Four Functionally Distinct Populations of Human Effector-Memory ...

The Journal of Immunology

Four Functionally Distinct Populations of Human Effector-Memory CD8ⴙ T Lymphocytes1 Pedro Romero,* Alfred Zippelius,*† Isabel Kurth,‡§ Mikae¨l J. Pittet,*¶ Ce´dric Touvrey,* Emanuela M. Iancu,‡ Patricia Corthesy,‡ Estelle Devevre,* Daniel E. Speiser,* and Nathalie Rufer2‡ In humans, the pathways of memory and effector T cell differentiation remain poorly defined. We have dissected the functional properties of ex vivo effector-memory (EM) CD45RAⴚCCR7ⴚ T lymphocytes present within the circulating CD8ⴙ T cell pool of healthy individuals. Our studies show that EM T cells are heterogeneous and are subdivided based on differential CD27 and CD28 expression into four subsets. EM1 (CD27ⴙCD28ⴙ) and EM4 (CD27ⴚCD28ⴙ) T cells express low levels of effector mediators such as granzyme B and perforin and high levels of CD127/IL-7R␣. EM1 cells also have a relatively short replicative history and display strong ex vivo telomerase activity. Therefore, these cells are closely related to central-memory (CD45RAⴚCCR7ⴙ) cells. In contrast, EM2 (CD27ⴙCD28ⴚ) and EM3 (CD27ⴚCD28ⴚ) cells express mediators characteristic of effector cells, whereby EM3 cells display stronger ex vivo cytolytic activity and have experienced larger numbers of cell divisions, thus resembling differentiated effector (CD45RAⴙCCR7ⴚ) cells. These data indicate that progressive up-regulation of cytolytic activity and stepwise loss of CCR7, CD28, and CD27 both characterize CD8ⴙ T cell differentiation. Finally, memory CD8ⴙ T cells not only include centralmemory cells but also EM1 cells, which differ in CCR7 expression and may therefore confer memory functions in lymphoid and peripheral tissues, respectively. The Journal of Immunology, 2007, 178: 4112– 4119.

U

pon productive interaction between mature Ag-presenting dendritic cells and specific but functionally naive T lymphocytes, the latter undergo both clonal expansion and differentiation into memory and effector type T cells (reviewed in Ref. 1). Although memory T cells acquire the ability to respond with an accelerated kinetic to a second encounter with Ag, effector T cells display functions such as lytic activity against Ag-expressing target cells and production of cytokines such as IFN-␥ that are measurable in short-term assays (2). Major efforts have been made in recent years to understand T cell differentiation pathways. An important task has been to define molecular markers that readily identify and isolate T cells sharing discrete stages of differentiation. The introduction of multimers of MHC/Ag peptide, that bind stably to specific TCR on the surface of T cells, has made it possible to carry out these types of analyses at the Ag-specific T cell level (3– 6). Nonetheless, this endeavor remains challenging due to the relatively low numbers of single Ag-specific T cells that can be

*Division of Clinical Onco-Immunology, Ludwig Institute for Cancer Research, Lausanne Branch, University Hospital of Lausanne, Lausanne, Switzerland; †Medical Oncology, Department of Internal Medicine, University Hospital Zurich, Zurich, Switzerland; ‡Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland; §Columbia University Medical Center, Irving Cancer Research Center, New York, NY 10032; and ¶Center for Molecular Imaging Research, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA 02129 Received for publication September 29, 2006. Accepted for publication January 12, 2007. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was sponsored and supported by the Swiss National Center of Competence in Research Molecular Oncology and Swiss National Science Foundation Grants 3100-068016 and 3100A0-105929. A.Z. was supported in part by the EmmyNoether Program of the Deutsche Forschungsgemeinshaft (Zi-685/2.3) and a grant from the Swiss National Foundation (3200B0-103608/1). 2 Address correspondence and reprint requests to Dr. Nathalie Rufer, Swiss Institute for Experimental Cancer Research, 155 ch. des Boveresses, CH-1066 Epalinges, Switzerland. E-mail address: [email protected]

www.jimmunol.org

retrieved from immune individuals and to the apparent complexity of the T cell differentiation process (7–13). Four major subsets of human CD8⫹ T lymphocytes have been delineated with the help of two cell surface markers, the high m.w. isoform of the common lymphocyte Ag CD45RA and the chemokine receptor CCR7 (14, 15). Although the relationship between the phosphatase activity of the former and T cell differentiation remains ill-defined, the CCR7 is involved in the molecular cascade leading to lymphocyte recirculation from peripheral blood to secondary lymphoid tissues (reviewed in Ref. 16). Thus, CCR7⫹ naive and central-memory (CM)3 T cells are characterized by the ability to repeatedly circulate into lymph nodes and eventually encounter Ag presented by incoming CCR7⫹ mature dendritic cells. In contrast, effector-memory (EM) and effector T lymphocytes down-regulate the CCR7 and appear specialized in migrating to peripheral nonlymphoid tissues. Although this two-marker procedure to identify functionally distinct CD8⫹ T cell subsets has proven popular, increasing evidence indicates the existence of highly heterogeneous functional CD8⫹ T subpopulations (7–13). For instance, five-color analysis including two additional surface Ags, CD27 and CD28, has proven useful in defining two additional subsets of pre-effector CD8⫹ T lymphocytes (17). In the present study, we uncovered additional heterogeneity among the EM subset by studying the functional attributes of pools of EM cells separated on the basis of the various combinations of costimulatory receptor (i.e., CD27 and CD28) cell surface expression. We propose the identification of four functionally distinct subsets of EM T cells, EM1, EM2, EM3, and EM4, and show data supporting the notion that these populations represent T cells with progressive differentiation toward lymphocytes with potent cytokine and lytic 3 Abbreviations used in this paper: CM, central-memory; EM, effector-memory; sj, signal joint; TRAP, telomerase repeat amplification protocol; FISH, fluorescence in situ hybridization; HD, healthy donor; TREC, TCR excision circle.

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00

The Journal of Immunology effector functions. Moreover, the EM1 subset is nearly identical with the CM one except for the lack of CCR7 cell surface expression, suggesting that this subset shares functional features with memory cells.

Materials and Methods Cell preparation and flow cytometry Peripheral blood samples were collected from 15 healthy donors, aged 22– 46 years, with a normal proportion of CD8⫹ T lymphocytes (average, 26%; range, 18 –34%). PBMC were obtained by density centrifugation using Ficoll-Hypaque (Pharmacia). Our experimental procedures involve two steps that exclude NK cell contamination. First, CD8⫹ T lymphocytes were positively enriched from cryopreserved or fresh PBMCs using antiCD8-coated magnetic microbeads (Miltenyi Biotec), a procedure that eliminates most NK cells because they are not efficiently retained by the magnet. Cells were stained with appropriate mAbs in PBS, 0.2% BSA, 50 ␮M EDTA for 20 min at 4°C and either directly analyzed or sorted into defined populations on a FACSVantage SE, using CellQuest software (BD Biosciences). Immediate reanalysis of the isolated populations revealed on average ⬎95% purity, and in the case of naive T cells, over 99% purity. Of note, as naive T cells represent a homogeneous subset that is distinctively CD45RA-, CCR7-, CD27-, and CD28-positive, contaminant naive cells were not present within sorted CM and EM1 cells in significantly high numbers (⬍1%), thus allowing us to exclude bias in the TCR excision circle (TREC) analysis of those cells and in their estimate of proliferative history (data not shown). Second, FACS analysis and sorting was performed on gated CD8 bright T cells, allowing the exclusion of any residual contaminating NK cells in the sorted populations (⬍2%). Intracellular content of granzyme B and perforin was measured in freshly isolated CD8⫹ T lymphocytes without previous stimulation as previously described (17). The following mAbs were purchased from BD Biosciences or BD Pharmingen: anti-CD27-FITC and -PE, anti-CD28-PE and -allophycocyanin, antiCD8-allophycocyanin/Cy7, anti-HLA-DR-FITC, and goat anti-rat IgG-allophycocyanin. Other sources of mAbs were: Beckman Coulter (anti-CD45RA-PE-Texas Red) and Caltag Laboratories (goat anti-rat IgG-PE). Anti-CCR7 rat IgG mAb 3D12 was provided by Dr. M. Lipp (Max Delbru¨ck Institute, Berlin, Germany). Anti-granzyme B-FITC and anti-perforin-FITC mAbs were obtained from Ho¨lzel Diagnostika and Alexis, respectively. Synthesis of PE-labeled HLA-A*0201/peptide multimers with Melan analog peptide26 –35 (ELAGIGILTV), and allophycocyanin-labeled HLA-A*0201/peptide multimers with Flu matrix protein58 – 66 (GILGFVTL), CMV pp65495–503 (NLVPMVATV), and EBV BMFL1280 –288 (GLCTLVAML), were prepared as described previously (5).

cDNA amplification and five-cell RT-PCR To avoid contamination of small populations by more abundant subsets, 10 ⫻ 103 T cells of each subset were sorted by flow cytometry and five-cell aliquots of the purified subsets were then resorted directly into wells of 96-V-bottom plates. The procedures for cDNA preparation, cDNA amplification as well as the RT-PCR were recently described in details (17, 18). Additional primers were used in the present study: CD27, 5⬘-ACGTGA CAGAGTGCCTTTTCG-3⬘; reverse-5⬘-TTTGCCCGTCTTGTAGCATG3⬘; CD127/IL-7R␣: 5⬘-ATCTTGGCCTGTGTGTTATGG-3⬘; reverse-5⬘ATTCTTCTAGTTGCTGAGGAAACG-3⬘.

Cytolytic activity Cytolytic activity was tested in a CD3 mAb-mediated “redirected” 51Cr release assay. In brief, FcR-expressing P815 target cells were radiolabeled with Na51CrO4 (PerkinElmer) for 1 h at 37°C. Sorted CD8⫹ T subsets were incubated with P815 target cells (103 cells/well) at varying effector-target cell ratios in the presence or absence of 300 ng/ml anti-CD3 mAb (OKT3). After 4 h at 37°C, supernatants were collected and counted on a gamma counter. Percent lysis was calculated as (experimental release ⫺ spontaneous release) ⫻ 100/(total release ⫺ spontaneous release).

Quantification of TRECs by real-time PCR The amount of signal joint (sj) TRECs in 5–15 ⫻ 104 sorted CD8⫹ T subsets was determined by real-time quantitative PCR using the ABI PRISM 7700 Sequence Detector TaqMan system (Applied Biosystems) as previously described (19, 20). In brief, after cell lysis in 100 mg/L proteinase K (Roche Diagnostics) for 2 h at 56°C followed by 15 min at 95°C, a PCR was performed in a final volume of 25 ␮l containing 5 ␮l of cell extract, 12.5 ␮l of TaqMan Universal Master Mix including AmpliTaq Gold (Applied Biosystems), 500 nM of each primer (sj-5⬘ forward: CA

4113 CATCCCTTTCAACCATGCT; sj-3⬘ reverse: GCCAGCTGCAGGGTT TAGG), and 125 nM TaqMan probe (FAM-ACACCTCTGGTTTTTGTA AAGGTGCCCACT-TAMRA). After one cycle of 2 min at 50°C followed by an initial 10-min denaturation at 95°C, 40 cycles of 30 s at 95°C and 1 min at 65°C were performed. The number of TRECs in a given sample was estimated by comparing the cycle threshold value obtained with a standard curve obtained from PCR performed with 10-fold serial dilutions of an internal standard provided by Dr. D. Douek (Vaccine Research Center, National Institutes of Health, Bethesda, MD). The dilutions contained between 107 and 101 copies of sjTREC and four reactions were run with each dilution. Considering that ⱖ50,000 cells were always analyzed per subset, and that the linear range of the external standard used starts at 10 copies, our lower TREC detection limit was 10 copies/50,000 cells or 0.02% (thus values of ⬍10 copies/sample were quoted as below the detection limit of the assay). In all PCR assays, the correlation coefficient of the standard curve was ⱖ0.997, whereas the slope varied between ⫺3.52 and ⫺3.67. The TREC analysis was performed on young healthy individuals (⬍35 years of age) because aging has been shown to inversely correlate with the TREC levels (19, 20).

Telomerase repeat amplification protocol assay Telomerase activity was measured with the telomerase repeat amplification protocol (TRAP) assay using a telomerase substrate primer as described previously (17). Cell extracts were obtained from 5 to 15 ⫻ 104 sorted CD8⫹ T cell subsets. As positive control we used extracts from CD8⫹ T lymphocytes stimulated for 5 days with 1 ␮g/ml PHA (Sodiag) and 150 U/ml rIL-2 in presence of 1 ⫻ 106/ml irradiated feeder cells. Extension of the telomerase substrate primer by telomerase was performed for 30 min at 30°C in the presence of [␣-32P]dGTP and the products generated were amplified by 27 cycles of PCR at 94°C for 30 s and 60°C for 30 s using the ACX-anchored return primer. One-half of the amplified products were resolved on a 15% polyacrylamide gel and visualized by a phosphoimaging system.

Telomere fluorescence in situ hybridization and flow cytometry The average length of telomere repeats at chromosome ends in individual cells was measured by fluorescence in situ hybridization (FISH) and flow cytometry as previously reported (18, 21, 22). Telomere fluorescence was calculated by subtracting the mean fluorescence of the background control (no probe) from the mean fluorescence obtained from cells hybridized with the telomere probe after calibration with FITC-labeled fluorescent beads (Quantum TM-24 Premixed; Bangs Laboratories) and conversion into molecules of equivalent soluble fluorochrome units (MESF). The following equation was performed to estimate the telomere length in base pair: bp ⫽ MESF ⫻ 0.495 (21, 22). Telomere length measurement was performed on in vitro-derived T cell clones by limiting dilution (23) sorted from MelanA-, Flu-, EBV-, and CMV-specific CD8⫹ T lymphocytes isolated from a single healthy individual, allowing for the recovery of a sufficient number of cells for flow FISH analysis. All T cell clones were expanded in identical in vitro culture conditions and have undergone approximately the same mean number of population doublings. Cell extracts obtained from several T cell clones with distinct Ag specificity were submitted to the TRAP assay. Although very low levels of telomerase activity could be detected, no significant differences among the stimulated cells were apparent (data not shown). After a single round of mitogenic stimulation, 2 ⫻ 105 cells were further processed by flow FISH. Because the average telomere fluorescence from all these clones was evaluated in the same experimental design, this allowed direct telomere comparison between each tested clonal condition.

Results

Ex vivo distribution of phenotypically distinct CD8⫹ T cell subsets from human peripheral blood Human CD8⫹ T lymphocytes can be separated into four functionally different populations on the basis of CD45RA and CCR7 expression: naive (RA⫹CCR7⫹), CM (RA⫺CCR7⫹), EM (RA⫺CCR7⫺), and effector (EMRA; RA⫹CCR7⫺) (Fig. 1A). The relationship between CD45RA/CCR7 and the expression of CD27 and CD28 was further assessed within each cell subset. Staining of peripheral blood CD8⫹ T lymphocytes with Abs to CD45RA, CCR7, CD27, and CD28 revealed the presence of nine discrete subpopulations in the blood from a representative healthy donor (Fig. 1B). Naive and CM cells uniformly coexpressed CD27 and CD28. In contrast, as based on

4114

EX VIVO HUMAN EM T CELL SUBSETS

FIGURE 1. Differential expression of CD45RA, CCR7, CD27, and CD28 cell surface molecules on total CD8⫹ T cells from healthy blood donors. A, CD8⫹ gated cells were separated into four subsets (naive, CM, EM, and effector) based on CD45RA and CCR7 labeling. B, Each of these subsets was analyzed for CD27 and CD28 coexpression and nine subpopulations of CD8⫹ T cells could be distinguished. A representative example is here depicted. C, Naive-CD27⫹, CMCD27⫹, EM-CD27⫹ (comprising EM1⫹EM2), EMCD27⫺ (EM3⫹EM4), and EMRA-CD27⫺ T cells were analyzed for their level of CD28 expression. Naive-CD28⫹, CM-CD28⫹, EM-CD28⫹ (comprising EM1⫹EM4), EM-CD28⫺ (EM2⫹EM3), and EMRACD28⫺ T cells were analyzed for their level of CD27 expression. D, The distribution of the nine defined CD8⫹ T cell subsets among 14 healthy individuals is shown as mean percentage (range).

our previous report (17), the RA⫹CCR7⫺ EMRA T cells were split into three functionally distinct subsets: pE1 (27⫹28⫹), pE2 (27⫹28⫺), and effector (27⫺28⫺) T cell subpopulation. EM T cells also exhibited high heterogeneity with a differential expression of CD27 and CD28 cell surface molecules (Fig. 1B). Thus, we identified, within the RA⫺CCR7⫺ EM T compartment, four phenotypically separate subsets referred to as EM1 (27⫹28⫹), EM2 (27⫹28⫺), EM3 (27⫺28⫺), and EM4 (27⫺28⫹). Naive T cells exhibited the highest levels of CD27, with a progressive down-regulation of CD27 cell surface expression from naive, through CM and EM1, to EM2 cells (Fig. 1C). In contrast, cell surface CD28 expression levels were higher in CM, EM1, and EM4 cells than in naive cells. Finally, the EM3 subset resembles to the effector one, because both subsets have lost CD27 and CD28 coexpression. These data are in line with reports showing that in vitro stimulation of CD8⫹ T cells induces down-regulation of CCR7 and CD27 and up-regulation of CD28 (14, 24, 25). To obtain an overview of the distribution of the nine CD8⫹ T cell subsets defined by a different pattern of CD45RA, CCR7, CD27, and CD28 expression, we analyzed PBMC from 14 healthy individuals, ranging in age from 22 to 46 years (Fig. 1D). Although the proportion of each T cell subset varied between donors, all nine subpopulations were found in every single individual, with the preferential dominance of naive, effector, and EM1 CD8⫹ T subpopulations. Progressive acquisition of granzyme B and perforin expression and of ex vivo killing activity within EM CD8⫹ T cell subsets Previous studies have proposed that CD27⫹CD28⫹ T cells differentiate through a CD27⫹CD28⫺ to a CD27⫺CD28⫺ stage (11, 17). According to this model, CD8⫹ T cells sequentially downregulate CCR7, CD28, and CD27 surface expression, while up-

regulating expression of molecules that confer cytolytic activity. To investigate whether EM1, EM2, and EM3 CD8⫹ T cell subsets corresponding to the proposed differentiation pathway (11, 17) differed in the expression of genes involved in T cell effector functions, we used a modified RT-PCR protocol that detects specific cDNAs after global amplification of expressed mRNAs from as few as five cells (17, 18). As expected, all naive and most CM T cell five-cell samples contained no detectable granzyme B, perforin, IFN-␥, or NK receptor CD94 mRNA (Fig. 2A). Interestingly, despite the loss of CCR7 expression, the gene expression profile of EM1 T cells resembled closely to the one of CM cells. Of note, we observed differences of IFN-␥ mRNA expression by EM1 T cells that may reflect biological variations in the proportions of IFN-␥expressing cells from the same subset among different healthy donors (Fig. 2A; see HD1, HD2, and HD3). In contrast, EM2 and EM3 T cell aliquots exhibited detectable levels of granzyme B, IFN-␥, and CD94 transcripts. This was particularly marked for granzyme B mRNA expression and associated with the high expression level for this protein (Fig. 2B). Moreover, most of naive, CM, and EM1 cells, but almost none of the aliquots of EM3 and effector T cells, yielded a detectable CD127-specific product, encoding for the ␣-chain of the IL-7R complex. According to both mRNA analysis (Fig. 2A) and intracellular staining (Fig. 2B), all three EM T subsets expressed perforin, but at lower levels than in the differentiated effector subset. When we compared the cytolytic activity of these distinct cell populations, using a CD3 mAb-mediated redirected 51Cr release assay (Fig. 2C), we found that both EM3 and effector T cells displayed high ex vivo lytic activity, whereas CM and EM1 T cells had comparable killing activity, which was ⬃10 times lower than that of the effector population.

The Journal of Immunology

FIGURE 2. Ex vivo analysis of expression of effector mediators within EM T cell subsets. A, Gene expression analysis was performed on sorted naive (RA⫹CCR7⫹27⫹28⫹), CM (RA⫺CCR7⫹27⫹28⫹), EM1 (RA⫺CCR7⫺27⫹28⫹), EM2 (RA⫺CCR7⫺27⫹28⫺), EM3 (RA⫺CCR7⫺27⫺28⫺) and effector (RA⫹CCR7⫺27⫺28⫺) CD8⫹ T cells using a modified RT-PCR protocol (18). Data from three or six independent five-cell aliquots per subset, and negative (⫺) and positive (⫹) controls, are depicted. Comparable results were obtained in three healthy individuals. For IFN-␥ expression, data obtained from HD1, HD2, and HD3 is depicted; na, not applicable. B, The proportion of granzyme B- and perforin-positive cells among CM, EM1, EM2, EM3, and effector T cells was determined by immunofluorescence. Note that the perforin signal is lower in EM1, EM2, and EM3 than in effector cells. Data are representative of four healthy donors. C, Ex vivo-sorted CD8⫹ naive, CM, EM1, EM2, EM3, and effector T cells were tested in a redirected cytolytic assay against 51 Cr-labeled P815 target cells. None of these subsets lysed P815 cells in absence of CD3 mAbs (lysis ⱕ10%; data not shown). Data are representative of two healthy donors.

Progressive telomere shortening and reduction in level of TRECs within EM CD8⫹ T cell subsets We next investigated the replicative history of EM1, EM2, and EM3 T cell subsets by quantifying their content of TRECs, which are stable DNA episomes formed during TCR-␣ gene rearrangement and are diluted out with each cell division (19). In all four healthy individuals tested (Fig. 3A), naive cells had the highest level of TRECs, whereas they were below the detection limit of the assay (ⱕ0.01 TREC copies/100 cells) in the EM3 and effector subsets. CM and EM2 T cells contained low but detectable TRECs in all healthy individuals. Intriguingly, TRECs in EM1 cells were detected in reduced levels in two healthy donors (HD), and not at all in the two other individuals, indicating that these cells had at least undergone six to seven more divisions than the bulk of naive cells.

4115

FIGURE 3. Ex vivo analysis of the replicative history and telomerase activity of EM T cell subsets. A, Real-time PCR quantification of TRECs was performed on sorted naive, CM, EM1, EM2, EM3, and effector CD8⫹ T cells from four healthy young individuals (age range, 22–35 years). ⴱ, Not detectable (sorted cell number was 5 ⫻ 104 to 105, lower quantification limit ⫽ 0.01– 0.02%). Of note, the levels of TRECs measured in sorted CM (0.4 ⫾ 0.2), EM1 (0.1 ⫾ 0.1), and EM2 (1 ⫾ 0.4) cells were not significantly different. B, Telomere fluorescence analysis in ex vivo-sorted naive, CM, EM1, EM2, EM3, and effector CD8⫹ T cells isolated from two healthy donors (HD1 and HD2). The mean telomere fluorescence (in FL1 channel) was converted to kilobase as described in Materials and Methods; n.a., not applicable. C, Telomerase activity in cell extracts of sorted 28⫹DR⫹, 28⫹DR⫺, naive, CM (28⫹DR⫹), CM (28⫹DR⫺), EM1 (28⫹DR⫹), EM1 (28⫹DR⫺), EM2, EM3, and effector CD8⫹ T cells. As positive controls, we used cell extracts of in vitro-PHA-activated CD8⫹ T cells (act. CD8⫹). Data are representative of two healthy individuals. CD45RA; RA, CCR7; R7.

To further characterize the relationship between the mitotic history of cells and their differentiation status, the average length of telomere repeats of ex vivo-sorted naive, CM, EM1, EM2, EM3, and effector T lymphocytes from two healthy donors, was measured by FISH and flow cytometry (Fig. 3B). Because telomeres progressively shorten as a function of cell division (26), telomere length is a powerful indicator of the replicative in vivo history of lymphocytes (21, 27). We observed a progressive reduction in the mean telomere fluorescence from naive through CM, EM1, EM2 to EM3 and effector T lymphocytes, that corresponded to a telomere shortening of ⬃4.5–5 kb. EM1 cells displayed shorter average telomere lengths than those observed in CM cells from HD1, suggesting additional cell divisions within the former subset. In contrast, in HD2, the telomeres of both subsets exhibited similar lengths. These results confirmed the heterogeneity observed when we characterized the content of TRECs within EM1 cells (Fig. 3A). Importantly, our data revealed that EM3 cells had relatively short telomeres, within the same range than those found in

4116

EX VIVO HUMAN EM T CELL SUBSETS

FIGURE 4. Comparison of the ex vivo expression of effector mediators between EM1 and EM4 T cell subsets. A, Gene expression analysis was performed on sorted CM, EM3, EM1, and EM4 CD8⫹ T cells by RTPCR. Data from three or eight independent five-cell aliquots are shown. (⫺), Negative; (⫹), positive controls. B, The proportion of granzyme B- and perforin-positive cells among CM, EM1, EM4, and effector CD8⫹ T cells was determined by immunofluorescence. All results are representative of two healthy individuals. CD45RA; RA, CCR7; R7.

differentiated effector cells. This agrees with our recently published results (12) showing progressive shortening of the telomeres along T cell differentiation, so that highly differentiated CD8⫹CCR7⫺CD27⫺CD57⫹ cells displayed the shortest telomeres, with lengths equivalent to those observed in primed CD8⫹ T cells from the elderly (22). Altogether, these data support the view that T cells evolve through extensive rounds of division as they differentiate further. A previous report (28) showed telomerase activity in ex vivoisolated CD8⫹ T cells that express both CD28 and the activation marker HLA-DR. To determine which if not all of the HLA-DRexpressing T cell subsets accounted for such activity, TRAP assays with the HLA-DR⫹ and HLA-DR⫺ fractions of CM and EM1 T cells were conducted (Fig. 3C). Remarkably, we observed positive telomerase activity selectively confined to the DR-positive fractions of both subsets. In contrast, naive, EM2, and EM3 T cells, that did not contain a sizeable fraction of HLA-DR⫹ cells, revealed no ex vivo detectable telomerase activity. Of note, telomerase activity was already detectable when EM1 T cells had been sorted without discriminating HLA-DR⫹ from HLA-DR⫺ cells (data not shown). Finally, within both CM and EM1 populations, the proportion of HLA-DR⫹ T cells represented between 5 and 15% with a trend toward CM cells. Both primed EM1 (27⫹28⫹) and EM4 (27⫺28⫹) T cell subsets express a similar pattern of genes and display low levels of effector mediators The fourth T cell population (EM4) was analyzed in additional experiments, as EM4 cells could not be fully investigated in parallel in the previously shown experiments due to technical limitations. To gain insight into the relationship between EM4 T cells and the EM1 subset (as outlined in Fig. 1B), we compared their expression of genes involved in effector functions within five cell-sorted samples (Fig. 4A). In both EM1 and EM4 cells, no granzyme B mRNA transcripts were detected, whereas CD127/ IL7R␣, perforin, and IFN-␥ transcripts were found in a significant proportion of these samples. Moreover, these results correlated with the analysis of granzyme B and perforin expression by intracellular staining for these molecules (Fig. 4B). Thus, our data indicate that despite the loss of CD27 expression within the EM4 compartment, these cells seem closely related to the EM1 T cells. Ag-specific CD8⫹ T lymphocytes exhibit distinct differentiation phenotypes and replicative history Several studies have reported that Ag-specific CD8⫹ T cells directed against the tumor-associated self-Ag Melan-A/MART-1 or against viral epitopes such as the influenza matrix protein (Flu),

BMFL1 (EBV), and pp65 (CMV) show different stages of cellular differentiation. We previously found that the majority of circulating Melan-A-specific T cells from HLA-A2 healthy donors is phenotypically and functionally naive despite their high frequencies (20). In contrast, influenza- and EBV-specific T cells respectively display a primed CM and EM phenotype, while CMV-specific T lymphocytes are mostly composed of differentiated effector cells (2, 8 –11, 17, 29). Here, we confirmed and extended these studies by comparing the coexpression of CCR7/CD45RA (Fig. 5A), and of CD27/CD28 (Fig. 5B) in Melan-A-specific T cells to that of primed Ag-specific T cells such as EBV and CMV in healthy individuals. Whereas Melan-A-specific T cells shared a homogeneous naive-like phenotype (CCR7⫹CD45RA⫹CD27⫹CD28⫹), the EBV-specific response consisted primarily of early differentiated T cells (EM1, mean ⫾ SD, 60 ⫾ 16%; n ⫽ 8). Consistent with our recent report (17), CMV-specific T lymphocytes displayed the phenotype of effector CD8⫹ T cells (mean ⫾ SD, 18 ⫾ 7%; n ⫽ 8), but also the phenotype of EM2 (19 ⫾ 8%) and EM3 cells (34 ⫾ 14%). Interestingly, our data further revealed that Melan-A-specific naive T cells, as compared with EBV-specific T cells, exhibited the highest levels of CD27, while they expressed lower levels of CD28 (Fig. 5C), in line with the notion that activation and priming of T cells involves down-regulation of CD27 and up-regulation of CD28 cell surface molecules (14, 24, 25). An important aspect, often neglected in those current studies, concerns the proliferative potential of the characterized Ag-specific T cells. Therefore, we investigated the replicative history of each of the four above-defined Ag-specific T cells that exhibit distinct differentiation phenotypes (Fig. 5, A and B). For this purpose, we measured the average length of telomeres in Melan-A/ MART-1, Flu-, EBV-, and CMV-specific T cell clones isolated from a single healthy individual (Fig. 5D). Strikingly, we observed a progressive reduction in mean telomere fluorescence from Melan-A- through Flu-, to EBV- and CMV-specific lymphocytes, consistent with the observation of progressive telomere shortening found within differentiated CD8⫹ T cell subsets (Fig. 3B). Due to the relative low frequencies of the four antigenic specificities within CD8⫹ T cells, we were unable to perform the flow FISH experiment on ex vivo-sorted T cells (see Materials and Methods). Nevertheless, these results confirmed our previous finding that Fluspecific T cells, displaying an early differentiated phenotype (CM and EM1; Fig. 5, A and B), showed significant reduced telomere lengths ex vivo compared with the naive Melan-A-specific cells (20). Collectively, our data favor the notion that Ag-specific T cells, as they differentiate, also undergo additional rounds of in vivo cell division.

The Journal of Immunology

4117

FIGURE 5. CD8⫹ T cells vary in phenotype and replicative history depending on Ag specificity. A and B, Melan-A, Flu-, EBV-, and CMV-multimer⫹ T cells were characterized ex vivo by flow cytometry for their cell surface expression of CD45RA, CCR7, CD28, and CD27. The dot plots show double staining for CD45RA/CCR7 (A) and CD28/CD27 (B) on Melan-A (HLA-A2), Flu (HLA-A2/Matrix protein), EBV (HLA-A2/BMFL1), and CMV (HLA-A2/pp65)-specific CD8⫹ T lymphocytes gated using the relevant multimers. For comparison, stainings on whole CD8⫹ T cells (CD8⫹) and the distribution of the different subsets according to cell surface marker expressions is depicted. Eight healthy individuals were included in these analyses. Representative data are shown. C, Enlarged dot plots of CD28 and CD27 costaining on Melan-A (multimer-allophycocyanin) and EBV (multimer-PE) specific CD8⫹ T cells. Dotted lines represent the reference for CD28 and CD27 positivity based on the fluorescent signal obtained after gating on the equivalent whole CD8⫹ T cell subset. D, Telomere fluorescence analysis was performed on 10 in vitro-generated T cell clones derived from either MelanA-, Flu-, EBV-, or CMV-specific CD8⫹ T lymphocytes isolated from a single healthy donor as described in Materials and Methods. The proportion of multimer-specific T cells in CD8⫹ T cells is indicated. The mean telomere fluorescence (in FL1 channel) was converted to kilobase as described in Materials and Methods.

Discussion Circulating naive T lymphocytes form a relatively homogeneous population expressing a well-defined set of cell surface glycoproteins and are characterized by the null expression of effector mediators (e.g., IFN-␥, granzyme B, perforin, Fas/CD95) and by high proliferative potential (e.g., long telomeres, high detectable levels of TREC copies). During the last three decades, primed Ag-experienced T lymphocytes have mostly been classified into two distinct subpopulations, e.g., effector and memory cells (30). Effectors are presumably rather short-lived, produce cytolytic effector molecules and are capable of migrating to the site of infection and of killing target cells directly ex vivo. In contrast, memory cells are long-lived, persist after pathogen clearance, and have increased survival properties and cell division capacities. However, several lines of evidence recently challenge this simplified view of defining primed T cells (reviewed in Ref. 31). First, using five-color flow cytometry, detailed analysis of human peripheral CD8⫹ (17) and CD4⫹ (32) T cells allowed their distribution into six major subpopulations, identified by the patterns of expression of CD45RA, CCR7, CD28, and/or CD27. Second, it has been shown that EBV-, CMV-, and HIV-specific T cells vary in differentiation phenotype during persistent viral infections, suggesting that Agspecific T cells present in each individual type of infection are very different (11). Third, several studies in mice revealed that effector-

type T cells were required in peripheral tissue before viral challenge to protect against vaccinia virus, whereas CM cells were most potent at protecting against systemic infection with lymphocytic choriomeningitis virus (33, 34), as they have a greater capacity than EM T cells to persist in vivo (35). Finally, Roberts et al. (36) recently described that both CM and EM T cells contributed to recall responses to Sendai virus infection in the lung, with a progressive increase in efficacy of the CM subset over time. These data reinforce the notion that the protective capacity of different subpopulations of primed T cells (i.e., effector vs memory) may vary depending on the nature of the challenging pathogen, and may change substantially over time. Altogether, these and other studies (7–13) indicate that primed Ag-experienced T lymphocytes are highly heterogeneous, varying in terms of their cell surface phenotype, functional capacities, and history of Ag encounter. Here, by combining the simultaneous analysis of surface markers by multiparameter flow cytometry with the analysis of gene expression and of replicative history, we show that EM CD8⫹CD45RA⫺CCR7⫺ T cells can be further divided into four distinct subsets, based on differential CD27 and CD28 expression patterns. CD27 and CD28 are costimulatory receptors involved respectively in the generation of Ag-primed cells and the regulation of T cell activation (37, 38). EM1 (27⫹28⫹), EM2 (27⫹28⫺), EM3 (27⫺28⫺), and EM4 (27⫺28⫹) T cell subsets are all present

4118 in the peripheral blood of healthy donors with a predominance toward EM1 cells (Fig. 1D). Functionally, at least three subsets can be clearly identified: EM1 (that includes the very similar EM4), EM2, and EM3. EM1 are memory-like, EM2 are intermediate with partial effector functions and replicative history, and EM3 are effector-like. Taken together, our data are in agreement with the model according to which there is a differentiation pathway with progressive loss of CCR7, CD28, and CD27 cell surface expression concomitant with up-regulation of cytolytic capacity (11). In particular, we show that both circulating EM2 and EM3 CD8⫹ T cells express mediators characteristic of effector cells, but gene and protein expression profiles of EM3 T cells more closely resemble that of effector cells. In line with this notion, EM3 cells display stronger ex vivo cytolytic activity and have experienced a larger number of cell division, similarly to differentiated effector T lymphocytes (Figs. 2 and 3). In contrast, EM2 cells contain low but yet detectable levels of TRECs (Fig. 3A). Unless EM2 cells differentiate from CM cells without cell division, our data indicate that these cells descend from naive T cells. Moreover, loss of CD28 cell surface expression is associated with the acquisition of granzyme B expression, allowing more differentiated cells (e.g., EM3) to kill their targets through perforin/granzymes pathways. Another major finding in this study is that EM1 T cells, despite their lack of CCR7 expression, have several functional features in common with CM cells. Both populations have a similar replicative history (Fig. 3), thus they have undergone more cell divisions than naive cells but fewer than effector cells. Remarkably, they express the enzyme telomerase, which is known to be involved in maintenance of telomere length and cell proliferation potential. Because telomerase activity was exclusively found within the HLA-DR-positive fraction of CM and EM1 T cells, our data demonstrate that these subsets account for the previously described telomerase in CD8⫹28⫹DR⫹ T cells (28). Speiser et al. (28) also showed that in vivo cycling CD8⫹ T lymphocytes expressed HLADR, thus supporting the idea that telomerase expression in HLADR⫹ CM and EM1 cells reflects proliferative activity while maintaining telomere lengths. This would be compatible with reduced levels of TRECs but only progressive shortening of telomere lengths as observed in the EM1 cells (Fig. 3). It is also noteworthy to mention that induction of telomerase activity in Ag-specific effector and memory CD8⫹ T cells from mice infected with lymphocytic choriomeningitis virus has been suggested to be important for the maintenance and longevity of the memory CD8⫹ T cell population (39). Finally, both CM and EM1 T cell subsets express high levels of the IL-7R␣ chain (CD127) necessary to memory cell survival (40), but only low levels of effector molecules such as IFN-␥, granzyme B, and perforin. Altogether, these results, obtained with cells analyzed directly ex vivo, indicate that the CM and EM1 subpopulations may comprise T cells that have been recently activated following antigenic challenge. Alternatively, telomerase-expressing T cells may consist of memory cells that expand following homeostatic maintenance of T cell numbers. Investigations of ongoing immune responses in vivo (41) as well as on T lymphocytes isolated from lymph nodes will be useful to specifically address these questions. Based on these findings, it is tempting to propose that human memory CD8⫹ T cells include in fact two types of functionally equivalent populations with identical cell surface phenotypes except for the expression of the chemokine receptor CCR7. CCR7⫹ cells (CD8⫹ TCM, for CM) have the ability to migrate from blood through secondary lymphoid organs, like naive T cells, whereas CCR7⫺ cells (CD8⫹ TPM, identified here as EM1, for peripheral memory) travel from blood to nonlymphoid tissues where they can directly re-encounter Ag. In both cases, the memory status of both

EX VIVO HUMAN EM T CELL SUBSETS subsets allows for rapid reactivation upon Ag challenge regardless of tissue location. Loss of CD27 during clonal expansion and differentiation presumably leads to the emergence of differentiated T cells with a more extensive replicative history and more complete effector functions than CD8⫹CD27⫹ cells (Refs. 17 and 42; Figs. 2 and 3). Recent analysis on sorted CD27⫺ HIV- and EBV-specific T cell clones followed by in vitro stimulation revealed that most of these cells had irreversibly lost CD27 expression (43). Yet, sorted CD27⫹ CD8⫹ T cells transiently down-regulated the expression of CD27 with the majority re-expressing CD27 at the end of the proliferative cycle. In the present study, we identified a small but significant proportion of EM T cells within the circulating blood that had down-regulated CD27 expression while maintaining the CD28 costimulatory molecule (EM4; CD27⫺CD28⫹). Similarly to EM1 cells, this subset displays low levels of effector-mediated molecules, while still expressing IL-7R␣ (Fig. 4). One possibility is that EM4 cells differentiate from the CM T cell pool and that such cells represent a transitory subset appearing during the expansion phase of a secondary immune response. Alternatively, one cannot formally exclude that these cells have emerged directly from the EM1 pool and have transiently or definitively down-regulated CD27 expression. Future analysis involving the careful evaluation of their replicative history combined to their cell cycle status is necessary to understand the role of this subset along the CD8⫹ T cell differentiation pathways. Ultimately, a conclusive answer to the issues concerning the function of the various EM T subsets described here and their relationship with other subpopulations will require the in vivo tracking of Ag-specific T cells in humans during the course of infection with viruses such as influenza, EBV, or CMV. Although timely access to such clinical situations is enormously difficult, they remain attractive because they would provide two major advantages from an experimental viewpoint. On one hand, infections by these viruses are frequent and independent of geographical location. On the other hand, there are currently well-defined MHC class I-restricted dominant T cell epitopes and enough tools exist, including fluorescent pMHC multimers, to identify and characterize the corresponding Ag-specific CD8⫹ T cell responses.

Acknowledgments We dedicate this work to our friend and colleague Dr. Pascal Batard who contributed indispensably to this study and passed away abruptly. We are thankful for Drs. Immanuel Luescher and Philippe Guillaume for synthesis of multimers; Dr. Daniel Douek for providing signal joint internal standard; and Dr. Martin Lipp for the anti-CCR7 mAb. We also thank the excellent technical and secretarial help of Dr. Lionel Arlettaz, Pierre Zaech, Martine van Overloop, and Se´verine Reynard.

Disclosures The authors have no financial conflict of interest.

References 1. Ahmed, R., and D. Gray. 1996. Immunological memory and protective immunity: understanding their relation. Science 272: 54 – 60. 2. Pittet, M. J., A. Zippelius, D. E. Speiser, M. Assenmacher, P. Guillaume, D. Valmori, D. Lienard, F. Lejeune, J. C. Cerottini, and P. Romero. 2001. Ex vivo IFN-␥ secretion by circulating CD8 T lymphocytes: implications of a novel approach for T cell monitoring in infectious and malignant diseases. J. Immunol. 166: 7634 –7640. 3. Altman, J. D., P. A. Moss, P. J. Goulder, D. H. Barouch, M. G. McHeyzer-Williams, J. I. Bell, A. J. McMichael, and M. M. Davis. 1996. Phenotypic analysis of antigenspecific T lymphocytes. Science 274: 94 –96. 4. Callan, M. F., L. Tan, N. Annels, G. S. Ogg, J. D. Wilson, C. A. O’Callaghan, N. Steven, A. J. McMichael, and A. B. Rickinson. 1998. Direct visualization of antigen-specific CD8⫹ T cells during the primary immune response to EpsteinBarr virus in vivo. J. Exp. Med. 187: 1395–1402. 5. Romero, P., P. R. Dunbar, D. Valmori, M. Pittet, G. S. Ogg, D. Rimoldi, J. L. Chen, D. Lienard, J. C. Cerottini, and V. Cerundolo. 1998. Ex vivo staining

The Journal of Immunology

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

of metastatic lymph nodes by class I major histocompatibility complex tetramers reveals high numbers of antigen-experienced tumor-specific cytolytic T lymphocytes. J. Exp. Med. 188: 1641–1650. Gallimore, A., A. Glithero, A. Godkin, A. C. Tissot, A. Pluckthun, T. Elliott, H. Hengartner, and R. Zinkernagel. 1998. Induction and exhaustion of lymphocytic choriomeningitis virus-specific cytotoxic T lymphocytes visualized using soluble tetrameric major histocompatibility complex class I-peptide complexes. J. Exp. Med. 187: 1383–1393. Gillespie, G. M., M. R. Wills, V. Appay, C. O’Callaghan, M. Murphy, N. Smith, P. Sissons, S. Rowland-Jones, J. I. Bell, and P. A. Moss. 2000. Functional heterogeneity and high frequencies of cytomegalovirus-specific CD8⫹ T lymphocytes in healthy seropositive donors. J. Virol. 74: 8140 – 8150. Hislop, A. D., N. H. Gudgeon, M. F. Callan, C. Fazou, H. Hasegawa, M. Salmon, and A. B. Rickinson. 2001. EBV-specific CD8⫹ T cell memory: relationships between epitope specificity, cell phenotype, and immediate effector function. J. Immunol. 167: 2019 –2029. Champagne, P., G. S. Ogg, A. S. King, C. Knabenhans, K. Ellefsen, M. Nobile, V. Appay, G. P. Rizzardi, S. Fleury, M. Lipp, et al. 2001. Skewed maturation of memory HIV-specific CD8 T lymphocytes. Nature 410: 106 –111. Gamadia, L. E., R. J. Rentenaar, P. A. Baars, E. B. Remmerswaal, S. Surachno, J. F. Weel, M. Toebes, T. N. Schumacher, I. J. ten Berge, and R. A. van Lier. 2001. Differentiation of cytomegalovirus-specific CD8⫹ T cells in healthy and immunosuppressed virus carriers. Blood 98: 754 –761. Appay, V., P. R. Dunbar, M. Callan, P. Klenerman, G. M. Gillespie, L. Papagno, G. S. Ogg, A. King, F. Lechner, C. A. Spina, et al. 2002. Memory CD8⫹ T cells vary in differentiation phenotype in different persistent virus infections. Nat. Med. 8: 379 –385. Papagno, L., C. A. Spina, A. Marchant, M. Salio, N. Rufer, S. Little, T. Dong, G. Chesney, A. Waters, P. Easterbrook, et al. 2004. Immune activation and CD8⫹ T-cell differentiation towards senescence in HIV-1 infection. PLoS Biol. 2: E20. Tomiyama, H., H. Takata, T. Matsuda, and M. Takiguchi. 2004. Phenotypic classification of human CD8⫹ T cells reflecting their function: inverse correlation between quantitative expression of CD27 and cytotoxic effector function. Eur. J. Immunol. 34: 999 –1010. Sallusto, F., D. Lenig, R. Forster, M. Lipp, and A. Lanzavecchia. 1999. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions. Nature 401: 708 –712. Sallusto, F., J. Geginat, and A. Lanzavecchia. 2004. Central memory and effector memory T cell subsets: function, generation, and maintenance. Annu. Rev. Immunol. 22: 745–763. Sallusto, F., C. R. Mackay, and A. Lanzavecchia. 2000. The role of chemokine receptors in primary, effector, and memory immune responses. Annu. Rev. Immunol. 18: 593– 620. Rufer, N., A. Zippelius, P. Batard, M. J. Pittet, I. Kurth, P. Corthesy, J. C. Cerottini, S. Leyvraz, E. Roosnek, M. Nabholz, and P. Romero. 2003. Ex vivo characterization of human CD8⫹ T subsets with distinct replicative history and partial effector functions. Blood 102: 1779 –1787. Rufer, N., P. Reichenbach, and P. Romero. 2005. Methods for the ex vivo characterization of human CD8⫹ T subsets based on gene expression and replicative history analysis. Methods Mol. Med. 109: 265–284. Douek, D. C., R. D. McFarland, P. H. Keiser, E. A. Gage, J. M. Massey, B. F. Haynes, M. A. Polis, A. T. Haase, M. B. Feinberg, J. L. Sullivan, et al. 1998. Changes in thymic function with age and during the treatment of HIV infection. Nature 396: 690 – 695. Zippelius, A., M. J. Pittet, P. Batard, N. Rufer, M. de Smedt, P. Guillaume, K. Ellefsen, D. Valmori, D. Lienard, J. Plum, et al. 2002. Thymic selection generates a large T cell pool recognizing a self-peptide in humans. J. Exp. Med. 195: 485– 494. Rufer, N., W. Dragowska, G. Thornbury, E. Roosnek, and P. M. Lansdorp. 1998. Telomere length dynamics in human lymphocyte subpopulations measured by flow cytometry. Nat. Biotechnol. 16: 743–747. Rufer, N., T. H. Brummendorf, S. Kolvraa, C. Bischoff, K. Christensen, L. Wadsworth, M. Schulzer, and P. M. Lansdorp. 1999. Telomere fluorescence measurements in granulocytes and T lymphocyte subsets point to a high turnover of hematopoietic stem cells and memory T cells in early childhood. J. Exp. Med. 190: 157–167. Speiser, D. E., P. Baumgaertner, C. Barbey, V. Rubio-Godoy, A. Moulin, P. Corthesy, E. Devevre, P. Y. Dietrich, D. Rimoldi, D. Lienard, et al. 2006. A

4119

24.

25.

26. 27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37. 38.

39.

40.

41.

42.

43.

novel approach to characterize clonality and differentiation of human melanomaspecific T cell responses: spontaneous priming and efficient boosting by vaccination. J. Immunol. 177: 1338 –1348. Geginat, J., A. Lanzavecchia, and F. Sallusto. 2003. Proliferation and differentiation potential of human CD8⫹ memory T-cell subsets in response to antigen or homeostatic cytokines. Blood 101: 4260 – 4266. Turka, L. A., J. A. Ledbetter, K. Lee, C. H. June, and C. B. Thompson. 1990. CD28 is an inducible T cell surface antigen that transduces a proliferative signal in CD3⫹ mature thymocytes. J. Immunol. 144: 1646 –1653. Harley, C. B., A. B. Futcher, and C. W. Greider. 1990. Telomeres shorten during ageing of human fibroblasts. Nature 345: 458 – 460. Weng, N. P., B. L. Levine, C. H. June, and R. J. Hodes. 1995. Human naive and memory T lymphocytes differ in telomeric length and replicative potential. Proc. Natl. Acad. Sci. USA 92: 11091–11094. Speiser, D. E., M. Migliaccio, M. J. Pittet, D. Valmori, D. Lienard, F. Lejeune, P. Reichenbach, P. Guillaume, I. Luscher, J. C. Cerottini, and P. Romero. 2001. Human CD8⫹ T cells expressing HLA-DR and CD28 show telomerase activity and are distinct from cytolytic effector T cells. Eur. J. Immunol. 31: 459 – 466. Wills, M. R., G. Okecha, M. P. Weekes, M. K. Gandhi, P. J. Sissons, and A. J. Carmichael. 2002. Identification of naive or antigen-experienced human CD8⫹ T cells by expression of costimulation and chemokine receptors: analysis of the human cytomegalovirus-specific CD8⫹ T cell response. J. Immunol. 168: 5455–5464. Volkert, M., O. Marker, and K. Bro-Jorgensen. 1974. Two populations of T lymphocytes immune to the lymphocytic choriomeningitis virus. J. Exp. Med. 139: 1329 –1343. Rocha, B., and C. Tanchot. 2006. The Tower of Babel of CD8⫹ T-cell memory: known facts, deserted roads, muddy waters, and possible dead ends. Immunol. Rev. 211: 182–196. Amyes, E., A. J. McMichael, and M. F. Callan. 2005. Human CD4⫹ T cells are predominantly distributed among six phenotypically and functionally distinct subsets. J. Immunol. 175: 5765–5773. Bachmann, M. F., P. Wolint, K. Schwarz, and A. Oxenius. 2005. Recall proliferation potential of memory CD8⫹ T cells and antiviral protection. J. Immunol. 175: 4677– 4685. Bachmann, M. F., P. Wolint, K. Schwarz, P. Jager, and A. Oxenius. 2005. Functional properties and lineage relationship of CD8⫹ T cell subsets identified by expression of IL-7 receptor ␣ and CD62L. J. Immunol. 175: 4686 – 4696. Wherry, E. J., V. Teichgraber, T. C. Becker, D. Masopust, S. M. Kaech, R. Antia, U. H. von Andrian, and R. Ahmed. 2003. Lineage relationship and protective immunity of memory CD8 T cell subsets. Nat. Immunol. 4: 225–234. Roberts, A. D., K. H. Ely, and D. L. Woodland. 2005. Differential contributions of central and effector memory T cells to recall responses. J. Exp. Med. 202: 123–133. Lenschow, D. J., T. L. Walunas, and J. A. Bluestone. 1996. CD28/B7 system of T cell costimulation. Annu. Rev. Immunol. 14: 233–258. Hendriks, J., L. A. Gravestein, K. Tesselaar, R. A. van Lier, T. N. Schumacher, and J. Borst. 2000. CD27 is required for generation and long-term maintenance of T cell immunity. Nat. Immunol. 1: 433– 440. Hathcock, K. S., S. M. Kaech, R. Ahmed, and R. J. Hodes. 2003. Induction of telomerase activity and maintenance of telomere length in virus-specific effector and memory CD8⫹ T cells. J. Immunol. 170: 147–152. Kaech, S. M., J. T. Tan, E. J. Wherry, B. T. Konieczny, C. D. Surh, and R. Ahmed. 2003. Selective expression of the interleukin 7 receptor identifies effector CD8 T cells that give rise to long-lived memory cells. Nat. Immunol. 4: 1191–1198. Maini, M. K., M. V. Soares, C. F. Zilch, A. N. Akbar, and P. C. Beverley. 1999. Virus-induced CD8⫹ T cell clonal expansion is associated with telomerase upregulation and telomere length preservation: a mechanism for rescue from replicative senescence. J. Immunol. 162: 4521– 4526. Hamann, D., S. Kostense, K. C. Wolthers, S. A. Otto, P. A. Baars, F. Miedema, and R. A. van Lier. 1999. Evidence that human CD8⫹CD45RA⫹CD27⫺ cells are induced by antigen and evolve through extensive rounds of division. Int. Immunol. 11: 1027–1033. Ochsenbein, A. F., S. R. Riddell, M. Brown, L. Corey, G. M. Baerlocher, P. M. Lansdorp, and P. D. Greenberg. 2004. CD27 expression promotes longterm survival of functional effector-memory CD8⫹ cytotoxic T lymphocytes in HIV-infected patients. J. Exp. Med. 200: 1407–1417.