CD4 CD25 Foxp3 Regulatory T Cells Are ...

The Journal of Immunology

CD4+CD25+Foxp3+ Regulatory T Cells Are Dispensable for Controlling CD8+ T Cell-Mediated Lung Inflammation Milena J. Tosiek,* Achim D. Gruber,† Sophie R. Bader,† Susanne Mauel,† Heinz-Gerd Hoymann,‡ Silvia Prettin,* Thomas Tschernig,x Jan Buer,{ Marcus Gereke,*,1 and Dunja Bruder*,1 Every person harbors a population of potentially self-reactive lymphocytes controlled by tightly balanced tolerance mechanisms. Failures in this balance evoke immune activation and autoimmunity. In this study, we investigated the contribution of self-reactive CD8+ T lymphocytes to chronic pulmonary inflammation and a possible role for naturally occurring CD4+CD25+Foxp3+ regulatory T cells (nTregs) in counterbalancing this process. Using a transgenic murine model for autoimmune-mediated lung disease, we demonstrated that despite pulmonary inflammation, lung-specific CD8+ T cells can reside quiescently in close proximity to selfantigen. Whereas self-reactive CD8+ T cells in the inflamed lung and lung-draining lymph nodes downregulated the expression of effector molecules, those located in the spleen appeared to be partly Ag-experienced and displayed a memory-like phenotype. Because ex vivo-reisolated self-reactive CD8+ T cells were very well capable of responding to the Ag in vitro, we investigated a possible contribution of nTregs to the immune control over autoaggressive CD8+ T cells in the lung. Notably, CD8+ T cell tolerance established in the lung depends only partially on the function of nTregs, because self-reactive CD8+ T cells underwent only biased activation and did not acquire effector function after nTreg depletion. However, although transient ablation of nTregs did not expand the population of self-reactive CD8+ T cells or exacerbate the disease, it provoked rapid accumulation of activated CD103+CD62Llo Tregs in bronchial lymph nodes, a finding suggesting an adaptive phenotypic switch in the nTreg population that acts in concert with other yet-undefined mechanisms to prevent the detrimental activation of self-reactive CD8+ T cells. The Journal of Immunology, 2011, 186: 6106–6118. lthough CD8+ T cells are crucial for maintaining protective immunity against viral infections and tumors, they can cause deleterious effects when they react against self tissues. To avoid such scenarios, the immune system has evolved complex and usually effective mechanisms of preventing undesirable responses to self components, thus protecting from autoimmunity. Although tolerance mechanisms are very effective, they are far from complete. Every person harbors a certain population of potentially autoaggressive lymphocytes that can, once they have escaped from peripheral control mechanisms,

A

*Immune Regulation Group, Helmholtz Centre for Infection Research, 38124 Braunschweig, Germany; †Department of Veterinary Pathology, Free University Berlin, 14163 Berlin, Germany; ‡Fraunhofer Institute for Toxicology and Experimental Medicine, 30625 Hannover, Germany; xInstitute of Anatomy and Cell Biology, Saarland University, 66421 Homburg, Germany; and {Institute of Medical Microbiology, University Hospital Essen, 45122 Essen, Germany 1

M.G. and D.B. contributed equally to this work.

Received for publication February 23, 2010. Accepted for publication March 23, 2011. This work was supported by a grant from the German Research Foundation (SFB 587 to D.B.) and by a stipend from the Wilhelm Hirte Foundation (to M.J.T.). The sequences presented in this article have been submitted to the GEO database under accession number GSE27379. Address correspondence and reprint requests to Dr. Dunja Bruder, Immune Regulation Group, Helmholtz Centre for Infection Research, Inhoffenstrasse 7, 38124 Braunschweig, Germany. E-mail address: [email protected] The online version of this article contains supplemental material. Abbreviations used in this article: AECII, alveolar epithelial cells type II; AHR, airway hyperresponsiveness; BALF, bronchoalveolar lavage fluid; BLN, bronchial lymph node; COPD, chronic obstructive pulmonary disease; HA, hemagglutinin; LIP, lymphoid interstitial pneumonia; MCh, methacholine; nTreg, CD4+CD25+ Foxp3+ regulatory T cell; SPC, surfactant protein C; Tc1, T cytotoxic type 1; Treg, regulatory T cell. Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1000632

cause autoimmune tissue destruction. For example, CD8+ T cellmediated tissue damage has been demonstrated in common organrelated autoimmune diseases such as type 1 diabetes and multiple sclerosis and has been postulated in many others (1, 2). Substantial evidence suggests that CD8+ T cells also contribute to the pathogenesis of a variety of lung diseases such as chronic obstructive pulmonary disease (COPD) and lymphoid interstitial pneumonia (LIP). This suggestion is based in part on the observation that these disorders are characterized by the preferential accumulation of CD8+ T cells in the alveolar space of the interstitium (3, 4). Moreover, an effector cytokine profile has been detected in CD8+ T cells isolated from patients with pulmonary disorders; the presence of these cytokines suggests that T cytotoxic type 1 (Tc1) cells make an important contribution to the pathologic findings associated with lung diseases (5, 6). This finding was further supported by experimental data obtained from the adoptive transfer of Tc1 cells specific for a neo self-antigen expressed in the lungs of recipient mice; the transfer led to severe pulmonary inflammation and lethality within a few days (7). Most recently, sustained activation of CTLs by pulmonary epithelial cells has been directly linked to the development of COPDlike diseases (8). Although the above-mentioned findings indicate the importance of CD8+ T cells in many pulmonary disorders, little is known about immune mechanisms preventing the expansion of autoreactive CD8+ T cells and thus the exacerbation of lung inflammation. Understanding the mechanisms involved in the generation and regulation of pathological CD8+ T cells may be crucial to therapeutic interventions for those respiratory diseases suggested to have an autoimmune cause. To study in more detail the role of CD8+ T cells in autoimmune respiratory disorders, we generated a novel transgenic murine model of CD8+ T cell-mediated autoimmune lung disease, which

The Journal of Immunology in some aspects resembles LIP in humans. Whereas especially in adults LIP is linked to autoimmune disorders, such as Sjo¨gren syndrome, rheumatoid arthritis, or systemic lupus erythematosus, in children LIP is mainly associated with systemic immunodeficiency states, such as HIV. Of note, LIP develops in, on average, 50% of HIV-positive children. In adults, LIP is postulated to be of autoimmune origin; in children, it may be of an indirect viral cause (9, 10). However, the pathogenesis of LIP remains poorly understood and requires further study. Although some animal models of virus-induced LIP have been described (11, 12), the transgenic murine model described in this study is, to our knowledge, the first animal model of autoimmune-induced LIP. To limit self-reactive tissue attack, the immune system constitutively produces a mature and functionally distinct CD4+ T lymphocyte population specialized in the suppression of immune responses; this population acts in secondary lymphoid tissues and at sites of inflammation (13, 14). These naturally arising CD4+ CD25+Foxp3+ regulatory T cells (nTregs) are produced in the thymus, cannot be induced de novo from naive T cells in the periphery, and serve as immune safeguards that keep self-reactive responses under control by bystander suppression (15, 16). The nTregs play a crucial role in preventing the priming and activation of self-reactive CD8+ and CD4+ T cells. Recent studies have extensively examined the role of nTreg numbers and nTreg suppressor activity in pulmonary disorders, such as asthma, allergic airway inflammation, and COPD in steadystate conditions, upon exacerbation, or after therapeutic treatment (17, 18). Studies have shown that the nTregs of pediatric asthma patients are fewer in number and are more likely to be dysfunctional than those of healthy subjects and corticosteroid-treated asthma patients (19). Similarly, decreased expression of Foxp3 in regulatory T cells (Tregs) has been observed in the small airways of patients with COPD (20). Tregs are an important target for immunotherapeutic approaches; such treatments may be an interesting alternative to the nonspecific drugs commonly used to treat pulmonary disorders. It has recently been shown that therapeutic transfers of Tregs can resolve the symptoms of chronic allergen-induced inflammation and can prevent the development of airway remodelling (18). Nevertheless, these findings relate primarily to nTreg control over CD4+ T lymphocytes, whereas the importance of nTreg-mediated control of lung-specific CD8+ T cells, especially in the context of ongoing autoimmunity, has not yet been completely elucidated. Using a novel murine model of chronic CD8+ T cell-mediated pulmonary inflammation, we found that nTregs are indeed involved in maintaining CD8+ T cell tolerance to lung Ag, because their depletion leads to partial activation of these T cells in the lung and in the lung-draining lymph nodes. However, transient ablation of nTregs results neither in the development of fully competent effector CD8+ T cells nor in the exacerbation of lung inflammation. This finding indicates the existence of additional powerful tolerance mechanisms that prevent uncontrolled expansion of the population of self-reactive cytotoxic CD8+ T lymphocytes in the lung.

Materials and Methods Mice CL4 transgenic mice expressing the abTCR specific for the H-2Kd–presented epitope hemagglutinin (HA) 512–520 have been described elsewhere (21, 22). SPC-HA mice expressing the influenza virus A/PR8/34 HA under the transcriptional control of surfactant protein C (SPC) promoter in alveolar epithelial cells type II (AECII) in the lung have been described previously (23). If not stated otherwise, animals of 4 mo of age were used, classified as stage III according to Table I. All transgenic mice used for the study were bred in the animal facility of the Helmholtz Centre for Infection

6107 Research and were kept under specific pathogen-free conditions. All animal experiments were performed according to national and institutional guidelines.

Histology Lungs were perfusion-fixed with neutral-buffered 4% formalin, embedded in paraffin, sectioned at 4 mm, and stained with H&E. Immunohistochemical analysis of paraffin sections was performed with biotinylated Abs to CD3 (CD3–12; Serotec, Du¨sseldorf, Germany) and CD45R/B220 (RA3–6B2; BD Biosciences, Heidelberg, Germany), as previously described (23). For histopathological and morphometrical analyses, lymphocytic aggregations in SPC-HA3CL4 mice were morphometrically quantified for stage I (0- or 3-d-old mice), stage II (21-d-old mice), and stage III (4-mo-old mice) in a blinded manner using the analySIS FIVE system (Digital Imaging System, Olympus Soft Imaging Solutions, Munster, Germany). CD3+ T cell and B220+ B cell staining was performed to identify lymphoid follicles. The lymphoid follicle areas were defined by the mean of the squared micrometer value for each cluster of lymphocytes encountered in the lung parenchyma. In such a manner, five randomly distributed histological sections per mouse (three to eight individuals) per group were examined in two independent experiments. Age-matched single-transgenic littermates were included as controls. Mean and SD of cross-sectional areas of individual lymphoid aggregates were calculated. The fraction of lung parenchyma that contained lymphoid follicles was determined by dividing the total area covered by follicles by the total lung area of all sections evaluated per group. Differences between the morphometrical data of the groups of stage II and III were determined by the Mann–Whitney U nonparametric test, and p values #0.05 were considered significant.

Analysis of bronchoalveolar lavage cells The lungs were rinsed twice with 0.75 ml ice-cold PBS via a cannula (B Braun, Melsungen, Germany) inserted into the trachea. The total cell number in the bronchoalveolar lavage fluid (BALF) was determined by staining with trypan blue dye and referenced to the volume of the lavage fluid. For examination of BALF cell composition, samples were subjected to Fc-block with anti-CD16/32 Ab (24G2; BD Biosciences) to avoid unspecific staining and were then stained with the following Ab mix: antiCD3 biotin (145-2C11; BD Biosciences), anti-CD11c FITC (HL3; BD Biosciences), anti-CD11b PE (M1/70; BD Biosciences), anti-CD19PerCPCy5.5 (eBioscience, Frankfurt, Germany), anti-Gr1 PE-Cy7 (BD Biosciences), and anti-CCR3 allophycocyanin (R&D Systems, Abingdon, UK). Samples were then analyzed with a FACSCanto flow cytometer (BD Biosciences). Alveolar macrophages were defined as a CD11bhiCD11chi population and were excluded from further gating. T lymphocytes were defined as CD3+CD192 cells, B lymphocytes as CD32CD19+ cells, neutrophils as CD32CD192CD11bhiGr1hi cells, and eosinophils as CD32 CD192CCR3+ cells. The percentage of each population was calculated by referencing to the number of viable cells in each measurement.

Lung function testing Pulmonary function of 4-mo-old mice (representative of stage III of the inflammation) was measured invasively in anesthetized (67 mg/kg i.p. propofol and 1.5% halothane), orotracheally intubated, spontaneously breathing mice by body plethysmography, as previously described (24). Briefly, mice were subjected to careful orotracheal intubation and placed spontaneously breathing in the supine position into temperature-controlled body plethysmographs (HSE-Harvard Apparatus type 871; Harvard Apparatus, March-Hugstetten, Germany). Lung resistance (RL) and dynamic lung compliance (Cdyn) were calculated from the measured differences in transpulmonary pressure, tidal flow, and volume signals over each complete breath cycle. The respiratory variables were continuously recorded under baseline conditions with HEM 4.2 software (Notocord, Croissy, France). Next, airway responsiveness to increasing doses of aerosolized methacholine (MCh) chloride (Sigma, Deisenhofen, Germany) was assessed by using a Bronchy type III generator (Fraunhofer ITEM) in combination with a feedback-dose control system as previously described (25). The effective dose of MCh that elicited an increase of RL above baseline of 100% or of 150% (ED100 and ED150, respectively) was calculated from the individual dose-response curve.

Abs and flow cytometry The following Abs (obtained from BD Biosciences, eBioscience, and BioLegend [Uithoorn, The Netherlands]) were used: anti-mouse mAbs anti-CD8 (Ly-2, 53-6.7), anti-CD25 (7D4), anti-CD62L (MEL-14), antiCD43 (1B11), anti-CD4 (L3T4, RM4-5), anti-CD44 (IM7), anti-CD122

6108 (TM-b1), anti-CD103, and anti-CD127. Pro5 MHC pentamer H-2Kd IYSTVASSL (Proimmune, Oxford, U.K.) was used to detect HA-specific T cells in the context of MHC class I. The anti-CD25 Ab (PC61) was purified from hybridoma supernatants. Flow cytometry was performed with a FACSCalibur or FACSCanto (BD Biosciences) flow cytometer. Compensated data were analyzed with FlowJo software (version 5.7.2 and 8.8.4; Treestar, Olten, Switzerland). Cell sorting was performed with a MoFlow cell sorter (DAKO Cytomation, Fort Collins, CO).

CD8+ T CELL TOLERANCE TO PULMONARY SELF-ANTIGEN GG-39, and for ribosomal protein S9 (RPS9) 59-CTG GAC GAG GGC AAG ATG AAG C-39, 59-ACA CGA GTA GGC GGT TGC AGT-39.

Gene expression microarray

Single-cell suspensions were prepared by passing cells through a 70-mm nylon mesh. Erythrocytes were removed from splenocytes by treatment with erythrocyte lysis buffer. Lymphocytes from the lung were isolated according to the protocol previously described (23).

For gene expression profiling, lymphocytes were isolated from lungs and separately from bronchial lymph nodes (BLNs) of CL4 and SPC-HA3CL4 mice, as well as from CL4 mice that had been infected with sublethal doses of influenza virus A/PR8/34 5 d earlier. The mRNA purified from the sorted lymphocytes was applied to Affymetrix Mouse Genome 430 2.0 chips (Affymetrix, Santa Clara, CA) and hybridized for 16 h, according to the manufacturer’s instructions. Data analysis was performed with gene expression software (MicroDB and Data Mining Tool; Affymetrix), and data have been deposited in the GEO database under accession number GSE27379 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE27379).

Intracellular cytokine staining

Statistical analysis

The intracellular production of cytokines was stimulated by incubating single-cell suspensions with PMA and ionomycin (Sigma) for 2 h. Incubation continued for the next 4 h; cytokine accumulation inside the endoplasmic reticulum was enhanced by the addition of 5 mg/ml brefeldin A (Sigma). Surface staining was performed with pentamer, anti-CD8 Ab, and anti-CD43 Ab, followed by 20 min of fixation with 2% paraformaldehyde, permeabilization with 0.01% Igepal CA-630 (Sigma), and 30 min of intracellular staining with a corresponding Ab: anti–IFN-g (16G6; BD Bioscience), a–IL-2 (JES6-5H4; eBioscience), or anti–IL-10 (JES5-16E3; eBioscience). Cytokine production by CD8+ pentamer-bound cells was determined by flow cytometry. When IFN-g production was assessed upon antigenic stimulation (Fig. 6B), splenocytes were adjusted to 2.5 3 105 HA-specific CD8+ T cells and cultured in a 96-well flat-bottom plate in the presence of HA 512–520 peptide (1 mg/ml) in IMDM/10% FCS. After 20 h of incubation at 37˚C, brefeldin A was added to the culture. To assess ex vivo IFN-g production, cells isolated from SPC-HA3CL4 mice and CL4 mice were incubated for 4 h with brefeldin A (without PMA and ionomycin) and then stained as described above. Foxp3 (FJK-16a; eBioscience), perforin (eBioOMAK-D; eBioscience), and granzyme B (16G6; eBioscience) intracellular staining was performed according to the manufacturer’s instructions.

If not stated otherwise, one-tailed, unpaired Student t test was used for comparison of two groups, whereas for statistical analysis of more than two groups, single one-way ANOVA, followed by one-tailed, unpaired Student t test was applied. Statistical significance was set at the level of p , 0.05.

Isolation of cells

Proliferation assay Sorted HA-specific CD8+ T lymphocytes (105) were cocultured with irradiated BALB/c splenocytes (4 3 105) in a 96-well flat-bottom plate in the presence or absence of HA 512–520 peptide. After 24 to 36 h, the cells were pulsed with 1 mCi [3H]thymidine for the final 6 to 12 h, and [3H] thymidine incorporation was measured by scintillation counting.

Adoptive transfer Splenocytes from SPC-HA3CL4 mice were enriched for CD8+ T cells by “untouched” column separation using the CD8 T cell isolation kit and autoMACS (both Miltenyi Biotec) according to the manufacturer’s instructions. Purity of CD8+ T cells and content of Ag-specific CD8+ T cells were confirmed by flow cytometry. Ag-specific CD8+ T cells (5 3 106) were injected into the tail veins of age- and sex-matched SPC-HA and control BALB/c mice. For adoptive transfers of Tc1 cells, splenocytes from SPC-HA3CL4 mice enriched in CD8+ T lymphocyte population as described earlier were activated in vitro. For this purpose, 5 3 105 CD8+ T cells and 5 3 106 irradiated BALB/c splenocytes were cultured for 3 d at 37˚C in a 24-well plate in IMDM/10% FCS in the presence of 1 mg/ml HA 512–520 peptide, 1 ng/ml IL-2, and 20 ng/ml IL-12 followed by addition of fresh IMDM/10% FCS medium with 1 ng/ml IL-2. Three days later, activated lymphocytes were enriched by Ficoll density centrifugation gradient. Their purity, Ag specificity, and activation status were confirmed by flow cytometry. HA-specific Tc1 cells (2.5 3 106) were adoptively transferred into SPC-HA and BALB/c control recipients. Seven days later, lungs were sampled for histological evaluation.

Real-time PCR Total RNA was prepared from sorted CD8+ pentamer+ T cells using the RNeasy kit (Qiagen, Hilden, Germany) after DNase digestion (Qiagen) and cDNA synthesis by SuperScript II reverse transcriptase and oligonucleotide mixed with random hexamer primers (Invitrogen, Darmstadt, Germany) according to the manufacturers’ instructions. Real-time RT-PCR was performed in an ABI PRISM cycler (Applied Biosystems, Darmstadt, Germany) using a Maxima SYBR Green Rox qPCR Master Mix (MBI Fermentas, St. Leon-Rot, Germany) and specific primers for Foxo1 59-TTT CTA AGT GGC CTG CGA GT-39, 59-GTA AGG GAC CAC ATA GGT

Results Expression of self-antigen in the lung in the presence of self-reactive CD8+ T cells results in pulmonary inflammation and airway hyperresponsiveness To investigate how autoimmune conditions in the lung influence the phenotype and function of self-reactive CD8+ T cells, we generated a novel transgenic mouse model for autoimmune-mediated pulmonary inflammation. This model was developed by crossing SPC-HA mice, which express the neo self-antigen influenza HA in the airway epithelium (23), with CL4 mice producing HA-specific CD8+ T lymphocytes (21). No specific phenotype develops in single transgenic SPC-HA or CL4 mice. In contrast, chronic selfantigen exposure in the lung of double transgenic SPC-HA3CL4 mice to HA-specific CD8+ T cells results in phenotypic alterations in the respiratory system. Histological changes were classified as three distinct stages of the disease (Table I). SPC-HA3CL4 embryos and mice younger than 7 d did not exhibit any histological abnormalities (stage I, Fig. 1A). Mice between 7 d and 3 wk of age developed increasing numbers of diffuse interstitial to perivascular lymphocytic infiltrates (stage II, Fig. 1Bb and Supplemental Fig. 1A) that expressed the CD3 Ag (Fig. 1Bd) and fewer lymphocytes expressing the B220 Ag (Fig 1Bf) compared with single transgenic littermates (Fig. 1Ca, c, e). Increasing age during this stage was associated with a tendency toward the appearance of follicle-like lymphocytic aggregates around the walls of medium-sized arteries and veins. In mice of 3 wk or older, multiple prominent lymphoid follicles were present around medium-sized arteries and veins and, less commonly, around bronchioles (stage III, Fig. 1Cb–d and Supplemental Fig. 1B). Airways were partially obstructed by peribronchial hyperplastic follicles (Fig. 1Cc). The more diffuse interstitial, interalveolar lymphocytic infiltrations were not present any more in stage III. No further development of this quiescent, resting stage was observed beyond 3 wk of age. Immunohistochemical characterization of the follicles revealed mainly CD3+ T cells but occasionally also accumulations of B220+ B lymphocytes within the lymphocytic aggregates (Fig. 1Ce, f) suggesting the formation of secondary BALT-like lymphoid follicles (26). We also quantified number and size of lymphoid follicle areas. In total, 530 follicle-like structures were measured in 500 sections in the group of 3-wk-old mice that already had developed distinguishable perivascular lymphocytic follicles representative of stage II and 713 follicle-like structures in the group of 4-mo-old mice representative of stage III. The average size of

The Journal of Immunology

6109

Table I. Stages of disease progression in SPC-HA3CL4 mice Stage

Age of Mice

Lesions

I II

,7 d 7 d to 3 wk

III

.3 wk

No histological changes. Diffuse interalveolar interstitial and perivascular infiltration with T lymphocytes; increasing perivascular lymphocytic follicles. Resting perivascular and fewer peribronchial lymphocytic follicles consisting of B and T lymphocytes, no diffuse interstitial infiltrates.

individual lymphoid follicles was significantly larger in mice of stage III than of stage II (Fig. 1Da). In contrast, no differences were observed between the percentages of lymphoid follicles in

the total lung tissue (Fig. 1Db). No follicles were found in newborn SPC-HA3CL4 mice (stage I) or any of the analyzed SPCHA and CL4 littermates (control). There was no evidence of further progression of the inflammation or any other histological changes in mice examined until 2 y of age. Inflammation was restricted exclusively to the lung with no other organs affected. We did not find any prerequisites for accumulation of immune cells in the spleen and in the distant lymph nodes of SPC-HA3CL4 mice. Though absolute cell numbers in lung-draining bronchial lymph nodes of SPC-HA3CL4 mice were slightly elevated, the difference was not statistically significant (Supplemental Fig. 2A). Analysis of BALF from 4-mo-old animals demonstrated a significantly higher number of cells in SPC-HA3CL4 mice than in control animals (Fig. 2A). Moreover and in line with data obtained by immunohistochemical characterization (Fig. 1C), SPCHA3CL4 mice exhibited a significantly higher number of T and B lymphocytes (lymphocytosis) in the BALF; this finding could

FIGURE 1. Lymphocytic infiltrates develop early in life in the lung of SPC-HA3CL4 mice. Paraffin-embedded sections of lungs from SPC-HA3CL4 and control SPC-HA mice were prepared and stained with H&E. Slides (Bc–f, Ce, Cf) were stained with anti-CD3 or anti-B220 Abs with diaminobenzidine as chromogen (brown) and hematoxylin as counterstain (blue). A, Representative lung tissue sections of stage I SPC-HA3CL4 mice at the day of birth (a) and 3 d after birth (b) not showing any signs of inflammation. Scale bars, 50 mm. B, Lungs of 9-d-old mice (representative of stage II). a, SPC-HA control mouse. b, d, and f, SPC-HA3CL4 mice. c and e, CL4 control mice. b, Interalveolar septa widely distended by lymphocytes. d, Prominent infiltration of the interalveolar septa with CD3+ T lymphocytes, sometimes with small lymphocytic aggregates surrounding arterial or venous blood vessels. f, Sparse B220+ B lymphocytes in the interalveolar septa. Scale bars, 50 mm. C, Lungs of 4-mo-old mice (representative of stage III). a, Healthy lungs of SPC-HA (right) and CL4 (left) control animals. b–f, SPC-HA3CL4 mice. b, Perivascular (#) and peribronchial (*) lymphoid follicles. c, Marked peribronchial lymphocytic infiltrate partly obstructing the bronchial lumen. d–f, Formation of a perivascular lymphoid follicle, consecutive serial sections from the same follicle: (d) H&E staining; (e) immunohistochemical detection of CD3+ lymphocytes (brown) and (f) detection of B220+ lymphocytes (brown). Scale bars: (a, b) 200 mm; (c) 50 mm; (d–f) 100 mm. Data represent individual samples collected in two (A), two (B), and five (C) independent experiments with similar results performed with 14 (A), 10 (B), and 50 (A) mice analyzed individually. D, Quantification of lymphoid aggregations. a, The mean and SD of cross-sectional areas of individual lymphoid aggregates measured in 4-mo-old SPC-HA3CL4 mice (representative of stage III) were significantly (p # 0.009) larger than those in 3-wk-old SPC-HA3CL4 mice (representative of stage II). No follicle formation was found in newborn mice (stage I) and any of the 4-mo-old single-transgenic littermates (control). b, In contrast, the relative areas of the lung that were covered by lymphoid follicles were similar in stage II and stage III SPC-HA3CL4 mice. Differences between the morphometrical data of the groups of stage II and III were determined by the Mann–Whitney U nonparametric test, and p values #0.05 were considered significant.

6110 indicate preferential migration of these cellular subsets to the lung (Fig. 2B). To investigate whether lung inflammation in SPC-HA3CL4 mice may be associated with impaired pulmonary function, the physiology of the lung was examined in adult 4-mo-old mice with invasive methods. In spontaneously breathing mice, no significant difference was detectable between the groups in baseline measurements (lung resistance and dynamic lung compliance) (Fig. 3A, 3B). In contrast, when challenged with MCh aerosol, SPCHA3CL4 mice exhibited significantly more airway hyperresponsiveness (AHR) than did the control group (BALB/c, CL4, and SPC-HA mice), as demonstrated by a remarkable decrease in the MCh dose (Fig. 3C, 3D). Collectively, expression of selfantigen in the lung epithelium upon the presence of lungspecific CD8+ T cells resulted in pulmonary disease in SPCHA3CL4 mice. Nonetheless, compared with wild-type and single transgenic animals, SPC-HA3CL4 mice do not exhibit increased mortality and reach a normal life span; this finding is highly suggestive of the existence of potent immune regulatory mechanisms that control self-reactive CD8+ T cells in the lung. To dissect the possible contribution of other immune cell types, for example, B cells and CD4+ T lymphocytes, to the phenotype of

FIGURE 2. SPC-HA3CL4 mice demonstrate elevated cell content and lymphocytosis in BALF. Total cell numbers in BALF from 4-mo-old SPCHA3CL4 and control mice were determined by cell counting. BALF cell composition was assessed by Ab staining and measured by FACS. A, Elevated cell infiltration in BALF from SPC-HA3CL4 mice. Single one-way ANOVA (for control groups: BALB/c, CL4, and SPC-HA, p . 0.05), followed by one-tailed, unpaired Student t test (controls versus SPCHA3CL4): p = 0.00013. One-way ANOVA comparing all four mouse groups: p = 0.003. B, Increase in T and B lymphocyte content in BALF from SPC-HA3CL4 mice. Single one-way ANOVA (for control groups: BALB/c, CL4, and SPC-HA, p . 0.05), followed by one-tailed, unpaired Student t test (controls versus SPC-HA3CL4): p = 0.006 for T cells; p = 0.047 for B cells; p . 0.05 for macrophages, eosinophils, and neutrophils. One-way ANOVA comparing all four mouse groups: p . 0.05 for all cell types analyzed. Data show the results obtained from three independent experiments with six to nine mice studied individually. Error bars: SD. *p , 0.05, **p , 0.001, ***p , 0.0001.

CD8+ T CELL TOLERANCE TO PULMONARY SELF-ANTIGEN SPC-HA3CL4 mice, we estimated ratios of T cells to B cells and CD4/CD8 T cells in SPC-HA3CL4 and in control CL4 mice. Proportions of CD3+ and B220+ lymphocytes in SPC-HA3CL4 mice resembled those of corresponding CL4 control (Supplemental Fig. 2B). Despite slightly elevated levels of total Igs IgA and IgG in the BALF of SPC-HA3CL4 mice, expression of activation markers (CD80, CD86, MHC class II) on SPC-HA3CL4– derived B cells did not differ from the corresponding control (Supplemental Fig. 3). The only obvious difference observed regarding the cell composition in SPC-HA3CL4 and control animals was elevated CD4/ CD8 T cell ratios in spleen, lymph nodes, and lung of SPCHA3CL4 mice. Nevertheless, further phenotypic characterization of nonregulatory CD4+Foxp32 T lymphocytes in SPC-HA3CL4 mice did not reveal any signs of recent activation of these cells (Supplemental Fig. 4). Thus, we largely excluded that B cells and CD4+ T cells may contribute to pulmonary pathology in adult SPC-HA3CL4 mice. Therefore, we focused on dissecting the role of autoaggressive CD8+ T cells in pulmonary disorder observed in SPC-HA3CL4 mice. Lung-specific CD8+ T lymphocytes accumulate in the lung but do not resemble effector T cells Although self-reactive lymphocytes are supposed to be eliminated in the thymus by the process of negative selection (27), and although SPC-HA mice express the self-antigen HA in the thymus (23), autoaggressive HA-specific CD8+ T cells escape from clonal deletion in SPC-HA3CL4 littermates (Fig. 4A). Moreover, we observed an increased percentage of self-reactive CD8+ T cells preferentially at the site of Ag expression (i.e., in the lungs of SPC-HA3CL4 mice). Although the difference was not statistically significant, the proportion of lung-specific CD8+ T cells was also higher in the Ag-draining BLNs of SPC-HA3CL4 mice than in those of control CL4 mice. In the spleen and in the mesenteric lymph nodes, which are distant and unrelated to the lung, the distribution of HA-specific CD8+ T cells in SPC-HA3CL4 mice resembled that in CL4 control mice (Fig. 4B). Assuming that the HA-specific CD8+ T cells infiltrating the lung were effector T cells, we determined their phenotype in more detail. To this end, we measured the expression of activation and homing markers on lung-specific CD8+ T cells. Strikingly, despite pulmonary inflammation, the phenotype of lung-specific CD8+ T cells in SPC-HA3CL4 mice resembled that of such cells from healthy control CL4 animals. To our surprise, this finding was also true for lung-specific CD8+ T lymphocytes that were isolated directly from the inflamed lung; that is, from the site of Ag exposure (Fig. 5A). Marginal expression of activation markers was associated with low production of proinflammatory cytokines, such as IL-2 or IFN-g by self-reactive CD8+ T cells in the lungs and BLNs (Fig. 5B, 5C). Notably, we also observed a tendency toward elevated IFN-g secretion by splenic CD8+ T cells from SPCHA3CL4 mice (Fig. 5C). The observation that lung-specific CD8+ T cells residing in direct proximity to their specific Ag exhibit no signs of activation led to the hypothesis that these cells may have developed toward a regulatory phenotype. However, screening of SPC-HA3CL4–derived HA-specific CD8+ T cells for the expression of marker molecules related to classical regulatory phenotype (e.g., Foxp3 and CTLA-4) (Supplemental Fig. 5) and to cytolysis-dependent function (e.g., perforin and granzyme B) (Supplemental Fig. 6A, 6B) did not reveal any differences compared with naive HA-specific CD8+ T cells from CL4 control mice. Moreover, production of anti-inflammatory cytokines by self-reactive CD8+ T cells such as TGF-b and IL-10 did not differ between SPC-HA3CL4 and CL4 control mice (Supplemental Fig.

The Journal of Immunology

6111

FIGURE 3. SPC-HA3CL4 mice reveal AHR. Pulmonary function was measured invasively in anesthetized, orotracheally intubated, and spontaneously breathing mice (4 mo of age corresponding with stage III; Table I) by body plethysmography at baseline and increasingly provoked with MCh aerosol (inhaled doses: 0.063, 0.125, 0.25, 0.50, 1.0, 2.0, and 4.0 mg MCh). A, Unaltered lung resistance upon baseline measurements. B, Unaffected dynamic compliance in SPC-HA3CL4 mice upon baseline measurements. C and D, AHR in SPC-HA3CL4 mice upon MCh challenge (ED100 and ED150). Data show the results obtained in two independent experiments with nine mice studied individually. Error bars indicate SEM. Single one-way ANOVA (for control groups: BALB/c, CL4, and SPC-HA, p . 0.05), followed by one-tailed, unpaired Student t test (controls versus SPC-HA3CL4): p = 0.019. Oneway ANOVA comparing all four mouse groups: p . 0.05. *p , 0.05. Cdyn, dynamic lung compliance; ED, effective MCh dose to produce a 100% (C) or 150% (D) increase in lung resistance (inhalational dose, mg MCh); RL, lung resistance.

6C, 6D). Together, these data largely exclude that lung-reactive CD8+ T cells may have converted into suppressor cells. Self-reactive CD8+ T lymphocytes exhibit features of memory T cells Because lung-specific CD8+ T lymphocytes from SPC-HA3CL4 mice were found to be neither recently activated T cells nor regulatory T cells, we hypothesized that they might exhibit a memory-like phenotype. To test this assumption, we assessed the ability of SPC-HA3CL4–derived CD8+ T cells to respond to HA Ag upon in vitro stimulation. As expected, CL4-derived; lymphocytes responded to their specific epitope by massive proliferation. Notably, the degree of clonal expansion was significantly higher when CD8+ T cells were isolated from SPC-

FIGURE 4. Autoreactive CD8+ T lymphocytes escape from clonal deletion and accumulate in the lungs of SPC-HA3CL4 mice. Lymphocytes from various organs of SPC-HA3CL4 and control CL4 mice were isolated, and the percentage of HAspecific CD8+ pentamer-bound T cells in each of the compartments was determined by flow cytometry. A, Potentially autoreactive T cells (CD8+ pentamer +) are not depleted in the thymus of SPC-HA3CL4 mice. FACS plots show one representative staining of CD8+ pentamer+ thymocytes in CL4 and SPC-HA3CL4 mice (left) and data from one experiment with seven mice analyzed individually (right) did not reveal any significant difference between both mouse lines (p . 0.05). B, Increased percentage of CD8+ pentamer+ T lymphocytes in the lungs of SPC-HA3CL4 mice. Data show the outcome of three independent experiments with at least two mice per group analyzed individually. Filled circles represent CD8+ pentamer+ T cells isolated from CL4 mice; open triangles represent those from SPC-HA3CL4 mice. *p , 0.05. Animals used in this experiment were 4 mo old. MLN, mesenteric lymph nodes.

HA3CL4 mice (Fig. 6A); this finding suggests that a certain proportion of self-reactive CD8+ T cells from autoimmune SPCHA3CL4 mice may be Ag-experienced and thus may resemble memory T cells. To corroborate further this hypothesis, we assessed IFN-g production, being another hallmark of memory T cell phenotype (28), in HA-specific CD8+ T cells from CL4 and SPC-HA3CL4 mice after antigenic stimulation in vitro. Indeed, the number of IFN-g–producing, HA-specific CD8+ T cells was nearly twice as high in SPC-HA3CL4 mice as in CL4 mice (Fig. 6B). Of note, ex vivo analysis showed that IFN-g production in unstimulated HA-specific CD8+ T cells from SPC-HA3CL4 mice was elevated even at baseline. In addition, the expression of the CD8+ T cell memory-related markers CD44 and CD122 on selfspecific CD8+ T cells from autoimmune-prone SPC-HA3CL4

6112

CD8+ T CELL TOLERANCE TO PULMONARY SELF-ANTIGEN

FIGURE 5. Self-reactive CD8+ T cells remain quiescent despite constant Ag exposure in the lung. Lymphocytes isolated from various organs of SPCHA3CL4 and control CL4 mice were stained with anti-CD8, pentamer, and either with anti-CD25, anti-CD43, or anti-CD62L, as indicated, or with corresponding isotype controls for anti-CD25, anti-CD43, or anti-CD62L and analyzed by FACS. A, Expression profile of a given marker on gated CD8+ pentamer+ T lymphocytes in SPC-HA3CL4 and CL4 mice. Data represent one of three experiments with similar outcome with three mice per group analyzed individually. Gray lines indicate expression of a given marker by CL4-derived cells; black lines indicate expression of a given marker by cells isolated from SPC-HA3CL4 mice; blue lines indicate isotype control for a given marker on CL4-derived cells; red lines indicate isotype control for a given marker on SPC-HA3CL4–derived cells. Cytokine production (IFN-g and IL-2) was determined in a CD8+ pentamer+ T lymphocyte population by intracellular staining and was measured with FACS. Percentage of cytokine-producing cells among gated CD8+ pentamer+ population is shown. B, Production of IL-2 by CD8+ pentamer-bound T cells. Gray bar, CL4 mice; black bar, SPC-HA3CL4 mice. Error bars: SD. C, IFN-g production by CL4-derived and SPC-HA3CL4–derived CD8+ pentamer-bound T cells. Gray bar, CL4 mice; black bar, SPC-HA3CL4 mice. Data (B, C) represent outcome of three independent experiments with three animals per group. Mice used for the experiments were 4 mo of age.

mice was also slightly increased (Fig. 6C). Notably, enhanced IFN-g production (Figs. 5C, 6B) and elevated expression of markers related to memory T cells was restricted to a small population of self-reactive CD8+ T cells found in the spleen. These observations led us to the conclusion that a certain proportion of cells within the potentially autoreactive CD8+ T cell pool in SPC-HA3CL4 mice may be memory-like T cells accumulating in the spleen (28–31). Self-reactive CD8+ T cells are quiescent and fail to provoke acute lung inflammation To investigate the quiescent phenotype of self-specific CD8+ T cells from BLNs and lungs of SPC-HA3CL4 mice in more detail, we performed gene expression profiling (Fig. 7A). Several genes related to CD8+ T cell activation and effector function were downregulated in SPC-HA3CL4–derived self-specific CD8+ T lymphocytes but not in activated and naive CD8+ T cells from CL4 controls. Autoreactive CD8+ T cells isolated from the inflamed area of autoimmune SPC-HA3CL4 mice demonstrated even lower expression of some activation-related genes than did naive CD8+ T cells from CL4 mice (Fig. 7A); the finding suggesting somehow quiescent status of self-specific CD8+ T cells in SPC-HA3CL4 mice. To corroborate further this hypothesis, we measured relative expression of Foxo1, a transcription factor associated with homeostasis of naive T cells and quiescence (32, 33), on self-reactive CD8+ T lymphocytes sorted from spleen, BLNs, and lung of SPC-HA3CL4 mice. Relative expression levels of Foxo1 mRNA in self-specific CD8+ T cells isolated from BLNs and lungs of SPC-HA3CL4 mice was as high as that found in naive CD8+ T cells from CL4 mice (Fig. 7B). In contrast and in line with the proposed memory cell phenotype, self-reactive CD8+

T lymphocytes from the spleen of SPC-HA3CL4 mice exhibited decreased Foxo1 mRNA expression in comparison with naive CL4 splenocytes. Highest downregulation of Foxo1 expression was, as expected, found in in vitro-activated HA-specific CD8+ T cells. To substantiate further the quiescent status of self-reactive CD8+ T cells, adoptive transfer of autoreactive CD8+ T cells from the spleen of SPC-HA3CL4 mice into SPC-HA mice was performed. Indeed, lung inflammation was provoked only in the case where CD8+ T cells were preactivated in vitro prior to transfer, whereas injection of unmanipulated self-reactive CD8+ T cells from SPCHA3CL4 mice did not induce any inflammatory changes in the lung of SPC-HA mice (Fig. 7C). Depletion of Tregs does not result in gain of effector function by self-reactive CD8+ T lymphocytes Because several clinical reports have suggested a key role for Tregs in the regulation of pulmonary disorders (17, 18), we addressed the putative nTreg contribution to CD8+ T cell tolerance in the inflamed lung. The numbers of CD4+CD25+Foxp3+ nTregs were not elevated in the thymus, where nTregs develop (14, 15), or in the BLNs and lungs of SPC-HA3CL4 mice (Supplemental Fig. 7A, 7B). As expected, the nTreg population in SPC-HA3CL4 mice proved to be fully functional and able to prevent efficiently the expansion of the population of self-reactive CD8+ T cells in vitro (Supplemental Fig. 7C). To investigate the potential contribution of nTregs to the observed CD8+ T cell tolerance to pulmonary self-antigen, we depleted the numbers of nTregs by injecting anti-CD25 Ab (PC61) as reported before (34, 35). Within the first 2 wk after Ab treatment, a substantial decline in the number of CD4+CD25+ nTregs was observed in the blood of SPCHA3CL4 mice and in that of control CL4 mice (data not shown).

The Journal of Immunology

6113

FIGURE 6. Autoaggressive CD8+ T lymphocytes exhibit features of memory cells. Transgenic CD8+ T lymphocytes from BLNs and spleen of SPCHA3CL4 and CL4 mice were sorted and cocultured with irradiated BALB/c splenocytes in the presence or absence of their specific HA Ag. A, Proliferation of lymphocytes in response to HA peptide was measured by thymidine incorporation. Data represent one of three experiments performed in triplicate with similar outcome using three mice per group. Gray bars, CL4 mice; black bars, SPC-HA3CL4 mice. Error bars: SD. *p , 0.05, **p , 0.01. B, IFN-g production on gated HA-specific CD8+ splenocytes from SPC-HA3CL4 and CL4 mice was determined by flow cytometry after 24 h of incubation with their specific Ag (HA stimulation) or directly ex vivo (no stimulation). Data represent one of three independent experiments with similar outcome with at least three animals per group analyzed individually. C, Expression of CD8+ T cell memory-related surface markers is shown in splenocytes from SPCHA3CL4 and CL4 mice gated on HA-specific CD8+ T lymphocyte population. Data represent one of two independent experiments with similar outcome with at least three animals per group analyzed individually. Numbers on histograms indicate percentage of CD8+ pentamer+ T cells upregulating a given marker. Contours on histograms, SPC-HA3CL4 mice; shadowed shapes, CL4 control. Mice 3–6 mo of age were used for these experiments.

This finding was further confirmed by a significant decrease in the number of CD4+CD25+Foxp3+ Tregs in the BLNs, spleen, and lungs (Fig. 8A). The depletion did not affect the circulating CD8+ T cell population, which remained CD252Foxp32 within the first week after depletion (Supplemental Fig. 8A). To exclude the possibility that the anti-CD25 Ab used to ablate nTregs may also eliminate those CD8+ T cells that may have underwent recent activation due to the release from Treg suppression and in consequence may have upregulated CD25 expression, we analyzed the dynamics between these two cell populations in more detail. For this purpose, CD25-depleted BALB/c animals were infected with the influenza A virus PR8/34 and subsequently received CFSE-labeled HA (and therefore virus-specific) CD8+ T cells. After virus-induced activation, adoptively transferred CD8+ T cells proliferated massively and acquired CD25 expression, clearly indicating that nTreg depletion by anti-CD25 treatment does not interfere with CD8+ T cell activation in the respiratory tract (Supplemental Fig. 8B–D). Four weeks after anti-CD25– mediated depletion of nTregs, we observed partially restored CD25 expression on CD4+Foxp3+ T cells in spleen and in lymph nodes (80% and 50%, respectively; data not shown) of SPCHA3CL4 and CL4 control mice. While nTreg depletion in CL4 mice did not affect the phenotype of CD8+ T cells, we observed elevated expression of the activation marker CD43 (36) on previously quiescent autoaggressive CD8+ T lymphocytes in the lungs and BLNs of nTreg-depleted SPC-HA3CL4 mice (Fig. 8B). Notably, this phenotypic alteration in CD8+ T cells was not accompanied by acquisition of effector function, because SPCHA3CL4–derived CD8+ T cells failed to produce IFN-g (Fig. 8B) or other CD8+ T cell-related effector molecules, such as perforin, granzyme B, Fas ligand (FasL), ICOS, or programmed death 1 (Fig. 8C and data not shown). There was no evidence for clonal

expansion of self-reactive CD8+ T cells upon Treg ablation, as the numbers of self-specific CD8+ cells were not increased (Fig. 8D). Moreover, unproductive activation of lung-reactive CD8+ T cells was further corroborated by the observation that Treg depletion did not exacerbate pulmonary inflammation (Supplemental Fig. 10B). Of note, analogous activation without acquisition of effector function by self-reactive CD8+ T cells also occurred in SPCHA3CL4 mice when CD4+ T cell help was provided or when local LPS-mediated immune stimulation was performed (Supplemental Fig. 9 and data not shown). Notably, transient ablation of nTregs in SPC-HA3CL4 mice resulted not only in phenotypic changes in self-reactive CD8+ T cells but also in alterations in the nTreg population itself after anti-CD25 Ab treatment of mice. In the steady state, nTregs derived from autoimmune SPC-HA3CL4 mice do not differ either in number or in phenotype from those derived from a control CL4 population (Fig. 9A, 9C and Supplemental Figs. 7B, 9A). However, 4 wk after nTreg depletion, the number of CD4+CD25+ Foxp3+ Tregs recovered from BLNs of SPC-HA3CL4 mice that expressed the CD103 effector/memory Treg marker (37) was twice as high as the number recovered from the corresponding CL4 controls (Fig. 9B). Whereas expression of other surface markers typically related to Treg phenotype and function (i.e., CD127, GITR, FR4, and Klrg-1) remained unaltered (Fig. 9 and Supplemental Fig. 10A), nTregs recovered after ablation in SPC-HA3CL4 mice showed low-level expression of CD62L (Fig. 9D). Thus, although in SPC-HA3CL4 mice the temporary depletion of nTregs did not result in uncontrolled expansion of selfreactive CD8+ T cells at the site of self-antigen exposure, it did provoke accumulation of activated CD103+ Tregs in the lymph node draining the inflamed lung tissue. Together with concomitant

6114

CD8+ T CELL TOLERANCE TO PULMONARY SELF-ANTIGEN

FIGURE 7. Self-reactive CD8+ T lymphocytes from SPC-HA3CL4 are quiescent. A, Autoreactive CD8+ T cells from lung and BLNs of SPC-HA3CL4 mice downregulate activation-related genes. Gene expression profiling (Affymetrix) was performed on CD8+ pentamer-bound lymphocytes isolated from lungs or BLNs of SPC-HA3CL4 mice (n = 15) and compared with corresponding cell populations from CL4 (n = 20) control mice (“naive”) or from CL4 mice (n = 10) infected with sublethal doses of influenza virus A/PR8/34 5 d prior to sorting (“active”). Depicted are the most strikingly downregulated or upregulated genes (values of relative signal log ratio [fold change] either lower than 24 or higher than 4) that directly refer to the effector phenotype of CD8+ T cells. Ctla, CTL Ag; Gzma, granzyme A; Gzmb, granzyme B; Icos, inducible T cell costimulator. B, Real-time RT-PCR analysis for Foxo1 expression in CD8+ pentamer+ T cells sorted from lung, BLN, and spleen of SPC-HA3CL4 (n = 15) and CL4 (n = 20) mice and in CD8+ T cells from CL4 spleen after 72-h activation with the corresponding HA Ag in vitro. Relative mRNA expression levels of Foxo1 and RPS9 (housekeeping gene as an internal control) were analyzed in duplicate real-time RT-PCR assays. Relative mRNA amounts were normalized with respect to expression levels in CL4 control (fold change = 1). Data represent one of three independent real-time RT-PCR assays performed in duplicate with similar outcome. Dark gray bars, CL4derived CD8+ pentamer+ T cells; black bars, SPC-HA3CL4–derived CD8+ pentamer+ T cells; light gray bars, CD8+ pentamer+ spleen T cells from CL4 mice after 72-h activation with specific HA Ag in vitro. C, SPC-HA3CL4–derived self-specific CD8+ T cells do not cause lung inflammation in SPC-HA mice. CD8+ T cells isolated from SPC-HA3CL4 spleens were adoptively transferred either immediately (a, b) or after in vitro differentiation with the corresponding HA Ag to cytotoxic CD8+ T cells (Tc1) (c, d) into SPC-HA or control BALB/c mice. Seven days posttransfer, paraffin-embedded sections of lungs from recipient mice were prepared and stained with H&E. a, Unaltered BALB/c lung after adoptive transfer of SPC-HA3CL4–derived CD8+ T cells. b, Unaltered SPC-HA lung after adoptive transfer of SPC-HA3CL4–derived CD8+ T cells. c, Unaffected lung tissue of BALB/c mice after transfer of Tc1 cells differentiated in vitro from CD8+ T cells isolated from SPC-HA3CL4 mice. d, Formation of perivascular lymphocytic cuffs in SPC-HA lung after transfer of Tc1 cells differentiated in vitro from SPC-HA3CL4–derived CD8+ T cells. Data depict representative results from two independent experiments with two to three mice per group treated individually. Scale bars, 100 mm (a–d); inset, 20 mm (d). Experiments were performed on 3- to 6-mo-old mice.

loss of CD62L expression, indicative for Tregs migrating from lymphoid tissue to inflamed sites, the observed accumulation of CD103+CD62Llo Tregs in BLN after nTreg ablation suggests their recently activated phenotype and readiness to suppress detrimental immune reactions at the site of autoimmune pulmonary inflammation (37–39).

Discussion Healthy persons have been shown to possess recirculating, selfreactive lymphocytes in the periphery; these lymphocytes remain harmless as long as immune regulatory mechanisms are intact (14). Although the number of studies examining the immunological processes that evoke autoimmunity has increased greatly, the issue of immune tolerance in autoimmune-prone persons is still under

intensive investigation. In this study, we used a novel mouse model that serves as a relevant experimental tool for dissecting the basic cellular mechanisms that underlie CD8+ T cell-mediated autoimmune pulmonary disorders. Our SPC-HA3CL4 mouse model shares important clinical and histological features with autoimmune-induced LIP in human patients. Histologically, the principal lesion of LIP is a diffuse T- and B-lymphocytic infiltration of the interstitium, an infiltration that expands and widens the interalveolar septa. In more progressed stages, prominent perivascular and peribronchial lymphoid follicles, consisting of central T cells with peripheral B cells, are present; these cells and patterns are also found in SPC-HA3CL4 mice. The mouse model also shares with LIP the features of compression of the airways by hyperplastic lymphoid tissue, which leads to reduced

The Journal of Immunology

6115

FIGURE 8. Foxp3+ regulatory CD4+ T cells contribute to maintenance of CD8+ T cell tolerance to pulmonary self-antigen. To deplete Foxp3+CD4+ Tregs, SPC-HA3CL4 and CL4 mice received i.v. injections of 250 mg anti-CD25 Ab (PC61). A, Seven days after the depletion, the population of CD4+ CD25+Foxp3+ T cells in the periphery had substantially decreased. Data represent outcome of two independent experiments with three animals per group analyzed individually. Groups treated with anti-CD25 Ab (+) or left untreated (2) as control groups are depicted in the legend below the graph. Gray bars, CL4 mice; black bars, SPC-HA3CL4 mice. Error bars: SD. *p , 0.05. B, Treg-depleted animals were put to death 4 wk later, and the activation status was assessed by analysis of HA-specific CD8+ T cells for CD43 expression and IFN-g production. Data show one of two independent experiments with similar outcome with at least three animals per group. Arrows indicate upregulation of CD43 activation marker after Treg depletion. C, Production of effector molecules granzyme B and perforin 4 wk after Treg depletion was determined on self-reactive CD8+ T cells by flow cytometry. Data represent one of two independent experiments with similar outcome with at least three animals per group. Contours on histograms, SPC-HA3CL4 mice after Treg depletion; shadowed shapes, untreated SPC-HA3CL4 control mice. D, Percentages of CD8+ pentamer+ T cells were analyzed by FACS in lung, BLNs, and spleen of SPC-HA3CL4 mice 4 wk after anti-CD25 Ab treatment and compared with those in untreated SPC-HA3CL4 control mice. Data show results from three to four independent experiments with two to three mice pooled within a group. Error bars: SD. Squares, SPC-HA3CL4 group after anti-CD25 Ab treatment; circles, untreated SPC-HA3CL4 control group. Experiments were performed using 4-mo-old mice.

ventilation capacity and impaired lung function, and lymphocytosis in BALF (9, 10). In this context, the basic mechanisms underlying maintenance and loss of self-tolerance will be of prime interest because it remains largely unclear why ∼50% of LIP patients experience a favorable outcome with stabilization or resolution of lung lesions, whereas in other patients the disease progresses inexorably and takes a lethal course within 5 y of diagnosis. To our surprise, in conditions of chronic pulmonary inflammation, we found immunologically quiescent self-reactive CD8+ T cells. These lung-specific CD8+ T lymphocytes did not exhibit an effector phenotype, either in the lung or in the periphery. However, sustained ignorance of the lung Ag appears rather unlikely, because mice exhibited lymphocytic infiltrates in the lung and AHR. Therefore, suspicious about the quiescent phenotype of autoaggressive CD8+ T cells, we speculated that these cells might have instead differentiated to a stage of anergy or into Tregs that could in turn contribute to the maintenance of immune tolerance. However, in contrast to the results obtained for CD8+ T cell-mediated intestinal inflammation (40) or CD4+ T lymphocyte-related pulmonary disease (41), lung-specific CD8+ T cells do not differentiate into Foxp3+ Tregs nor do they become anergic, as they proliferate massively upon antigenic stimulation in vitro. Notably, when the autoreactive lymphocytes infiltrating the lung are CD4+ T cells (23, 41, 42), their fate, function, and phenotype are quite opposite those we observed for CD8+ T cells.

These findings point to fundamental differences between the lung and the gut mucosa, not only with regard to the efficient priming of self-reactive CD8+ T cells but also with regard to the regulation of CD8+ and CD4+ lymphocyte reactivity to mucosal self-antigen. Instead of the presumed regulatory function, we demonstrated that self-specific CD8+ T cells from the spleen of SPC-HA3CL4 mice are more potent producers of IFN-g than those from healthy control animals. The disposition of CD8+ T cells to produce higher amounts of IFN-g upon antigenic stimulation has been described as an indicator of a memory T cell phenotype (28, 43). In combination with hyperresponsiveness to the Ag in vitro, which is another hallmark of memory CD8+ T cells (28), and elevated expression of memory-related surface markers (31, 44), these findings suggest that at least a portion of autoaggressive cells found in the spleen of SPC-HA3CL4 mice may have been Agexperienced and converted to a memory-like phenotype. The remainder of the cells, including self-reactive CD8+ T cells residing in direct proximity to the self-antigen, may be constantly prevented from being activated. The massive downregulation of genes related to the effector function together with preserved high Foxo1 expression in self-reactive CD8+ T cells isolated from the lungs and BLNs of SPC-HA3CL4 mice suggest a unique quiescent phenotype. We propose that the activation of autoimmune CD8+ T cells and the subsequent establishment of immune tolerance in the lungs must evolve very early in life. This suggestion is in line with our

6116

CD8+ T CELL TOLERANCE TO PULMONARY SELF-ANTIGEN

FIGURE 9. Recovered Tregs display effector/ memory phenotype in SPC-HA3CL4 mice. Expression of surface markers CD103, CD127, and GITR related to Treg function and expression of the homing marker CD62L was determined on SPC-HA3CL4–derived and CL4-derived CD25+ CD4+Foxp3+ T cells by flow cytometry. A, Levels of CD103 and CD127 on nTregs isolated from untreated SPC-HA3CL4 and CL4 mice. B, Expression of CD103 and CD127 on SPC-HA-CL4– derived and CL4-derived Tregs 4 wk after antiCD25 Ab treatment. C, Levels of CD62L and GITR on nTregs isolated from untreated SPCHA3CL4 and CL4 mice. D, Expression of CD62L and GITR on SPC-HA–derived and CL4derived Tregs 4 wk after anti-CD25 Ab treatment. Depicted are representative data from one of three independent experiments with similar outcome with at least three mice per group. Arrows indicate changes in expression of a given marker. Mice 4 mo of age were used for these experiments.

observation that, although lymphoid infiltrates are detectable in the lungs of even 7-d-old SPC-HA3CL4 newborns, in adult mice the inflammation reaches the steady state and does not progress inexorably. One might speculate that after an early effector phase, which may be the initial step for the observed lymphocytic infiltrates in the lung, autoaggressive activated CD8+ T lymphocytes are largely eliminated by a so-called “contraction” (29, 30), whereas the remaining cells develop into memory CD8+ T cells. Such memory cells can be maintained throughout the life of the host in a dynamic memory turnover (44, 45). This speculation could explain why we detected IFN-g–producing lung-specific CD8+ T cells in the spleen of SPC-HA3CL4 mice and suggests that these cells were Ag-experienced. Although this theory may explain the presence of lung inflammation in the absence of recently activated CD8+ effector T cells, we still do not understand why lung-specific CD8+ T cells can be primed and can confer effector function in early life or why this priming is efficiently prevented in the adult organism. One possible mechanism contributing to the maintenance of immune tolerance to self-antigen in the lung may be Foxp3+ Tregmediated control of self-reactive CD8+ T cells. Numerous reports have discussed the contribution of naturally occurring Tregs to the immune control of CD4+ T lymphocytes in pulmonary disorders (19, 42). The absence of nTregs has been found to provoke the random development of autoimmune diseases in athymic mice (46) in vivo. Importantly, in contrast to naive CD4+ and CD8+ T cells, nTregs have been suggested to develop later in the thymus, beginning 3 d after birth (47); therefore, the appearance of thymus-derived Tregs in the circulation is delayed. Hence, there may be a gap in immune regulation that is sufficient to allow priming of self-reactive CD8+ T cells and the resulting onset of immunopathology in the lung of newborn SPC-HA3CL4 mice. Why inflammation starts to develop only shortly after birth and not already during pregnancy remains elusive. Notably, it takes about 1 wk before signs of inflammation become evident in newborn mice, which matches the period of time required to establish T cell effector function after Ag recognition. Due to the lack of antigenic stimulation in utero, the adaptive immune system of the perinatal organism is inexperienced, and prenatal activation of self-reactive T cells may be prevented due to the lack of suf-

ficient costimulation provided by APCs. After birth, T lymphocytes normally show a massive expansion in neonates, probably due to the overwhelming antigenic stimulation from the environment (48, 49). The increase of environmental stimuli (“danger signals”) may result in maturation of professional APCs, which may lead to improved priming of self-reactive T cells in the respiratory tract at a time where the Treg pool is still insufficiently filled. To dissect a possible interplay between nTregs and self-reactive CD8+ T cells in pulmonary autoimmune-mediated inflammation, we eliminated nTregs by injection of anti-CD25 Ab (34, 35). CD25 (IL-2R) is required for nTreg function and survival (47, 50), and the anti-CD25 mAb (clone PC61) used in our study was reported to block IL-2 binding to the IL-2R (51). Injection of antiCD25 may thus lead to loss of nTreg function or death by cytokine deprivation. Several reports (34, 52) propose that nTreg depletion using the PC61 Ab application may also result from nTreg elimination by phagocytic cells in a process of Ab-dependent cellmediated cytotoxicity after recognition of Fc Ab fragments bound on the nTreg surface. Although ∼50–60% of nTregs are depleted, the remaining Tregs substantially downregulate CD25, which impairs their function and activation as demonstrated in accelerated onset of immunity in experimental autoimmune encephalomyelitis (34). Thus, we assume that regardless of the underlying mechanism, the use of anti-CD25 Ab results in transient impairment of nTreg function and represents an appropriate tool to investigate CD8+ T cell function in the absence of functional nTregs. The in vivo elimination of nTregs from adult SPC-HA3CL4 mice triggered a phenotypic change of self-specific CD8+ T cells in BLNs and lungs. Nevertheless, despite elevated expression of the CD43 activation marker, these autoaggressive CD8+ T lymphocytes failed to produce proinflammatory cytokines or cytotoxic molecules. CD8+ T cell numbers in SPC-HA3CL4 mice remained unaffected after Treg depletion, and extensive histological examination did not reveal any signs of autoimmune disease exacerbation in the lung. Thus, we propose that immune regulatory mechanisms established in mature SPC-HA3CL4 mice efficiently prevent undesired expansion of the numbers of autoreactive CD8+ T cells and thereby prevent exacerbation of the disease that was initiated very early in life.

The Journal of Immunology Beyond the putative contribution of nTregs to immune tolerance in SPC-HA3CL4 mice, more complex mechanisms obviously exist that may work together to limit CD8+ T cell reactivity to pulmonary self-antigen. On the basis of our findings, we suggest that, instead of one given suppressor cell population, there may be additional cellular or molecular mechanisms that exert compatible suppressive functions against pulmonary autoimmunity. One of the possible candidates may be tolerogenic dendritic cells that can prevent the activation of autoaggressive lymphocytes either directly or by interacting with Tregs (53, 54). However, in vitro assays of the stimulating capability of APCs derived from BLNs of SPC-HA3CL4 mice did not demonstrate that the function of these cells was different from that of APCs derived from healthy controls (data not shown). The suppressive function of nTregs may also be supported by other cell populations that can acquire immunomodulatory function. Such activity has been described for mast cells, which are indispensable in transplant tolerance (55), and for mesenchymal stromal cells, which display complex effects on several immune cell types, such as dendritic cells, Tregs, and effector cells (56). The suppressive effects of these cells are usually exerted via the secretion of anti-inflammatory cytokines (56, 57). However, screening on the protein level (FACS) and on the transcription level (RT-PCR) showed that the production of anti-inflammatory cytokines (IL-10 and TGF-b) in the lungs and BLNs of SPC-HA3CL4 mice was not higher than that in healthy controls. Because also self-reactive CD8+ T cells did not exhibit any signs for immune suppressive function (i.e., production of anti-inflammatory cytokines and/or expression of Treg-related markers), we largely exclude the possibility that these cells may be actively involved in the control of autoimmune-mediated lung inflammation. We are, however, aware of the fact that the list of experiments performed to prove this is not exhaustive. In this context, it is interesting to note that in rheumatoid arthritis, IL-16– producing regulatory CD8+ T cells have been shown to accumulate in the synovial lesions. Strikingly, adoptive transfer of these IL-16+ CD8+ T cells dramatically decreased the production of inflammatory cytokines in the tissue. In line with this, treatment of animals suffering from rheumatoid arthritis with recombinant IL16 resulted in almost complete inhibition of IFN-g and TNF-a production in the inflammatory lesions (58). Although IL-16 expression was not found to be upregulated in CD8+ T cells from the lung of SPC-HA3CL4 mice, it may be worth analyzing the cells in more detail for the production of this particular cytokine in future studies, especially in light of the fact that IL-16 has been described to play an immunomodulatory role in allergic airway inflammation (59). Because we and others have reported that AECII cells are bona fide autoantigen-expressing cells with immune-modulating capability (41, 60), we speculated that these cells may also contribute to immune tolerance in SPC-HA3CL4 mice. Thus, we have tested the inhibitory potential of these cells to CD8+ T lymphocytes in vitro. Nevertheless, in contrast to CD4+ T cells, AECII cells from SPC-HA or SPC-HA3CL4 mice endogenously expressing the HA self-antigen failed to activate HA-specific CD8+ T cells in vitro (data not shown). In general, for technical reasons, most of the functional tests for putative suppressor cell populations that may keep in check self-reactive CD8+ T cells in SPC-HA3CL4 mice could be conducted only in vitro, whereas the immune regulation processes in the lung and in the lung-draining lymph nodes probably involve a complex and difficult-to-dissect network consisting of multiple synergizing mechanisms. Our in vivo data clearly demonstrate the partial contribution of nTregs to pulmonary tolerance against self-reactive CD8+ T cells. The fact that nTregs do not have to be the exclusive protective

6117 factor against autoimmune disease was previously demonstrated in autoimmunity-prone NOD mice, which surprisingly generate nTregs more effectively than do healthy mice (61). Notably, in our autoimmune SPC-HA3CL4 mice, nTregs in the steady state do not differ in number or in phenotype from control nTregs in CL4 mice. However, 4 wk after in vivo nTreg depletion, the number of CD4+CD25+Foxp3+ T cells isolated from BLNs of SPC-HA3CL4 mice that expressed CD103 was twice as high as the number of such cells from the corresponding CL4 controls. Additionally, this effect was accompanied by downregulation of CD62L on SPC-HA3CL4–derived nTregs in BLN and spleen. Of note, elevated expression of CD103 (integrin aEb7) together with loss of CD62L (L-selectin) expression have been previously demonstrated to be indicators of highly potent regulatory CD4+ T cells, and CD103 has been proposed as a marker of activated or memory “inflammation-seeking Tregs” (37–39). In experimental disease models, CD103-positive CD25+CD4+Foxp3+ Tregs have been shown to be the most potent suppressors because of their high suppressive capability and their enhanced specific migration into inflamed areas (37). Elevated CD103 expression on nTregs 4 wk after depletion in autoimmune SPC-HA3CL4 mice may thus indicate rapid activation, migration, and accumulation of new nTregs in the lymph node draining the inflamed lung. This action may prevent uncontrolled local expansion in the numbers of lungspecific CD8+ T cells and, thus, may prevent exacerbation of the disease. The absence of fully functional CD8+ T cells upon Treg depletion might be the outcome of an adaptive phenotypic switch in the nTreg population, a switch that acts in concert with other supporting but as yet undefined peripheral tolerance mechanisms.

Acknowledgments We thank Beata Zygmunt for scientific discussions, Lothar Gro¨be for cell sorting, and Sarah Herzog, Birthe Ellinghusen, and Elena Reinhard for expert technical assistance.

Disclosures The authors have no financial conflicts of interest.

References 1. Liblau, R. S., F. S. Wong, L. T. Mars, and P. Santamaria. 2002. Autoreactive CD8 T cells in organ-specific autoimmunity: emerging targets for therapeutic intervention. Immunity 17: 1–6. 2. Walter, U., and P. Santamaria. 2005. CD8+ T cells in autoimmunity. Curr. Opin. Immunol. 17: 624–631. 3. Costabel, U., and J. Guzman. 2001. Bronchoalveolar lavage in interstitial lung disease. Curr. Opin. Pulm. Med. 7: 255–261. 4. Newman, L. S., C. S. Rose, and L. A. Maier. 1997. Sarcoidosis. N. Engl. J. Med. 336: 1224–1234. 5. Prasse, A., C. G. Georges, H. Biller, H. Hamm, H. Matthys, W. Luttmann, and J. C. Virchow, Jr. 2000. Th1 cytokine pattern in sarcoidosis is expressed by bronchoalveolar CD4+ and CD8+ T cells. Clin. Exp. Immunol. 122: 241–248. 6. Wahlstro¨m, J., K. Katchar, H. Wigzell, O. Olerup, A. Eklund, and J. Grunewald. 2001. Analysis of intracellular cytokines in CD4+ and CD8+ lung and blood T cells in sarcoidosis. Am. J. Respir. Crit. Care Med. 163: 115–121. 7. Enelow, R. I., A. Z. Mohammed, M. H. Stoler, A. N. Liu, J. S. Young, Y. H. Lou, and T. J. Braciale. 1998. Structural and functional consequences of alveolar cell recognition by CD8(+) T lymphocytes in experimental lung disease. J. Clin. Invest. 102: 1653–1661. 8. Borchers, M. T., S. C. Wesselkamper, V. Curull, A. Ramirez-Sarmiento, A. Sanchez-Font, J. Garcia-Aymerich, C. Coronell, J. Lloreta, A. G. Agusti, J. Gea, et al. 2009. Sustained CTL activation by murine pulmonary epithelial cells promotes the development of COPD-like disease. J. Clin. Invest. 119: 636– 649. 9. Cha, S. I., M. B. Fessler, C. D. Cool, M. I. Schwarz, and K. K. Brown. 2006. Lymphoid interstitial pneumonia: clinical features, associations and prognosis. Eur. Respir. J. 28: 364–369. 10. Swigris, J. J., G. J. Berry, T. A. Raffin, and W. G. Kuschner. 2002. Lymphoid interstitial pneumonia: a narrative review. Chest 122: 2150–2164. 11. Fitzpatrick, E. A., M. Avdiushko, A. M. Kaplan, and D. A. Cohen. 1999. Role of T cell subsets in the development of AIDS-associated interstitial pneumonitis in mice. Exp. Lung Res. 25: 671–687.

6118 12. Lairmore, M. D., J. M. Poulson, T. A. Adducci, and J. C. DeMartini. 1988. Lentivirus-induced lymphoproliferative disease. Comparative pathogenicity of phenotypically distinct ovine lentivirus strains. Am. J. Pathol. 130: 80–90. 13. Sakaguchi, S., N. Sakaguchi, M. Asano, M. Itoh, and M. Toda. 1995. Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor alpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J. Immunol. 155: 1151–1164. 14. Sakaguchi, S. 2004. Naturally arising CD4+ regulatory t cells for immunologic self-tolerance and negative control of immune responses. Annu. Rev. Immunol. 22: 531–562. 15. Sakaguchi, S. 2005. Naturally arising Foxp3-expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self. Nat. Immunol. 6: 345– 352. 16. Zwar, T. D., I. R. van Driel, and P. A. Gleeson. 2006. Guarding the immune system: suppression of autoimmunity by CD4+CD25+ immunoregulatory T cells. Immunol. Cell Biol. 84: 487–501. 17. Barcelo´, B., J. Pons, J. M. Ferrer, J. Sauleda, A. Fuster, and A. G. Agustı´. 2008. Phenotypic characterisation of T-lymphocytes in COPD: abnormal CD4+CD25+ regulatory T-lymphocyte response to tobacco smoking. Eur. Respir. J. 31: 555– 562. 18. Kearley, J., D. S. Robinson, and C. M. Lloyd. 2008. CD4+CD25+ regulatory T cells reverse established allergic airway inflammation and prevent airway remodeling. J. Allergy Clin. Immunol. 122: 617–624, e6. 19. Hartl, D., B. Koller, A. T. Mehlhorn, D. Reinhardt, T. Nicolai, D. J. Schendel, M. Griese, and S. Krauss-Etschmann. 2007. Quantitative and functional impairment of pulmonary CD4+CD25hi regulatory T cells in pediatric asthma. J. Allergy Clin. Immunol. 119: 1258–1266. 20. Isajevs, S., I. Taivans, G. Strazda, U. Kopeika, M. Bukovskis, V. Gordjusina, and A. Kratovska. 2009. Decreased FOXP3 expression in small airways of smokers with COPD. Eur. Respir. J. 33: 61–67. 21. Morgan, D. J., R. Liblau, B. Scott, S. Fleck, H. O. McDevitt, N. Sarvetnick, D. Lo, and L. A. Sherman. 1996. CD8(+) T cell-mediated spontaneous diabetes in neonatal mice. J. Immunol. 157: 978–983. 22. Kirberg, J., A. Baron, S. Jakob, A. Rolink, K. Karjalainen, and H. von Boehmer. 1994. Thymic selection of CD8+ single positive cells with a class II major histocompatibility complex-restricted receptor. J. Exp. Med. 180: 25–34. 23. Bruder, D., A. M. Westendorf, R. Geffers, A. D. Gruber, M. Gereke, R. I. Enelow, and J. Buer. 2004. CD4 T Lymphocyte-mediated lung disease: steady state between pathological and tolerogenic immune reactions. Am. J. Respir. Crit. Care Med. 170: 1145–1152. 24. Glaab, T., W. Mitzner, A. Braun, H. Ernst, R. Korolewitz, J. M. Hohlfeld, N. Krug, and H. G. Hoymann. 2004. Repetitive measurements of pulmonary mechanics to inhaled cholinergic challenge in spontaneously breathing mice. J. Appl. Physiol. 97: 1104–1111. 25. Hoymann, H. G. 2006. New developments in lung function measurements in rodents. Exp. Toxicol. Pathol. 57(Suppl 2): 5–11. 26. Chiavolini, D., J. Rangel-Moreno, G. Berg, K. Christian, L. OliveiraNascimento, S. Weir, J. Alroy, T. D. Randall, and L. M. Wetzler. 2010. Bronchus-associated lymphoid tissue (BALT) and survival in a vaccine mouse model of tularemia. PLoS ONE 5: e11156. 27. Takahama, Y., E. W. Shores, and A. Singer. 1992. Negative selection of precursor thymocytes before their differentiation into CD4+CD8+ cells. Science 258: 653– 656. 28. Cho, B. K., C. Wang, S. Sugawa, H. N. Eisen, and J. Chen. 1999. Functional differences between memory and naive CD8 T cells. Proc. Natl. Acad. Sci. USA 96: 2976–2981. 29. Badovinac, V. P., and J. T. Harty. 2006. Programming, demarcating, and manipulating CD8+ T-cell memory. Immunol. Rev. 211: 67–80. 30. Harty, J. T., and V. P. Badovinac. 2008. Shaping and reshaping CD8+ T-cell memory. Nat. Rev. Immunol. 8: 107–119. 31. Pihlgren, M., C. Arpin, T. Walzer, M. Tomkowiak, A. Thomas, J. Marvel, and P. M. Dubois. 1999. Memory CD44(int) CD8 T cells show increased proliferative responses and IFN-gamma production following antigenic challenge in vitro. Int. Immunol. 11: 699–706. 32. Kerdiles, Y. M., D. R. Beisner, R. Tinoco, A. S. Dejean, D. H. Castrillon, R. A. DePinho, and S. M. Hedrick. 2009. Foxo1 links homing and survival of naive T cells by regulating L-selectin, CCR7 and interleukin 7 receptor. Nat. Immunol. 10: 176–184. 33. Ouyang, W., O. Beckett, Q. Ma, J. H. Paik, R. A. DePinho, and M. O. Li. 2010. Foxo proteins cooperatively control the differentiation of Foxp3+ regulatory T cells. Nat. Immunol. 11: 618–627. 34. McNeill, A., E. Spittle, and B. T. Ba¨ckstro¨m. 2007. Partial depletion of CD69low-expressing natural regulatory T cells with the anti-CD25 monoclonal antibody PC61. Scand. J. Immunol. 65: 63–69. 35. Onizuka, S., I. Tawara, J. Shimizu, S. Sakaguchi, T. Fujita, and E. Nakayama. 1999. Tumor rejection by in vivo administration of anti-CD25 (interleukin-2 receptor alpha) monoclonal antibody. Cancer Res. 59: 3128–3133. 36. Ellies, L. G., A. T. Jones, M. J. Williams, and H. J. Ziltener. 1994. Differential regulation of CD43 glycoforms on CD4+ and CD8+ T lymphocytes in graftversus-host disease. Glycobiology 4: 885–893.

CD8+ T CELL TOLERANCE TO PULMONARY SELF-ANTIGEN 37. Huehn, J., K. Siegmund, J. C. Lehmann, C. Siewert, U. Haubold, M. Feuerer, G. F. Debes, J. Lauber, O. Frey, G. K. Przybylski, et al. 2004. Developmental stage, phenotype, and migration distinguish naive- and effector/memory-like CD4+ regulatory T cells. J. Exp. Med. 199: 303–313. 38. Huehn, J., and A. Hamann. 2005. Homing to suppress: address codes for Treg migration. Trends Immunol. 26: 632–636. 39. Zhao, D., C. Zhang, T. Yi, C. L. Lin, I. Todorov, F. Kandeel, S. Forman, and D. Zeng. 2008. In vivo-activated CD103+CD4+ regulatory T cells ameliorate ongoing chronic graft-versus-host disease. Blood 112: 2129–2138. 40. Westendorf, A. M., D. Fleissner, S. Deppenmeier, A. D. Gruber, D. Bruder, W. Hansen, R. Liblau, and J. Buer. 2006. Autoimmune-mediated intestinal inflammation-impact and regulation of antigen-specific CD8+ T cells. Gastroenterology 131: 510–524. 41. Gereke, M., S. Jung, J. Buer, and D. Bruder. 2009. Alveolar type II epithelial cells present antigen to CD4(+) T cells and induce Foxp3(+) regulatory T cells. Am. J. Respir. Crit. Care Med. 179: 344–355. 42. Hadeiba, H., and R. M. Locksley. 2003. Lung CD25 CD4 regulatory T cells suppress type 2 immune responses but not bronchial hyperreactivity. J. Immunol. 170: 5502–5510. 43. Kreuwel, H. T., S. Aung, C. Silao, and L. A. Sherman. 2002. Memory CD8(+) T cells undergo peripheral tolerance. Immunity 17: 73–81. 44. Prlic, M., L. Lefrancois, and S. C. Jameson. 2002. Multiple choices: regulation of memory CD8 T cell generation and homeostasis by interleukin (IL)-7 and IL15. J. Exp. Med. 195: F49–F52. 45. Homann, D., L. Teyton, and M. B. Oldstone. 2001. Differential regulation of antiviral T-cell immunity results in stable CD8+ but declining CD4+ T-cell memory. Nat. Med. 7: 913–919. 46. Sakaguchi, S., M. Toda, M. Asano, M. Itoh, S. S. Morse, and N. Sakaguchi. 1996. T cell-mediated maintenance of natural self-tolerance: its breakdown as a possible cause of various autoimmune diseases. J. Autoimmun. 9: 211–220. 47. Fontenot, J. D., and A. Y. Rudensky. 2005. A well adapted regulatory contrivance: regulatory T cell development and the forkhead family transcription factor Foxp3. Nat. Immunol. 6: 331–337. 48. Comans-Bitter, W. M., R. de Groot, R. van den Beemd, H. J. Neijens, W. C. Hop, K. Groeneveld, H. Hooijkaas, and J. J. van Dongen. 1997. Immunophenotyping of blood lymphocytes in childhood. Reference values for lymphocyte subpopulations. J. Pediatr. 130: 388–393. 49. Hannet, I., F. Erkeller-Yuksel, P. Lydyard, V. Deneys, and M. DeBruyere. 1992. Developmental and maturational changes in human blood lymphocyte subpopulations. Immunol. Today 13: 215, 218. 50. Furtado, G. C., M. A. Curotto de Lafaille, N. Kutchukhidze, and J. J. Lafaille. 2002. Interleukin 2 signaling is required for CD4(+) regulatory T cell function. J. Exp. Med. 196: 851–857. 51. Lowenthal, J. W., P. Corthe´sy, C. Tougne, R. Lees, H. R. MacDonald, and M. Nabholz. 1985. High and low affinity IL 2 receptors: analysis by IL 2 dissociation rate and reactivity with monoclonal anti-receptor antibody PC61. J. Immunol. 135: 3988–3994. 52. Couper, K. N., D. G. Blount, J. B. de Souza, I. Suffia, Y. Belkaid, and E. M. Riley. 2007. Incomplete depletion and rapid regeneration of Foxp3+ regulatory T cells following anti-CD25 treatment in malaria-infected mice. J. Immunol. 178: 4136–4146. 53. Mahnke, K., E. Schmitt, L. Bonifaz, A. H. Enk, and H. Jonuleit. 2002. Immature, but not inactive: the tolerogenic function of immature dendritic cells. Immunol. Cell Biol. 80: 477–483. 54. Tang, Q., J. Y. Adams, A. J. Tooley, M. Bi, B. T. Fife, P. Serra, P. Santamaria, R. M. Locksley, M. F. Krummel, and J. A. Bluestone. 2006. Visualizing regulatory T cell control of autoimmune responses in nonobese diabetic mice. Nat. Immunol. 7: 83–92. 55. de Vries, V. C., A. Wasiuk, K. A. Bennett, M. J. Benson, R. Elgueta, T. J. Waldschmidt, and R. J. Noelle. 2009. Mast cell degranulation breaks peripheral tolerance. Am. J. Transplant. 9: 2270–2280. 56. Nauta, A. J., and W. E. Fibbe. 2007. Immunomodulatory properties of mesenchymal stromal cells. Blood 110: 3499–3506. 57. Grimbaldeston, M. A., S. Nakae, J. Kalesnikoff, M. Tsai, and S. J. Galli. 2007. Mast cell-derived interleukin 10 limits skin pathology in contact dermatitis and chronic irradiation with ultraviolet B. Nat. Immunol. 8: 1095–1104. 58. Klimiuk, P. A., J. J. Goronzy, and C. M. Weyand. 1999. IL-16 as an antiinflammatory cytokine in rheumatoid synovitis. J. Immunol. 162: 4293–4299. 59. Little, F. F., J. de Bie, A. van Oosterhout, H. Kornfeld, D. M. Center, and W. W. Cruikshank. 2003. Immunomodulatory effect of interleukin-16 on allergic airway inflammation. Chest 123(3, Suppl): 431S–432S. 60. Gereke, M., L. Gro¨be, S. Prettin, M. Kasper, S. Deppenmeier, A. D. Gruber, R. I. Enelow, J. Buer, and D. Bruder. 2007. Phenotypic alterations in type II alveolar epithelial cells in CD4+ T cell mediated lung inflammation. Respir. Res. 8: 47. 61. Feuerer, M., W. Jiang, P. D. Holler, A. Satpathy, C. Campbell, M. Bogue, D. Mathis, and C. Benoist. 2007. Enhanced thymic selection of FoxP3+ regulatory T cells in the NOD mouse model of autoimmune diabetes. Proc. Natl. Acad. Sci. USA 104: 18181–18186.