KLRG1+NKG2A+ CD8 T cells mediate protection and participate in memory responses during γ-herpesvirus infection

Functional CD8 T cell effector and memory responses are generated and maintained during murine γ-herpesvirus 68 (γHV68) persistent infection despite continuous presentation of viral lytic Ags. However, the identity of the CD8 T cell subpopulations that mediate effective recall responses and that can participate in the generation of protective memory to a γ-herpesvirus infection remains unknown. During γHV68 persistence, ～75% of γHV68-specific CD8 T cells coexpress the NK receptors killer cell lectin-like receptor G1 (KLRG1) and NKG2A. In this study, we take advantage of this unique phenotype to analyze the capacity of CD8 T cells expressing or not expressing KLRG1 and NKG2A to mediate effector and memory responses. Our results show that γHV68-specific KLRG1(+)NKG2A(+) CD8 T cells have an effector memory phenotype as well as characteristics of polyfunctional effector cells such us IFN-γ and TNF-α production, killing capacity, and are more efficient at protecting against a γHV68 challenge than their NKG2A(-)KLRG1(-) counterparts. Nevertheless, γHV68-specific NKG2A(+)KLRG1(+) CD8 T cells express IL-7 and IL-15 receptors, can survive long-term without cognate Ag, and subsequently mount a protective response during antigenic recall. These results highlight the plasticity of the immune system to generate protective effector and proliferative memory responses during virus persistence from a pool of KLRG1(+)NKG2A(+) effector memory CD8 T cells.     

Influenza-specific T cells from older people are enriched in the late effector subset and their presence inversely correlates with vaccine response

T cells specific for persistent pathogens accumulate with age and express markers of immune senescence. In contrast, much less is known about the state of T cell memory for acutely infecting pathogens. Here we examined T cell responses to influenza in human peripheral blood mononuclear cells from older (>64) and younger (<40) donors using whole virus restimulation with influenza A (A/PR8/34) ex vivo. Although most donors had pre-existing influenza reactive T cells as measured by IFNγ production, older donors had smaller populations of influenza-responsive T cells than young controls and had lost a significant proportion of their CD45RA-negative functional memory population. Despite this apparent dysfunction in a proportion of the older T cells, both old and young donors' T cells from 2008 could respond to A/California/07/2009 ex vivo. For HLA-A2+ donors, MHC tetramer staining showed that a higher proportion of influenza-specific memory CD8 T cells from the 65+ group co-express the markers killer cell lectin-like receptor G1 (KLRG1) and CD57 compared to their younger counterparts. These markers have previously been associated with a late differentiation state or immune senescence. Thus, memory CD8 T cells to an acutely infecting pathogen show signs of advanced differentiation and functional deterioration with age. There was a significant negative correlation between the frequency of KLRG1(+)CD57(+) influenza M1-specific CD8 T cells pre-vaccination and the ability to make antibodies in response to vaccination with seasonal trivalent inactivated vaccine, whereas no such trend was observed when the total CD8(+)KLRG1(+)CD57(+) population was analyzed. These results suggest that the state of the influenza-specific memory CD8 T cells may be a predictive indicator of a vaccine responsive healthy immune system in old age.     

Expression of killer cell lectin-like receptor G1 on antigen-specific human CD8+ T lymphocytes during active, latent, and resolved infection and its relation with CD57

Killer cell lectin-like receptor G1 (KLRG1) is one of several inhibitory killer cell lectin-like receptors expressed by NK cells and T lymphocytes, mainly CD8(+) effector/memory cells that can secrete cytokines but have poor proliferative capacity. Using multiparameter flow cytometry, we studied KLRG1 expression on CD8(+) T cells specific for epitopes of CMV, EBV, influenza, and HIV. Over 92% of CD8(+) cells specific for CMV or EBV expressed KLRG1 during the latent stage of these chronic infections. CD8(+) T cell cells specific for HIV epitopes were mostly (72-89%) KLRG1(+), even though not quite at the level of predominance noted with CMV or EBV. Lower frequency of KLRG1 expression was observed among CD8(+) cells specific for influenza (40-73%), a resolved infection without a latent stage. We further observed that CD8(+) cells expressing CD57, a marker of replicative senescence, also expressed KLRG1; however, a population of CD57(-)KLRG1(+) cells was also identified. This population may represent a "memory" phenotype, because they also expressed CD27, CD28, CCR7, and CD127. In contrast, CD57(+)KLRG1(+) cells did not express CD27, CD28, and CCR7, and expressed CD127 at a much lower frequency, indicating that they represent effector cells that are truly terminally differentiated. The combination of KLRG1 and CD57 expression might thus aid in refining functional characterization of CD8(+) T cell subsets.     

Gamma interferon and perforin control the strength, but not the hierarchy, of immunodominance of an antiviral CD8+ T cell response

The two major antiviral effector mechanisms of CD8(+) T cells are thought to be perforin (Prf)-mediated cell lysis and gamma interferon (IFN-γ)-mediated induction of an antiviral state. By affecting the expression of proteins involved in antigen presentation, IFN-γ is also thought to shape the magnitude and specificity of the CD8(+) T cell response. Here we studied the roles of Prf and IFN-γ in shaping the effector and memory CD8(+) T cell responses to vaccinia virus (VACV). IFN-γ deficiency resulted in increased numbers of anti-VACV effector and memory CD8(+) T cells, which were partly dependent on increased virus loads. On the other hand, Prf-deficient mice showed an increase in the number of VACV-specific CD8(+) T cells only in the memory phase. Treatment of the mice with the antiviral drug cidofovir reduced the numbers of effector and memory cells closer to wild-type levels in IFN-γ-deficient mice and reduced the numbers of memory CD8(+) T cells to wild-type levels in Prf-deficient mice. These data suggest that virus loads are the main reason for the increased strength of the CD8 response in IFN-γ- and Prf-deficient mice. Neither Prf deficiency nor IFN-γ deficiency had an effect on the immunodominance hierarchy of five K(b)-restricted CD8(+) T cell determinants either during acute infection or after recovery. Thus, our work shows that CD8(+) T cell immunodominance during VACV infection is not affected by the effects of IFN-γ on the antigen presentation machinery.     

Pathogen-induced inflammatory environment controls effector and memory CD8+ T cell differentiation

In response to infection, CD8(+) T cells integrate multiple signals and undergo an exponential increase in cell numbers. Simultaneously, a dynamic differentiation process occurs, resulting in the formation of short-lived effector cells (SLECs; CD127(low)KLRG1(high)) and memory precursor effector cells (CD127(high)KLRG1(low)) from an early effector cell that is CD127(low)KLRG1(low) in phenotype. CD8(+) T cell differentiation during vesicular stomatitis virus infection differed significantly than during Listeria monocytogenes infection with a substantial reduction in early effector cell differentiation into SLECs. SLEC generation was dependent on Ebi3 expression. Furthermore, SLEC differentiation during vesicular stomatitis virus infection was enhanced by administration of CpG-DNA, through an IL-12-dependent mechanism. Moreover, CpG-DNA treatment enhanced effector CD8(+) T cell functionality and memory subset distribution, but in an IL-12-independent manner. Population dynamics were dramatically different during secondary CD8(+) T cell responses, with a much greater accumulation of SLECs and the appearance of a significant number of CD127(high)KLRG1(high) memory cells, both of which were intrinsic to the memory CD8(+) T cell. These subsets persisted for several months but were less effective in recall than memory precursor effector cells. Thus, our data shed light on how varying the context of T cell priming alters downstream effector and memory CD8(+) T cell differentiation.     

Viral infections induce abundant numbers of senescent CD8 T cells.

Viral infections are often accompanied by extensive proliferation of reactive CD8 T cells. After a defined number of divisions, normal somatic cells enter a nonreplicative stage termed senescence. In the present study we have identified the inhibitory killer cell lectin-like receptor G1 (KLRG1) as a unique marker for replicative senescence of murine CD8 T cells. KLRG1 expression was induced in a substantial portion (30-60%) of CD8 T cells in C57BL/6 mice infected with lymphocytic choriomeningitis virus (LCMV), vesicular stomatitis virus, or vaccinia virus. Similarly, KLRG1 was found on a large fraction of LCMV gp33 peptide-specific TCR-transgenic (tg) effector and memory cells activated in vivo using an adoptive transfer model. Transfer experiments with CFSE-labeled TCR-tg cells into LCMV-infected hosts further indicated that induction of KLRG1 expression required an extensive number of cell divisions. Most importantly, KLRG1(+) TCR-tg effector/memory cells could efficiently lyse target cells and secrete cytokines, but were severely impaired in their ability to proliferate after Ag stimulation. Thus, this study demonstrates that senescent CD8 T cells are induced in abundant numbers during viral infections in vivo.     

CD4+ T Cell-Dependent IFN-γ Production by CD8+ Effector T Cells in Mycobacterium tuberculosis Infection

Both CD4(+) and CD8(+) T cells contribute to immunity to tuberculosis, and both can produce the essential effector cytokine IFN-γ. However, the precise role and relative contribution of each cell type to in vivo IFN-γ production are incompletely understood. To identify and quantitate the cells that produce IFN-γ at the site of Mycobacterium tuberculosis infection in mice, we used direct intracellular cytokine staining ex vivo without restimulation. We found that CD4(+) and CD8(+) cells were predominantly responsible for production of this cytokine in vivo, and we observed a remarkable linear correlation between the fraction of CD4(+) cells and the fraction of CD8(+) cells producing IFN-γ in the lungs. In the absence of CD4(+) cells, a reduced fraction of CD8(+) cells was actively producing IFN-γ in vivo, suggesting that CD4(+) effector cells are continually required for optimal IFN-γ production by CD8(+) effector cells. Accordingly, when infected mice were treated i.v. with an MHC-II-restricted M. tuberculosis epitope peptide to stimulate CD4(+) cells in vivo, we observed rapid activation of both CD4(+) and CD8(+) cells in the lungs. Indirect activation of CD8(+) cells was dependent on the presence of CD4(+) cells but independent of IFN-γ responsiveness of the CD8(+) cells. These data provide evidence that CD4(+) cell deficiency impairs IFN-γ production by CD8(+) effector cells and that ongoing cross-talk between distinct effector T cell types in the lungs may contribute to a protective immune response against M. tuberculosis. Conversely, defects in these interactions may contribute to susceptibility to tuberculosis and other infections.     

IL-12 signaling drives CD8+ T cell IFN-gamma production and differentiation of KLRG1+ effector subpopulations during Toxoplasma gondii Infection

IFN-gamma-producing CD8(+) T lymphocytes are essential effector cells that mediate protective immunity during murine toxoplasmosis, and yet their effector development remains poorly characterized. Vaccination with the carbamoyl phosphate synthase (CPS) mutant strain of Toxoplasma gondii was used to examine the CD8(+) T cell response in the peritoneal effector site. Four CTL subpopulations with varying effector potentials were defined based on the expression of effector molecules and the cell surface activation markers CD62L and killer cell lectin-like receptor G1 (KLRG1). Further phenotypic analysis revealed that the acquisition of KLRG1 among effector subpopulations correlated with the down-regulation of both IL-7R and CD27, suggesting that KLRG1 marks dominant, end-stage effector cells. Using gene-targeted mice, we tested the in vivo requirements of key IL-12 signaling components for effector CTL differentiation. Contrary to established models of viral and bacterial infection, CD8(+) T cell-intrinsic IL-12 signaling was required for the generation of IFN-gamma-producing CTLs in response to T. gondii. Importantly, the development of the KLRG1(+) effector subpopulations, but not the memory precursor-containing KLRG1(-) effector subset, was critically reliant on IL-12. Furthermore, IL-12 signaling-dependent T-bet expression was also found to be important for differentiation of KLRG1(+) effectors. Our results underscore a vital role for IL-12 in not only the induction of IFN-gamma expression but also in the development of heterogeneous subpopulations of effector CD8(+) T cells generated in response to the intracellular parasite T. gondii.     

Bcl-2 allows effector and memory CD8+ T cells to tolerate higher expression of Bim

As acute infections resolve, most effector CD8(+) T cells die, whereas some persist and become memory T cells. Recent work showed that subsets of effector CD8(+) T cells, identified by reciprocal expression of killer cell lectin-like receptor G1 (KLRG1) and CD127, have different lifespans. Similar to previous reports, we found that effector CD8(+) T cells reported to have a longer lifespan (i.e., KLRG1(low)CD127(high)) have increased levels of Bcl-2 compared with their shorter-lived KLRG1(high)CD127(low) counterparts. Surprisingly, we found that these effector KLRG1(low)CD127(high) CD8(+) T cells also had increased levels of Bim compared with KLRG1(high)CD127(low) cells. Similar effects were observed in memory cells, in which CD8(+) central memory T cells expressed higher levels of Bim and Bcl-2 than did CD8(+) effector memory T cells. Using both pharmacologic and genetic approaches, we found that survival of both subsets of effector and memory CD8(+) T cells required Bcl-2 to combat the proapoptotic activity of Bim. Interestingly, inhibition or absence of Bcl-2 led to significantly decreased expression of Bim in surviving effector and memory T cells. In addition, manipulation of Bcl-2 levels by IL-7 or IL-15 also affected expression of Bim in effector CD8(+) T cells. Finally, we found that Bim levels were significantly increased in effector CD8(+) T cells lacking Bax and Bak. Together, these data indicate that cells having the highest levels of Bim are selected against during contraction of the response and that Bcl-2 determines the level of Bim that effector and memory T cells can tolerate.     

Distinct effector memory CD4+ T cell signatures in latent Mycobacterium tuberculosis infection, BCG vaccination and clinically resolved tuberculosis

Two billion people worldwide are estimated to be latently infected with Mycobacterium tuberculosis (Mtb) and are at risk for developing active tuberculosis since Mtb can reactivate to cause TB disease in immune-compromised hosts. Individuals with latent Mtb infection (LTBI) and BCG-vaccinated individuals who are uninfected with Mtb, harbor antigen-specific memory CD4(+) T cells. However, the differences between long-lived memory CD4(+) T cells induced by latent Mtb infection (LTBI) versus BCG vaccination are unclear. In this study, we characterized the immune phenotype and functionality of antigen-specific memory CD4(+) T cells in healthy BCG-vaccinated individuals who were either infected (LTBI) or uninfected (BCG) with Mtb. Individuals were classified into LTBI and BCG groups based on IFN-γ ELISPOT using cell wall antigens and ESAT-6/CFP-10 peptides. We show that LTBI individuals harbored high frequencies of late-stage differentiated (CD45RA(-)CD27(-)) antigen-specific effector memory CD4(+) T cells that expressed PD-1. In contrast, BCG individuals had primarily early-stage (CD45RA(-)CD27(+)) cells with low PD-1 expression. CD27(+) and CD27(-) as well as PD-1(+) and PD-1(-) antigen-specific subsets were polyfunctional, suggesting that loss of CD27 expression and up-regulation of PD-1 did not compromise their capacity to produce IFN-γ, TNF-α and IL-2. PD-1 was preferentially expressed on CD27(-) antigen-specific CD4(+) T cells, indicating that PD-1 is associated with the stage of differentiation. Using statistical models, we determined that CD27 and PD-1 predicted LTBI versus BCG status in healthy individuals and distinguished LTBI individuals from those who had clinically resolved Mtb infection after anti-tuberculosis treatment. This study shows that CD4(+) memory responses induced by latent Mtb infection, BCG vaccination and clinically resolved Mtb infection are immunologically distinct. Our data suggest that differentiation into CD27(-)PD-1(+) subsets in LTBI is driven by Mtb antigenic stimulation in vivo and that CD27 and PD-1 have the potential to improve our ability to evaluate true LTBI status.     

Heterogeneous and tissue-specific regulation of effector T cell responses by IFN-gamma during Plasmodium berghei ANKA infection

IFN-γ and T cells are both required for the development of experimental cerebral malaria during Plasmodium berghei ANKA infection. Surprisingly, however, the role of IFN-γ in shaping the effector CD4(+) and CD8(+) T cell response during this infection has not been examined in detail. To address this, we have compared the effector T cell responses in wild-type and IFN-γ(-/-) mice during P. berghei ANKA infection. The expansion of splenic CD4(+) and CD8(+) T cells during P. berghei ANKA infection was unaffected by the absence of IFN-γ, but the contraction phase of the T cell response was significantly attenuated. Splenic T cell activation and effector function were essentially normal in IFN-γ(-/-) mice; however, the migration to, and accumulation of, effector CD4(+) and CD8(+) T cells in the lung, liver, and brain was altered in IFN-γ(-/-) mice. Interestingly, activation and accumulation of T cells in various nonlymphoid organs was differently affected by lack of IFN-γ, suggesting that IFN-γ influences T cell effector function to varying levels in different anatomical locations. Importantly, control of splenic T cell numbers during P. berghei ANKA infection depended on active IFN-γ-dependent environmental signals--leading to T cell apoptosis--rather than upon intrinsic alterations in T cell programming. To our knowledge, this is the first study to fully investigate the role of IFN-γ in modulating T cell function during P. berghei ANKA infection and reveals that IFN-γ is required for efficient contraction of the pool of activated T cells.     

OX40 costimulatory signals potentiate the memory commitment of effector CD8+ T cells

A T cell costimulatory molecule, OX40, contributes to T cell expansion, survival, and cytokine production. Although several roles for OX40 in CD8(+) T cell responses to tumors and viral infection have been shown, the precise function of these signals in the generation of memory CD8(+) T cells remains to be elucidated. To address this, we examined the generation and maintenance of memory CD8(+) T cells during infection with Listeria monocytogenes in the presence and absence of OX40 signaling. We used the expression of killer cell lectin-like receptor G1 (KLRG1), a recently reported marker, to distinguish between short-lived effector and memory precursor effector T cells (MPECs). Although OX40 was dispensable for the generation of effector T cells in general, the lack of OX40 signals significantly reduced the number and proportion of KLRG1(low) MPECs, and, subsequently, markedly impaired the generation of memory CD8(+) T cells. Moreover, memory T cells that were generated in the absence of OX40 signals in a host animal did not show self-renewal in a second host, suggesting that OX40 is important for the maintenance of memory T cells. Additional experiments making use of an inhibitory mAb against the OX40 ligand demonstrated that OX40 signals are essential during priming, not only for the survival of KLRG1(low) MPECs, but also for their self-renewing ability, both of which contribute to the homeostasis of memory CD8(+) T cells.     

Differential regulation of killer cell lectin-like receptor G1 expression on T cells

The killer cell lectin-like receptor G1 (KLRG1) is the mouse homologue of the rat mast cell function-associated Ag and contains a tyrosine-based inhibitory motif in its cytoplasmic domain. It has been demonstrated that KLRG1 is induced on activated NK cells and that KLRG1 can inhibit NK cell effector functions. In this study, we show that in naive C57BL/6 mice KLRG1 is expressed on a subset of CD44(high)CD62L(low) T cells. KLRG1 expression can be detected on a small number of V(alpha)14i NK T cells but not on CD8alphaalpha(+) intraepithelial T cells that are either TCRgammadelta(+) or TCRalphabeta(+). We also show that KLRG1 expression is dramatically induced on approximately 50% of the CD8(+) T cells during both a viral and a parasitic infection. Interestingly, during Toxoplasma gondii infection, KLRG1 is up-regulated on CD4(+) T cells. Although KLRG1 expression can be induced on both NK cells and T cells, the molecular mechanism leading to the induction of KLRG1 differs in these two subsets of cells. Indeed, the up-regulation of KLRG1 on NK cells can be driven in vivo by cytokines, whereas KLRG1 cannot be induced on CD8(+) T cells by cytokines. In addition, although induction of KLRG1 on T cells appears to require TCR engagement in vivo, TCR engagement is not sufficient for KLRG1 induction in vitro. Taken together, these data suggest that the expression and induction of KLRG1 on T cells are tightly regulated. This could have important biological consequences on T cell activation and homeostasis.     

Stable CD8+ T cell memory during persistent Trypanosoma cruzi infection

CD8(+) T cell responses to persistent infections caused by intracellular pathogens are dominated by resting T effectors and T effector memory cells, with little evidence suggesting that a T central memory (T(CM)) population is generated. Using a model of Trypanosoma cruzi infection, we demonstrate that in contrast to the T effector/T effector memory phenotype of the majority of T. cruzi-specific CD8(+) T cells, a population of cells displaying hallmark characteristics of T(CM) cells is also present during long-term persistent infection. This population expressed the T(CM) marker CD127 and a subset expressed one or more of three other T(CM) markers: CD62L, CCR7, and CD122. Additionally, the majority of CD127(high) cells were KLRG1(low), indicating that they have not been repetitively activated through TCR stimulation. These CD127(high) cells were better maintained than their CD127(low) counterparts following transfer into naive mice, consistent with their observed surface expression of CD127 and CD122, which confer the ability to self-renew in response to IL-7 and IL-15. CD127(high) cells were capable of IFN-gamma production upon peptide restimulation and expanded in response to challenge infection, indicating that these cells are functionally responsive upon Ag re-encounter. These results are in contrast to what is typically observed during many persistent infections and indicate that a stable population of parasite-specific CD8(+) T cells capable of Ag-independent survival is maintained in mice despite the presence of persistent Ag.     

Effector memory Th1 CD4 T cells are maintained in a mouse model of chronic malaria

Protection against malaria often decays in the absence of infection, suggesting that protective immunological memory depends on stimulation. Here we have used CD4(+) T cells from a transgenic mouse carrying a T cell receptor specific for a malaria protein, Merozoite Surface Protein-1, to investigate memory in a Plasmodium chabaudi infection. CD4(+) memory T cells (CD44(hi)IL-7Rα(+)) developed during the chronic infection, and were readily distinguishable from effector (CD62L(lo)IL-7Rα(-)) cells in acute infection. On the basis of cell surface phenotype, we classified memory CD4(+) T cells into three subsets: central memory, and early and late effector memory cells, and found that early effector memory cells (CD62L(lo)CD27(+)) dominated the chronic infection. We demonstrate a linear pathway of differentiation from central memory to early and then late effector memory cells. In adoptive transfer, CD44(hi) memory cells from chronically infected mice were more effective at delaying and reducing parasitemia and pathology than memory cells from drug-treated mice without chronic infection, and contained a greater proportion of effector cells producing IFN-γ and TNFα, which may have contributed to the enhanced protection. These findings may explain the observation that in humans with chronic malaria, activated effector memory cells are best maintained in conditions of repeated exposure.     

Unified immune modulation by 4-1BB triggering leads to diverse effects on disease progression in vivo

4-1BB (CD137) is a powerful T-cell costimulatory molecule in the treatment of virus infections and tumors, but recent studies have also uncovered regulatory functions of 4-1BB signaling. Since 4-1BB triggering suppresses autoimmunity by accumulating indoleamine 2,3-dioxygenase (IDO) in dendritic cells (DCs) in an interferon (IFN)-γ-dependent manner, we asked whether similar molecular and cellular changes were induced by 4-1BB triggering in virus-infected mice. 4-1BB triggering increased IFN-γ and IDO, and suppressed CD4(+) T cells, in C57BL/6 mice infected with the type 1 KOS strain of Herpes simplex virus (HSV-1), as it does in an autoimmune disease model. Detailed analysis of the CD4(+) T suppression showed that freshly activated CD62L(high) T cells underwent apoptosis in the early phase of suppression, and CD62L(low) effector/memory T cells in the later phase. Although 4-1BB triggering resulted in similar cellular changes - increased CD8(+) T and decreased CD4(+) T cells, it had different effects on mortality in mice infected with HSV-1 RE, influenza, and Japanese encephalitis virus (JEV); it increased mortality in influenza-infected mice but decreased it in JEV-infected mice. Since the dominant type of immune cell generated to protect the host was different for each virus - CD4(+) T cells and neutrophils in HSV-1 RE infection, both CD4(+) T and CD8(+) T cells in influenza infection, and a crucial role for B cells in JEV infection, 4-1BB triggering resulted in different therapeutic outcomes. We conclude that the therapeutic outcome of 4-1BB triggering is determined by whether the protective immunity generated against the virus was beneficially altered by the 4-1BB triggering.     

CD94/NKG2A expression is associated with proliferative potential of CD8 T cells during persistent polyoma virus infection

Memory CD8 T cells comprise a critical component of durable immunity because of their capacity to rapidly proliferate and exert effector activity upon Ag rechallenge. During persistent viral infection, memory CD8 T cells repetitively encounter viral Ag and must maintain a delicate balance between limiting viral replication and minimizing immunopathology. In mice infected by polyoma virus, a natural mouse pathogen that establishes long-term persistent infection, the majority of persistence-phase antiviral CD8 T cells express the inhibitory NK cell receptor CD94/NKG2A. In this study, we asked whether CD94/NKG2A expression is associated with Ag-specific recall of polyoma virus-specific CD8 T cells. During the persistent phase of infection, polyoma virus-specific CD8 T cells that express CD94/NKG2A were found to preferentially proliferate; this proliferation was dependent on cognate Ag both in vitro and in vivo. In addition, CD94/NKG2A(+) polyoma-specific CD8 T cells have a markedly enhanced capacity to produce IL-2 upon ex vivo Ag stimulation compared with CD94/NKG2A(-) polyoma-specific CD8 T cells. Importantly, CD94/NKG2A(+) anti-polyoma virus CD8 T cells appear to be essential for Ag-specific recall responses in mice persistently infected by polyoma virus. Because of its higher proliferative potential and capacity to produce IL-2, we propose that the CD94/NKG2A(+) subpopulation represents a less differentiated state than the CD94/NKG2A(-) subpopulation. Identification of proliferation-competent subpopulations of memory CD8 T cells should prove valuable in designing therapeutic vaccination strategies for persistent viral infections.