IL-7- and IL-15-Rich Environment Cells That Are Activated in An

Home , Interleukin 7

Initial Antigen Encounter Programs CD8+ T Cells Competent to Develop into Memory Cells That Are Activated in an Antigen-Free, IL-7- and IL-15-Rich Environment This information is current as of September 26, 2021. Roberto Carrio, Oliver F. Bathe and Thomas R. Malek J Immunol 2004; 172:7315-7323; ; doi: 10.4049/jimmunol.172.12.7315 http://www.jimmunol.org/content/172/12/7315 Downloaded from

References This article cites 52 articles, 26 of which you can access for free at: http://www.jimmunol.org/content/172/12/7315.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 26, 2021 *average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Initial Antigen Encounter Programs CD8؉ T Cells Competent to Develop into Memory Cells That Are Activated in an Antigen-Free, IL-7- and IL-15-Rich Environment1

Roberto Carrio,* Oliver F. Bathe,† and Thomas R. Malek2*

Although much is known concerning the immunobiology of CD8؉ T memory cells, the initial events favoring the generation of CD8؉ T memory cells remain poorly deﬁned. Using a culture system that yields memory-like CD8؉ T cells, we show that 1 day after Ag encounter, Ag-activated T cells developed into memory-like T cells, but this optimally occurred 3 days after Ag encounter. Key phenotypic, functional, and molecular properties that typify central memory T cells were expressed within 48 h when the activated CD8؉ T cells were cultured with IL-7 or IL-15 in the absence of Ag or following transfer into normal mice. These data support a model whereby Ag activation of naive CD8؉ T cells not only programs effector cell expansion and contraction but the Downloaded from potential to develop into a memory cell which ensues in an Ag-free environment containing IL-7 or IL-15. The Journal of Immunology, 2004, 172: 7315–7323.

ver the last several years there has been progress in de- Current data favor a linear differentiation model for memory cell fining features that typify the induction of a CD8ϩ T cell development. The main tenet in this model is that memory CD8ϩ immune response, including the production of long- T cells are direct descendants of effector CTL. Thus, after Ag- O ϩ ϩ http://www.jimmunol.org/ lived Ag-specific CD8 T memory cells (1). Upon an initial en- activated naive CD8 T cells expand and differentiate into CTL, counter with Ag, naive CD8ϩ T cells undergo programmed ex- some of these effector cells escape apoptosis and differentiate over pansion, followed by differentiation into effector CTL, and then several weeks to fully express properties of central memory cells. programmed contraction through apoptosis (2Ð5). Ag-activated This process is complex, requiring the regulation of many genes ϩ CD8 T cells that escape apoptosis go on to differentiate into (26) and optimally occurs after Ag clearance. However, it is still memory cells that are detected in vivo long after the initial Ag unclear at what point effector cells optimally acquire competency ϩ encounter (6Ð12). The long-term persistence of memory CD8 T to develop into memory cells. This issue is difficult to study in vivo cells depends upon constant slow turnover mediated by IL-7 and during an immune response, as the magnitude of the expansion of

IL-15 (10Ð17). However, very little is known concerning signals the effector CTL potentially obscures the initial emergence of by guest on September 26, 2021 that favor memory cell development. Understanding the nature of memory cells. Typically, memory cell development in this setting such signals is fundamental toward improving the efficacy of has been assessed only after the contraction of effector cells, where vaccines. the detection of memory-like T cells becomes practical. ϩ Two types of memory CD8 T cells have been identified. One The inherent complexity of investigating memory cell develop- is designated effector-memory cells as they exhibit phenotypic and ment solely in vivo as a consequence of Ag activation of naive functional properties similar to effector CTL, but persist after Ag Ag-specific precursor cells accentuates a need for other experi- clearance, predominately in nonlymphoid tissue. The other is mental systems, including in vitro models, that simplify and em- called central memory cells, which are primarily found in second- ulate one or more aspects of memory cell development. There is ary lymphoid tissue and are functionally and phenotypically dis- especially a need for more flexibility in investigating the impact of tinct from effector CTL, effector-memory cells, and naive T cells individual agents or conditions on the cellular and molecular basis (18, 19). Characteristic properties of central memory cells include ϩ of T cell memory. In this regard, we previously demonstrated that expression of CD44high, CD62Lhigh, IL-2R␤ , Ly-6Chigh, and ϩ ϩ short-term (4Ð5 days) in vitro Ag-activated CD8 T cells readily CCR7 and enhanced sensitivity to Ag that leads to a rapid rein- developed into persistent central memory cells upon adoptive duction of the effector program (20Ð24). Recent studies indicate transfer into normal syngeneic mice in the absence of Ag (24). that effector-memory cells eventually convert to central memory This approach directly permits assessing the development of mem- cells which have the greatest potential to persist in vivo (25). ory cells from effector cells without complication of an ongoing immune response. Furthermore, von Andrian and colleagues (17, 27) demonstrated that T cells with properties of central memory cells were obtained when Ag-activated CD8ϩ T cells were cul- *Department of Microbiology and Immunology, University of Miami School of Med- icine, Miami, FL 33101; and †Department of Surgery and Oncology, Tom Baker tured for an extended time with IL-15, a cytokine already impli- Cancer Center, University of Calgary, Calgary, Canada cated in regulating the homeostasis of memory CD8ϩ T cells. Received for publication November 19, 2003. Accepted for publication April 9, 2004. These experiments illustrate the ability to model memory cell de- The costs of publication of this article were defrayed in part by the payment of page velopment solely in vitro. charges. This article must therefore be hereby marked advertisement in accordance In the present study, we have refined an in vitro system that with 18 U.S.C. Section 1734 solely to indicate this fact. leads to the generation of central memory-like CD8ϩ T cells. By 1 This research was supported by National Institutes of Health Grant R01 AI40114. using this system, we evaluated the competency of effector CTL to 2 Address correspondence and reprint requests to Dr. Thomas R. Malek, Department of Microbiology and Immunology (R138), University of Miami School of Medicine, develop into memory T cells. Our data support the view that CTL P.O. Box 016960, Miami, FL 33101. E-mail address: [email protected] optimally develop into central memory cells in the absence of Ag

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00 7316 INDUCTION OF CD8ϩ MEMORY T CELLS IN VITRO during a discrete time frame and that some changes toward the Proliferation assay memory phenotype are induced within days. IL-7 was shown to be Ag-driven proliferation was determined essentially as previously described as efficient as IL-15 in promoting memory phenotypic cells. These (33). In brief, to measure Ag-driven proliferation, naive OT-I spleen cells ϩ data support a model whereby Ag activation of naive CD8 T cells (1 ϫ 105 cells/well), or cytokine-expanded OT-I T cells (2 ϫ 104/well) not only programs effector cell expansion and contraction but the with T-depleted mitomycin C-treated normal C57BL/6 spleen cells (8 ϫ 104cells/well), as a source of APC, were cultured in 96-well plates with potential to develop into memory cells, which ensues in an Ag-free 3 OVA257Ð264 (0.1 nM). [ H]Thymidine was added during the last4hofa environment containing IL-7 or IL-15. This latter finding raises the 48-h culture. possibility that IL-7 and IL-15 not only promote the survival and homeostasis of CD8ϩ T memory cells but also instruct aspects of Adoptive transfer a developmental program that typifies these cells. The indicated population of cultured OT-I T cells (10 ϫ 106 cells) was injected via the tail vein in 0.5 ml of HBSS into nonirradiated mice. Materials and Methods mRNA analysis Mice Apoptosis and cell cycle gene expression array systems were purchased OT-I TCR-transgenic mice (28) were maintained by breeding heterozygous from SuperArray Bioscence (Frederick, MD). Total RNA was extracted OT-I mice to wild-type C57BL/6J mice. B6.SJL-Ptprc/BoAiTac mice, con- from the indicated cell population with TRIzol (Life Technologies, Grand genic for CD45 and expressing the CD45.1 allele were purchased from Island, NY). cDNA was prepared from this total RNA and hybridized to Taconic Farms (Germantown, MD). C57BL/6J mice were purchased from the arrayed filters according to the manufacturer’s instruction. The result- The Jackson Laboratory (Bar Harbor, ME). All mice were housed and bred ing hybridization signal was visualized by chemiluminescence. Data were under viral-Ag-free conditions. subjected to densitometric analysis using Scion Image Software (Scion, Downloaded from Frederick, MD) (34). RNA levels were expressed as relative OD measure- Cell culture conditions ment after normalizing to the hybridization signals to ␤-actin. Spleen cells from OT-I mice (1 ϫ 106/well) were cultured in 24-well plates 3 in 1 ml of complete RPMI 1640 medium (CM) (29), with OVA257Ð264 Results peptide (0.1 nM; synthesized by Research Genetics, Huntsville, AL) and Modeling the development of CD8ϩ central memory T cells in either IL-2, IL-7, or IL-15 (all cytokines (10 ng/ml) from PeproTech, vitro Rocky Hill, NJ). After 1 or 3 days in culture, the cells were harvested, ϩ http://www.jimmunol.org/ washed three times with RPMI 1640, and the CD8 OT-I T cells were Previous studies have demonstrated that when Ag-activated CD8ϩ purified by positive selection using anti-CD8 magnetic beads according to T cells were further cultured for 6 days with IL-15, the majority of the manufacturer’s instructions (Miltenyi Biotec, Auburn, CA). These cells were typically Ͼ98% CD8ϩ. Positive selection was always used to study the cells expressed a cell surface phenotype and the functional day 1 Ag-activated OT-I T cells for subsequent in vitro or in vivo exper- characteristics of central memory cells (17, 27). In the current iments and the day 3 cultured cells for adoptive transfer experiments. study, we used OVA-specific MHC class I-restricted CD8ϩ T cells Given the high proportion (Ͼ90%) of activated CD8ϩ T cells after the day ϩ from the OT-I TCR-transgenic mouse to more precisely model the 3 culture, positive selection of the CD8 T cells was found not to be induction of memory-like T cells in vitro. Initially, we determined required and not performed for in vitro experiments. These activated OT-I T cells were recultured in T25 culture flasks at 105cells/ml in 10 ml of CM the kinetics by which CTL expressed a memory phenotype. The without OVA257Ð264, but with the indicated cytokines. After an additional ability of IL-7 and IL-15 to induce such changes was also com- by guest on September 26, 2021 2 days in culture, the cells were harvested and recultured, if required, with pared, as both cytokines function redundantly in vivo to promote the same cytokines for 2Ð3 more days. In some experiments to generate the homeostasis of memory T cells. large number of cells, the day 3 cultures were expanded using T75 tissue culture flasks containing 30 ml of CM. These cultures were initiated at 3 ϫ Naive OT-I T cells were cultured with OVA257Ð264 and exog- 105 T cells/ml when using IL-7 or IL-15. Viability was determined at enous IL-2 for 3 days to generate effector cells. Since memory cell various time points by trypan blue exclusion or flow cytometry as de- development in vivo is best visualized after Ag has been elimi- scribed below. nated, these cells were washed to remove Ag. In some cases the ϩ Flow cytometry CD8 T cells were purified by positive selection to minimize the presence of APC and then recultured with IL-2, IL-7, or IL-15. We The following Abs were used for flow cytometry and purchased from BD confirmed that the day 3 OT-I effector T cells were not contami- PharMingen (San Diego, CA): FITC-CD8 (53.6.7), FITC-CD45.2 (104), nated with residual OVA and APC as the coculture of naive PE-V␣2 (B20.1), Cyc-CD8 (53.6.7), biotin-CD44 (Pgp-1), biotin-CD62L 257Ð264 (MEL-14), biotin-CD69 (H1.2F3), biotin-Ly-6C (AL-21), biotin-V␤5.1,5.2 OT-I spleen cells with unfractionated or purified OT-I effector (MR9-4), and biotin-CD25 (IL-2R␤) (7D4). The cells were analyzed using cells did not induce a proliferative response by the naive OT-I T a FACScan flow cytometer (BD Biosciences, San Jose, CA) and CellQuest cells (Table I). After the initial 3-day culture, Ͼ90% of the viable ϫ 5 software (BD Biosciences) as previously described (24). Typically 1 10 recovered cells were CD8ϩ T cells that coexpressed V␣2 and V␤5, cells/spleen or lymph node sample and 1 ϫ 104 cultured cells/sample were analyzed. Dead cells were visualized with 7-aminoactinomycin D (BD characteristic of the OT-I TCR (data not shown). At that time, PharMingen). Cultured cells were labeled using CFSE (Molecular Probes; Eugene, OR) as previously described (30). For cell cycle analysis, T cells were harvested, washed, and lysed in a buffer containing 0.1% sodium a citrate, 0.01% Triton X-100, and 0.1 mg/ml propidium iodide and incu- Table I. OT-I effector cells lack residual Ag bated overnight before FACS analysis. Stimulus cpm ϫ 10Ϫ3 CTL and cytokine assays

51 None 1.3 Cytotoxicity was measured by a standard Cr release assay as previously Ϫ10 OVA257Ð264 10 M 550.0 described (31) against EL4 targets alone or after a 1-h incubation at 37¡C Ϫ 10 11 M 80.6 with 0.1 nM OVA (32). IFN-␥ secretion was assessed for cytokine- Ϫ 257Ð264 10 12 M 1.7 stimulated OT-I T cells by reculture of washed cells (1 ϫ 106 cells/well) OT-I CTL 1 ϫ 105 1.3 in 1 ml of CM in 24-well plates previously coated with anti-CD3 (1 ␮g/ml) 3 ϫ 105 0.9 and anti-CD28 (3 ␮g/ml). Supernatant fluids were collected 48 h later and Purified OT-I CTL 1 ϫ 105 1.0 IFN-␥ was measured by ELISA using Ab pairs from BD PharMingen ac- 3 ϫ 105 0.6 cording to the manufacturer’s instructions. a OT-I spleen cells (2 ϫ 105/well) were cultured with the indicated concentration of OVA257Ð267 or washed mitomycin c-treated unfractionated or purified OT-I effec- 3 Abbreviation used in this paper: CM, complete medium. tor cells for 3 days. [3H]Thymidine was added during the last4hofculture. The Journal of Immunology 7317 activated OT-I T cells expressed CD69, CD25, and high levels of Role for Ag, IL-7, and IL-15 in the development of memory-like CD44, while CD62L was down-regulated (Fig. 1A). These cells T cells in vitro exhibited potent CTL activity against OVA257Ð264-pulsed EL4 (H- We next evaluated to what extent Ag-activated OT-I T cells re- b 2 ) targets (Fig. 1B). Thus, after 3 days in culture, these T cells mained competent to develop into memory-like T cells when typified effector CTL. maintained for different periods of time as effector cells. For these Upon subsequent culture in IL-2 for 2 days, the T cells expanded experiments, OT-I T cells were primed with OVA and APC ϳ 257Ð264 10-fold and maintained phenotypic and the functional properties for 1 or 3 days and then recultured in the absence of Ag with either of CTL. In marked contrast, culture of the CTL in either IL-7 or IL-2, IL-7, IL-15, or no exogenous cytokine for an additional 2 IL-15 resulted in a time-dependent decrease in cell size and a shift days (Fig. 2, A and B). Alternatively, the 3-day OVA-primed OT-I in phenotype toward a central memory cell, i.e., CD69 and CD25 T cells were further cultured without Ag, using IL-2 for an addi- were substantially down-regulated while CD62L was up-regulated tional 2Ð4 days before switching to IL-7 and IL-15 (Fig. 2, C and to a uniformly high level of expression. CTL activity was essentially undetectable (Fig. 1B). High expression of Ly-6C represents another property of memory CD8ϩ T cells (22), and a higher per- centage of OT-I T cells expressed Ly-6Chigh after culture with IL-7 or IL-15 (Fig. 1A). This shift in phenotype required ϳ48 h, as cells cultured for only 24 h in IL-7 and IL-15 expressed an intermediate phenotype. Furthermore, these phenotypic changes were accompanied by an ϳ4- to 8-fold increase in the number of OT-I T cells, Downloaded from indicating that the cells which expressed these properties did not simply represent the survival of a rare subpopulation of T cells that was present within the initial priming culture. Thus, these data demonstrate that a population of effector cells is rather uniformly and rapidly converted to express several properties characteristic of central memory T cells. http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 2. Temporal requirements for the induction of memory-like T cells by IL-7 or IL-15 in vitro. Splenocytes from naive OT-I mice were

activated with OVA257Ð264 and IL-2 for 1 (A)or3(B–D) days and further cultured with IL-2 for in the absence of Ag for 2 (C)or4(D) days. After these cultures, the OT-I T cells were recultured for 2 additional days with either IL-2, IL-7, or IL-15 as indicated. On the day indicated, cell viability FIGURE 1. IL-7 or IL-15 rapidly induced CTL to express phenotypic and surface phenotype for the indicated markers were assessed by FACS and functional properties of memory cells in vitro. Splenocytes from naive for the OT-I T cells. The FACS gating for CD69ϩ or CD62Lhigh is shown

OT-I mice were activated with OVA257Ð264 and IL-2 for 3 days and then in Fig. 1A, day 0. The expression for CD25 is represented as the mean the CD8ϩ OT-I T cells were purified and recultured in only the indicated fluorescent intensity (MFI) for all cells. Data represent the mean Ϯ SD of cytokine for an additional 1Ð2 days. Cell size and surface phenotype (A), three experiments. B, Splenocytes activated with OVA257Ð264 and IL-2 for as assessed by FACS, or CTL activity using OVA257Ð264- pulsed EL4 cells 1 day were recultured in the indicated cytokine for 1Ð2 days (day 2 or day (B) was determined for OT-I T cells at the indicated time point after cul- 3, respectively), and CTL activity was assessed using OVA257Ð264-pulsed ture. Data are representative of five experiments. EL4 cells. Data are representative of two experiments. 7318 INDUCTION OF CD8ϩ MEMORY T CELLS IN VITRO

D). Optimal conversion into memory-like cells occurred after 3 fold over this period. After this time, there was minimal expansion days of priming with IL-2 and Ag as these cells were most polar- by IL-7 or IL-15 (Fig. 3, C and D). Collectively, our data indicate ized by IL-7 or IL-15 with respect to down-regulation of CD69 and that optimal generation of memory T cells by IL-7 and IL-15, i.e., CD25 and up-regulation of CD62L (Fig. 2B). For CTL that were between days 3 and 5 of cell culture, occurs after at least ﬁve to six solely cultured in medium, cell expansion did not occur and via- cell divisions, based on these cell yields. bility was low. For the viable cells, some trends toward a memory All of the preceding experiments tested the capacity of IL-7 and phenotype were observed, e.g., down-regulation of CD69 and IL-15 to promote memory cell development after removal of Ag. CD25, while others were not, e.g., up-regulation of CD62L (Fig. To investigate whether Ag removal was in fact necessary, we ex- 2B). At later time points there we too few viable cells to analyze. amined the phenotype and CTL activity of OT-I T cells that were

When the Ag-activated cells were expanded with IL-2 for an primed for 3 days with OVA257Ð264 in the continuous presence of additional 2Ð4 days, IL-7 and IL-15 were less apt to encourage IL-2, IL-7, or IL-15. Under these conditions, all cytokines sup- memory-like T cell development. This was apparent based on the ported similar cell recoveries (data not shown). Each condition phenotype, as progressively fewer CD62Lhigh cells were detected, also generated activated OT-I T cells with a cell surface phenotype and on the cell recoveries, as OT-I T cell viability progressively characteristic of effector T cells, with strong CTL activity (Fig. 4, decreased (Fig. 2, C and D). Interestingly, IL-7 and IL-15 also favored the development of memory-like cells when OT-I T cells were primed for as little as 24 h with Ag and APC, as the majority of cells 2 days later were CD62Lhigh and exhibited progressively lower CTL activity when further cultured with IL-7 or IL-15 (Fig. Downloaded from 2, A and E). During these cultures, there was at least a 9-fold cell expansion using either IL-2, IL-7, or IL-15. Therefore, even after the initial encounter with Ag, when cells are programmed for expansion and differentiation into CTL, IL-7 and IL-15 inﬂuenced this program to promote properties characteristic of memory cells.

However, as apparent by OT-I T cells that were primed for 3 days http://www.jimmunol.org/ with OVA257Ð264, this is most efficient after several days as an effector CTL. OT-I T cells always exhibited the best growth when cultured with IL-2 (Fig. 3). For the first week in culture, IL-2 generated an approximate 1000-fold increase, or at least 10 cell doublings, of OT-I T cells. As shown in Fig. 2D, after 7 days in culture with IL-2, OT-I CTL were inefficiently converted to memory-like cells when placed in IL-7 or IL-15. At this point OT-I T cells were somewhat responsive to IL-2, but cell recoveries in IL-7 or IL-15 by guest on September 26, 2021 were Ͻ50% of the input (Fig. 3D). Furthermore, after a 1 (Fig.

3A)- or 3 (Fig. 3B)-day priming with IL-2 and OVA257Ð264, subsequent culture of the primed cells with either IL-7 or IL-15 always resulted in a lower growth rate, with cell expansion of ϳ40-

FIGURE 4. Effect of IL-7 and IL-15 during the initial Ag priming and subsequent development of memory-like T cells in vitro. A, Splenocytes

from naive OT-I mice were activated with OVA257Ð264 and the indicated cytokine for 3 days and then recultured in only the indicated cytokine for an additional 2Ð4 days. On the indicated day, viability and cell surface phenotype for the indicated markers were assessed by FACS for the OT-I FIGURE 3. Proliferative properties of cytokine-stimulated OT-I T cells. T cells. FACS gating is as described in the legend to Fig. 2. Data represent Naive OT-I splenocytes were cultured exactly as described in Fig. 2. On the mean Ϯ SD of three experiments. B, On days 3 and 5, CTL activity was the day indicated (A–D), cell numbers were assessed by trypan blue ex- assessed using OVA257Ð264-pulsed EL4 cells. Data are representative of clusion. Data represent the mean Ϯ SD of two to three experiments. three experiments. MFI, Mean ﬂuorescence intensity. The Journal of Immunology 7319

cultured in the cytokines used during priming, 2 days later, the OT-I T cells cultured in IL-7 or IL-15 resembled central memory cells (Fig. 4A, day 5), with minimal CTL activity (Fig. 4B, day 5). In contrast, those cells cultured with IL-2 behaved as effector CTL. For the most part, extending the culture period for 7 days resulted in a cell surface phenotype that was somewhat more uniformly characteristic of activated effector cells in IL-2 or memory cells in IL-7 or IL-15 (Fig. 4A, day 7) with relatively high cell viability in all cultures. Furthermore, OT-I T cells primed in the presence of

OVA257Ð264 and IL-7 or IL-15 for 3 days and maintained with these cytokines for an additional 6 days (data not shown) exhibited growth rates similar to those in Fig. 3 for cells primed in the presence of IL-2 and then shifted to IL-7 or IL-15. Further analysis of cells that were cultured for 5 days indicated that, upon restimulation, each group secreted IFN-␥ (Fig. 5A), although much higher levels were noted for OT-I T cells cultured with exogenous IL-2. Continued exposure of 3-day OVA-primed OT-I T cells to IL-7 or IL-15 for 2 additional days resulted in Downloaded from strong proliferative responses when rechallenged with OVA257Ð264 FIGURE 5. Functional properties of cytokine-stimulated OT-I T cells. (Fig. 5B). By contrast, the OT-I effector cells maintained with IL-2

Naive OT-I splenocytes were activated and cultured as described in Fig. 4. not only failed to proliferate to OVA257Ð264, but upon examining On day 5, the cells were restimulated with plate-bound anti-CD3 and anti- the cultures, most T cells appeared to have undergone activation- ␥ CD28 or OVA257Ð264 and APC. IFN- secretion (A) and proliferation (B), induced cell death. Thus, CTL exposed to IL-7 or IL-15, but not respectively, were assessed 48 h latter. Data are representative of three IL-2, exhibited intensiﬁed responsiveness to Ag, a property char- experiments. acteristic of central memory cells. http://www.jimmunol.org/ day 3) that was best in the cultures containing exogenous IL-2. Properties of in vitro effector and memory-like T cells upon Therefore, in the presence of Ag, IL-7 and IL-15 were largely adoptive transfer to normal syngeneic mice ineffective in promoting memory-like T cells. However, as before, The preceding experiments indicated that Ag-activated CD8ϩ T if Ag was removed by washing, and the cells were then further cells developed into memory-like T cells in the absence of Ag, but by guest on September 26, 2021

FIGURE 6. Temporal requirements for the engraftment of in vitro-generated effector and memory-like OT-I T cells upon adoptive transfer to normal mice. Naive OT-I splenocytes were activated for 1 and 3 days with OVA257Ð264 and IL-2 and after washing were further expanded in IL-2 without Ag 2Ð5 additional days. The activated effector OT-I T cells were adoptively transferred into normal CD45.1 C57BL/6-congenic mice. A and B, After 4 wk, the number of donor cells in the spleen was determined by FACS by coexpression CD8 and CD45.2. Representative FACS dot plots (A) and summary of donor engraftment (B) for all mice that received OT-I T cells cultured in vitro for the indicted number of days. Data in B are the mean Ϯ SD of six to seven mice per group from three experiments. C, Naive OT-I splenocytes were activated and cultured for 5 and 8 days as described in the legend to Fig. 4 using exogenous IL-2, IL-7, or IL-15, as indicated, to prepare effector vs memory-like OT-I T cells. Four weeks after adoptive transfer of these cultured OT-I cells to CD45.1 C57BL/6-congenic mice, engraftment of OT-I T cells in the spleen was determined by FACS, as in A, for coexpression of CD8 and CD45.2. Data are the mean Ϯ SD of seven mice per group from three experiments. D, The cell surface phenotype of the donor cells was determined for the indicated markers by FACS analysis after gating on CD8ϩCD45.2ϩ cells. For simplicity, shown are only the persistent OT-I T cells derived from day 3 effector CTL. Regardless of the level of engraftment and the type of cells transferred, the phenotype was the same as shown for both the spleen and lymph node. 7320 INDUCTION OF CD8ϩ MEMORY T CELLS IN VITRO in the presence of IL-7 or IL-15, and this differentiation optimally occurred during a discrete time frame after initial activation. To validate some of these issues in vivo, we next examined the capacity of effector OT-I T cells to engraft and persist in vivo when generated for various periods of time in vitro. Effector cells were generated by priming with OVA257Ð264 and Ag for 1 or 3 days and then further expanded in IL-2 as necessary. Regardless of when the CTL were adoptively transferred into syngeneic CD45.1-congenic C57BL/6 mice (Fig. 6A), donor-derived OT-I T cells were detected when assayed 4 wk later. Optimal persistence was noted for those effector cells that were generated during a 3-day in vitro culture (Fig. 6B). Importantly, the persistent cells uniformly expressed a phenotype, i.e., CD69neg, CD25neg, CD62Lhigh, and Ly-6Chigh, that typiﬁes central memory cells when assayed in either the spleen (Fig. 6D) or lymph nodes (data not shown) of recipient mice. Therefore, this time course for memory cell development in vivo largely parallels that for the in vitro conversion of IL-2-generated effector cells into memory-like cells.

We also compared the ability of in vitro-maintained effector vs Downloaded from memory-like T cells to persist upon adoptive transfer. Somewhat surprisingly, after 5 days in culture, there was no measurable dif- ference in the capacity of OT-I T cells cultured in IL-2, IL-7, or FIGURE 7. In vivo surface phenotype of adoptively transferred OT-I IL-15 to persist as memory phenotypic T cells (Fig. 6C, left), even effector T cells. After 5 days in culture, IL-2-induced CTL cells were though the latter two cell populations already expressed properties washed, labeled with CFSE to follow the donor cells, and adoptively trans- of memory cells. Therefore, at this juncture, OT-I CTL exhibited ferred into normal C57BL/6 mice. Two and 5 days after adoptive transfer, http://www.jimmunol.org/ a high potential to develop and persist as memory cell in vivo. FACS analysis was performed on splenocytes for the indicated markers However, after 8 days in culture, only the transferred IL-7 or IL-15 after gating on the CFSEϩCD8ϩ cells. Data are derived from six mice per memory-like T cells were readily detected in vivo 28 days after the group from two experiments. transfer (Fig. 6D, right). Therefore, the poor persistence of day 8 in vitro effector cells suggests that these OT-I T cells are less capable of developing into memory cells the longer they expand as and TNFR1) or decreased (TNFR-associated factor 6) in the mem- effector cells. ory-like IL-15-cultured OT-I T cells. Thus, these 10 mRNAs may To determine the rapidity by which some of these changes to- be directly and rapidly regulated by IL-15 as effector CTL develop ward memory cells occurred in vivo, the cell surface phenotype into memory T cells. The remaining four mRNAs (Bak, caspase-3, by guest on September 26, 2021 was examined 2 and 5 days after effector OT-I CTL were adop- caspase-8, and TNF-␣) were highly increased after continued cul- tively transferred to normal mice (Fig. 7). The expression of CD69 ture of CTL in IL-2. This pattern of mRNA expression is consis- and CD25 was both substantially down-regulated 2 days after tent with a cell that is a polarized effector cell poised for apoptosis. adoptive transfer in a manner essentially identical to that of effec- The IL-2-dependent CTL are likely protected from apoptosis tor cells cultured with IL-7 or IL-15. Somewhat surprisingly, through expression of heightened levels of Bcl-x. Ly-6C was expressed at a near uniform high level that character- On the cell cycle array, 42 mRNAs were not detected whereas izes CD8ϩ memory cells. This level of conversion was much 23 of the 54 remaining mRNAs were differentially expressed by at greater than that seen for in vitro-cultured cells. The expression of least 3-fold. Two of the differentially expressed mRNAs were co- CD62L was somewhat intermediate with only a subpopulation of ordinately up-regulated by the day 5 effector and memory-like transferred cells that expressed this molecule at high levels. There- OT-I T cells (data not shown). However, 21 of the remaining dif- fore, similar to in vitro, some changes occurred rapidly while other ferentially expressed mRNAs appeared to distinguish effector from required additional time. memory cells (Fig. 8B). When compared with the day 3 or day 5

effector cells, 15 mRNAs (cyclin G2 and B, Mcm family, Dp1, Molecular profiles of in vitro-generated effector vs memory cells Pcna, Csk1, Cdk1 and 6, Cdc20, and Prc1), which primarily pro- To begin to assess some of the initial molecular changes during the mote cell cycle progression, were highly decreased in the IL-15 development of memory CD8ϩ T cells, the expression of mRNAs memory-like T cells. Four other mRNAs showed the highest ex- that primarily regulate apoptosis (Fig. 8A) and the cell cycle (Fig. pression in the IL-15-induced memory-like cells (E2F2, Cdk8, 8B) were compared for day 3 effector OT-I T cells vs cells cultured p27, and p15). The two most highly and differentially expressed for 2 additional days with IL-2 or IL-15. For the mRNA of the 96 mRNA were p15 and p27, which both promote cell cycle arrest. genes on the apoptosis array, 26 mRNAs were not detected, When compared with the IL-2-cultured OT-I T cells, cell cycle whereas 30 of the 70 remaining mRNAs were differentially ex- analysis revealed fewer cells in S-G2-M for the IL-15 memory-like pressed by at least 3-fold. Sixteen of the differentially expressed cells (Fig. 8C). Collectively, these molecular changes support the mRNA were coordinately up- (12) or down-regulated (4) by the notion that the day 3 CTL consist of effector cells that have de- day 5 effector and memory-like OT-I T cells (data not shown). creased proliferative and enhanced survival potential after expo- However, 14 of the remaining differentially expressed mRNAs ap- sure to IL-15. These characteristics typify slowing dividing, long- peared to distinguish effector from memory cells (Fig. 8A). When lived memory cells. compared with day 3 effector cells, seven mRNAs (Bcl-x, Sur- vivin, Bnip3, TNF-␤, CD30, 4-1BB, and Rpa) were increased in Discussion the IL-2-expanded CTL but decreased in the IL-15 memory-like T Immunological memory is operationally defined as an extremely cells. Three other mRNAs were either selectively increased (CD27 rapid and efficient immune response upon a second encounter with The Journal of Immunology 7321

FIGURE 8. Differential mRNA expression of cytokine-induced effector and memory-like OT-I T cells. Splenocytes from naive OT-I mice were activated with

OVA257Ð264 and IL-2 for 3 days (D3) and then recultured with the indicated cytokine for an additional 2

days (D5). Total RNA was prepared from these cells Downloaded from and the reverse-transcribed cDNA was hybridized to the apoptosis (A) or cell cycle (B) GE array. Shown is the relative expression of those mRNAs that varied by Ͼ3-fold after densitometric analysis of the hybridized arrays and were uniquely characteristic of effector or memory-like OT-I T cells. Data represent the mean Ϯ

SD from RNA isolated from four distinct cultures for http://www.jimmunol.org/ the day 5 cultured cells or mean of 1 (B)or2(A) distinct cultures for the day 3 CTL. These day 3 data are shown as a reference and were internally controlled by parallel analysis to day 5 cultured cells. C, Cell cycle analysis for day 5 cultured cells. Data are representative of three experiments. by guest on September 26, 2021

an immunogen. Some of the key principles bearing on the cellular activated T cells are poised to develop into memory cells. Activa- basis of immunological memory have emerged by analysis of the tion of naive CD8ϩ T cells by Ag for as little as 24 h was sufficient properties of Ag-specific lymphocytes that persist after a success- to endow the capability to develop into memory-like T cells. Op- ful primary immune response in vivo (1). A critical aspect of our timal transition into memory-like CD8ϩ T cells required 3 days of study is that the in vitro system utilized closely recapitulates key activation in the presence of Ag. However, more extensive IL-2- phenotypic and functional properties of CD8ϩ T memory cells. driven expansion and polarization into a CTL rendered the cells Therefore, conclusions derived from this culture system should less capable of expressing properties of memory T cells. Second, directly relate to the mechanism of memory cell development in most, if not all, Ag-activated CD8ϩ T cells were competent to vivo. In this regard, most of our key findings were verified when develop into memory cells. This was especially evident when na- the phenotype and persistence of in vitro-derived effector cells ive OT-I T cells were cultured for 3 days with OVA257Ð264 and were analyzed following adoptive transfer into normal unmanipu- IL-2, leading to a population of effector CTL, which were essen- lated recipient mice. tially uniformly converted to memory-like cells within 48 h by One of the most difficult issues to study in vivo is the initial IL-7 or IL-15. This conversion was accompanied by a 4- to 8-fold events that control the development of memory T cells from Ag- expansion of T cells, which was nearly comparable to expansion activated lymphocytes. For this reason, many of our experiments driven by IL-2. Thus, the expression of memory cell properties relied on a system where memory-like T cells were generated in cannot be ascribed to a minor subset of Ag-activated T cells. vitro. Several important conclusions are evident from such exper- Lastly, memory T cell development did not occur unless the Ag- iments. First, there appears to be a window of time in which Ag- activated T cells were in the correct environment, i.e., with IL-7 or 7322 INDUCTION OF CD8ϩ MEMORY T CELLS IN VITRO

IL-15, but in the absence of Ag. Importantly, in the presence of transferred in vivo to normal mice. This quick conversion in vivo Ag, during the initial priming, IL-7 or IL-15 was ineffective in may at least partially explain why the transfer of either CTL or generating memory-like T cells. Although both cytokines are memory-like cells to normal mice, after a 5-day culture, persisted broadly and constitutively produced in an Ag-independent manner to essentially identical levels when assayed 4 wk later (Fig. 6C). by nonlymphoid cells, this finding suggests that IL-7 or IL-15 can- The short time frame for these changes appears to conflict with not subvert the effector phase of an immune response until Ag is several recent studies that favor a mechanism whereby the prop- eliminated or the effector cells are within a niche that is Ag free. erties of memory cells are gradually acquired over several weeks Recently, Wherry et al. (25) demonstrated that, after lympho- after the induction of an effector CTL response (2, 20, 26). Our cytic choriomeningitis virus infection, central memory CD8ϩ T data, however, also indicate a somewhat extended time for full cells persisted much longer than effector-memory T cells. These conversion to memory CD8ϩ T cells. Even after 5Ð8 days in cul- investigators further demonstrated that, upon clearance of Ag, ef- ture under memory cell conditions or 5 days after adoptive transfer fector-memory T cells converted to central memory cells. We be- in vivo, the phenotype of these OT-I T cells was not identical to the lieve it is noteworthy that in both our in vitro and in vivo model persistent memory cells characterized 4 wk after adoptive transfer systems, T cells with an effector-memory phenotype were essen- to normal mice (Figs. 1 and 7 vs Fig. 6D) or after development in tially not detected. This result indicates that production of effector- vivo after virus infection (21, 26, 45). Therefore, our findings sup- memory T cells is not a prerequisite for the development of long- port the notion that the cell fate decision whereby an effector CTL lived central memory cells and suggests that Ag removal is pivotal differentiates into a memory T cell is critically dependent upon a to produce central memory T cells. 48-h time period, at which point a number of the key properties of ϩ Both IL-7 and IL-15 are critical cytokines for the survival and a memory CD8 T cell are apparent. An extended time, however, Downloaded from homeostasis of memory CD8ϩ T cells (10Ð17, 35, 36). Our data is still necessary for full development into a long-term persistent suggest a much more active role for each of these cytokines in memory cell. promoting the development of memory CD8ϩ T cells. In the pres- A gene expression profile has been described for naive, effector, ence of either cytokine, but in the absence of nominal Ag, cell and memory cells by isolating the appropriate cell populations af- surface phenotypic, functional, and molecular changes occurred ter infection (26, 45). This type of analysis provides an important ϩ that were characteristic of central memory cells, including down- snapshot of gene expression by the CD8 T cells as a consequence http://www.jimmunol.org/ regulation of CTL activity and reacquisition of Ag responsiveness. of an immune response, but it does not ascribe particular changes These data are consistent with a direct role for IL-7 and IL-15 in in profile to a specific signal. A distinct feature of in vitro-gener- regulating these changes. However, it is highly unlikely that IL-7 ated memory-like CD8ϩ T cells is that large numbers of T cells are or IL-15 are the only extrinsic factors responsible for the devel- available for cellular and molecular analysis to assay the initial opment of memory T cells. Notably, Ly-6C expression was highly, events as memory T cells develop from effector CTL. During this rapidly, and uniformly induced upon adoptive transfer of CTL to effector to memory transition, our molecular analysis indicates a normal mice while this characteristic of CD8ϩ memory T cells was large number of changes occurred during this 48-h culture period. markedly less efficient in vitro with either IL-7 or IL-15. Further- Some of these changes, e.g., the relatively high mRNA expression more, some properties of memory T cells, such as the down-reg- for TNF-␣, TNF-␤, 4-1BB, CD30, and OX40 by effector CTL and by guest on September 26, 2021 ulation of CD69 and CD25, might simply represent the absence of CD27 by memory-liked T cells, has been noted by others (46Ð50). IL-2 or Ag signals, as these changes were also favored when ef- More importantly, IL-15 down-regulated a number of genes im- fector CTL were simply cultured in medium. Given the array of portant for cell cycle progression and apoptosis that was accom- cellular and molecular modifications that occur as CTL transit to panied by a decrease in cell size. Correspondingly, cell cycle anal- memory cells, it seems likely that both passive and active cyto- ysis revealed fewer cells in the S-G2-M phase of the cell cycle for kine-dependent mechanisms are operative. these IL-15-stimulated memory-like T cells. Thus, IL-15 sup- The ability of both IL-7 or IL-15 to redundantly favor memory- ported alteration in the molecular profile of effector CTL to one like activities in vitro may be related to the fact that receptors for with lower potential for cell growth and apoptosis, properties of both cytokines induce overlapping signaling pathways, notably Ja- memory CD8ϩ T cells. nus kinase 1, Janus kinase 3, Stat3, and Stat5 (37Ð41). The mere Two models have been proposed to explain the development of induction of these pathways, however, cannot explain their efficacy memory CD8ϩ T cells. In one, effector CTL and memory T cells in promoting memory cell development. The IL-2 signaling path- are derived from separate lineages (9, 27). In the other the devel- way essentially entirely overlaps with that induced by IL-15, as opment of effector CTL is a prerequisite for production of memory both the IL-2 and IL-15 receptors share identical ␤ and common cells (25, 26, 51, 52). Our data support tenets inherent to both of ␥-chain subunits (42), yet continued culture with IL-2 favors ef- these models. We favor the view that, during the initial encounter fector rather than memory T cells. This dichotomy between IL-2 with Ag, a naive CD8ϩ T cell is programmed for expansion, con- and IL-15 signaling has been noted previously (43, 44). One ex- traction, and the potential to develop into an effector or memory planation for the distinct outcomes of IL-2 vs IL-15 or IL-7 for cells. The potential for a CTL to develop into memory cells exists CD8ϩ T cells may simply be related to the strength of signal, as for a discrete window of time. These fate outcomes of these two other studies have shown that the culture of Ag-activated CD8ϩ T cells are dictated by environmental factors, with the presence or cells with a suboptimal dose of IL-2 in vitro primarily yielded absence of Ag as one factor favoring effector vs memory cells, memory phenotypic T cells (27). The alternative possibility, that respectively. However, the mere absence of Ag may not in itself be cannot yet be entirely excluded, is that there is a unique signaling sufficient to promote memory cell development. Other factors, in- element, shared between IL-7 and IL-15, that favors the production cluding but not limited to IL-7 or IL-15, are necessary to either of memory CD8ϩ T cells. rescue the CTL from apoptosis and/or to induce properties of One very striking aspect of this study is the rapidity, i.e., days, memory cells. Our view predicts that, once activated with Ag, a by which both IL-7 or IL-15 supported features of memory CD8ϩ CD8ϩ T cell has the potential to travel down the memory cell T cells in vitro. Rapid changes toward memory phenotype were not pathway without becoming a CTL. This was readily visualized by just limited to in vitro culture, as several features of memory T removing Ag and then either culturing the Ag-activated T cells in cells were also noted 2Ð5 days after effector CTL were adoptively IL-7 or IL-15 or by adoptively transferring the cells into normal The Journal of Immunology 7323 mice. In the physiological situation of an immune response to an notherapy by persistent memory phenotypic CD8ϩ T cells. J. Immunol. infectious agent, removal of Ag requires a substantial immune re- 167:4511. 25. Wherry, E. J., V. Teichgraber, T. C. Becker, D. Masopust, S. M. Kaech, R. Antia, sponse, including effector CTL. Upon removing the bulk of Ag, U. H. von Andrian, and R. Ahmed. 2003. Lineage relationship and protective CTL that escape apoptosis then develop into memory cells. In immunity of memory CD8 T cell subsets. Nat. Immunol. 4:225. 26. Kaech, S. M., S. Hemby, E. Kersh, and R. Ahmed. 2002. Molecular and func- practical terms, memory cells do not develop until after a vigorous tional profiling of memory CD8 T cell differentiation. Cell 111:837. effector response which is responsible for decreasing Ag levels to 27. Manjunath, N., P. Shankar, J. Wan, W. Weninger, M. A. Crowley, K. Hieshima, a permissive level. Therefore, the linear differentiation model of T. A. Springer, X. Fan, H. Shen, J. Lieberman, and U. H. von Andrian. 2001. ϩ Effector differentiation is not prerequisite for generation of memory cytotoxic T naive to effector to memory CD8 T cells describes most immune lymphocytes. J. Clin. Invest. 108:871. responses in vivo. 28. 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. Acknowledgments 29. Dalyot-Herman, N., O. F. Bathe, and T. R. Malek. 2000. Reversal of CD8ϩ T cell We thank Lin Kong and Aixin Yu for technical assistance. ignorance and induction of anti-tumor immunity by peptide-pulsed APC. J. Im- munol. 165:6731. 30. Lyons, A. B. 2000. Analysing cell division in vivo and in vitro using flow cy- References tometric measurement of CFSE dye dilution. J. Immunol. Methods 243:147. 1. Sprent, J., and C. D. Surh. 2002. T cell memory. Annu. Rev. Immunol. 20:551. 31. Liu, B., E. R. Podack, J. P. Allison, and T. R. Malek. 1996. Generation of primary 2. Kaech, S. M., and R. Ahmed. 2001. Memory CD8ϩ T cell differentiation: initial tumor-specific CTL in vitro to immunogenic and poorly immunogenic mouse antigen encounter triggers a developmental program in naive cells. Nat. Immunol. tumors. J. Immunol. 156:1117. 2:415. 32. Clarke, S. R., M. Barnden, C. Kurts, F. R. Carbone, J. F. Miller, and W. R. Heath. 3. Badovinac, V. P., B. B. Porter, and J. T. Harty. 2002. Programmed contraction of 2000. Characterization of the ovalbumin-specific TCR transgenic line OT-I: ϩ

CD8 T cells after infection. Nat. Immunol. 3:619. MHC elements for positive and negative selection. Immunol. Cell Biol. 78:110. Downloaded from 4. van Stipdonk, M. J., E. E. Lemmens, and S. P. Schoenberger. 2001. Naive CTLs 33. Malek, T. R., A. Yu, P. Scibelli, M. G. Lichtenheld, and E. K. Codias. 2001. require a single brief period of antigenic stimulation for clonal expansion and Broad programming by IL-2 receptor signaling for extended growth to multiple differentiation. Nat. Immunol. 2:423. cytokines and functional maturation of antigen-activated T cells. J. Immunol. 5. Veiga-Fernandes, H., U. Walter, C. Bourgeois, A. McLean, and B. Rocha. 2000. 166:1675. ϩ Response of naive and memory CD8 T cells to antigen stimulation in vivo. Nat. 34. Cress, A. E. 2000. Quantitation of phosphotyrosine signals in human prostate cell Immunol. 1:47. adhesion sites. BioTechniques 29:776. 6. Opferman, J. T., B. T. Ober, and P. G. Ashton-Rickardt. 1999. Linear differen- 35. Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers, tiation of cytotoxic effectors into memory T lymphocytes. Science 283:1745. N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al. 2000. Reversible defects 7. Ahmed, R., and D. Gray. 1996. Immunological memory and protective immunity: in natural killer and memory CD8 T cell lineages in interleukin 15-deﬁcient mice. http://www.jimmunol.org/ understanding their relation. Science 272:54. J. Exp. Med. 191:771. 8. Murali-Krishna, K., J. D. Altman, M. Suresh, D. J. Sourdive, A. J. Zajac, 36. Judge, A. D., X. Zhang, H. Fujii, C. D. Surh, and J. Sprent. 2002. Interleukin 15 J. D. Miller, J. Slansky, and R. Ahmed. 1998. Counting antigen-speciﬁc CD8 T controls both proliferation and survival of a subset of memory-phenotype CD8ϩ cells: a reevaluation of bystander activation during viral infection. Immunity T cells. J. Exp. Med. 196:935. 8:177. 37. Zeng, Y. X., H. Takahashi, M. Shibata, and K. Hirokawa. 1994. JAK3 Janus 9. Jacob, J., and D. Baltimore. 1999. Modelling T-cell memory by genetic marking kinase is involved in interleukin 7 signal pathway. FEBS Lett. 353:289. of memory T cells in vivo. Nature 399:593. 38. Sharfe, N., H. K. Dadi, and C. M. Roifman. 1995. JAK3 protein tyrosine kinase 10. Ku, C. C., M. Murakami, A. Sakamoto, J. Kappler, and P. Marrack. 2000. Control Ј ϩ mediates interleukin-7-induced activation of phosphatidylinositol-3 kinase. of homeostasis of CD8 memory T cells by opposing cytokines. Science Blood 86:2077. 288:675. 39. Johnston, J. A., C. M. Bacon, D. S. Finbloom, R. C. Rees, D. Kaplan, K. Shibuya, 11. Zhang, X., S. Sun, I. Hwang, D. F. Tough, and J. Sprent. 1998. Potent and

ϩ J. R. Ortaldo, S. Gupta, Y. Q. Chen, J. D. Giri, et al. 1995. Tyrosine phosphor- by guest on September 26, 2021 selective stimulation of memory-phenotype CD8 T cells in vivo by IL-15. Im- ylation and activation of STAT5, STAT3, and Janus kinases by interleukins 2 and munity 8:591. 15. Proc. Natl. Acad. Sci. USA 92:8705. 12. Schluns, K. S., W. C. Kieper, S. C. Jameson, and L. Lefrancois. 2000. Interleu- 40. Foxwell, B. M., C. Beadling, D. Guschin, I. Kerr, and D. Cantrell. 1995. Inter- kin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nat. leukin-7 can induce the activation of Jak 1, Jak 3 and STAT 5 proteins in murine Immunol. 1:426. T cells. Eur. J. Immunol. 25:3041. 13. Becker, T. C., E. J. Wherry, D. Boone, K. Murali-Krishna, R. Antia, A. Ma, and 41. Tagaya, Y., J. D. Burton, Y. Miyamoto, and T. A. Waldmann. 1996. Identifica- R. Ahmed. 2002. Interleukin 15 is required for proliferative renewal of virus- tion of a novel receptor/signal transduction pathway for IL-15/T in mast cells. specific memory CD8 T cells. J. Exp. Med. 195:1541. EMBO J. 15:4928. 14. Goldrath, A. W., P. V. Sivakumar, M. Glaccum, M. K. Kennedy, M. J. Bevan, 42. Waldmann, T., Y. Tagaya, and R. Bamford. 1998. Interleukin-2, interleukin-15, C. Benoist, D. Mathis, and E. A. Butz. 2002. Cytokine requirements for acute and ϩ and their receptors. Int. Rev. Immunol. 16:205. basal homeostatic proliferation of naive and memory CD8 T cells. J. Exp. Med. 43. Lodolce, J., P. Burkett, R. Koka, D. Boone, M. Chien, F. Chan, M. Madonia, 195:1515. S. Chai, and A. Ma. 2002. Interleukin-15 and the regulation of lymphoid ho- 15. Kieper, W. C., J. T. Tan, B. Bondi-Boyd, L. Gapin, J. Sprent, R. Ceredig, and meostasis. Mol. Immunol. 39:537. C. D. Surh. 2002. Overexpression of interleukin (IL)-7 leads to IL-15-independent generation of memory phenotype CD8ϩ T cells. J. Exp. Med. 195:1533. 44. Waldmann, T. A., S. Dubois, and Y. Tagaya. 2001. Contrasting roles of IL-2 and 16. Tan, J. T., B. Ernst, W. C. Kieper, E. LeRoy, J. Sprent, and C. D. Surh. 2002. IL-15 in the life and death of lymphocytes: implications for immunotherapy. Interleukin (IL)-15 and IL-7 jointly regulate homeostatic proliferation of memory Immunity 14:105. ϩ ϩ 45. Grayson, J. M., K. Murali-Krishna, J. D. Altman, and R. Ahmed. 2001. Gene phenotype CD8 cells but are not required for memory phenotype CD4 cells. ϩ J. Exp. Med. 195:1523. expression in antigen-specific CD8 T cells during viral infection. J. Immunol. 17. Weninger, W., M. A. Crowley, N. Manjunath, and U. H. von Andrian. 2001. 166:795. Migratory properties of naive, effector, and memory CD8ϩ T cells. J. Exp. Med. 46. Bilsborough, J., C. Uyttenhove, D. Colau, P. Bousso, C. Libert, B. Weynand, T. Boon, and B. J. van den Eynde. 2002. TNF-mediated toxicity after massive 194:953. ϩ 18. Sallusto, F., D. Lenig, R. Forster, M. Lipp, and A. Lanzavecchia. 1999. Two induction of specific CD8 T cells following immunization of mice with a tumor- subsets of memory T lymphocytes with distinct homing potentials and effector specific peptide. J. Immunol. 169:3053. functions. Nature 401:708. 47. Tan, J. T., J. K. Whitmire, R. Ahmed, T. C. Pearson, and C. P. Larsen. 1999. 19. Walzer, T., C. Arpin, L. Beloeil, and J. Marvel. 2002. Differential in vivo per- 4-1BB ligand, a member of the TNF family, is important for the generation of sistence of two subsets of memory phenotype CD8 T cells defined by CD44 and antiviral CD8 T cell responses. J. Immunol. 163:4859. CD122 expression levels. J. Immunol. 168:2704. 48. Cannons, J. L., P. Lau, B. Ghumman, M. A. DeBenedette, H. Yagita, 20. Murali-Krishna, K., L. L. Lau, S. Sambhara, F. Lemonnier, J. Altman, and K. Okumura, and T. H. Watts. 2001. 4-1BB ligand induces cell division, sustains R. Ahmed. 1999. Persistence of memory CD8 T cells in MHC class I-deficient survival, and enhances effector function of CD4 and CD8 T cells with similar mice. Science 286:1377. efficacy. J. Immunol. 167:1313. 21. Goldrath, A. W., L. Y. Bogatzki, and M. J. Bevan. 2000. Naive T cells transiently 49. Sun, Y., M. Song, M. J. Maeurer, and D. Schadendorf. 2001. Segregation of ϩ acquire a memory-like phenotype during homeostasis-driven proliferation. effector mechanisms in a tumour-specific CD8 T-cell clone correlates with J. Exp. Med. 192:557. CD30 expression. Scand. J. Immunol. 54:314. 22. Walunas, T. L., D. S. Bruce, L. Dustin, D. Y. Loh, and J. A. Bluestone. 1995. 50. Hendriks, J., L. A. Gravestein, K. Tesselaar, R. A. van Lier, T. N. Schumacher, Ly-6C is a marker of memory CD8ϩ T cells. J. Immunol. 155:1873. and J. Borst. 2000. CD27 is required for generation and long-term maintenance 23. Curtsinger, J. M., D. C. Lins, and M. F. Mescher. 1998. CD8ϩ memory T cells of T cell immunity. Nat. Immunol. 1:433. (CD44high, Ly-6Cϩ) are more sensitive than naive cells to (CD44low, Ly-6CϪ)to 51. Farber, D. L. 1998. Differential TCR signaling and the generation of memory T TCR/CD8 signaling in response to antigen. J. Immunol. 160:3236. cells. J. Immunol. 160:535. 24. Bathe, O. F., N. Dalyot-Herman, and T. R. Malek. 2001. IL-2 during in vitro 52. Sallusto, F., and A. Lanzavecchia. 2001. Exploring pathways for memory T cell priming promotes subsequent engraftment and successful adoptive tumor immu- generation. J. Clin. Invest. 108:805.