T Cells Produce + Antigen-Specific Nai - The Journal of Immunology

T Cells Produce + Antigen-Specific Nai - The Journal of Immunology

Antigen-Specific Naive CD8+ T Cells Produce a Single Pulse of IFN- γ In Vivo within Hours of Infection, but without Antiviral Effect This information ...

2MB Sizes 0 Downloads 0 Views

Recommend Documents

TL1A - The Journal of Immunology
di Investimento per la Ricerca di Base, and 60%), Fondazione Cassa di Risparmio, ...... Arnett, F. C., S. M. Edworthy, D

Author Index - The Journal of Immunology
Berry, Daniel, 6730. Besra, Gurdyal S., 6650. Beum, Paul ... Chang, Alfred E., 6713. Chang, Serena, 6758 ... Friedmann,

Author Index - The Journal of Immunology
Braun, Michel Y., 6790. Brazil, Jennifer C., 7186. Bresson .... Kasper, Lloyd H., 7022. Kaveri, Srinivas, 6585 ... Lloyd

Author Index - The Journal of Immunology
Carlson, John H., 8063. Caron, Gersende, 7595. Carter, Robert H., 7343. Castro, Iris, 7729. Chamberlain ...... Caplan, B

Author Index - The Journal of Immunology
Chagnon, Miguel, 5193. Chang, Chih-Chao, 5208. Chao, Linda H., 5586. Chen, An-Jing, 5250. Chen, Kuan, 5363. Chen, Liepin

Author Index - The Journal of Immunology
Araújo-Santos, Théo, 1196. Arlaud, Gérard J., 808. Armstrong, Abraham, 818 ... Brown, Cheryl Y., 1071. Bublitz, DeAnn

Author Index - The Journal of Immunology
Backert, Linus, 2639. Bagnall, James, 2652. Bakke, Siril S., 2910. Bakos, Eszter ... Drube, Sebastian, 2948. Dudeck, Ann

Ligand-Independent Down-Regulation of - The Journal of Immunology
We thank Violette Paragas for running the RNase protection assays and. Robyn Kizer for her ... Szabo, S. J., A. S. Dighe

Cells - The Journal of Biological Chemistry
Feb 17, 1981 - Incubation of mouse T lymphoma (5-49) cells with the inosinate dehydrogenase inhibitor mycophenolic acid

Table of Contents (PDF) - The Journal of Immunology
George Kollias, Sijie Lu, Daniel Graf, and Chen Dong ..... Huett, Aylwin Ng, Zhifang Cao, Petric Kuballa, Masaaki Komats

Antigen-Specific Naive CD8+ T Cells Produce a Single Pulse of IFN- γ In Vivo within Hours of Infection, but without Antiviral Effect This information is current as of May 15, 2018.

Martin P. Hosking, Claudia T. Flynn and J. Lindsay Whitton J Immunol published online 11 July 2014 http://www.jimmunol.org/content/early/2014/07/11/jimmun ol.1400348

• 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 *average

Subscription Permissions Email Alerts

Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html 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 © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606.

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

Why The JI? Submit online.

Published July 11, 2014, doi:10.4049/jimmunol.1400348 The Journal of Immunology

Antigen-Specific Naive CD8+ T Cells Produce a Single Pulse of IFN-g In Vivo within Hours of Infection, but without Antiviral Effect Martin P. Hosking, Claudia T. Flynn, and J. Lindsay Whitton

ntigen-specific CD8+ effector T cells play a key role in controlling and eradicating primary infections with many viruses, intracellular bacteria, and protozoa. These cells develop from naive precursors, and the pathway from naive to effector to memory CD8+ T cells is currently thought to reflect the following distinct phases of cellular differentiation: 1) upon encountering cognate Ag and appropriate costimulatory molecules on professional APCs, naive CD8+ T cells become activated. These cells are thought to be functionally quiescent; and only after 2) clonal expansion do they 3) acquire their effector functions (cytokine synthesis/cytotoxicity); and 4) develop the ability to migrate to inflamed and infected tissues, where they exert their antimicrobial effects. Following the successful resolution of infection, Ag-specific CD8+ T cells (5) contract and form stable, long-lived memory T cells (reviewed in Refs. 1, 2), which constitute one of the cornerstones of protective immunity against secondary viral challenge. In naive wild-type mice infected with lymphocytic choriomeningitis virus (LCMV; Armstrong strain [LCMV-Arm]), virus titers increase exponentially during the first 12 h following infection, continue to rise until ∼72 h postinfection (p.i.), and remain readily detectable for several days thereafter, until cleared by the burgeoning virus-specific CD8+ T cell response. In contrast, we have recently shown that LCMV replication within LCMV-immune mice is suppressed as soon as 6 h p.i. (3). This extraordinarily rapid protection conferred by virus-specific CD8+ memory T cells has


Department of Immunology and Microbial Science, The Scripps Research Institute, La Jolla, CA 92037 Received for publication February 5, 2014. Accepted for publication June 5, 2014. This work was supported by National Institutes of Health Grants AI077607 and AI052351 (to J.L.W.) and T32 HL007195-34 (to M.P.H.). This is manuscript number 26057 from the Scripps Research Institute. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Address correspondence and reprint requests to Martin P. Hosking, Department of Immunology and Microbial Science, SP30-2110, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. E-mail address: [email protected] Abbreviations used in this article: gMFI, geometric mean fluorescence intensity; ICCS, intracellular cytokine staining; LCMV, lymphocytic choriomeningitis virus; LCMV-Arm, LCMV Armstrong strain; TSRI, The Scripps Research Institute. Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1400348

been attributed to a combination of several factors. First, memory cells outnumber their naive precursors by a large margin (often ∼1000:1). Second, they are better able to respond to low levels of cognate Ag (i.e., to have an increased functional avidity) (4), allowing them to detect an individual infected cell earlier in the course of its infection. Third, memory cells very rapidly express key effector functions following secondary exposure to virus. In contrast, when studied in vitro, naive CD8+ T cells are largely incapable of rapid effector function. For example, in response to in vitro stimulation with cognate peptide, naive CD8+ T cells fail to produce IFN-g (5) or kill Ag-expressing target cells in vitro (5– 7). Moreover, the in vitro development of these effector functions and/or phenotypes requires one or more rounds of cell division (8– 10). Fourth, memory cells had long been thought to initiate cell division sooner than their naive counterparts, although this is now doubtful (11). Fifth, memory cells express several proteins that permit them to enter, and transit through, solid tissues, allowing them to more readily access sites of infection (12, 13). In this study, we wished to better delineate the response of naive T cells within the hours following virus infection in vivo, thereby allowing us to assess the relative contributions of each of the above factors to antiviral protection. Our in vitro analyses were consistent with those published by other groups: naive CD8+ T cells failed to synthesize IFN-g following in vitro exposure to cognate peptide. However, and in stark contrast, when analyzed in vivo (i.e., when the naive CD8+ T cells are in their natural microenvironment, encountering authentic [virus-derived] Ag together with normal costimulatory signals), a large proportion of the cells rapidly produced IFN-g, which began long before the initiation of proliferation in vivo. The up- and downregulation of IFN-g synthesis was paralleled by TcRmediated changes in the expression of T-bet, a master regulator of IFN-g and other gene activity. Coincident with the waning of in vivo IFN-g production, and still before extensive proliferation had occurred, the recently activated cells expressed abundant granzyme B and perforin, key components of the cytolytic machinery and developed strong in vivo cytolytic activity. Surprisingly, the robust production of IFN-g from naive CD8+ T cells within hours of infection, and the substantial rise in cytolytic capacity, had no detectable impact on viral expansion during the first 48 h p.i.; a

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

In vitro studies have shown that naive CD8+ T cells are unable to express most of their effector proteins until after at least one round of cell division has taken place. We have reassessed this issue in vivo and find that naive CD8+ T cells mount Ag-specific responses within hours of infection, before proliferation has commenced. Newly activated naive Ag-specific CD8+ T cells produce a rapid pulse of IFN-g in vivo and begin to accumulate granzyme B and perforin. Later, in vivo cytolytic activity is detectable, coincident with the initiation of cell division. Despite the rapid development of these functional attributes, no antiviral effect was observed early during infection, even when the cells are present in numbers similar to those of virus-specific memory cells. The evolutionary reason for the pulse of IFN-g synthesis by naive T cells is uncertain, but the lack of antiviral impact suggests that it may be regulatory. The Journal of Immunology, 2014, 193: 000–000.

2 reduction in viral load was detected only after the CD8+ T cells had begun to rapidly expand and express proteins that facilitate the cells’ migration toward, and entry into, sites of infection.

NAIVE CD8+ T CELLS RAPIDLY EXPRESS IFN-g IN VIVO above [see Detection of in vivo IFN-g production by CD8+ T cells (direct intracellular cytokine staining)]. Proliferation was determined by CFSE dilution of transferred P14s and compared with in vivo cytokine production.

LCMV real-time PCR

Materials and Methods Ethics statement All animal experiments were approved by The Scripps Research Institute (TSRI) Institutional Animal Care and Use Committee and carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Mouse, virus, and adoptive transfers

Indirect intracellular cytokine staining (standard intracellular cytokine staining) A total of 2 3 106 isolated splenocytes was stimulated for 6 h ex vivo with GolgiPlug (BD Biosciences) and 1 mM synthetic peptide GP33–41 (GenScript). Following stimulation, the cells were Fc blocked (anti-CD16/32; BD Biosciences), surface stained with CD8a and CD45.1, fixed and permeabilized with CytoFix/CytoPerm (BD Biosciences), and stained for the intracellular cytokines IFN-g (XMG1.2; BioLegend) and TNF (MP6XT22; BioLegend). Samples that were not stimulated with peptide were used to assess specific individual cytokine production.

Detection of in vivo IFN-g production by CD8+ T cells (direct intracellular cytokine staining)

In vivo CTL Mice that had received naive P14 cells (see above) were either sham infected or infected with LCMV-Arm. To determine the in vivo cytolytic capacity of P14 cells at various different times p.i., mice were inoculated with peptidecoated, CFSE-labeled target cells (naive mice that had not received P14s also received target cells). Two hours after target cell delivery, the recipient mice were sacrificed, the transferred splenocytes were identified by staining for CD45.1, and the proportions of CFSE-low and -high cells were determined. The percentage of specific cytotoxicity was calculated as previously described (20): (1 2 [percent LCMV peptide labeled in infected mice/ percent control peptide labeled in infected mice]/[percent LCMV peptide labeled in naive mice/percent control peptide labeled in naive mice]) 3 100. Target cells were prepared as follows. Splenocytes from congenic Ly5a (CD45.1+) mice were incubated for 1 h at 37˚C with 1 mM either the LCMV peptide GP33–41, or the irrelevant control peptide NP366–374 from Influenza A (GenScript). After washing, the two populations were labeled for 7 min with differing concentrations of CFSE (GP33–41, 2.0 mM CFSE; NP366–374, 0.1 mM CFSE). These target cells were mixed 1:1, and a total of 107 cells was injected i.v. into mice, as described above. E:T ratios were calculated by determining the number of transferred P14s in the spleens of individual mice and comparing them with the number of transferred GP33–41loaded target cells, normalized to naive, control mice. Nonlinear regression lines were calculated using GraphPad Prism 6 (GraphPad).


As previously described (3, 11, 16, 17), 250 mg brefeldin A (Sigma-Aldrich, St. Louis, MO) was injected i.v. into LCMV- or sham-infected mice at the indicated time points. Six hours later, mice were sacrificed, and single-cell suspensions were prepared from spleens on ice in the presence of 10 mg/ml brefeldin A. Splenocytes were immediately Fc-blocked, surface stained, fixed, permeabilized, and stained for intracellular cytokines as above. Splenocytes were not stimulated with any exogenous synthetic peptide.

Statistical significance was determined (GraphPad Prism 6; GraphPad) by one-way ANOVA (Kruskal–Wallis with Dunn posttest), repeated unpaired t tests (Holm–Sidak method), or two-way ANOVA tests (Dunnett or Sidak posttest). The p values ,0.05 were considered significant and, unless indicated otherwise, are indicated in figures as follows: *0.05 $ p . 0.01, **0.01 $ p . 0.001, ***0.001 $ p . 0.0001, ****p # 0.0001.

Flow cytometry


A total of 2 3 106 splenocytes was Fc blocked and immunophenotyped with fluorescently labeled Abs to the following surface proteins (BioLegend and BE Biosciences): CD8a (53-6.7), Thy1.1 (OX-7), Ly5a (A20), CD44 (1M7), CD69 (H1.2F3), CD25 (PC61), CD62L (MEL-14), and CXCR3 (CXCR3-173). Intracellular marker expression was determined following fixation and permeabilization with either Cytofix/Cytoperm (BD Biosciences) or FoxP3/Transcription Factor Staining Buffer Set (eBioscience) with fluorescently conjugated Abs to Granzyme B (GB12; Invitrogen), T-bet (4B10; BioLegend), or Eomes (Dan11mag; eBioscience). Appropriately conjugated isotype control Abs were used in all experiments. Data were acquired on a BD Biosciences LSR II and analyzed using FlowJo (Tree Star). Geometric mean fluorescence intensity (gMFI) was normalized to the average expression following sham infection (0 h) and expressed as fold induction.

In vivo proliferation P14-Thy1.1 CD8+ T cells (isolated from naive mice) were loaded with 5 mM CFSE (Invitrogen) and injected into naive C57BL/6J hosts. After resting, recipient mice were injected with 2 3 106 PFU LCMV-Arm i.p. At various times following infection, mice were injected i.v. with 250 mg brefeldin A, and 6 h later, mice were sacrificed and processed as described

In vivo, TcR-activated naive CD8+ T cells produce IFN-g, but not TNF, within hours of viral infection Naive mice contain ∼100–500 naive CD8+ T cells that are specific for any given epitope (21, reviewed in Ref. 22), and the scarcity of these cells necessarily prevents analyses of their functional responses in the hours immediately following infection. Therefore, to assess in vivo responses mounted by naive T cells, we relied on the adoptive transfer of naive transgenic P14 CD8+ T cells, which are specific for the LCMV epitope GP33–41. We chose to transfer a large number of naive P14 cells for two reasons. First, it allowed us to carry out functional and phenotypic analyses of the cells immediately p.i. (i.e., without requiring the cells to multiply to a detectable level). An added benefit of this approach is that the cells are still synchronous at the time of analysis; such synchronicity is lost as individual cells, and their progeny, undergo cycles of cell division. Second, we wished to determine if naive CD8+ T cells could exert rapid antiviral protective effects in vivo and to compare this to the virus control imposed by CD8+ memory T cells, which suppress

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

P14-transgenic mice (C57BL/6J-Thy1.1+ or C57BL/6J-Ly5a/CD45.1+), specific for the H-2b LCMV epitope GP33–41 (14), OT-I transgenic mice (C57BL/6J-Thy1.1+), specific for the H-2k OVA epitope OVA257–264 (15), and congenic Ly5a mice were bred in-house. For most experiments, 1.7– 2.5 3 106 P14-transgenic CD8+ T cells (isolated from naive P14 mice) were transferred i.v. into 5- to 6-wk-old naive C57BL/6J mice (TSRI breeding colony). Mice were rested for 48 h before being injected i.p. with 2 3 106 PFU LCMV-Arm. For cotransfer experiments, 2 3 106 P14 and 2 3 106 OT-I transgenic CD8+ T cells were transferred i.v. into 5- to 6-wkold naive C57BL/6J mice before i.p. infection with 2 3 106 PFU LCMVArm 24 h later. In some experiments, naive CD8+ T cells were highly purified (.94%) via negative selection from P14-transgenic mice using the Naive CD8a+ T Cell Isolation Kit, mouse (Miltenyi Biotec, San Diego, CA), according to the manufacturer’s instructions. A total of 1.2 3 106 highly purified naive P14 CD8+ T cells was subsequently transferred i.v. into 5- to 6-wk-old naive C57BL/6J mice before i.p. infection with LCMVArm (2 3 106 PFU) 48 h later.

As previously described (3, 18) small (∼5 mg) pieces of spleen were harvested and stored in RNAlater (Qiagen). RNA was isolated using RNeasy mini (Qiagen) with on-column DNA digestion according to the manufacturer’s instructions. cDNA was generated using the primer 59CAGGGTGCAAGTGGTGTGGTAAGA-39 (Valuegene), specific for the genomic (negative) sense of the LCMV nucleoprotein RNA, at 200 nM with Multiscribe-RT (Applied Biosystems). TaqMan real-time PCR was performed with the following primers (Valuegene) and probe (Applied Biosystems): forward primer (900 nM) 59-CGCTGGCCTGGGTGAAT-39, reverse primer (900 nM) 59-ATGGGAAAACACAACAATTGATCTC-39, and probe (200 nM) 6FAM-59-CTGCAGGTTTCTCGC-39-MGBNFQ. LCMV genome copy number was calculated from standard dilutions of the plasmid pT7-S that encodes the S segment of LCMV (19) (a kind gift from Prof. Juan Carlos de la Torre, TSRI). Data were acquired on a Bio-Rad iCycler and analyzed with iCycler iQ software (Bio-Rad).

The Journal of Immunology viral replication within ∼6 h (3). To do so, it was necessary to generate mice in which the abundance of naive cells was approximately equivalent to the natural abundance of epitope-specific memory CD8+ T cells in LCMV-immune mice (1–2 3 105 cells/spleen). We therefore employed the experimental protocol shown diagrammatically in Fig. 1A: resting, naive transgenic P14 CD8+ T cells (CD44lowCD62LhighCD11alowCD127+CD692) were adoptively transferred into naive mice and allowed to “take” for 48 h; as shown in Fig. 3A (0-h time point), immediately prior to infection, the spleen contained ∼1–2 3 105 naive P14 cells. Recipient mice were sham-infected or received LCMV, and at various time points, brefeldin A was directly injected i.v. into mice, and the

3 in vivo production of IFN-g by transferred naive P14s was measured [see Materials and Methods, Detection of in vivo IFN-g production by CD8+ T cells (direct intracellular cytokine staining)]. Direct intracellular cytokine staining (ICCS) enhances the ability to detect in vivo cytokine production, producing a snapshot of activity during the 6-h period prior to harvesting (16, 17). Representative individual mice are shown in Fig. 1B (red numerals show the percentages of P14 cells that were producing IFN-g in vivo), and cumulative data are shown in Fig. 1C. As expected, mice receiving a sham infection contained very few IFN-g+ P14 cells (Fig. 1B, 0-h time point). Soon after LCMV infection, the naive P14s began to produce IFN-g; between 6 and 12 h p.i. (i.e.,

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 1. In vivo, TcR-activated naive CD8+ T cells produce IFN-g, but not TNF, within hours of viral infection. (A) Naive LCMV-specific CD8+ T cells (P14s) were adoptively transferred into naive mice. Representative FACS plots of marker expression upon transferred P14s (gated on CD8a+Va2+) are shown. Recipient mice were subsequently infected with LCMV-Arm (2 3 106 PFU) or given a sham infection, and the in vivo production of IFN-g by the transferred CD8+ T cells was directly determined within the spleen (see Materials and Methods). CD8+ T cells were not stimulated in vitro with exogenous peptide. (B) Representative FACS plots of total splenocytes at the indicated hour post–LCMV infection (gated on CD8a+ T cells). The red bolded numbers are the frequency of IFN-g production by the transferred P14 T cells (CD8a+Thy1.1+). (C) Cumulative data from four independent experiments show IFN-g production by naive P14 CD8+ T cells at seven time points p.i. Each data point represents an individual mouse. (D) In vivo production of TNF also was assessed by direct ICCS. In (B) and (D), representative contour plots (gated on CD8a+Thy1.1+, i.e., P14 cells) are shown. (E) Naive P14 (CD45.1+) and OT-I (Thy1.1+) transgenic CD8+ T cells were cotransferred into naive mice that were subsequently infected with LCMV, and the in vivo production of IFN-g and TNF by each population was determined. Representative FACS plots from matched samples, gated on CD8a+CD45.1+ (P14s) or CD8a+Thy1.1+ (OT-Is), are shown. Cumulative data are shown in (F). Significance was determined by one-way ANOVA [(C) Kruskal–Wallis with Dunn posttest] or two-way ANOVA [(F) Sidak posttest]. **0.01 $ p . 0.001, ****p # 0.0001.


In vitro Ag exposure triggers naive CD8+ T cells to produce TNF, but not IFN-g In summary, upon encountering naturally presented cognate Ag in vivo, naive CD8+ T cells produce IFN-g within hours of viral infection, but TNF appears to not be synthesized. This is the obverse of current understanding, which is based on analyses of in vitro exposure to cognate peptide (5–7). Consequently, we considered the possibility that the adoptive transfer process, and/or subsequent events within the recipient mouse, had altered the in vitro responsiveness of the P14 cells. To determine if this was the case, P14s were transferred to naive mice, and recipient mice were subsequently sham or LCMV infected. At the indicated time points (Fig. 2), transgenic CD8+ T cells were recovered from the recipient mice and cultured in vitro with 1 mM of the LCMV peptide GP33–41. Recipient mice were not injected with brefeldin A prior to sacrifice. Naive P14 cells recovered from sham-infected animals and incubated in vitro with cognate peptide (Fig. 2, 0-h time point, bottom panel) produced TNF, but not IFN-g. These results are entirely consistent with published data (5) but are, in essence, the opposite of what is observed in vivo. IFN-g remains

FIGURE 2. In vitro Ag exposure triggers naive CD8+ T cells to produce TNF, but not IFN-g. LCMV-naive mice, previously transferred with naive LCMV-specific T cells (P14s), were infected with LCMV-Arm (2 3 106 PFU) or given a sham infection, and spleens were harvested at the indicated times following infection. Splenocytes were incubated for 6 h without or with 1 mM of the synthetic LCMV peptide GP33–41, and then the in vitro production of IFN-g and TNF was determined by standard ICCS. Representative contour plots (gated on CD8a+CD45.1+, i.e., P14 cells) are shown.

detectable from unstimulated P14s (Fig. 2, top panel) harvested from animals at 24 or 48 h p.i., consistent with Fig. 1; of note, absolute frequencies are relatively lower compared with Fig. 1, because brefeldin A was not administered directly to mice 6 h prior to sacrifice, akin to previous reports (16, 17). Peptide stimulation of these P14s at either 24 or 48 h p.i. yielded little additional IFN-g and no TNF. Thus, naive T cells, harvested within ∼ 48 h of infection, appear largely inert in vitro, failing to specifically make IFN-g in response to in vitro peptide stimulation; in contrast, at each of the 12-, 24-, 36-, and 48-h time points, a large proportion of those cells actively produces IFN-g in response to Ag in vivo (Fig. 1B–F). When harvested at 72 or 96 h p.i., the P14 cells had acquired the standard in vitro phenotype of primary effector T cells, rapidly producing IFN-g 6 TNF in response to peptide (Fig. 2). IFN-g production precedes T cell division in vivo Next, we assessed the population dynamics of transferred P14 and OT-I transgenic T cells during the first 4 to 5 d of primary LCMV infection. In accordance with previous observations with regards to type I IFN–mediated attrition of naive CD8+ T cells (25), the absolute number of both P14 and OT-I CD8+ T cells declined ∼10-fold within the spleen during the first 36–48 h of LCMV infection. Our method of direct ICCS precludes enzymatic digestion of the spleen (16); however, because both populations of transgenic CD8+ T cells decline during this period, it is unlikely that they are being sequestered within areas of the spleen that are difficult to disassociate. P14s explosively expanded thereafter, whereas OT-Is failed to accumulate (Fig. 3A). To determine exactly when this TcR-dependent proliferation was initiated, naive P14s were labeled with CFSE and transferred into naive hosts that were subsequently sham-infected or infected with LCMV-Arm. Mice were also injected with brefeldin A (see Materials and Methods) and sacrificed 6 h later to simultaneously assess CFSE dilution and in vivo cytokine production. As expected, P14s did not proliferate following sham infection (Fig. 3B, 0-h time point). Within LCMV-infected mice, robust proliferation was not observed until ∼48 h p.i. (Fig. 3B, 48-h time point), indicating that proliferation initiated approximately between 36 and 48 h p.i. By 72 h p.i., P14s had proliferated extensively and completely diluted out their CFSE, similar to previous in vivo observations (11, 26). During the peak of IFN-g production at 24 h p.i. (Fig. 1C), P14s remained undivided (Fig. 3B, 3C), demonstrating that proliferation is not a necessary requirement to produce IFN-g in vivo. Moreover, as proliferation commenced between 36 and 48 h p.i., the production of IFN-g was significantly associated with dividing CD8+ T cells (Fig. 3D). Therefore, naive CD8+ T cells, following

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

12-h time point), ∼13% of all P14s within the spleen were actively producing this cytokine, and at the 24-h time point, IFN-g was being actively produced by ∼20% of all P14s. Similar numbers of actively responding cells were observed at the 36 and 48 h p.i. time points. However, the number of IFN-g–producing cells declined precipitously thereafter; at 72 h p.i., only 2.5% of P14s contained IFN-g, and by 120 h p.i., IFN-g+ cells were almost undetectable. We also determined whether the naive virus-specific CD8+ T cells produced TNF, in addition to IFN-g, in vivo. In contrast to IFN-g, which is rapidly and sustainably produced by the naive P14s, TNF was undetectable in vivo (Fig. 1D). This was surprising, given the well-established capacity of naive T cells to make this cytokine following in vitro stimulation (5). We next determined whether this burst of IFN-g was the result of TcR-specific or nonspecific activation [e.g., cytokine mediated (23) or pattern recognition (24)]. First, we took advantage of the fact that, at early time points p.i., the vast majority of endogenous CD8+ T cells (Thy1.12) are not virus specific; minimal IFN-g synthesis was observed within these endogenous cells (Fig. 1B, lower quadrants). Second, we cotransferred approximately equal numbers of P14 cells and OT-I (specific for the OVA epitope OVA257–264) transgenic CD8+ T cells into naive mice and measured in vivo cytokine production during subsequent LCMV infection. As shown in Fig. 1E and 1F, in vivo IFN-g production is strictly limited to the virus-specific P14 CD8+ T cells rather than the nonspecific OT-I CD8+ T cells.


The Journal of Immunology

their natural encounter with cognate Ag in vivo, rapidly synthesize IFN-g long before T cell proliferation is observed. After 48 h p.i., transferred P14s cease to produce IFN-g in vivo, rapidly expand, and acquire the capability to secrete cytokines during in vitro peptide stimulation. In vivo activated naive CD8+ T cells rapidly upregulate activation markers and develop an effector phenotype A number of activation markers have previously been shown to be expressed upon CD8+ T cells following TcR stimulation in vivo or in vitro. Thus, we next determined the expression kinetics of several of these proteins on naive T cells in the hours following infection (Fig. 4), allowing us to evaluate whether changes in expression were associated with: 1) production of IFN-g in vivo (Fig. 1) and/or in vitro (Fig. 2); and 2) subsequent in vivo cell division (Fig. 3). The activation marker CD69 (Fig. 4A) was quickly upregulated; within 12 h of LCMV infection, ∼25% of the P14s expressed CD69 (not shown), and by 24 h p.i., the great majority of P14s expressed high levels of this protein; the magnitude of the change in expression is presented in the graph at the right side of Fig. 4A; at 24 h p.i., CD69 expression has increased ∼105-fold. The rapid rise was followed by a precipitous decline; by 72 h p.i., CD69 levels on transferred P14s had reverted to background. Thus, the kinetics of CD69 up- and downregulation closely follow those reported above

for in vivo IFN-g synthesis. Similar to previous in vivo observations (26), the expression of CD25, the a-chain of the IL-2R, also was markedly but transiently increased, although the kinetics were slightly delayed compared with CD69 and IFN-g (Fig. 4B). Within 24 h, .50% of transferred P14s upregulated CD25, and by 72 h p.i., nearly all of the P14s (.90%) expressed high levels of CD25. CD44, CD62L, CD11a, and CD127 are often used to differentiate naive (CD44LowCD62LHighCD11aLowCD127+) from effector (CD44HighCD62LlowCD11aHighCD1272/Low) CD8+ T cells. We next attempted to determine when, following in vivo TcR stimulation, naive P14s acquire these markers of effector differentiation. Sham infection does not alter the expression profile of the transferred naive P14s; the cells remain CD44LowCD62LHigh CD11aLowCD127+ in vivo (Fig. 4C–F, 0-h time point), similar to their input expression (Fig. 1A). Moreover, P14s remain CD44Low CD62LHigh for the first 48 h of LCMV infection before rapidly acquiring high levels CD44 expression and downregulating CD62L at 72 h p.i. (Fig. 4C, 4D). CD11a progressively increases upon transferred P14s over the course of LCMV infection (Fig. 4E), whereas CD127 expression oscillates during LCMV infection: first rapidly reducing between during the first 36 h of infection, before briefly increasing at 48 h, and, again, downregulating thereafter (Fig. 4F). Finally, the transition of P14s to a classical effector CD8+ T cell phenotype at ∼72 h p.i. is accompanied by the upregulation of the chemokine receptor CXCR3 (Fig. 4G), important for lymphocyte transit to infected peripheral organs. Naive CD8+ T cells quickly upregulate both T-bet and Eomes; only T-bet is Ag specific The transcription factors T-bet and Eomes are considered to be master regulators of CD8+ T cell differentiation and effector activity (including IFN-g production) (27–31). We next sought to determine the expression profile of these important mediators of function within CD8+ T cells encountering and responding to viral Ag in vivo. Following LCMV infection, T-bet is specifically upregulated within transferred P14s in two distinct waves (Fig. 5A, top panel, fold induction shown in Fig. 5B, filled circles). Initially, P14s rapidly upregulate T-bet within 24 h p.i. Thereafter, T-bet expression wanes between 24 and 48 h p.i. within transferred P14s, before, again, rapidly accumulating within expanding P14s at 72 and 96 h p.i. Thus, the kinetics of the first wave of T-bet expression are coincident with IFN-g production in vivo. Although we cannot exclude the possibility that some of the changes in T-bet expression may be nonspecific, endogenous CD8+ T cells showed only a limited increase in T-bet expression over the course of the experiment (Fig. 5A, bottom panel, Fig. 5B, open circles). Moreover, T-bet expression in OT-I cells was barely altered by LCMV infection (Fig. 5B, closed triangles). Therefore, we conclude that the observed changes in T-bet expression in virus-specific cells are driven largely by specific TcR signaling, both before and after the initiation of cell proliferation. In contrast (Fig. 5C, 5D), the kinetics of Eomes expression were very similar in all three cell populations: transferred P14 cells, transferred OT-I cells, and host CD8+ T cells. Eomes was markedly upregulated within 24 h p.i., followed by a reversion to background level. Hence, unlike T-bet, Eomes expression may be controlled by non–Ag-specific signaling. Naive CD8+ T cells begin to express granzyme B and perforin within hours of infection, but acquisition of in vivo cytolytic activity is delayed In TcR-stimulated naive CD8+ T cells in vitro, T-bet binds directly to the granzyme B promoter, regulating transcription of the mRNA encoding this important component of the cytolytic machinery (24,

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 3. IFN-g production precedes T cell division in vivo. (A) The absolute number of P14 or OT-I transgenic CD8+ T cells within the spleens of recipient mice was determined at the indicated time points following LCMV infection. (B) CFSE-labeled naive P14 CD8+ T cells were transferred into naive mice that were subsequent infected with LCMV, and proliferation and in vivo cytokine production was determined [see Materials and Methods, Detection of in vivo IFN-g production by CD8+ T cells (direct intracellular cytokine staining)]. Representative histograms (gated on P14 CD8+ T cells) are shown. Red bolded numbers indicate the percent undivided (CFSEHigh) and percent divided (CFSElow). (C) Representative pseudocolor dot plots (gated on P14s) comparing CFSE and IFN-g are shown. (D) The IFN-g gMFI between undivided and divided P14 CD8+ T cells was also determined (n = 3 to 4). Data presented in (A) are a summation of up to eight independent experiments; each data point represents an individual mouse. Significant differences between undivided and divided P14s in (D) were determined by repeated unpaired t tests (Holm– Sidak method). **0.01 $ p . 0.001, ****p # 0.0001. nd, not determined.




32). Therefore, we investigated whether the Ag-specific changes in T-bet expression was accompanied by parallel changes in granzyme B expression by naive virus-specific CD8+ T cells in vivo. Within 24 h of LCMV infection, approximately one third of all P14 T cells expressed granzyme B (Fig. 6A, 6C), and granzyme B was enriched, both on a percentage and per-cell basis, up to 72 h p.i. (Fig. 6B). We next determined whether granzyme B expression was related to the production of IFN-g in vivo, because they both are regulated by T-bet. Most of the P14 cells were granzyme B negative at 12 h after LCMV infection (Fig. 6C); this was true even for the IFN-g+ cells, indicating that the TcR-dependent production of IFN-g and granzyme B are not inextricably linked. Granzyme B expression steadily increased between 12 and 48 h p.i., in both IFN-g+ and IFN-g2 P14s, and by 48 h p.i., before proliferation had begun, .75% of all IFN-g–producing P14s also expressed granzyme B. Perforin protein, too, was rapidly upregulated in the transferred P14 cells (Fig. 6D). Thus, soon after the

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 4. In vivo–activated naive CD8+ T cells rapidly upregulate activation markers and develop an effector phenotype. Total of 2 3 106 naive LCMVspecific CD8+ T cells was transferred into LCMVnaive mice that were subsequently infected with LCMV. At various times post–LCMV infection, splenocytes were isolated and the expression of activation markers determined among transferred P14 T cells. Representative FACS plots (gated on P14s, CD8a+Thy1.1+, or CD8a+Ly5a+) at the indicated time points and the fold induction of gMFI are shown for CD69 (A), CD25 (B), CD44 (C), CD62L (D), CD11a (E), CD127 (F), and CXCR3 (G). Data presented are summations of four (A, n = 8–16), four [(C) n = 5–16], two [(B), (D), (F), and (G) n = 4–8], or one [(E) n = 4] independent experiment(s). Isotype controls for each Ab are indicated in gray. Significance was assessed by one-way ANOVA (Kruskal–Wallis with Dunn posttest). *0.05 $ p . 0.01, **0.01 $ p . 0.001, ***0.001 $ p . 0.0001, ****p # 0.0001.

burst of IFN-g synthesis, there is a marked increase in expression of these two key components of the cytolytic machinery. Because we observed upregulation of portions of the cytolytic machinery within hours of LCMV infection, we next evaluated the Ag-specific in vivo cytolytic activity of the cells. Naive mice received 2 3 106 naive P14s and, 24 h later, were infected with LCMV. At various points following infection, CFSE-loaded target cells were inoculated, and the mice were sacrificed 2 h later, allowing us to determine, for each time point p.i., the in vivo specific cytolytic activity during the preceding 2-h period. Similar to previous observations (9), transferred naive P14 CD8+ T cells remained cytolytic-negative following a sham infection (Fig. 6E, 0-h time point). Cytolytic activity remained essentially undetectable at 24 and 36 h p.i.; this was surprising, because large numbers of P14 cells expressed granzyme B and had upregulated perforin at these time points (Fig. 6A, 6B, 6D). Significant CTL activity (∼35%) was, however, detectable at 48 h p.i., and, by 72 h p.i., the

The Journal of Immunology


P14 cells were capable of rapidly eliminating nearly all of the GP33–41-labeled target cells within a 2-h period (Fig. 6E, 6F). Finally, the cytolytic activity of individual P14 cells was estimated by correlating total cytolytic activity against the estimated E:T ratio (Fig. 6G). At 48 h p.i. (Fig. 6G, closed squares), recently activated naive P14 cells showed cytolytic activity that was comparable to that reported for CD8+ memory T cells (3, 20), but only after multiple further rounds of cell division (72- and 96-h p.i. time points) was P14 cytotoxicity similar to that demonstrated by endogenous effector cells (20). The above observations are recapitulated in MACS-sorted naive CD8+ T cells For several reasons, we considered it likely that the above findings accurately reflected authentic in vivo TcR-mediated responses of naive CD8+ T cells. In vivo, the cells displayed surface and intracellular markers that are characteristic of naivety (Figs. 4, 5), and they failed to produce IFN-g following peptide stimulation ex vivo (Fig. 2). Nevertheless, we considered the possibility that the responding cells might not be truly naive and instead might be virtual memory cells, as described by others (33, 34). To address this concern, we repeated the above analyses with MACS-sorted P14 cells; a summary of the data is presented in Fig. 7. To minimize the risk that the sorting procedure would cause inadvertent activation of the P14 cells, naive CD8+ T cells were negatively selected (see Materials and Methods). After purification, the cells were characterized for expression of CD44, CD62L, CD11a, CD69, CD127, and CD25, and, by all criteria, the isolated P14s were phenotypically naive and unactivated (Fig. 7A). The cells were transferred into recipient mice and later infected with LCMV. Purified naive P14s produced, at 24 and 48 h p.i., a pulse of IFN-g, but no TNF in vivo (Fig. 7B). Moreover, the expression kinetics of IFN-g (Fig. 7C), granzyme B (Fig. 7D), and perforin (Fig. 7E) paralleled those reported above for nonsorted P14 cells. Finally, in response to infection, the MACS-sorted naive cells also showed the same delayed, but explosive, proliferation (Fig. 7F), as well as similar

changes in expression of T-bet and Eomes (Fig. 7G, 7H, respectively), and of CD44, CD62L, CD11a, and CD127 (Fig. 7I–L). We are, therefore, confident that the above findings (Figs. 1, 3–7) accurately represent the early Ag-specific in vivo responses of naive T cells to infection with cognate virus. Naive T cell suppression of virus in vivo: rapidity, correlation with effector functions, cell division, and marker phenotype Finally, we determined whether the early rapid production of IFN-g or the later development of CTL capabilities by transferred P14s provided any protective benefit to the host. We used a quantitative real-time PCR approach to determine genome copies within the spleens of LCMV-infected mice that previously received or did not receive naive P14s before infection. Unexpectedly, given the large number of IFN-g–producing P14 cells that were present in the spleens (Fig. 1B, 1C) and the presence of per-cell cytolytic capacity similar to that shown by memory T cells (Fig. 6F), there was no significant difference in LCMV splenic genome content during the first 48 h of infection, when compared with mice that had not received P14 T cells (Fig. 8A). However, a highly significant reduction in viral genome copy number occurred between 48 and 72 h p.i. (gray boxes, Fig. 8), a time point when P14s had: 1) largely terminated IFN-g synthesis (Figs. 1C, 8B, blue); 2) expressed high levels of granzyme B (Figs. 6A–C, 8B, green); 3) acquired significant CTL capabilities (Figs. 6E, 6F, 8B, red); and 4) undergone an ∼100-fold increase in number (Figs. 3A, 8B, black). Moreover, at this time, the cells also had: 5) increased their T-bet content in an Ag-specific manner (Figs. 5B, 8C, blue); 6) further downregulated CD62L (Figs. 4D, 8C, green); and 7) upregulated the chemokine receptor CXCR3 (Figs. 4G, 8C, red).

Discussion During a natural viral infection in vivo, Ag-specific naive CD8+ T cells are activated, differentiate, proliferate, and acquire effector functions that become readily measurable ∼5 d p.i. Prior to this

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 5. Naive CD8+ T cells quickly upregulate both T-bet and Eomes; only T-bet is Ag specific. Naive mice, having previously received 2 3 106 LCMV-specific naive CD8+ T cells (P14s) 6 irrelevant OVA-specific naive CD8+ T cells (OT-I) by adoptive transfer, were infected with 2 3 106 PFU of LCMV-Arm. Spleens were harvested from infected mice, and the expression of the transcription factors T-bet and Eomes was determined. Representative histograms from transferred P14s or host CD8a+ for T-bet (A) and Eomes (C) are shown. Isotype controls are shaded in gray. The gMFI fold induction of transferred P14s and host CD8a+ T cells for T-bet (B) and Eomes (D) was also enumerated. Data presented in (B) and (D) are the summations of three (P14 and Host CD8+, n = 4–12) or one (OT-I, n = 4) independent experiment(s). Significant differences among P14s, host CD8+, and OT-Is in (B) and (D) were determined by two-way ANOVA (Dunnett posttest). **0.01 $ p . 0.001, ****p # 0.0001.



time, the low frequency of virus-specific CD8+ T cells hampers the straightforward analyses of their phenotypic and functional attributes, and this difficulty increases as we regress toward the time of infection, at which point the extreme scarcity of naive precursor CD8+ T cells (22, 35) renders their study infeasible. In this study, we have circumvented this numerical limitation by transferring a large number of P14 cells into naive mice, permitting us to assess their responses during the earliest stages of virus infection. We feel that this in vivo model has at least two advantages over the more

commonly employed in vitro systems. First, it reveals the changes that occur in naive CD8+ T cells after exposure to naturally processed viral Ag and in their normal microenvironment. Second, it allows us to evaluate the major biological function of CD8+ T cells, suppression of viral replication in vivo, and this in turn permits us to investigate how this suppression of viral replication is temporally related to the expression of effector proteins (e.g., IFN-g and granzyme B) and to the presence of measurable effector activity (e.g., in vivo cytotoxicity).

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 6. Naive CD8+ T cells begin to express granzyme B and perforin within hours of infection, but acquisition of in vivo cytolytic activity is delayed. Total of 2 3 106 naive LCMV-specific CD8+ T cells (P14s) was transferred into LCMV-naive mice that were subsequently infected with LCMVArm (2 3 106 PFU). At the indicated times following infection, spleens were harvested, and the granzyme B expression was determined. (A) Representative histogram plots from transferred P14s are shown (gated on CD8a+Thy1.1+); isotype control is indicated in gray. (B) The fold change in gMFI of granzyme B (GrB), compared with 0 h, was calculated throughout LCMV infection. Data presented in (B) are a summation of two independent experiments (n = 4–8/ experiment). (C) Representative contour plots (gated on transferred P14s, CD8a+Thy1.1+) comparing the in vivo production of IFN-g [see Materials and Methods, Detection of in vivo IFN-g production by CD8+ T cells (direct intracellular cytokine staining)] and the expression of granzyme B. (D) The fold change in gMFI of perforin, relative to 0 h, was also calculated for transferred naive P14s. To determine in vivo cytotoxic activity, CFSE-labeled Ly5a+ splenocytes, loaded with the LCMV peptide GP33–41 or the irrelevant Db control peptide from influenza A NP366–374, were inoculated into naive mice (negative control) or into LCMV-infected mice at the indicated times p.i. Two hours later, the mice were sacrificed, spleens were harvested, and the percent specific cytotoxicity of GP33–41-loaded cells was determined by flow cytometry. (E) Representative histograms (gated on Ly5a+CFSE+ cells) are shown. (F) The total GP33–41 percent specific cytotoxicity present at each time point is shown. (G) The observed specific cytotoxicity was correlated with the E:T ratio (number of P14 cells/number of transferred GP33–41-loaded target cells). The symbol key in (G) applies also to (F). Data presented in (B) (n = 4–8), (F), and (G) are the summations of two independent experiments. Data in (D) are from one experiment (n = 3 to 4). Each data point in (F) and (G) represents an individual mouse. Significance in (B), (D), and (F) was determined by one-way ANOVA (Kruskal–Wallis with Dunn posttest). **0.01 $ p . 0.001, ****p # 0.0001.

The Journal of Immunology


Lack of expression of effector molecules is a defining and wellestablished feature of naive CD8+ T cells. When stimulated in vitro, effector and/or memory CD8+ T cells are immediately cytolytic and produce TNF, IFN-g, and IL-2. In contrast, their naive precursors have been thought to be unable to rapidly elaborate effector proteins in vitro, making only TNF in response to synthetic peptide (5–8). However, in this study, we report that, when assessed in vivo, naive CD8+ T cells quickly express IFN-g, granzyme B, and perforin (Figs. 1, 6, 7). These responses are TcR dependent, because they do not occur in the majority of naive endogenous cells nor in cotransferred naive cells specific to an irrelevant epitope (i.e., OT-I

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 7. The above observations are recapitulated in MACS-sorted naive CD8+ T cells. Naive P14 CD8+ T cells were MACS purified by negative selection and transferred into naive mice. (A) Representative contour plots (gated on CD8a+ Thy1.1+Va2+) demonstrating activation and effector marker expression on the input P14s are shown. (B and C) Recipient mice were infected with LCMV and the in vivo production of IFN-g and TNF was determined via direct ICCS (see Materials and Methods). Representative contour plots (gated on CD8a+Thy1.1+) are shown in (B). The fold change in expression of granzyme B (GrB) (D) and perforin (E) was also determined. (F) The absolute number of P14s within the spleen following LCMV infection was calculated. The fold change in expression of T-bet (G) and Eomes (H) was determined for transferred P14s and host CD8+ T cells. The gMFI fold expression, relative to sham-infected mice, was determined upon transferred P14s for CD44 (I), CD62L (J), CD11a (K), and CD127 (L). Significance was determined by two-way ANOVA (G, H, Sidak posttest) or oneway ANOVA [(C–E), (I–L) Kruskal–Wallis with Dunn posttest]. n = 4 for (C)–(L). *0.05 $ p . 0.01, **0.01 $ p . 0.001, ****p # 0.0001.

cells). The naive P14 cells began to produce IFN-g within 12 h of LCMV infection; this is barely sufficient time for LCMV to complete a round of replication (∼8–12 h) (36), demonstrating, for the first time, to our knowledge, that the in vivo Ag-specific naive CD8+ T cell response is remarkably rapid. IFN-g–expressing P14 cells remain abundant in vivo up to 48 h p.i., at which time cytokine production is terminated (Figs. 1, 7). Moreover, within 24 h of infection, naive CD8+ T cells also begin to upregulate granzyme B and perforin, key components of the cytolytic machinery (Figs. 6, 7). Thus, the in vitro responses of naive T cells (Fig. 2) differ from those observed in vivo. Furthermore, this difference


was maintained until LCMV-specific CD8+ T cells rapidly expanded; newly activated transgenic CD8+ T cells that were actively producing IFN-g in vivo (i.e., cells harvested during the first 48 h of infection) were unable to produce appreciable amounts of IFN-g in vitro in response to peptide stimulation. Following in vivo expansion ($72 h p.i.), transgenic CD8+ T cells gained the capability to produce both IFN-g and TNF in vitro (Fig. 2). Taken together, these data show that: 1) naive T cells express key effector molecules much faster than previously thought; and 2) prior to P14 expansion, there is a signifi-

cant discordance between the in vitro activity of newly activated CD8+ T cells and their actual in vivo responses. The transcription factors T-bet and/or Eomes are critical for CD8+ T cell–mediated effector function, including IFN-g production and granzyme B expression (27–31). When stimulated in vitro, naive CD8+ T cells rapidly express T-bet following TcR ligation, whereas Eomes expression is delayed for several days and is dependent upon IL-2 (32, 37, 38). More generally, strong inflammatory signals, such as IL-12 or CpG, promote T-bet expression, and reciprocally repress Eomes (37, 38), and the relative level of T-bet versus Eomes expression determines whether an effector T cell will become a shortlived effector cell or a memory precursor effector cell (39, 40). We show in this study that, following viral infection, naive LCMVspecific CD8+ T cells specifically upregulate T-bet in two distinct phases in vivo: the first wave occurs between 12 and 24 h p.i., coincident with IFN-g production, and the second takes place between 48 and 96 h p.i., along with rapid in vivo expansion. By comparing Tbet regulation in P14 cells (all of which are LCMV specific) with those in endogenous CD8+ T cells and in naive OT-I cells, we conclude that both waves of T-bet upregulation are driven largely by TcR stimulation. This contrasts with in vitro activation of CD4+ T cells, which also results in dual waves of T-bet expression, but in which only the first is TcR mediated; the second wave appears to be driven by IL-12 (41). Others have reported that IL-12 upregulates T-bet, and coordinately downregulates Eomes, in CD8+ T cells during infection with the bacterium Listeria monocytogenes (38). However, it is worth noting that LCMV infection does not induce much, if any, IL-12 expression (42, 43), consistent with the proposal that the second wave of T-bet upregulation observed in this study does not rely on this cytokine. In contrast to T-bet, changes in Eomes expression during the hours following infection follow very similar kinetics in P14 cells, endogenous CD8+ T cells, and naive OT-I cells, suggesting that the driving forces regulating these events are largely not Ag specific. Interestingly, the early burst of IFN-g production by newly activated naive CD8+ T cells did not have any measurable antiviral effect. Moreover, although granzyme B and perforin were upregulated within 24 h of infection (Figs. 6, 7), and significant CTL activity was detected by 48 h of infection (Fig. 6), LCMV viral loads remained unaffected at this time point (Fig. 8A). Only between 48 and 72 h p.i. did modest, but highly statistically significant, suppression of viral replication begin. As summarized in Fig. 8 (gray boxes), during this 24-h period, there was a dramatic increase in the number of P14 cells in vivo (∼100-fold, indicative of six to seven cycles of cell division). Furthermore, the great majority of the P14 CD8+ T cells had, by 72 h p.i., acquired a classical effector CD8+ T cell phenotype (CD44HighCD62LLow CD11aHighCD1272/LowCXCR3+), and they had markedly increased their CTL capabilities. Strikingly, IFN-g production declined dramatically during the same interval. Thus, these in vivo analyses reveal three previously unappreciated facts about CD8+ T cells. First, as is true of memory cells, but was thought not to apply for naive cells, cytokine production and granzyme B accumulation are activated in naive T cells very rapidly p.i., before the cells have begun to divide. Second, despite the expression of antiviral effector functions, the newly activated naive cells cannot immediately suppress virus replication, even when the virus-specific cells are present in numbers equivalent to that of a normal memory T cell pool. Taken together, these data suggest that the enhanced and accelerated CD8+ T cell–mediated antiviral protection that is observed in immune mice is not mediated primarily by the increased abundance of virus-specific CD8+ T cells or by the memory cells’ ability to rapidly produce cytokines in vivo. Therefore, third, we speculate that the biological benefits of CD8+ memory T cells can be attributed largely to three features in which they differ from naive

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

FIGURE 8. Naive T cell suppression of virus in vivo: rapidity, and correlation with effector functions, cell division. (A) The level of LCMV viral genome copies within the spleens of naive mice receiving P14s or empty transfers was determined using quantitative real-time PCR. Data presented are a summation of three (+P14s, n = 4–12) and two (no P14s, n = 4–8) independent experiments. (B) Normalized percent maximum effector responses by P14 CD8+ T cells were calculated for in vivo IFN-g production (blue), granzyme B expression (green), and in vivo CTL (red) and plotted versus the absolute number of P14s (black) within the spleen (right axis). (C) Normalized percent maximums for T-bet (blue), CD62L (green), and CXCR3 (red) are also shown. The gray-shaded box identifies a critical period between 48 and 72 h (see text). Significance in (A) was determined by repeated unpaired t tests (Holm–Sidak method). **0.01 $ p . 0.001, ****p # 0.0001.


The Journal of Immunology

changes, including: histone acetylation, nucleosome depletion, histone methylation, and DNA demethylation of effector loci (7, 56–62). Moreover, epigenetic alterations of effector genes such as IFN-g and granzyme B are associated with the greater abundance of effector transcripts within both in vitro TcR-stimulated and days 7 to 8 p.i. effector CD8+ T cells, compared with naive CD8+ T cells (6, 60, 61). These changes begin early: comparisons of naive and effector CD8+ T cells have revealed that the latter contain several epigenetic modifications (8, 58, 61, 62) that are associated with the faster transcription, and greater transcript abundance, of effector-related genes (6, 61). It is interesting to note that others have reported that the epigenetic changes that occur at the IFN-g promoter following Ag contact in vitro differs from the changes observed following in vivo contact (63); perhaps this explains the in vivo/in vitro disjunction reported in this study (Figs. 1, 2). It is thought that these performance-enhancing epigenetic changes are complete after three to four rounds of cell division (64). However, epigenetic modifications continue to accrue; CD8+ memory T cells display stepwise transcriptome changes in response to repeated stimulation with cognate Ag (65). In our recent studies of CD8+ memory T cells, we showed that three characteristics of T cell exhaustion—failure to produce IFN-g in response to Ag contact, hierarchical decline in cytokine multifunctionality, and upregulation of inhibitory receptors—occur within hours of infection, and, based on these data, we proposed that these phenotypic changes do not constitute a slowly developing T cell dysfunction, but instead represent a rapid-onset feature of normal T cell physiology. We further speculated that, with each successive exposure to stimulatory Ag, epigenetic changes would occur that gradually imprinted and extended these inhibitory regulatory effects, ultimately causing the substantial degradation of T cell responsiveness that is observed during chronic infection. In this light, we now speculate that the T cell exhaustion of chronic infection may, in fact, represent accelerated T cell aging, as timed by the epigenetic clock. Consistent with this notion, Decman and colleagues (66) have implicated epigenetic changes in age-related functional immune decline in uninfected mice. The majority of CD8+ T cells from aged mice were CD44hi and expressed a variety of inhibitory receptors; transcriptome analysis indicated that these cells were similar to exhausted cells that are found during chronic infection. As noted above, based on the findings reported in this study, we conclude that: 1) the production of IFN-g by naive T cells is unlikely to exert any detectable antiviral effector activity and instead is likely to be regulatory; and 2) that the cell’s rapid termination of IFN-g synthesis and subsequent inability to produce it in the presence of immunostimulatory Ag together suggest that although initially potentially beneficial, the continued production of this cytokine may be undesirable.

Acknowledgments We thank Sheila Silverstein and Pamela Cryer for excellent secretarial support and the National Institutes of Health Tetramer Core Facility for providing MHC class I tetramers.

Disclosures The authors have no financial conflicts of interest.

References 1. Haring, J. S., V. P. Badovinac, and J. T. Harty. 2006. Inflaming the CD8+ T cell response. Immunity 25: 19–29. 2. Kaech, S. M., E. J. Wherry, and R. Ahmed. 2002. Effector and memory T-cell differentiation: implications for vaccine development. Nat. Rev. Immunol. 2: 251–262. 3. Hosking, M. P., C. T. Flynn, J. Botten, and J. L. Whitton. 2013. CD8+ memory T cells appear exhausted within hours of acute virus infection. J. Immunol. 191: 4211–4222.

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

cells: their increased expression of surface proteins that facilitate entry into, and transit through, solid tissues, their capacity to immediately exert cytolytic activity, and their increased sensitivity to low levels of cognate Ag (increased functional avidity), which allows them to detect infected cells very soon after infection has taken place (4). What, then, is the purpose of the rapid burst of IFN-g by newly activated naive CD8+ T cells? Three observations suggest it is unlikely to be directly antiviral. First, the early burst of synthesis has no detectable impact on the virus; second, IFN-g synthesis is terminated while viral replication is increasing exponentially; and, third, at 48–72 h p.i., when the T cells are first imposing their in vivo antiviral effects, the proportion of virus-specific T cells that are producing IFN-g is declining. Instead of performing any detectable antiviral roles, it is likely that the IFN-g derived from naive CD8+ T cells is regulatory. We have previously shown that CD8+ T cells that quickly initiate IFN-g synthesis are more likely to become immunodominant (44), and others have found that early IFN-g production, as detected by reporter gene intensity, directly influences subsequent in vitro functionality of the progeny effector CD8+ T cells (45). CD8+ T cells that lack the receptor for IFN-g expand poorly and accumulate to orders of magnitude lower than IFN-gR–sufficient CD8+ T cells, demonstrating the importance of direct IFN-g signaling into CD8+ T cells (46, 47). Indeed, as naive LCMV-specific CD8+ T cells initiated proliferation between 36 and 48 h p.i., the production of IFN-g was positively associated with dividing, rather than undivided, P14s (Fig. 3D). In addition to directly promoting CD8+ T cell function, IFN-g can also act upon DCs, promoting and altering MHC class I presentation (48–50) (reviewed in Ref. 51), thereby further promoting CD8+ T cell activation. Rather than performing a direct antiviral function, we propose that the pulse of IFN-g that is produced by naive CD8+ T cells serves to shape and sculpt the burgeoning CD8+ T cell immune response. Supporting this interpretation, Ab-mediated neutralization of IFN-g in vivo on day 1 following LCMV infection reduced both the generation of virus-specific CTLs and subsequent viral control. However, neutralization on day 3 p.i. (a time point at which virus titers are rising exponentially) had no discernable effect, suggesting not only that IFN-g is not directly antiviral, but also that its regulatory impact occurs very early in the expansion phase, consistent with our observations above (52, 53). Although it is tempting to focus mainly on the fact that IFN-g is rapidly produced, we also find it intriguing that the newly activated naive CD8+ T cells also very quickly downregulate IFN-g synthesis (Fig. 1), even though there is a concurrent exponential increase in virus replication (Fig. 8). This suggests that termination of IFN-g synthesis is an active regulatory process, rather than merely the consequence of Ag unavailability, and the cells remain IFN-g silent despite the presence of Ag. We have recently reported a similar phenomenon in CD8+ memory T cells responding to LCMV or to vaccinia virus (3); memory cells produced a short-lived burst of IFN-g, followed by a refractory phase during which they failed to produce cytokine in response to Ag. We proposed (3) that one reason for this cellular behavior may be the avoidance of immunopathology; IFNs are remarkably toxic molecules. In this study, we speculate that there may be an additional reason for the rapid termination of IFN-g synthesis: to regulate T cell aging. Most studies of cellular aging have focused on senescence, a process associated with telomere shortening. However, a second independent cellular aging clock has very recently been proposed, based on epigenetic changes as reflected by DNA methylation; and, at least in principle, this clock may apply to immune cells (54, 55). There is no doubt that T cells accumulate epigenetic changes throughout their lifespan. Following TcR-mediated activation and differentiation, effector CD8+ T cells acquire heritable epigenetic



33. Lee, J. Y., S. E. Hamilton, A. D. Akue, K. A. Hogquist, and S. C. Jameson. 2013. Virtual memory CD8 T cells display unique functional properties. Proc. Natl. Acad. Sci. USA 110: 13498–13503. 34. Bou Ghanem, E. N., C. C. Nelson, and S. E. D’Orazio. 2011. T cell-intrinsic factors contribute to the differential ability of CD8+ T cells to rapidly secrete IFN-g in the absence of antigen. J. Immunol. 186: 1703–1712. 35. Moon, J. J., H. H. Chu, M. Pepper, S. J. McSorley, S. C. Jameson, R. M. Kedl, and M. K. Jenkins. 2007. Naive CD4(+) T cell frequency varies for different epitopes and predicts repertoire diversity and response magnitude. Immunity 27: 203–213. 36. Lehmann-Grube, F., M. Popescu, H. Schaefer, and H. H. Gschwender. 1975. LCM virus infection of cells in vitro. Bull. World Health Organ. 52: 443–456. 37. Pipkin, M. E., J. A. Sacks, F. Cruz-Guilloty, M. G. Lichtenheld, M. J. Bevan, and A. Rao. 2010. Interleukin-2 and inflammation induce distinct transcriptional programs that promote the differentiation of effector cytolytic T cells. Immunity 32: 79–90. 38. Takemoto, N., A. M. Intlekofer, J. T. Northrup, E. J. Wherry, and S. L. Reiner. 2006. Cutting Edge: IL-12 inversely regulates T-bet and eomesodermin expression during pathogen-induced CD8+ T cell differentiation. J. Immunol. 177: 7515–7519. 39. Joshi, N. S., W. Cui, A. Chandele, H. K. Lee, D. R. Urso, J. Hagman, L. Gapin, and S. M. Kaech. 2007. Inflammation directs memory precursor and short-lived effector CD8(+) T cell fates via the graded expression of T-bet transcription factor. Immunity 27: 281–295. 40. Rao, R. R., Q. Li, K. Odunsi, and P. A. Shrikant. 2010. The mTOR kinase determines effector versus memory CD8+ T cell fate by regulating the expression of transcription factors T-bet and Eomesodermin. Immunity 32: 67–78. 41. Schulz, E. G., L. Mariani, A. Radbruch, and T. Ho¨fer. 2009. Sequential polarization and imprinting of type 1 T helper lymphocytes by interferon-gamma and interleukin-12. Immunity 30: 673–683. 42. Dalod, M., T. P. Salazar-Mather, L. Malmgaard, C. Lewis, C. Asselin-Paturel, F. Brie`re, G. Trinchieri, and C. A. Biron. 2002. Interferon alpha/beta and interleukin 12 responses to viral infections: pathways regulating dendritic cell cytokine expression in vivo. J. Exp. Med. 195: 517–528. 43. Orange, J. S., and C. A. Biron. 1996. An absolute and restricted requirement for IL-12 in natural killer cell IFN-g production and antiviral defense. Studies of natural killer and T cell responses in contrasting viral infections. J. Immunol. 156: 1138–1142. 44. Liu, F., J. L. Whitton, and M. K. Slifka. 2004. The rapidity with which virusspecific CD8+ T cells initiate IFN-g synthesis increases markedly over the course of infection and correlates with immunodominance. J. Immunol. 173: 456–462. 45. Beuneu, H., F. Lemaıˆtre, J. Deguine, H. D. Moreau, I. Bouvier, Z. Garcia, M. L. Albert, and P. Bousso. 2010. Visualizing the functional diversification of CD8+ T cell responses in lymph nodes. Immunity 33: 412–423. 46. Whitmire, J. K., B. Eam, N. Benning, and J. L. Whitton. 2007. Direct interferong signaling dramatically enhances CD4+ and CD8+ T cell memory. J. Immunol. 179: 1190–1197. 47. Whitmire, J. K., J. T. Tan, and J. L. Whitton. 2005. Interferon-g acts directly on CD8+ T cells to increase their abundance during virus infection. J. Exp. Med. 201: 1053–1059. 48. Hickman, H. D., K. Takeda, C. N. Skon, F. R. Murray, S. E. Hensley, J. Loomis, G. N. Barber, J. R. Bennink, and J. W. Yewdell. 2008. Direct priming of antiviral CD8+ T cells in the peripheral interfollicular region of lymph nodes. Nat. Immunol. 9: 155–165. 49. Norbury, C. C., D. Malide, J. S. Gibbs, J. R. Bennink, and J. W. Yewdell. 2002. Visualizing priming of virus-specific CD8+ T cells by infected dendritic cells in vivo. Nat. Immunol. 3: 265–271. 50. Willis, R. A., J. W. Kappler, and P. C. Marrack. 2006. CD8 T cell competition for dendritic cells in vivo is an early event in activation. Proc. Natl. Acad. Sci. USA 103: 12063–12068. 51. Groettrup, M., S. Khan, K. Schwarz, and G. Schmidtke. 2001. Interferon-g inducible exchanges of 20S proteasome active site subunits: why? Biochimie 83: 367–372. 52. Utermo¨hlen, O., A. Dangel, A. Ta´rnok, and F. Lehmann-Grube. 1996. Modulation by gamma interferon of antiviral cell-mediated immune responses in vivo. J. Virol. 70: 1521–1526. 53. Wille, A., A. Gessner, H. Lother, and F. Lehmann-Grube. 1989. Mechanism of recovery from acute virus infection. VIII. Treatment of lymphocytic choriomeningitis virus-infected mice with anti-interferon-g monoclonal antibody blocks generation of virus-specific cytotoxic T lymphocytes and virus elimination. Eur. J. Immunol. 19: 1283–1288. 54. Horvath, S. 2013. DNA methylation age of human tissues and cell types. Genome Biol. 14: R115. 55. Bocklandt, S., W. Lin, M. E. Sehl, F. J. Sa´nchez, J. S. Sinsheimer, S. Horvath, and E. Vilain. 2011. Epigenetic predictor of age. PLoS ONE 6: e14821. 56. Araki, Y., M. Fann, R. Wersto, and N. P. Weng. 2008. Histone acetylation facilitates rapid and robust memory CD8 T cell response through differential expression of effector molecules (eomesodermin and its targets: perforin and granzyme B). J. Immunol. 180: 8102–8108. 57. Fitzpatrick, D. R., K. M. Shirley, and A. Kelso. 1999. Cutting edge: stable epigenetic inheritance of regional IFN-gamma promoter demethylation in CD44highCD8+ T lymphocytes. J. Immunol. 162: 5053–5057. 58. Fitzpatrick, D. R., K. M. Shirley, L. E. McDonald, H. Bielefeldt-Ohmann, G. F. Kay, and A. Kelso. 1998. Distinct methylation of the interferon gamma (IFN-gamma) and interleukin 3 (IL-3) genes in newly activated primary CD8+ T lymphocytes: regional IFN-gamma promoter demethylation and mRNA expression are heritable in CD44(high)CD8+ T cells. J. Exp. Med. 188: 103–117.

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018

4. Slifka, M. K., and J. L. Whitton. 2001. Functional avidity maturation of CD8(+) T cells without selection of higher affinity TCR. Nat. Immunol. 2: 711–717. 5. Brehm, M. A., K. A. Daniels, and R. M. Welsh. 2005. Rapid production of TNF-a following TCR engagement of naive CD8 T cells. J. Immunol. 175: 5043– 5049. 6. Kaech, S. M., S. Hemby, E. Kersh, and R. Ahmed. 2002. Molecular and functional profiling of memory CD8 T cell differentiation. Cell 111: 837–851. 7. Kersh, E. N., D. R. Fitzpatrick, K. Murali-Krishna, J. Shires, S. H. Speck, J. M. Boss, and R. Ahmed. 2006. Rapid demethylation of the IFN-gamma gene occurs in memory but not naive CD8 T cells. J. Immunol. 176: 4083–4093. 8. Denton, A. E., B. E. Russ, P. C. Doherty, S. Rao, and S. J. Turner. 2011. Differentiation-dependent functional and epigenetic landscapes for cytokine genes in virus-specific CD8+ T cells. Proc. Natl. Acad. Sci. USA 108: 15306–15311. 9. Oehen, S., and K. Brduscha-Riem. 1998. Differentiation of naive CTL to effector and memory CTL: correlation of effector function with phenotype and cell division. J. Immunol. 161: 5338–5346. 10. van Stipdonk, M. J., G. Hardenberg, M. S. Bijker, E. E. Lemmens, N. M. Droin, D. R. Green, and S. P. Schoenberger. 2003. Dynamic programming of CD8+ T lymphocyte responses. Nat. Immunol. 4: 361–365. 11. Whitmire, J. K., B. Eam, and J. L. Whitton. 2008. Tentative T cells: memory cells are quick to respond, but slow to divide. PLoS Pathog. 4: e1000041. 12. Weninger, W., M. A. Crowley, N. Manjunath, and U. H. von Andrian. 2001. Migratory properties of naive, effector, and memory CD8(+) T cells. J. Exp. Med. 194: 953–966. 13. Lefranc¸ois, L. 2006. Development, trafficking, and function of memory T-cell subsets. Immunol. Rev. 211: 93–103. 14. Pircher, H., K. B€ urki, R. Lang, H. Hengartner, and R. M. Zinkernagel. 1989. Tolerance induction in double specific T-cell receptor transgenic mice varies with antigen. Nature 342: 559–561. 15. Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, and F. R. Carbone. 1994. T cell receptor antagonist peptides induce positive selection. Cell 76: 17–27. 16. Foster, B., C. Prussin, F. Liu, J. K. Whitmire, and J. L. Whitton. 2007. Detection of intracellular cytokines by flow cytometry. Curr. Prot. Immunol. Unit 6:24 (Suppl 78): 1–21. 17. Liu, F., and J. L. Whitton. 2005. Cutting edge: re-evaluating the in vivo cytokine responses of CD8+ T cells during primary and secondary viral infections. J. Immunol. 174: 5936–5940. 18. Kemball, C. C., C. T. Flynn, M. P. Hosking, J. Botten, and J. L. Whitton. 2012. Wild-type coxsackievirus infection dramatically alters the abundance, heterogeneity, and immunostimulatory capacity of conventional dendritic cells in vivo. Virology 429: 74–90. 19. Sa´nchez, A. B., and J. C. de la Torre. 2006. Rescue of the prototypic Arenavirus LCMV entirely from plasmid. Virology 350: 370–380. 20. Barber, D. L., E. J. Wherry, and R. Ahmed. 2003. Cutting edge: rapid in vivo killing by memory CD8 T cells. J. Immunol. 171: 27–31. 21. Blattman, J. N., R. Antia, D. J. Sourdive, X. Wang, S. M. Kaech, K. MuraliKrishna, J. D. Altman, and R. Ahmed. 2002. Estimating the precursor frequency of naive antigen-specific CD8 T cells. J. Exp. Med. 195: 657–664. 22. Jenkins, M. K., and J. J. Moon. 2012. The role of naive T cell precursor frequency and recruitment in dictating immune response magnitude. J. Immunol. 188: 4135–4140. 23. Yang, Y., J. C. Ochando, J. S. Bromberg, and Y. Ding. 2007. Identification of a distant T-bet enhancer responsive to IL-12/Stat4 and IFNgamma/Stat1 signals. Blood 110: 2494–2500. 24. Geng, D., L. Zheng, R. Srivastava, N. Asprodites, C. Velasco-Gonzalez, and E. Davila. 2010. When Toll-like receptor and T-cell receptor signals collide: a mechanism for enhanced CD8 T-cell effector function. Blood 116: 3494–3504. 25. Bahl, K., S. K. Kim, C. Calcagno, D. Ghersi, R. Puzone, F. Celada, L. K. Selin, and R. M. Welsh. 2006. IFN-induced attrition of CD8 T cells in the presence or absence of cognate antigen during the early stages of viral infections. J. Immunol. 176: 4284–4295. 26. Kalia, V., S. Sarkar, S. Subramaniam, W. N. Haining, K. A. Smith, and R. Ahmed. 2010. Prolonged interleukin-2Ralpha expression on virus-specific CD8+ T cells favors terminal-effector differentiation in vivo. Immunity 32: 91–103. 27. Intlekofer, A. M., N. Takemoto, E. J. Wherry, S. A. Longworth, J. T. Northrup, V. R. Palanivel, A. C. Mullen, C. R. Gasink, S. M. Kaech, J. D. Miller, et al. 2005. Effector and memory CD8+ T cell fate coupled by T-bet and eomesodermin. Nat. Immunol. 6: 1236–1244. 28. Mayer, K. D., K. Mohrs, W. Reiley, S. Wittmer, J. E. Kohlmeier, J. E. Pearl, A. M. Cooper, L. L. Johnson, D. L. Woodland, and M. Mohrs. 2008. Cutting edge: T-bet and IL-27R are critical for in vivo IFN-g production by CD8 T cells during infection. J. Immunol. 180: 693–697. 29. Pearce, E. L., A. C. Mullen, G. A. Martins, C. M. Krawczyk, A. S. Hutchins, V. P. Zediak, M. Banica, C. B. DiCioccio, D. A. Gross, C. A. Mao, et al. 2003. Control of effector CD8+ T cell function by the transcription factor Eomesodermin. Science 302: 1041–1043. 30. Sullivan, B. M., A. Juedes, S. J. Szabo, M. von Herrath, and L. H. Glimcher. 2003. Antigen-driven effector CD8 T cell function regulated by T-bet. Proc. Natl. Acad. Sci. USA 100: 15818–15823. 31. Zhu, Y., S. Ju, E. Chen, S. Dai, C. Li, P. Morel, L. Liu, X. Zhang, and B. Lu. 2010. T-bet and eomesodermin are required for T cell-mediated antitumor immune responses. J. Immunol. 185: 3174–3183. 32. Cruz-Guilloty, F., M. E. Pipkin, I. M. Djuretic, D. Levanon, J. Lotem, M. G. Lichtenheld, Y. Groner, and A. Rao. 2009. Runx3 and T-box proteins cooperate to establish the transcriptional program of effector CTLs. J. Exp. Med. 206: 51–59.


The Journal of Immunology 59. Mukasa, R., A. Balasubramani, Y. K. Lee, S. K. Whitley, B. T. Weaver, Y. Shibata, G. E. Crawford, R. D. Hatton, and C. T. Weaver. 2010. Epigenetic instability of cytokine and transcription factor gene loci underlies plasticity of the T helper 17 cell lineage. Immunity 32: 616–627. 60. Scharer, C. D., B. G. Barwick, B. A. Youngblood, R. Ahmed, and J. M. Boss. 2013. Global DNA methylation remodeling accompanies CD8 T cell effector function. J. Immunol. 191: 3419–3429. 61. Zediak, V. P., J. B. Johnnidis, E. J. Wherry, and S. L. Berger. 2011. Cutting edge: persistently open chromatin at effector gene loci in resting memory CD8+ T cells independent of transcriptional status. J. Immunol. 186: 2705–2709. 62. Zhou, W., S. Chang, and T. M. Aune. 2004. Long-range histone acetylation of the Ifng gene is an essential feature of T cell differentiation. Proc. Natl. Acad. Sci. USA 101: 2440–2445.

13 63. Bandyopadhyay, S., H. Z. Qui, and A. J. Adler. 2009. In vitro and in vivo differentiated effector CD8 T cells display divergent histone acetylation patterns within the Ifng locus. Immunol. Lett. 122: 214–218. 64. Russ, B. E., A. E. Denton, L. Hatton, H. Croom, M. R. Olson, and S. J. Turner. 2012. Defining the molecular blueprint that drives CD8(+) T cell differentiation in response to infection. Front. Immunol. 3: 371. 65. Wirth, T. C., H. H. Xue, D. Rai, J. T. Sabel, T. Bair, J. T. Harty, and V. P. Badovinac. 2010. Repetitive antigen stimulation induces stepwise transcriptome diversification but preserves a core signature of memory CD8(+) T cell differentiation. Immunity 33: 128–140. 66. Decman, V., B. J. Laidlaw, T. A. Doering, J. Leng, H. C. Ertl, D. R. Goldstein, and E. J. Wherry. 2012. Defective CD8 T cell responses in aged mice are due to quantitative and qualitative changes in virus-specific precursors. J. Immunol. 188: 1933–1941.

Downloaded from http://www.jimmunol.org/ by guest on May 15, 2018