J. Immunol. 175: 8320–8326. 2. Fukiwake, N., N. Furusyo, N. Kubo, H. Takeoka, K. Toyoda, K. Morita, 27. Grewe, M., S. Walther, K. Gyufko, W. Czech, E. Schopf, and J. Krutmann. 1995. S. Shibata, T. Nakahara, M. Kido, S. Hayashida, et al. 2006. Incidence of atopic Analysis of the cytokine pattern expressed in situ in inhalant allergen patch test dermatitis in nursery school children: a follow-up study from 2001 to 2004, reactions of atopic dermatitis patients. J. Invest. Dermatol. 105: 407–410. Kyushu University Ishigaki Atopic Dermatitis Study (KIDS). Eur. J. Dermatol. 28. Yawalkar, N., S. Schmid, L. R. Braathen, and W. J. Pichler. 2001. Perforin and 16: 416–419. granzyme B may contribute to skin inflammation in atopic dermatitis and psori- 3. Bieber, T., and D. Y. Leung. 2002. Atopic dermatitis. Marcel Dekker, New York. asis. Br. J. Dermatol. 144: 1133–1139. 4. Leung, D. Y. 2000. Atopic dermatitis: new insights and opportunities for thera- 29. Kehren, J., C. Desvignes, M. Krasteva, M.-T. Ducluzeau, O. Assossou, peutic intervention. J. Allergy Clin. Immunol. 105: 860–876. F. Horand, M. Hahne, D. Ka¨gi, D. Kaiserlian, and J.-F. Nicolas. 1999. Cytotox- ϩ 5. Samochocki, Z., W. Owczarek, and S. Zabielski. 2006. Can atopy patch tests with icity is mandatory for CD8 T cell-mediated contact hypersensitivity. J. Exp. aeroallergens be an additional diagnostic criterion for atopic dermatitis? Eur. Med. 189: 779–786. J. Dermatol. 16: 151–154. 30. Miyahara, N., K. Takeda, T. Kodama, A. Joetham, C. Taube, J. W. Park, 6. Leung, D. Y., M. Boguniewicz, M. D. Howell, I. Nomura, and Q. A. Hamid. S. Miyahara, A. Balhorn, A. Dakhama, and E. W. Gelfand. 2004. Contribution of ϩ 2004. New insights into atopic dermatitis. J. Clin. Invest. 113: 651–657. antigen-primed CD8 T cells to the development of airway hyperresponsiveness 7. Mamessier, E., and A. Magnan. 2006. Cytokines in atopic diseases: revisiting the and inflammation is associated with IL-13. J. Immunol. 172: 2549–2558. Th2 dogma. Eur. J. Dermatol. 16: 103–113. 31. Miyahara, N., B. J. Swanson, K. Takeda, C. Taube, S. Miyahara, T. Kodama, 8. Akdis, M., A. Trautmann, S. Klunker, I. Daigle, U. C. Kucuksezer, W. Deglmann, A. Dakhama, V. L. Ott, and E. W. Gelfand. 2004. Effector CD8ϩ T cells mediate R. Disch, K. Blaser, and C. A. Akdis. 2003. T helper (Th)2 predominance in atopic inflammation and airway hyper-responsiveness. Nat. Med. 10: 865–869.