Signal 3 requirement for memory CD8 1 T-cell activation is determined by the infectious pathogen

Similar documents
SUPPLEMENTARY INFORMATION

Naive and memory CD8 T cell responses after antigen stimulation in vivo

Therapeutic PD L1 and LAG 3 blockade rapidly clears established blood stage Plasmodium infection

Lineage relationship and protective immunity of memory CD8 T cell subsets

Out-of-sequence signal 3 as a mechanism for virusinduced immune suppression of CD8 T cell responses

IMMUNOLOGICAL MEMORY. CD4 T Follicular Helper Cells. Memory CD8 T Cell Differentiation

Supplementary Figure S1. PTPN2 levels are not altered in proliferating CD8+ T cells. Lymph node (LN) CD8+ T cells from C57BL/6 mice were stained with

Supplementary Figure 1. Characterization of basophils after reconstitution of SCID mice

NK1.1 þ CD8 þ T cells escape TGF-b control and contribute to early microbial pathogen response

IMMUNOLOGICAL MEMORY. May 28, 2017

MATERIALS AND METHODS. Neutralizing antibodies specific to mouse Dll1, Dll4, J1 and J2 were prepared as described. 1,2 All

Micro 204. Cytotoxic T Lymphocytes (CTL) Lewis Lanier

Peer review correspondence

Supplemental Figure 1. Activated splenocytes upregulate Serpina3g and Serpina3f expression.

Supplemental Figure 1

During acute infections, antigen-specific CD8 T cells are activated

Supplementary Figures

Memory NK cells during mousepox infection. Min Fang, Ph.D, Professor Institute of Microbiology, Chinese Academy of Science

Blocking antibodies and peptides. Rat anti-mouse PD-1 (29F.1A12, rat IgG2a, k), PD-

T-cell activation T cells migrate to secondary lymphoid tissues where they interact with antigen, antigen-presenting cells, and other lymphocytes:

T-cell activation T cells migrate to secondary lymphoid tissues where they interact with antigen, antigen-presenting cells, and other lymphocytes:

Effector mechanisms of cell-mediated immunity: Properties of effector, memory and regulatory T cells

CD4 T-Cell Memory Generation and Maintenance

Immunological memory.

Scott Abrams, Ph.D. Professor of Oncology, x4375 Kuby Immunology SEVENTH EDITION

Recombinant adeno-associated virus vectors induce functionally impaired transgene product specific CD8 + T cells in mice

Tracking total polyclonal CD8 T cell responses in inbred and outbred hosts after infection

Dendritic cell subsets and CD4 T cell immunity in Melanoma. Ben Wylie 1 st year PhD Candidate

Protective Capacity of Memory CD8 + TCells Is Dictated by Antigen Exposure History and Nature of the Infection

T cell memory Friday, February 10, 17

Recombinant adeno-associated virus vectors induce functionally impaired transgene product specific CD8 + T cells in mice

Origin of CD8 + Effector and Memory T Cell Subsets

NIH Public Access Author Manuscript Nature. Author manuscript; available in PMC 2009 July 28.

Supplemental Figure 1. Signature gene expression in in vitro differentiated Th0, Th1, Th2, Th17 and Treg cells. (A) Naïve CD4 + T cells were cultured

NK cell flow cytometric assay In vivo DC viability and migration assay

SUPPLEMENTARY INFORMATION

Supplemental Table I.

SUPPLEMENTARY INFORMATION

Antigen Presentation and T Lymphocyte Activation. Abul K. Abbas UCSF. FOCiS

Carmen Gerlach 1, Jeroen W.J van Heijst 1 and Ton N.M. Schumacher 1

Natural Killer Cells: Development, Diversity, and Applications to Human Disease Dr. Michael A. Caligiuri

Blimp-1 Transcription Factor Is Required for the Differentiation of Effector CD8 + T Cells and Memory Responses

CD8 T cells leave the thymus 4 5 days after undergoing positive

A Slfn2 mutation causes lymphoid and myeloid immunodeficiency due to loss of immune cell quiescence

Supplementary Figure 1

Essential role of the Wnt pathway effector Tcf-1 for the establishment of functional CD8 T cell memory

Supplemental Information. Gut Microbiota Promotes Hematopoiesis to Control Bacterial Infection. Cell Host & Microbe, Volume 15

Inflammatory IL-15 is required for optimal memory T cell responses

Dendritic cells in cancer immunotherapy Aimin Jiang

Resolution of a chronic viral infection after interleukin-10 receptor blockade

Scott Abrams, Ph.D. Professor of Oncology, x4375 Kuby Immunology SEVENTH EDITION

Control of Virus-Specific CD8 T-Cell Exhaustion and Immune-Mediated Pathology by E3 Ubiquitin Ligase Cbl-b during Chronic Viral Infection

Nature Immunology: doi: /ni Supplementary Figure 1. Gene expression profile of CD4 + T cells and CTL responses in Bcl6-deficient mice.

The encephalitogenicity of TH17 cells is dependent on IL-1- and IL-23- induced production of the cytokine GM-CSF

B220 CD4 CD8. Figure 1. Confocal Image of Sensitized HLN. Representative image of a sensitized HLN

TITLE: MODULATION OF T CELL TOLERANCE IN A MURINE MODEL FOR IMMUNOTHERAPY OF PROSTATIC ADENOCARCINOMA

CD40-Activated B Cells Can Efficiently Prime Antigen- Specific Naïve CD8 + T Cells to Generate Effector but Not Memory T cells

Reprogramming of antiviral T cells prevents inactivation and restores T cell activity during persistent viral infection

Memory CD4 T Cells Enhance Primary CD8 T-Cell Responses

Canberra, Australia). CD11c-DTR-OVA-GFP (B6.CD11c-OVA), B6.luc + and. Cancer Research Center, Germany). B6 or BALB/c.FoxP3-DTR-GFP mice were

Increased IL-12 induced STAT-4 signaling in CD8 T cells. from aged mice

Medical Virology Immunology. Dr. Sameer Naji, MB, BCh, PhD (UK) Head of Basic Medical Sciences Dept. Faculty of Medicine The Hashemite University

Supplementary Figure 1. Normal T lymphocyte populations in Dapk -/- mice. (a) Normal thymic development in Dapk -/- mice. Thymocytes from WT and Dapk

IL-15 regulates memory CD8 + T cell O-glycan synthesis and affects trafficking

Pathogenic virus-specific T cells cause disease during treatment with the calcineurin inhibitor FK506: implications for transplantation

Chapter 22: The Lymphatic System and Immunity

New insights into CD8+ T cell function and regulation. Pam Ohashi Princess Margaret Cancer Centre

ABSTRACT. Title of Document: GENERATING FUNCTIONAL CYTOTOXIC T LYMPHOCYTES THROUGH REPETITIVE PEPTIDE BOOSTING.

L-selectin Is Essential for Delivery of Activated CD8 + T Cells to Virus-Infected Organs for Protective Immunity

RAISON D ETRE OF THE IMMUNE SYSTEM:

NKTR-255: Accessing The Immunotherapeutic Potential Of IL-15 for NK Cell Therapies

Supplementary Figure 1. mrna expression of chitinase and chitinase-like protein in splenic immune cells. Each splenic immune cell population was

TGF-β receptor maintains CD4 T helper cell identity during chronic viral infections

Differentiation and Protective Capacity of Virus- Specific CD8 + T Cells Suggest Murine Norovirus Persistence in an Immune-Privileged Enteric Niche

Tolerance, autoimmunity and the pathogenesis of immunemediated inflammatory diseases. Abul K. Abbas UCSF

Absence of mouse 2B4 promotes NK cell mediated killing of activated CD8 + T cells, leading to prolonged viral persistence and altered pathogenesis

Proteasome activity regulates CD8 + T lymphocyte metabolism and fate specification

TCR, MHC and coreceptors

Unexpected positive control of NFκB and mir-155 by DGKα and ζ ensures effector and memory CD8 + T cell differentiation

chapter 17: specific/adaptable defenses of the host: the immune response

of whole cell cultures in U-bottomed wells of a 96-well plate are shown. 2

Immune response to infection

Nature Immunology: doi: /ni Supplementary Figure 1. Id2 and Id3 define polyclonal T H 1 and T FH cell subsets.

CD4 + T cells recovered in Rag2 / recipient ( 10 5 ) Heart Lung Pancreas

Supplemental Figure 1. Protein L

Inflammation Directs Memory Precursor and Short-Lived Effector CD8 + T Cell Fates via the Graded Expression of T-bet Transcription Factor

General Overview of Immunology. Kimberly S. Schluns, Ph.D. Associate Professor Department of Immunology UT MD Anderson Cancer Center

HD1 (FLU) HD2 (EBV) HD2 (FLU)

The Adaptive Immune Response. B-cells

Cell-mediated Immunity

Determinants of Immunogenicity and Tolerance. Abul K. Abbas, MD Department of Pathology University of California San Francisco

days days and gbt-i.cd Recipient 20

Shenghua Zhou, Rong Ou, Lei Huang, and Demetrius Moskophidis*

Programmed death-1 (PD-1) defines a transient and dysfunctional oligoclonal T cell population in acute homeostatic proliferation

Time-dependent alterations in memory CD8 T cell function after infection

and Nucleic Acid-Sensing TLR Pathways in Controlling a Chronic Viral Infection

Recombinant Vaccinia Virus-Induced T-Cell Immunity: Quantitation of the Response to the Virus Vector and the Foreign Epitope

ILC1 and ILC3 isolation and culture Following cell sorting, we confirmed that the recovered cells belonged to the ILC1, ILC2 and

Supplementary Figures

Supplementary Information. Tissue-wide immunity against Leishmania. through collective production of nitric oxide

Transcription:

3176 DOI 1.12/eji.21141537 Eur. J. Immunol. 211. 41: 3176 3186 Signal 3 requirement for memory CD8 1 T-cell activation is determined by the infectious pathogen Selina J. Keppler 1,2 and Peter Aichele 1 1 Institute for Medical Microbiology and Hygiene, Department of Immunology, University of Freiburg, Freiburg, Germany 2 University of Freiburg, Faculty of Biology, Freiburg, Germany The relevance of direct inflammatory signals (signal 3) for the activation of memory CD8 1 T cells during recall responses is so far unknown. We therefore investigated the direct impact of IL-12 and type I IFN on the formation, recall potential and protective capacity of memory T cells. Using CD8 1 T cells deficient for IL-12 or type I IFN receptors in an adoptive transfer system, we generated memory populations after infection with vaccinia virus, lymphocytic choriomeningitis virus or Listeria monocytogenes. The results demonstrate that in the absence of signal 3 cytokines during primary infection, functional memory T cells were formed. After retransfer into naïve mice, signal 3-deficient memory T cells were able to specifically lyse target cells in vivo under non-infectious conditions. However, after reinfection, secondary effector CD8 1 T cells lacking signal 3 were impaired in expansion and protective capacity dependent on the nature of the pathogen. We conclude that memory CD8 1 T cells depend on a signal 3 for expansion, independent of signals obtained during priming, thereby being influenced by the pathogen-induced inflammatory milieu during secondary infection. In summary, our results reveal an essential role for direct inflammatory cytokine signaling in secondary T-cell responses. Key words: IL-12. Infection. Memory T cells. Signal 3. Type I IFN Supporting Information available online Introduction The ability of the adaptive immune system to remember pathogens via the generation of memory cells improves the host immune response in the case of reinfection with the same pathogen and is the basis for protective immunity after vaccination. The relationship between effector and memory T cells is still a matter of debate although it is widely accepted that memory T cells transit through an effector phase after Correspondence: Dr. Peter Aichele e-mail: peter.aichele@uniklinik-freiburg.de antigen contact [1 3]. Multiple signals determining which effector T cells survive to become memory cells have been proposed: asymmetric cell division, TCR signaling strength, duration of antigenic stimulation, metabolic fitness, and availability of proinflammatory cytokines and growth or survival factors [4 8]. Investigating the requirements for primary activation of naïve CD8 1 T cells is therefore crucial to understanding the effector-to-memory transition. In recent years, it became clear that, during acute infection, CD8 1 T-cell fate is regulated by three signals: TCR engagement (signal 1), costimulation (signal 2), and an inflammatory stimulus (signal 3) via cytokines such as IL-12 or type I interferons (type I IFN) [9 11]. Direct signaling via IL-12 and type I IFN receptors thereby favors terminal differentiation of effector CD8 1 T cells, & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Immunol. 211. 41: 3176 3186 Cellular immune response 3177 which demonstrate high cytolytic activity, produce IFN-g but are limited in survival. This subpopulation of terminal effector CD8 1 T cells expresses killer cell lectin-like receptor G1 (KLRG1) downregulates CD127 and is not able to establish memory after clearance of the infection. In contrast, memory precursor effector CD8 1 T cells demonstrating comparable cytolytic activity and IFN-g expression but producing IL-2 and exhibiting a phenotypically more naïve status (low levels of KLRG1 and high levels of CD127 expression) are supposed to survive and establish memory [3, 12]. We and others demonstrated that CD8 1 T cells with deficiencies in receptors for IL-12 or type I IFN mainly differentiate into KLRG1 lo /CD127 hi memory precursor effector T cells [13 16]. The importance of type I IFN versus IL-12 in a given T-cell response is thereby dictated by the pathogen-induced cytokine milieu. Interestingly, a recent study could demonstrate that direct type I IFN signaling to CD8 1 T cells substituted for T-cell help during vaccinia virus (VV) infection. In this experimental setting, type I IFN replaces T-cell help by directly supporting survival and differentiation of virus-specific CD8 1 T cells independent of IL-2 and IL-15 signals [17]. It was suggested that the preferential differentiation of CD8 1 T cells into memory precursors in the absence of signal 3 cytokines during priming leads to higher numbers and more active memory T cells. In line with this, it was demonstrated that missing the IL-12 signal during priming favors memory CD8 1 T-cell formation after infection [18]. However, this finding is so far controversial, since another study suggested that programming of memory T-cell development strictly requires signal 3 in the context of VV or Listeria infections and thus, in the absence of signal 3, no memory T-cell formation occurs [19]. In contrast, we showed recently that CD8 1 T cells lacking IL-12 or type I IFN signals were able to form stable memory T-cell populations [13, 15]. Altogether, a conclusive study addressing the influence of signal 3 availability during priming on memory formation of T cells is missing. Memory CD8 1 T cells have been classified into CD44 hi / CD62L lo /CCR7 lo effector memory T cells (T EM ), that exhibit high granzyme B expression levels, direct ex vivo cytolytic activity and an increased capacity to migrate to peripheral tissues, and CD44 hi /CD62L hi /CCR7 hi central memory T (T CM ) cells, that have strong proliferative capacities, produce IL-2, and reside mainly in lymph nodes [2]. Memory CD8 1 T cells were shown to be more efficient in response to antigens compared with naïve CD8 1 T cells, thereby demonstrating rapid proliferation, an accelerated cytolytic activity and cytokine production. Based on this faster recall potential and on in vitro activation and in vivo peptide immunization studies, it was proposed that memory T cells are independent of a third, inflammatory signal for recall responses [1, 21]. However, the direct influence of the proinflammatory cytokines IL-12 and type I IFN on the activation and recall capacity of memory/secondary effector CD8 1 T cells during infections has not been addressed so far. In this study, we analyzed the influence of IL-12 and type I IFN on formation, recall potential and protective capacity of memory/secondary effector T cells. Using CD8 1 T cells with defined antigen-specificity lacking receptors for IL-12 or type I IFN in an adoptive transfer (AdTf) system, we generated memory populations after infection with VV, lymphocytic choriomeningitis virus (LCMV), or Listeria. We first analyzed the formation of memory cell populations, demonstrating that the presence or absence of signal 3 cytokines does not prevent memory formation and function of CD8 1 T cells, even in the situations where primary T-cell expansion was limited. Second, we investigated signal 3 dependency for recall and protective capacity of memory/secondary effector CD8 1 T cells. We found that reactivation of memory CD8 1 T cells is controlled by signal 3 cytokines for expansion and protective capacity, dependent on the nature of the challenge infection. Results CD8 1 T cells form functional memory cells independent of type I IFN after VV infection As previously described, CD8 1 T cells lacking signal 3 cytokine receptors were able to establish memory populations after AdTf and infection with different pathogens even in the situation when expansion of T cells in the primary response was limited. First, we analyzed the contribution of signal 3 cytokines to memory formation of P14 T cells competent (P14.WT) or deficient for type I IFN signaling (P14.IFNARKO) after infection with recombinant vaccinia virus expressing LCMV glycoprotein (rvv GP ) (expressing the LCMV-GP33 epitope). This infection model was chosen since CD8 1 T cells are largely independent of type I IFN signals for expansion after primary VV infection (Fig. 1A, [13]) and thus, WT and IFNARKO memory T cells are supposed to display similar characteristics and functions. Forty days after primary infection P14.WT or P14.IFNARKO memory T cells constituted 7 8% of total CD8 1 T cells in the blood (Fig. 1A). To distinguish different memory T-cell subsets, we examined expression of KLRG1 and CD127, as well as CD62L and CD44 (Fig. 1B and C). Only CD8 1 T cells expressing high levels of CD127 survived and formed memory cells (9% of P14.WT and P14.IFNARKO T cells), with 25% of P14.WT and 19% of P14.IFNARKO T cells still expressing KLRG1 (Fig. 1B and C), reflecting the slightly impaired KLRG1 expression on P14.IFNARKO T cells during primary infection (our unpublished results). No differences in the development into central or effector memory T cells were observed, as 3% of P14.WT and P14.IFNARKO T cells exhibited a T CM and 7% a T EM phenotype. Furthermore, the expression of the transcription factors T-box transcription factor (T-bet) and Eomes was analyzed. About 95% of both P14 memory T-cell populations expressed T-bet after VV infection, with comparable expression levels on a per cell basis (see MFI, Fig. 1C). After VV infection, 3% of P14.WT and 25% of P14.IFNARKO memory T cells expressed Eomes (Fig. 1C). The comparison of cytokine production after short-term peptide restimulation in vitro revealed that 8% of both memory & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3178 Eur. J. Immunol. 211. 41: 3176 3186 A 6 5 4 3 2 1 d4 1 8 6 4 2 d4 B KLRG1 P14.WT P14.IFNARKO 3.7 22.9.91 12.9 1.4 72.9.82 85.4 CD127 C T cell differentiation % KLRG1 hi of P14 5 4 3 2 1 % CD127 hi of P14 1 8 6 4 2 % CD44hi CD62L hi of P14 5 4 3 2 1 % CD44hi CD62L lo of P14 1 8 6 4 2 D % T-bet of P14 1 8 6 4 2 MFI T-bet 2 18 16 14 12 1 8 6 4 2 5 6 4 5 4 3 3 2 2 of P14 % Eomes 1 effector functions mes MFI Eom 1 % IFN-γ of P14 1 8 6 4 2 25 2 15 1 5 % TNF-α of P14 1 8 6 4 2 MFI TNF-α 7 6 5 4 3 2 1 14 % IL-2 of P1 5 4 3 2 1-2 MFI IL- 12 1 8 6 4 2 % CD17a of P14 1 8 6 4 2 7a MFI CD1 12 1 8 6 4 2 Figure 1. Analysis of memory T cells day 4 after VV infection. (A) 1 5 naïve P14.WT () or P14.IFNARKO (J) T cells were adoptively transferred into B6 mice which were infected with 2 1 6 PFU rvv GP. Expansion of P14 T-cell populations was analyzed in the acute (day 4 p.i.; left) and in the memory phase (day 4 p.i.; right). Plots show percentages of P14 T cells of total CD8 1 T cells in the blood. (B) Analysis of KLRG1 and CD127 expression on P14.WT (left) and P14.IFNARKO (right) memory T cells in the spleen. Plots are gated on CD8 1, Thy1.1 1 (P14) cells. (C) Analysis of differentiation of splenic memory P14.WT (black bars and symbols) and P14.IFNARKO (white bars and symbols) T cells. Percentages of KLRG1 and CD127 positive as well as percentages of central (CD44 hi /CD62L hi ) and effector (CD44 hi /CD62L lo ) memory P14 T cells of total P14 T cells are depicted (top, left to right). Percentages of T-bet and Eomes-positive P14 T cells of total P14 memory T cells, as well as mean fluorescence intensity (MFI) of T-bet and Eomes of P14 T cells are indicated (bottom). (D) Effector functions of P14 memory T cells were analyzed after short-time restimulation ex vivo. Percentages of cytokine-positive P14 memory T cells of total P14 T cells, as well as MFI of cytokine expression are indicated. (A, C, D) Values are expressed as mean (1SEM), n 5 5. Results are representative of three independent experiments., not significant (Student s unpaired t-test). T-cell populations produced IFN-g and TNF-a, about 4% of both populations expressed IL-2 and almost all T cells degranulated (CD17a surface translocation). Interestingly, on a per cell basis, P14.IFNARKO memory T cells produced less IFN-g, TNF-a, and IL-2 (but not statistically significant) and were somewhat reduced in degranulation compared with P14.WT memory T cells (see MFI, Fig. 1D). These data indicate that CD8 1 T-cell memory formation was possible in the absence of type I IFN signaling after VV infection. Memory CD8 1 T cells depend on type I IFN as a signal 3 after LCMV recall infection In a first approach, we studied in vivo recall capacities and effector functions of memory T cells, lacking type I IFN signaling, in a retransfer system. We used memory P14.WT and P14.IFNARKO T cells generated in the context of VV infection since T cells were largely independent of type I IFN signals during primary infection (comparable expansion, differentiation, and & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Immunol. 211. 41: 3176 3186 Cellular immune response 3179 effector functions) and therefore leading to memory cells with the same experience. To examine the contribution of type I IFN signaling to memory CD8 1 T-cell recall responses, equal numbers of purified P14.WT and P14.IFNARKO memory T cells, generated by VV infections (Fig. 1), were transferred into naïve recipients (experimental setup, Fig. 2A). The transfer of equal numbers of memory T cells guaranteed a direct comparability of the two T-cell populations. To analyze in vivo cytolytic activity under non-infectious conditions, we performed an in vivo killer assay. After AdTf of P14.WT and P14.IFNARKO memory T cells, target cells were transferred into recipient mice. Splenocytes of H8 mice that ubiquitously express GP33 and are recognized by P14 T cells were labeled with a low amount of CFSE and used as target cells. Spleen cells of C57BL/6 (B6) mice were marked with a high concentration of CFSE and used as a control. In total, 28 h after transfer of a 1:1 mixture of the two target cell populations, the specific cytolytic activity of P14.WT and P14.IFNARKO memory T cells reduced the number of H8 target cells in the blood to 32% (from originally 5%, Fig. 2B). Thereafter, H8 target cells were gradually reduced, whereas control target cells were still detectable in the blood 66 h after transfer. From these results, we concluded that both memory T-cell populations exhibited comparable, cytolytic activity in vivo under non-infectious conditions independent of type I IFN-signaling. As memory CD8 1 T cells were thought to be independent of a third signal for re-expansion and survival, we next tested the recall capacity of P14.WT and P14.IFNARKO memory T cells in vivo. After AdTf of low numbers (2 1 4 ) of memory T cells, recipient mice were infected with either rvv GP or LCMV (Fig. 2C and D). We decided to use these two infection models because CD8 1 T cells were largely independent of type I IFN during primary VV infection, whereas fully depended on type I IFN signals for expansion during primary LCMV infection [13]. On day 5 after VV infection, analysis of the P14 memory T cells in the spleen revealed that P14.IFNARKO secondary A acute 1 5 naive CD8 T cells P14.WT or P14.IFNARKO (Thy1.1) B 28 hrs 42 hrs 66 hrs day 4 analysis VV infection (Fig. 1) memory CD8 T cells no transfer B6 H8 P14.wt CFSE memory B in vivo killer C/D viral infections E LCMV infection P14.IFNARKO E PFU/spleen 1 1 1 1 1 1 1 killing? expansion? protection? spleen * C % IFN-γ + of P14 % KLRG1 hi of P14 1 8 6 4 2 1 8 6 4 2 1 8 6 4 2 VV infection (d5) total P14 + per spleen 1 1 1 1 1 D 8 6 4 2 LCMV infection (d7) 1 1 8 8 % CD127 hi of P14 6 4 2 25 2 15 1 5 % KLRG1 hi of P14 % IFN-γ + of P14 6 4 2 1 8 6 4 2 * total P14 + per spleen % CD127 hi of P14 1 1 1 1 1 1 8 6 4 2 4 3 2 1 Figure 2. Functional analysis of VV-experienced P14.WT and P14.IFNARKO memory T cells. (A) Experimental setup. Forty days after AdTf and primary VV infection, MACS-purified P14.WT and P14.IFNARKO memory T cells (Thy1.1) were retransferred into B6 mice (Thy1.2) and analyzed for functionality. (B) Cytolytic activity of P14 memory T cells was determined by an in vivo cytotoxicity assay. In total, 2 1 5 purified P14.WT and P14.IFNARKO memory T cells were transferred into naïve B6. Spleen cells of H8 mice (CFSE low) and B6 mice (CFSE high) were used as target cells. Cytolytic activity was measured 28, 42, and 66 h after transfer of target cells by analyzing PBLs. Histograms show target cell populations, with numbers indicating percentages of CFSE-positive cells. (C, D) Analysis of re-expansion of VV-experienced P14 memory T cells. Briefly, 2 1 4 Thy1.1 1 P14.WT or P14.IFNARKO memory T cells were transferred into B6 mice (Thy1.2 1 ) followed by secondary infection with (C) rvv GP or (D) LCMV. Expansion of P14 T cells was analyzed in the spleen (C) at day 5 after rvv GP infection or (D) at day 7 after infection with LCMV. Percentages and absolute numbers of P14.WT () and P14.IFNARKO (J) T cells are shown (top). Percentages of KLRG1 and CD127 1 memory P14.WT (black bars) and P14.IFNARKO (white bars) T cells of total P14 T cells are shown (middle). Percentages of IFN-g-positive P14 T cells of total P14 memory T cells, as well as MFI of IFN-g of P14.WT (black bars and symbols) and P14.IFNARKO (white bars and symbols) memory T cells are indicated (bottom). (E) Protective capacity of P14 memory T cells was tested after AdTf of 2 1 5 P14.WT or P14.IFNARKO memory T cells in naïve B6, followed by infection with LCMV-WE. Viral titers in the spleen at day 4 after challenge are shown. Values are expressed as mean (1SEM), n 5 3. po.5, po.1, po.1,, not significant (Student s unpaired t-test). Results are representative of two independent experiments. & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

318 Eur. J. Immunol. 211. 41: 3176 3186 effector T cells were slightly reduced in expansion compared with P14.WT T cells, constituting only 2% of total CD8 1 T cells (compared with 5% of P14.WT T cells). This two to three-fold lower expansion was also seen in absolute cell numbers in the spleen. It is worth noting that both secondary effector T-cell populations demonstrated strongly diminished expansion compared with primary infection with VV, where both T-cell populations expanded to 45% of CD8 1 T cells in the blood (Fig. 1A). This impaired secondary expansion of VV-experienced memory T cells was also observed when higher numbers of P14.WT and P14.IFNARKO T cells were transferred and the challenging dose was increased (Supporting Information Fig. 1). One may speculate that the VV-induced inflammatory milieu during T-cell priming negatively influences the fitness of the memory T cells or that antigen availability during VV reinfection is limited leading to diminished memory T-cell triggering and expansion. Despite this difference in expansion, analysis of surface markers KLRG1 and CD127, as well as CD44 and CD62L revealed no differences between the two secondary effector T-cell populations after VV infection. About 7 8% of P14 secondary effector T cells expressed KLRG1 and 9% demonstrated a T EM phenotype (CD44 hi /CD62L lo, Fig. 2C and Supporting Information Fig. 2A), suggesting highly activated T cells after challenge infection. Additionally, P14.WT and P14.IFNARKO secondary effector T cells showed comparable cytokine production after short-time restimulation in vitro, with 7% of P14 T cells producing IFN-g, 2% of P14 T cells producing TNF-a and only 2 3% demonstrated IL-2 production (Fig. 2C and Supporting Information Fig. 2A). No differences were detectable in the degranulation capacity according to CD17a staining of both secondary effector T-cell populations (Supporting Information Fig. 2A). Thus, after secondary VV infection, CD8 1 T cells were dependent on type I IFN as a signal 3 for expansion, but were independent for differentiation into highly activated, cytokineproducing cells. After retransfer of equal numbers of memory T cells and LCMV challenge infection, P14.WT secondary effector T cells expanded to 6% of CD8 1 T cells, whereas P14.IFNARKO secondary effector T cells were severely impaired in their reexpansion, as seen in relative and total cell numbers on day 7 postinfection in the spleen (Fig. 2D). The drastically reduced expansion correlated with diminished KLRG1 expression, as only 4% of P14.IFNARKO secondary effector T cells expressed KLRG1 compared with 9% of P14.WT T cells (Fig. 2D). In addition, only 8% of P14.IFNARKO secondary effector T cells demonstrated a T EM phenotype (compared with 95% of P14.WT T cells) and were reduced in cytokine production. Less P14.IFNARKO T cells produced IFN-g, TNF-a, and IL-2 compared with P14.WT cells after short-time restimulation in vitro, demonstrating a less activated phenotype (Fig. 2D and Supporting Information Fig. 2B). Interestingly, on a per-cell basis, expression levels of all cytokines analyzed were higher with P14.IFNARKO secondary effector T cells, probably due to a compensatory mechanism of the few P14.IFNARKO T cells left for in vitro restimulation. To test the protective capacity of P14 secondary effector T cells in vivo, we adoptively transferred equal numbers of P14.WT and P14.IFNARKO memory T cells in B6 mice and challenged the recipient mice with low-dose LCMV-WE. Four days later, viral titers in the spleen of recipient mice were analyzed. P14.WT secondary effector T cells were able to clear the virus close to detection limit until day 4, whereas P14.IFNARKO T cells were markedly impaired in virus control (Fig. 2E). In summary, our results show that P14.IFNARKO secondary effector T cells are severely impaired in expansion and protective capacity during LCMV infection, although fully functional under non-infectious conditions or during VV recall infection. The reverse approach, transfer of memory P14.IFNARKO T cells that were initially activated with LCMV and then challenged with rvv GP is an obvious complementary experimental setting. However, such an experiment is technically not possible, since P14.IFNARKO T cells are drastically limited in expansion after primary LCMV infection (there are virtually no effector T cells present) and the frequencies of P14.IFNARKO memory T cells are too low for retransfer experiments. We concluded from these experiments that memory CD8 1 T cells are regulated by type I IFN as signal 3 for recall expansion dependent on the inflammatory milieu induced by different pathogens, probably irrespective of signals obtained originally during primary infection. Memory CD8 1 T cells depend on IL-12 as a signal 3 during Listeria recall infection To further support our findings that signal 3 cytokines modulate memory CD8 1 T-cell recall responses dependent on the infection, we analyzed, in a second approach, memory CD8 1 T cells lacking IL-12 as signal 3. For the generation of memory T cells, naïve P14.WT and P14.IL-12RKO T cells were adoptively transferred and recipient mice were infected with LCMV (experimental setup, Fig. 3A). We chose LCMV as primary infection since P14.IL-12RKO T cells expanded and gained effector functions like P14.WT cells and therefore memory T cells with comparable experience are generated [15]. About 5 days after primary infection, memory T cells were purified and equal numbers of P14.WT and P14. IL-12RKO memory T cells transferred into B6 mice. First, in vivo cytolytic activity of transferred P14.WT and P14.IL-12RKO memory T cells was tested in an in vivo killer assay with CFSE-labeled H8 and B6 spleen cells as targets. Both memory T-cell populations lysed the specific target cells with comparable kinetics, which is indicated as a reduction in H8 splenocytes in the blood (Fig. 3B). In a second approach, to investigate cytolytic activity of memory T cells under non-infectious conditions over a longer period of time, control or B16 melanoma cells expressing the GP33-epitope were used as read-out system. After AdTf of P14.WT or P14.IL-12RKO memory T cells, B16 GP33 melanoma cells were injected i.v. into recipient mice. Two weeks after transfer of tumor cells, a high number of B16 metastases were found in the lungs of mice, which did not receive memory T cells. & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Immunol. 211. 41: 3176 3186 Cellular immune response 3181 In contrast, lungs of recipient mice which received P14.WT or P14.IL-12RKO memory T cells were almost free of metastases (Fig. 3C). Thus, P14.IL-12RKO and P14.WT memory T cells exhibited comparable short-term and long-term in vivo cytolytic activities under non-infectious conditions. We conclude that memory T cells are independent of IL-12 (or type I IFN) as signal 3 for the development of cytolytic effector functions under noninfectious conditions. To analyze expansion and effector functions under infectious conditions, recipient mice of memory T cells were challenged with either LCMV (no dependency on IL-12 signals during primary infection) or recombinant Listeria GP33 (dependency on IL-12 signals during primary infection). On day 7 after secondary LCMV infection, P14.WT and P14.IL-12RKO secondary effector T cells showed vigorous expansion in the spleen with about 5% P14 T cells in the CD8 1 compartment. Although equally well in expansion, P14.IL-12RKO secondary effector T cells were slightly but significantly impaired in KLRG1 expression (only 6% expressing KLRG1 compared with 8% of P14.WT T cells) but not in T-bet expression on a per cell basis (Fig. 3D, Supporting Information Fig. 3A). Despite the lower expression of KLRG1, no differences were observed between P14.WT and P14.IL-12RKO secondary effector T cells concerning expression of other surface markers (CD127, CD44, and CD62L) or cytokine production. About 8% of both secondary effector T-cell populations produced IFN-g, 5% TNF-a, and 3 4% showed IL-2 production after short-term in vitro restimulation, with no significant differences in the MFI (Fig. 3D, Supporting Information Fig. 3A). After secondary Listeria infection, however, P14.IL-12RKO secondary effector T cells exhibited severely reduced expansion in A acute D LCMV infection (d7) E Listeria infection (d6) 1 naive CD8 T cells P14.WT or LCMV infection P14.IL-12RKO 12RKO (Thy1.1) B C 17 hrs 4 hrs no transfer B6 H8 no transfer P14.wt CFSE P14.wt memory CD8 T cells memory B in vivo killer C B16 challenge C/D infections P14.IL12RKO P14.IL12RKO killing? protection? expansion? % IFN-γ of P14 % KL RG1 hi of P14 P14/Thy1.1.21 28.9 44.74 26.1 1.27 28.8 45.2 1 8 6 4 2 1 8 6 4 2 CD8 24.7 % CD D127 hi of P14 total P14 per spleen % P14 o f CD8 6 5.6 3.66 4 3 2 1 1 1 1 1 8 P14/Thy1.1 RG1 hi of P14 85.7 1.6.3.53 88.2 11.2 1 8 CD8 % CD D127 hi of P14 total P14 per spleen 3 25 2 15 1 5 1 1 1 * 1 6 6 6 4 4 4 2 2 2 35 1 3 3 8 25 * 25 2 2 6 15 15 4 1 1 5 2 5 % KL % IFN-γ of P14 8 Figure 3. Functional analysis of LCMV-experienced P14.WT and P14.IL-12RKO memory T cells. (A) Experimental setup. Forty days after AdTf and primary LCMV infection, MACS-purified P14.WT and P14.IL-12RKO memory T cells (Thy1.1) were retransferred into B6 mice (Thy1.2) and analyzed for functionality. (B) Cytolytic activity of P14 memory T cells was determined by an in vivo cytotoxicity assay. In total, 5 1 5 purified P14.WT or P14.IL-12RKO memory T cells were transferred into naïve B6. Spleen cells of H8 mice (CFSE low) and B6 mice (CFSE high) were used as target cells. Cytolytic activity was measured 17 and 4 h after transfer of target cells by analyzing PBLs. Histograms show target cell populations, with numbers indicating percentages of CFSE-positive cells. (C) Analysis of long-term protective capacity using B16 GP melanoma cells. Two wks after AdTf of 2 1 4 P14 memory cells and 5 1 5 B16 GP melanoma cells, lungs of recipient mice were analyzed for metastases. (D, E) Analysis of re-expansion of LCMV-experienced P14 memory T cells. In total, 2 1 4 Thy1.1 1 P14.WT (black symbols) or P14.IL-12RKO (white symbols) memory T cells were transferred into B6 mice (Thy1.2 1 ) followed by infection with (D) LCMV or (E) rlisteria GP33. Expansion of P14 T cells was analyzed in the spleen (D) at day 7 after LCMV infection or (E) at day 6 after infection with rlisteria GP33. Dot plots demonstrate expansion of P14.WT (top, upper row) and P14. IL-12RKO LCMV memory cells (top, lower row). Percentages and absolute numbers of P14.WT (, black bars) and P14.IL-12RKO (J, white bars) T cells are shown (top). Percentages of KLRG1 and CD127 1 memory P14 T cells of total P14 T cells are shown (middle). Percentages of IFN-g-positive P14 T cells of total P14 memory T cells, as well as MFI of IFN-g of P14.WT (black bars and symbols) and P14.IL-12RKO (white bars and symbols) memory T cells are indicated (bottom). Values are expressed as mean (1SEM), n 5 3. po.5, po.1, po.1,, not significant (Student s unpaired t-test). Results are representative of two independent experiments. & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3182 Eur. J. Immunol. 211. 41: 3176 3186 the spleen compared with P14.WT cells (5 versus 22% P14 T cells of total CD8 1 ). Results were confirmed by absolute cell numbers (Fig. 3E). In addition, P14.IL-12RKO secondary effector T cells were impaired in KLRG1 expression and downregulated CD127 to a lesser extend than P14.WT T cells. The diminished expression of T-bet on a per cell basis correlated with the impaired KLRG1 expression on P14.IL-12RKO secondary effector T cells. In addition, more P14.IL-12RKO secondary effector T cells expressed IL-2 after restimulation, altogether indicating a less activated secondary effector phenotype. However, no differences were observed in differentiation into T CM and T EM subpopulations, as almost 1% of P14.WT and P14.IL-12RKO secondary effector T cells showed a T EM phenotype. Again, expression of IFN-g and TNF-a as well as degranulation capacity was comparable in both secondary effector T-cell populations (Fig. 3E and Supporting Information Fig. 3B). In line with our findings of memory CD8 1 T cells lacking type I IFN signals, these results indicate that memory CD8 1 T cells depend on a signal 3 for expansion and differentiation into highly activated secondary effector cells under defined infectious conditions. Memory T cells with different experience exhibit equal proliferative capacity after LCMV rechallenge So far, we examined memory T cells competent or deficient for signal 3 cytokines that were established during primary infection conditions leading to the same experience (comparable expansion, differentiation, and effector functions). Therefore, we analyzed in a third approach the impact of signal 3 cytokines on memory T-cell re-expansion in situations where already the primary T-cell response was determined by signal 3 and thus memory T cells with different experience were generated. To this end, P14.WT and P14.IL-12RKO memory T cells were retransferred into naïve recipient mice 5 days after primary Listeria infection (experimental setup, Supporting Information Fig. 4A). During primary Listeria infection, P14.IL-12RKO T cells are diminished in their expansion (Fig. 4A), show severely reduced KLRG1 expression, limited downregulation of CD127, and produce less IFN-g on a per cell basis compared with P14.WT T cells. Nevertheless, P14.IL-12RKO T cells form memory populations of comparable size to P14.WT T cells, but are slightly reduced in IFN-g and TNF-a production after ex vivo restimulation [15]. Thus, after primary Listeria infection, P14.WT and P14.IL-12RKO develop into memory T cells with different experience. To investigate re-expansion capacities of P14.WT and P14.IL-12RKO memory T cells generated after primary Listeria infection, equal numbers of memory T cells were retransferred into B6 mice and recipients challenged with either LCMV or rlisteria GP33. In accordance to our previous findings with naïve T cells, P14.WT and P14.IL-12RKO secondary effector T cells expanded to the same extent after LCMV infection (5 6% of CD8 1 in the spleen) (Fig. 4B). In contrast after Listeria infection, P14.IL-12RKO secondary effector T cells demonstrated a five-fold reduced expansion compared with P14.WT T cells (Fig. 4C). P14.IL-12RKO secondary effector T cells showed lower KLRG1 expression and diminished production of IFN-g during both LCMV and Listeria infections. In addition, during Listeria infection, more P14.IL-12RKO secondary effector T cells produced IL-2, thereby indicating a less activated phenotype compared with P14.WT T cells (Fig. 4C). However, production of TNF-a and degranulation were comparable in P14.WT and P14.IL-12RKO secondary effector T cells after LCMV and Listeria infections (Supporting Information Fig. 4B and C). In summary, while Listeria- and LCMV-experienced P14.IL-12RKO memory T cells developed robust secondary immune responses following LCMV infection, both rlisteria- and LCMV-experienced P14.IL-12RKO memory T cells were limited in expansion after rlisteria infection. Thus, these results indicate that the availability of signal 3 cytokines during priming does not significantly influence the recall capacity of memory T cells and that memory T cells are modulated by signal 3 cytokines for expansion dependent on the inflammatory milieu induced by the pathogens used for challenge. Discussion The ability to develop and sustain memory T-cell populations after infection or immunization is a hallmark of the adaptive immune response. The aim is to manipulate the primary activation of CD8 1 T cells in a way to improve formation and efficiency of memory responses to challenge. Here, we analyzed the direct influence of the proinflammatory cytokines IL-12 and type I IFN on T-cell memory formation as well as the recall responses and protective capacity of memory CD8 1 T cells after antigen re-encounter. We therefore followed two different approaches: first, we applied equal priming conditions for WT CD8 1 T cells and T cells lacking the receptor for IL-12 or type I IFN, using infectious conditions under which CD8 1 T cells were largely independent of the respective signal 3 cytokine (LCMV infection for IL-12RKO T cells and VV infection for IFNARKO T cells). We thus generated wild-type and signal 3-deficient memory T cells with the same experience and comparable effector to memory transition. Second, we analyzed memory T cells from WT and IL-12Rdeficient T cells that developed differently during primary Listeria infection. In this setting, during acute phase of infection, IL-12Rdeficient CD8 1 T cells are diminished in their expansion, show severely reduced KLRG1 expression and downregulation of CD127, and produce less IFN-g on a per cell basis compared with WT T cells. Additionally, CD8 1 T cells that lack IL-12 signals during priming with Listeria differentiated preferentially into a memory precursor phenotype. Thus, memory cells with different experience and different history of effector to memory transition were generated. The analysis of memory T cells at day 4 after infection generated from the two described settings revealed that CD8 1 T cells formed memory populations of comparable size & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Immunol. 211. 41: 3176 3186 Cellular immune response 3183 2 infection 1 infection A Listeria (d6) B LCMV (d7) C P14.wt P14.IL12RKO 12.12 5.7 85.3 8.9.1 1.9 88.5 9.5 CD8 (39.) (16,7) retransfer of sorted memory cells day 4 p.i. P14/Thy1.1 % KLRG1 hi of P14.9 54.2 1 8 6 CD8 18.6.16 25.1 53.6 4 2 27.1 21.1 total P14 + per spleen 8 6 4 2 1 1 1 5 4 3 Listeria (d6) 3 25 *.4 2.8 * P14/Thy1.1 % KLRG1 hi of P14 89.1 1 8 6 CD8 * * * * 7 hi of P14 % CD127 2 1 8.1.6.42 91.3 4 2 8.23 hi of P14 % CD127 total P14 + per spleen 2 15 1 5 1 1 1 5 4 3 2 1 * % IFN-γ + of P14 1 8 6 4 2 * 8 6 4 2 % IFN-γ + of P14 1 8 6 4 2 * 2 15 1 5 % IL-2 pos of P14 1 8 6 4 2 MFI IL-2 18 16 14 12 1 8 6 4 2 IL-2 pos of P14 % 5 4 3 2 1 MFI IL-2 5 * * 4 3 2 1 Figure 4. Functional analysis of Listeria-experienced P14.WT and P14.IL-12RKO memory T cells. (A) Expansion of naïve P14.WT (top) and P14. IL-12RKO (bottom) T cells day 6 after AdTf and primary rlisteria GP33 infection. Numbers in brackets indicate percentages of P14 T cells of total CD8 1 T cells in the spleen. Forty days after primary AdTf, 2 1 4 MACS-purified P14.WT and P14.IL-12RKO memory T cells (Thy1.1) were retransferred into B6 mice (Thy1.2) and analyzed for recall capacity. Expansion of P14 T cells in the spleen (B) at day 7 after LCMV infection or (C) at day 6 after infection with rlisteria GP33. Dot plots demonstrate expansion of P14.WT (upper row) and P14.IL-12RKO memory cells (lower row). Percentages and absolute numbers of P14.WT (, black bars) and P14.IL-12RKO (J, white bars) T cells are shown (top, right of dot plots). Percentages of KLRG1 and CD127 1 memory P14 T cells of total P14 T cells are depicted (middle). Percentages of IFN-g and IL-2-positive P14 T cells of total P14 memory T cells, as well as MFI of IFN-g and IL-2 of P14.WT (black symbols and bars) and P14.IL-12RKO (white symbols and bars) memory T cells are indicated (lower rows). Values are expressed as mean (1SEM), n 5 4. po.5, po.1, po.1,, not significant (Student s unpaired t-test). Results are representative of three independent experiments. independent of signal 3 availability during priming. Wild-type and signal 3-deficient memory T cells were able to produce cytokines after ex vivo restimulation (Fig. 1, [15]). Our finding that memory T cells formed independently of signal 3 contrasts previously published results which demonstrated that OT-1 T cells lacking type I IFN receptor or IL-12 receptor failed to form memory populations in response to VV or Listeria infections [19]. In this study, OT-1 T cells deficient in the IL-12Rb1 chain were used, which lack both IL-12 and IL-23 signals, whereas in the current article we worked with P14 T cells deficient in the IL-12Rb2 chain which exclusively lack IL-12 signals. Furthermore, the use of different TCR transgenic T cells (varying in TCR avidity), infection with different recombinant VV and different application routes may have an impact on signal 3 dependency of T-cell responses. In our hands, as demonstrated in the current and our previously published article, the generation of functional P14 memory T cells lacking IL-12 or type I IFN signals was observed after infections with LCMV, recombinant vesicular stomatitis virus expressing LCMV glycoprotein (rvsv GP ), rvv GP and rlisteria GP33 [13, 15] (our unpublished results). Even under conditions, in which severely impaired expansion occurred during primary infection (IFNAR-deficient T cells during LCMV infection) a small number of memory T cells were detectable at late time points [13, 14]. & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3184 Eur. J. Immunol. 211. 41: 3176 3186 It was suggested that altering the ratio of effector subpopulations during primary infections toward less activated KLRG1 lo effector cells (memory precursors) by reducing the inflammatory milieu might lead to higher numbers of efficient memory T cells [3, 7, 22, 23]. Findings of one study thus indicated that CD8 1 T cells in IL-12p35-deficient mice, and therefore in the absence of IL-12, differentiated more into a memory precursor T-cell phenotype and generated a higher frequency of long-term memory T cells with an enhanced protective capacity after rechallenge infection [22]. However, IL-12p35-deficient mice exhibit an altered cytokine pattern after infections as other cytokines are produced to compensate the general lack of IL-12. Thus, CD8 1 T cells in IL-12p35-deficient mice might be influenced by other cytokines and therefore conclusions of direct effects of IL-12 on memory formation of CD8 1 T cells are difficult to draw. In contrast, we used IL-12Rb2-deficient T cells adoptively transferred into B6 mice, thereby depriving only CD8 1 T cells from IL-12 signals during primary Listeria infection. Thus, we were able to study the transition of KLRG1 lo memory precursor effector cells to memory cells in an intact environment concerning the ability to produce an inflammatory milieu. In the context of primary Listeria infection, IL-12R-deficient T cells were limited in their expansion but demonstrated preferentially a memory precursor phenotype when compared with WT T cells. Thereafter, the stronger contraction of WT T cells mainly correlated with the disappearance of the terminal effector cell population over time. As the total number of memory precursor T cells was largely the same in WT and IL-12R-deficient populations during the peak of infection, it is not surprising that both cell populations formed memory pools of comparable size and the same ratio of T CM and T EM subpopulations until day 4 postinfection [15]. A similar memory precursor phenotype of CD8 1 T cells during primary infection was described in studies, in which antibiotic treatment limited inflammation after Listeria infection. In this setting, the availability of inflammatory cytokines was reduced due to a shortened duration of infection and memory-like CD8 1 Tcells were able to rapidly expand in response to a booster challenge at day 7 post primary infection. In contrast, CD8 1 T cells primed during Listeria infection without antibiotic treatment were not able to re-expand at this early time point after primary infection [7]. These and our findings clearly demonstrate that under reduced inflammatory conditions the generation of memory precursor T cells is favored. It was suggested that the rapid transition to memory phenotype of cells lacking inflammatory signals during priming might lead to a faster recall capacity compared with terminally differentiated effector cells [7, 24]. However, conclusive results are still lacking. In our model, the rapid transition into the memory precursor phenotype seems to have no influence on the size of the memory T-cell population formed at later time points. Altogether, these findings suggest that the quantity and quality of proinflammatory cytokines during priming determine (i) the expansion capacity of T cells and (ii) whether CD8 1 T cells exhibit a terminal effector phenotype to successfully combat an acute infection or rapidly progress to a memory status. To investigate the recall potential and protective capacity of memory T cells at day 4 after primary infection, we decided to retransfer normalized numbers of memory CD8 1 T cells into B6 mice, as the direct challenge of memory mice does not allow discrimination between individual variations in precursor frequencies and direct impact of signal 3 on reactivation of memory T cells. Thus being able to analyze the functionality of memory cells on a per cell basis, we found that signal 3-deficient memory CD8 1 T cells showed comparable cytolytic activity to wild-type cells under non-infectious conditions. This is in line with in vivo data from peptide immunization experiments revealing comparable expansion of memory cells in the presence or absence of IL-12 signals during secondary antigen contact [21]. Under infectious conditions during recall responses, we observed comparable effector functions and differentiation into secondary effector T-cell subpopulations independent of IL-12 or type I IFN signaling. However, memory CD8 1 T cells lacking signal 3 cytokine receptors were impaired in their re-expansion during secondary infections, dependent on the nature of the pathogen. In addition, secondary effector CD8 1 T cells lacking the type I IFN receptor were not only reduced in their proliferative, but also in their protective capacity after LCMV challenge infection. From this, we concluded that memory T cells exert immediate effector functions independent of signal 3 cytokines, and thus mount more efficient immune responses than naïve T cells under non-infectious conditions. However, protective immunity against rechallenge infection requires expansion of memory T cells, dictated by sensing inflammatory signals. Thus, memory/secondary effector T cells during secondary immune responses demonstrated a signal 3 dependency for expansion similar to naïve T cells during priming. Signal 3 dependency is thereby determined by the pathogen induced inflammatory milieu. Although inflammatory signals obtained during priming did not affect formation and re-expansion capacity of memory T cells, results of IL-12R-deficient memory T cells suggest that IL-12 signals during priming are able to imprint CD8 1 T cells for KLRG1 upregulation and T-bet expression. Lower KLRG1 expression on IL-12RKO secondary effector T cells after secondary infection correlated with increased IL-2 production and therefore points at less activated secondary effector CD8 1 T cells. These findings were independent of the experience of memory T cells during priming and therefore the effector to memory transition in both settings analyzed. In summary, our data clearly show that even in the absence of signal 3 cytokines during priming, functional memory T cells are formed, which exhibit cytokine production and in vivo cytolytic activity. However, depending on the rechallenge infection, signal 3 cytokines were crucial to support expansion and protective capacity of secondary effector CD8 1 T cells after pathogen reencounter. To our knowledge, this is the first study to demonstrate a signal 3 dependency of secondary effector T cells. Our results therefore reveal an essential role for inflammatory cytokines in secondary immune responses, thereby contributing to the understanding of the developmental fate of naïve and memory T cells. & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Immunol. 211. 41: 3176 3186 Cellular immune response 3185 Materials and methods T-bet and Eomes was done directly ex vivo. All flow cytometry was analyzed on a FACSCalibur or a FACSCanto (BD). Mice C57BL/6 (B6) mice were obtained from Harlan Winkelmann, IL-12Rb2-deficient (IL-12RKO) (Il12rb2 tm1jm ) mice and IFN-adeficient IFNARKO (B6.129S7-Ifnar1 tm1agt ) from the Jackson Laboratory [25, 26]. P14.WT TCR transgenic (B6.D2- Tg(TcrLCMV)327Sdz/JDvsJ), P14.IFNARKO, and P14.IL-12RKO mice on a B6.Thy1.1 background were generated by breeding. Mice were kept under specific pathogen-free conditions and used at 8 16 wk of age. Animal care and use was approved by the Regierungspräsidium Freiburg. In vivo cytolytic activity Cytolytic activity of memory P14 T cells was tested in vivo. Memory P14 T cells were MACS-purified and 2 5 1 5 P14 T cells (effector cells) transferred into B6. CFSE-labeled spleen cells from H8 mice (CFSE low), which ubiquitously express the LCMV-GP and are therefore recognized by P14 T cells, and B6 cells (CFSE high, as control) were used as targets. Target cells were mixed at a ratio of 1:1 and 2 1 7 cells injected i.v. into recipient mice. Cytolytic activity was detected at indicated time points in the blood [3]. Infections LCMV-WE and rvv GP [27] expressing the LCMV-GP were grown on L929 or BSC-1 cells. rlisteria GP33 [28], expressing the LCMV- GP epitope gp33-41, were grown in TSB medium. Mice were infected with 2 PFU of LCMV-WE, 2 1 6 PFU of rvv GP, and 2 1 4 CFU of rlisteria GP33 i.v. Protection was determined 4 days after infection with 2 PFU LCMV-WE in the spleen of recipient mice using a focus forming assay as described previously [29]. Acknowledgements: The authors thank Dr. H. Pircher for comments on the manuscript. This work was supported by Deutsche Forschungsgemeinschaft grant AI 34/2-1 (P. A.) and the Boehringer Ingelheim Fonds (S. J. K.). Conflict of interest: The authors declare no financial or commercial conflict of interest. AdTf experiments P14 T cells express a transgenic TCR specific for the LCMV-GP epitope gp33-41 in the context of H-2 D b. Splenic P14 T cells (P14.WT, P14.IFNARKO, or P14.IL-12RKO) from donor mice were purified and equal numbers were transferred i.v. into sexmatched B6 mice as described previously [15]. After AdTf of P14 T cells, recipient mice were infected with LCMV, rvv GP, or rlisteria GP33. For AdTf of memory cells, P14 memory T cells were isolated 4 5 days after primary infections. After MACS purification, 2 1 4 P14 memory T cells were retransferred into B6 mice for expansion and B16 GP protection experiments; 2 5 1 5 P14 memory T cells were retransferred for the analysis of cytolytic activity (in vivo killer) and protection (LCMV-WE challenge). For tumor protection experiments, 5 1 5 B16 GP melanoma cells were injected i.v. and 2 wk later lungs of recipient mice were analyzed. Flow cytometry All antibodies were purchased from ebioscience. For the analysis of intracellular cytokines, 1 6 lymphocytes per well were stimulated with 1 7 M LCMV-GP33-41 peptide in the presence of Brefeldin A for 4 h, followed by surface staining for CD8 and Thy1.1 and intracellular staining for IFN-g and TNF-a using the Cytofix/Cytoperm kit (BD Bioscience). Intracellular staining for References 1 Kaech, S. M., Hemby, S., Kersh, E. and Ahmed, R., Molecular and functional profiling of memory CD8 T cell differentiation. Cell 22. 111: 837 851. 2 Opferman, J. T., Ober, B. T. and Ashton-Rickardt, P. G., Linear differentiation of cytotoxic effectors into memory T lymphocytes. Science 1999. 283: 1745 1748. 3 Sarkar, S., Kalia, V., Haining, W. N., Konieczny, B. T., Subramaniam, S. and Ahmed, R., Functional and genomic profiling of effector CD8 T cell subsets with distinct memory fates. J. Exp. Med. 28. 25: 625 64. 4 D Souza, W. N. and Hedrick, S. M., Cutting edge: latecomer CD8 T cells are imprinted with a unique differentiation program. J. Immunol. 26. 177: 777 781. 5 Rao, R. R., Li, Q., Odunsi, K. and Shrikant, P. A., The mtor kinase determines effector versus memory CD81T cell fate by regulating the expression of transcription factors T-bet and Eomesodermin. Immunity 32: 67 78. 6 Araki, K., Turner, A. P., Shaffer, V. O., Gangappa, S., Keller, S. A., Bachmann, M. F., Larsen, C. P. et al., mtor regulates memory CD8 T-cell differentiation. Nature 29. 46: 18 112. 7 Badovinac, V. P. and Harty, J. T., Manipulating the rate of memory CD81 T cell generation after acute infection. J. Immunol. 27. 179: 53 63. 8 Chang, J. T., Palanivel, V. R., Kinjyo, I., Schambach, F., Intlekofer, A. M., Banerjee, A., Longworth, S. A. et al., Asymmetric T lymphocyte division in the initiation of adaptive immune responses. Science 27. 315: 1687 1691. & 211 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim