Homeostatic regulation of CD8+ T cells by perforin.

To prevent uncontrolled expansion, the massive proliferation of T cells during an acute immune response has to be followed by controlled deletion. Here we show that similar to Fas, perforin is not only an important effector molecule of cytotoxic T lymphocytes (CTL) but also involved in down-regulating peripheral T cells. Mice deficient for both the CTL effector molecule perforin and the apoptosis-inducing Fas ligand spontaneously develop infiltration of highly activated CD8(+) T cells in kidney and liver and die between 5 and 12 weeks of age. Injection of staphylococcal enterotoxin B (SEB) into perforin-deficient mice results in dramatically increased selective expansion and prolonged persistence of CD8(+), but not CD4(+), SEB-reactive T cells. Also, secondary immunization of TCR transgenic perforin-deficient mice with the lymphocytic choriomeningitis virus glycoprotein-derived epitope peptide leads to an increased proliferation of transgenic CD8(+) T cells, that is not explained by failure to deplete professional antigen-presenting cells. These results point to a novel mechanism of T cell homeostasis in which the acquisition of perforin-dependent cytotoxic activity regulates the expansion and persistence of CD8(+) effector T cells in vivo.     

Persistence of viral infection despite similar killing efficacy of antiviral CD8 T cells during acute and chronic phases of infection

Why some viruses establish chronic infections while others do not is poorly understood. One possibility is that the host's immune response is impaired during chronic infections and is unable to clear the virus from the host. In this report, we use a recently proposed framework to estimate the per capita killing efficacy of CD8⁺ T cells, specific for the polyoma virus (PyV), which establishes a chronic infection in mice. Surprisingly, the estimated per cell killing efficacy of PyV-specific effector CD8⁺ T cells during the acute phase of the infection was very similar to the efficacy of effector CD8⁺ T cells specific to lymphocytic choriomeningitis virus (LCMV-Armstrong), which is cleared from the host. Our results suggest that persistence of PyV does not result from the generation of an inefficient PyV-specific CD8⁺ T cell response, and that other host or viral factors are responsible for the ability of PyV to establish chronic infection.     

Differential requirement of perforin and IFN-gamma in CD8 T cell-mediated immune responses against B16.F10 melanoma cells expressing a viral antigen

Perforin-mediated lysis and secretion of IFN-gamma belong to the key effector functions of CD8 T cells. To compare the anti-tumor activity of these two mechanisms, we used B16.F10 melanoma cells (B16GP33) expressing the cytotoxic T cell epitope GP33 and T cell receptor transgenic (TCR-tg) mice specific for GP33 and deficient in perforin or IFN-gamma. B16GP33 tumor cells, injected either i.v. to induce experimental metastases or s.c., were similarly controlled in both wild-type and perforin-deficient, but not in IFN-gamma-deficient TCR-tg mice. A similar result was obtained when the therapeutic efficacy of adoptively transferred TCR-tg effector cells from these mice was examined in the corresponding perforin- or IFN-gamma-deficient C57BL/6 hosts. Criss-cross experiments further revealed that IFN-gamma production by the host strongly influenced the efficiency of the adoptively transferred effector cells. In contrast to the potent ability of GP33-specific effector cells to mediate B16GP33 tumor regression without perforin, GP33-specific memory cells, induced with recombinant vaccinia virus expressing GP33, failed to control B16GP33 tumor growth in the absence of perforin. In conclusion, our data demonstrate a crucial role for IFN-gamma in B16GP33 tumor cell elimination in vivo and indicate a differential requirement of perforin by effector versus memory CD8 T cells in anti-tumor immunity.     

Perforin and IFN-gamma do not significantly regulate the virus-specific CD8+ T cell response in the absence of antiviral effector activity

Using gene-targeted mice we have investigated whether perforin and/or interferon-gamma exert a direct regulatory effect on the expansion and contraction of antigen-specific CD8(+) T cells following infection with a virus (vesicular stomatitis virus) which is not controlled through these molecular effector systems. Unlike what has been observed when these molecules are essential for pathogen clearance, neither molecule was found to play an important role in regulating the kinetics of the virus-specific CD8(+) T cell response in the absence of antiviral effector activity.     

Study of diverse mechanisms of cell-mediated cytotoxicity in gene-targeted mice using flow cytometric cytotoxicity assay

Cytotoxic lymphocytes kill tumor or virus-infected target cells utilizing two mechanisms-(1) release of lytic granules (containing perforin and granzymes), and (2) Fas ligand (FasL)/Fas or TNF-initiated apoptosis. We have examined mechanisms of target cell lysis by effector T cells from gene-targeted and mutant mice using a new Flow Cytometric Cytotoxicity Assay (FC Assay). Target cells were labeled with PKH67 dye. Cell death was estimated by 7-AAD inclusion and annexin V-PE binding. A direct correlation has been found between the percentage of dead target cells in FC Assay and the results of 111In release cytotoxicity assay when effector T cells from either Pfp -/- (perforin knockout) or gld (non-functional Fas Ligand) mice were used. As shown by the 4 h FC assay, the granule-mediated mechanism was utilized by T cells from gld mice. In contrast, T cells from Pfp -/- mice used death receptor-mediated lysis. Therefore, cytotoxic cells from gene-targeted and mutant mice can serve as valuable tools for studying different mechanisms of cell-mediated cytotoxicity, and the FC assay could be applied irrespective of which cytotoxic effector pathway is involved.     

Nasal T/NK cell lymphomas commonly express perforin and Fas ligand: important mediators of tissue damage

 

Aims:

Two molecular mechanisms of T/natural killer (NK) cell-mediated cytotoxicity, one perforin based and the other Fas based, have been demonstrated, and both systems induce cytotoxicity in the target cells. The Fas-based mechanism involves the transducing molecule Fas and its ligand (FasL). In addition, perforin and/or FasL are also expressed in the cytotoxic T/NK cells. This study was thus designed to investigate the Fas and perforin pathways of the cytotoxic T/NK lymphoma cells in the nasal cavity.  

Methods and results:

Eight patients with nasal lymphoma were analysed using immunohistochemical staining methods. Two cases were CD3+ CD56+ (T/NK cell) type, and six were CD3− CD56+ (NK cell) type. All cases showed Epstein–Barr virus genomes by in-situ hybridization. In addition, all cases showed the expression of TIA-1 (GMP-17), which is a marker of cytotoxic T and NK cells. FasL was expressed in the majority of the lymphoma cells and some histiocytes, while Fas was found in lymphoma cells and many non-neoplastic cells. In addition, the expression of perforin was detected in almost all lymphoma cells. In the double stainings, lymphoma cells expressed both FasL and perforin. Based on these findings, both the perforin- and Fas-based pathway of the cytotoxic T/NK lymphoma cells are thus considered to play an important role in the clinical features.  

Conclusions:

Tissue damage is a common morphological feature in nasal T/NK cell lymphoma. The above findings therefore support the theory that tissue damage is due to both the cytotoxicity of T/NK lymphoma cells as well as to angiocentricity.     

Genetic deletion of granzyme B does not confer resistance to the development of spontaneous diabetes in non‐obese diabetic mice

Granzyme B (GzmB) and perforin are proteins, secreted mainly by natural killer cells and cytotoxic T lymphocytes that are largely responsible for the induction of apoptosis in target cells. Because type 1 diabetes results from the selective destruction of β cells and perforin deficiency effectively reduces diabetes in non‐obese diabetic (NOD) mice, it can be deduced that β cell apoptosis involves the GzmB/perforin pathway. However, the relevance of GzmB remains totally unknown in non‐obese diabetic (NOD) mice. In this study we have focused on GzmB and examined the consequence of GzmB deficiency in NOD mice. We found that NOD.GzmB^–/– mice developed diabetes spontaneously with kinetics similar to those of wild‐type NOD (wt‐NOD) mice. Adoptive transfer study with regulatory T cell (T_reg)‐depleted splenocytes (SPCs) into NOD‐SCID mice or in‐vivo T_reg depletion by anti‐CD25 antibody at 4 weeks of age comparably induced the rapid progression of diabetes in the NOD.GzmB^–/– mice and wt‐NOD mice. Expression of GzmA and Fas was enhanced in the islets from pre‐diabetic NOD.GzmB^–/– mice. In contrast to spontaneous diabetes, GzmB deficiency suppressed the development of cyclophosphamide‐promoted diabetes in male NOD mice. Cyclophosphamide treatment led to a significantly lower percentage of apoptotic CD4⁺, CD8⁺ and CD4⁺CD25⁺ T cells in SPCs from NOD.GzmB^–/– mice than those from wt‐NOD mice. In conclusion, GzmB, in contrast to perforin, is not essentially involved in the effector mechanisms for β cell destruction in NOD mice.     

Perforin, a cytotoxic molecule which mediates cell necrosis, is not required for the early control of mycobacterial infection in mice.

Host defense against mycobacterial infection requires the participation of monocytes and T cells. Both CD4+ and CD8+ T cells have been shown to be important in resistance to mycobacterial infection in vivo. The main contribution of CD4+ T cells to the protective antituberculosis response involves the production of Th1-type cytokines, including interleukin-2 (IL-2) and gamma interferon (IFN-gamma). CD8+ T cells have been considered to be responsible primarily for cytotoxicity mediated by toxic molecules, including perforin. CD8+ T cells may also elaborate Th1-type cytokines, such as IFN-gamma, in response to the infection. To elucidate the contribution of perforin-mediated target cell death to the control of mycobacterial infection in vivo, mice with a disruption in the perforin gene (P-/-) were infected with Mycobacterium bovis BCG or M. tuberculosis Erdman for 5 and 13 weeks, respectively. At 1, 3, 5, and 13 weeks postinfection, the number of viable mycobacteria in the lungs, spleens, and livers of mice were determined by CFU assay. The infected tissues were examined histologically, and cytokine mRNA levels in the spleens of these mice were determined. Similar studies were carried out in Fas receptor-defective (CBA/lpr(cg)) mice to evaluate the contribution of this alternative cytotoxic pathway to the control of mycobacterial infection. The absence of either perforin gene function or Fas receptor gene function did not modify the course of experimental mycobacterial infection in these mice. In addition, both P-/- and Fas receptor-defective mice appeared to have a compensatory activation of cytokine genes, even in the absence of the experimental infection. P-/- mice had a mean 3.4- to 5-fold increase in mRNA levels for IL-10, IL-12p35, IL-6, and IFN-gamma. Similarly, Fas receptor-defective mice had a mean 3- to 3.6-fold increase in mRNA levels for IFN-gamma, IL-12p35, and IL-10. Our results indicate that both perforin-mediated cytotoxicity and Fas-mediated cytotoxicity do not appear to be necessary for the early control of mycobacterial infection in vivo.     

Preferential utilization of the perforin/granzyme pathway for lysis of Epstein-Barr virus-transformed lymphoblastoid cells by virus-specific CD4+ T cells

In this report, we show that Epstein-Barr virus (EBV)-infected lymphoblastoid cell lines (LCL) express Fas and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptor 2 and that LCL are lysed following engagement of these receptors by agonist Fas and TRAIL receptor-specific monoclonal antibodies (MAb). We also show that EBV-specific CD4+ T cells mediate bystander lysis of susceptible targets through both the Fas/Fas ligand (FasL) and the TRAIL pathways, but find that the dominant mechanism of lysis following cognate, HLA class II-restricted recognition of LCL is the perforin/granzyme pathway. Killing of LCL by EBV-specific CD4+ T cells was strongly inhibited by concanamycin A, an agent that elevates granule pH, resulting in accelerated destabilization and degradation of perforin. In contrast, blocking anti-FasL MAb showed only limited inhibition of LCL killing. Blocking anti-TRAIL MAb had no effect on lysis of LCL by EBV-specific CD4+ T cells. We further show that culture of EBV-specific CD4+ T cells in the presence of interleukin 4 markedly abrogates effector cytotoxic function against LCL through direct depletion of intracellular perforin, with no evidence of a Th1 to Th2 shift in patterns of cytokine expression.     

Homeostatic regulation of CD8+ T cells after antigen challenge in the absence of Fas (CD95)

The role of Fas in the homeostatic regulation of CD8⁺ T cells after antigen challenge was analyzed in the murine model of lymphocytic choriomeningitis virus (LCMV) infection. Mice homozygous for the lpr mutation and carrying T cell receptor (TCR) αβ transgenes specific for the LCMV glycoprotein peptide aa 33–41 in the context of H-2D^b were used. Five main results emerged: first, development of lymphadenopathy and of CD4⁻CD8⁻ double-negative B220⁺ T cells in lpr mice was not inhibited by the αβ TCR transgenes; second, tolerance induction and peripheral deletion of CD8⁺ T cells induced by LCMV glycoprotein peptide injection was independent of Fas expression; third, clonal down-regulation of Fas-deficient TCR-transgenic CD8⁺ T cells after acute LCM virus infection was identical to the decline of transgenic T cells expressing Fas; fourth, in vivo activated CD8⁺ effector T cells from TCR transgenic and TCR-lpr/lpr mice were equally susceptible to activation-induced cell death in vitro; and fifth, transgenic effector T cells from lpr/lpr mice were cleared in the declining phase of the immune response in vivo without giving rise to CD4⁻CD8⁻ double-negative T cells. Taken together, these data suggest that the homeostatic regulation of CD8⁺ T cells after antigen challenge in vivo is regulated by mechanisms that do not require Fas.     

Strain-dependent requirement for IFN-γ for respiratory control and immunotherapy in murine gammaherpesvirus infection

Interferon-γ (IFN-γ) and perforin (pfp) are important effector mechanisms used by CD8 T cells to clear virus-infected cells. In this study, we used IFN-γ/pfp double knockout mice to address if these two effector molecules play redundant roles in the control of acute infection with murine gammaherpesvirus-68 (MHV-68) in BALB/C mice. Perforin knockout (KO) mice and wild-type mice cleared infectious virus from the lungs, even following high-dose infection. However, the IFN-γ KO and IFN-γ/pfp double knockout (DKO) groups had higher virus titers in the lungs at day 10 post-infection, and both groups had higher mortality rates. In IFN-γ/pfp DKO mice, the virus titer and mortality rate were significant higher than in IFN-γ KO mice, indicating a role for perforin in protection from disease. WT mice given IFN-γ blocking antibody also showed significantly higher viral titers. In contrast, IFN-γ KO mice on a C57BL/6 background controlled respiratory infection comparably to wild-type mice. These data show that perforin plays a redundant role in the control of virus replication, but IFN-γ plays an essential role in BALB/C mice infected with MHV-68. We conclude that there is a marked strain-dependent difference in the effector mechanisms needed to control acute MHV-68 infection between C57BL/6 and BALB/C mice. In addition we show that immune therapy that re-establishes viral control after spontaneous reactivation in CD4-deficient mice depends upon perforin in C57BL/6 mice but IFN-γ in BALB/C mice.     

Tumor necrosis factor alpha production from CD8+ T cells mediates oviduct pathological sequelae following primary genital Chlamydia muridarum infection

The immunopathogenesis of Chlamydia trachomatis-induced oviduct pathological sequelae is not well understood. Mice genetically deficient in perforin (perforin(-/-) mice) or tumor necrosis factor alpha (TNF-α) production (TNF-α(-/-) mice) displayed comparable vaginal chlamydial clearance rates but significantly reduced oviduct pathology (hydrosalpinx) compared to that of wild-type mice. Since both perforin and TNF-α are effector mechanisms of CD8(+) T cells, we evaluated the role of CD8(+) T cells during genital Chlamydia muridarum infection and oviduct sequelae. Following vaginal chlamydial challenge, (i) mice deficient in TAP I (and therefore the major histocompatibility complex [MHC] I pathway and CD8(+) T cells), (ii) wild-type mice depleted of CD8(+) T cells, and (iii) mice genetically deficient in CD8 (CD8(-/-) mice) all displayed similar levels of vaginal chlamydial clearance but significantly reduced hydrosalpinx, compared to those of wild-type C57BL/6 mice, suggesting a role for CD8(+) T cells in chlamydial pathogenesis. Repletion of CD8(-/-) mice with wild-type or perforin(-/-), but not TNF-α(-/-), CD8(+) T cells at the time of challenge restored hydrosalpinx to levels observed in wild-type C57BL/6 mice, suggesting that TNF-α production from CD8(+) T cells is important for pathogenesis. Additionally, repletion of TNF-α(-/-) mice with TNF-α(+/+) CD8(+) T cells significantly enhanced the incidence of hydrosalpinx and oviduct dilatation compared to those of TNF-α(-/-) mice but not to the levels found in wild-type mice, suggesting that TNF-α production from CD8(+) T cells and non-CD8(+) cells cooperates to induce optimal oviduct pathology following genital chlamydial infection. These results provide compelling new evidence supporting the contribution of CD8(+) T cells and TNF-α production to Chlamydia-induced reproductive tract sequelae.     

Perforin and gamma interferon are critical CD8+ T-cell-mediated responses in vaccine-induced immunity against Leishmania amazonensis infection

Previous studies have demonstrated that protection against New World leishmaniasis caused by Leishmania amazonensis can be elicited by immunization with the developmentally regulated Leishmania amastigote antigen, P-8. In this study, several independent experimental approaches were employed to investigate the protective immunological mechanisms involved. T-cell subset depletion experiments clearly indicate that elicitation of CD8(+) (as well as CD4(+)) effector responses is required for protection. Further, mice lacking beta(2)-microglobulin (and hence deficient in major histocompatibility complex class I antigen presentation) were not able to control a challenge infection after vaccination, indicating an essential protective role for CD8(+) T effector responses. Analysis of the events ongoing at the cutaneous site of infection indicated a changing cellular dynamic involved in protection. Early postinfection in protectively vaccinated mice, a predominance of CD8(+) T cells, secreting gamma interferon (IFN-gamma) and expressing perforin, was observed at the site of infection; subsequently, activated CD4(+) T cells producing IFN-gamma were primarily found. As protection correlated with the ratio of total IFN-gamma-producing cells (CD4(+) and CD8(+) T cells) to macrophages found at the site of infection, a role for IFN-gamma was evident; in addition, vaccination of IFN-gamma-deficient mice failed to provide protection. To further assess the effector mechanisms that mediate protection, mice deficient in perforin synthesis were examined. Perforin-deficient mice vaccinated with the P-8 antigen were unable to control infection. Thus, the elicitation of CD8(+) T cell effector mechanisms (perforin, IFN-gamma) are clearly required in the protective immune response against L. amazonensis infection in vaccinated mice.     

Virus clearance and immunopathology by CD8(+) T cells during infection with respiratory syncytial virus are mediated by IFN-gamma

CD8(+) T cells (CTL) are important effector cells for virus control and immunopathology after primary infection with respiratory syncytial virus (RSV). To investigate the effector mechanisms involved, we set up an adoptive transfer model, in which effector CTL specific for p82-90 of RSV M2 were generated in vivo, followed by short-term restimulation in vitro and transfusion into infected recipients. A total of 4 x 10(4) donor-derived p82-specific CTL homing to the lung within 4 days after transfusion were sufficient to completely eliminate a virusinoculum of 1.5 x 10(6) pfu. This was accompanied by significant lung pathology. Surprisingly, virus control and immunopathology proceeded unimpaired when donor cells lacking perforin, CD95 ligand or TNF were transfused. By contrast, treatment of recipient mice with a neutralizing antibody against IFN-gamma or transfusion of IFN-gamma-deficient effector CTL largely abolished virus control and significantly reduced CD8(+) T cell-mediated pathology. In IFN-gamma-deficient mice, high-dose primary infection experiments revealed attenuated immunopathology, but only slightly delayedvirus clearance, suggesting that other cells and molecules can partly substitute for the effects of CTL-derived IFN-gamma on virus clearance. These experiments identify IFN-gamma as a key molecule in RSV-induced immunopathology and in CD8(+) T cell-mediated control of RSV infection.     

Progression of spontaneous autoimmune diabetes is associated with a switch in the killing mechanism used by autoreactive CTL

Autoimmune (type 1) diabetes mellitus results from the destruction of insulin-producing pancreatic beta-cells by T lymphocytes. Beta-cell death that is induced by autoreactive CTL in diabetes involves both Fas/Fas ligand (FasL)- and perforin/granzyme-mediated pathways, although their relative contributions during the progression of the disease remain unknown. We demonstrate here that despite the preferential use of the Fas/FasL pathway for cytolysis of beta-cell targets, transgenic beta-cell-specific CTL were able to kill targets via the perforin pathway when triggered by a higher affinity stimulus. In addition, we show that the killing mechanism used by islet-associated CD8(+) T cells from non-obese diabetic mice changed as the mice aged and correspondingly, with the stage of diabetes. These results provide direct evidence for age-related changes in the cytotoxic pathways used by diabetogenic T cells during the progression of autoimmune diabetes.     

Antigen-induced cell death of T effector cells in vitro proceeds via the Fas pathway,requires endogenous interferon-gamma and is independent of perforin and granzymes

Activation of resting T cells usually leads to their proliferation and differentiation into effector cells and a subsequent decline following elimination of the antigen. A situation of excessive antigen density may result in T cell receptor (TCR)-induced deletion of T effector cells, a process termed antigen-induced cell death (AgICD). Previous studies indicate that AgICD of cytotoxic T cells may be induced by either of the two key cytotoxic processes, granule exocytosis, including perforin and granzymes, or the Fas ligand (FasL)/Fas pathway. By using in vitro-polyclonally activated or ex vivo-derived virus-induced T cell populations from mice with mutations or targeted gene defects in one or more components of the two key cytolytic pathways we now show that TCR-induced apoptosis is only impaired in the absence of FasL and/or Fas, but not in the absence of perforin and/or granzymes. Furthermore, antibody-blockage of FasL alone is sufficient to inhibit early T cell death. Inhibition of both, FasL and tumor necrosis factor (TNF-alpha) is required to abrogate late apoptosis by AgICD. The fact that antibodies to IFN-gamma also inhibit AgICD suggests that the perforin plus granzyme-independent and FaSL and/or TNF-alpha facilitated process of AgICD of T effector cells is tightly regulated by endogenous IFN-gamma.     

Local immune responses to influenza virus infection in mice with a targeted disruption of perforin gene

The role of perforin in the local defense mechanisms against influenza virus infection was investigated. Mice deficient in the perforin gene (perforin(-/-)) were more susceptible to influenza virus infection than the ordinary wild-type C57BL/6 mice, showing an increased mortality with elevated virus growth and prolonged virus shedding. The lung parenchyma cells of perforin(-/-) mice showed no cytolytic activities of natural killer cells or virus-specific cytotoxic T lymphocytes in vitro, although the local antibody production system in the respiratory tract functioned well. In perforin(-/-) mice, the appearance of apoptotic degeneration in virally infected lung cells was delayed. This might cause cellular infiltration, including CD4, CD8, and CD19 positive cells, in the lung, peaking at day 8 after infection and maintaining a high level for a longer period. In the virus-induced local cytokine production, interferon-gamma (IFN-gamma) was prominent. The adoptive transfer of immune-competent spleen cells from wild-type C57BL/6 mice achieved a complete protection of perforin(-/-) mice against lethal challenge infection. These results suggest that perforin plays a crucial role in the host defense system against influenza virus infection, especially in its early stage, by inducing apoptosis of virus-infected cells.     

Human umbilical cord blood NK T cells kill tumors by multiple cytotoxic mechanisms

Natural killer (NK) T cells are restricted by CD1d and play an important role in the rejection of malignant tumors, but how kill these tumors is unclear. To investigate this, we cultured Valpha24+CD4+ NK T cells in human umbilical cord blood, which was enriched by immunomagnetic beads. In short-term (4 h) cytotoxicity assays, the NK T cells could kill only those targets expressing CD1d. In longer cytotoxicity assays (20 h), however, the NK T cells were able to kill all the tumors, regardless of CD1d expression. When each of the perforin, Fas-FasL, and TNF-alpha cytotoxic mechanisms were blocked, it was apparent that perforin killing dominated in both the short- and long-term assays. In the short-term assay, perforin killing required that the targets expressed CD1d, but killing was more efficient if Fas was present because then the Fas-FasL mechanism was also used. Thus, cells that lacked Fas and CD1d and were not killed in the 4-h assay, were instead lysed in 20-h assay through a combination of perforin and TNF-alpha killing. NK T cells can kill tumor targets by perforin, Fas-FasL, and TNF-alpha mechanisms. TNF-alpha killing requires longer contact between effectors and targets, suggesting that TNF-alpha acts by enhancing perforin killing.     

Properties of murine (CD8+)CD27- T cells

In humans, loss of CD27 expression is associated with the stable acquisition of effector functions by CD8+ T cells. We found that murine (CD8+)CD27- T cells were confined to the primed CD62L(dull/-)CD44(bright)CCR7- T cell population. (CD8+)CD27- T cells were absent from lymph nodes but could be found in blood, spleen and in non-lymphoid organs such as lung and liver. Late after primary influenza virus infection, low percentages of antigen-specific CD27- cells emerged in the lung and spleen. After recovery from secondary influenza virus infection, high percentages of influenza-specific CD27- T cells were found in the lung and the loss of CD27 on lung CD8+ T cells coincided with high granzyme B expression. After murine cytomegalovirus infection, loss of CD27 expression on virus-specific CD8+ T cell populations was sustained and especially marked in liver and lung. We suggest that in mice, CD27 is lost from CD8+ T cells only after repetitive antigenic stimulation. Moreover, the high expression of both granzyme B and perforin in the CD27- T cells suggests that the lack of CD27 on murine CD8+ T cells can be used to identify memory T cells with expression of cytotoxic effector molecules.