Four Functionally Distinct Populations of Human
Effector-Memory CD8?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) CD45RA?CCR7?T lymphocytes present within the circulating CD8?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(CD27?CD28?) and EM4(CD27?CD28?) T cells express low levels of effector mediators such
as granzyme B and perforin and high levels of CD127/IL-7R?. EM1cells also have a relatively short replicative history and display
strong ex vivo telomerase activity. Therefore, these cells are closely related to central-memory (CD45RA?CCR7?) cells. In
contrast, EM2(CD27?CD28?) and EM3(CD27?CD28?) cells express mediators characteristic of effector cells, whereby EM3cells
display stronger ex vivo cytolytic activity and have experienced larger numbers of cell divisions, thus resembling differentiated
effector (CD45RA?CCR7?) cells. These data indicate that progressive up-regulation of cytolytic activity and stepwise loss of
CCR7, CD28, and CD27 both characterize CD8?T cell differentiation. Finally, memory CD8?T cells not only include central-
memory cells but also EM1cells, 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.
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-express-
ing 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 differentia-
tion. The introduction of multimers of MHC/Ag peptide, that bind
stably to specific TCR on the surface of T cells, has made it pos-
sible 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
pon productive interaction between mature Ag-present-
ing dendritic cells and specific but functionally naive T
lymphocytes, the latter undergo both clonal expansion
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 chemo-
kine 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 sec-
ondary lymphoid tissues (reviewed in Ref. 16). Thus, CCR7?na-
ive and central-memory (CM)3T 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 lympho-
cytes down-regulate the CCR7 and appear specialized in migrating
to peripheral nonlymphoid tissues. Although this two-marker pro-
cedure 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 sub-
set by studying the functional attributes of pools of EM cells sep-
arated on the basis of the various combinations of costimulatory
receptor (i.e., CD27 and CD28) cell surface expression. We pro-
pose 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
*Division of Clinical Onco-Immunology, Ludwig Institute for Cancer Research, Lau-
sanne 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, Switzer-
land;§Columbia University Medical Center, Irving Cancer Research Center, New
York, NY 10032; and¶Center for Molecular Imaging Research, Massachusetts Gen-
eral Hospital, Harvard Medical School, Charlestown, MA 02129
Received for publication September 29, 2006. Accepted for publication January
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.
1This work was sponsored and supported by the Swiss National Center of Compe-
tence in Research Molecular Oncology and Swiss National Science Foundation
Grants 3100-068016 and 3100A0-105929. A.Z. was supported in part by the Emmy-
Noether Program of the Deutsche Forschungsgemeinshaft (Zi-685/2.3) and a grant
from the Swiss National Foundation (3200B0-103608/1).
2Address 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: Nathalie.Rufer@isrec.ch
3Abbreviations 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 EM1subset is nearly identical
with the CM one except for the lack of CCR7 cell surface expres-
sion, suggesting that this subset shares functional features with
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 anti-
CD8-coated magnetic microbeads (Miltenyi Biotec), a procedure that elim-
inates most NK cells because they are not efficiently retained by the mag-
net. 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 Bio-
sciences). 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 EM1cells 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 contam-
inating NK cells in the sorted populations (?2%). Intracellular content of
granzyme B and perforin was measured in freshly isolated CD8?T lympho-
cytes without previous stimulation as previously described (17). The follow-
ing mAbs were purchased from BD Biosciences or BD Pharmingen:
anti-CD27-FITC and -PE, anti-CD28-PE and -allophycocyanin, anti-
CD8-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 allo-
phycocyanin-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 pre-
cDNA amplification and five-cell RT-PCR
To avoid contamination of small populations by more abundant subsets,
10 ? 103T 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 ampli-
fication as well as the RT-PCR were recently described in details (17, 18).
Additional primers were used in the present study: CD27, 5?-ACGTGA
3?; CD127/IL-7R?: 5?-ATCTTGGCCTGTGTGTTATGG-3?; reverse-5?-
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 (103cells/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 ? sponta-
neous release) ? 100/(total release ? spontaneous release).
Quantification of TRECs by real-time PCR
The amount of signal joint (sj) TRECs in 5–15 ? 104sorted 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 pro-
teinase 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
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 be-
tween 107and 101copies of sjTREC and four reactions were run with each
dilution. Considering that ?50,000 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 ?0.997, 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 ? 104sorted
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 re-
solved on a 15% polyacrylamide gel and visualized by a phosphoimaging
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 mol-
ecules 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 Melan-
A-, 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 ? 105cells
were further processed by flow FISH. Because the average telomere flu-
orescence from all these clones was evaluated in the same experimental
design, this allowed direct telomere comparison between each tested clonal
Ex vivo distribution of phenotypically distinct CD8?T cell
subsets from human peripheral blood
Human CD8?T lymphocytes can be separated into four func-
tionally 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 expres-
sion 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 pres-
ence of nine discrete subpopulations in the blood from a rep-
resentative healthy donor (Fig. 1B). Naive and CM cells uni-
formly coexpressed CD27 and CD28. In contrast, as based on
4113 The Journal of Immunology
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 ex-
pression of CD27 and CD28 cell surface molecules (Fig. 1B).
Thus, we identified, within the RA?CCR7?EM T compart-
ment, 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 pro-
gressive down-regulation of CD27 cell surface expression from
naive, through CM and EM1, to EM2cells (Fig. 1C). In contrast,
cell surface CD28 expression levels were higher in CM, EM1, and
EM4cells than in naive cells. Finally, the EM3subset 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 ob-
tain an overview of the distribution of the nine CD8?T cell sub-
sets defined by a different pattern of CD45RA, CCR7, CD27, and
CD28 expression, we analyzed PBMC from 14 healthy individu-
als, 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 EM1CD8?T
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 differ-
entiate through a CD27?CD28?to a CD27?CD28?stage (11,
17). According to this model, CD8?T cells sequentially down-
regulate CCR7, CD28, and CD27 surface expression, while up-
regulating expression of molecules that confer cytolytic activity.
To investigate whether EM1, EM2, and EM3CD8?T cell subsets
corresponding to the proposed differentiation pathway (11, 17) dif-
fered in the expression of genes involved in T cell effector func-
tions, 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, per-
forin, IFN-?, or NK receptor CD94 mRNA (Fig. 2A). Interestingly,
despite the loss of CCR7 expression, the gene expression profile of
EM1T 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 do-
nors (Fig. 2A; see HD1, HD2, and HD3). In contrast, EM2and
EM3T 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 ex-
pression level for this protein (Fig. 2B). Moreover, most of naive,
CM, and EM1cells, but almost none of the aliquots of EM3and
effector T cells, yielded a detectable CD127-specific product, en-
coding 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 cyto-
lytic activity of these distinct cell populations, using a CD3
found that both EM3and effector T cells displayed high ex vivo
lytic activity, whereas CM and EM1T cells had comparable
killing activity, which was ?10 times lower than that of the
51Cr release assay (Fig. 2C), we
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 label-
ing. 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?, CM-
CD27?, EM-CD27?(comprising EM1?EM2), EM-
CD27?(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 EMRA-
CD28?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).
Differential expression of CD45RA,
4114EX VIVO HUMAN EM T CELL SUBSETS
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
EM3T cell subsets by quantifying their content of TRECs, which
are stable DNA episomes formed during TCR-? gene rearrange-
ment 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 (?0.01 TREC copies/100 cells) in the EM3and effector
subsets. CM and EM2T cells contained low but detectable TRECs
in all healthy individuals. Intriguingly, TRECs in EM1cells 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
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 do-
nors, 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 ob-
served a progressive reduction in the mean telomere fluores-
cence from naive through CM, EM1, EM2to EM3and effector
T lymphocytes, that corresponded to a telomere shortening of
?4.5–5 kb. EM1cells 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 EM1cells (Fig. 3A).
Importantly, our data revealed that EM3cells had relatively
short telomeres, within the same range than those found in
within EM T cell subsets. A, Gene expression analysis was performed
on sorted naive (RA?CCR7?27?28?), CM (RA?CCR7?27?28?),
(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 indepen-
dent five-cell aliquots per subset, and negative (?) and positive (?) con-
trols, are depicted. Comparable results were obtained in three healthy in-
dividuals. 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 EM3than 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
51Cr-labeled P815 target cells. None of these subsets lysed P815 cells in
absence of CD3 mAbs (lysis ?10%; data not shown). Data are represen-
tative of two healthy donors.
Ex vivo analysis of expression of effector mediators
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). ?, Not detectable (sorted cell number was 5 ? 104to 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, Telom-
erase 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.
Ex vivo analysis of the replicative history and telomerase
4115The Journal of Immunology
differentiated effector cells. This agrees with our recently pub-
lished results (12) showing progressive shortening of the telo-
meres along T cell differentiation, so that highly differentiated
CD8?CCR7?CD27?CD57?cells displayed the shortest telo-
meres, with lengths equivalent to those observed in primed
CD8?T cells from the elderly (22). Altogether, these data sup-
port the view that T cells evolve through extensive rounds of
division as they differentiate further.
A previous report (28) showed telomerase activity in ex vivo-
isolated CD8?T cells that express both CD28 and the activation
marker HLA-DR. To determine which if not all of the HLA-DR-
expressing T cell subsets accounted for such activity, TRAP assays
with the HLA-DR?and HLA-DR?fractions of CM and EM1T
cells were conducted (Fig. 3C). Remarkably, we observed positive
telomerase activity selectively confined to the DR-positive frac-
tions of both subsets. In contrast, naive, EM2, and EM3T 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 EM1T cells had been sorted without
discriminating HLA-DR?from HLA-DR?cells (data not shown).
Finally, within both CM and EM1populations, 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
The fourth T cell population (EM4) was analyzed in additional
experiments, as EM4cells could not be fully investigated in par-
allel in the previously shown experiments due to technical limita-
tions. To gain insight into the relationship between EM4T cells
and the EM1subset (as outlined in Fig. 1B), we compared their
expression of genes involved in effector functions within five
cell-sorted samples (Fig. 4A). In both EM1and EM4cells, no
granzyme B mRNA transcripts were detected, whereas CD127/
IL7R?, perforin, and IFN-? transcripts were found in a signif-
icant proportion of these samples. Moreover, these results cor-
related 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 EM4compartment, these cells seem closely related to
the EM1T cells.
Ag-specific CD8?T lymphocytes exhibit distinct differentiation
phenotypes and replicative history
Several studies have reported that Ag-specific CD8?T cells di-
rected 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 circulat-
ing Melan-A-specific T cells from HLA-A2 healthy donors is phe-
notypically 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 in-
dividuals. Whereas Melan-A-specific T cells shared a homoge-
neous naive-like phenotype (CCR7?CD45RA?CD27?CD28?),
the EBV-specific response consisted primarily of early differ-
entiated T cells (EM1, mean ? SD, 60 ? 16%; n ? 8). Con-
sistent with our recent report (17), CMV-specific T lympho-
cytes displayed the phenotype of effector CD8?T cells
(mean ? SD, 18 ? 7%; n ? 8), but also the phenotype of EM2
(19 ? 8%) and EM3cells (34 ? 14%). Interestingly, our data
further revealed that Melan-A-specific naive T cells, as com-
pared 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 in-
volves 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-spe-
cific 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 pur-
pose, 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 Flu-
specific 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.
effector mediators between EM1and EM4T cell sub-
sets. A, Gene expression analysis was performed on
sorted CM, EM3, EM1, and EM4CD8?T cells by RT-
PCR. Data from three or eight independent five-cell ali-
quots 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.
Comparison of the ex vivo expression of
4116 EX VIVO HUMAN EM T CELL SUBSETS
Circulating naive T lymphocytes form a relatively homogeneous
population expressing a well-defined set of cell surface glycopro-
teins and are characterized by the null expression of effector me-
diators (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-expe-
rienced T lymphocytes have mostly been classified into two dis-
tinct subpopulations, e.g., effector and memory cells (30). Effectors
are presumably rather short-lived, produce cytolytic effector mol-
ecules 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 defin-
ing 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 Ag-
specific 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 chal-
lenge to protect against vaccinia virus, whereas CM cells were
most potent at protecting against systemic infection with lympho-
cytic choriomeningitis virus (33, 34), as they have a greater ca-
pacity 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 phe-
notype, functional capacities, and history of Ag encounter.
Here, by combining the simultaneous analysis of surface mark-
ers 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 regula-
tion 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
phenotype and replicative history de-
pending on Ag specificity. A and B,
Melan-A, Flu-, EBV-, and CMV-mul-
timer?T cells were characterized ex
vivo by flow cytometry for their cell
surface expression of CD45RA, CCR7,
CD28, and CD27. The dot plots show
and CD28/CD27 (B) on Melan-A
(HLA-A2), Flu (HLA-A2/Matrix pro-
tein), 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 ex-
pressions is depicted. Eight healthy in-
dividuals were included in these analy-
ses. Representative data are shown. C,
Enlarged dot plots of CD28 and CD27
costaining on Melan-A (multimer-allo-
phycocyanin) and EBV (multimer-PE)
resent the reference for CD28 and
CD27 positivity based on the fluores-
cent 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 Melan-
A-, 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 con-
verted to kilobase as described in Ma-
terials and Methods.
CD8?T cells vary in
4117The Journal of Immunology
in the peripheral blood of healthy donors with a predominance
toward EM1cells (Fig. 1D). Functionally, at least three subsets can
be clearly identified: EM1(that includes the very similar EM4),
EM2, and EM3. EM1are memory-like, EM2are intermediate with
partial effector functions and replicative history, and EM3are ef-
fector-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 expres-
sion concomitant with up-regulation of cytolytic capacity (11).
In particular, we show that both circulating EM2and EM3
CD8?T cells express mediators characteristic of effector cells, but
gene and protein expression profiles of EM3T cells more closely
resemble that of effector cells. In line with this notion, EM3cells
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, EM2cells contain low but
yet detectable levels of TRECs (Fig. 3A). Unless EM2cells dif-
ferentiate 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 EM1T cells, despite
their lack of CCR7 expression, have several functional features in
common with CM cells. Both populations have a similar replica-
tive 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 EM1T cells, our data dem-
onstrate 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 HLA-
DR, thus supporting the idea that telomerase expression in HLA-
DR?CM and EM1cells reflects proliferative activity while main-
taining telomere lengths. This would be compatible with reduced
levels of TRECs but only progressive shortening of telomere
lengths as observed in the EM1cells (Fig. 3). It is also noteworthy
to mention that induction of telomerase activity in Ag-specific ef-
fector and memory CD8?T cells from mice infected with lym-
phocytic choriomeningitis virus has been suggested to be impor-
tant for the maintenance and longevity of the memory CD8?T cell
population (39). Finally, both CM and EM1T 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, ob-
tained with cells analyzed directly ex vivo, indicate that the CM
and EM1subpopulations may comprise T cells that have been
recently activated following antigenic challenge. Alternatively, te-
lomerase-expressing T cells may consist of memory cells that ex-
pand following homeostatic maintenance of T cell numbers. In-
vestigations 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 ex-
cept 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
subsets allows for rapid reactivation upon Ag challenge regardless
of tissue location.
Loss of CD27 during clonal expansion and differentiation pre-
sumably 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 pro-
liferative 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
EM1cells, this subset displays low levels of effector-mediated
molecules, while still expressing IL-7R? (Fig. 4). One possibility
is that EM4cells differentiate from the CM T cell pool and that
such cells represent a transitory subset appearing during the ex-
pansion phase of a secondary immune response. Alternatively, one
cannot formally exclude that these cells have emerged directly
from the EM1pool and have transiently or definitively down-reg-
ulated 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 ad-
vantages from an experimental viewpoint. On one hand, infections
by these viruses are frequent and independent of geographical lo-
cation. 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 character-
ize the corresponding Ag-specific CD8?T cell responses.
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.
The authors have no financial conflict of interest.
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 ap-
proach for T cell monitoring in infectious and malignant diseases. J. Immunol.
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 antigen-
specific 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 Epstein-
Barr 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
4118 EX VIVO HUMAN EM T CELL SUBSETS
of metastatic lymph nodes by class I major histocompatibility complex tetramers Download full-text
reveals high numbers of antigen-experienced tumor-specific cytolytic T lympho-
cytes. J. Exp. Med. 188: 1641–1650.
6. Gallimore, A., A. Glithero, A. Godkin, A. C. Tissot, A. Pluckthun, T. Elliott,
H. Hengartner, and R. Zinkernagel. 1998. Induction and exhaustion of lympho-
cytic choriomeningitis virus-specific cytotoxic T lymphocytes visualized using
soluble tetrameric major histocompatibility complex class I-peptide complexes.
J. Exp. Med. 187: 1383–1393.
7. 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 het-
erogeneity and high frequencies of cytomegalovirus-specific CD8?T lympho-
cytes in healthy seropositive donors. J. Virol. 74: 8140–8150.
8. 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.
9. 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.
10. 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.
11. 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.
12. 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.
13. 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.
14. 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.
15. Sallusto, F., J. Geginat, and A. Lanzavecchia. 2004. Central memory and effector
memory T cell subsets: function, generation, and maintenance. Annu. Rev. Im-
munol. 22: 745–763.
16. Sallusto, F., C. R. Mackay, and A. Lanzavecchia. 2000. The role of chemokine
receptors in primary, effector, and memory immune responses. Annu. Rev. Im-
munol. 18: 593–620.
17. 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.
18. Rufer, N., P. Reichenbach, and P. Romero. 2005. Methods for the ex vivo char-
acterization of human CD8?T subsets based on gene expression and replicative
history analysis. Methods Mol. Med. 109: 265–284.
19. 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.
20. 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.
21. 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.
22. 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.
23. 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
novel approach to characterize clonality and differentiation of human melanoma-
specific T cell responses: spontaneous priming and efficient boosting by vacci-
nation. J. Immunol. 177: 1338–1348.
24. Geginat, J., A. Lanzavecchia, and F. Sallusto. 2003. Proliferation and differen-
tiation potential of human CD8?memory T-cell subsets in response to antigen or
homeostatic cytokines. Blood 101: 4260–4266.
25. 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.
26. Harley, C. B., A. B. Futcher, and C. W. Greider. 1990. Telomeres shorten during
ageing of human fibroblasts. Nature 345: 458–460.
27. 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.
28. 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.
29. 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:
30. Volkert, M., O. Marker, and K. Bro-Jorgensen. 1974. Two populations of T
lymphocytes immune to the lymphocytic choriomeningitis virus. J. Exp. Med.
31. 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.
32. 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.
33. Bachmann, M. F., P. Wolint, K. Schwarz, and A. Oxenius. 2005. Recall prolif-
eration potential of memory CD8?T cells and antiviral protection. J. Immunol.
34. Bachmann, M. F., P. Wolint, K. Schwarz, P. Jager, and A. Oxenius. 2005. Func-
tional properties and lineage relationship of CD8?T cell subsets identified by
expression of IL-7 receptor ? and CD62L. J. Immunol. 175: 4686–4696.
35. 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.
36. 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:
37. Lenschow, D. J., T. L. Walunas, and J. A. Bluestone. 1996. CD28/B7 system of
T cell costimulation. Annu. Rev. Immunol. 14: 233–258.
38. 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.
39. 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.
40. 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:
41. 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 up-
regulation and telomere length preservation: a mechanism for rescue from rep-
licative senescence. J. Immunol. 162: 4521–4526.
42. 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. Immu-
nol. 11: 1027–1033.
43. Ochsenbein, A. F., S. R. Riddell, M. Brown, L. Corey, G. M. Baerlocher,
P. M. Lansdorp, and P. D. Greenberg. 2004. CD27 expression promotes long-
term survival of functional effector-memory CD8?cytotoxic T lymphocytes in
HIV-infected patients. J. Exp. Med. 200: 1407–1417.
4119 The Journal of Immunology