Peripheral Deletion of CD8 T Cells Requires p38 MAPK in Cross-Presenting Dendritic Cells

This information is current as Trevor Smith, Xiaotian Lin, Marielle Mello, Kristi of September 25, 2021. Marquardt, Jocelyn Cheung, Binfeng Lu, Linda A. Sherman and Grégory Verdeil J Immunol published online 1 September 2017 http://www.jimmunol.org/content/early/2017/09/01/jimmun

ol.1700427 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2017/09/01/jimmunol.170042 Material 7.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 25, 2021

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2017 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published September 1, 2017, doi:10.4049/jimmunol.1700427 The Journal of Immunology

Peripheral Deletion of CD8 T Cells Requires p38 MAPK in Cross-Presenting Dendritic Cells

Trevor Smith,* Xiaotian Lin,* Marielle Mello,† Kristi Marquardt,* Jocelyn Cheung,* Binfeng Lu,‡ Linda A. Sherman,* and Gre´gory Verdeil*,†,x

Peripheral tolerance mechanisms exist to prevent autoimmune destruction by self-reactive T cells that escape thymic deletion. Dominant tolerance imposed by CD4+Foxp3+ T regulatory cells can actively control autoaggressive T cell responses. Tolerance mechanisms that act endogenous to the T cell also exist. These mechanisms include T cell inactivation (anergy) and deletion. A major difference between anergic T cells and T cells undergoing peripheral deletion is the capacity of the latter to still signal through MAPKs upon TCR stimulation, suggesting these signals may be required for T deletion. In this study, we used several different models of CD8 T cell deletion to investigate the contribution of MAPK activation. Using chemical inhibitors, we

established that inhibition of p38, but not ERK or JNK, rescue T cells from undergoing peripheral deletion both in vitro and Downloaded from in vivo. Using T cell–specific murine lines genetically altered in expression of p38a, and mice in which p38a was deleted only in CD11c-expressing cells, we surprisingly found that CD8 T cell–intrinsic p38a activation was not responsible for increased survival, but rather that inhibition of p38a in the Ag-presenting dendritic cells prevented CD8 T cell deletion. The Journal of Immunology, 2017, 199: 000–000. http://www.jimmunol.org/ entral tolerance purges the immune system of T cells Under noninflammatory conditions, CD8 T cells undergo tol- displaying high-affinity TCRs specific for self-peptide– erance induction in the periphery. This occurs after interacting with C MHC complexes. However, this process is not complete, cognate peptide–MHC class I complexes (signal 1) expressed on and significant numbers of T cells recognizing peripheral self- quiescent APCs that fail to provide sufficiently high levels of peptide–MHC molecules emigrate from the thymus (1). Fortu- costimulatory molecules (signal 2) or inflammatory cytokines nately, peripheral tolerance mechanisms are in place to protect (signal 3) for proper CD8 T cell activation with development of against dangerous autoreactive T cell responses. These peripheral effector cell functions such as cytotoxicity and cytokine produc- mechanisms of tolerance are especially important for the control tion (5, 6). Defined tolerance mechanisms intrinsic to the T cell include deletion (cell death) and inactivation (anergy) of autore- by guest on September 25, 2021 of CD8 T cell immunity because of widespread expression of active T cells (7–9). T cell anergy is dependent upon the contin- MHC class I molecules on all nucleated cells. There are several uous presence of Ag in the absence of inflammation. This is a defined mechanisms of peripheral tolerance. Dominant tolerance + + reversible process, and T cells regain function after a period of rest imposed by CD4 Foxp3 T regulatory cells (Tregs) can actively in the absence of Ag (9–11). During peripheral deletion, T cells prevent or control autoaggressive T cell responses (2–4). There proliferate in response to Ag, but then die of apoptotic cell death. also exist tolerance mechanisms that act endogenously in self- This has been demonstrated using TCR-transgenic CD8+ T cells reactive T cells. specific for islet b cell Ags that are continuously cross-presented in the pancreatic lymph nodes (pLNs) by dendritic cells (DCs). *Department of Immunology and Microbiology, Scripps Research Institute, La Jolla, Interaction with the cross-presented Ag in this system abortively CA 92037; †Centre d’Immunologie de Marseille-Luminy, Aix-Marseille Universite´, activates autoreactive T cells, leading to their deletion (12, 13). CNRS, INSERM, 13009 Marseille, France; ‡Department of Immunology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261; and xDepartment of Fun- Using in vivo models, we and others have shown that differential damental Oncology, University of Lausanne, 1066 Epalinges, Switzerland peripheral T cell tolerance mechanisms can be induced by varying ORCIDs: 0000-0002-9938-7633 (X.L.); 0000-0002-5454-3559 (M.M.); 0000-0001- the dose of soluble peptide injected under noninflammatory con- 6422-7069 (L.A.S.); 0000-0001-9445-8474 (G.V.). ditions. Using several different TCR transgenic models, including Received for publication March 22, 2017. Accepted for publication August 8, 2017. Clone 4 (CL4) specific for a peptide of hemagglutinin (HA) and This work was supported by National Institutes of Health Grant R0I DK050824 (to OT-1 cells specific for a peptide of OVA, we previously demon- L.A.S.); institutional grants from INSERM, CNRS and Aix-Marseille University to the strated that high levels of peptide favor the induction of a form of Centre d’Immunologie de Marseille Luminy; a Retour Post-Doctorants 2009 from the Agence Nationale de la Recherche (to G.V.); and a graduate student fellowship from anergy characterized by the loss of capacity of the cells to signal The Scripps Research Institute Kellogg School of Science and Technology (to X.L.). through MAPKs, including p38 and ERK (14, 15). Low levels of Address correspondence and reprint requests to Dr. Gre´gory Verdeil, Department of peptides favor deletional tolerance, which has been shown to be Fundamental Oncology, Universite´ de Lausanne, Chemin des Boveresses, 155, 1066 BIM dependent (16, 17). Cells committed toward deletional tol- Epalinges, Switzerland. E-mail address: [email protected] erance are still capable of phosphorylating p38 and JNK after The online version of this article contains supplemental material. TCR stimulation. The proximal TCR signaling events that result Abbreviations used in this article: CL4, Clone 4; DC, dendritic cell; HA, hemagglu- tinin; HDA, high-dose Ag; ILA, isoleucine, leucine, alanine; KO, knockout; LCMV, in peripheral deletion rather than anergy of T cells are still poorly lymphocytic choriomeningitis virus; LDA, low-dose Ag; LN, lymph node; pLN, defined. In this study, we tested the hypothesis that MAPK sig- pancreatic lymph node; RIP, rat insulin ; Treg, T regulatory cell; WT, naling may be required to initiate deletion. wild-type. To this end, we established an in vivo/in vitro model that mimics Copyright Ó 2017 by The American Association of Immunologists, Inc. 0022-1767/17/$35.00 peripheral deletion and allowed us to test the effects of MAPK

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1700427 2 REGULATION OF CD8 PERIPHERAL DELETION BY p38 MAPK inhibitors on CD8 T cell . We established that p38 in- Flow cytometry hibitors, but not ERK or JNK inhibitors, could partially prevent Abs were ordered from BD Biosciences (CD8, Thy1.1, Thy1.2, CD25, CD8 T cell deletion. We further confirmed those results in vivo CD44). Cells were analyzed on a FACSCalibur or a FACSLSRII cytometer using two different protocols to induce peripheral deletion (BD Biosciences). Data were analyzed using FlowJo (Tree Star) or Diva through cross-presentation of Ag. Surprisingly, we found that CD8 (BD Biosciences) software. For phospho-Ab labeling, after ex vivo T cell–intrinsic p38 activation was not responsible for survival, but restimulation cells were fixed in ice-cold 1.6% paraformaldehyde for 15 min. Cells were permeabilized with 90% methanol and stored at 220˚C. rather that inhibition of p38 in the cross-presenting DCs prevented Cells were labeled with anti-CD8, anti-Thy1.1, and anti–phospho-Abs CD8 T cell deletion. Consequently, p38 signaling in DCs appears specific for Erk and p38 (BD Biosciences) for 1 h in FACS buffer as a central regulator of peripheral deletion. (HBSS, 1% FCS, 0.1 mM EDTA). Loading of b2M knockout spleen cells with ILA peptide Materials and Methods ILA Ag-loaded spleen cells were prepared by osmotic shock as previously Mice described (21). In brief, splenocytes from b2M knockout (KO) mice were washed in RPMI 1640 and resuspended in hypertonic medium (0.5 M CL4 (18) TCR transgenic mice express a TCR specific for the HA518–526 epitope restricted by MHC class I H-2Kd. C57BL/6 Bim2/2 mice were sucrose, 10% w/v polyethylene glycol 1000, and 10 mM HEPES in RPMI 2 2 kindly provided by Dr. D. Green. Bim / and CL4 TCR rat insulin 1640 [pH 7.2]). Prewarmed hypotonic medium (40% H2O, 60% RPMI promoter (RIP)-Tag2-HA mice were each backcrossed with B10.D2 mice 1640) was added. Cells were pelleted by centrifugation and washed twice 3 6 for greater than eight generations. P14 T cells express a TCR specific for with ice-cold HBSS, and 30 10 in 0.2 ml of ILA-loaded splenocytes was injected i.v. into each recipient mouse. the GP33–41 epitope of lymphocytic choriomeningitis virus (LCMV). OT- 1 cells express a TCR specific for the SIINFEKL peptide from chicken 2 2 2 2 Statistical test Downloaded from OVA. C57BL/6 Gadd45g / and Gadd45b / mice were kindly pro- vided by Dr. B. Lu (Department of Immunology, University of Pitts- Data are expressed as mean + SD for each group. Unless mentioned, burgh, School of Medicine, Pittsburgh, PA). C57BL/6 p38a and p38b statistical differences between groups were evaluated with a nonpara- Tyr323-mutated mice were kindly providedbyDr.J.Ashwell(Laboratory metric Mann–Whitney t test using GraphPad Prism 5.0b software; p . of Immune Cell Biology, National Cancer Institute, National Institutes of 0.05 was considered statistically significant (*p , 0.05, **p , 0.01, Health, Bethesda, MD). Permission to use C57BL/6 Mapk14fl/fl (p38afl/fl) ***p , 0.001). mice (19) was kindly given by Dr. K. Otsu (Department of Cardio- vascular Medicine, Osaka University Graduate School of Medicine, http://www.jimmunol.org/ Osaka, Japan), and mice were crossed on CreERT2 mice obtained from Results Taconic. To induce the activity of the Cre recombinase, mice were ad- CL4 T cells undergoing peripheral deletion are ministrated 2 mg of tamoxifen by oral gavage for 4 d and used the day partially anergized after as a recipient or a source of P14 T cells. RIP-OVA mice were obtained from Dr. W. Heath (University of Melbourne, Melbourne, VIC, Using a previously reported model of CD8 tolerance (8), we wished Australia) and crossed on CD11c-Cre Mapk14fl mice. Animals were to compare the capacity of anergized T cells and T cells under- housed at The Scripps Research Institute animal facility, and all proce- going peripheral deletion to activate MAPK signaling upon dures were performed according to theNIHGuidefortheCareandUse restimulation. B10.D2 recipients that previously received purified of Laboratory Animals. TCR-transgenic CL4 CD8 T cells (CL4 cells) were injected i.v. Immunizations daily for 3 d with 1 mg (low-dose Ag [LDA]) or 100 mg (high-dose by guest on September 25, 2021 Ag [HDA]) of cognate KdHA peptide to induce peripheral deletion Recombinant vaccinia virus expressing the OVA of chicken OVA was as described previously (20). Mice were infected with 106 PFU of or anergy, respectively (15). We monitored the activation status recombinant vaccinia virus by i.v. injection. Amounts of the various and numbers of CL4 cells in the spleen between days 3 and 6 after peptides were injected in PBS. Peptides used in these studies included peptide injection. Flow cytometry analysis revealed comparable the HA518–526 (IYSTVASSL) epitope, isoleucine, leucine, alanine expression levels of the activation marker CD44, whereas CD25 (ILA)-HA (ILAIYSTVASSL), which extends the sequence of the nominal peptide by including three amino-proximal residues that ap- was not expressed (Fig. 1A). Cell numbers peaked on day 3 pear in the natural sequence of the , GP33M (KAVYNFATM) (Fig. 1B). As anticipated, there was a rapid loss in the numbers of and ILA-OVA257–264 (ILASIINFEKL). All peptides were synthe- CL4 cells thereafter in mice treated with LDA (Fig. 1B). The rapid sized at The Peptide Synthesis Core Facility at The Scripps Research decrease in cell number was not due to an absence of peptide Institute. between days 3 and 6, because CL4 T cell numbers contract with Preparation and adoptive transfer of naive TCR-transgenic the same kinetics if 1 mg of peptide is administered daily T cells throughout the duration of the experiment to day 6 (Supplemental

+ Fig. 1A). Treatment with HDA was not associated with a decrease CD8 T cells were isolated from the LNs of CL4, P14, or OT-1 TCR mice in CL4 T cell number after day 3, because similar numbers of CL4 (6–8 wk of age) by negative selection using the MACS CD8+ T cell iso- lation kit (Miltenyi Biotec). T cell purity was .85% with no contami- cells were detected in the spleens on days 3–6 (Fig. 1B). The nating CD4+ cells. For adoptive transfer experiments, the indicated number number of CL4 cells in mice treated with PBS did not change of cells was injected i.v. in a volume of 100 ml of PBS. significantly throughout the duration of the experiment (Fig. 1B). Cell culture condition Anergy is associated with an inability to signal through the TCR after restimulation (9). We measured by flow cytometry the ability 6 For in vitro culture after in vivo peptide injections, 5 3 10 total splenocytes of CL4 cells to phosphorylate the MAPKs Erk and p38 after Kd were cultured for 3 additional days in 10% FCS complete RPMI 1640 medium, without addition of any peptide. For flow cytometry experiments, HA peptide restimulation in vitro. HDA treatment prevented Erk stimulation was performed with peptides at 1026 M or PMA (10 ng/ml) phosphorylation as early as 24 h after the first peptide injection and Ionomycin (200 ng/ml). (5% of pERK+ CL4 cells; Fig. 1C, 1D). CL4 cells from mice treated with a low amount of peptide are partially anergized but In vitro and in vivo kinase inhibition can still phosphorylate Erk to some extent (from 29% of pERK+ SKF86002 (10 mM), SB220025 (10 mM), SP600125 (JNK inhibitor, 10 CL4 cells at 24 h to 10.5% at 72 h; Fig. 1C, 1D). The results are mM), JNK Inhibitor II (CAS 129-56-6; 5 mM), PD98059 (5 mM), and similar for p38 phosphorylation at 72 h posttreatment (Fig. 1E, U0126 (5 mM) (Sigma-Aldrich) were used at the indicated concentra- + tion in vitro. For in vivo inhibition, mice received twice-daily 1F). Forty-one percent of naive CL4 cells are phospho-p38 versus SB220025 (0.6 mg per injection), SB203580 (1 mg per injection), or 28 and 14% of CL4 cells after low dose or high dose of antigenic vehicle. peptide injection, respectively (Fig. 1D). Thus, HDA treatment The Journal of Immunology 3 Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 1. Injection of low or high dose of soluble Ag induces deletion or anergy of CD8 T cells. CL4 Thy1.1+ CD8 T cells (1 3 106) were adoptively transferred into B10D2 mice. Mice received daily injections of low (1 mg) or high (100 mg) dose of antigenic HA peptide for 3 d. (A) Levels of expression of CD25 and CD44 at day 3 on naive CL4 cells (N: shaded) or CL4 cells coming from mice injected with low (LDA: thin line) or high dose (HDA: thick line) of antigenic peptide are shown. Results are representative of at least three independent experiments. (B) CL4 cell numbers in splenocytes from mice that received low (black bars) or high (gray bars) dose of antigenic peptide or PBS (white bars) were determined at indicated days after the first peptide injection. (C–F) Mice were sacrificed 24, 48, or 72 h after the first injection of antigenic peptide. Splenocytes from mice treated with PBS (black triangles), LDA (black squares), or HDA (white circles) of antigenic HA peptide were stimulated for 10 min with medium alone, antigenic peptide (HA), or PMA and ionomycin (P/I), fixed and labeled for FACS analysis. (C) Representative histograms of phospho-ERK+ CL4 cells from the various conditions and median fluorescence intensity 6 SD (average of four mice from one representative experiment per condition) for pERK labeling are shown. (D) Average of four mice per condition per time point of one representative experiment out of two for p-ERK+ population is shown. (E) Representative histograms of phospho- p38+ CL4 cells from the various conditions and median fluorescence intensity 6 SD (average of three mice per condition from one representative ex- periment) are shown. (F) Pooled percentages of phospho-p38+ CL4 cells from two independent experiments are shown. *p , 0.05, ***p , 0.001.

induces almost complete blockade of signaling through Erk or p38 hour after the last Ag injection, splenocytes were isolated and MAPKs, whereas this inhibition is only partial after LDA treat- cultured in vitro for 72 h (Fig. 2A). As a consequence, both APCs ment of CL4. and T cells are present in the culture well. Numbers of non- apoptotic (annexin V2) CL4 cells at the beginning and the end of The p38 MAPK pathway is involved in peripheral deletion the culture period were compared. Most of the CL4 cells from Because the capacity to activate the MAPK pathway is an early untreated animals remained annexin V2 throughout the experi- difference between T cells undergoing peripheral deletion or ment. Only 16 and 54% of the initial pool of CL4 cells from mice anergy, we assessed the role of MAPK signaling pathways in the treated with LDA or HDA, respectively, remained annexin V2 at regulation of peripheral deletion. JNK and p38 MAPK pathways the end of the culture (Fig. 2B). Cell death in this system was BIM have been shown to regulate cell death in various cell types, in- dependent, because LDA-treated bim2/2 CL4 cells did not un- cluding T cells (22–25). We analyzed the effects of chemical in- dergo apoptosis as previously described (Fig. 2B) (16). To assess hibitors of MAPKs in an in vivo/in vitro model of deletion. CL4 the importance of MAPK signaling pathways in apoptosis in- cells were transferred into B10D2 mice, and the host mice were duction in this system, we added small-molecule inhibitors of treated with LDA or HDA for 3 d as previously described. One ERK, p38, and JNK during the in vitro culture phase (Fig. 2C, 4 REGULATION OF CD8 PERIPHERAL DELETION BY p38 MAPK

FIGURE 2. p38 inhibition rescues LDA-induced deletion in vitro. (A and B)106 WT or bim2/2 CL4 Thy1.1+ CD8 T cells were adoptively transferred into B10D2 mice. For 3 d, mice received daily injections of PBS, 1 (LDA) or 100 mg (HDA) of HA peptide. One hour after the last injection of peptide, mice were sacrificed and 5 3 106 total splenocytes were cultured for 3 additional days in the absence or presence of the indicated MAPK inhibitor. (B) Percent of survival by comparing the final number of annexin V2 CL4 cells with the initial number at the beginning of the culture is shown (three mice per condition from one representative experiment out of three). (C) Ratio of CL4 cell number between each inhibitor (p38 inhibitors [p38i], JNK inhibitors [JNKi], and ERK in- hibitors [ERKi]) and mock condition at the end of the culture after LDA treatment is shown. Pooled results from two to six independent experiments are shown. Each dot represents data from one mouse. ANOVA using Kruskal–Wallis nonpara- metric test with multiple comparison was used. (D) A repre- Downloaded from sentative FACS plot from experiment in (C) with the indicated markers is shown (gated on CD8+, Topro32, lymphocytic cells using forward scatter/side scatter). *p , 0.05, **p , 0.01. http://www.jimmunol.org/ 2D). Neither JNK nor ERK inhibition increased cell survival. until day 6, when most of the cells are gone (Supplemental Fig. Inhibitors for p38 were associated with a .2-fold increase in 1B). However, the dose of peptide required for deletion is lower survival rate of CL4 cells compared with control. than the one used for CL4 cells (0.2 versus 1 mg). We confirmed We next assessed the in vivo role of p38 in LDA-induced that most of the cells have been deleted by infecting the mice with deletional tolerance by injecting soluble p38 inhibitors (SB220025 LCMV 30 d after LDA treatment (Supplemental Fig. 1C). This or SB203580) in mice harboring adoptively transferred CL4 T cells process is BIM dependent, because bim-KO P14 T cells do not (Fig. 3A). Data from two independent experiments using each undergo LDA-induced deletion (Fig. 4A; Supplemental Fig. 1B). inhibitor showed that p38 inhibition partially protected CL4 cells This was further confirmed by using the same in vivo/in vitro model from deletion during LDA treatment, with a 4- to 7-fold increase used in Fig. 2, and we observed that the p38 inhibitor SB220025 by guest on September 25, 2021 in survival (Fig. 3A, 3B). rescues P14 T cells from undergoing deletion (Supplemental To confirm that the p38 pathway was not specific to LDA- Fig. 1D). induced deletion, we analyzed its role in another model of dele- Gadd45b and Gadd45g have been demonstrated to be important tional tolerance, in which cell death is induced after activation by molecules in the potentiation of MAPK p38 signaling in T cells cross-presented Ag in vivo. We had previously shown that CL4 after exposure to inflammatory mediators (27). In the absence of cells were deleted after activation by an ILA-extended HA Ag Gadd45b and Gadd45g, p38 signaling is attenuated in T cells. We variant that could not directly bind H-2Kd, but must be cross- investigated the effect such deficiency had on the induction of presented by DCs to stimulate CL4 cells (21). Treatment with deletional tolerance. In the absence of Gadd45b and Gadd45g, the cross-presented Ag resulted in efficient CL4 deletion that P14 T cells are still efficiently deleted (Fig. 4B), indicating that could be inhibited by the p38 inhibitor SB203580 in vivo (Fig. 3C). the conventional pathway of activation of p38 in T cells was not These data demonstrate that inhibition of the p38 pathway rescues required for deletion. CL4 cells from deletion after LDA treatment with nominal or a To assess whether the alternative pathway was critical for LDA- cross-presented Ag. induced deletion, we obtained mice that carried a mutation in the

+ p38a at position 323 that prevents TCR-mediated p38a activation Deficiency of CD8 T cell–intrinsic p38 signaling does not (26). P14 T cells carrying this mutation were adoptively transferred inhibit LDA-mediated peripheral deletion into B6 mice, and the hosts were treated with LDA. P14 T cells Two pathways of p38 signaling have been described for T cells: (1) lacking the alternative p38a activation pathway were efficiently the conventional pathway involving molecules; and (2) the deleted (Fig. 4C). Although p38a is the major isoform expressed in alternative pathway initiated after signaling through the TCR, T cells, p38b has been shown to possess some activity in T cells leading to the autophosphorylation of p38 (26). To determine (28). To rule out a compensatory role for p38b, we used P14 cells whether p38 signaling within the CD8+ T cell was critically in- containing Tyr323 mutations on both p38 a and b. These cells were volved in LDA-induced deletional tolerance, we obtained mice also efficiently deleted after LDA treatment (Fig. 4C). deficient for key signaling molecules involved in p38 signaling Finally, to fully exclude a critical role for p38a signaling in- pathways. Because these various KO/knockin mice were only trinsic to T cell deletion, we obtained ERT2-Cre Mapk14fl/fl P14 available on the C57BL/6 background, we used the C57BL/6 CD8 cells (p38adel). After tamoxifen treatment, p38 activation is P14 TCR transgenic model in these experiments. The P14 TCR completely abrogated upon TCR stimulation in these T cells recognizes the LCMV-derived epitope gp33m in the context of (Supplemental Fig. 1E). p38adel cells were efficiently deleted H2-Kb. P14 T cells undergo a similar kinetic of peripheral deletion after LDA treatment (Fig. 4D). Taken together, these results in- upon treatment with LDA compared with CL4 cells, with a peak dicate that p38 signaling intrinsic to CD8 T cells is not involved in of proliferation at day 3 and a gradual decrease in cell number peripheral deletion. The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 3. p38 inhibition rescues LDA-induced deletion in vivo. (A and B) Mice harboring CL4 T cells received daily i.v. injections of 1 mgofHA peptide and were injected twice a day with 600 mg of SB220025, 1 mg of SB203580, or vehicle alone at the indicated time points. CL4 cell numbers from spleen were determined 3 and 6 d after the first HA peptide injection (left graph), and data at day 6 were normalized to data from day 3 (right graph). Pooled results from two independent experiments for each inhibitor are shown. (B) Representative dot plots at days 3 and 6 are shown. (C) Mice harboring CL4 T cells received 30 3 106 apoptotic splenocytes loaded with ILA-HA peptide and were treated twice a day with 1 mg of SB203580 or vehicle alone as indicated. CL4 cell numbers at days 3 and 6 (left graph), and data from day 6 normalized to data from day 3 (right graph) are shown. Pooled results from two independent experiments are shown. *p , 0.05, **p , 0.01.

Role for T cell–extrinsic p38 signaling in CD8+ T cell Role of p38 in DCs to induce peripheral T cell deletion deletional tolerance It was previously established that cross-presenting CD8a+CD103+ The fact that CD8 T cell deficiency in p38 had no significant effect DCs are required to induce T cell peripheral deletion (12, 13). To on deletion was surprising, because p38 inhibition using small- determine whether deficiency of p38a in DCs affects T cell pe- molecule inhibitors saved CD8+ T cells from undergoing dele- ripheral deletion, we used CD11c-Cre Mapk14fl/fl mice with a tion after LDA treatment (Fig. 2C). We next considered whether specific inactivation of the p38a gene in CD11c-expressing cells p38 signaling in a non–T cell population was required for deletion. (p38del-CD11c mice). We tested two models in which peripheral Wild-type (WT) P14 cells were transferred into p38a-deficient deletion is induced by Ag cross-presented by DCs. We previously hosts. P14 cells were deleted significantly less efficiently after showed that when the cognate peptide recognized by OT-1 cells is receiving LDA treatment in a p38a-deficient host compared with extended at the N-terminal three residues (ILA), it must undergo WT host (Fig. 5A). Furthermore, upon challenge with LCMV TAP-dependent processing and cross-presentation by DCs before virus 30 d after LDA treatment, a robust recall response is ob- it can be recognized by CD8 cells, and that such activation results served 7 d postinfection in the p38a-deficient host compared with in T cell deletion (22). As a second model of deletion, we used a negligible response in the WT host (Fig. 5B). There was no RIP-OVAhi mice, which express OVA under the RIP in the b cells significant difference in the deletion efficiency between p38a- of the pancreas (30). OT-1 cells injected into these mice are de- deficient and WT P14 cells in a p38a-deficient host (Fig. 5), leted through recognition of cross-presented Ag in the pLNs (31). further suggesting intrinsic p38 MAPK signaling is not essential in It was necessary to employ a third TCR transgenic model, that is, the T cell population undergoing deletional tolerance. We noticed OT-1 T cells, to use this model of endogenous deletion. a 10-fold decrease in the amplification of virus-specific T cells Mice received 1 3 104 purified CD8+ Thy1.1 OT-1 cells, fol- when the host is p38a deficient. Systemic p38 inhibition has been lowed by injection of apoptotic splenocytes from B6 mice loaded shown to inhibit viral replication and viral protein synthesis (29). with the long ILA-SIINFEKL peptide. This peptide requires cross- This could explain those results. presentation to induce deletion of OT-1 cells, as previously de- 6 REGULATION OF CD8 PERIPHERAL DELETION BY p38 MAPK

FIGURE 4. CD8 T cell–intrinsic p38 deficiency does not rescue them from peripheral deletion. C57/BL6 mice were adoptively transferred with 1 3 106 WT (A–D) and bim2/2 (A), GADD45b2/2, GADD45g2/2, and GADD45bg2/2 (B), p38a(Y323F) and p38ab(Y323F) (C), and p38aflox (D) P14 Thy1.1+ CD8 T cells. For 3 days, host mice received daily injections of PBS or 0.2 mg (LDA) of GP33m peptide. Mice were sacrificed on days 3 and 6, and P14 T cell numbers in the spleen were determined by flow cytometry. Survival of P14 T cells was determined by comparing day 6 cell numbers with day 3 cell numbers (% of day 3). Cell numbers at day 6 are indicated above the histograms. Pooled results from two to three independent experiments are shown. Each dot represents data from one mouse. (B) PBS KO stands for a pool of the Downloaded from three KO mice (two per KO). http://www.jimmunol.org/

scribed (21). At day 4 after injection, OT-1 cells have undergone letion is dependent on p38a expression in CD11c+ cells (Fig. 6A; proliferation in both WT and p38del-CD11c hosts (Supplemental Supplemental Fig. 2E). To further confirm those results, CD11c- Fig. 2A–D). Mice were infected 4 wk posttreatment to expand Cre p38afl/fl were crossed onto RIP-OVA mice, which express the surviving OT-1 T cells for easier detection by flow cyto- OVA under the RIP in the b cells of the pancreas. It has been metry. In PBS-treated mice, OT-1 T cells represented ∼1% of demonstrated in this model that deletion of OT-1 cells requires CD8 T cells after infection-induced proliferation (Fig. 6A; cross-presentation of OVA by CD11c DCs (12). To measure the by guest on September 25, 2021 Supplemental Fig. 2E). In mice treated with ILA-SIINFEKL, this efficiency of peripheral deletion, we infected mice with vaccinia- percentage was significantly reduced. Numbers of OT-1 detected SIINFEKL 4 wk after OT-1 transfer to expand the numbers of OT- in CD11c-Cre Mapk14fl/fl mice were similar to the numbers ob- 1 cells for detection. Very few OT-1 cells were recovered from served without treatment, strongly suggesting that peripheral de- inguinal LNs, pLNs, or spleens from WT RIP-OVA mice (Fig. 6B;

FIGURE 5. Host p38 deficiency rescues CD8+ T cells from LDA-induced deletion. WT (A) and (B) or p38adel (C)and(D) P14 Thy1.1 CD8 T cells (106) were adoptively transferred into B6 WT or p38adel hosts. For 3 days, host mice received daily i.v. in- jections of PBS or 0.2 mg (LDA) of Gp33m peptide. (A and C) Mice were sacrificed on days 3 and 6, and P14 T cell numbers in the spleen were determined by flow cytometry. (B and D) On day 30, mice were challenged with LCMV-Armstrong i.p., and 6 d later mice were sacrificed and P14 T cell numbers in the spleen were determined by flow cytometry. Pooled results for three independent experiments are shown in (A)–(D). **p , 0.01, ***p , 0.001. The Journal of Immunology 7

FIGURE 6. p38 deficiency in DCs rescues CD8+ T cells from peripheral deletion. (A) OT-1 Thy1.1 CD8 T cells (104) were adoptively transferred into B6 WT or CD11c-Cre Mapk14flox/flox hosts. The day after, mice were injected with ILA-SIINFEKL (5 mg)–loaded splenocytes (30 3 106). Mice were infected 4 wk later with vaccinia-OVA, and spleens were analyzed 7 d postinfection. Pooled data from three experiments are shown. (B) OT-1 Thy1.1 CD8 T cells (3 3 106) were adoptively transferred into RIP-OVAhigh 3 CD11c-Cre Mapk14flox/flox mice or control littermates. After 4 wk, the indicated organs (pLNs [pancLN] and inguinal LNs [iLN]) were analyzed by flow cytometry. Pooled data from two independent experiments are shown. *p , 0.05, **p , 0.01, ***p , 0.001.

Supplemental Fig. 2F). In contrast, a significantly greater per- tion of DCs by upregulating p38 MAPK activity in DC precursors centage of OT-1 cells was still present in those organs from in both murine tumor models and cancer patients (35, 36). DCs fl/fl

CD11c-Cre Mapk14 RIP-OVA mice. Taken together, these re- from melanoma-bearing mice secreted significantly more IL-10 Downloaded from sults demonstrate that p38a expression in DCs is necessary for the and less IL-12p70, and showed a decreased capacity to activate induction of peripheral deletion of CD8 T cells. T cells compared with DCs from tumor-free animals. This cy- tokine profile was linked to the strong activation of p38 MAPK Discussion in those DCs (37). A more recent study shows that the inhibition It is well established that the strength of TCR signal as determined of p38 MAPK resulted in decreased PPARg expression in DCs, thus reducing the functional inhibition of p50 transcriptional by the amount of Ag (7–9) and the affinity for Ag (15) are critical http://www.jimmunol.org/ factors to direct a CD8 T cell toward anergy or peripheral deletion. activities by PPARg. The activation of p50 upregulated surface The proximal events that translate those differences into various expression of OX40L on DCs, which increases immunostimu- cellular signaling pathways and their consequences on the fate of latory potency of these DCs to activate Ag-specific effector the CD8 T cells are still poorly characterized. We hypothesized T cells and to inhibit Treg conversion and function, thus facili- that MAPK signaling could affect T cell survival. Using an tating tumor rejection (38). Altogether, p38 deletion can have in vitro/in vivo model for T cell tolerance, we established that p38, multiple effects on DCs. One of these parameters, or a combi- but not ERK or JNK signaling, was involved in peripheral dele- nation of them, could be responsible for the rescue in deletional tion. We also showed that p38 inhibition could rescue CD8 T cells tolerance observed in our models. Considering this complexity, from Bim-dependent apoptosis. Activation of the MAPK kinase further investigations are required to discriminate between these by guest on September 25, 2021 p38 signaling pathway had previously been reported to selectively various mechanisms in DCs and their effects on T cell deletional induce apoptosis in CD8 T cells in vivo (23). Furthermore, T cell– tolerance. endogenous p38 has been reported to be required for Fas-induced death in CD8 T cells (22). However, through the use of various Acknowledgments mouse strains deficient in components of the p38 signaling path- We thank Anne-Marie Schmitt-Verhulst and Daniel Speiser for support. We ways, we found that the role of p38 during T cell peripheral de- thank Jonathan Ashwell for providing reagents and for advice. We thank letion is not intrinsic to T cells, but rather an effect mediated by Sara Trabanelli for critical reading of the manuscript. DCs. To our knowledge, this is the first report to implicate p38 signaling extrinsic to the T cell in deletional tolerance. Our current Disclosures hypothesis is that activation of the p38 pathway supports periph- The authors have no financial conflicts of interest. eral deletion in an APC population, possibly by sustaining a tol- erogenic profile in the DC. The role of p38 in DCs has been studied in different contexts. References During immune tolerance it was shown that p38a is constitutively 1. Liu, G. Y., P. J. Fairchild, R. M. Smith, J. R. Prowle, D. Kioussis, and activated in cross-presenting DCs from the mesenteric LNs, D. C. Wraith. 1995. Low avidity recognition of self-antigen by T cells permits escape from central tolerance. Immunity 3: 407–415. leading to the generation of induced Tregs and inhibition of Th1 2. Fontenot, J. D., M. A. Gavin, and A. Y. Rudensky. 2003. Foxp3 programs the T cells through a TGF-b–dependent mechanism (32). The loss of development and function of CD4+CD25+ regulatory T cells. Nat. Immunol. 4: p38a signaling in DCs was reported to result in impaired induc- 330–336. 3. Thornton, A. M., and E. M. Shevach. 1998. CD4+CD25+ immunoregulatory tion of oral tolerance and generation of Ag-specific Tregs in vivo. T cells suppress polyclonal T cell activation in vitro by inhibiting interleukin 2 Another study showed that the loss of p38a protected mice from production. J. Exp. Med. 188: 287–296. 4. Itoh, M., T. Takahashi, N. Sakaguchi, Y. Kuniyasu, J. Shimizu, F. Otsuka, and development of autoimmune encephalomyelitis, and demonstrated S. Sakaguchi. 1999. Thymus and autoimmunity: production of CD25+CD4+ that p38a acts in DCs to drive Th17 differentiation and inflam- naturally anergic and suppressive T cells as a key function of the thymus in mation, by reciprocal regulation of IL-6 and IL-27 (33). A very maintaining immunologic self-tolerance. J. Immunol. 162: 5317–5326. 5. Curtsinger, J. M., and M. F. Mescher. 2010. Inflammatory cytokines as a third recent paper reported that p38a has a role in Ag processing by signal for T cell activation. Curr. Opin. Immunol. 22: 333–340. conventional CD8+ DCs, whereas p38a has no function in Ag 6. Hernandez, J., S. Aung, W. L. Redmond, and L. A. Sherman. 2001. Phenotypic uptake by these DCs or their costimulation for T cells. This study and functional analysis of CD8(+) T cells undergoing peripheral deletion in response to cross-presentation of self-antigen. J. Exp. Med. 194: 707–717. also showed that p38a inactivation led to reduced production of 7. Redmond, W. L., and L. A. Sherman. 2005. Peripheral tolerance of CD8 IL-12p40 and IL-12p70 by conventional DCs (34). T lymphocytes. Immunity 22: 275–284. 8. Redmond, W. L., B. C. Marincek, and L. A. Sherman. 2005. Distinct require- In cancer-associated studies, it was established that tumor- ments for deletion versus anergy during CD8 T cell peripheral tolerance in vivo. derived suppressive factors inhibit the differentiation and func- J. Immunol. 174: 2046–2053. 8 REGULATION OF CD8 PERIPHERAL DELETION BY p38 MAPK

9. Schwartz, R. H. 2003. T cell anergy. Annu. Rev. Immunol. 21: 305–334. mediated BIM phosphorylation potentiates BAX-dependent apoptosis. Neuron 10. Rocha, B., C. Tanchot, and H. Von Boehmer. 1993. Clonal anergy blocks in vivo 38: 899–914. growth of mature T cells and can be reversed in the absence of antigen. J. Exp. 25. Ley, R., K. E. Ewings, K. Hadfield, and S. J. Cook. 2005. Regulatory phos- Med. 177: 1517–1521. phorylation of Bim: sorting out the ERK from the JNK. Cell Death Differ. 12: 11. Stamou, P., J. de Jersey, D. Carmignac, C. Mamalaki, D. Kioussis, and 1008–1014. B. Stockinger. 2003. Chronic exposure to low levels of antigen in the periphery 26. Salvador, J. M., P. R. Mittelstadt, T. Guszczynski, T. D. Copeland, causes reversible functional impairment correlating with changes in CD5 levels H. Yamaguchi, E. Appella, A. J. Fornace, Jr., and J. D. Ashwell. 2005. Alter- in monoclonal CD8 T cells. J. Immunol. 171: 1278–1284. native p38 activation pathway mediated by T cell receptor-proximal tyrosine 12. Kurts, C., H. Kosaka, F. R. Carbone, J. F. Miller, and W. R. Heath. 1997. Class I- kinases. Nat. Immunol. 6: 390–395. restricted cross-presentation of exogenous self-antigens leads to deletion of 27. Lu, B., H. Yu, C. Chow, B. Li, W. Zheng, R. J. Davis, and R. A. Flavell. 2001. autoreactive CD8(+) T cells. J. Exp. Med. 186: 239–245. GADD45gamma mediates the activation of the p38 and JNK MAP kinase 13. Morgan, D. J., C. Kurts, H. T. Kreuwel, K. L. Holst, W. R. Heath, and pathways and cytokine production in effector TH1 cells. Immunity 14: 583–590. L. A. Sherman. 1999. Ontogeny of T cell tolerance to peripherally expressed 28. Ashwell, J. D. 2006. The many paths to p38 mitogen-activated protein kinase antigens. Proc. Natl. Acad. Sci. USA 96: 3854–3858. activation in the immune system. Nat. Rev. Immunol. 6: 532–540. 14. Safford, M., S. Collins, M. A. Lutz, A. Allen, C. T. Huang, J. Kowalski, 29. Hirasawa, K., A. Kim, H. S. Han, J. Han, H. S. Jun, and J. W. Yoon. 2003. Effect A. Blackford, M. R. Horton, C. Drake, R. H. Schwartz, and J. D. Powell. 2005. of p38 mitogen-activated protein kinase on the replication of encephalomyo- Egr-2 and Egr-3 are negative regulators of T cell activation. Nat. Immunol. 6: carditis virus. J. Virol. 77: 5649–5656. 472–480. 30. Behrens, G. M., M. Li, G. M. Davey, J. Allison, R. A. Flavell, F. R. Carbone, and 15. Smith, T. R., G. Verdeil, K. Marquardt, and L. A. Sherman. 2014. Contribution W. R. Heath. 2004. Helper requirements for generation of effector CTL to islet of TCR signaling strength to CD8+ T cell peripheral tolerance mechanisms. J. beta cell antigens. J. Immunol. 172: 5420–5426. Immunol. 193: 3409–3416. 31. Davey, G. M., C. Kurts, J. F. Miller, P. Bouillet, A. Strasser, A. G. Brooks, 16. Redmond, W. L., C. H. Wei, H. T. Kreuwel, and L. A. Sherman. 2008. The F. R. Carbone, and W. R. Heath. 2002. Peripheral deletion of autoreactive CD8 apoptotic pathway contributing to the deletion of naive CD8 T cells during the T cells by cross presentation of self-antigen occurs by a Bcl-2-inhibitable induction of peripheral tolerance to a cross-presented self-antigen. J. Immunol. pathway mediated by Bim. J. Exp. Med. 196: 947–955. 180: 5275–5282. 32. Huang, G., Y. Wang, and H. Chi. 2013. Control of T cell fates and immune tolerance 17. Parish, I. A., S. Rao, G. K. Smyth, T. Juelich, G. S. Denyer, G. M. Davey, by p38a signaling in mucosal CD103+ dendritic cells. J. Immunol. 191: 650–659. Downloaded from A. Strasser, and W. R. Heath. 2009. The molecular signature of CD8+ T cells 33. Huang, G., Y. Wang, P. Vogel, T. D. Kanneganti, K. Otsu, and H. Chi. 2012. undergoing deletional tolerance. Blood 113: 4575–4585. Signaling via the kinase p38a programs dendritic cells to drive TH17 differ- 18. Morgan, D. J., R. Liblau, B. Scott, S. Fleck, H. O. McDevitt, N. Sarvetnick, entiation and autoimmune inflammation. Nat. Immunol. 13: 152–161. D. Lo, and L. A. Sherman. 1996. CD8(+) T cell-mediated spontaneous diabetes 34. Zhou, Y., J. Wu, C. Liu, X. Guo, X. Zhu, Y. Yao, Y. Jiao, P. He, J. Han, and in neonatal mice. J. Immunol. 157: 978–983. L. Wu. 2016. p38a has an important role in antigen cross-presentation by den- 19. Nishida, K., O. Yamaguchi, S. Hirotani, S. Hikoso, Y. Higuchi, T. Watanabe, dritic cells. Cell. Mol. Immunol. DOI: 10.1038/cmi.2016.49. T. Takeda, S. Osuka, T. Morita, G. Kondoh, et al. 2004. p38alpha mitogen- 35. Wang, S., S. Hong, J. Yang, J. Qian, X. Zhang, E. Shpall, L. W. Kwak, and Q. Yi.

activated protein kinase plays a critical role in cardiomyocyte survival but not 2006. Optimizing immunotherapy in multiple myeloma: restoring the function of http://www.jimmunol.org/ in cardiac hypertrophic growth in response to pressure overload. Mol. Cell. Biol. patients’ monocyte-derived dendritic cells by inhibiting p38 or activating MEK/ 24: 10611–10620. ERK MAPK and neutralizing interleukin-6 in progenitor cells. Blood 108: 4071– 20. Wong, S. B., R. Bos, and L. A. Sherman. 2008. Tumor-specific CD4+ T cells 4077. render the tumor environment permissive for infiltration by low-avidity CD8+ 36. Oosterhoff, D., S. Lougheed, R. van de Ven, J. Lindenberg, H. van Cruijsen, T cells. J. Immunol. 180: 3122–3131. L. Hiddingh, J. Kroon, A. J. van den Eertwegh, B. Hangalapura, R. J. Scheper, 21. Wei, C. H., and L. A. Sherman. 2007. N-terminal trimer extension of nominal and T. D. de Gruijl. 2012. Tumor-mediated inhibition of human dendritic cell CD8 T cell epitopes is sufficient to promote cross-presentation to cognate CD8 differentiation and function is consistently counteracted by combined p38 T cells in vivo. J. Immunol. 179: 8280–8286. MAPK and STAT3 inhibition. OncoImmunology 1: 649–658. 22. Farley, N., G. Pedraza-Alva, D. Serrano-Gomez, V. Nagaleekar, A. Aronshtam, 37. Armbruster, N. S., J. R. Richardson, J. Schreiner, J. Klenk, M. Gunter,€ T. Krahl, T. Thornton, and M. Rinco´n. 2006. p38 mitogen-activated protein ki- D. Kretschmer, S. Po¨schel, K. Schenke-Layland, H. Kalbacher, K. Clark, and nase mediates the Fas-induced mitochondrial death pathway in CD8+ T cells. S. E. Autenrieth. 2016. PSM peptides of Staphylococcus aureus activate the p38-

Mol. Cell. Biol. 26: 2118–2129. CREB pathway in dendritic cells, thereby modulating cytokine production and by guest on September 25, 2021 23. Merritt, C., H. Enslen, N. Diehl, D. Conze, R. J. Davis, and M. Rinco´n. 2000. T cell priming. J. Immunol. 196: 1284–1292. Activation of p38 mitogen-activated protein kinase in vivo selectively induces 38. Lu, Y., M. Zhang, S. Wang, B. Hong, Z. Wang, H. Li, Y. Zheng, J. Yang, apoptosis of CD8(+) but not CD4(+) T cells. Mol. Cell. Biol. 20: 936–946. R. E. Davis, J. Qian, et al. 2014. p38 MAPK-inhibited dendritic cells induce 24. Putcha, G. V., S. Le, S. Frank, C. G. Besirli, K. Clark, B. Chu, S. Alix, superior antitumour immune responses and overcome regulatory T-cell-mediated R. J. Youle, A. LaMarche, A. C. Maroney, and E. M. Johnson, Jr.. 2003. JNK- immunosuppression. Nat. Commun. 5: 4229.