BH3-only protein Puma contributes to death of antigen-specific T cells during shutdown of an immune response to acute viral infection

Abstract

During acute T cell immune responses to viral infection, antigen-specific T cells first proliferate and differentiate into effector cells, but after pathogen clearance most are deleted by apoptosis. The developmentally programmed death of antigen-specific T cells during shutdown of a T cell response is mediated by the Bcl-2-regulated apoptotic pathway and partly depends on the proapoptotic BH3-only protein Bim. However, loss of Bim enhanced survival of antigen-activated T cells to a lesser extent than Bcl-2 overexpression, indicating that other proapoptotic factors must contribute to T cell killing. In this study, we investigated the contributions of several BH3-only proteins to the shutdown of an acute T cell immune response in vivo. After infection with human herpes simplex virus (HSV-1), mice lacking Noxa, Bid, or Bad had a normal increase and subsequent decline in the numbers of antigen-specific CD8⁺ T cells. In contrast, Puma-deficient mice showed an abnormally prolonged persistence of antigen-specific CD8⁺ T cells in the spleen, associated with enhanced in vitro survival of these cells in the absence of cytokines. Puma was dispensable for viral clearance and also did not play a role in proliferation or activation of HSV-1-specific CD8⁺ T cells in vivo. Collectively, these findings show that Puma contributes to the death of antigen-specific T cells during shutdown of an immune response.

T cell immune responses to foreign antigens involve the generation of effector as well as memory cells through a process of activation, proliferation, and differentiation (1). In the case of acute infections, pathogen clearance (with concomitant loss of antigenic stimulation and inflammation) is followed by apoptosis of the majority of activated, antigen-specific T lymphocytes, and this process of cell killing appears to be critical to the maintenance of T cell homeostasis and the prevention of immunopathology (1, 2).

Analysis of the molecular mechanisms involved in T cell death is necessary for our understanding of how immune responses develop and how autoimmune tissue destruction is prevented (3). Mammals have two distinct, albeit ultimately converging, apoptotic pathways: “death receptor signaling” and “Bcl-2 regulated” (also called “intrinsic” or “mitochondrial”) apoptosis signaling (4). Based on the observation that FasL-Fas signaling mediates apoptosis of activated T cells elicited by TCR restimulation in vitro (5), it has been speculated that this pathway is critical for the shutdown of T cell immune responses in vivo. Although some early studies provided support for this model (6), it has been shown that Fas and TNF-R1 death receptor signaling are dispensable for killing of antigen-activated T cells during shutdown of an immune response to a single injection with staphylococcal enterotoxin B (SEB) (7) or acute viral (HSV-1) infection (8). The currently widely supported model postulates that T cell death during termination of acute immune responses is triggered by a decline in the levels of cytokines and mediated by the Bcl-2-regulated apoptotic pathway (2), initiated by the proapoptotic BH3-only subgroup of the Bcl-2 protein family (9). Eight BH3-only proteins have been found in mammals: Bim, Puma, Noxa, Bmf, Bad, Bid, Bik, and Hrk. They all share with each other and the Bcl-2 family at large only the short BH3 region and trigger apoptosis when overexpressed (9, 10). Individual BH3-only proteins vary in their binding to their prosurvival Bcl-2-like relatives (11, 12), but all require Bax/Bak-like proteins, members of the so-called multi-BH domain proapoptotic subgroup of the Bcl-2 family, for cell killing (13, 14). Bim (15) and Puma (16, 17) have particularly prominent roles in apoptosis initiation because, in contrast to all other BH3-only proteins, they bind with high affinity to all prosurvival Bcl-2-like proteins (11, 12).

Expression analysis and studies with gene-targeted mice have identified cell-type- and death-stimulus-specific roles for individual BH3-only proteins. Bim is expressed in many tissues, including lymphoid and myeloid cells (18), and is essential for their normal homeostasis (19), deletion of autoreactive thymocytes (20) and B cells (21), and death of antigen-activated T cells during shutdown of an immune response (7, 8). Puma is a direct transcriptional target of the tumor suppressor p53 (16, 17) and is required for p53-mediated apoptosis, such as that induced by γ-irradiation (22, 23). Puma is also essential for apoptosis induced by a range of p53-independent stimuli, such as cytokine deprivation (22, 23). Compared with Puma, the other p53-regulated BH3-only protein Noxa has a more restricted role in apoptosis (24), mainly being responsible for UV-radiation-induced cell killing (25). Cytokine receptor stimulation was shown to inhibit the proapoptotic activity of Bad by causing its phosphorylation and sequestration to 14-3-3 scaffold proteins (26, 27). However, most cells from bad^−/− mice died at a normal rate after cytokine withdrawal (28), indicating that Bad plays only a minor role in this pathway. The BH3-only protein Bid has a cell-type-restricted role in linking death-receptor-induced apoptosis to the Bcl-2-regulated pathway, because hepatocytes from bid^−/− mice are protected from Fas-induced apoptosis, although lymphocytes are normally sensitive (29, 30).

We have previously shown that Bim plays a critical role in the killing of antigen-specific CD8⁺ T cells during shutdown of an acute response to HSV-1 (8). However, loss of Bim afforded less protection than Bcl-2 overexpression, indicating that additional BH3-only proteins may also contribute. Here, we investigated the contribution of several BH3-only proteins to T cell death during the shutdown of an acute T cell immune response to HSV-1 infection. Using mice that lack different BH3-only proteins, we confirm that Bim plays the dominant role and show that Puma contributes significantly to apoptosis of activated T cells.

Results

Loss of Puma Causes Accumulation of HSV-1-Specific CD8⁺ T Cells in the Spleen.

Because Bcl-2 overexpression inhibited deletion of antigen-activated T cells during shutdown of an immune response to an acute infection or single injection of SEB to a greater extent than loss of Bim (7, 8, 31), it appears likely that additional BH3-only proteins contribute to T cell killing. To identify the BH3-only proteins involved in this process, we infected mice deficient for Puma, Noxa, Bid, or Bad, and as controls WT as well as bim^−/− mice, with HSV-1 in both hind feet. After 1 week, there was no significant alteration of total leukocyte and CD8⁺ T cell numbers in the spleens of infected mice compared with naïve mice for any of the genotypes (Fig. 1 A and B). Consistent with previous studies (8, 19), naïve as well as HSV-1-infected bim^−/− mice showed 2- to 5-fold increased numbers of total leukocytes and CD8⁺ T cell in their spleens (Fig. 1 A and B). The total numbers of CD4⁺ T cells and B220⁺ cells did not change over the entire observation period in mice of any of the genotypes, with the exception that bim^−/− animals, as previously reported (8, 19), contained 2- to 4-fold increased numbers of these cell types at all time points. Although loss of Puma does not cause any obvious defects in lymphocyte development (32), cell counting revealed a significant increase in total leukocyte and CD8⁺ T cell numbers in puma^−/− mice compared with WT mice after 2 and 4 weeks of infection with HSV-1 (Fig. 1 A and B). This led us to examine the role of Puma in the T cell immune response in more detail. Kinetics of expansion and contraction of the pool of HSV-1-specific CD8⁺ T cells were monitored by staining with K^b-gB tetramers containing the immunodominant HSV gB_498–505 peptide or by stimulating cells in vitro with gB_498–505 HSV peptide followed by intracellular staining for IFN-γ [supporting information (SI) Fig. 7]. Strikingly, in the spleen we found that compared with WT controls, puma^−/− mice had 2- to 4-fold increased numbers of K^b-gB⁺ CD8⁺ cells after 2 and 4 weeks of infection (Fig. 1 C and D). The percentages of K^b-gB⁺ CD8⁺ cells within the CD8⁺ cell population were also significantly increased in puma^−/− mice (SI Fig. 8). Consistent with our previous studies (8), Bim-deficient mice had abnormally increased numbers of HSV-1-specific T cells in their spleens at all time points examined after infection (Fig. 1 C and D). HSV-1-activated CD8⁺ T cells were also found in the draining (popliteal) lymph nodes, but no significant differences in their numbers were evident between mice of the genotypes tested (SI Fig. 9). This is consistent with the previous observation that abnormally surviving HSV-1-activated CD8⁺ T cells accumulate in the spleen but not the lymph nodes of bim^−/− mice (8). Loss of any of the other BH3-only proteins tested—Noxa, Bid, and Bad—had no effect on the numbers of HSV-1-specific CD8⁺ T cells throughout the experimental period (Fig. 1 C and D). Collectively, these results show that Puma and Bim play major roles in the death of antigen-activated CD8⁺ T cells during shutdown of an immune response to an acute viral infection.

Fig. 1.

HSV-1-infected puma^−/− mice have abnormally increased numbers of leukocytes, CD8⁺ T cells, and HSV-1-specific CD8⁺ T cells in the spleen. (A and B) Wild-type and mice of the given genotypes were injected with HSV-1 in both hind feet. Total leukocyte numbers (A) and numbers of CD8⁺ T cells (B) in the spleen were quantified by immunofluorescent staining with surface marker-specific mAbs and cell counting at the specified time points after infection. (C and D) Total numbers of HSV-1-specific CD8⁺ T cells in spleen of these mice were quantified at the indicated time points after HSV-1 infection by using surface staining with K^b-gB_498–505 tetramers (C) or by staining for intracellular IFN-γ in cells that had been restimulated for 5 h with HSV gB_498–505 peptide in vitro (D). Data represent mean ± SD of three to eight mice of each genotype for each time point. *, P < 0.05.

Loss of Puma Enhances Survival of Antigen-Specific CD8⁺ T Cells Cultured in the Absence of Cytokines.

It is widely believed that T cell apoptosis during termination of an acute immune response is triggered by the drop in the levels of cytokines that promote T cell survival and proliferation (3, 33). Because Puma contributes to growth-factor-withdrawal-induced apoptosis of lymphoid and myeloid cells in culture (22, 23, 34, 35), we hypothesized that loss of Puma promoted the accumulation of HSV-1-specific CD8⁺ T cells in infected mice by rendering them resistant to cytokine deprivation. To examine this idea, we enriched HSV-1-specific CD8⁺ T cells from spleens of WT or puma^−/− mice 7 or 14 days after infection, cultured them in medium with or without IL-7, and measured their survival by flow-cytometric analysis after staining for CD8 plus either TCRVα2 (expressed by most HSV-1-activated CD8⁺ T cells; Fig. 2A) or K^b-gB⁺ (SI Fig. 10). Antigen-specific T cells from WT mice died rapidly in the absence of cytokines with only ≈40% surviving at 48 h and <20% after 72 h of culture (Fig. 2A). In contrast, puma^−/− cells survived considerably better with ≈60% survival after 48 h and ≈35% persisting at 72 h (Fig. 2A). As previously reported (8), HSV-1-specific CD8⁺ T cells from bim^−/− mice showed nearly complete resistance to cytokine-withdrawal-induced apoptosis, but those from mice lacking Noxa, Bid, or Bad died at a normal rate (Fig. 2B). These results demonstrate that both Puma and Bim play critical roles in cytokine-deprivation-induced apoptosis of antigen-activated CD8⁺ T cells.

Fig. 2.

HSV-1-specific CD8⁺ T cells from puma^−/− mice are abnormally resistant to cytokine deprivation in vitro. CD8⁺ T cells from HSV-1-infected WT, puma^−/−, or bim^−/− mice (A) or WT, noxa^−/− bid^−/−, or bad^−/− mice (B) were cultured in vitro in the presence (black lines) or absence (dashed lines) of IL-7. Viability of HSV-1-specific CD8⁺ T cells was determined by staining with anti-CD8 mAb, anti-TCRVα2 mAb, and FITC-annexin V after 48 and 72 h in culture. Live cells were considered as FITC-annexin V⁻. Data represent mean ± SD of viable cells from three to eight mice of each genotype. *, P < 0.005.

Puma-Deficient Mice Show Normal Viral Clearance and Contain Normal Numbers of HSV-1-Specific CD8⁺ T Cells in Nonlymphoid Organs.

One possible explanation for the elevated numbers of antigen-specific CD8⁺ T cells in puma^−/− mice could be abnormally protracted viral clearance. To exclude this, we performed viral plaque-forming assays on extracts from the hind feet of mice at the indicated time points after infection. Puma^−/− mice and mice deficient for any of the other BH3-only proteins examined cleared the HSV-1 virus nearly completely by day 7 after infection, and there was no virus detectable after 14 days (Fig. 3). Abnormal distribution of antigen-specific CD8⁺ T cells to nonlymphoid organs could also result in elevated numbers of antigen-specific CD8⁺ T cells in the spleen. However, at day 7 after infection, puma^−/− mice contained numbers of HSV-1-specific CD8⁺ T cells in their lungs and livers that were similar to those in WT animals (SI Fig. 11). Thus, abnormal HSV-1 persistence or abnormal antigen-activated CD8⁺ T cell trafficking are unlikely to explain the increased accumulation of these cells in the spleens of puma^−/− mice.

Fig. 3.

Loss of Puma does not affect HSV-1 clearance in infected mice. Mice were infected with HSV-1 and killed after 3, 7, 14, or 28 days. Feet from HSV-1-infected WT, puma^−/−, or bim^−/− mice (A) or WT, noxa^−/− bid^−/−, or bad^−/− mice (B) were homogenized, and extracts were serially diluted for viral plaque-forming assays. Data represent mean ± SD of three to eight mice of each genotype for each time point. PFU, plaque-forming units.

Loss of Puma Does Not Affect in Vivo Activation and Proliferation of HSV-1-Specific CD8⁺ T Cells.

Excessive proliferation could also be a cause for the abnormal accumulation of HSV-1-specific CD8⁺ T cells in the spleens of puma^−/− mice. Therefore, we analyzed the expression of the classic activation markers CD25 and CD44 in HSV-1-specific CD8⁺ T cells in the draining lymph nodes of WT and puma^−/− mice on day 5 (Fig. 4A) and day 7 (Fig. 4B) after infection. Antigen-activated CD8⁺ T cells from puma^−/− mice up-regulated expression of CD25 and CD44 early in the immune response and to an extent comparable to WT T cells (Fig. 4A). In HSV-1-specific CD8⁺ T cells from both WT and puma^−/− mice, expression of CD44 remained up-regulated whereas CD25 was down-regulated after 7 days of infection (Fig. 4B). HSV-1-specific CD8⁺ T cells from the spleens of these mice showed changes in expression of activation markers similar to those from the draining lymph nodes (SI Fig. 12).

Fig. 4.

HSV-1-specific CD8⁺ T cells from puma^−/− mice show normal expression of cell-surface activation markers. (A) Cells from popliteal (draining) lymph nodes of WT or puma^−/− mice that were either naïve or had been infected with HSV-1 5 (A) or 7 (B) days earlier were stained with antibodies to CD8 plus CD25 or CD44 and in the case of infected mice also with HSV gB_498–505 peptide tetramers. Histograms showing CD44 or CD25 staining intensity of total CD8⁺ T cells from naïve mice (gray area) are plotted together with profiles of CD8⁺HSV gB_498–505⁺ (gated) T cells from HSV-1-infected mice (black line). The FACS histograms shown are representative of several experiments with three to six mice of each genotype.

To examine whether the abnormal accumulation of antigen-specific CD8⁺ T cells in puma^−/− mice was associated with abnormally enhanced proliferation, we performed in vivo proliferation assays by pulsing infected mice for 20 h with BrdU, which is incorporated into dividing cells. Seven days after HSV-1 infection, the proportions of proliferating cells were comparable between puma^−/− (45–52%) and WT (41–50%) mice (Fig. 5). These results demonstrate that the increased accumulation of antigen-specific CD8⁺ T cells in puma^−/− mice is unlikely to be a consequence of abnormally increased activation or proliferation.

Fig. 5.

Loss of Puma does not affect proliferation of HSV-1-specific CD8⁺ T cells in infected mice. Seven days postinfection, WT or puma^−/− mice were injected and then fed for 20 h with BrdU, and spleens were then harvested. Cells were stained with anti-CD8 mAb plus K^b-gB_498–505 tetramers and sorted, followed by intracellular staining with anti-BrdU mAb to identify cells undergoing division. Histograms are representative of four to six mice of each genotype (Upper) Wild-type mice. (Lower) puma^−/− mice.

Loss of Puma Does Not Affect Cytotoxic T Cell-Mediated Target Cell Killing in Vivo.

To examine the impact of Puma deficiency on cytotoxic killing of target cells in vivo, WT and puma^−/− mice were injected at 1, 2, or 4 weeks after HSV-1 infection with a mixture of WT spleen cells that were either loaded with a HSV-1-specific or a control peptide and labeled with different intensities of the fluorochrome carboxylfluorescein succinimidyl ester (CFSE). Specific killing of these peptide-pulsed cells was measured after 4 h by flow-cytometric analysis of spleen cells. After 1 and 2 weeks, only a minor fraction of pulsed cells were killed (probably nonspecifically) in naïve mice whereas >98% were destroyed in HSV-1-infected mice, and no difference was apparent between WT and puma^−/− mice (SI Fig. 13). At 4 weeks after infection, more target cell killing was seen in puma^−/− mice compared with WT mice, possibly reflecting the prolonged survival of effector cells afforded by loss of Puma (Fig. 6). These results indicate that loss of Puma prolongs the survival of cytolytic effector T cells in vivo but does not protect target cells from CTL attack.

Fig. 6.

Loss of Puma does not affect cytotoxic activity of HSV-1-specific CD8⁺ T cells in infected mice. Spleen cells from naïve C57BL/6 mice were either loaded with HSV-1-specific peptide (gB_498–505) or incubated without peptide (control). Peptide-pulsed cells were labeled with high levels of CFSE, and control cells were labeled with low levels of CFSE. 10⁷ CFSE^high plus 10⁷ CFSE^low cells were injected into the tail vein of naïve or infected mice (WT or puma^−/− mice infected with HSV-1 28 days earlier). Spleens were harvested 4 h after injection of labeled cells, and the percentages of CFSE^high and CFSE^low cells were determined by flow-cytometric analysis. Histograms are representative of four mice of each genotype.

Discussion

Upon clearance of an infectious pathogen or injected antigen, T cell immune responses are down-regulated, and this involves developmentally programmed death of the majority of the antigen-activated T cells (1, 3). How this pathway to apoptosis is activated is not fully understood, and much confusion has been created by using the term AICD to understand the mechanisms for shutdown of T cell immune responses (2). In the AICD model, T cells are activated in culture by mitogens or antigen and as proliferating blasts restimulated through their TCR. Studies using Fas ligand (FasL) antagonists (5) or T cells from Fas (lpr) or FasL mutant (gld) mice (36) have shown that this death is mediated (at least in part) by FasL-Fas signaling. Confusion arose from the assumption that this mechanism is generally responsible for the shutdown of T cell immune responses in vivo. This appears an unlikely process for the death of T cells during termination of acute immune responses because AICD requires signaling through the TCR and because no antigen (needed for TCR stimulation) is left when these T cells are dying in vivo (2). Accordingly, it has been found that loss of Fas, TNF-R1, or even both does not impede the death of T cells during shutdown of immune responses to a single injection of Staphylococcus enterotoxin B (SEB) (7) or HSV-1 infection (8). It is, however, possible that death receptor signaling is involved in the death of T cells during chronic immune responses when antigens are continually present, because abnormal accumulation of activated T cells was found in lpr mice that had been injected repeatedly with SEB (4, 6, 37).

For the developmentally programmed death of antigen-activated T cells during shutdown of acute immune responses, it is now widely believed that apoptosis is activated through the Bcl-2-regulated pathway because it can be inhibited by overexpression of Bcl-2 (31) or loss of Bim (7, 8) or Bax/Bak (38). Because loss of Bim protected T cells from apoptosis to a lesser extent than Bcl-2 overexpression (7) or Bax/Bak loss (38), it appeared likely that additional BH3-only proteins contribute to this process. In this study, we analyzed side-by-side the roles of five BH3-only proteins—Puma, Bim, Bad, Bid, and Noxa—in the shutdown of the T cell immune response to HSV-1 infection. We demonstrate that Puma contributes significantly to the death of antigen-specific CD8⁺ T cells and therefore plays a physiological role during the shutdown of an immune response. We also confirmed that Bim plays a critical role in this process, but no abnormalities were seen in mice lacking Bid, Bad, or Noxa. Puma-deficient mice cleared the HSV-1 infection at a rate similar to that of WT animals, had normally active cytolytic T cells, and had no abnormalities in the activation status or proliferation of these cells during the course of the infection. This indicates that the abnormal accumulation of HSV-1-specific CD8⁺ T cells in puma^−/− mice cannot be attributed to persistent antigenic stimulation or abnormal proliferation.

So how is Puma activated in antigen-activated T cells during shutdown of an acute immune response? Puma is transcriptionally activated by the tumor suppressor p53 in response to DNA damage (16, 17) and critical for p53-mediated apoptosis (22, 23). However, it appears unlikely that this mechanism activates Puma in T lymphocytes undergoing programmed cell death because these cells are not known to sustain stresses that cause p53 activation and because no major defects in T cell immune responses have been observed in p53^−/− mice (39). Puma can also be activated by p53-independent death stimuli, such as growth factor deprivation (40). Therefore, and because loss of Puma protects a range of hemopoietic cell types (including T cells) against growth factor withdrawal in vitro (22, 23), we believe that puma^−/− mice accumulate excess antigen-activated T cells because they have increased resistance to growth factor deprivation. It has recently been shown that after cytokine deprivation, Puma, like Bim (41), is activated by the transcription factor FOXO3a (35), and it appears likely that this mechanism regulates Puma expression in antigen-activated T cells undergoing programmed cell death. Because of the small number of these cells, FOXO3a activation could, however, not be examined directly by Western blot analysis.

Upon cytokine deprivation FOXO3a promotes not only expression of Puma (35) but also Bim (41). Therefore, and because loss of either Puma or Bim alone causes less accumulation of antigen-activated T cells than Bcl-2 overexpression (7, 8, 31), it appears likely that these two BH3-only proteins have overlapping function in shutdown of acute T cell immune responses. This is supported by the finding that T lymphocytes and many other hemopoietic cells from bim^−/−puma^−/− mice were more resistant to a broad range of apoptotic stimuli, including cytokine withdrawal, than cells lacking only one of these BH3-only proteins (34). However, there were no effects of Puma loss on numbers of antigen-activated T cells after SEB challenge in vivo (34). One possible explanation for this discrepancy might be that because numbers of antigen-activated T cells were abnormally increased relatively late after HSV-1 infection, differences between puma^−/− and WT mice would have become apparent at later times after SEB challenge. Consistent with this idea, our previous work showed that loss of Puma prolonged the survival of SEB-activated T cells cultured in the absence of cytokines. Immune responses in bim^−/−puma^−/− mice could unfortunately not be analyzed because production of the large numbers of animals needed for such studies is precluded by their high incidence of embryonic and early postnatal lethality (34).

Although cytokine deprivation, which is now widely believed to cause contraction of acute T cell immune responses (42, 43), is known to activate the BH3-only protein Bad (26), no defects in T cell immune response shutdown were observed in bad^−/− mice. Therefore, Bad might only have a minor or no role in T cell killing during immune response shutdown, consistent with the fact that it binds only to Bcl-2, Bcl-x_L, and Bcl-w but not to A1 or Mcl-1, whereas the more potent Puma and Bim can neutralize all of these prosurvival proteins (11, 12). Conversely, the finding that Noxa can only bind to Mcl-1 and A1 but not the other Bcl-2-like prosurvival family members (11, 12) may explain why no defects in T cell immune response shutdown were seen in noxa^−/− mice. Bid can bind with high affinity to most prosurvival Bcl-2 family members (with the possible exception of Bcl-2) (11, 12). The finding that its deletion does not cause an increase in HSV-1-specific CD8⁺ T cells is probably due to the fact that Bid needs to be activated by caspase-8 downstream of death receptor signaling, a pathway that is not involved in the shutdown of acute immune responses (7, 8). Bid has also been shown to be activated by and critical for granzyme B/perforin-mediated killing of target cells (44, 45). Our finding that HSV-1 virus is cleared normally from bid^−/− mice demonstrates that this process on its own cannot account for the killing of virus-infected cells in vivo. Puma, Bim, Bad, and Noxa by themselves are also not critical for this cytolytic activity because HSV-1 was also cleared normally from mice lacking these proteins. Moreover, Bcl-2 overexpression also does not protect target cells from CTL-mediated death (31, 46), indicating that the intrinsic apoptotic pathway per se cannot be solely responsible for this process.

Taken together, this study and our previous experiments with bim^−/− mice (8) show that the programmed death of antigen-activated CD8⁺ T cells during termination of an acute T cell immune response requires Puma and Bim, but that other BH3-only proteins, including Bad, Bid, and Noxa, play no significant role. As such, these findings define the physiologically relevant pathways that regulate the homeostasis of antigen-specific T cells and therefore have implications for T cell memory, immunopathology, and vaccination.

Materials and Methods

Mice and Infection.

All mice were bred and experiments performed at the Walter and Eliza Hall Institute under approval from the Royal Melbourne Hospital Animal Ethics Committee. The generation and genotyping of bim^−/− (19), puma^−/− and noxa^−/− (22), bad^−/− (28), and bid^−/− mice (30) have been described previously. Mice were injected s.c. with HSV-1 KOS strain (4 × 10⁵ plaque-forming units) in each hind foot (47).

Immunofluorescent Staining and Flow-Cytometric Analysis.

CD8⁺ cells were enriched by staining with a mixture of rat mAbs to Mac-1 (MI/70), F4/80 (F4/80), Ter-119 (TER119), Gr-1 (RB6–8C5), class II MHC (M5/114), and CD4 (YTA321) followed by incubation with goat anti-rat IgG antibody-coupled magnetic beads (Dynal). Purified CD8⁺ T cells were labeled with APC-conjugated anti-CD8 mAb (53–6.7) and PE-streptavidin-conjugated K^b-gB tetramers containing the immunodominant HSV gB_498–505 peptide (47). Expression of T cell activation markers was analyzed by staining cells with FITC-conjugated mAbs to CD44 (Pgp-1) or CD25 (PC61).

Intracellular FACS Analysis of Cytokine Production.

Spleen cells from HSV-1-infected mice were incubated for 5 h in 5 μg/ml brefeldin A plus 10 units/ml recombinant mouse IL-2 together with 1 μM HSV gB_498–505 (SSIEFARL) or control peptide (the K^b-OVA_257–264 peptide SIINFEKL). Cells were then stained with anti-CD8 mAb and HSV gB_498–505 as described above, fixed with 1% formalin, permeabilized in BSS containing 0.5% saponin and 1% BSA, and stained with PE-coupled rat anti-IFN-γ mAb or an isotype-matched control mAb.

In Vitro Apoptosis Assay.

Spleen cells from HSV-1-infected mice were cultured in the presence or absence of recombinant mouse IL-7. Cells were then stained with APC-coupled anti-CD8 mAb, PE-coupled anti-TCRVα2 mAb (B20.1), or PE-streptavidin-conjugated K^b-gB tetramers containing the immunodominant HSV gB_498–505 peptide (47) plus FITC-coupled annexin V. CTL responses to HSV-1 in C57BL/6 mice show a very strong bias to TCRVα2 plus TCRVβ10 or TCRVβ8 usage (48).

Measuring Viral Titers.

Feet from HSV-1-infected mice were dissected and homogenized in 1 ml of MEM. Aliquots from diluted extracts were cultured at 37°C for 4 days on a VERO cell monolayer previously treated with MEM containing 5 mM Hepes, 0.1 mg/ml streptomycin, 100 units/ml benzylpenicillin, 4 mM l-glutamine, 50 μM 2-mercaptoethanol, 2% FBS, and 3% agarose. Cultures were fixed in 10% formalin, and agarose layers were removed. Cell monolayers were stained with 0.2% crystal violet.

Analysis of T Cell Proliferation in Vivo.

HSV-1-infected mice were injected i.p. at 12 and 20 h before with 100 μg of BrdU (Sigma) per g of body weight and were also given BrdU in the drinking water (1 mg/ml). Splenocytes were stained with anti-CD8 mAb and HSV gB_498–505 MHC tetramers as described. CD8⁺K^b-gB⁺ T cells were sorted on a MoFlo (Cytomation) or FACSDiVa (BD Biosciences), fixed in 0.5% paraformaldehyde, permeabilized, and DNA-denatured in 3 M HCl with 0.5% Tween 20. Cells were then resuspended in 0.1 M Na₂B₄O₇ (pH 9.0) for 3 min at room temperature and stained with FITC-coupled anti-BrdU mAb (Becton Dickinson).

In Vivo T Cell Cytotoxicity Assay.

Spleen cells from naïve C57BL/6 mice were either loaded with HSV-specific peptide (gB_498–505) or incubated without peptide (control). Peptide-pulsed cells were labeled with high carboxylfluorescein succinimidyl ester (CFSE) levels and control cells with low CFSE levels. Cells were injected into the tail vein of naïve puma^−/− mice or into WT or puma^−/− mice infected with HSV-1. Spleens were harvested 4 h after injection of labeled cells, and the percentages of CFSE^high and CFSE^low cells were determined by analysis in a FACScan flow cytometer (BD Biosciences). Specific lysis was calculated by using the formula: % specific lysis = [1 − (r unprimed/r primed)] × 100, where r unprimed and r primed = (% CFSE^low/CFSE^high) for uninfected control and HSV-1-infected mice, respectively.

Statistical Analysis.

Statistical analysis was performed with Student's t test. P values of 0.05 or less were considered significant.