The Dual Role of IL-2 in the Generation and Maintenance of CD8 + Memory T Cells Zhenhua Dai, Bogumila T. Konieczny and Fadi G. Lakkis This information is current as J Immunol 2000; 165:3031-3036; ; of September 30, 2021. doi: 10.4049/jimmunol.165.6.3031 http://www.jimmunol.org/content/165/6/3031 Downloaded from References This article cites 31 articles, 17 of which you can access for free at: http://www.jimmunol.org/content/165/6/3031.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/ • 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 by guest on September 30, 2021 Subscription Information about subscribing to The Journal of 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 © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Dual Role of IL-2 in the Generation and Maintenance of CD8؉ Memory T Cells1

Zhenhua Dai, Bogumila T. Konieczny, and Fadi G. Lakkis2

The mechanisms responsible for the generation and maintenance of memory are unclear. In this study, we tested the role of IL-2 in allospecific CD8؉ T cell memory by analyzing the long-term survival, phenotype, and functional characteristics of IL-2-replete (IL-2؉/؉) and IL-2-deficient (IL-2؊/؊) CD8؉ TCR-transgenic in an adoptive transfer model. We found that IL-2 is not essential for the in vivo generation, maintenance, or recall response of CD8؉ memory T cells. However, IL-2 increased the size of the CD8؉ memory pool if present at the time of initial T cell activation but reduced the size of the pool if present during memory maintenance by inhibiting the proliferation of CD8؉ memory T cells. Thus, IL-2-based vaccine strategies or immunosuppressive regimens that target IL-2 should take into account the divergent roles of IL-2 in CD8؉ T cell

immunity. The Journal of Immunology, 2000, 165: 3031–3036. Downloaded from

defining feature of the adaptive immune response is its In this study, we tested whether IL-2 is required for generating ability to generate memory lymphocytes (1–3). Upon and maintaining CD8ϩ memory T cells by analyzing the long-term encounter with a foreign Ag, Ag-specific T lymphocytes survival, phenotype, and functional characteristics of Ag-activated A ϩ/ϩ Ϫ/Ϫ proliferate and differentiate into effector cells. The majority of ef- IL-2-replete (IL-2 ) (3) and IL-2 gene-knockout (IL-2 ) fector cells undergo after the Ag is eliminated (4). The CD8ϩ TCR-transgenic (TCR-tg)3 lymphocytes (2C) in an adop- http://www.jimmunol.org/ few that survive become memory cells that persist for a long pe- tive transfer model. The 2C TCR recognizes the Ld MHC class I riod of time, sometimes throughout the life of an animal (5–7). Ag and can be tagged by a clonotypic Ab (1B2) that permits the Memory T cells are specific to the Ag they encountered during the detection of small numbers of Ag-specific cells in vivo (20). We primary immune response and react rapidly upon re-encounter demonstrate here that CD8ϩ memory T cells can be efficiently with the same Ag. generated and maintained in the absence of IL-2. However, IL-2 The mechanisms responsible for the persistence of T cell mem- increased the size of the CD8ϩ memory population if present dur- ory are unclear. It appears that memory T cell populations are ing T cell activation but reduced its size if present during the main- maintained through the homeostatic replication of a subgroup of tenance period. These findings indicate that IL-2 has a dual role in ϩ memory cells and through the intrinsic ability of some memory the generation and maintenance of CD8 T cell memory. by guest on September 30, 2021 cells to survive in the resting state for an extended duration (8, 9). Although earlier studies suggested that continued antigenic stim- Materials and Methods ulation is required for maintaining T cell memory (9–12), recent Mice ϩ ϩ evidence indicates that CD4 and CD8 memory T cell popula- Ϫ Ϫ Ϫ Ϫ BALB/c (H-2d), C57BL/6 (H-2b) Rag1 / , and C57BL/6 IL-2 / mice tions persist in the absence of specific or cross-reactive Ags pre- were purchased from The Jackson Laboratory (Bar Harbor, ME). C57BL/6 sented by MHC molecules (13–15). Therefore, it is possible that 2C TCR-tg mice were provided by Dr. Dennis Loh and bred at the Veterans Ag-independent factors, such as , are critical for the sur- Affairs Medical Center animal facility (Atlanta, GA) (20). IL-2-deficient Ϫ/Ϫ vival and proliferation of memory T cells. IL-2 enhances the sur- 2C mice were generated by cross-breeding C57BL/6 IL-2 mice with C57BL/6 2C mice. Genotyping was performed by PCR amplification of vival of naive T cells and stimulates the proliferation of primary DNA extracted from tail clippings. Transgenic and gene-knockout mice activated lymphocytes (16, 17). In addition, it has been proposed were housed under specific pathogen-free conditions. that IL-2 is required for generating and maintaining memory T ϩ cells (18, 19). This hypothesis, however, is challenged by the find- In vitro generation of effector CD8 T cells ings that IL-2 programs T cells for activation-induced apoptosis Splenocytes were isolated from IL-2ϩ/ϩ and IL-2Ϫ/Ϫ 2C mice and stimu- and that IL-2Ϫ/Ϫ mice are not immunodeficient but instead display lated in vitro with mitomycin C-treated BALB/c splenocytes in RPMI 1640 exaggerated T cell immunity (4, 17). medium supplemented with 1 mM L-glutamine, 1% sodium pyruvate, 50 ␮M 2-ME, 100 U/ml penicillin, 100 ␮g/ml streptomycin (all from Life Technologies, Grand Island, NY), and 10% FCS (Sigma, St. Louis, MO). A total of 50 U/ml recombinant mouse IL-2 (Genzyme, Cambridge, MA) The Carlos and Marguerite Mason Transplantation Research Center, Renal Division, was added to all cultures. After 96 h of mixed culture, live Department of Medicine, Emory University and Veterans Affairs Medical Center, cells were isolated by ficoll density centrifugation using Lympholyte-M Atlanta, GA 30033 (Cedarlane, Hornby, Ontario) and were subsequently enriched for T cells by nonadherence to nylon wool columns (Polysciences, Warrington, PA). Received for publication April 17, 2000. Accepted for publication June 28, 2000. IL-2ϩ/ϩ and IL-2Ϫ/Ϫ 2C T cell-enriched preparations (a total of ϳ2.5 ϫ The costs of publication of this article were defrayed in part by the payment of page 107 2C T cells in each) containing 5 ϫ 106 CD8ϩ1B2ϩ cells (quantitated charges. This article must therefore be hereby marked advertisement in accordance by two-color flow cytometry as described below) were mixed with 3-fold with 18 U.S.C. Section 1734 solely to indicate this fact. more (ϳ7.5 ϫ 107) -enriched IL-2ϩ/ϩ and IL-2Ϫ/Ϫ non- 1 This work was supported by National Institutes of Health Grants AI41643 (to TCR-tg C57BL/6 splenocytes, respectively, and were then adoptively F.G.L.) and AI44644 (to F.G.L.), and the Carlos and Marguerite Mason Trust for transferred by i.v. injection into Rag1Ϫ/Ϫ mice. Control Rag1Ϫ/Ϫ mice Transplantation (to F.G.L.). were adoptively transferred with 5 ϫ 106 naive IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2 Address correspondence and reprint requests to Dr. Fadi G. Lakkis, Veterans Affairs Medical Center and Emory University, Research 151N, 1670 Clairmont Road, At- lanta, GA 30033. E-mail address: [email protected] 3 Abbreviations used in this paper: tg, transgenic; BrdU, 5-bromo-2-deoxyuridine.

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00 3032 THE ROLE OF IL-2 IN CD8ϩ T CELL MEMORY

CD8ϩ1B2ϩ cells mixed with ϳ7.5 ϫ 107 naive T cell-enriched IL-2ϩ/ϩ or 5-Bromo-2-deoxyuridine (BrdU) labeling Ϫ/Ϫ IL-2 non-TCR-tg C57BL/6 splenocytes, respectively. 2C lymphocytes ϩ/ϩ Ϫ/Ϫ were mixed with non-TCR-tg, syngeneic splenocytes before transfer to Naive IL-2 or IL-2 2C T cell-enriched splenocyte preparations that contain 5 ϫ 106 CD8ϩ1B2ϩ cells were mixed with ϳ7.5 ϫ 107 naive T simulate physiologic settings in which Ag-specific cells are present at ϩ/ϩ Ϫ/Ϫ much lower frequency than in TCR-tg mice. Mice were sacrificed at either cell-enriched IL-2 or IL-2 non-TCR-tg splenocytes, respectively, and were subsequently adoptively transferred to Rag1Ϫ/Ϫ mice by i.v. in- 1 wk or 10 wk after adoptive transfer and spleen and lymph node cells Ϫ/Ϫ were harvested to test for the presence of CD8ϩ1B2ϩ cells. Memory jection. Twenty-four hours later, the Rag1 hosts were immunized i.p. cells were identified according to the following criteria: 1) survival ad- with mitomycin C-treated BALB/c splenocytes. At 10 wk, 0.8 mg/ml BrdU vantage over naive cells; 2) high CD44 and low CD62L expression was added to the drinking water for 7 days, after which the mice were high low sacrificed and lymph node cells were harvested. A separate group of (CD44 CD62L ); and 3) ability to mount a recall CTL response. Ab- Ϫ/Ϫ Ϫ/Ϫ sence of IL-2 production in Rag1Ϫ/Ϫ hosts that received IL-2Ϫ/Ϫ lympho- Rag1 mice harboring IL-2 2C memory lymphocytes was injected ␮ cytes was confirmed by 35 cycles of RT-PCR amplification of splenic and with 5 g recombinant murine IL-2 i.p. twice during the BrdU labeling lymph node mRNA. period. Surface staining of lymph node cells with 1B2 and anti-CD8 was performed as described in the previous sections. Cells were then fixed in 70% ethanol followed by 1% paraformaldehyde and incubated with 50 In vivo generation of memory T cells Kunitz/ml of DNAase I (Sigma) for 10 min at room temperature. The cells

ϩ/ϩ Ϫ/Ϫ were subsequently stained with FITC-conjugated anti-BrdU Ab (Becton Naive IL-2 or IL-2 2C T cell-enriched splenocyte preparations (a Dickinson) and three-color flow analysis was performed. The percentage of ϳ ϫ 7 ϫ 6 ϩ ϩ total of 2.5 10 2C T cells in each) that contain 5 10 CD8 1B2 BrdUϩ cells was determined after gating on the CD8ϩ1B2ϩ population. cells were mixed with 3-fold more (ϳ7.5 ϫ 107) naive T cell-enriched IL-2ϩ/ϩ or IL-2Ϫ/Ϫ non-TCR-tg splenocytes, respectively, and were sub- sequently adoptively transferred to Rag1Ϫ/Ϫ mice by i.v. injection. Twen- ty-four hours later, the Rag1Ϫ/Ϫ hosts were immunized i.p. with mitomycin Results d Downloaded from C-treated BALB/c splenocytes (which express the L MHC class I Ag IL-2 is not required for the maintenance of CD8ϩ memory T recognized by the 2C TCR). Control mice were injected with PBS i.p. Ten cells weeks after , mice were sacrificed and lymph node cells and ϩ ϩ spleen cells were harvested to test for the presence of CD8 1B2 cells. To study the role of IL-2 in the maintenance of Ag-specific CD8ϩ Memory cells were identified according to the same criteria described in T cell memory, IL-2ϩ/ϩ and IL-2Ϫ/Ϫ 2C lymphocytes were acti- the previous section. To test the influence of exogenous IL-2 on the generation of CD8ϩ vated in vitro in the presence of excess IL-2 and were then trans- Ϫ/Ϫ Ϫ/Ϫ

memory T cells, 5 ␮g recombinant murine IL-2 (Genzyme, Boston, MA) ferred to Rag1 hosts. Control Rag1 mice were adoptively http://www.jimmunol.org/ was administered i.p. to Rag1Ϫ/Ϫ mice harboring IL-2Ϫ/Ϫ 2C lymphocytes transferred with naive IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes. Both on days 0 and 2 after immunization with BALB/c splenocytes. Control activated and naive 2C lymphocytes were mixed with naive, non- mice received PBS injections i.p. On day 4, lymph node cells were har- vested and adoptively transferred to a second Rag1Ϫ/Ϫ host. Ten weeks TCR-tg, syngeneic T lymphocytes before adoptive transfer to sim- later, CD8ϩ1B2ϩ cells were quantitated as described in the previous para- ulate physiologic settings in which Ag-specific T cells are present graph. To test the influence of IL-2 on the maintenance of CD8ϩ memory at much lower frequency than in TCR-tg mice. Because 2C lym- Ϫ Ϫ Ϫ Ϫ T cells, Rag1 / mice harboring IL-2 / 2C lymphocytes were immu- phocytes can be identified by a clonotypic Ab (1B2), this experi- nized with BALB/c splenocytes in the absence of exogenous IL-2. Four mental model allowed us to compare the in vivo survival, pheno- days later, lymph node cells were harvested and adoptively transferred to ϩ a second Rag1Ϫ/Ϫ host which was subsequently injected i.p. with 5 ␮g type, and functional characteristics of previously activated CD8 recombinant IL-2 twice weekly. Ten weeks later, CD8ϩ1B2ϩ cells were T cells to that of naive cells in the presence or absence of IL-2. In by guest on September 30, 2021 quantitated as described in the previous paragraph. the experiments that follow, similar findings were obtained whether lymph node or spleen cells were analyzed. Therefore, only Flow cytometry the results of lymph node analysis are shown. ϩ ϩ One week after adoptive transfer, the number of previously ac- Adoptively transferred 2C lymphocytes (CD8 1B2 ) were identified in ϩ/ϩ Ϫ/Ϫ ϩ ϩ Rag1Ϫ/Ϫ hosts by incubating pooled lymph nodes cells or splenocyte with tivated IL-2 or IL-2 2C (CD8 1B2 ) lymphocytes present Ϫ Ϫ PE-conjugated rat anti-mouse CD8a (53–6.7; PharMingen, San Diego, in the lymph nodes of Rag1 / hosts was similar to that of naive CA), mouse IgG1 anti-2C TCR clonotypic Ab (1B2), and FITC-conjugated IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes (Fig. 1, A and B). No signif- rat anti-mouse IgG1 (PharMingen) or, in some experiments, biotinylated icant difference in the number of CD8ϩ1B2ϩ cells was observed rat anti-mouse IgG1 followed by streptavidin-PerCP (PharMingen). To de- ϩ/ϩ Ϫ/Ϫ tect cell surface markers, the following Abs (all from PharMingen) were between the IL-2 and IL-2 2C populations (Fig. 1, A and used: FITC-conjugated rat anti-mouse CD44 (1 M7) and FITC-conjugated B). However, at 10 wk after adoptive transfer, the number of pre- rat anti-mouse CD62L (MEL-14). The appropriate isotype control Abs viously activated IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes signifi- were used as negative controls (all from PharMingen). Stained cells were cantly exceeded that of naive IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes washed and analyzed by two- or three-color flow cytometry on a FACScan (Fig. 1, A and B), indicating that previously activated lymphocytes (Becton Dickinson, Mountain View, CA). The CD44 and CD62L pheno- type of naive and memory T cells was determined by gating on CD8ϩ1B2ϩ have a survival advantage over their naive counterparts. Like the ϩ ϩ cells. naive 2C TCR-tg (CD8 1B2 ) population, the naive non-TCR-tg (CD8ϩ1B2Ϫ) population had diminished dramatically by 10 wk CTL activity (Fig. 1A). Moreover, at 10 wk after adoptive transfer, previously Ϫ/Ϫ Ϫ/Ϫ ϩ/ϩ activated IL-2 2C lymphocytes were present in significantly Rag1 mice were adoptively transferred with naive or activated IL-2 ϩ/ϩ Ϫ Ϫ greater numbers than previously activated IL-2 2C lympho- and IL-2 / 2C lymphocytes as described above and challenged in the footpads with mitomycin C-treated BALB/c splenocytes 10 wk later. After cytes (Fig. 1, A and B). These findings suggest that long-lived, ϩ Ϫ Ϫ 3 days, popliteal and inguinal lymph node cells were pooled and immedi- Ag-specific, CD8 memory T cells are present in Rag1 / mice ately analyzed in an ex vivo CTL assay. Allospecific CTL activity was adoptively transferred with previously activated 2C lymphocytes, d measured by incubating the lymph node cells with either P815 (H-2 ) tar- and that the number of memory cells is increased, rather than di- get cells (American Type Culture Collection (ATCC), Manassas, VA) or third party cells LK35.2 (H-2k) (ATCC) for 3 h. Target cells were pre- minished, in the absence of IL-2. ϩ/ϩ loaded with calcein-AM (Molecular Probes, Eugene, OR), and calcein re- To confirm that long-lived, previously activated IL-2 and lease was measured in a LS50B luminescence spectrometer (Perkin-Elmer, IL-2Ϫ/Ϫ 2C lymphocytes are memory cells, we tested their phe- Foster City, CA) (21). Experiments in which spontaneous calcein release notypic and functional characteristics. As shown in Fig. 1C, pre- was more than 25% of maximal release were excluded. Ag-specific cyto- ϩ/ϩ Ϫ/Ϫ toxic activity was calculated according to the following formula: % specific viously activated IL-2 and IL-2 2C lymphocytes, present at lysis ϭ 100 ϫ [(sample release Ϫ spontaneous release)/(maximum re- 10 wk after adoptive transfer, express high levels of CD44 lease Ϫ spontaneous release)]. (CD44high) and low levels of CD62L (CD62Llow), a phenotype The Journal of Immunology 3033 Downloaded from http://www.jimmunol.org/ by guest on September 30, 2021

FIGURE 1. Long-term survival of CD8ϩ memory T cells in the absence of IL-2. Naive or in vitro-activated IL-2ϩ/ϩ and IL-2Ϫ/Ϫ 2C lymphocytes (5 ϫ 106 CD8ϩ1B2ϩ cells) were mixed with ϳ7.5 ϫ 107 naive non-TCR-tg IL-2ϩ/ϩ and IL-2Ϫ/Ϫ T cells, respectively, and were adoptively transferred to syngeneic Rag1Ϫ/Ϫ mice. At 1 or 10 wk after transfer, Rag1Ϫ/Ϫ hosts were sacrificed and lymph node cells were analyzed by flow cytometry for the presence of CD8ϩ1B2ϩ cells. A, Flow cytometric analysis showing the percentage of naive or previously activated IL-2ϩ/ϩ and IL-2Ϫ/Ϫ CD8ϩ1B2ϩ cells present at 1 and 10 wk after adoptive transfer. Density plots shown are representative of four experiments. B, Absolute number (mean Ϯ SD, n ϭ 4) of ;p Ͻ 0.05 compared with naive cells ,ء .naive or previously activated IL-2ϩ/ϩ and IL-2Ϫ/Ϫ CD8ϩ1B2ϩ cells present at 1 and 10 wk after adoptive transfer †, p Ͻ 0.05 compared with IL-2ϩ/ϩ activated cells. C, Flow cytometric analysis of CD62L and CD44 expression on naive (solid histograms) or previously activated (open histograms) IL-2ϩ/ϩ and IL-2Ϫ/Ϫ CD8ϩ1B2ϩ cells present at 10 wk after adoptive transfer. Histograms shown are representative of four experiments. D, CTL recall response of Rag1Ϫ/Ϫ hosts at 10 wk after adoptive transfer. Ex vivo CTL activity was measured 3 days after challenging Rag1Ϫ/Ϫ hosts with BALB/c splenocytes (mean Ϯ SD, n ϭ 4).

associated with CD8ϩ memory T cells in mice. We then asked Fig. 1D, Ag-specific stimulation induced a rapid and strong CTL whether a recall immune response can be elicited in Rag1Ϫ/Ϫ mice response in mice that harbored previously activated IL-2ϩ/ϩ or that received previously activated 2C lymphocytes. As shown in IL-2Ϫ/Ϫ lymphocytes but not in those that harbored naive cells. 3034 THE ROLE OF IL-2 IN CD8ϩ T CELL MEMORY

These findings prove that neither the long-term maintenance nor weekly injections of either PBS or recombinant IL-2. In both ex- the recall response of CD8ϩ memory T cells is dependent on IL-2. periments, the number of CD8ϩ1B2ϩ was quantitated 10 wk after

ϩ lymphocyte activation. As shown in Fig. 3A, administering IL-2 to IL-2 is not required for in vivo generation of CD8 memory T mice at the time of IL-2Ϫ/Ϫ 2C lymphocyte activation resulted in cells a significantly larger number of long-lived CD8ϩ1B2ϩ cells (38% To test whether IL-2 is essential for generating CD8ϩ T cell mem- increase over mice injected with PBS), indicating that IL-2 poten- ϩ ory, we adoptively transferred naive IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lym- tiates the generation of CD8 T cell memory. In contrast, repeated phocytes to Rag1Ϫ/Ϫ hosts. As in the previous model, naive 2C administration of IL-2 to mice harboring previously activated IL- Ϫ Ϫ lymphocytes were mixed with naive, non-TCR-tg, syngeneic T 2 / 2C lymphocytes resulted in a significantly smaller number of ϩ ϩ lymphocytes before adoptive transfer. The adoptively transferred long-lived CD8 1B2 cells (30% decrease compared with mice mice were then challenged with either PBS or Ld-bearing alloge- injected with PBS) (Fig. 3B), suggesting that IL-2 interferes with ϩ neic splenocytes. Ten weeks later, lymph node cells were analyzed the maintenance of CD8 T cell memory. Unlike its effects on by flow cytometry. A significantly larger number of CD8ϩ1B2ϩ memory T cells, IL-2 administration did not alter the number of ϩ ϩ cells was detected in allostimulated than in naive mice (Fig. 2). naive CD8 1B2 T cells (Fig. 3, A and B). The number of long-lived CD8ϩ1B2ϩ cells was comparable in The T cell memory pool is maintained by the continuous but mice that harbored either IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes. slow division of memory cells (8, 9, 14). Therefore, we asked Moreover, long-lived CD8ϩ1B2ϩ cells present in allostimulated whether IL-2 interferes with memory maintenance by inhibiting ϩ mice, but not those present in naive mice, exhibited a the proliferation of CD8 memory T cells. To answer this ques- Ϫ/Ϫ CD44highCD62Llow phenotype (data not shown). These findings tion, BrdU was added for 7 days to the drinking water of Rag1 Downloaded from ϩ ϩ Ϫ Ϫ indicate that CD8ϩ memory T cells can be generated in vivo in the mice harboring either IL-2 / or IL-2 / 2C memory lympho- ϩ ϩ absence of IL-2. Similar findings were obtained when spleen cells cytes, and BrdU uptake by Ag-specific (CD8 1B2 ) cells was were anlayzed (data not shown). quantitated by flow analysis. As shown in Fig. 3C, the percentage of BrdUϩ cells was 2-fold higher in IL-2Ϫ/Ϫ than IL-2ϩ/ϩ mem- The effects of exogenous IL-2 on the in vivo generation and ory cells. Importantly, administering recombinant IL-2 to mice ϩ Ϫ/Ϫ maintenance of CD8 memory T cells harboring IL-2 2C memory lymphocytes reduced BrdU uptake http://www.jimmunol.org/ We have observed in this study that the size of the CD8ϩ T cell significantly (Fig. 3C). These data indicate that IL-2 interferes with memory population is increased if 2C lymphocytes are activated in memory maintenance by decreasing the proliferation rate of the ϩ vitro in the presence of IL-2 and then maintained in vivo in its CD8 memory T cell population. absence (Fig. 1). However, when 2C lymphocytes were activated and maintained in vivo in the absence of IL-2, the size of the resulting CD8ϩ T cell memory population was similar to that ob- Discussion served if IL-2 was present (Fig. 2). These findings suggest that The goal of this study was to investigate the role of IL-2 in the IL-2 has divergent effects on the generation and maintenance of generation and maintenance of CD8ϩ memory T cells. To do so, memory T cells. To study the effect of IL-2 on memory generation, we followed the fate of activated, Ag-specific, IL-2ϩ/ϩ and IL- by guest on September 30, 2021 IL-2Ϫ/Ϫ 2C lymphocytes were activated in vivo in the presence or 2Ϫ/Ϫ TCR-tg CD8ϩ T cells (2C) in a lymphocyte-deficient host. absence of exogenous IL-2 and were then maintained in a second We found that long-lived populations of previously activated 2C Rag1Ϫ/Ϫ host without further IL-2 administration. To study the lymphocytes, which have phenotypic and functional characteristics effect of IL-2 on memory maintenance, IL-2Ϫ/Ϫ 2C lymphocytes of memory cells, can be efficiently generated and maintained in the were activated in vivo in the absence of exogenous IL-2 and were absence of IL-2. However, IL-2 increased the size of the CD8ϩ then maintained in a second Rag1Ϫ/Ϫ host that received twice memory pool if present at the time of initial T cell activation but

FIGURE 2. In vivo generation of CD8ϩ memory T cells in the absence of IL-2. Naive IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes (5 ϫ 106 CD8ϩ1B2ϩ cells) were mixed with ϳ7.5 ϫ 107 naive non-TCR-tg IL-2ϩ/ϩ and IL-2Ϫ/Ϫ T cells, respectively, and were adoptively transferred to syngeneic Rag1Ϫ/Ϫ mice. Twenty-four hours later, Rag1Ϫ/Ϫ hosts were challenged with allogeneic BALB/c splenocytes (activated) or PBS (naive). Ten weeks after challenge, mice were sacrificed and lymph node cells were analyzed by flow cytometry for the presence of CD8ϩ1B2ϩ cells. Percentage (A) and absolute number (B)of .p Ͻ 0.05 compared with naive cells ,ء .(CD8ϩ1B2ϩ cells present at 10 wk are shown (mean Ϯ SD, n ϭ 3 The Journal of Immunology 3035

of TCR-tg lymphocytes to lymphocyte-deficient hosts is an estab- lished approach to studying CD4ϩ and CD8ϩ T cell memory in vivo (2, 3) as it allows for the quantitation and analysis of Ag- specific T cells in a manner which is not possible in non-TCR-tg systems. The dual role of IL-2 in the generation and maintenance of CD8ϩ T cell memory is consistent with its divergent effects on naive/primary activated T cells and repeatedly activated T cells. IL-2 promotes the survival of naive T cells and enhances their clonal expansion upon primary antigenic stimulation (22, 23). In contrast, IL-2 promotes the Ag-induced apoptosis of repeatedly activated T cells and thus contributes to the clonal contraction of effector lymphocytes (24–28). Our finding that IL-2 enhances memory generation if present during initial T cell activation is consistent with the hypothesis that the size of the memory pool is directly proportional to the magnitude of T cell clonal expansion during the primary immune response (1). However, we identified in this study an inhibitory effect of IL-2 on the proliferation of ϩ CD8 memory T cells which leads to a smaller memory pool. Downloaded from Although it is possible that IL-2 decreases BrdU uptake by mem- ory T cells by promoting activation-induced cell death (AICD), this is an unlikely event because memory T cells are thought to be resistant to AICD (2). Furthermore, our finding that IL-2 sup- presses rather than stimulates memory T cell proliferation is con- FIGURE 3. IL-2 modulates the generation and maintenance of CD8ϩ Ϫ/Ϫ Ϫ/Ϫ sistent with a recent report showing that in vivo inhibition of IL-2 http://www.jimmunol.org/ memory T cells. A, Rag1 mice harboring IL-2 2C lymphocytes re- ϩ Ϫ Ϫ increases the cycling of memory-phenotype CD8 T cells in mice ceived either recombinant murine IL-2 (IL-2 / (ϩIL-2)) or PBS (IL- 2Ϫ/Ϫ) during in vivo T cell activation. Four days later, lymph node cells (29). Therefore, the results suggest that IL-2 could serve as a vac- were adoptively transferred to a second Rag1Ϫ/Ϫ host and CD8ϩ1B2ϩ cine adjuvant if used at the time of primary antigenic stimulation cells were quantitated by flow analysis at 10 wk (mean Ϯ SD, n ϭ 3). The but will lead to a paradoxical decline in the memory population if effect of exogenous IL-2 on mice harboring naive IL-2Ϫ/Ϫ 2C lymphocytes used during the posteffector phase of the immune response. Con- p Ͻ 0.05. B, Rag1Ϫ/Ϫ mice harboring previ- versely, when designing immunosuppresive or tolerance-inducing ,ء .is shown for comparison Ϫ Ϫ ously activated IL-2 / 2C lymphocytes were given either recombinant strategies, one should take into account the possibility that long- ϩ ϩ murine IL-2 or PBS twice weekly. Ten weeks later, CD8 1B2 cells were term IL-2 blockade could amplify the pool of unwanted allospe- quantitated by flow analysis (mean Ϯ SD, n ϭ 3). The effect of exogenous by guest on September 30, 2021 Ϫ/Ϫ cific memory T cells. It is important to note, however, that the net IL-2 on mice harboring naive IL-2 2C lymphocytes is shown for com- effect of endogenous IL-2 on the size of the CD8ϩ memory pool p Ͻ 0.05. C, Ten weeks after in vivo T cell activation, BrdU was ,ء .parison Ϫ Ϫ ϩ ϩ in an immunized animal is negligible (Fig. 2), possibly because the added to the drinking water of Rag1 / mice harboring either IL-2 / or IL-2Ϫ/Ϫ 2C memory lymphocytes. Seven days later, BrdU uptake was enhancing effect of IL-2 on memory generation is offset by its measured by flow analysis after gating on the CD8ϩ1B2ϩ population. A inhibitory effect on memory maintenance. separate group of Rag1Ϫ/Ϫ mice harboring IL-2Ϫ/Ϫ 2C memory lympho- Our demonstration that IL-2 is not required for the maintenance ϩ cytes was injected with recombinant murine IL-2 during the BrdU labeling of CD8 T cell memory contradicts the results of Ke et al. (18), period (IL-2Ϫ/Ϫ (ϩIL-2)). Histograms shown are representative of three who observed that the number of CD8ϩ memory T cells declines experiments. significantly in the absence of IL-2. Unlike our study, Ke et al. employed an in vivo model based on the adoptive transfer of an IL-2-dependent CD8ϩ T cell line. The requirements for survival reduced the size of the pool if present during memory maintenance and proliferation of IL-2-dependent T cell lines differ from those of by inhibiting the proliferation of CD8ϩ memory T cells. primary T cells because prolonged in vitro culture with IL-2 may We employed in this study two adoptive transfer models to in- modulate the expression of receptors and survival genes vestigate CD8ϩ T cell memory. The first model was designed to in T lymphocytes. Saparov et al. (19), in contrast, observed that examine the role of IL-2 in memory maintenance; therefore, IL- IL-2 expression during primary T cell activation correlates with 2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes were activated in vitro in the the magnitude of the secondary immune response elicited 1 wk presence of excess IL-2 and then transferred to Rag1Ϫ/Ϫ hosts to after primary activation, leading the authors to suggest that IL-2 follow their long-term survival. Rag1Ϫ/Ϫ mice do not produce promotes effector/memory T cell generation. Our findings extend IL-2 (confirmed by RT-PCR in our study) because they lack lym- this observation by demonstrating that IL-2 enhances CD8ϩ mem- phocytes. Therefore, the only potential source of IL-2 in these ory T cell generation in an established model of immunologic mice is the adoptively transferred lymphocyte population. The sec- memory. However, our results identify a paradoxical effect of ond model was designed to examine the role of IL-2 in memory IL-2, specifically that it interferes with the maintenance of long- generation; therefore, naive IL-2ϩ/ϩ or IL-2Ϫ/Ϫ 2C lymphocytes lived CD8ϩ memory T cells. were transferred to Rag1Ϫ/Ϫ hosts, activated in vivo, and their Cytokines other than IL-2 may be critical for generating and survival followed over an extended period of time. We confirmed sustaining T cell memory. Exogenous IL-15 is a potent and selec- in our experiments that long-lived, previously activated 2C lym- tive stimulator of memory-phenotype CD8ϩ T cells in vivo (29, phocytes are CD8ϩ memory T cells by demonstrating that they 30). Moreover, memory-phenotype CD8ϩ T cells are selectively have a significant survival advantage over adoptively transferred reduced in number in IL-15R␣ or IL-15 gene-knockout mice (31, naive lymphocytes, are CD44highCD62Llow, and exhibit a rapid 32). Unlike IL-2, IL-15 is ubiquitously expressed by stromal cells CTL response upon antigenic restimulation. The adoptive transfer in multiple tissues making it a likely cytokine to maintain the 3036 THE ROLE OF IL-2 IN CD8ϩ T CELL MEMORY memory T cell pool in the absence of ongoing inflammation or 17. Theze, J., P. M. Alzari, and J. Bertoglio. 1996. 2 and its receptors: immunity (31). recent advances and new immunological functions. Immunol. Today 17:481. 18. Ke, Y., H. Ma, and J. A. Kapp. 1998. is required for activation of effector activities, whereas is required for the maintenance of memory in Acknowledgments ovalbumin-specific, CD8ϩ cytotoxic T lymphocytes. J. Exp. Med. 187:49. We thank Dr. Rafi Ahmed (Emory University) for his critical comments. 19. Saparov, A., F. H. Wagner, R. Zheng, J. Oliver, H. Maeda, R. D. Hocket, and C. T. Weaver. 1999. Interleukin-2 expression by a subpopulation of primary T cells is linked to enhanced memory/effector function. Immunity 11:271. References 20. Sha, W. C., C. A. Nelson, R. D. Newberry, D. M. Kranz, J. H. Russel, and 1. Ahmed, R., and D. Gray. 1996. and protective immunity: D. Y. Loh. 1988. Selective expression of an antigen receptor on CD8-bearing T Understanding their relation. Science 272:54. lymphocytes in transgenic mice. Nature 335:271. 2. Dutton, R. W., L. M. Bradley, and S. L. Swain. 1998. T cell memory. Annu. Rev. 21. Biddison, W. E., R. Lichtenfels, M. Adibzadeh, and R. Martin. 1994. Measure- Immunol. 16:201. ment of CTL activity by the calcein release assay. In Current Protocols in Im- 3. Goldrath, A. W., and M. J. Bevan. 1999. Selecting and maintaining a diverse munology, Vol. 1. J. Coligan, A. M. Kruisbeek, D. H. Margulies, E. M. Shevach, T-cell repertoire. Nature 402:255. and W. Strober, eds. Wiley, New York, p. 7.17.4. 4. Lenardo, M., F. K.-M. Chan, F. Hornung, H. McFarland, R. Siegel, J. Wan, and 22. Boise, L. H., A. J. Minn, C. H. June, T. Lindsten, and C. B. Thompson. 1995. L. Zheng. 1999. Mature T lymphocyte apoptosis: Immune regulation in a dy- Growth factors can enhance lymphocyte survival without committing the cell to namic and unpredictable antigenic environment. Annu. Rev. Immunol. 17:221. undergo cell division. Proc. Natl. Acad. Sci. USA 92:5491. 5. Murali-Krishna, K., J. D. Altman, M. Suresh, D. J. D. Sourdive, A. J. Zajac, 23. Boise, L. H., A. J. Minn, P. J. Noel, C. H. June, M. A. Accavitti, t. Lindsten, and J. D. Miller, J. Slansky, and R. Ahmed. 1998. Counting antigen-specific CD8 T C. B. Thompson. 1995. CD28 costimulation can promote T cell survival by cells: a re-evaluation of bystander activation during viral infections. Immunity enhancing the expression of Bcl-xL. Immunity 3:87. 8:177. 24. Lenardo, M. J. 1991. Interleukin-2 programs mouse ␣/␤ T lymphocytes for ap- 6. Opferman, J. T., B. T. Ober, and P. G. Ashton-Rickardt. 1999. Linear differen- optosis. Nature 353:858. tiation of cytotoxic effectors into memory T lymphocytes. Science 283:1745. 25. Kneitz, B., T. Herman, S. Yonehara, and A. Schimpl. 1995. Normal clonal ex- 7. Jacob, J., and D. Baltimore. 1999. Modelling T-cell memory by genetic marking pansion but impaired Fas-mediated cell death and anergy in IL-2 deficient mice. Downloaded from of memory T cells in vivo. Nature 399:593. Eur. J. Immunol. 25:2572. 8. Tough, D. F., and J. Sprent. 1994. Turnover of naive- and memory- phenotype T 26. Van-Parijs, L., A. Biuckians, A. Ibragimov, F. W. Alt, D. M. Willerford, and cells. J. Exp. Med. 179:1127. A. K. Abbas. 1997. Functional responses and apoptosis of CD25 (IL-2R␣)-de- 9. Bachmann, M. F., B. Odermatt, H. Hengartner, and R. M. Zinkernagel. 1996. ficient T cells expressing a transgenic antigen receptor. J. Immunol. 158:3738. Induction of long-lived germinal centers associated with persisting antigen after 27. Dai, Z., A. Arakelov, M. Wagener, B. T. Konieczny, and F. G. Lakkis. 1999. The viral infection. J. Exp. Med. 183:2259. ␥ ␥ role of the common cytokine receptor chain ( c) in regulating IL-2- dependent, 10. Gray, D., and P. Matzinger. 1991. T cell memory is short-lived in the absence of activation-induced CD8ϩ T cell death. J. Immunol. 163:3131. antigen. J. Exp. Med. 174:969. 28. Dai, Z., B. T. Konieczny, F. K. Baddoura, and F. G. Lakkis. 1998. Impaired http://www.jimmunol.org/ 11. Oehen, S., H. Waldner, T. M. Kundig, H. Hengartner, and R. M. Zinkernagel. alloantigen-mediated T cell apoptosis and failure to induce long term allograft 1992. Antivirally protective memory to lymphocytic choriomen- survival in interleukin-2-deficient mice. J. Immunol. 161:1659. ingitis virus is governed by persisting antigen. J. Exp. Med. 176:1273. 29. Ku, C. C., M. Murakami, A. Sakamoto, J. Kappler, and P. Marrack. 2000. Control 12. Kundig, T. M., M. F. Bachmann, S. Oehen, U. W. Hoffman, J. J. Simard, ϩ C. P. Kalberer, H. Pircher, P. S. Ohashi, H. Hengartner, and R. M. Zinkernagel. of homeostasis of CD8 memory T cells by opposing cytokines. Science 288: 1996. On the role of antigen in maintaining cytotoxic T-cell memory. Proc. Natl. 675. 30. Zhang, X., S. Sun, I. Hwang, D. F. Tough, and J. Sprent. 1998. Potent and Acad. Sci. USA 93:9716. ϩ 13. Mullbacher, A. 1994. The long-term maintenance of cytotoxic T cell memory selective stimulation of memory-phenotype CD8 T cells in vivo by IL-15. Im- does not require persistence of antigen. J. Exp. Med. 179:317. munity 8:591. 14. Murali-Krishna, K., L. L. Lau, S. Sambhara, F. Lemonnier, J. Altman, and 31. Lodolce, J. P., D. L. Boone, S. Chai, R. E. Swain, T. Dassopoulos, S. Trettin, and R. Ahmed. 1999. Persistence of memory CD8 T cells in MHC class I-deficient A. Ma. 1998. IL-15 receptor maintains lymphoid homeostasis by supporting lym- mice. Science 286:1377. phocyte homing and proliferation. Immunity 9:669. by guest on September 30, 2021 15. Swain, S. L., H. Hu, and G. Huston. 1999. Class II-independent generation of 32. Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers, CD4 memory T cells from effectors. Science 286:1381. N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al. 2000. Reversible defects 16. Taniguchi, T., and Y. Minami. 1993. The IL-2/IL-2 receptor system: a current in natural killer and memory CD8 T cell lineages in interleukin-15-deficient mice. overview. Cell 73:5. J. Exp. Med. 191:771.