Role of Bim in Regulating CD8⁺ T-Cell Responses during Chronic Viral Infection

Abstract

Apoptosis is critical for the development and maintenance of the immune system. The proapoptotic Bcl-2 family member Bim is important for normal immune system homeostasis. Although previous experiments have shown that Bim is critical for the apoptosis of antigen-specific CD8⁺ T cells during acute viral infection, the role of Bim during chronic viral infection is unclear. Using lymphocytic choriomeningitis virus clone 13 infection of mice, we demonstrate a role for Bim in CD8⁺ T-cell apoptosis during chronic viral infection. Enumeration of antigen-specific CD8⁺ T cells by major histocompatibility complex class I tetramer staining revealed that CD8⁺ D^bNP396-404⁺ T cells, which undergo extensive deletion in wild-type mice, exhibited almost no decrease in Bim mutant mice. This contrasts with CD8⁺ D^bGP33-41⁺ and CD8⁺ D^bGP276-286⁺ T cells that underwent similar decreases in numbers in both Bim mutant and wild-type mice. Increased numbers of CD8⁺ D^bNP396-404⁺ T cells in Bim mutant mice were due to lack of apoptosis and could not be explained by altered proliferation, differential homing to tissues, or increased help from CD4⁺ T cells. When viral titers were examined, high levels were initially observed in both groups, but in Bim mutant mice, clearance from the spleen and sera was slightly accelerated. These experiments demonstrate the critical role of Bim during chronic viral infection to down-regulate CD8⁺ T-cell responses and have implications for designing strategies for optimizing immunotherapies during situations where antigen persists, such as chronic infection, autoimmune syndromes, and cancer.

Apoptosis plays a key role in the development and maintenance of multicellular organisms. For proper physiological functioning, the total amount of cell death must be tightly controlled. Excess apoptosis has been linked to degenerative syndromes such as Alzheimer's disease (9, 22, 44) and Parkinson's disease (11, 34), while a failure to complete apoptosis has also been implicated in the pathogenesis of cancer (16). In the immune system, apoptosis plays a critical role not only in development by balancing the generation of a diverse and functional repertoire but also by maintaining relatively constant numbers of cells throughout most of the life of the animal.

CD8⁺ T cells are critical for the clearance of intracellular pathogens such as viruses, certain bacteria, and some tumors. These cells proceed through three developmental states. After their initial thymic emigration, naive CD8⁺ T cells exist in a quiescent state, undergoing little to no proliferation in secondary lymphoid organs (10). After encounter with an antigen-presenting cell-expressing cognate peptide major histocompatibility complex (MHC) class I complexes and costimulatory molecules, T cells become activated and begin to proliferate and differentiate. Naive cells become effector cells, which are capable of producing cytokines such as gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), and interleukin 2 (IL-2) (24, 40). In addition to cytokine production, effector CD8⁺ T cells can also lyse infected cells through perforin- and Fas-mediated pathways. At the same time as they begin to produce cytokines, effector cells divide multiple times to rapidly increase their numbers. Studies have documented that this expansion is a programmed event that occurs even when cells are exposed to antigen only for a brief period (20). The fate of the effector CD8⁺ T-cell population differs between acute and chronic viral infections. In acute lymphocytic choriomeningitis virus (LCMV) Armstrong infection, the vast majority of the effector T cells undergo apoptosis (26) through a process termed contraction. The surviving cells are memory cells and can provide lifelong immunity to the pathogen. These cells possess two cardinal properties: the ability to undergo homeostatic proliferation to maintain numbers (27) and the ability to rapidly respond to antigen (26). During chronic viral infection, a large-scale expansion of effector CD8⁺ T cells occurs. However, virus is not cleared and persists for months in some organs, while remaining at high levels in others throughout the remainder of the animal's life (12, 46). Similar to acute infection, cell numbers also begin to decrease by day 8 of chronic infection. This process has been referred to as deletion (47).

The purpose of the experiments presented here was to determine the role of Bim in downregulating CD8⁺ T-cell responses during chronic viral infection. Prior studies have demonstrated that contraction of antigen-specific CD8⁺ T cells does not occur following acute human herpes simplex virus type 1 infection of Bim mutant mice (33). Bim is a BH₃-only proapoptotic member of the Bcl-2 superfamily. Originally identified as a protein that interacted with Bcl-2 (19, 30), Bim has been shown to be a key regulator of apoptosis in both lymphocytes (5) and neurons (4). Lymphocytes from Bim mutant mice are resistant to negative selection, growth factor withdrawal-induced apoptosis, and to a lesser degree, dexamethasone-induced cell death (5, 6). The exact mechanisms by which Bim functions to cause cell death remain unclear, as some studies have suggested that it resides in a complex with dynein microtubules in the cytosol and it translocates to the mitochondria upon cell stress, initiating death (35). In lymphocytes, it appears that the vast majority of Bim is already localized to the mitochondria and complexed with Bcl-2 (50). Indeed, activation does not appear to alter Bim levels in CD8⁺ T cells (18), but instead, it appears to alter the association between Bcl-2 and Bim (50). We found that Bim mutation almost completely blocked the deletion of CD8⁺ D^bNP396-404⁺ T cells while leaving other specificities less affected.

MATERIALS AND METHODS

Mice and infection.

Bim null and wild-type mice were bred in our facility. The Bim mutant mice were originally derived on a C57BL/6/129 background (5) and have been backcrossed onto C57BL/6 for 12 generations. Wild-type and null mice were derived from littermates of heterozygous matings. Mice were infected with 2 × 10⁶ PFU of LCMV clone 13 intravenously or 2 × 10⁵ PFU of LCMV Armstrong intraparentally and used at the indicated time points. LCMV clone 13 stocks were grown and quantitated as described previously (1).

Preparation of MHC class I tetramers.

The construction and purification of D^bGP33-41, D^bNP396-404, and D^bGP276-286 have been described previously (26).

Surface and intracellular staining.

All antibodies were purchased from BD Pharmingen (San Diego, CA). In this study, the following antibodies were used: rat anti-mouse CD8α-phycoerythrin, rat anti-mouse CD44-fluorescein isothiocyanate, rat anti-mouse TNF-α-allophycocyanin, rat anti-mouse CD4-phycoerythrin, and rat anti-mouse IFN-γ-fluorescein isothiocyanate. Surface staining was performed by incubation of antibodies at a 1:100 dilution in fluorescence-activated cell sorter buffer (2% fetal calf serum, phosphate-buffered saline [PBS]) for 30 min at 4°C. After three washes, cells were intracellularly stained with the appropriate antibody using the BD Cytofix/Cytoperm kit according to the manufacturer's instructions. Bromodeoxyuridine (BrdU) staining was performed as described by Tebo et al. (45). Samples were acquired on a FACSCalibur instrument and analyzed with FloJo software (TreeStar, San Francisco, CA).

Cell isolation.

Spleens were removed from mice after cervical dislocation. After being teased apart on a wire mesh, red blood cells were removed by osmotic lysis in ACK buffer. Splenocytes were then resuspended in RPMI supplemented with 10% fetal calf serum and l-glutamine, penicillin-streptomycin, and β-mercaptoethanol. For isolation of nonlymphoid tissues, each mouse was euthanized, its abdomen was opened, the hepatic vein was cut, and 5 ml ice-cold PBS was injected directly into the hepatic artery to perfuse the liver. The gall bladder was removed, and the entire liver was excised. The liver tissue was homogenized using a wire screen and incubated in 0.25 mg/ml collagenase B (Boehringer Mannheim, Federal Republic of Germany) and 1 U/ml DNase (Sigma, St. Louis, MO) at 37°C for 60 min. The digested liver was centrifuged, and the pellet was resuspended in 5 ml 44% Percoll (Sigma). This solution was underlaid with 56% Percoll and spun at 2,000 rpm for 20 min at 20°C. The intrahepatic lymphocyte populations were harvested from the interface, and red blood cells were lysed using 0.83% ammonium chloride, washed, and counted.

Annexin V and 7-AAD staining.

For analysis of direct ex vivo apoptosis, splenocytes were isolated and then surface stained, as described above, incubated with annexin V and 7-actinomycin D (7AAD; BD Pharmingen) at room temperature for 15 min, and acquired immediately.

Statistical analysis.

Data from wild-type and mutant mice were analyzed using two-tailed Student's t test, and a P value of <0.05 was considered significant.

RESULTS

Immunodominance hierarchy is altered during chronic viral infection of Bim mutant mice.

To determine the role of Bim in antigen-driven expansion and death during chronic viral infection, we infected wild-type and null mutant mice with LCMV clone 13. This strain of LCMV induces a chronic viral infection that takes 60 to 90 days to be cleared from most organs, but virus persists in the kidneys and brain. Animals were sacrificed at various time points postinfection, and splenocytes were stained with anti-CD8α antibodies and D^b MHC class I tetramers for three LCMV epitopes: GP33-41, NP396-404, and GP276-286. Throughout acute primary LCMV infection, the hierarchy of CD8⁺ T-cell responses is NP396-404 > GP33-41 > GP276-286 (8, 26). In chronic viral infection, the responding CD8⁺ T cell hierarchy is altered to GP276-286 > GP33-41 > NP396-404 (46) (Fig. 1, panel A). In mutant mice, NP396-404 remained the immunodominant epitope during the deletion phase and, by day 60, comprised almost 16% of CD8⁺ T cells compared to 1.8% in wild-type mice.

FIG. 1.

Chronic viral infection of Bim mutant mice results in alteration of the immunodominance hierarchy and increased numbers of NP396-404-specific CD8⁺ T cells. Wild-type or Bim mutant mice were infected with LCMV clone 13. (A) At the indicated time points, mice were sacrificed, the spleen was removed, and cells were stained with anti-CD8α, anti-CD44, and either D^bGP33-41, D^bNP396-404, or D^bGP276-286 MHC class I tetramer. The numbers in the upper quadrants indicate the percentages of total CD8⁺ T cells that each population comprises. The numbers of activated (B) or antigen-specific (C to E) CD8⁺ T cells were quantitated, and the averages and standard deviations are shown. Five to twelve mice were analyzed at each time point. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

After infection, the number of CD8⁺ CD44^high (activated/memory phenotype) cells increased in both groups of mice but was four- to fivefold higher in mutant mice until day 15 (panel B). Afterwards, the number of cells decreased in both groups and remained relatively constant from day 30 to 111. In this later phase of infection, there were always twice as many CD44^high cells in the mutant mice. When the number of antigen-specific CD8⁺ T cells was determined, the epitopes appeared to fall into two groups. In the first group, which included GP33-41 (panel C)- and GP276-286 (panel E)-specific cells, the number of cells at the peak of the response was very similar between wild-type and mutant mice. In wild-type mice, the number of antigen-specific CD8⁺ T cells declined from day 10 to 30 and was stably maintained for up to 111 days. In mutant mice, cell numbers remained elevated from day 8 to 15 and then began a period of deletion, reaching a nadir at day 30. At this point, numbers were stably maintained. It is important to note that the final set point in mutant mice was only 1.5- to 2-fold higher than in wild-type mice. The behavior of these two epitopes contrasts with that observed for NP396-404-specific cells (panel D). By the time the peak of the response was reached on day 8 postinfection, 4 to 5 times as many effector cells were present in mutant mice. This difference became more pronounced as wild-type cells underwent a significant deletion from day 10 to 30, while the number of NP396-404-specific cells remained constant in mutant mice. Thus, loss of Bim preferentially affects the deletion of some epitope specificities (NP396-404) while leaving others (GP33-41 and GP276-286) less affected.

Bim mutant mice exhibit a delay in the contraction of all CD8 epitopes during acute LCMV infection.

Our results with chronic LCMV infection prompted us to examine whether antigen-specific CD8⁺ T-cell responses would be affected in a similar manner during acute LCMV infection of Bim mutant mice. Briefly, mice were infected with LCMV Armstrong, which induces an acute infection and is cleared from tissues by 7 to 9 days. Animals were sacrificed, and the number of activated and antigen-specific CD8⁺ T cells in the spleen was quantitated (Fig. 2). Regardless of whether activated (panel A) or antigen-specific (panels B to D) cells were examined, CD8⁺ responses reached a maximum on day 8 in wild-type mice and then declined until day 35, where stable numbers were maintained until day 116. This contrasts with Bim mutant mice, where both activated and antigen-specific CD8⁺ T cells increased from day 8 to 15 and then began a slow decline until day 116, where numbers were elevated compared to wild-type mice. It is critical to note that during acute infection, contraction of all three antigen specificities (GP33-41, NP396-404, and GP276-286) was equally affected. Thus, Bim mutation results in a delay in contraction for all epitopes following acute infection, with increased antigen-specific CD8⁺ T cells compared to wild-type mice.

FIG. 2.

Acute viral infection of Bim mutant mice results in increased numbers of antigen-specific CD8⁺ memory T cells. Wild-type or Bim mutant mice were infected with LCMV Armstrong. (A) At the indicated time points, mice were sacrificed, spleens were removed, and cells were stained with anti-CD8α, anti-CD44, and either D^bGP33-41, D^bNP396-404, or D^bGP276-286 MHC class I tetramer. The numbers of activated (A) or antigen-specific (B to D) CD8⁺ T cells were quantitated, and the averages and standard deviations are shown. Five to ten mice were analyzed at each time point. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

Bim mutant mice contain increased numbers of functional NP396-404-specific CD8⁺ T cells during chronic infection.

Although we determined that Bim mutation resulted in greater numbers of some CD8⁺ epitope specificities during chronic infection, it was unclear whether these cells were functional. In Fig. 3 (panel A) we compared the ability of CD8⁺ T cells from clone 13-infected wild-type and mutant mice to produce IFN-γ and TNF-α after peptide stimulation. Previous studies have demonstrated that, during chronic viral infection, there is a progressive loss of function of CD8⁺ T cells and a decrease in their ability to produce both cytokines (12, 46, 47). On day 8 postinfection, stimulation of splenocytes with either GP33-41 or NP396-404 peptide induced a greater percentage of CD8⁺ T cells to make cytokines in Bim mutant mice. This contrasts with GP276-286 or NP205-212 stimulation, where the opposite trend was observed. Following peptide stimulation, increased percentages of CD107a&b⁺ CD8⁺ T cells were also observed in mutant mice (data not shown).

FIG. 3.

Chronic viral infection of Bim mutant mice results in increased numbers of IFN-γ⁺ and TNF-α⁺ NP396-404-specific CD8⁺ T cells. Wild-type or Bim mutant mice were infected with LCMV clone 13. (A) On day 8 postinfection, mice were sacrificed, the spleens were removed, and cells were stimulated with the indicated peptide and then stained with anti-CD8α, anti-IFN-γ, and anti-TNF-α. The dot plots are gated on CD8⁺ T cells, and the number in the plot indicates the percentage of CD8⁺ T cells that are present in that region. (B) Mice were sacrificed at other time points, and the numbers of cytokine-producing CD8⁺ T cells were quantitated. The averages and standard deviations are shown. Five to twelve mice were analyzed at each time point. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

When the number of IFN-γ⁺ and IFN-γ⁺ TNF-α⁺ CD8⁺ T cells was determined, similar groupings of epitopes were observed as for MHC class I tetramer staining. In the first group, which contained cytokine-producing cells after stimulation with GP33-41, GP276-286, or NP205-212 peptide (Fig. 3, panel B), the number of either IFN-γ⁺ or IFN-γ⁺ TNF-α⁺ cells was increased slightly at the peak of the response between days 8 and 10. As the response is shut off from day 10 to day 30, the number of functional cells remained slightly elevated (GP33-41 and NP205-212 stimulation) or was equivalent (GP276-286 stimulation) between wild-type and mutant mice. The behavior of these specificities contrasts with that of NP396-404-specific cells. When the number of IFN-γ⁺ cells was examined, the number of functional cells was always increased during the expansion phase of the response. As the number of functional cells dropped in wild-type mice from day 8 to 60, the number of IFN-γ⁺ cells in mutant mice remained relatively constant. When the numbers of IFN-γ⁺ TNF-α⁺ cells was examined, this trend became even more pronounced. In mutant mice, the number of cells remained constant from day 8 until day 111 postinfection, while in wild-type mice, it decreased significantly. At any given time point during the deletion phase, there were 10- to 100-fold more IFN-γ⁺ TNF-α⁺ cytokine-producing cells after NP396-404 stimulation. Thus, loss of Bim function results in greater numbers of cytokine producing cells after stimulation with some epitopes (NP396-404) while having minimal effects on others (GP33-41, GP276-286, and NP205-212).

Chronic infection of Bim mutant mice results in greater numbers of NP396-404-specific CD8⁺ T cells in the liver.

Prior studies have demonstrated that during viral infections large numbers of virus-specific CD8⁺ T cells can be found in both lymphoid and nonlymphoid tissues. In clone 13 infection of mice, the liver becomes one of the tissues with the largest number of virus-specific CD8⁺ T cells (46). To determine if the differences in deletion were specific to the spleen, we enumerated activated and virus-specific CD8⁺ T cells in the liver during chronic viral infection. When the number of activated/memory CD44^high CD8⁺ T cells was determined (Fig. 4, panel A), similar numbers were observed at the peak of the response on day 8 postinfection. As the response decreased, the number of cells was very similar between wild-type and mutant mice. When the number of antigen-specific CD8⁺ T cells was determined by MHC class I tetramer staining, the epitopes again fell into two groups. In the first group, which contained GP33-41 (panel B)- and GP276-286 (panel D)-specific CD8⁺ T cells, very little difference was observed between wild-type and mutant mice except for day 60 postinfection, when the number of cells was twofold greater in mutant mice. By day 111 postinfection, this difference was no longer apparent. This contrasted with the numbers of NP396-404-specific CD8⁺ T cells. At all time points examined, the number of cells was elevated in mutant compared to wild-type mice and, at its final set point, was almost 20-fold greater in mutant mice. Thus, Bim mutation preferentially increases the number of NP396-404-specific CD8⁺ T cells in nonlymphoid organs, such as the liver.

FIG. 4.

Chronic viral infection of Bim mutant mice results in increased numbers of NP396-404-specific CD8⁺ T cells in the liver. Wild-type or Bim mutant mice were infected with LCMV clone 13. At the indicated time points, mice were sacrificed, livers were removed, and cells were stained with anti-CD8α and anti-CD44 (A) and either D^bGP33-41 (B), D^bNP396-404 (C), or D^bGP276-286 (D) MHC class I tetramer. The numbers of activated (A) or antigen-specific (B to D) CD8⁺ T cells were quantitated, and the averages and standard deviations are shown. Five to twelve mice were analyzed at each time point. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

Bim mutation does not increase the number of activated or functional CD4⁺ T cells.

CD4⁺ T cells play a critical role in controlling and shaping the immune response. These cells play a particularly important role in the generation and maintenance of CD8⁺ memory T cells following infection (39, 41, 42). Prior studies have demonstrated that CD4⁺ T cells also undergo a progressive loss of function and deletion during chronic LCMV infection (12). To determine the effect of Bim mutation, we determined the number of activated/memory CD44^high CD4⁺ T cells in the spleen. At the peak of infection (Fig. 5, panel A) similar numbers of activated CD4⁺ T cells were detected in both strains of mice, and these numbers were maintained throughout infection. When the number of functional cells was determined after GP61-80 peptide stimulation (panels B and C), similar numbers were detected at all times postinfection, except for day 30, when approximately 1.5- to 2-fold more functional cells were detected. It is important to note that, by day 60, similar numbers of cells were detected, and this was maintained throughout the infection. Responses for NP309-328-specific CD4⁺ T cells were below the limit of detection at all time points (data not shown). Thus, Bim mutation has minimal effects on the number of activated or antigen-specific CD4⁺ T cells during chronic viral infection.

FIG. 5.

Bim mutation does not affect CD4⁺ T-cell responses during chronic viral infection. Wild-type or Bim mutant mice were infected with LCMV clone 13. At the indicated time point postinfection, mice were sacrificed, spleens were removed, and cells were either stained with anti-CD4 and anti-CD44 antibodies (A) or stimulated with GP61-80 peptide and then stained with anti-CD4 anti-IFN-γ and anti-TNF-α (B and C). The number of cytokine-producing CD4⁺ T cells was quantitated, and the averages and standard deviations are shown. Five to twelve mice were analyzed at each time point. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

Accelerated viral clearance during chronic infection of Bim mutant mice.

To determine whether the increased numbers of CD8⁺ T cells could lead to decreased viral load, we examined viral titers in the serum and several organs. We observed that both wild-type and mutant mice achieved a high serum viremia (Fig. 6, panel A) that peaked between days 8 and 10 postinfection and gradually declined. By day 30, titers were higher in mutant mice than in wild-type mice, but by day 60, virus levels dropped below the limit of detection in mutant mice but remained elevated in wild-type mice. By day 111, there was no detectable virus in the sera of either mouse strain. When the spleen was examined (panel B), both strains of mice had a high level of viral replication on day 8 postinfection. This dropped slightly by day 15 and remained comparable in Bim mutant mice at day 30 but dropped below the limit of detection by day 60. This contrasted with wild-type mice, where virus was still present at day 60, but by day 111, levels were below the limit of detection. When levels of virus were determined in the liver (panel C), both groups had a very high level of virus at the peak of the response, but this gradually dropped and was cleared from the mice by day 60. This contrasts with results from the kidney. In this organ, high titers were observed on day 8 in both groups. Over 111 days, the overall level of virus declined, but even at day 111, high levels of virus could be detected in both groups. Thus, accelerated viral clearance is observed in the serum and spleen of Bim mutant mice, while no effect was observed in the kidneys and liver.

FIG. 6.

Clone 13 infection generates high levels of virus in Bim mutant mice. Wild-type or Bim mutant mice were infected with LCMV clone 13. At the indicated time point, mice were sacrificed and the serum (A), spleen (B), liver (C), and kidneys (D) were isolated, and virus was quantitated by plaque assay. Each symbol indicates the value for one mouse. Five to twelve mice were analyzed at each time point. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

Antigen-specific CD8⁺ T cells from Bim mutant mice enter the early stages of apoptosis during chronic viral infection.

The number of antigen-specific CD8⁺ T cells during chronic viral infection is controlled by rates of cell death and proliferation. To determine the contribution of apoptosis, we stained cells with annexin V and 7-AAD. Cells that are in the early part of apoptosis are annexin V^high and 7-AAD⁻. Those that are in the final stages are annexin V^high and 7-AAD⁺. In Fig. 7, we examined activated/memory phenotype and virus-specific CD8⁺ T cells and determined the portions of cells that were in the various stages of apoptosis. CD8⁺ CD44^low cells are viable and serve as a control (panel A). The majority of these cells (∼96%) are annexin V^low and 7-AAD⁻, while only 0.3% are annexin V^high and 7-AAD⁺. When CD8⁺ D^bNP396-404⁺ T cells were examined on day 15 postinfection, the vast majority of cells were annexin V^hi and 7-AAD⁻ (85.5% +/+ versus 88.5% −/−). Similar results were obtained when activated (CD44^high) or other antigen-specific CD8⁺ T cells were examined (panels B to E). Thus, Bim mutation does not prevent cells from entering the early stages of apoptosis.

FIG. 7.

Antigen-specific CD8⁺ T cells from Bim mutant mice enter the early stages of apoptosis. Wild-type or Bim mutant mice were infected with LCMV clone 13. On day 15 postinfection, mice were sacrificed, spleens were removed, and cells were either stained with anti-CD8 and anti-CD44 (A) or MHC class I tetramer (B to E). A representative dot plot is presented for D^bNP396-404⁺ CD8⁺ T cells. Direct ex vivo apoptosis was assessed by staining with annexin V and 7-AAD. The numbers in each dot plot indicate the percentage of each population that falls into that region. The numbers of cells that fall into each quadrant were determined for six mice and are plotted as averages and standard deviations. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

Antigen-specific CD8⁺ T cells from Bim mutant mice undergo less proliferation during the deletion phase than those from wild-type mice.

In addition to apoptosis, cell numbers are also controlled by proliferation. To determine the contribution of proliferation, we injected mice with BrdU from day 15 to 25 of infection during the deletion phase. Background staining for mice that had received PBS injection was ∼3%. In wild-type mice, 92% of CD8⁺ D^bNP396-404⁺ cells were BrdU⁺ compared with 25% in mutant mice (Fig. 8, panel A). Examination of CD8⁺ CD44^high (activated/memory phenotype) cells revealed ∼2.6-fold-higher BrdU incorporation in wild-type mice than in Bim mutant mice (panel B, 80% [+/+] versus 30% [−/−]). When BrdU incorporation in antigen-specific CD8⁺ T cells was determined, results similar to those for the CD44^high cells were observed. For all three epitopes (GP33-41, NP396-404, and GP276-286), increased incorporation was observed in wild-type mice compared to Bim mutant mice. Thus, these data combined with our annexin V staining demonstrate that the increase in NP396-404-specific CD8⁺ T cells is not due to excess proliferation but a failure of these cells to progress from early apoptosis to cell death.

FIG. 8.

Reduced proliferation of antigen-specific CD8⁺ T cells from Bim mutant mice during the deletion phase. Wild-type or Bim mutant mice were infected with LCMV clone 13. Proliferation during the deletion phase was assessed by including BrdU (0.8 mg/ml) in the drinking water of mice from day 15 to day 25 after infection. Splenocytes were isolated on day 25 postinfection and stained with anti-CD8α, D^bNP396-404, and anti-BrdU antibody. The plots are gated on CD8⁺ D^bNP396-404⁺ cells, and the number on the dot plot indicates the percentage of antigen-specific cells that are BrdU⁺. To determine the proliferation in other populations, splenocytes were stained with either anti-BrdU and anti-CD8α antibodies (A) or anti-CD44 antibodies (B) and D^bGP33-41, D^bNP396-404, or D^bGP276-286 MHC class I tetramers. The percentages of BrdU⁺ cells were determined and plotted as averages and standard deviations. Five mice were analyzed for each genotype. *, significant difference between wild-type and Bim mutant mice, with a P value of ≤0.05.

DISCUSSION

In this study, we have determined the effects of Bim on CD8⁺ T-cell apoptosis during chronic viral infection. Specifically, loss of Bim results in maintenance of NP396-404-specific CD8⁺ T cells, while in wild-type mice, these cells undergo a major reduction in numbers. This contrasts with GP33-41- and GP276-286-specific CD8⁺ T cells, in which less of an effect was observed. In addition to documenting increased numbers of antigen-specific CD8⁺ T cells, we also demonstrate that loss of Bim results in a major increase only in NP396-404-specific IFN-γ- and TNF-α-producing cells. Increases in CD8⁺ D^bNP396-404⁺ T cells were not restricted to the spleen, as we determined they also occurred in the liver. Unlike CD8⁺ T cells, no increase in activated phenotype or antigen-specific cells was observed for CD4⁺ T cells. Failure to undergo deletion was not due to reduced viral replication because both Bim wild-type and mutant mice had very high viral loads; however, viral clearance was slightly more rapid in mutant mice. Finally, we observed that, during the deletion phase, a similar portion of antigen-specific CD8⁺ T cells were in the early stages of apoptosis, but these cells were undergoing less proliferation in Bim mutant mice than in wild-type mice.

What are the implications of our study for antiviral immune responses? Prior studies have demonstrated that during acute LCMV Armstrong infection, CD8⁺ T cells become activated and undergo massive proliferation such that 50% of the CD8⁺ T population can be LCMV specific (26). At the peak of the expansion phase, many of the antigen-specific CD8⁺ T cells become Bcl-2 (15) and IL-7 receptor α (21) low and are susceptible to apoptosis. Indeed, after the peak of expansion, a “contraction” phase lasting 2 to 3 weeks ensues. Contraction has been documented to occur for multiple infections, including LCMV (26), vaccinia virus (17), vesicular stomatitis virus (38), simian virus 5 (14), herpes simplex virus (33), and Listeria monocytogenes (7). At the conclusion of the contraction phase, the remaining antigen-specific CD8⁺ T cells are memory cells that are Bcl-2 and IL-7 receptor α high and can provide an accelerated secondary response. The fate of antigen-specific CD8⁺ T cells following acute infection contrasts with that during chronic viral infection. Following an initial expansion phase, high levels of virus lead to silencing of the immune response by functional inactivation (anergy) or death (deletion) of antigen-specific CD8⁺ T cells (47). This has important implications for human infection. In human immunodeficiency virus infection, after the peak of primary viremia, the expanded pool of antigen-specific CD8⁺ T cells declines and reaches a reduced set point (23, 31). A positive correlation between the number of antigen-specific CD8⁺ T cells and the time to disease progression has been observed (32). The death of antigen-specific CD8⁺ T cells that occurs during chronic viral infection has been termed deletion (47). Prior to this study, the extent of overlap between the molecular mechanisms of contraction and deletion was unclear.

During acute viral infection, where antigen exposure is more limited, it appears that intrinsic pathways of apoptosis predominate. There are three major pieces of evidence that support this hypothesis. First, acute LCMV infection of lpr or gld mice results in a relatively normal expansion of antigen-specific CD8⁺ T cells, followed by contraction to normal levels (25, 36). When TNF receptor I (TNFRI) and Fas were examined in adoptive transfer experiments of transgenic CD8⁺ T cells, no effect on contraction was observed either as single or double mutants (28, 37). Death receptors are also not required for the contraction following herpes simplex virus infection of mice (33) or after administration of staphylococcal enterotoxin B superantigen. A third piece of evidence is that contraction occurs even when pathogen can be cleared by antibiotics, suggesting that repetitive antigen stimulation is not required for this process (3). While death receptors appear to play a lesser role in contraction, there is an absolute requirement for Bim in the contraction process. Following herpes simplex virus infection, antigen-specific CD8⁺ T cells fail to undergo contraction in numbers for up to 38 days in mutant mice (33). Additionally, deletion of cells expanded by staphylococcal enterotoxin B superantigen is dependent upon Bim (18). During acute LCMV Armstrong infection, contraction is also altered such that initial contraction is blocked and the number of antigen-specific CD8⁺ T cells for all epitope specificities is ∼4-fold greater, even at day 116 postinfection. This equal effect across all epitopes is different from chronic LCMV infection, where the number of NP396-404 specific T cells is affected to the greatest extent. How Bim facilitates the contraction process is unclear because treatment with caspase inhibitors does not block contraction of cells following LCMV infection (29).

In chronic viral infection, where virus persists at high levels for extended periods, death receptors would be predicted to play a role because of activation-induced cell death (48). This process has been shown to be dependent on death receptors, including Fas. Typically, cells are activated in vitro and then, after a period of proliferation, are exposed to T-cell receptor stimulation again and undergo apoptosis. In vivo this can occur during high-level infection or after repeated administration of antigen. Infection of mice with clone 13 results in high levels of virus for up to 60 to 90 days in the spleen, liver, and lung. In some organs, such as the brain and kidneys, virus persists indefinitely. Infection of Bim mutant mice resulted in high viral titers in all tissues examined at early time points, similar to wild-type mice. In the later phase of infection on day 30, titers were slightly higher in the spleens and sera of mutant mice than in those of wild-type mice. This may have been due to reduced apoptosis of virus-producing cells in Bim mutant mice. By day 60, however, the situation had reversed and virus was below the detection limit in mutant mice but still detectable in the sera and spleens of wild-type mice. Repeated exposure to viral antigen may cause T cells to undergo apoptosis. Infection of TNFRI and II double mutants with clone 13 results in a delay in deletion, but eventually, the number of antigen-specific CD8⁺ T cells declines for all epitope specificities, including NP396-404 (43). Additionally, infection of lpr mice with strains that induce persistently high levels of virus in the spleen, such as docile and aggressive, demonstrated a delay in deletion, but eventually, all epitopes reach a level similar to that of the wild type (49). This contrasts with results from Bim-deficient mice where NP396-specific CD8⁺ T cells remain elevated for extended periods but other specificities reach a set point similar to that of wild-type mice. The slight increase in the number of GP33-41- and GP276-286-specific CD8⁺ T cells may be due to the fact that naive Bim mice contain ∼2.5 times as many naive CD8⁺ T cells, which could increase the precursor frequency for any epitope slightly. Taken together, our studies and the ones mentioned above suggest that deletion of CD8⁺ T cells during chronic viral infection is a multilayered process that may require both death receptors and Bim.

Why is the apoptosis of CD8⁺ D^bNP396-404⁺ T cells blocked to the greatest extent? Previous studies have demonstrated that, with supraoptimal peptide stimulation, cytotoxic T lymphocytes can undergo a TNFRII-dependent apoptosis that is accompanied by decreases in Bcl-2 (2). Bim was identified through its interaction with Bcl-2 (19, 30) and has been shown to be complexed with Bcl-2 in the mitochondria of activated T cells (50). Repeated stimulation of NP396-specific CD8⁺ T cells could activate TNFRII-mediated apoptosis that does not proceed in the absence of Bim because Bcl-2 is free to exert its antiapoptotic function. It is important to note that not only did absence of Bim prevent deletion of NP396-404⁺ T cells, it resulted in the surviving cells retaining function. This suggests that the mechanisms that induce cell death may be linked to those that modulate anergy. In the repetitive stimulation model, antigen exposure determines the initiation of apoptosis, but it requires Bim for completion. This would imply that NP396-404 peptide is present in vivo at higher concentrations than GP33-41 or GP276-286. There are three pieces of evidence to support this conjecture. First, because nucleoprotein mRNA accumulates before glycoprotein mRNA, the NP396-404 epitope is presented first (13). Second, in clone 13 infection of LCMV Armstrong immune mice, NP396-404-specific memory cells are recruited into responses to a greater extent than other epitope-specific cells (45). Third, splenocytes from clone 13-infected mice are able to stimulate CD8⁺ D^bNP396-404⁺ T cells to proliferate to a much greater extent than D^bGP3-41⁺- or D^bGP276-286⁺-specific cells (46). In this model there is less of an effect on GP33-41- and GP276-286-specific CD8⁺ T cells due to decreased presentation of the epitope.

In conclusion, we demonstrate that Bim mutation results in maintenance of CD8⁺ D^bNP396-404⁺ T cells during chronic viral infection. This effect is observed in both the spleen and liver and is accompanied by slightly accelerated viral clearance. These results suggest that targeting of Bim with small-molecule therapeutics may allow immunodominant responses to be maintained in diseases where antigen persists, such as chronic viral and bacterial infections.