The acute phase of many viral infections is associated with the induction of a pronounced CD8 T cell response which plays a principle role in clearing the infection. By contrast, certain infections are not as readily controlled. In this study, we have used the well-defined system of lymphocytic choriomeningitis virus (LCMV) infection of mice to determine quantitative and qualitative changes in virus-specific CD8 T cell responses that rapidly resolve acute infections, more slowly control protracted infections, or fail to clear chronic infections. Acute LCMV infection elicits potent, functional, multi-epitope-specific CD8 T cell responses. Virus-specific CD8 T cells also expand, albeit to a lesser extent, during protracted LCMV infection. Under these conditions, there is a progressive diminution in the capacity to produce IL-2, TNF-α, and IFN-γ. Changes in cytotoxic activities are also detectable but differ depending upon the specificity of the responding cells. As the infection is slowly resolved, a resurgence of cytokine production by virus-specific CD8 T cells is observed. CD4-deficient mice cannot control infection with certain strains of LCMV, but do mount multi-epitope-specific CD8 T cell responses that also lose effector capabilities; however, they are not maintained indefinitely in an unresponsive state as these cells become deleted over time. Overall, our findings suggest that constant high viral loads result in the progressive diminution of T cell effector functions and subsequent physical loss of the responding cells, whereas if the viral load is brought under control a partial restoration of CD8 T cell functions can occur.
The induction of an appropriate adaptive immune response is critical for the resolution of viral infections, and maintenance of immunological memory is necessary for long-term protection against viral re-exposure. In recent years, it has become widely accepted that an overwhelming CD8 T cell response is elaborated during the initial phase of many viral infections (1, 2, 3, 4, 5, 6, 7, 8). As naive virus-specific CD8 T cells encounter presented Ag, they proliferate and differentiate as they attain phenotypic and functional attributes which enable them to play a principle role in controlling the infection (9, 10). Activation-associated changes include alterations in cell surface markers, the production of effector cytokines, as well as the generation of cytolytic granules that enable virus-specific CD8 T cells to directly kill infected targets (5, 8, 11, 12). Following the successful resolution of the infection, CD8 T cell responses do not remain in a persistently activated state. Instead, a contraction phase ensues during which the majority of virus-specific CD8 T cells die and subsequently a pool of resting memory cells emerges. Long-term immunological protection is conferred by the physical presence of elevated numbers of virus-specific CD8 T cells in both lymphoid and nonlymphoid organs as well as by the capacity of these cells to rapidly elaborate effector activities following re-exposure to viral Ag (5, 6, 8, 13). Thus, a transient infection results in the induction and maintenance of an exceptionally effectual virus-specific CD8 T cell response.
Although numerous infections are successfully cleared by the host’s immune response, certain viral infections are only slowly or never resolved. This inability to rapidly control an infection suggests that the immune response is ineffective. Certain viruses can deploy immune evasion strategies or confine replication to immunoprivileged sites (14, 15). Furthermore, although CD8 T cells can usually exquisitely target cells that harbor replicating virus, the defective induction or maintenance of these responses can be detrimental for clearing the infection. Given the public health importance of chronic infections, defining how and why antiviral immune responses become dysregulated and fail to resolve viral infections may facilitate the development of improved therapies and vaccines to combat these pathogens.
CD4 T cells also play a prominent role in the elaboration of robust antiviral immune responses. These cells activate dendritic cells, help both B and T cells, can acquire cytotoxic effector functions, and can also negatively regulate immune activities (5, 16, 17). Notably, many persistent viral infections are better controlled in hosts that elaborate more vigorous CD4 T cell responses (18, 19, 20, 21, 22). CD4 T cells are required for the development of functionally robust CD8 T cells and also serve to sustain CD8 T cell effector functions in chronically infected hosts (23, 24, 25, 26, 27, 28, 29). Thus, inadequate CD4 T cell help may impair CD8 T cell responses resulting in difficulties in controlling the infection that consequently further promotes virus-persistence.
How CD8 T cell responses are impacted by changing viral loads, as well as the long-term consequences of sustained exposure to high viral burdens, is not well understood. In this study, we have used the well-defined system of lymphocytic choriomeningitis virus (LCMV)3 infection of mice to determine quantitative and qualitative changes in virus-specific CD8 T cell responses that rapidly resolve acute infection, more slowly control protracted infection, or that fail to clear chronic infection (30). We document that protracted and chronic infections result in the step-wise inactivation of virus-specific CD8 T cell responses; however, if viral loads can be brought under control, functionally competent CD8 T cells reemerge. By contrast, the inability to resolve the infection is associated with the physical loss of all specificities of antiviral CD8 T cells.
Materials and Methods
Mice and virus
C57BL/6J (B6), C57BL/6-Cd4tm1Mak (CD4−/−) (31), and B6.MRL-Tnfrsf6lpr/J (B6-lpr) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Animals were bred and maintained in accredited facilities at the University of Alabama (Birmingham, AL). Male and female mice between 6 and 10 wk of age were used.
The Armstrong and clone 13 strains of LCMV were kindly provided by Dr. R. Ahmed (Emory University, Atlanta, GA) (32). Plaque-purified viral isolates were propagated in BHK-21 cells. To establish acute infections, B6 mice received 2 × 105 PFU LCMV-Armstrong in a volume of 0.5 ml by i.p. inoculation. Protracted and chronic infections were established by i.v. inoculation with 4 × 106 PFU LCMV-clone 13 in a volume of 0.5 ml. The titers of viral stocks, serum samples, and tissue homogenates were determined by plaque assay (32).
Livers and spleens were explanted from mice following perfusion with 10 ml of PBS. Spleens were disrupted into single-cell suspensions and erythrocytes removed by lysis using 0.83% (w/v) NH4Cl (33). Livers were disrupted into single-cell suspensions as previously described (34). Viable liver lymphocytes were collected by centrifugation over a layer of Histopaque-1083 (Sigma-Aldrich, St. Louis, MO). After washing, cell preparations were finally resuspended in RPMI 1640 medium supplemented with 10% FCS, 50 μM 2-ME, 100 U/ml penicillin, and 100 μg/ml streptomycin. Bone marrow was collected by flushing femurs with RPMI 1640 medium, supplemented as above but containing only 1% FCS (35). Erythrocytes were removed by treatment with 0.83% (w/v) NH4Cl, and after washing, cells were finally resuspended as described above.
Preparation of MHC class I tetramers
MHC class I tetramers were prepared as previously described (33). Briefly, monomeric MHC-peptide complexes were refolded with defined peptide epitopes and subsequently enzymatically biotinylated using BirA. Tetramers were formed by the step-wise addition of allophycocyanin-conjugated streptavidin (Molecular Probes, Eugene, OR).
MHC class I tetramer staining
Suspensions of spleen, bone marrow, and liver cells were pretreated with anti-CD16/CD32 mAb (clone 2.4G2). Costaining was then performed with either anti-CD44-FITC (clone IM7) or anti-CD43-FITC (clone 1B11) mAbs (BD PharMingen, San Diego, CA) together with allophycocyanin-conjugated MHC class I tetramers and anti-CD8-PE mAbs. The anti-CD8-PE clone 53-6.7 (BD PharMingen and eBioscience, San Diego, CA) was used in conjunction with H-2Db tetramers. For costains with H-2Kb tetramers, the anti-CD8-PE mAb clone, CT-CD8a (Caltag Laboratories, Burlingame, CA), was used. Staining procedures were performed at 4°C in PBS containing 2% (w/v) BSA and 0.2% (w/v) NaN3. After incubation with Abs, samples were washed and then fixed in PBS containing 2% (w/v) paraformaldehyde.
Intracellular cytokine staining
Intracellular cytokine staining was performed as previously described (33). Briefly, cells were either left untreated or stimulated with LCMV-derived peptide epitopes (1 μg/ml for H-2Db or Kb epitopes and 10 μg/ml for I-Ab epitopes) for 5–6 h at 37°C. The intracellular accumulation of cytokines was facilitated by the addition of brefeldin A for CD8 T cell responses (Golgi-plug; BD PharMingen) or monensin for CD4 T cell responses (Golgi-stop; BD PharMingen). Surface and intracellular staining were then performed using the mAbs anti-CD8-PE (clone 53-6.7), anti-CD4-PE (clone RM4.5), anti-IL-2-FITC (clone JES6-5H4), anti-IFN-γ-FITC or -ALLOPHYCOCYANIN (clone XMG1.2), and anti-TNF-α-ALLOPHYCOCYANIN (clone MP6-XT22). Conjugated mAbs were purchased from either BD PharMingen or eBioscience.
In vivo CTL assays
Splenocytes from naive B6 mice or B6-lpr mice were prepared as previously described and used as target cells for in vivo CTL assay (36, 37). Splenocytes (107 cells/ml) were labeled with either 2 or 0.2 μM CFSE for 5 min at 25°C in PBS. CFSE labeling was then quenched by the addition of FCS to a final concentration of 20% (v/v). Splenocytes labeled with 2 μM CFSE were then sensitized for lysis by pulsing with gp33 peptide (1 μg/ml) for 1 h at 37°C. Splenocytes labeled with 0.2 μM CFSE were not peptide-pulsed and used as a control population of target cells. To determine nucleoprotein (np) 396-specific killing, separate populations of splenocytes suspended in RPMI 1640 medium supplemented with 10% FCS (5 × 106 cells/ml) were labeled with either 5 or 0.5 μM CellTracker Orange (CTO; Molecular Probes, Eugene, OR) for 1 h at 37°C. Cells labeled with 5 μM CTO were concurrently pulsed with the LCMV np396 peptide (1 μg/ml). After labeling and peptide pulsing, all populations of cells were washed and resuspended in HBSS without calcium and magnesium. All four populations of target cells were then mixed together such that recipient mice received 107 of each population in a single i.v. injection. Target cells were administered to naive B6 and CD4−/− mice, and to mice at various time points following acute, protracted, or chronic LCMV infections. Recipient mice were sacrificed 8 h following cell transfer and single-cell suspensions of spleens and livers were prepared as described above. Flow cytometry was used to determine the recoveries of peptide-sensitized and non-peptide-treated dye-labeled target cells. Detection of CFSE- and CTO-labeled cells was used to determine gp33- and np396-specific killing, respectively. The percentage of specific lysis was determined using the following formulas (36): ratio of recovery of non-peptide-treated control splenocytes to peptide-sensitized splenocytes = (percentage of CFSElow or CTOlow cells)/(percentage of CFSEhigh or CTOhigh cells); percent specific lysis = 100 × [1 − (ratio of cells recovered from naive mice/ratio of cells recovered from infected mice)].
Statistical analysis was performed using the Student’s two-tailed t test. Statistical significance was defined as p < 0.05.
Establishment of acute, protracted, and chronic LCMV infections
In this study, we used the well-defined system of LCMV infection of mice to assess the impact of viral load and duration of infection on virus-specific T cell responses. We have adopted the nomenclature acute, protracted, and chronic to describe the different durations of infection. As expected, an acute infection ensued following inoculation of adult B6 mice with the Armstrong strain of LCMV. This infection was rapidly resolved, and by 8 days after inoculation, viral titers were below the limit of detection in the serum and liver and the majority of mice had resolved the infection from the lungs (data not shown). Infection of adult B6 mice with the clone 13 strain of LCMV results in a widely disseminated infection (38, 39). High viral titers were initially observed in the serum, liver, and lungs; however, titers subsided between days 16 and 60 as the infection was brought under control (Fig. 1, upper panels). We have used the term protracted to describe this type of infection that is more gradually resolved. A chronic infection was established following inoculation of CD4−/− mice with LCMV-clone 13. Under these conditions, a lifelong infection with persistent viral shedding ensued and viral titers remained high in the serum, liver, and lungs at all time points tested (Fig. 1, lower panels).
Maintenance, loss, and resurgence of CD8 T cell effector functions
We determined how the cytokine-producing capacity of virus-specific CD8 T cells changed during the course of acute, protracted, and chronic infections. Virus-specific CD8 T cells that produced both IFN-γ and TNF-α became readily detectable following acute LCMV infection (Fig. 2,A). Cytokine-producing CD8 T cells were also elicited in protractedly infected hosts, but these responses were less robust than those elicited by acute infection (Fig. 2, A and B). Notably, the majority of the virus-specific cells failed to produce TNF-α. As the protracted infection progressed, the capacity to produce IFN-γ was further reduced and TNF-α production was almost completely abolished (Fig. 2,B). Nevertheless, sufficient functional activity was retained to bring the infection under control (Fig. 1, upper panels). At later time points, the clearance of the replicating virus was associated with the reacquisition of cytokine-producing activities as IFN-γ- and also TNF-α-producing CD8 T cells were again detected (Fig. 2 B).
Virus-specific CD8 T cell responses were also measurable in chronically infected CD4−/− mice (Fig. 2,C). By 8 days postinfection, similar patterns of IFN-γ and TNF-α production were apparent in both protractedly and chronically infected hosts (Fig. 2, B and C). In chronically infected hosts, no restoration of cytokine production was detected; instead the failure to control the virus resulted in a step-wise and progressive abolition of cytokine production.
IL-2 production by splenic virus-specific CD8 T cells was also evaluated at 8 days following acute, protracted, and chronic LCMV infections (Fig. 3). In acutely infected hosts, IL-2-producing CD8 T cells were detectable; however, the numbers of these cells was significantly lower following protracted and chronic infection (p < 0.01). Depending upon the epitope recognized, a 2- to 50-fold lower number of IL-2-producing CD8 T cells developed in protractedly infected hosts (Fig. 3, ▪ and □). By contrast, in chronically infected mice barely detectable numbers of IL-2-producing CD8 T cells were observed for all specificities examined (Fig. 3,▩).
Dynamics of virus-specific CD8 T cell responses
To further define the kinetics of virus-specific CD8 T cell responses during the course of acute, protracted, and chronic infections, we used both MHC class I tetramer staining and additional intracellular cytokine analysis. Representative tetramer staining profiles are shown in Fig. 2, D–F, and Fig. 4 summarizes the longitudinal analysis. During acute LCMV infection, virus-specific CD8 T cells were clearly discernable by tetramer staining (Fig. 2,D) and distinct expansion and contraction phases were apparent (Fig. 4 A). Memory virus-specific CD8 T cells that developed were maintained at remarkably stable levels and there was good correlation between the numbers of tetramer-binding and cytokine-producing CD8 T cells.
During both protracted and chronic LCMV infections, virus-specific CD8 T cells initially expanded, albeit to a lesser extent than that observed in acutely infected hosts (Figs. 2, E and F and 4, B and C). In protractedly infected hosts, a discordance between the numbers of tetramer-binding and cytokine-producing virus-specific CD8 T cells was initially discernable and became more pronounced during the first month after infection. Not all virus-specific CD8 T cells attained the ability to produce IFN-γ and few, if any, produced TNF-α by day 29 postinfection (Figs. 2,E and 4,B). Absolute numbers of virus-specific CD8 T cells in the spleen also declined during this period, but the degree of contraction differed depending on the epitope specificity of the responding cells. The down-regulation was most marked for np396- and np205-specific responses. The control of clone 13 infection in B6 mice, which occurred over a period of ∼2 mo (Fig. 1, upper panels), was associated with a resurgence in the cytokine-producing capacity of these cells (Figs. 2, B and E, and 4 B). By 184 days following protracted infection of B6 mice, the proportion of gp33-, np396-, and gp276-specific CD8 T cells producing IFN-γ was significantly higher than that observed at 29 and 55 days postinfection (p < 0.05). The reacquisition of cytokine production was more apparent for gp33- and gp276-specific responses and in each case IFN-γ production was preferentially restored. As the infection was resolved, the modest increase in the overall numbers of tetramer-binding virus-specific CD8 T cells was not statistically significant.
In chronically infected CD4−/− hosts, a discordance between physical presence and functional activity of virus-specific CD8 T cells was also observed (Figs. 2, C and F, and 4,C). The responding cells failed to attain the capacity to produce IL-2 (Fig. 3,▩), a rapid reduction of TNF-α production was observed, and by 3 mo after infection IFN-γ production was extinguished for all epitopes tested (Fig. 4,C). Longitudinal analysis revealed that virus-specific CD8 T cells were not indefinitely maintained in chronically infected hosts. The numbers of virus-specific CD8 T cells declined over time; however, the rate of attrition differed depending on the viral epitope recognized. np396- and np205-specific T cells were rapidly lost, but gp276-specific T cells were deleted more slowly (Figs. 2,F and 4 C).
In acutely infected hosts, virus-specific CD8 T cells acquired and maintained expression of high levels of the surface molecule CD44 (Fig. 2,D). Following both protracted and chronic LCMV infection, the levels of expression of CD44 on virus-specific CD8 T cells were initially high but then declined slightly as CD8 T cells expressing intermediate levels of CD44 (CD44int) developed (Fig. 2, E and F). In chronically infected CD4−/− hosts, these CD44int cells were deleted overtime, whereas in protractedly infected hosts, CD44high virus-specific CD8 T cells reemerged as the infection was controlled.
CD8 T cell resurgence in the bone marrow
Examination of the bone marrow from protractedly infected hosts also revealed changes in the abundance and functional activities of virus-specific CD8 T cells (Fig. 5). Virus-specific CD8 T cells were detectable by tetramer staining at 8 days postinfection; however, the majority failed to produce IFN-γ and even fewer produced TNF-α. By day 16 postinfection, changes in the pattern of immunodominance were apparent as only low levels of np396-specific CD8 T cells were detected, but gp33- and gp276-specific T cells were maintained at higher levels, albeit in a functionally impaired state. By 2 mo postinfection, np396-specific responses remained ablated while the proportion of functionally competent gp33- and gp276-specific T cells increased. This restoration of effector functions was further evident by 139 days postinfection and at this time point the proportion of IFN-γ- and TNF-α-producing gp33- and gp276-specific CD8 T cells was significantly higher than that observed at day 16 postinfection (p < 0.05).
Alterations in CD43 (1B11) expression are associated with changing viral loads
The Ag-driven activation of CD8 T cells is associated with the up-regulation of the 1B11 isoform of CD43 (40, 41). Flow cytometric analysis was performed to determine how CD43 expression changes on virus-specific CD8 T cells during acute, protracted, and chronic LCMV infection (Fig. 6). In acutely infected hosts, CD43 was up-regulated on virus-specific CD8 T cells by 8 days after infection (Fig. 6, C and D, ▪). CD43 expression was then rapidly down-regulated on all virus-specific CD8 T cells as the acute infection was resolved. Following protracted and chronic infection, virus-specific CD8 T cells in the spleen and bone marrow also acquired a CD43high phenotype. Over time, CD43 expression on virus-specific CD8 T cells in protractedly infected hosts gradually declined (Fig. 6, C and D, □), and this reduction in expression was coincident with the resolution of infection and the rebound of cytokine-producing activities (Figs. 1, 4,B, and 5). In chronically infected CD4−/− mice, CD43 expression was not down-regulated and by day 35 postinfection, the levels of CD43 expression were significantly higher on gp33-, np396-, and gp276-specific CD8 T cells in chronically infected mice compared with acutely infected hosts (p < 0.05). These high levels of expression were sustained until the virus-specific CD8 T cells succumbed to deletion and became undetectable by tetramer staining (Fig. 6, C and D,▩).
Functional impairment and deletion of CD8 T cells in the liver
Virus-specific CD8 T cell responses were compared in the livers of mice undergoing acute, protracted, or chronic LCMV infection. A pronounced multi-epitope-specific CD8 T cell response was elicited during the initial phase of each type of infection (Fig. 7). During acute infection, the overall kinetics of the virus-specific CD8 T cell response in the liver resembled that observed in the spleen (Figs. 7,A and 4 A). Distinct expansion and contraction phases were apparent and virus-specific CD8 T cells subsequently remained detectable for over 1 year after infection. CD8 T cells in the liver of acutely infected hosts also produced IFN-γ and TNF-α; however, unlike their splenic counterparts, cytokine production could not necessarily be ascribed to all of the virus-specific T cells present.
In protractedly infected mice, virus-specific CD8 T cell numbers were maximal ∼1 mo after infection, which was later than that observed during acute infection (Fig. 7,B). In protractedly infected mice few, if any, of the virus-specific CD8 T cells produced TNF-α, and IFN-γ production was lost over time. Whereas in the spleens of protractedly infected hosts, virus-specific CD8 T cells regained the capacity to produce cytokines as the infection was brought under control (Fig. 4,B), no concomitant rebound of cytokine production was observed in the liver (Fig. 7 B).
Virus-specific CD8 T cells were detectable by MHC tetramer staining in the livers of chronically infected CD4−/− mice (Fig. 7 C); however, few, if any, of these cells registered in assays for TNF-α and IFN-γ production. As in the spleens of chronically infected mice, the number of virus-specific CD8 T cells also declined over time. By comparison with acutely infected hosts, the differential expansion and deletion of virus-specific CD8 T cells in the spleens and livers of protractedly and chronically infected hosts resulted in an altered pattern of epitope immunodominance. During protracted and chronic infections, gp276-specific T cells predominated, whereas the number of np396- and np205-specific cells declined more rapidly.
CTL function in vivo
In vivo CTL assays were used to evaluate how cytotoxic effector activities changed during the course of acute, protracted, and chronic viral infections. This approach enables CTL activity to be monitored in the intact host without the requirement for any in vitro culture or restimulation (36, 37). At various time points following infection, gp33- and np396-specific killing was checked by determining the clearance of adoptively transferred syngeneic splenocytes pulsed with either gp33 or np396 peptide epitopes (Fig. 8). By 8 h posttransfer, both peptide-pulsed and control nonpulsed target cells were recoverable from naive mice; however, acute LCMV infection elicited a potent CTL response which was revealed by the loss of peptide-pulsed target cells but the maintenance of non-peptide-treated control cells (Fig. 8,A). Consistent with cytokine production data, efficient gp33 and np396 killing activity was detectable in the spleen over 1 year following acute infection (Fig. 8,B). In acutely infected hosts, reduced levels of killing were detectable in the livers over time following acute LCMV infection which concurs with both the lower number of virus-specific CD8 T cells and the inability to ascribe cytokine production to all of the cells in this organ (Figs. 8,C and 7).
During protracted and chronic LCMV infection, altered in vivo CTL profiles were observed as by 28 days following infection, np396-specific lysis in the spleen was reduced to 29.3 ± 14.3% and 17.3 ± 1.3%, respectively. This is consistent with the previously reported tetramer staining and cytokine production data (Figs. 2 and 4). Although np396-specific cytotoxicity was dramatically decreased during protracted and chronic LCMV infection, gp33-specific killing remained relatively high for at least 90 days postinfection with similar patterns observed in both the spleens and livers (Fig. 8). Notably, a statistically significant decline in gp33-specific CTL activity was apparent in chronically infected hosts between days 90 and 287 postinfection (p < 0.05), but a similar reduction was not detectable in protractedly infected mice. Consequently, CTL activity more closely parallels the persistence of the epitope-specific CD8 T cells visualized by tetramer staining following chronic infection. Additional studies using B6-lpr target cells demonstrated that the observed killing was not Fas dependent (data not shown).
Aberrant CD4 T cell responses
Since CD4 T cells can augment the functional activity of pathogen-specific CD8 T cells (25, 26, 27, 28, 29), we compared CD4 T cell responses in B6 mice undergoing acute and protracted LCMV infection. As expected, I-Ab-restricted gp61-specific responses were readily detectable and np309-specific responses were subdominant (Fig. 9, A and C). As the acute infection was cleared, >50% of the virus-specific CD4 T cells that produced IFN-γ were also capable of producing IL-2. Although memory CD8 T cells were stably maintained following acute LCMV infection (Fig. 4,A), virus-specific CD4 T cell responses gradually declined over time (Fig. 9 A).
In protractedly infected B6 mice, virus-specific CD4 T cell responses were also elicited, but the magnitude of these responses was lower than that observed in acutely infected hosts (Fig. 9, B and D). As the virus persisted, these responses declined in number with the subdominant np309-specific response becoming undetectable by 29 days following infection. The gp61-specific response also continued to decline, and was below the limit of detection by day 200 postinfection. Unlike virus-specific CD8 T cell responses that reemerged in the spleen and bone marrow as the infection was controlled, a similar rebound of CD4 T cell responses was not observed. Thus, the reacquisition of CD8 T cell effector capabilities does not appear to be aided by a parallel rise in CD4 T cell activities.
In this study we have investigated the dynamics of Ag-specific T cell responses to viral infections which are rapidly resolved, those that are more slowly brought under control, and those that persist with chronic viral shedding. Taken together, our findings reveal two key points regarding virus-specific CD8 T cell responses to infections that are only slowly or never controlled. First, although the majority of virus-specific CD8 T cells may lose the capacity to produce antiviral cytokines, these activities may be partially restored if the viral load can be brought under control. Second, virus-specific CD8 T cells are not maintained indefinitely in a functionally unresponsive state. Thus, if high viral loads persist, there is a continued attrition of virus-specific CD8 T cells as these cells become deleted from the chronically infected host. Our data show that the functional robustness of the virus-specific CD8 T cell response is determined in part by the duration of exposure to high viral loads. Also, our findings suggest that there is an interplay between the presence of viral Ag and CD4 T cell responses that also impacts the competency and long-term maintenance of virus-specific CD8 T cell responses.
Numerous previous studies have documented that following LCMV-Armstrong infection, immunocompetent mice elaborate a massive CD8 T cell response that rapidly clears the virus (1, 2, 33, 42). The clone 13 strain of LCMV differs from the parental Armstrong strain by only two amino acid substitutions but causes a protracted infection in immunocompetent mice and a chronic infection in CD4−/− mice (20, 32, 38, 39, 42, 43, 44, 45). Despite the higher initial viral burden during LCMV-clone 13 infection, the burst size of the CD8 T cell response is lower than that observed during acute LCMV-Armstrong infection (Figs. 2 and 4). The targeting of dendritic cells and macrophages by LCMV-clone 13 as well as inadequate help from virus-specific CD4 T cells may contribute to the reduced early Ag-driven expansion of virus-specific CD8 T cells (44, 45, 46, 47, 48, 49) and subsequent clearance of the infection is likely to be hindered by the lower numbers of virus-specific CD8 T cells, as well as by loss of effector functions. Tropism differences between LCMV-Armstrong and LCMV-clone 13 may contribute to this reduced ability to resolve the infection (44, 45, 46, 47); however, the much broader effects of high viral load, resulting in repetitive T cell activation, also probably drive the step-wise and progressive impairment of antiviral T cells. Perforin-deficient mice cannot control LCMV-Armstrong infection and virus-specific T cell responses are also aberrant under these conditions (33, 50). Conversely, an acute infection occurs if immunocompetent mice are infected with lower doses (102 PFU) of LCMV-clone 13 and this is associated with a multi-epitope-specific CD8 T cell response and robust IFN-γ production (42). The consistent finding of these studies is that the identical virus can elicit different types of cellular immune responses depending on the inoculum size and subsequent duration of infection. This supports the notion that although viral tropism may influence CD8 T cell responses, viral load and length of infection represent more global regulatory mechanisms.
The dysregulation of cytokine production by virus-specific CD8 T cells during protracted and chronic LCMV infection does not appear to be stochastic, but instead occurs in a step-wise and predictable manner as the responding cells fail to produce IL-2 and subsequently display reduced TNF-α and then IFN-γ production. The resulting impairment is, however, not necessarily a permanent state of functional unresponsiveness and during the course of protracted LCMV infection the presence of viral Ag appeared to act as a rheostatic regulator of virus-specific CD8 T cell effector potential. During protracted LCMV infection, the viral loads slowly declined and a sequential reemergence of CD8 T cell effector activities was observed as first IFN-γ-producing CD8 T cells and then TNF-α/IFN-γ-producing cells were again detected. This rebound of cytokine production by virus-specific CD8 T cells did not correlate with a concomitant rebound of CD4 T cell functions suggesting that CD8 T cells can regain certain activities without considerable CD4 T cell help. The reacquisition of cytokine-producing activities also differed depending upon the viral-epitope recognized as well as the anatomical location of the cells. In the spleen and bone marrow, effector functions were more rapidly and prominently regained by gp33- and gp276-specific CD8 T cells but the impairment of np396-specific T cell functions was more sustained. By contrast, no such resurgence in effector activities was observed in the livers of mice undergoing protracted infection.
By using sensitive in vivo assays, we have revealed the rather surprising result that CTL activity to at least one viral epitope (gp33) remains detectable for prolonged periods in chronically infected hosts. Why the hosts remain chronically infected despite this maintenance of CTL function is not yet defined. It has been reported that IFN-γ is more critical for the control of persistent LCMV infection, and thus the impairment of IFN-γ production that we report may be the primary determinant of chronicity (42, 51, 52, 53). In addition, np396-specific cells may be more effective at viral control. Unpublished findings from our laboratory, as well as other published reports, suggest that this specificity of CD8 T cells may play a principle role in controlling acute LCMV infection (54) (A. E. Tebo and A. J. Zajac, unpublished observations). The observation that the physical loss of np396-specific CD8 T cells correlates with impaired virus-control is consistent with this notion. Decreased viral glycoprotein expression and inadequate presentation of the gp33 epitope has been documented during persistent LCMV infection (39, 55). Therefore, even though the cytotoxic effector functions of gp33-specific CD8 T cells may be detectable for extended periods in chronically infected hosts, they may be unable to clear the hosts’ infected cells because of lack of Ag expression and presentation. Thus, CD8 T cell fate and function are likely to not only be determined by the overall viral load, but also the abundance of individual viral epitopes that are presented.
Nonlymphoid, tertiary tissues harbor significant numbers of Ag-specific T cells which accumulate in these sites even after infections that do not widely disseminate (6, 13). Since CD8 T cells purge virus-infected cells by elaborating their effector activities locally, this distribution of responding cells is critical for the control of systemic infections. As shown in this study, virus-specific CD8 T cells are present in the livers of mice undergoing acute, protracted, and chronic LCMV infections; however, these cells are not as functionally robust as their splenic counterparts and no reacquisition of effector capabilities is detectable as protracted infection is controlled. Impairment of cytokine production by liver-derived CD8 T cells has also been reported during influenza infection, and Ag presentation by liver endothelial cells has been shown to induce T cell tolerance (56, 57). Since the liver has been proposed as a location for the elimination of activated T cells, it is plausible that loss of effector functions precedes deletion (58). This would be consistent with the situation in the spleens and livers of chronically infected mice where effector-function-negative CD8 T cells emerge that are then also removed over time (Figs. 4,C and 7 C).
The phenotypic alterations that occur in virus-specific CD8 T cells during the course of protracted and chronic infections not only manifest as changes in their functional attributes but are also accompanied by shifts in cell surface marker expression. In the case of chronic LCMV infection, the virus-specific CD8 T cells express CD43 (1B11) at sustained high levels (Fig. 6), become CD44int (Fig. 2 F)(39), are predominately CD69high (20), and tend to remain CD62Llow (30, 59) before becoming physically deleted. This expression profile indicates that Ag-specific T cells remain activated despite their inability to elaborate effector activities. Initially during the course of both protracted and chronic LCMV infection, the phenotype of virus-specific CD8 T cells is similar, but as protracted LCMV infection is controlled, changes in their cell surface and functional characteristics occur. As the antigenic stimuli are withdrawn, the properties of virus-specific CD8 T cells change as they begin to resemble normal functionally competent memory cells. Thus, intervention strategies to lower viral loads or rest CD8 T cells in the absence of viral Ag may boost the effector potential of virus-specific CD8 T cells.
The inability to ascribe effector capabilities to Ag-specific CD8 T cells has now been documented in numerous systems, particularly during the course of infections where the replicating virus persists for extended periods (60, 61, 62, 63). Appay et al. (64) reported HIV-specific CD8 T cells that produced IFN-γ and macrophage-inflammatory protein-1β but not TNF-α and these cells also failed to elaborate cytolytic functions. Impairments of CD8 T cell effector functions have also been shown to occur during persistent LCMV infection of adult mice (20, 33, 39, 42, 65). The overall pattern that emerges is that CD8 T cells experience distinct stages of functional impairment as they lose effector activities in a step-wise manner. However, if viral loads can be sufficiently contained, then Ag-specific CD8 T cells may reacquire effector capabilities. During the course of acute hepatitis C virus infection, hepatitis C virus-specific CD8 T cells appear to undergo a transient inability to produce IFN-γ, termed a stunned phenotype, but recover from this as the viremia drops (61). Initiation of highly active antiretroviral therapy during the acute phase of HIV infection is associated with more vigorous CD4 T cell responses and, following periods of cessation of treatment, often a more broadly reactive CD8 T cell response emerges and the viral load is contained (66, 67, 68). Thus, early intervention strategies to lower viral loads in chronically infected hosts may have beneficial effects on CD8 T cell responses. Nevertheless, enhancing virus-specific CD8 T cell activities under these conditions may be temporally restricted as the potentially responsive cells may pass a point of no return after which they cannot reacquire effector capabilities. Interruptions of highly active antiretroviral therapy in patients with established chronic HIV infection did not result in more successful control of viral loads (69, 70). Additionally, the results presented in this report show that prolonged exposure to high viral loads leads to the silencing of effector functions and attrition of virus-specific CD8 T cells. Several mechanisms may contribute to the attrition of these T cells including a more vigorous programmed death phase, the inability of the CD8 T cells to respond to survival factors such as IL-7 and IL-15, or a direct or indirect requirement for CD4 T cell functions (71, 72, 73).
CD4 T cells are key regulators of the immune response and are pivotal for the development of functionally robust subsets of CD8 T cells. Reduced CD4 T cell responses, including diminished IL-2 production, are often associated with protracted and chronic infections, and in the case of HIV infection, more prominent CD4 T cell responses correlate with better viral control (18, 19, 21, 22, 33, 66). This report shows that the failure to rapidly control LCMV-clone 13 infection in immunocompetent mice is associated with both diminished CD4 and CD8 T cell activities whereas LCMV-clone 13 infection of CD4−/− hosts results in a chronic infection and the functional inactivation of the responding CD8 T cells. Several reports have documented that the functional quality of CD8 T cells is reduced if CD4 T cell help is absent during the primary response, even if the initiating Ag is rapidly cleared (25, 26, 27, 28, 29). However, it is unlikely that the dysregulation of virus-specific CD8 T cells in chronically infected hosts is solely due to inadequate CD4 T cell help. Studies from our laboratory demonstrate that acute LCMV infection of CD4−/− mice results in the emergence of phenotypically and functionally aberrant virus-specific CD8 T cells (74). Nevertheless, by comparison with the results of this report, the extent of the functional impairment of “helpless” CD8 T cells in acutely infected hosts is far less severe than that observed during chronic LCMV infection and marked deletion is also not necessarily apparent. Taken together, the overall pattern that emerges suggests that although poor CD4 T cell help may reduce the quality of responding CD8 T cells, prolonged exposure to high viral loads is associated with severe functional inhibition. This is consistent with a model that high viral loads impair the functional capacity of CD4 T cells and then inadequate CD4 T cell help coupled with repetitive antigenic stimulation lead to the functional unresponsiveness and subsequent deletion of virus-specific CD8 T cells.
Overall, this report shows that the functional competency and developmental fate of virus-specific CD8 T cells is determined by multiple factors. This includes the degree of CD4 T cell help as well as the magnitude and duration of infection. The specificities of CD8 T cells that recognize preferentially presented viral epitopes appear to be more prone to rapid functional inactivation and deletion. Conversely, CD8 T cells that detect epitopes that are less favorably presented are more inclined to regain effector activities as protracted infections are brought under control. Critically, the degree of CD4 T cell help coupled with the strength of signals that result from Ag-driven stimulation are likely to determine the outcome of the CD8 T cell response.
We thank Kyoko Kojima for excellent technical assistance, Laurie Harrington for helpful discussions, and Rafi Ahmed and John Altman for supplying reagents. We also thank the University of Alabama Molecular Biology Core and Fermentation Facility for assistance with MHC tetramer production.
This work was supported by Grant AI49360-03 (to A.J.Z.), and M.J.F was supported in part by Training Grant AI07150, both from the National Institutes of Health.
Abbreviations used in this paper: LCMV, lymphocytic choriomeningitis virus; np, nucleoprotein; CTO, CellTracker Orange.