Activation-Induced Cell Death CD137-Induced CD8 T Cell Dendritic ...

The Journal of Immunology

Dendritic Cells and Stat3 Are Essential for CD137-Induced CD8 T Cell Activation-Induced Cell Death Benyue Zhang,* Yuanyuan Zhang,* Liguo Niu,*,1 Anthony T. Vella,† and Robert S. Mittler*,‡ Agonistic anti-CD137 mAbs either positively or negatively regulate T cell function. When administered at the beginning of lymphocytic choriomeningitis virus Armstrong infection anti-CD137 induced immunosuppression and T cell deletion, and in the case of influenza infection led to increased mortality. In contrast, 72 h delay in anti-CD137 treatment led to an enhanced virus-specific CD8 T cell response and rapid viral clearance. Virus-specific CD8 T cells in anti-CD137–injected mice rapidly upregulate Fas expression, and although necessary, was insufficient to induce CD8 T cell deletion. Strikingly, CD137 signaling in T cells was found to be insufficient to induce suppression or deletion. Rather, immunosuppression and T cell deletion was only observed if CD137 signals were provided to T cells and dendritic cells (DCs). In vitro CD137 crosslinking in DCs led to phosphorylation of Stat3, and importantly, anti-CD137 treatment of lymphocytic choriomeningitis virus Armstrong infected Stat3 conditional knock-out mice induced neither immune suppression or T cell deletion. Taken together, these data suggest that CD137 signaling in DCs can regulate CD8 T cell survival through a Stat3 and Fas-mediated pathway. The Journal of Immunology, 2010, 184: 4770–4778.

T

he activation inducible member of the TNFR superfamily, CD137 (4-1BB/tnfrsf9), is expressed by most hematopoietic cell lineages (1), serves as a T cell costimulatory molecule (2), and represents a potential clinical target for treatment of cancer and autoimmune disease (3–5). The ligand for CD137, an inducible type II transmembrane protein belonging to TNF superfamily is expressed or upregulated on activated professional APCs (6, 7). Anti-CD137 mAbs can preferentially activate CD8 T cells in vitro and in vivo (8) and CD137 signaling functions as a T cell survival factor (9–13) leading to potent antitumor immunity and enhanced or prolonged CTL function (14). In contrast to CD8 T cell activation, anti-CD137 treatment can induce suppression of T-dependent humoral immunity (15), and reverse established autoimmune disease (16–18). In an earlier study, we found that anti-CD137 mAbs induced activation induced cell death (AICD) or costimulation in CD4 or CD8 T cells and that the outcome was dependent on the timing of anti-CD137 injection (4). Given within 48 h of infection, anti-CD137 induced T cell deletion, immune tolerance, and persistent viral infection in lymphocytic choriomeningitis virus (LCMV) infected mice; and in A/PR8/34 influenza virus infected mice, led to high levels of mortality. In contrast, when anti-CD137 treatment was delayed by 72 h postinfection antiviral T cell immunity was enhanced (4).

In this report, we show that the majority of endogenous virusspecific CD8 T cells and adoptively transferred LCMV-specific TCR transgenic P14 CD8 T cells (referred to as “P14 T cells”) undergo deletion after early anti-CD137 mAb treatment of LCMV-infected mice and that this coincides with marked Fas upregulation on virusactivated T cells. Despite the induction of deletion-inducing gene programs, P14 T cells undergo measurable expansion before undergoing progressive exhaustion that culminates in their deletion, a process that becomes evident around 21 d postinfection and complete within 90 d postinfection. Most surprising was the observation that T cell loss of function and deletion was not observed after adoptive transfer of CD137 sufficient wild-type (WT) or P14 T cells into CD137-deficient mice prior to infection and anti-CD137 treatment. This observation suggested that in addition to, or independent of CD137 signaling in T cells, another CD137-sufficient cell lineage was needed to induce suppression and T cell deletion. Sequential adoptive cell transfer experiments in which P14 T cells were cotransferred with WT CD137-sufficient hematopoietic cell lineages showed that in addition to T cells, CD137-sufficient dendritic cells (DCs) were essential for immune suppression of antiviral immunity. Moreover, immune suppression mediated by DCs was found to be CD137-induced Stat3 activation-dependent.

Materials and Methods ‡

*Emory Vaccine Center and Department of Surgery, Emory University School of Medicine, Atlanta, GA 30329; and †Department of Immunology, University of Connecticut Health Center, Farmington, CT 06030

Mice

Address correspondence and reprint requests to Dr. Robert S. Mittler, Emory University School of Medicine, Yerkes National Primate Research Center and Emory Vaccine Center, 954 Gatewood Road, Atlanta, GA 30329. E-mail address: rmittler@ rmy.emory.edu

The 8- to 12-wk-old female C57BL/6 mice, CD45.1 congenic BL/6 mice, CD42/2 BL/6 mice, CD82/2 BL/6 mice, RAG-12/2 BL/6 mice, and FasL2/2 BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The BL/6 Thy 1.1 congenic P14 BL/6 TCR transgenic mice specific for the GP33-41 epitope of LCMV were a gift from Dr. Rafi Ahmed. The BL/6 CD1372/2 mice, a gift of Dr. Byoung Kwon (19). Stat3 conditional knock-out (KO) mice were kindly provided by Dr. Eduardo Sotomayor (19). All mice were bred and/or housed at the Yerkes National Primate Research Center Vivarium according to Emory University Institutional Animal Care and Use Committee guidelines.

The online version of this article contains supplemental material.

Virus infection and in vivo Ab treatment

Abbreviations used in this paper: AICD, activation induced cell death; DC, dendritic cell; FasL, Fas ligand; HDAC, histone deacetylase; IDO, indoleammine 2, 3 deoxygenase; KO, knockout; LCMV, lymphocytic choriomeningitis virus; Mø, macrophage; R-IgG, rat IgG; sFasL, soluble Fas ligand; Treg, T regulatory cell; WT, wild-type.

BL/6 mice were infected by i.p. inoculation with 2 3 105 PFU LCMV Armstrong CA 1371 strain, as described previously (4). LCMV-infected WT and CD1372/2 BL/6 mice were injected i.p. with 200 mg rat anti-mouse CD137 mAb, clone 3H3, a rat IgG2a (20), or 200 mg rat IgG in 500 ml PBS at 24 h postinfection.

1

Current address: Department of Microbiology and Immunology, University of Miami School of Medicine, Miami, FL. Received for publication August 18, 2009. Accepted for publication March 4, 2010. This work was supported by National Institutes of Health Grant AI059290 (to R.S.M.).

Copyright Ó 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.0902713

The Journal of Immunology

4771

Cell enrichment and adoptive transfer

Intracellular staining for IFN-g and TNF-a

Thy1.1 TCR transgenic P14 T cells were isolated by two-step negative selection from the spleens of P14 TCR transgenic mice using mouse pan T cells isolation kit first and then mouse CD8+ T cells isolation kit (Miltenyi Biotec, Auburn, CA). The 1 3 104 P14 T cells with purity .97% were injected i.v. into WT or CD1372/2 BL/6 mice. In some experiments, P14 T cells were cotransferred with RAG-12/2 whole spleen cells, MACSpositively (CD11c) selected splenic DCs, CD11b+ macrophages isolated from MACS column CD3 and CD20-depleted spleen cells following the manufacturer’s instructions, and NK1.1+ NK cells or FACS sorted DC subsets into the above-mentioned mice. In some experiments, to increase the DC number, BL/6 mice or CD1372/2 BL/6 mice were injected i.p. with 30 mg Flt-3 ligand for 9 d prior to euthanasia on day 10.

IFN-g staining has been described previously (4). Spleen cells were stimulated in vitro with medium, GP33-41 or NP396-404 at 1 mg/ml for 5 h in vitro at the presence of GolgiPlug (BD Biosciences). Cells were surface stained with PerCP-conjugated anti-CD8a; PE-Cy7 conjugated anti-Thy1.2, Pacific blue-conjugated anti-Thy1.1, and then stained for intracellular IFN-g with APC-conjugated anti–IFN-g according to the manufacturer’s recommended protocol. In some experiments, DCs were stained with PE-Cy7–conjuaged CD11c and intracellular TNF-a were stained with PE-conjugated monoclonal hamster anti-mouse TNF-a. Fluorochrome-matched isotype controls for the previously described Abs or mAbs were also used for intracellular staining.

Cell staining, flow cytometry, and data processing

H-2Db tetramers loaded with LCMV peptide GP33-41 or NP396-404 were generously provided by Dr. John Altman and the National Institutes of Health Tetramer Core Facility at Emory University and used as described previously (4). Spleen cells were stained with APC-conjugated MHC class I GP33-41 or NP396-404 peptide-loaded tetramers, fluorochromeconjugated monoclonal rat anti-mouse CD8a and analyzed by multicolor flow cytometry.

Mice were euthanized at the indicated times, and spleens were harvested, single-cell suspensions were prepared by mincing, and passage through a 70-mm nylon screen (BD Biosciences, San Jose, CA), washed twice by low-speed centrifugation in cold PBS, and blocked with 10 mg/ml antiCD16/anti-CD32 (clone 2.4G2 ATCC) for 10 min in FACS-buffer (PBS supplemented with 0.5% BSA and 0.04% sodium azide). Aliquots of cells (1 3 106) were suspended in 0.1 ml BD FACS-buffer and incubated on ice for 20 min with FITC-, PE-, PerCP-, APC-, PE-Cy7-, Alexa Fluor 700-, and Pacific blue-conjugated mAbs as indicated to detect the following surface Ags: CD3, CD4, CD8a, CD11c, CD11b, B220, PD-1 (BD Biosciences), CD90.1 (Thy1.1), CD90.2 (Thy1.2), Fas, Fas ligand (FasL) (eBioscience, San Diego, CA), TNFR1 (Biolegend, San Diego, CA), and TNFR2 (Serotec, Oxford, U.K.). Expression of CD137 was measured with FITC-conjugated anti-CD137 mAb (clone 3H3) produced in our laboratory (20). Annexin V and propidium iodide staining was performed according to the manufacturer’s protocol (Molecular Probes/Invitrogen,San Diego, CA). Stained cells were analyzed on a BD LSR II flow cytometer. Data analysis was carried using FlowJo software (TreeStar, Ashland, OR).

FIGURE 1. CD137 signaling in CD8 T cells is insufficient to induce CD8 T cell AICD. MACS purified P14 T cells (1 3 104) were adoptively transferred to C57BL/6 WT or CD1372/2 BL/6 mice. The mice were infected with LCMV and injected with antiCD137 (a-CD137) or rat IgG (R-IgG) at day 1 postinfection. On day 8 postinfection, spleen cells were phenotyped by FACS for P14 CD8 T cell frequency (A), absolute number of viable P14 T cells (B), and absolute number of endogenous GP33–41-tetramer+ cells (C).

MHC class I tetramer staining

ELISA assays Sera were collected as previous described (4). Serum soluble Fas ligand (sFasL) measurements were performed using ELISA kits from R&D Systems (Minneapolis, MN), following the manufacturer’s instructions.

FACS sorting of DC subsets CD82, CD8+, and B220+ plasmacytoid DCs were purified by FACS sorting on an 11-color BD Aria Fluorescence Activated Cell Sorter by members of the Emory Vaccine Center Flow Cytometry Core Facility. Whole spleen cells were prepared with cell sorting buffer (1% FCS, 1 mM EDTA), blocked with anti-CD16/32 (2.4G2) and stained with fluorochrome-

4772

CD137-MEDIATED T CELL AICD DURING VIRAL INFECTION

conjugated CD11c, CD8, and B220. Stained cells were then FACS sorted. DC subsets with .97% purity were used to inject mice i.v.

Stat3 phosphorylation in DCs DCs were positively selected by CD11c expression from T cell and B celldepleted spleen cells obtained from Flt3-L–injected mice. DCs were incubated at 2 3 106 cells/100 ml D-PBS containing rat IgG or anti-CD137 at 10 mg/ml in Eppendorf centrifuge tubes set in a 4˚C ice water bath for 10 min, pelleted by centrifugation at 4˚C, and resuspended in 100 ml D-PBS at 37˚C containing 5 mg/ml sheep anti-rat IgG. Cells were incubated at 37˚C for the indicated times. The cells were pelleted by centrifugation for 15 s at 14,000 rpm in a refrigerated Eppendorf centrifuge, decanted, and resuspended in RIPA lysis buffer and stored until cellular proteins were separated by SDS-PAGE under reducing conditions. Proteins were electrotransferred onto nitrocellulose membranes and Western blotted with anti-Stat3 and anti–phospho-tyrosine (Y705)-Stat3 (R&D Systems).

Results CD137 expression on virus-specific CD8 T cells is insufficient to induce AICD Implicit in published studies of CD137-regulated T cell function is the notion that CD137 signaling in T cells is solely responsible for how the T cell responds. In this study, we attempted to confirm this concept by showing that anti-CD137 mAbs induced immune suppression in LCMV-infected CD1372/2 mice that had received CD137-sufficent P14 (or BL/6 CD8) T cells prior to infection. Surprisingly, rather than being suppressed, and undergoing deletion, P14 T cells (and BL/6 CD8 T cells, data not shown) underwent enhanced proliferation and expansion (Fig. 1A, 1B), whereas adoptive transfer of P14 T cells into WT recipients (Fig. 1A, 1B), or Rag2/ 2 mice (data not shown) led to their deletion within 8 d of infection. As expected, endogenous CD1372/2 T cells were unaffected by antiCD137 treatment (Fig. 1C); and this was also true when CD1372/2 T cells were transferred into BL/6 or Rag2/2 recipients (4). CD137 expression by CD8 T cells and DCs is necessary for induction of CD8 T cell AICD Because most hematopoietic- derived cell lineages, including DCs (1, 21, 22), can express CD137, it is possible that CD137 signaling in one or more of these lineages is needed to induce immune suppression. To test this possibility, we adoptively transferred 1 3 104 CD137-sufficient Thy 1.1 congenic P14 T cells with 1 3 106 spleen cells from BL/6 Rag2/2 mice into CD1372/2 recipients. P14 T cell recipients were infected with LCMV the following day and anti-CD137 or rat IgG treated on day 1 postinfection. The frequency and absolute numbers P14 T cells, but not endogenous Thy 1.2+ CD1372/2 virus-specific T cells, were markedly reduced in the anti-CD137 treated mice (Fig. 2A, 2B). To show that the number of P14 cells adoptively transferred did not influence the outcome, we carried out similar studies in which the number of P14 T cells transferred into the recipients varied from 1 3 103 to 1 3 105 (Supplemental Fig. 1). To determine which spleen cell lineages might be required to induce immune suppression T cell deletion, we cotransferred P14 T cells together with BL/6 CD45.1 congenic NK cells, macrophages, or DCs into Thy 1.2+ CD45.2+ BL/6 CD1372/2 mice, followed by LCMV infection the following day and injection of anti-CD137 mAb or rat IgG 1 d thereafter. Of the cell lineages transferred, only DCs had the capacity to reduce the frequency of P14 T cells, whereas leaving endogenous CD1372/2 GP33–41-specific CD8 T cells in the spleen unaffected (Fig. 3A), an observation that correlated with a reduction in the absolute number of splenic P14 T cells, but not endogenous virusspecific T cells (Fig. 3B); although we only show data from the spleen, identical observations were made in lymph nodes after anti-CD137 treatment. We next questioned whether a specific DC subset induced P14 T cell deletion in virus-infected anti-CD137–

FIGURE 2. CD137-sufficient CD8 T cells and Rag2/2 spleen cells are required for anti-CD137– induced T cell AICD in CD1372/2 mice. The 1 3 104 P14 T cells were coinjected with 1 3 106 Rag-1 KO spleen cells into CD1372/2 mice. The mice were infected with LCMVArmstrong and treated with anti-CD137 or rat IgG at day 1 postinfection. On day 8 postinfection, the frequency of splenic P14 T cells and endogenous GP33–41-tetramer stained CD8 T cells were measured by FACS analysis (A), absolute numbers of viable P14 T cells and endogenous GP33–41-tetramer stained CD8 T cells were enumerated by FACS and trypan blue exclusion microscopy (B).

injected mice. BL/6 mice were injected daily with Flt-3 ligand-Ig and 10 d later splenic CD11c+ Thy1.22 (non-T cells) spleen cells were FACS sorted based on the presence or absence of CD8a or B220 expression (Fig. 3C). The three sorted cell populations were individually coinjected with P14 T cells into CD1372/2 mice 1 d prior to LCMV infection, followed by anti-CD137 injection 1 d thereafter. On day 8 postinfection, the absolute numbers of viable P14 T cells in the spleens of each group was determined. The results of this study demonstrated that all three DC subsets were able to induce P14 T cell deletion (Fig. 3D), an observation that is consistent with the view that the induction of tolerance and Ag priming are not functions carried out by specialized subsets of DCs (23, 24). To verify that the deletional effect of DCs on T cells was dependent on the ability of DCs to express CD137, we isolated splenic DCs from Flt-3 ligand-Ig–treated CD1372/2 mice and coinjected them with P14 T cells into CD1372/2 BL/6 mice prior to LCMV infection and anti-CD137 treatment. In contrast to adoptive transfer of WT DCs, no difference was seen between rat IgG- and anti-CD137–treated groups on P14 T cells expansion in mice receiving CD1372/2 DCs (Fig. 3E). In our previous studies, we had shown that anti-CD137 must be given during the early phase of T cell activation if it is to induce AICD (4), a period when naive T cells are being primed by DCs. Therefore, given that populations of splenic DCs that constitutively express CD137 ex vivo can be found in naive mice (1, 21, 22), we considered DCs as a potential candidate in driving CD137-dependent T cell AICD. To address this possibility, we first confirmed that both conventional and pDCs expressed CD137 in naive uninfected, and LCMV-infected mice during the first 72 h of viral infection. In naive mice, the frequency of CD137+ DCs is highest in pDCs and CD8a2 DCs, respectively, and is markedly reduced in these populations within 24 h of infection. In contrast, the frequency of CD137+ CD8a+ DCs is low in naive mice and remains so 24 h after infection, but more than doubles over the next 48 h, whereas the frequency of CD137+ pDCs and CD8a2

The Journal of Immunology DCs remain unchanged postinfection (Supplemental Fig. 2). Thus, DCs express CD137 in naive mice as previously reported (21, 22) and after LCMV infection as shown in this study. At present it is not clear whether the change in frequency of CD137+ DC subsets postinfection reflect the emigration of these DCs out of the spleen, their deletion, or CD137 downregulation. Anti-CD137 treatment markedly increased the frequency of TNF-a–producing DCs To determine whether anti-CD137 mAb treatment altered DC function, we examined the effect of anti-CD137 mAb treatment on TNF-a production by DCs obtained from LCMV-infected antiCD137 or rat Ig-injected mice as it has been shown that TNF-a is a potent inducer of T cell death (25–27), and because TNF-a plays an important role in the contraction of LCMV-specific CD8 T cells (28, 29), P14 T cells were adoptively transferred into naive WT or CD1372/2 BL/6 mice prior to infection with LCMV. The following day, the mice were injected with anti-CD137 or rat IgG and on day 8 postinfection, spleen cells were stained with mAbs specific for Thy1.1, Thy1.2, CD11c, and CD11b. The cells were then permeabilized and stained with mAbs specific for TNF-a. Thy1.1+ and Thy1.2+ T cells were excluded from analysis during flow cytometry and CD11c+ DCs were analyzed for intracellular TNF-a. Approximately 8% of splenic DCs obtained from rat IgGinjected mice expressed TNF-a, whereas .38% of DCs from antiCD137 mAb-injected mice expressed elevated levels of TNF-a (Fig. 4A) consistent with our earlier observation of heightened and sustained levels of TNF-a in the serum of virus-infected antiCD137–injected mice (4). It had been shown that both FasL and TNF-a are mediators of T cell deletion, and that whereas the TNF-a–TNFR2 signaling axis was largely responsible for deletion

FIGURE 3. CD137 cross-linking of DCs and CD8 T cells is essential for CD8 AICD. BL/6 CD45.1 congenic NK cells, DCs, or macrophages (Mø) were MACS purified and 1 3 106 cells were coinjected with 1 3 104 P14 T cells into CD1372/2 BL/6 mice. The mice were infected with LCMV Armstrong and injected with anti-CD137 or rat IgG at day 1 postinfection. On day 8 postinfection, the frequency of P14 T cells and endogenous GP33–41+ CD8 T cells in the spleen were determined by FACS analysis (A) and the absolute numbers of viable P14 T cells and endogenous GP33–41+ CD8 T cells was determined (B). BL/6 mice were Flt-3 ligand injected on a daily basis for 9 d. On day 10 splenic CD11c+ DCs were FACS-sorted to obtain purified CD8+, CD82 or B220+ DCs (C). FACS-sorted DCs (0.3 3 106) were coinjected with 1 3 104 P14 T cells into CD1372/2 BL/6 mice prior to infection with LCMV Armstrong and injection with antiCD137 or rat IgG at day 1 postinfection. On day 8 postinfection, spleen cells were phenotyped for absolute numbers of viable P14 T cells (D). As a control, 1 3 106 MACS purified CD1372/2 pan DC were coinjected with 1 3 104 P14 T cells into CD1372/2 BL/ 6 mice and treated and analyzed as above (E).

4773 of most CD8 T cells, the FasL-Fas axis led to deletion of CD4 T cells (26, 27). However, we did not find this to be the case. Certainly, TNF-a plays a significant role in mediating CD8 T cell deletion in LCMV-infected anti-CD137–injected mice. Nevertheless, we still observed significant P14 T cell deletion in LCMVinfected anti-CD137–injected TNFR1/2 double-deficient mice (Fig. 4B) suggesting that the FasL-Fas death pathway was largely responsible for the loss of these cells as we had previously reported (4), and that the extent to which AICD was diminished was due to the now recognized need of TNFR2 signaling in augmenting the Fas-mediated death pathway (30). Anti-CD137 induced Stat3 phosphorylation in DCs Although DCs are essential for Ag processing and presentation of pathogen-derived peptides to naive T cells, it is also clear that DCs are instrumental in maintaining self-tolerance (23, 31). It has recently been shown that Stat3 activation in DCs is essential for the induction of Ag-specific T cell tolerance (32). Given this report, and because we had previously shown that high concentrations of IL-10 appear in the serum of LCMV-infected anti-CD137–treated mice within 2 d of infection (4), and IL-10 and Stat3 regulate one another, we questioned whether CD137 cross-linking on DCs might induce Stat3 activation. To address this possibility, splenic DCs were isolated from Flt3-L-Ig injected mice stimulated in vitro with antiCD137 or rat IgG and cross-linked with sheep anti-rat IgG. Cell lysates were prepared and proteins separated by SDS-PAGE and Western blotted with anti-Stat3 and anti–phosphotyrosine-Stat3 as described in Materials and Methods. The results of this experiment demonstrated that anti-CD137 cross-linking induced detectable Stat3 Y705 phosphorylation, and that Ab-mediated cross-linking of cell surface bound anti-CD137 further amplified Stat3 tyrosine

4774

CD137-MEDIATED T CELL AICD DURING VIRAL INFECTION

FIGURE 4. DCs produce elevated levels of TNFa following LCMV infection and anti-CD137 injection. A, WT or CD1372/2 BL/6 mice were injected with P14 T cells, infected with LCMV Armstrong and injected with anti-CD137 as above. On day 8 postinfection, spleen cells were cultured for 5 h in complete RPMI 1640 media at the presence of Golgi plug and surfacestained with fluorochrome-conjugated mAbs specific for Thy1.1, Thy1.2, CD8, NK1.1, CD11c, CD11b prior to intracellular staining for intracellular TNF. B, AntiCD137–mediated immune suppression is partially rescued in TNFR1/2 double KO mice. BL/6 and TNFR1/2 double KO BL/6 mice were infected with LCMV and treated with anti-CD137 or Rat IgG on day 1 postinfection. On day 8 postinfection, the frequency of endogenous GP33–41+ CD8 T cells in the spleen were determined by FACS analysis, and the absolute numbers of endogenous GP33–41+ CD8 T cells was determined.

phosphorylation (Fig. 5A). To determine whether CD137-mediated Stat3 activation correlated with anti-CD137–induced immune suppression and CD8 T cell deletion during LCMV infection, BL/6 and BL/6 Stat3 conditional KO mice (32) were infected with LCMV and injected the following day with either rat IgG or anti-CD137. On day 8 postinfection, spleen cells were stimulated in vitro with LCMV-derived GP33-41 or NP396-404 immunodominant class I-restricted peptides for 5 h and then phenotyped for CD8 expression, permeabilized, and stained with fluorochrome-conjugated anti–IFN-g mAbs to reveal intracellular IFN-g production. FACS analysis revealed that CD8 T cells in BL/6 infected mice, but not CD8 T cells from BL/6 Stat3 conditional KO mice, failed to produce IFN-g fol-

lowing peptide stimulation. In fact, we observed that CD8 T cell response to LCMV was significantly higher in the Stat3 conditional KO mice (Fig. 5B). Soluble FasL is elevated in infected anti-CD137 injected mice Anti-CD137–mediated T cell AICD is Fas and FasL-dependent (4). However, we could not detect significant increases in FasL expression on the surface, or after intracellular staining of permeablized spleen cells or P14 T cells isolated from LCMV-infected anti-CD137–injected mice. FasL can be cleaved from the plasma membrane by metalloproteinases releasing an active soluble form of FasL that induces Fas-mediated cell death (33, 34). Given that we

FIGURE 5. CD137 cross-linking induced Stat3 activation in DCs. A, Splenic DC from Flt3-L–injected mice were positively selected by CD11c expression after T cell and B cell depletion on Miltenyi columns. The cells were stimulated in vitro with anti-CD137 or rat IgG and subjected to secondary anti-rat Ig cross-linking or not. Proteins from cell lysates were separated by SDS-PAGE and Western blotted with Abs specific for nontyrosine phosphorylated, or tyrosine phosphorylated Stat3. Lane 4 contains phosphorylated Stat3 (a positive control provided by the manufacturer). B, Anti-CD137 failed to induce CD8 T cell AICD in Stat3 conditional KO mice. BL/6 Stat3 conditional KO mice lacking Stat3 expression in DCs and WT BL/6 mice were infected with LCMV and injected with either rat IgG or anti-CD137 the following day. On day 8 postinfection, spleen cells were stimulated for 5 h at 37˚C with either GP33–41 or NP396–404 peptides representing immunodominant CD8 restricted epitopes of LCMV. The cells were surface stained with fluorochrome-conjugated anti-CD8 mAbs, permeabilized and intracellular stained with fluorochrome-conjugated anti–IFN-g mAbs, washed and analyzed using a BD ARIA II flow cytometer and FlowJo software.

The Journal of Immunology could not detect significantly elevated levels cell associated FasL, we hypothesized that the level of sFasL in the serum of LCMV-infected, anti-CD137–injected mice would be elevated relative to controls either because it was enzymatically cleaved from the cell, or because it was released by apoptotic cells, and this indeed was found to be the case (Fig. 6A). Because TNF-a potently upregulates Fas expression we measured serum TNF-a levels in virus-infected anti-CD137 or rat IgGinjected mice and as expected, we observed an ∼9-fold increase in the level of TNF-a in the serum of anti-CD137–injected mice (Fig. 6B). To identify whether P14 T cells might be a source of sFasL, we adoptively transferred these T cells into FasL2/2 mice, infected the mice with LCMV and injected them with rat IgG or anti-CD137 mAb the following day. On day 8 postinfection, the frequency (Fig. 6C) and absolute number of splenic P14 T cells (data not shown) was significantly reduced in the anti-CD137–injected mice, whereas the frequency of FasL+ P14 T cells from infected anti-CD137–injected mice had increased (Fig. 6C) but not their absolute number (data not shown). Moreover, the amount of sFasL was significantly increased in the serum of P14 T cell transferred virus-infected, anti-CD137–treated

4775 FasL2/2 recipient mice over that observed in rat IgG controls indicating that high levels of sFasL was shed from P14 T cells (Fig. 6D). Fate of anti-CD137–treated CD8 T cells Although P14 T cells (as well as virus-specific endogenous WT CD8 T cells) in LCMV-infected anti-CD137–injected mice undergo Fas and TNF-a–dependent AICD (4), by day 8 of infection the number of P14 T cells had nevertheless increased 5- to 8-fold over input. It is routinely observed that ∼10% of adoptively transferred P14 T cells can be accounted for within 24 h after cell transfer, with the fate of the remaining 90% of P14 T cells being unaccounted for (35). Thus, the level of day 8 expansion of CD8 T cells in anti-CD137–injected LCMV-infected mice may actually be as high as 100-fold. On the other hand, it has been reported that Ag-specific T cells are not lost on adoptive transfer. Rather, during their initial stage of activation, they become refractory to extraction from the spleen and therefore not present in single-cell suspensions used for FACS analysis. However, these T cells can be readily identified following in situ using immunohistochemical procedures (36). Regardless, the observed increase in P14 T cells is two logs lower than that observed in rat IgG-injected LCMV-infected mice. To study the fate and function of these T cells beyond day 8 postinfection, we adoptively transferred P14 T cells into BL/6 mice, infected them with LCMV and injected them with anti-CD137 mAb or rat IgG on day 1 postinfection. On days 8, 14, 21, and 90 postinfection, the frequency and absolute number of viable splenic P14 T cells were measured by FACS analysis. Assuming 10% P14 T cell take, by day 8 of infection the number of P14 T cells in LCMV-infected rat IgG-treated mice increased by 6300-fold, whereas those in anti-CD137 injected LCMV-infected mice increased by ∼70-fold. By day 14 postinfection the number of P14 T cells in rat IgG injected mice decreased by ∼2-fold, whereas those in anti-CD137–injected mice remained essentially unchanged. By day 21 postinfection, the number of P14 T cells in the spleens of rat IgG-injected mice had again contracted by ∼2-fold, whereas the number of P14 T cells in the anti-CD137–injected mice decreased by almost 5-fold; and by day 90 postinfection, P14 T cells were all but absent in the antiCD137–injected mice (Fig. 7A). Furthermore, whereas .96% of the P14 T cells retrieved from rat IgG-injected LCMV-infected mice produced IFN-g after in vitro peptide restimulation over a 21-d period, we observed a progressive loss of P14 T cells having the capacity to produce IFN-g in response to peptide stimulation from 77% to 50% over this interval (Fig. 7B), and by day 90 postinfection these T cells were all but undetectable. Thus, although P14 T cells underwent significant expansion and persisted in the spleens of antiCD137–injected mice through 21 d postinfection, they developed an exhausted phenotype (expressing PD-1 and CD137, data not shown); lost the capacity to produce IFN-g in response to peptide stimulation, and underwent accelerated contraction leading to a fail to clear an acute viral infection as we have previously shown (4). Regulatory T cells are not required for virus-specific CD8 T cell AICD

FIGURE 6. Anti-CD137 treatment elevates sFasL levels. CD8 T cellderived sFasL is elevated in LCMV-infected anti-CD137–injected mice. BL/6 mice were infected with LCMV Armstrong and injected with antiCD137 or rat IgG on day 1 postinfection. On day 8 postinfection, sera were collected and analyzed for sFasL (A) and TNF-a (B) by ELISA. P14 T cells were adoptively transferred to WT BL/6 mice or FasL2/2 mice. The mice were infected with LCMV Armstrong and injected with antiCD137 on day 1 postinfection. On day 8 postinfection, spleen cells were FACS analyzed for surface expression of Fas and FasL on P14 T cells (C); sera were collected from both groups of mice and assayed for the presence of sFasL by ELISA (D).

Having demonstrated that signaling via CD137 in virus-specific T cells and DCs was essential for the induction of their deletion, we sought to determine whether regulatory T cells might play a role the suppression of P14 T cell function and their deletion. To test this possibility P14 T cells were adoptively transferred into CD42/2, CD82/2 BL/6 mice prior to infection with LCMV. On day 1 postinfection, the mice were injected with anti-CD137 mAbs or rat IgG, and on day 8 postinfection the frequency (Fig. 8A) and absolute numbers (Fig. 8B) of splenic P14 T cells in CD42/2 and CD82/2 mice were analyzed by flow cytometry. The results of this study demonstrated that regulatory CD4 do not appear to be required for

4776

FIGURE 7. Persistence of effector CD8 T cells after induction of AICD. P14 T cells were adoptively transferred into BL/6 mice prior to infection with LCMV Armstrong. The mice were injected with anti-CD137 mAbs (dotted lines) or rat IgG (solid lines) on day 1 postinfection. On days 8, 14, 21, and 90 postinfection, the absolute number of viable P14 T cells was measured (A). Spleen cells from the above treated mice were stimulated in vitro for 5 h with GP33–41 peptides, surface stained with anti-Thy1.1 mAbs and permeabilized prior to incubation with anti–IFN-g mAbs and analyzed by FACS for the frequency of IFN-g–producing P14 T cells (B).

suppression of P14 T cell function or their deletion. Studies carried out in P14 T cell containing CD82/2 mice suggest that if regulatory CD8 T cells play a role in the suppression of P14 T cells, these putative CD8 T regulatory cells (Tregs) must derive from P14 T cells. Although we think this is unlikely, we cannot rule out this possibility based on our current studies.

Discussion Costimulatory receptors provide signals needed to sustain T cells during expansion, and contribute to T cell survival and memory formation. The TNFR family member CD137 is one such example (9, 12). CD137-mediated signaling in vitro and in vivo can provide T cell costimulation (2, 10, 14, 20, 37, 38) or induce immune

FIGURE 8. Tregs do not contribute to anti-CD137induced T cell anergy. Tregs are not required for virusspecific CD8 T cell AICD. P14 T cells were adoptively transferred into BL/6, CD42/2 BL/6 or CD82/2 BL/6 mice prior to infection with LCMV and injection of anti-CD137 or rat IgG as described previously. On day 8 postinfection, spleen cells were phenotyped by FACS to assess P14 CD8 T cell frequencies (A), absolute number of viable splenic P14 T cells (B).

CD137-MEDIATED T CELL AICD DURING VIRAL INFECTION nonresponsiveness (5, 16–18). This dichotomous behavior, first thought to reflect differential regulation of help and CTL activity proved incorrect when we found through analysis of in vivo antiCD137 signaling showed the immunopotentiating and immunosuppressive outcome of the immune response after CD137 engagement applied equally to Ag-specific CD4 and CD8 T cells. Rather, this fate decision appeared to be largely, or solely dependent on the timing of the administration of anti-CD137 relative to the course of the immune response (4). Thus, enforced CD137 signaling during the initial phase of an immune response to viral infection led to immune suppression and T cell AICD, whereas postponement of CD137 signaling until postpriming led to enhanced T cell costimulation (4). CD8 T cell AICD was found to be IL-10, Fas, and TNF-a–dependent (4). Based on existing studies, it has been implicitly viewed that CD137-directed regulation of T cell function was a consequence of direct CD137 signaling in Ag-activated T cells. However, given that anti-CD137–induced T cell deletion only occurred when anti-CD137 was injected during the early stages of immunization (15), or viral infection (4), and because CD137 expression can be expressed by multiple lineages of hematopoietic cells, including NK cells (39) and DCs (21, 22), we tested, and found, that CD137 signaling in CD8 T cells, whereas necessary, was insufficient to induce their immunosuppression or deletion of CD8 T cells. Subsequent studies revealed that CD137-mediated signaling in any of the major subsets of DCs in the spleen was essential for immunosuppression and deletion of CD8 T cells, whereas neither CD137-sufficient NK cells, nor CD137-sufficient macrophages could substitute for DCs. In vitro CD137 receptor cross-linking on DCs rapidly induced tyrosine phosphorylation of Stat3. In vivo studies carried out in LCMV-infected anti-CD137–injected Stat3 conditional KO mice in which DCs and macrophages lacked Stat3 expression revealed that anti-CD137 signaling induced neither virus-specific CD8 T cell suppression nor deletion. Although we do not know how CD137 signaling induces Stat3 activation, we have previously shown that LCMV-infected anti-CD137– injected mice rapidly generate very high levels of IL-10 that are detectable in their serum within 2 d of infection, a time when CD137-mediated signaling is instrumental to the induction of immune suppression (4). IL-10–mediated signaling leads to Stat3 phosphorylation and nuclear translocation (40, 41) and activated Stat3 regulates IL-10 gene activation. Whether Stat3 phosphorylation is required for anti-CD137– induced immune suppression still remains to be determined. This is important because inhibition of histone deacetylase (HDAC) induces Stat3-mediated indoleammine 2, 3 deoxygenase (IDO) production in DCs through a pathway that does not induce Stat3 tyrosine phosphorylation. However, although HDAC did not directly induce Stat3 tyrosine phosphorylation, it did not adversely affect Stat3 phosphorylation

The Journal of Immunology (42). Thus, it is unclear whether Stat3 phosphorylation contributed to IDO synthesis. Nevertheless, generation of IDO through this pathway coincides with the observation that anti-CD137 mAbs ameliorates collagen-induced arthritis in mice through IFN-g and IDO-dependent mechanisms (17). Whether IFN-g is required for IDO production by DCs after HDAC inhibition is not known. However, we found that anti-CD137 effectively induced CD8 T cell suppression and AICD in LCMV-infected IFN-gR2/2 mice (data not shown) indicating that if IDO is a central mediator of CD137induced, DC-mediated immune suppression and/or AICD, it occurs independent of IFN-g. Therefore, it is worth noting that microglia can produce IDO through an IFN-g–independent pathway (43). Given the above, we are working to determine whether anti-CD137– mediated signaling in DCs inhibits HDAC and through this process, induces Stat3-mediated IDO production. The role of CD137-mediated signaling in DCs has yet to be thoroughly studied. However, follicular DCs were reported to be lost from the spleens of anti-CD137–injected mice during immunization but the absence of follicular DC in these mice was not linked to a specific deletional process (44). Contrary to inducing DC deletion, CD137-mediated signaling can provide antiapoptotic prosurvival signals to DCs (45). The differences posed by these two reports might be reconciled if anti-CD137–mediated and CD137 receptor-ligand–mediated signals induce nonidentical signals, or if CD137 signals selectively induce survival or deletion programs in cells that express them in a context, and/or timedependent manner. The studies described in this report confirmed our earlier observations showing that CD137 signaling during Ag priming of virus-specific CD8 T cells induced their upregulation of Fas, and that anti-CD137–induced AICD was Fas-dependent as it did not occur in Fas2/2 mice (4). Unexpectedly, we found CD137 signaling in CD8 T cells while necessary, was insufficient to induce suppression of virus-specific T cell expansion, or AICD. Rather, we found that CD137-mediated signaling in DCs like that in CD8 T cells was essential, but insufficient to induce T cell suppression and deletion. Instead, CD137 signaling in both lineages was needed. A second noteworthy observation made during the course of this study was the finding that, whereas CD137-mediated T cell AICD was completely absent in Fas2/2 mice, this was not found to be the case in FasL2/2 mice where T cell deletion still occurred, albeit less efficiently than that observed in WT mice. Thus, anti-CD137 induction of CD8 T cell AICD seems to require at least two contributory signaling pathways where Fas is the major contributor, whereas TNF plays a contributing role. This finding is consistent with previous observations showing that TNF is important for LCMV-specific CD8 T cell contraction (28, 29), and as we have already shown, it is required for CD137-induced upregulation of Fas on virus-specific CD8 T cells (4); moreover Fasmediated CD8 T cell deletion is compromised in TNFR2-deficient mice (30). In summary, our studies demonstrate that CD137 signaling in DCs may play a pivotal role in dampening the proliferation of naive CD8 T cells during the priming phase of T cell activation in response to viral infection, and suggest that this pathway might participate in the early events that program T cell contraction described by Badovinac et al. (46). Precisely how CD137-mediated signals, affect DC function remains to be determined. Previous studies have shown that Stat3 activation blocks DC maturation and induces T cell tolerance (32). It is clear that immature DCs play an important role in maintaining self-tolerance (47, 48), and some subsets of DCs have been shown to have regulatory properties (49–52). Whether CD137 signaling modulates DC maturation, limits their survival, or induces regulatory DCs is not known at

4777 this time and we are currently trying to address these possibilities. Given the clinical potential of using agonistic CD137 ligands as therapeutics for treating cancer, autoimmune disease, and chronic infections, such as HIV, understanding the effects of administration of these reagents in animal models has important clinical implications.

Acknowledgments We thank the members of the Emory University School of Medicine Flow Cytometry Core for assistance in cell sorting, and the National Institutes of Health Tetramer Core Facility at Emory University for assistance in providing MHC class I tetramers used in our studies.

Disclosures The authors have no financial conflicts of interest.

References 1. Watts, T. H. 2005. TNF/TNFR family members in costimulation of T cell responses. Annu. Rev. Immunol. 23: 23–68. 2. Pollok, K. E., Y. J. Kim, Z. Zhou, J. Hurtado, K. K. Kim, R. T. Pickard, and B. S. Kwon. 1993. Inducible T cell antigen 4-1BB. Analysis of expression and function. J. Immunol. 150: 771–781. 3. Croft, M. 2003. Co-stimulatory members of the TNFR family: keys to effective T-cell immunity? Nat. Rev. Immunol. 3: 609–620. 4. Zhang, B., C. H. Maris, J. Foell, J. Whitmire, L. Niu, J. Song, B. S. Kwon, A. T. Vella, R. Ahmed, J. Jacob, and R. S. Mittler. 2007. Immune suppression or enhancement by CD137 T cell costimulation during acute viral infection is time dependent. J. Clin. Invest. 117: 3029–3041. 5. Foell, J. L., B. I. Diez-Mendiondo, O. H. Diez, U. Holzer, P. Ruck, A. S. Bapat, M. K. Hoffmann, R. S. Mittler, and G. E. Dannecker. 2004. Engagement of the CD137 (4-1BB) costimulatory molecule inhibits and reverses the autoimmune process in collagen-induced arthritis and establishes lasting disease resistance. Immunology 113: 89–98. 6. Goodwin, R. G., W. S. Din, T. Davis-Smith, D. M. Anderson, S. D. Gimpel, T. A. Sato, C. R. Maliszewski, C. I. Brannan, N. G. Copeland, N. A. Jenkins, et al. 1993. Molecular cloning of a ligand for the inducible T cell gene 4-1BB: a member of an emerging family of cytokines with homology to tumor necrosis factor. Eur. J. Immunol. 23: 2631–2641. 7. Alderson, M. R., C. A. Smith, T. W. Tough, T. Davis-Smith, R. J. Armitage, B. Falk, E. Roux, E. Baker, G. R. Sutherland, and W. S. Din. 1994. Molecular and biological characterization of human 4-1BB and its ligand. Eur. J. Immunol. 24: 2219–2227. 8. Loo, D. T., N. J. Chalupny, J. Bajorath, W. W. Shuford, R. S. Mittler, and A. Aruffo. 1997. Analysis of 4-1BBL and laminin binding to murine 4-1BB, a member of the tumor necrosis factor receptor superfamily, and comparison with human 4-1BB. J. Biol. Chem. 272: 6448–6456. 9. Takahashi, C., R. S. Mittler, and A. T. Vella. 1999. Cutting edge: 4-1BB is a bona fide CD8 T cell survival signal. J. Immunol. 162: 5037–5040. 10. Cannons, J. L., P. Lau, B. Ghumman, M. A. DeBenedette, H. Yagita, K. Okumura, and T. H. Watts. 2001. 4-1BB ligand induces cell division, sustains survival, and enhances effector function of CD4 and CD8 T cells with similar efficacy. J. Immunol. 167: 1313–1324. 11. Bertram, E. M., P. Lau, and T. H. Watts. 2002. Temporal segregation of 4-1BB versus CD28-mediated costimulation: 4-1BB ligand influences T cell numbers late in the primary response and regulates the size of the T cell memory response following influenza infection. J. Immunol. 168: 3777–3785. 12. Lee, H. W., S. J. Park, B. K. Choi, H. H. Kim, K. O. Nam, and B. S. Kwon. 2002. 4-1BB promotes the survival of CD8+ T lymphocytes by increasing expression of Bcl-xL and Bfl-1. J. Immunol. 169: 4882–4888. 13. Pulle, G., M. Vidric, and T. H. Watts. 2006. IL-15-dependent induction of 4-1BB promotes antigen-independent CD8 memory T cell survival. J. Immunol. 176: 2739–2748. 14. Melero, I., W. W. Shuford, S. A. Newby, A. Aruffo, J. A. Ledbetter, K. E. Hellstro¨m, R. S. Mittler, and L. Chen. 1997. Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors. Nat. Med. 3: 682–685. 15. Mittler, R. S., T. S. Bailey, K. Klussman, M. D. Trailsmith, and M. K. Hoffmann. 1999. Anti-4-1BB monoclonal antibodies abrogate T cell-dependent humoral immune responses in vivo through the induction of helper T cell anergy. J. Exp. Med. 190: 1535–1540. 16. Sun, Y., X. Lin, H. M. Chen, Q. Wu, S. K. Subudhi, L. Chen, and Y. X. Fu. 2002. Administration of agonistic anti-4-1BB monoclonal antibody leads to the amelioration of experimental autoimmune encephalomyelitis. J. Immunol. 168: 1457–1465. 17. Seo, S. K., J. H. Choi, Y. H. Kim, W. J. Kang, H. Y. Park, J. H. Suh, B. K. Choi, D. S. Vinay, and B. S. Kwon. 2004. 4-1BB-mediated immunotherapy of rheumatoid arthritis. Nat. Med. 10: 1088–1094. 18. Foell, J., S. Strahotin, S. P. O’Neil, M. M. McCausland, C. Suwyn, M. Haber, P. N. Chander, A. S. Bapat, X. J. Yan, N. Chiorazzi, et al. 2003. CD137 costimulatory T cell receptor engagement reverses acute disease in lupus-prone NZB x NZW F1 mice. J. Clin. Invest. 111: 1505–1518.

4778 19. Kwon, B. S., J. C. Hurtado, Z. H. Lee, K. B. Kwack, S. K. Seo, B. K. Choi, B. H. Koller, G. Wolisi, H. E. Broxmeyer, and D. S. Vinay. 2002. Immune responses in 4-1BB (CD137)-deficient mice. J. Immunol. 168: 5483–5490. 20. Shuford, W. W., K. Klussman, D. D. Tritchler, D. T. Loo, J. Chalupny, A. W. Siadak, T. J. Brown, J. Emswiler, H. Raecho, C. P. Larsen, et al. 1997. 4-1BB costimulatory signals preferentially induce CD8+ T cell proliferation and lead to the amplification in vivo of cytotoxic T cell responses. J. Exp. Med. 186: 47–55. 21. Futagawa, T., H. Akiba, T. Kodama, K. Takeda, Y. Hosoda, H. Yagita, and K. Okumura. 2002. Expression and function of 4-1BB and 4-1BB ligand on murine dendritic cells. Int. Immunol. 14: 275–286. 22. Wilcox, R. A., A. I. Chapoval, K. S. Gorski, M. Otsuji, T. Shin, D. B. Flies, K. Tamada, R. S. Mittler, H. Tsuchiya, D. M. Pardoll, and L. Chen. 2002. Cutting edge: Expression of functional CD137 receptor by dendritic cells. J. Immunol. 168: 4262–4267. 23. Belz, G. T., G. M. Behrens, C. M. Smith, J. F. Miller, C. Jones, K. Lejon, C. G. Fathman, S. N. Mueller, K. Shortman, F. R. Carbone, and W. R. Heath. 2002. The CD8alpha(+) dendritic cell is responsible for inducing peripheral selftolerance to tissue-associated antigens. J. Exp. Med. 196: 1099–1104. 24. Lanzavecchia, A. 1998. Immunology. Licence to kill. Nature 393: 413–414. 25. Sytwu, H. K., R. S. Liblau, and H. O. McDevitt. 1996. The roles of Fas/APO-1 (CD95) and TNF in antigen-induced programmed cell death in T cell receptor transgenic mice. Immunity 5: 17–30. 26. Zheng, L., G. Fisher, R. E. Miller, J. Peschon, D. H. Lynch, and M. J. Lenardo. 1995. Induction of apoptosis in mature T cells by tumour necrosis factor. Nature 377: 348–351. 27. Chan, F. K., and M. J. Lenardo. 2000. A crucial role for p80 TNF-R2 in amplifying p60 TNF-R1 apoptosis signals in T lymphocytes. Eur. J. Immunol. 30: 652–660. 28. Suresh, M., A. Singh, and C. Fischer. 2005. Role of tumor necrosis factor receptors in regulating CD8 T-cell responses during acute lymphocytic choriomeningitis virus infection. J. Virol. 79: 202–213. 29. Singh, A., and M. Suresh. 2007. A role for TNF in limiting the duration of CTL effector phase and magnitude of CD8 T cell memory. J. Leukoc. Biol. 82: 1201–1211. 30. Teh, H. S., A. Seebaran, and S. J. Teh. 2000. TNF receptor 2-deficient CD8 T cells are resistant to Fas/Fas ligand-induced cell death. J. Immunol. 165: 4814–4821. 31. Heath, W. R., and F. R. Carbone. 2001. Cross-presentation, dendritic cells, tolerance and immunity. Annu. Rev. Immunol. 19: 47–64. 32. Cheng, F., H. W. Wang, A. Cuenca, M. Huang, T. Ghansah, J. Brayer, W. G. Kerr, K. Takeda, S. Akira, S. P. Schoenberger, et al. 2003. A critical role for Stat3 signaling in immune tolerance. Immunity 19: 425–436. 33. Alcouffe, J., N. Therville, B. Se´gui, D. Nazzal, N. Blaes, R. Salvayre, M. Thomsen, and H. Benoist. 2004. Expression of membrane-bound and soluble FasL in Fas- and FADD-dependent T lymphocyte apoptosis induced by mildly oxidized LDL. FASEB J. 18: 122–124. 34. Green, D. R., and C. F. Ware. 1997. Fas-ligand: privilege and peril. Proc. Natl. Acad. Sci. USA 94: 5986–5990. 35. Blattman, J. N., R. Antia, D. J. Sourdive, X. Wang, S. M. Kaech, K. MuraliKrishna, J. D. Altman, and R. Ahmed. 2002. Estimating the precursor frequency of naive antigen-specific CD8 T cells. J. Exp. Med. 195: 657–664. 36. Maxwell, J. R., R. J. Rossi, S. J. McSorley, and A. T. Vella. 2004. T cell clonal conditioning: a phase occurring early after antigen presentation but before clonal expansion is impacted by Toll-like receptor stimulation. J. Immunol. 172: 248–259.

CD137-MEDIATED T CELL AICD DURING VIRAL INFECTION 37. Takahashi, C., R. S. Mittler, and A. T. Vella. 2001. Differential clonal expansion of CD4 and CD8 T cells in response to 4-1BB ligation: contribution of 4-1BB during inflammatory responses. Immunol. Lett. 76: 183–191. 38. DeBenedette, M. A., N. R. Chu, K. E. Pollok, J. Hurtado, W. F. Wade, B. S. Kwon, and T. H. Watts. 1995. Role of 4-1BB ligand in costimulation of T lymphocyte growth and its upregulation on M12 B lymphomas by cAMP. J. Exp. Med. 181: 985–992. 39. Melero, I., J. V. Johnston, W. W. Shufford, R. S. Mittler, and L. Chen. 1998. NK1.1 cells express 4-1BB (CDw137) costimulatory molecule and are required for tumor immunity elicited by anti-4-1BB monoclonal antibodies. Cell. Immunol. 190: 167–172. 40. Murray, P. J. 2007. The JAK-STAT signaling pathway: input and output integration. J. Immunol. 178: 2623–2629. 41. Hirata, N., Y. Yanagawa, K. Iwabuchi, and K. Onoe´. 2009. Selective regulation of interleukin-10 production via Janus kinase pathway in murine conventional dendritic cells. Cell. Immunol. 258: 9–17. 42. Sun, Y., Y. E. Chin, E. Weisiger, C. Malter, I. Tawara, T. Toubai, E. Gatza, P. Mascagni, C. A. Dinarello, and P. Reddy. 2009. Cutting edge: Negative regulation of dendritic cells through acetylation of the nonhistone protein STAT-3. J. Immunol. 182: 5899–5903. 43. Wang, Y., M. A. Lawson, R. Dantzer, and K. W. Kelley. 2010. LPS-induced indoleamine 2,3-dioxygenase is regulated in an interferon-gamma-independent manner by a JNK signaling pathway in primary murine microglia. Brain Behav. Immun. 24: 201–209. 44. Sun, Y., S. E. Blink, J. H. Chen, and Y. X. Fu. 2005. Regulation of follicular dendritic cell networks by activated T cells: the role of CD137 signaling. J. Immunol. 175: 884–890. 45. Choi, B. K., Y. H. Kim, P. M. Kwon, S. C. Lee, S. W. Kang, M. S. Kim, M. J. Lee, and B. S. Kwon. 2009. 4-1BB functions as a survival factor in dendritic cells. J. Immunol. 182: 4107–4115. 46. Badovinac, V. P., B. B. Porter, and J. T. Harty. 2004. CD8+ T cell contraction is controlled by early inflammation. Nat. Immunol. 5: 809–817. 47. Steinman, R. M., and M. C. Nussenzweig. 2002. Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance. Proc. Natl. Acad. Sci. USA 99: 351–358. 48. Wakkach, A., N. Fournier, V. Brun, J. P. Breittmayer, F. Cottrez, and H. Groux. 2003. Characterization of dendritic cells that induce tolerance and T regulatory 1 cell differentiation in vivo. Immunity 18: 605–617. 49. Kronin, V., K. Winkel, G. Su¨ss, A. Kelso, W. Heath, J. Kirberg, H. von Boehmer, and K. Shortman. 1996. A subclass of dendritic cells regulates the response of naive CD8 T cells by limiting their IL-2 production. J. Immunol. 157: 3819– 3827. 50. Kronin, V., L. Wu, S. Gong, M. C. Nussenzweig, and K. Shortman. 2000. DEC205 as a marker of dendritic cells with regulatory effects on CD8 T cell responses. Int. Immunol. 12: 731–735. 51. Su¨ss, G., and K. Shortman. 1996. A subclass of dendritic cells kills CD4 T cells via Fas/Fas-ligand-induced apoptosis. J. Exp. Med. 183: 1789–1796. 52. Zhang, M., H. Tang, Z. Guo, H. An, X. Zhu, W. Song, J. Guo, X. Huang, T. Chen, J. Wang, and X. Cao. 2004. Splenic stroma drives mature dendritic cells to differentiate into regulatory dendritic cells. Nat. Immunol. 5: 1124–1133.