Shutdown of an acute T cell immune response to viral infection is mediated by the proapoptotic Bcl-2 homology 3-only protein Bim

Abstract

We used mutant Fas-deficient (lpr) or Bim-deficient mice to investigate the role of the death receptor and Bcl-2-regulated apoptotic pathways in terminating a physiological T cell response to herpes simplex virus infection. In WT and lpr mice CD8⁺ antigen-specific T cells were deleted after viral clearance. In contrast, the immune response was not terminated in Bim-deficient mice despite viral clearance, and CD8⁺ antigen-specific T cells accumulated in the spleen. Thus, Bim is dispensable for viral clearance but is necessary for the death of activated T cells when immune responses are terminated. These findings have implications for the therapeutic manipulation of immune responses to infections and immunization.

The magnitude of an acute immune response to infection is directly linked to the amount of foreign antigen present in the host. An antigen-driven burst of activation and proliferation of T and B lymphocytes characterizes the initiation and effector phase of the immune response, whereas clearance of antigen is associated with a down-regulation of the immune response (1). This is achieved by switching off proliferation and activating apoptosis in the specific immune responder population (1, 2).

Mammals have two distinct pathways leading to apoptosis (3) that ultimately converge to activate effector caspases. These cysteine proteases destroy vital cellular substrates, activate enzymes that degrade DNA, and thereby mediate the systematic demolition and subsequent engulfment of the dying cell (4). One of the apoptotic pathways involves activation of certain members of the tumor necrosis factor (TNF) receptor (TNF-R) family, called death receptors, including Fas (APO-1/CD95) and TNF-R1 (5). The death receptors, through the adaptor molecule FADD/Mort1 (6–8), activate the initiator caspase, caspase-8 (9). The Bcl-2 family of proteins is not involved in the regulation of this pathway in lymphocytes (3) and myeloid cells (10). In certain cell types, however, the proapoptotic Bcl-2 family member Bid is critical for Fas-induced apoptosis. Bid is proteolytically activated by caspase-8 and can then activate the Bcl-2-regulated apoptotic pathway, thereby amplifying the death signal (11, 12).

The Bcl-2-regulated apoptosis signaling pathway is controlled by the balance between proapoptotic and antiapoptotic members of the Bcl-2 protein family (11–13) and can be activated by developmental cues, cytokine withdrawal, or cytotoxic drugs. The death receptors, with their downstream signal transducers FADD and caspase-8, are not required in this pathway (3, 6, 8, 9). In the Bcl-2-regulated pathway mitochondrial release of cytochrome c causes the adaptor protein Apaf-1 to activate caspase-9, which in turn promotes effector caspase-mediated cell demolition (12). In lymphocytes, myeloid cells, and fibroblasts, caspase-9 and Apaf-1 are not essential for Bcl-2-regulated developmentally programmed cell death and stress-induced apoptosis, but rather, they function to amplify the caspase activation cascade (14).

Bcl-2 is a member of a large family of proapoptotic and antiapoptotic proteins that share between one and four regions of homology, called Bcl-2 homology (BH) domains (11, 13). Bcl-2 is the founding member of the antiapoptotic faction (Bcl-2, Bcl-x_L, Bcl-w, A1, Mcl-1, and Boo/Diva), all of which have three or four BH domains. The proapoptotic members of the family can be subdivided into two groups. Bax, Bak, Bok/Mtd, Bcl-G_L, Bfk, and Bcl-x_S all contain two or three BH regions and, surprisingly, at least Bax shares remarkable structural similarity with the prosurvival relatives (11–13). Members of the other proapoptotic subgroup (Bik/Blk/Nbk, Bad, Hrk/DP5, Bid, Bim/Bod, Noxa, Bmf, and Puma/Bbc3) possess only the BH3 domain and are therefore often referred to as BH3-only proteins (15). Genetic and biochemical studies have shown that BH3-only proteins are required for initiation of apoptosis, whereas the Bax/Bak-like proteins play an essential role in downstream signaling events (15). Experiments with Bim-deficient mice have shown that this BH3-only protein is required for lymphocyte apoptosis induced by cytokine withdrawal, deregulated calcium flux, and taxol treatment (16). Moreover, loss of Bim protects self-antigen-specific thymocytes (17) and B cells (18) against negative selection and prolongs the in vivo survival of T cells activated by the superantigen staphylococcal enterotoxin B (SEB) (19) or cross-presented antigen (20).

Defects in Fas or its cognate ligand, FasL, in mice or humans cause T cell hyperplasia and autoantibody production, which leads to significant morbidity and mortality (21). Experiments with cultured T cells have shown that FasL–Fas signaling is required for apoptosis induced by repeated ligation of the T cell antigen receptor (TCR) (21). On the basis of these results it has been predicted that FasL–Fas signaling plays a critical role in shutting down T cell responses in vivo. Indeed, it was reported that T cell responses to peptide or superantigen are abnormally enhanced and prolonged in FasL- or Fas-deficient mice (22–25). Other studies, however, have shown that activated T cells can undergo apoptosis in mutant lpr and gld mice (19, 26–28) and even in mice deficient in Fas, TNF receptor 1, and TNF receptor 2 (19).

Proliferation and survival of antigen-activated T cells in vivo and in vitro are cytokine dependent. Because pathogen clearance is followed by a decrease in cytokine levels (29), we postulate that during termination of an immune response activated T cells undergo apoptosis because of cytokine deprivation and that this death is Bim dependent but Fas independent. To test this hypothesis, we examined the T cell response to human herpes simplex virus (HSV-1) infection in Fas mutant lpr mice and Bim-deficient mice. The importance of CD8⁺ T cells in clearing this infection is highlighted by the high morbidity and mortality seen in HSV-infected rag-1^–/– mice and the fact that this can be prevented by transplanting HSV-specific CD8⁺ T cells (F. Carbone, personal communication).

The HSV infection model also enabled us to study the roles of Fas and Bim in viral clearance. Apoptosis functions as an important defense against infection, and cells can engage a number of mechanisms to activate the apoptotic program when infection is sensed (30). In turn, viruses have appropriated many antiapoptotic regulators to keep host cells alive including viral FLIP, which inhibits death receptor-induced caspase 8 activation, and several Bcl-2 homologs (30). If their innate defense is overwhelmed, infected cells can still be killed by cytotoxic T lymphocytes (CTLs), which initiate apoptosis by a contact-dependent mechanism, either by FasL or TNF, or by release of the pore-forming protein perforin plus granzymes (30). The serine protease granzyme B can activate caspases directly, by proteolysis, or indirectly by processing and activating Bid (30). Granzyme B may also bypass caspases and proteolyze essential cellular proteins directly (30).

Our results show that lpr mutant and Bim-deficient mice are able to clear virus and virally infected cells efficiently and that Bim but not Fas is essential for down-regulation of T cell immune responses in vivo. These results indicate how the immune response could potentially be modulated by impairing T cell death without affecting viral elimination.

Methods

Infection of Mice. All experiments with animals were performed according to the guidelines of the Melbourne Directorate Animal Ethics Committee. The bim^–/– (16) and lpr mice (31), both backcrossed to C57BL/6 mice for >12 generations, were age- and sex-matched with C57BL/6 mice and analyzed between the ages of 6 and 10 weeks. Animals were injected s.c. with 4 × 10⁵ plaque-forming units of HSV-1 KOS strain (32) in each hind foot after anaesthetic induction. At the specified times, mice were killed, and spleen and bilateral popliteal lymph nodes were harvested for analysis. Feet were stored at –70°C for viral plaque assays.

Immunofluorescence Staining and Flow Cytometry. Approximately 0.5–1 × 10⁶ cells from organ single-cell suspensions were used to purify CD8⁺ T cells by magnetic bead depletion using antibodies M1/70 (anti-Mac-1), F4/80, Ter 119, RB6-8C5 anti-Gr-1, M5/114 anti-class II MHC, and GK 1.5 anti-CD4 followed by goat anti-rat IgG-coupled magnetic beads (Dynal, Oslo). Enriched CD8⁺ T cells were labeled with anti-CD8-APC (53-6.7) mAb for 30 min on ice in balanced salt solution containing 2% FBS (Trace Biosciences, Melbourne). After washing, cells were stained for 45 min at 37°C with R-phycoerythrin (PE)-streptavidin-conjugated K^b-gB tetramer containing the immunodominant HSV gB_498–505 peptide, which was prepared as described (32, 33). For analysis of T cell activation marker expression, cells were additionally labeled with FITC-conjugated anti-CD44 mAb (IM78.1) or anti-CD25 mAb (PC61). Propidium iodide (PI) (2 μg/ml; Sigma) staining was used to exclude dead cells. Data on at least 10⁴ viable (PI^–) cells were acquired on a LSR flow cytometer (Becton Dickinson) and analyzed with cellquest software (BD Biosciences, San Jose, CA).

Intracellular Staining for Cytokines. Approximately 5 × 10⁵ cells were incubated in DMEM supplemented with 50 μM 2-mercaptoethanol, 10% FBS, 5 μg/ml brefeldin A, and 50 units/ml recombinant mouse IL-2 along with 1 μM HSV gB_498–505 (SSIEFARL) or control peptide (the K^b-OVA_257–264 peptide SIINFEKL) for 5 h at 37°C. Peptides were synthesized by Mimotopes (Melbourne). All subsequent washes and incubations until the completion of fixation were performed in buffers containing 5 μg/ml brefeldin A (Epicentre Technologies, Madison, WI). After incubation cells were washed, stained as described above, washed in balanced salt solution, and fixed with 1% formalin in PBS for 15 min at 4°C before permeablization in PBS containing 0.5% saponin (Sigma) and 1% BSA (Sigma) for 10 min at room temperature (RT). Cells were then washed and stained with PE-coupled rat anti-IFN-γ mAb (Cl.XMG1.2) or an isotype-matched control mAb (PE-rat-anti-mouse IgG₁ Cl.R3–34) for 30 min on ice and then analyzed by using a FACScan analyzer (Becton Dickinson). The frequency of virus-specific CD8⁺ T cells was determined by subtracting the frequency of CD8⁺ IFN-γ⁺ T cells in samples stimulated with control peptide from the frequency of CD8⁺ IFN-γ⁺ T cells in samples stimulated with the HSV peptide.

Assays for Measuring Cell Proliferation in Vivo. HSV-infected mice were injected i.p. 20 and 12 h before harvesting organs with 100 μg/g of body weight of BrdUrd dissolved in normal saline with 7 mN NaOH. Mice were also given drinking water containing 1 mg/ml of BrdUrd and 2% glucose for the 20-h period before organ harvest. Cells from spleen and lymph nodes were stained, and CD8⁺K^b-gB⁺ T cells were analyzed by fluorescence-activated cell sorter (FACS) and fixed in 0.5% paraformaldehyde for 30 min at RT. After two washes in PBS, cells were permeabilized, and the DNA was denatured in 3 M HCL with 0.5% Tween 20 (Sigma) for 20 min at RT. After pelleting, cells were resuspended in 0.1 M Na₂B₄O₇, pH 9.0, for 3 min at RT. After two washes in PBS, cells were stained with FITC-coupled anti-BrdUrd mAb (Becton Dickinson) for 30 min at 4°C in PBS containing 2% FBS and 0.5% Tween 20, then washed with PBS containing 0.5% Tween 20 and 2% FBS and analyzed on a FACScan flow cytometer.

Viral Titer Assays. Feet were dissected and homogenized for 30 s in MEM (GIBCO 61100-061) supplemented with 10% FBS by using a Polytron homogenizer (Kinematic, Lucerne, Switzerland). The homogenate was serially diluted up to 1/10^–6 in serum-free MEM. Aliquots of each dilution were added to a VERO cell monolayer, previously washed with serum-free MEM, and then incubated at RT for 60 min. 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 was added to each monolayer and allowed to set before incubating cells at 37°C for 4 days. Cultures were fixed for 30 min in 10% formalin in PBS, and agarose layers were then removed and the cell monolayer was stained with 0.2% crystal violet for 1 min before rinsing with water. Monolayers were dried at RT overnight, and plaques were counted.

Cell Death Assays. Enriched CD8⁺ T cells were cultured in DMEM supplemented with 50 μM 2-mercaptoethanol plus 10% FBS (Trace Biosciences, Melbourne) at 37°C in the presence or absence of recombinant mouse IL-7 (from IL-7 expression vector-transfected J558/L cells; a gift from A. Rolink, University of Basel, Basel, Switzerland). At the specified times, cells were washed, resuspended in balanced salt solution containing 2% FBS, and stained as described above with APC-anti-CD8 mAb, PE anti-TCRVα2 mAb (B20.1, Becton Dickinson) plus FITC-annexin V to quantify dead and dying cells. CTL responses to HSV in C57BL/6 mice show a very strong bias to TCRVα2 plus TCRVβ10 or TCRVβ8 usage, and previous studies have validated the use of antibodies to these TCR-variable regions for following a CTL response to HSV (33). Data on 10⁴ cells were collected on a LSR f low cytometer and analyzed with cellquest software.

Results

bim^–/– Mice Have Abnormally Increased Numbers of CD8⁺ T Cells. Previous studies have shown that bim^–/– mice accumulate abnormally large numbers of B cells, T cells, monocytes, and granulocytes (16). When bim^–/– mice were infected with HSV, their numbers of total leukocytes and CD8⁺ T cells did not change compared with naïve bim^–/– mice. At all time points before and after infection (0–28 days) we observed a 2- to 5-fold increase in total leukocyte numbers and numbers of CD8⁺ T cells in popliteal lymph nodes, draining the site of infection, and in spleens compared with WT mice (Fig. 1; P < 0.05).

Fig. 1.

Accumulation of leukocytes and CD8⁺ T cells in Bim-deficient mice. WT, bim^–/–, and lpr mice were injected with HSV-1 in the hind foot. Total leukocyte numbers and numbers of CD8⁺ T cells in draining popliteal lymph nodes (LN) (A and B, respectively) and spleen (C and D, respectively) were quantified by cell counting and immunofluorescent staining with surface marker-specific mAbs at specified time points after infection. Two lymph nodes were harvested from each mouse, and the lymph nodes from three mice of each genotype were pooled to facilitate analysis. Data shown represent numbers of lymph node leukocytes or CD8⁺ T cells per mouse and are mean ± SD for nine WT and nine bim^–/– mice but only for three lpr mice for which SD is not shown. Spleens were not pooled, and data represent mean ± SD for three to seven mice of each genotype.

Kinetics of CD8⁺ T Cell Response to HSV in the Draining Lymph Nodes. We used two techniques to follow the kinetics of a CD8⁺ T cell response to localized (hind foot) infection with HSV in mice. In the first, virus-specific CD8⁺ T cells were identified by staining with PE-conjugated MHC class I tetrameric complexes incorporating the gB_498–505 HSV glycoprotein peptide, the major epitope from HSV recognized by CD8⁺ T cells (33). Alternatively, cells from spleens or lymph nodes of infected animals were stimulated in culture with the gB_498–505 peptide, and antigen-specific CD8⁺ T cells were then enumerated by intracellular immunofluorescent staining with anti-IFN-γ antibodies and FACS analysis. Both techniques produced similar results and are presented here (Fig. 2 A and B). In the draining popliteal lymph nodes of WT mice there was a rapid increase in CD8⁺ HSV-specific T cells. The peak of the response, with an ≈50-fold increase in HSV gB_498–505-specific CD8⁺ T cells, was reached on day 7. This was followed by a rapid decline such that by 14 days most of these T cells had disappeared (Fig. 2 A and B). The kinetics of the CD8⁺ T cell responses to HSV in lpr and bim^–/– mice were similar to that found in WT animals. We believe that the decline in CD8⁺ T cell numbers is caused by a redistribution of HSV-specific CTLs from the lymph node to the spleen (see below).

Fig. 2.

Kinetics of an acute immune response to HSV infection in the draining popliteal nodes and accumulation of HSV-specific CD8⁺ T cells in the spleens of bim^–/– mice. The total numbers of HSV-specific CD8⁺ T cells in draining lymph nodes (LN) of WT, lpr, and bim^–/– mice were quantified at the specified time points before and after HSV infection by using surface staining with K^b-gB_498–505 tetramers (A) or by staining for intracellular IFN-γ in cells that had been restimulated in vitro with HSV gB_498–505 peptide (B). (The gB_498–505 peptide is the major HSV-derived epitope recognized by CD8⁺ T cells in infected C57BL/6 mice.) Lymph nodes from three mice of each genotype were pooled (this was repeated three times in the case of WT and bim^–/– mice, i.e., nine mice). Data points represent average cell numbers (mean ± SD) per mouse. The total numbers of HSV-specific CD8⁺ T cells in spleens of WT, lpr, and bim^–/– mice were quantified at the specified times before and after HSV infection by surface staining with K^b-gB tetramers (C) or by intracellular staining for IFN-γ in cells that had been restimulated in vitro with HSV gB_498–505 peptide (D). Data represent mean ± SD from three to nine mice of each genotype.

Accumulation of HSV-Specific CD8⁺ T Cells in the Spleens of bim^–/– Mice. When we studied the kinetics of HSV-specific CD8⁺ T cell responses in the spleens of infected animals, we found that bim^–/– mice differed significantly from both WT and lpr mice; the latter two were indistinguishable. In WT and lpr mice there was an increase in HSV-specific CD8⁺ T cells with a peak on day 7 after infection, followed by loss of most (>90%) of these cells. The bim^–/– mice showed a substantially augmented accumulation of HSV-specific CD8⁺ T cells, which was maintained even up to 4 weeks postinfection, a time when very few of these cells were detectable in WT or lpr mice (Fig. 2 C and D). HSV-specific CD8⁺ T cells from both WT and bim^–/– mice were capable of killing splenocytes presenting the major HSV epitope (gB_498–505) in vivo (Fig. 7, which is published as supporting information on the PNAS web site). The accumulation of HSV-specific CD8⁺ T cells in bim^–/– mice could potentially be explained by abnormally prolonged proliferation of bim^–/– T cells, incomplete clearance of virus in bim^–/– mice causing a persistent antigendriven response or a defect in apoptosis in bim^–/– T cells.

Proliferation and Activation of HSV-Specific CD8⁺ T Cells in Vivo Are Indistinguishable Among WT, lpr, and bim^–/– Mice. We pulsed HSV-infected WT, lpr, and bim^–/– mice for 20 h with BrdUrd before harvesting organs for analysis of proliferation of HSV-specific CD8⁺ T cells in vivo. Virus-specific CD8⁺ T cells were FACS-sorted, fixed, permeabilized, and then stained with an anti-BrdUrd antibody to identify cells that had been synthesizing DNA, and hence were undergoing proliferation. We found that the proliferative rates of HSV-specific CD8⁺ T cells from the popliteal lymph nodes (days 4 and 7 after infection) and spleens (day 7) of WT, bim^–/–, and mutant lpr mice were all comparable (Fig. 3). For example, 7 days after infection 66% WT, 67% lpr, and 65% bim^–/– antigen-specific splenic T cells were synthesizing DNA in vivo during the time of BrdUrd administration. Similar results were obtained when HSV-specific T cells were stained with the DNA dye 7AAD; no difference in cell cycle distribution was observed between cells from WT and bim^–/– mice (data not shown). HSV-specific CD8⁺ T cells harvested 20 days after infection from spleens of bim^–/– mice (a time when numbers of such cells in WT and lpr mice were too low for analysis) did not incorporate BrdUrd (Fig. 3). This finding demonstrates that the HSV-specific CD8⁺ T cells that persist in bim^–/– mice are noncycling, indicating that they are no longer being stimulated in vivo by antigen and/or cytokines.

Fig. 3.

Rates of in vivo proliferation of HSV-specific CD8⁺ T cells are indistinguishable among WT, lpr, and bim^–/– mice. HSV infected WT, lpr, and bim^–/– mice were pulsed for 20 h with BrdUrd before harvesting organs for analysis. HSV-specific CD8⁺ T cells were FACS-sorted after surface labeling with anti-CD8 mAb plus with K^b-gB_498–505 tetramers and were then stained intracellularly with an anti-BrdUrd mAb to identify cells synthesizing DNA and hence undergoing proliferation. HSV-specific CD8⁺ T cells from WT, lpr, and bim^–/– mice were collected from lymph nodes 4 and 7 days after HSV infection. HSV-specific CD8⁺ T cells from spleen were harvested 7 days after infection (and 20 days after infection only in bim^–/– mice because numbers in WT and lpr animals are too low for analysis). Comparisons are made between control CD8⁺ T cells from uninfected mice (gray lines) and HSV-specific T cells from infected mice (black lines) of the same genotype.

Cell surface staining with specific mAbs was performed to investigate the expression of two classical activation markers, CD25 (IL-2Rα chain) and CD44, on HSV-specific CD8⁺ T cells from infected WT, bim^–/–, and mutant lpr mice. Our analysis showed that in all three genotypes of mice the HSV-specific CD8⁺ T cells from the draining popliteal lymph nodes upregulated their expression of both CD44 and CD25 early in the immune response (Fig. 4). In all three mouse strains, CD44 levels remained high in HSV-specific T cells 7 days after infection, but expression of CD25 returned to basal levels. CD25^–44⁺ was the dominant phenotype of HSV-specific CD8⁺ T cells found in the spleen, and there was no difference in activation marker expression between mice of the three genotypes. Taken together, these data indicate that prolonged activation and/or excessive proliferation are not responsible for the abnormal accumulation of HSV-specific CD8⁺ T cells in bim^–/– mice.

Fig. 4.

Expression of cell surface activation markers on HSV-specific CD8⁺ T cells is indistinguishable among WT, lpr, and bim^–/– mice. CD8⁺ T cells from popliteal nodes (LN) at day 5 (A) or day 7 (B) post-HSV infection and from spleen at day 7 postinfection (C) were stained with HSV gB_498–505 peptide tetramers and either mAbs to CD44 or CD25. Profiles of HSV-specific CD8⁺ T cells (black lines) are plotted with profiles from control CD8⁺ T cells of uninfected mice (gray lines).

Viral Clearance Kinetics Are Similar in WT, lpr, and bim^–/– Mice. Viral plaque-forming assays were performed to examine the kinetics of viral clearance from the hind feet of infected WT, bim^–/–, and mutant lpr mice. We found that mice from all three genotypes cleared HSV with similar efficiency. By day 7 after infection, nearly all virus was cleared (Fig. 5A), consistent with our finding that at this time HSV-specific CD8⁺ T cells had down-regulated the activation marker CD25 and CTL numbers began to decline in both lpr and WT mice. This finding indicates that persistent viral infection was not causing the accumulation of antigen-specific CD8⁺ T cells in bim^–/– mice.

Fig. 5.

No differences in viral clearance kinetics and distribution of HSV-specific CD8⁺ T cells to nonlymphoid organs were found among HSV-infected WT, lpr, and bim^–/– mice. (A) Feet from HSV-infected mice were homogenized in medium, and extracts were serially diluted for plaque-forming assays. Data represent arithmetic mean (pair of feet) ± SD of three mice from each genotype. PFU, plaque-forming unit. (B) CD8⁺ T cells were enriched from pooled lungs and livers of three WT and three bim^–/– mice 7 days post-HSV infection. Data represent numbers of HSV-specific CD8⁺ T cells per organ per mouse.

Distribution of HSV-Specific CD8⁺ T Cells into Nonlymphoid Organs Are Similar in WT and bim^–/– Mice. The persistence of virus-specific CD8⁺ T cells in bim^–/– compared with WT and lpr mutant mice might also be caused by differences in their trafficking. We therefore measured the numbers of HSV-specific CD8⁺ T cells in livers and lungs of WT and bim^–/– mice 7 days after infection, the time when the numbers of these cells began to fall in the lymph nodes and spleens of WT animals. We found only small numbers of virus-specific CD8⁺ T cells in these two nonlymphoid organs, and there was no significant difference between WT and bim^–/– mice (Fig. 5B).

Loss of Bim Renders HSV-Specific CD8⁺ T Cells Resistant to Cytokine Withdrawal-Induced Apoptosis. Based on the results described above, it appeared likely that apoptosis was responsible for the deletion of HSV-specific T cells from the spleens of WT and lpr mice and that this was inhibited by the loss of Bim. Because cytokine levels decline after virus clearance (1) and injection of certain cytokines delays the deletion of activated T cells (34), this wave of apoptosis is likely caused by cytokine withdrawal. We therefore enriched CD8⁺ T cells from spleens and lymph nodes of HSV-infected WT and bim^–/– mice, cultured them in the presence or absence of IL-7, and then stained them with antibodies to CD8 and TCRVα2 [most HSV-specific CD8⁺ T cells express Vα2-containing TCR (33)] and FITC-annexin V to measure survival of antigen-specific CD8⁺ T cells. HSV-specific CD8⁺ T cells from WT and bim^–/– mice survived well in the presence of IL-7 with >80% survival after 3 days in culture (Fig. 6). In the absence of cytokine, WT T cells died very rapidly, and <5% were found alive after 2 days. In contrast, the bim^–/– T cells survived in simple medium just as well as they did in the presence of IL-7 with >80% survival after 3 days of growth factor starvation (Fig. 6). To investigate the possibility that bim^–/– T cells were secreting a prosurvival factor, for example IL-7, we mixed activated WT (Ly5.1) and activated bim^–/– (Ly5.2) T cells and found that bim^–/– T cells did not promote survival of cocultured WT T cells (Fig. 8 A and B, which is published as supporting information on the PNAS web site). Furthermore, an IL-7R-blocking antibody (35) did not alter the survival of activated bim^–/– T cells (Fig. 8 C and D). These results indicate that the accumulation and persistence of HSV-specific CD8⁺ T cells in bim^–/– mice is attributable to a cell-intrinsic resistance to cytokine withdrawal-induced apoptosis.

Fig. 6.

HSV-specific CD8⁺ T cells from bim^–/– mice are resistant to cytokine withdrawal. CD8⁺ T cells from lymph nodes (LN) (A) or spleen (B) of HSV-infected WT or bim^–/– mice were cultured in the presence or absence of IL-7. Viability of HSV-specific CD8⁺ T cells was determined by staining with anti-CD8 mAb, anti-TCRVα2 mAb [most HSV-specific CD8⁺ T cells in C57BL/6 mice express Vα2-containing TCR; (33)], and FITC-annexin V. Live cells were considered as being FITC-annexin V^–. Data represent mean ± SD of HSV-specific CD8⁺ T cells from three mice of each genotype.

Discussion

In an immune response to a pathogen, clearance of antigen is associated with a decline in the numbers of specific CTLs. We investigated the role of two apoptosis regulators, the death receptor Fas and the proapoptotic BH3-only Bcl-2 family member Bim, in the immune response to HSV infection. We found that the kinetics of the increase and subsequent disappearance of HSV-specific CD8⁺ T cells in the draining lymph nodes were similar in WT, mutant lpr, and bim^–/– mice, but there was a massive accumulation and persistence of such cells in the spleens of bim^–/– mice. The disappearance of HSV-specific CD8⁺ T cells from the draining lymph nodes in all three mouse strains is most likely caused by redistribution of cells to the spleen, and apoptosis does not appear to play a significant role. Consistent with this notion, previous work, which followed carboxyfluorescein succinimidyl ester-labeled adoptively transferred TCR transgenic HSV-specific T cells in WT-infected mice, showed that the decline in numbers of these cells in the draining lymph nodes coincided with their appearance in the spleen (33). In addition, HSV-specific CD8⁺ T cells from lymph nodes of bim^–/– mice were resistant to cytokine withdrawal-induced apoptosis to the same degree as HSV-specific CD8⁺ T cells that accumulated in the spleen of bim^–/– mice (Fig. 6), yet the former population of T cells disappeared from the popliteal nodes much like the WT and lpr cells (Fig. 2). No difference was observed between WT and bim^–/– HSV-specific T cells in their in vivo proliferation (Fig. 3), expression of activation markers (Fig. 4), or their trafficking to nonlymphoid organs (Fig. 5B). We therefore believe that the kinetics of the CD8⁺ T cell response to HSV is best explained by an initial antigen-driven proliferation of CD8⁺ T cells in draining lymph nodes, followed by redistribution of these cells to the spleen where, after viral clearance, most of these cells die by a Bim-dependent but Fas-independent mechanism. Our findings that death of activated T cells in vivo is Bim dependent but Fas independent is consistent with previous studies of T cells stimulated with nonreplicating, and therefore less physiological antigens, such as SEB (19) or a transgenic neo-self-antigen cross-presented by class I MHC (20).

There are several ways in which Bim could be involved in the down-regulation of a CD8⁺ T cell response. It has been demonstrated that pathogen clearance leads to a decrease in cytokine levels and that this is associated with and precedes apoptosis of activated T cells in vivo (29). We found that HSV-specific CD8⁺ T cells from WT mice can be kept alive in culture with IL-7 (Fig. 6). Moreover, previous in vivo studies showed that SEB-activated T cells could be rescued from apoptosis by injection of cytokines that, like IL-7, signal through γ_c-containing receptors (34, 36). Cytokine receptor stimulation increases the levels of Bcl-2 (and some of its functional homologs) in T cells (37). Interestingly, the levels of cytokines (29) and the levels of Bcl-2 both decrease around the time SEB-activated T cells undergo apoptosis in mice (19), and their death can be inhibited by Bcl-2 overexpression (19, 38). Other studies have shown that the levels of Bim, the proapoptotic BH3-only protein required for growth factor withdrawal-induced lymphocyte apoptosis (16), increase in cytokinestarved hemopoietic cells (39, 40). Collectively, these results demonstrate that cytokine withdrawal causes the balance between proapoptotic versus antiapoptotic Bcl-2 family members to shift in the direction of apoptosis. We speculate that during termination of an immune response the wave of CD8⁺ T cell apoptosis, which is Bim mediated, is a consequence of the decreasing levels of cytokines, which in turn results from pathogen clearance.

We have shown that Fas is not required for the down-regulation of a physiological immune response to HSV infection. Consistent with our data, it has been reported that Fas and the two TNF receptors are not required for in vivo death of SEB-activated T cells (19). However, mice with defects in Fas or FasL do develop T cell hyperplasia and autoimmunity (31). It is possible that in persistent infections associated with repeated TCR stimulation, Fas and perhaps other death receptors could be important in the deletion of chronically activated T cells. It will be interesting to test whether compound mutant animals with loss of Bim plus defects in Fas (lpr) or FasL (gld) have a more severe abnormality than bim^–/– mice in down-regulating acute immune responses and deleting chronically activated T cells.

Apoptosis plays a role in the killing of virally infected cells (41). Several CD4⁺ and CD8⁺ T cell-mediated mechanisms are involved in this process, including cytolysis mediated by perforin plus granzymes, ligands for death receptors (e.g., FasL or TNF) and interferons (41). Moreover, upon sensing pathogen-induced perturbations in cell physiology, infected cells can initiate apoptosis in a cell-autonomous manner (41). So far, Bim has not been shown to be activated by any of the cytotoxic T cell effector mechanisms, but it might function as part of the innate tripwires that activate apoptosis in virally infected cells. We found that HSV clearance was not impaired in lpr mutant or Bim-deficient mice (Fig. 5A). Fas and Bim therefore do not appear to be required for either cell intrinsic apoptosis or cytotoxic T cell-mediated killing of HSV-infected cells. A redundant role for Fas and Bim in this process is, however, possible because mammals and other multicellular organisms have evolved several, apparently overlapping, apoptotic pathways to cope with the threat posed by intracellular pathogens.

Our findings that Bim loss promotes accumulation of activated T cells but does not affect HSV clearance indicate that it might be possible to manipulate the apoptotic system selectively, allowing for modulation of the CD8⁺ T cell response without compromising pathogen elimination. These results may have implications for the maintenance and/or manipulation of immune responses to infections in humans. Because reduction of Bim levels by even just 50% has significant impact on T cell survival in vitro and in vivo (16, 42), Bim antagonists or factors that down-regulate Bim could have therapeutic value in boosting immune responses in vaccination or infection without affecting apoptotic clearance of infected cells.