Abstract
This study documents a striking dichotomy between CD4 and CD8 T cells in terms of their requirements for CD40-CD40 ligand (CD40L) costimulation. CD40L-deficient (−/−) mice made potent virus-specific CD8 T cell responses to dominant as well as subdominant epitopes following infection with lymphocytic choriomeningitis virus. In contrast, in the very same mice, virus-specific CD4 T cell responses were severely compromised. There were 10-fold fewer virus-specific CD4 T cells in CD40L−/− mice compared with those in CD40L+/+ mice, and this inhibition was seen for both Th1 (IFN-γ, IL-2) and Th2 (IL-4) responses. An in vivo functional consequence of this Th cell defect was the inability of CD40L−/− mice to control a chronic lymphocytic choriomeningitis virus infection. This study highlights the importance of CD40-CD40L interactions in generating virus-specific CD4 T cell responses and in resolving chronic viral infection.
Many viral infections induce a potent CD8 response (1). Infections such as HIV, EBV, and CMV in humans and vaccinia virus, vesicular stomatitis virus, murine γ-herpesvirus, and lymphocytic choriomeningitis virus (LCMV)4 in mice induce CD8 expansion (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21). These CD8 T cells are responsible for controlling the infection. They recognize virally infected cells via TCR engagement with MHC class 1-peptide complexes and kill the infected cells by secreting hole-punching perforin molecules and granzymes and attenuate viral spread by secreting IFN-γ and other cytokines (14, 18, 21, 22). This initial expansion of CD8 T cells is not only responsible for eliminating the primary infection, but it may shape the memory T cell repertoire found in immune mice, since activated T cells and memory T cells share similar TCR repertoires (18). Therefore, it is important to understand the mechanisms by which these cells are activated.
CD4 T cells also contribute to viral clearance by producing IL-2, which facilitates CD8 T cell activation and expansion, and by secreting IFN-γ and TNF to activate macrophages and inhibit viral replication. In addition to secreting cytokines, CD4 Th cells mediate help to B cells by direct cell-to-cell interaction involving CD40L (expressed on activated CD4 T cells) and CD40 (expressed on B cells). This interaction is required for B cell proliferation, Ab class switching, and memory B cell development (16, 23, 24, 25, 26, 27, 28). Signaling through CD40 also activates macrophages to produce TNF-α, nitric oxide, and IL-12 and to express costimulatory molecules (29, 30, 31, 32, 33, 34). Some CTL responses are dependent upon CD4 T cells, and recently, several reports documented the role of CD40-CD40L interaction for priming these Th-dependent CTL responses (35, 36, 37).
Protective viral immunity requires Ag-specific T cells be activated sufficiently to proliferate and eliminate the infection. In the conventional model of T cell activation, peripheral T cells require two signals: 1) the first signal is delivered by the cognate interaction of the TCR complex with peptide-bearing MHC, resulting in up-regulation of CD28 and induction of CD40L expression; 2) the second signal is delivered by the interaction of CD28 with B7 and of CD40L with CD40. The purpose of these secondary (costimulatory) signals is 2-fold. First, they mediate changes in the CD40-expressing APC such as up-regulation of MHC class II and B7 molecules to make it a more effective APC (38, 39, 40, 41, 42). Secondly, these interactions strongly activate Ag-specific T cells to proliferate and secrete cytokines (43, 44, 45, 46, 47, 48). The role of CD40-CD40L interactions in driving T cell responses has been characterized in experimental models of autoreactive CD4 T cells and T cell responses against Leishmania infections, where it was found to be important for activating these cells and eliminating the infection (44, 46, 47, 48). However, the primary CD8 CTL response to LCMV is normal in CD40L-deficient mice, and these mice resolve the acute LCMV infection as well as +/+ mice (16, 23, 24), so it is unclear whether the rules that govern CD4 activation are the same as those responsible for CD8 activation.
The CD8 T cell response to LCMV is relatively CD4 independent, as CD4-deficient mice are able to mount potent CD8 responses to LCMV (49, 50, 51). The CD8 response is also not affected by defects in the B cell response, since B cell-deficient (μIg−/−) mice generate normal CD8 responses (15). These features make acute LCMV infection a useful system for addressing the role of the CD40-CD40L interaction in inducing T cell responses, because the effect of this interaction on CD8 T cells can be distinguished from its effects on CD4 T cells. In addition, recent advances in the quantitation of T cell responses at the single cell level (ELISPOT and intracellular staining for cytokines) allow simultaneous quantitation of CD4 and CD8 responses, so not only are the effects of CD40-CD40L interaction measured in the same animal, but they can be analyzed using the same assays. Following infection of CD40L-deficient mice with LCMV, we found that there was a differential requirement for CD40-CD40L interaction for CD8 and CD4 T cells. CD40L-deficient mice generated large numbers of IFNγ-secreting CD8 T cells, which were specific to dominant and subdominant epitopes during acute infection, but anti-viral CD4 T cell expansion was 10-fold lower than that in +/+ mice. This deficiency was found for IL-2-IFN-γ, and IL-4-secreting CD4 T cells, indicating that both Th1 and Th2 responses were weaker in CD40L-deficient mice. An important consequence of this deficiency in the CD4 response was that CD40L-deficient mice were unable to resolve a chronic LCMV infection. Investigation of the CD8 response in chronically infected mice revealed an absence of virus-specific CD8 T cell responses. These results show that the activation requirements of CD4 T cells are different from those of CD8 T cells and highlight a role for CD4 T cells and CD40-CD40L costimulation in resolving chronic viral infections.
Materials and Methods
Mice
C57BL/6 × 129 (F2) and C57BL/6 (H-2b) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The CD40L-deficient mice used in these experiments were generated by targeted gene disruption, which abrogates surface expression of the CD40L molecule (27, 28). The C57BL/6 carrier mice used in these experiments were bred at Emory University (Atlanta, GA).
Virus
Mice were infected i.p. with 2 × 105 PFU of the Armstrong CA 1371 strain of LCMV (52). LCMV variants, clone 13 (52) and t1b (53), were injected i.v. in the tail vein of mice at 2 × 106 PFU to cause chronic infection. Infectious virus was quantitated by plaque assay (52).
CD4 enrichment
CD4 enrichment was performed using CD4 enrichment columns. Mouse CD4 Subset Column Kits were purchased from R & D Systems (Minneapolis, MN), and the manufacturer’s suggested protocol was used.
Quantitation of virus-specific IFN-γ-secreting CD8 and CD4 T cells by ELISPOT
Virus-specific CD8 and CD4 T cell responses were measured by IFN-γ ELISPOT assay as described previously (14, 54). For stimulation, LCMV carrier mouse spleen cells (congenitally infected mice that express virus protein in the context of MHC class I and MHC class II), purified LCMV peptides that bind to MHC class I and stimulate CD8 T cell responses specifically (55), or LCMV peptides that bind to MHC class II (NP309–328 and GP61–80 of Armstrong) and stimulate only I-Ab-restricted CD4 T cell responses (56) were used. Peptides NP396–404, GP33–41, GP276–296, NP205–212, and GP92–101 were used at 0.1 μg/ml final concentration and GP61–80 and NP309–328 were used at a 1.0 μg/ml final concentration. The results shown for the CD8 response as measured by ELISPOT assay are lower than recently reported (14). The number of Ag-specific CD8 T cells shown in Fig. 2⇓ is about 2- to 5-fold lower than what we have reported recently for LCMV-infected C57BL/6 mice using MHC class I tetramers and intracellular cytokine staining. The main reason for this discrepancy is that several of the experiments in the present study were performed with C57BL6 × 129 mice. These mice show some variability and, in general, exhibit slightly lower responses than C57BL/6 mice. Also, the experiments shown in Fig. 2⇓ were performed at a time when our IFN-γ ELISPOT assay had not been fully optimized. This systematic difference in the magnitude of the response due to variation in the assay does not affect the comparisons between +/+ and CD40L-deficient CD8 responses, as they were measured at the same time by the same assay.
Intracellular staining for IFN-γ
The method for intracellular IFN-γ staining has been described previously (14, 54). Spleen cells (106 cells/well in 96-well flat-bottom plates) were stimulated in vitro with medium or GP61–80 (for CD4 T cells) or NP396–404 (for CD8 T cells) for 5 h in vitro with brefeldin A (Golgistop, PharMingen, La Jolla, CA). They were then harvested, washed once in FACS buffer, stained with allophycocyanin-conjugated monoclonal anti-CD4 (clone RM4-5, PharMingen) or allophycocyanin-conjugated monoclonal anti-CD8 (clone 53-6.7, PharMingen), and stained for intracellular IFN-γ using the Cytofix/Cytoperm staining kit (PharMingen) according to the manufacturer’s recommended protocol. FITC-conjugated monoclonal rat anti-mouse IFN-γ (clone XMG1.2) and its control isotype Ab (rat IgG1; PharMingen) were used for intracellular IFN-γ staining. Intracellular staining for CD40L was performed in the same fashion using PE-conjugated anti-CD40L (clone MR1, PharMingen).
Quantitation of virus-specific IL-2- and IL-4-secreting CD4 T cells
ELISPOT assays for measuring IL-2- or IL-4-secreting CD4 T cells were performed in the same fashion as the IFN-γ ELISPOT assay and have been described previously (54). For the IL-2 ELISPOT, the capture Ab was clone JES6-1A12, and the detection Ab was clone JES6-5H4. For the IL-4 ELISPOT, the capture Ab was clone BVD4-1D11, and the detection Ab was clone BVD6-24G2. All Abs were purchased from PharMingen. ELISPOT plates were purchased from Millipore (Bedford, MA).
Cytokine ELISAs
Cytokine ELISAs were conducted using cytokine-specific ELISA kits purchased from Genzyme Diagnostics (Cambridge, MA) and were performed and analyzed as recommended by the manufacturer. The ELISAs were read using a Bio-Rad Microplate Reader 3550 (Bio-Rad, Hercules, CA) using appropriate filters.
Results
T cell activation in CD40L-deficient mice
The level of T cell activation in the spleen was analyzed following infection of +/+ and CD40L-deficient mice with LCMV-Armstrong (Fig. 1⇓). Consistent with previous observations (16, 25, 27, 28), there were no intrinsic differences between uninfected +/+ and CD40L-deficient mice in the percentages of CD8 and CD4 T cells in the spleen. CD8 T cells became strongly activated after infection in both +/+ and CD40L-deficient mice. In contrast, CD4 T cell activation was lower in CD40L-deficient mice than in +/+ mice. The ratio of activated (CD44high) to resting (CD44low) CD4 T cells was approximately 0.5 in both groups of uninfected mice. On day 8 after the infection, the ratio of activated CD4 T cells changed to 1.8 in +/+ mice, but remained relatively unchanged (0.4) in CD40L-deficient mice. These data indicated that there may be a difference in the requirement of CD40-CD40L interaction for activation of virus-specific CD8 and CD4 T cells. To address this, virus-specific CD8 and CD4 T cell responses were further characterized.
T cell activation in the spleen after LCMV infection. Spleen cells from LCMV-infected mice were analyzed for expression of CD4, CD8, and T cell activation marker CD44 by flow cytometry. Numbers represent the percentage of cells in each quadrant. Data are representative of five experiments with 10–20 mice/group.
Antiviral CD8 responses are normal in the absence of CD40-CD40L interaction
The number of virus-specific CD8 T cells was quantitated by intracellular IFN-γ staining on day 8 postinfection. Consistent with previous observations of CTL generation (16, 23, 24), high numbers of virus-specific CD8 T cells were generated in CD40L-deficient mice, and the virus was eliminated (Table I⇓ and data not shown). The CD8 T cell response to the dominant LCMV epitopes NP396–404 and GP33–41 was normal in CD40L-deficient mice, and both groups generated 5–7 × 106 specific cells/spleen on day 8 (Table I⇓). Responses to subdominant epitopes constitute only a minority of the CD8 T cell response (14, 55, 57). Since these epitopes may bind to MHC class I and/or TCR with lower affinity, initiating a CD8 T cell response to them might require costimulatory interactions. However, there was no defect in the response to GP276–296, GP92–101, or NP205–212. Both CD40L-deficient and +/+ mice contained 2 × 106 cells specific to GP276–296, 4 × 105 cells specific to GP92–101, and 1 × 106 cells specific to NP205–212 (Table I⇓). Overall, there was no deficit in responses to dominant or subdominant epitopes in the absence of CD40-CD40L interaction.
CD8+ T cell responses to dominant and subdominant epitopesa
Generation of anti-viral CD4 T cells requires CD40L
The IFN-γ ELISPOT assay was used to analyze the CD4 T cell response in the same mice in which the CD8 response was measured. CD4 T cells were purified from spleen cells of +/+ and CD40L-deficient mice on day 8 after LCMV infection and were restimulated with syngeneic LCMV-infected carrier spleen cells or with syngeneic uninfected spleen cells. Fig. 2⇓ shows that CD40L-deficient mice that generated potent CD8 T cell responses lacked normal anti-viral CD4 T cell responses. CD40L-deficient mice had 9- to 11-fold fewer IFN-γ-secreting virus-specific CD4 T cells per spleen, whereas they had in some cases more IFN-γ-secreting CD8 T cells (compare Expt. 1 and 2 in Fig. 2⇓A with Expt. 1 and 2 in Fig. 2⇓B). The frequency of virus-specific CD4 T cells was 7.7-, 9.4-, and 14.9-fold lower in Expt. 1, 2, and 3, respectively.
Differential requirement of CD40-CD40L interaction for CD8 and CD4 T cells. Total splenic CD8 and CD4 T cell responses to LCMV were quantitated by IFN-γ ELISPOT in the same mice on day 8 postinfection. A shows that the number of LCMV-specific CD4 T cells per spleen was ≥10-fold lower in CD40L-deficient mice (open bars). Purified CD4 T cells were analyzed by IFN-γ ELISPOT assay following restimulation with virus-infected carrier spleen cells. Virus-specific T cells were determined by subtracting the frequencies of IFN-γ-secreting CD4 T cells in unstimulated wells from those of cells responding to virus. The numbers shown are the percent decrease in the number of CD4 T cells in CD40L-deficient mice compared with that in +/+ mice. The frequency of virus-specific CD4 T cells per CD4 was 1 in 96 for +/+ and 1in 741 for −/− mice in Expt. 1, 1 in 500 for +/+ and 1 in 4700 for −/− in Expt. 2, and 1 in 556 for +/+ and <1 in 8333 for −/− in Expt. 3. B shows that there was no difference in the total CD8 T cell response to LCMV between +/+ and CD40L-deficient mice in the same mice that showed differences in the CD4 response. Each experiment represents the average of two mice.
Fig. 3⇓ shows intracellular-IFN-γ staining, illustrating that the CD4 T cell response was defective in CD40L-deficient mice, while the CD8 response was normal. The frequency of peptide-specific CD4 T cells was 10-fold lower in the CD40L-deficient mice compared with that in the +/+ mice. In eight mice analyzed by this assay, CD40L-deficient mice had 1.9 ± 1.2 × 105 GP61–80-specific cells, whereas +/+ mice had 1.7 ± 0.5 × 106. By the same assay the frequency of CD8 T cells that were specific to NP396–404 was not significantly different from that found in the +/+ mice. CD40L-deficient mice had 4.5 × 106 specific cells/spleen, and +/+ mice had 7.5 × 106 specific cells/spleen.
Intracellular staining for IFN-γ following peptide stimulation of CD8 and CD4 T cells. Spleen cells were stimulated with either LCMV NP396–404 (MHC class I restricted) or GP61–80 (MHC class II restricted), and CD8 and CD4 T cell production of IFN-γ was analyzed by flow cytometry. The numbers shown are the frequency of peptide-specific CD4 T cells per total CD4 T cells or of peptide-specific CD8 T cells per total CD8 T cells. Note that while the frequency of peptide-specific CD8 T cells was minimally affected by the absence of CD40-CD40L interaction, the frequency of virus-specific CD4 T cells in the same mouse was reduced 10-fold. Data are representative of eight mice per group analyzed in four independent experiments.
To further examine the difference between the requirement of CD40-CD40L interaction for CD8 vs CD4 T cell responses, supernatants from cells stimulated with NP396–404 were analyzed for IFN-γ production by ELISA. As shown in Table II⇓, there was no intrinsic defect in IFN-γ production, and CD8 T cells from CD40L-deficient mice produced similar levels of IFN-γ when stimulated with NP396–404 peptide compared with CD8 T cells from +/+ mice (18.9 vs 18.6 ng/ml). To analyze CD4 responses in these mice, CD4 T cells were purified using a negative selection column and then stimulated with LCMV carrier spleen cells for 24 h. As shown in Table II⇓, CD4 T cells from CD40L-deficient mice made less IFN-γ than those from +/+ mice. Again, by the same assay, it can be seen that there is a differential requirement for CD40-CD40L interaction for inducing CD8 and CD4 T cell responses: CD40L is dispensable for generating primary antiviral CD8 T cell responses, but is required for LCMV-specific CD4 T cell responses.
Cytokine production by +/+ and CD40L-deficient cellsa
This trend of poor CD4 T cell activation and expansion was also found by IL-2 ELISPOT. The frequency of virus-specific CD4 T cells that secrete IL-2 was 3- to 9-fold greater in +/+ mice than in CD40L-deficient mice, and there were 3- to 8-fold more IL-2-secreting CD4 T cells in +/+ mice than in CD40L-deficient mice (Fig. 4⇓A). Table II⇑ shows that less IL-2 was made by CD40L-deficient CD4 T cells as determined by ELISA. In addition, the IL-4 ELISPOT assay (Fig. 4⇓B) demonstrated that there was also a defect in generating IL-4-secreting CD4 T cells in mice lacking CD40L. There were 2–3 × 104 IL-4-secreting CD4 T cells/spleen in +/+ mice vs only 3–7 × 103/spleen in CD40L-deficient mice. Table II⇑ also shows that CD4 T cells from CD40L-deficient mice made less IL-4 than those from +/+ mice. Overall, CD4 Th1 and Th2 responses were compromised in the absence of CD40-CD40L interaction.
CD40L-deficient mice generate few virus-specific IL-2- and IL-4-secreting CD4 T cells. Purified CD4 T cells were stimulated with or without LCMV-infected carrier mouse spleen cells and analyzed by IL-2 ELISPOT assay (A) or IL-4 ELISPOT assay (B). The total number of LCMV-specific CD4 T cells per spleen was determined as described in Fig. 2⇑. Numbers in parentheses are the frequency of LCMV-specific CD4 T cells per CD4 T cell. Spleen cells were pooled from two mice in each group.
To examine why there is a differential requirement for CD40-CD40L interaction for CD8 and CD4 T cells, the level of CD40L expression was analyzed on these cells after LCMV infection. While CD40L was difficult to identify after surface staining, intracellular levels of CD40L could be found and quantitated by flow cytometry. Fig. 5⇓A shows spleen cells from LCMV-infected mice that were stimulated with NP396–404. The NP396–404-specific T cells that made IFN-γ were identified and analyzed for intracellular levels of CD40L as depicted in Fig. 5⇓B. Among the Ag-specific CD8 T cells, 18-56% expressed CD40L, and these had a mean fluorescence intensity (MFI) of 86. For comparison, resting CD8 T cells from uninfected mice were predominantly CD40L negative, with a MFI of ≤10. GP61–80-specific CD4 T cells were also identified as illustrated in Fig. 5⇓C, and their level of CD40L expression is shown in Fig. 5⇓D. While resting CD4 T cells did not show any staining, almost all Ag-specific CD4 T cells (98%) expressed CD40L, and the MFI of the CD40L staining for these cells was 507. These results demonstrate that Ag-specific CD4 T cells contain much higher levels of CD40L than Ag-specific CD8 T cells.
Levels of CD40L in CD8 and CD4 T cells. The levels of CD40L in CD8 and CD4 T cells were compared by intracellular staining for CD40L. On day 8 postinfection, spleen cells from +/+ LCMV-infected mice were stimulated with NP396–404, GP61–80, or medium alone for 5 h, surface stained for CD8 or CD4, and then stained for intracellular IFN-γ and CD40L. A shows CD8 T cells that made IFN-γ (indicated by the box) after NP396–404 stimulation. The level of CD40L in these virus-specific, IFN-γ+ve CD8 T cells is shown by the dark histogram in B. The level of CD40L staining without stimulation is indicated by the bold line, and the level in uninfected mice is indicated by the dotted line. In C, GP61–80-specific CD4 T cells were identified by intracellular staining for IFN-γ. D shows the level of intracellular CD40L made by virus-specific, IFN-γ+ve CD4 T cells (dark histogram). The level of CD40L staining for unstimulated CD4 T cells (bold line) and that in uninfected mice (dotted line) are also shown for comparison. The numbers shown in A and C are the percentages of CD8 or CD4 T cells that were IFN-γ positive. The MFI of CD40L staining ranged from 32–86 for virus-specific CD8 T cells and from 434–564 for virus-specific CD4 T cells. CD8 T cells from infected mice without stimulation or from uninfected mice had an MFI from 5–6. CD4 T cells from uninfected mice had an MFI of 8, and those from day 8 mice that were not stimulated had an MFI of 23. Data are representative of four mice analyzed.
Infection of CD40L-deficient mice with LCMV clone 13 or clone t1b
The results presented so far have shown that CD40L-deficient mice exhibit a selective defect in generating CD4 T cell responses during acute LCMV infection. However, these CD40L-deficient mice generate a potent virus-specific CD8 T cell response that is capable of controlling infection by the LCMV-Armstrong strain (16, 23). In contrast to the LCMV-Armstrong strain that is eliminated within a week by both +/+ and CD40L-deficient mice, infection of +/+ adult mice with macrophage tropic strains of LCMV, such as clone 13, results in a disseminated infection that can last for several months or longer (58). We next examined the ability of CD40L-deficient mice to control infection by the more virulent LCMV clone 13 (see Fig. 6⇓). Groups of +/+ and CD40L-deficient mice were infected with LCMV clone 13, and levels of virus in the serum were monitored over time. High levels of virus were present in the sera of both groups of mice (>105 PFU/ml) for the first 2–3 wk following infection. However, the level of virus dropped following this period in +/+ mice, but not in CD40L-deficient mice. The serum levels of virus in +/+ mice eventually dropped to below detection by plaque assay (<50 PFU/ml) by day 77, but the level of virus in CD40L-deficient mice never decreased, and on day 77 there remained between 5 × 104 to 105 PFU/ml of clone 13 in the serum. Similar results were seen after another LCMV variant, clone t1b. Both +/+ and CD40L-deficient mice initially showed high levels of virus in the serum (3–10 × 104 PFU/ml) and in other tissues on day 8. However, +/+ mice controlled the infection in approximately 1 mo in the serum and liver, whereas CD40L-deficient mice became chronically infected.
CD40L-deficient mice are unable to control infection with LCMV clone-13. CD40L-deficient or +/+ mice were infected on day 0 with 2 × 106 PFU of LCMV strain clone-13. Individual mice were bled at various times postinfection, and levels of virus in serum were quantitated by plaque assay. Data shown are from one of three independent experiments and are representative of 15 mice analyzed. The dashed line indicates the limit of detection of this assay.
Since CD40L-/- mice were unable to control chronic LCMV infection, it was of interest to determine whether virus-specific CD8 T cell responses were still present in these mice. In an analysis performed at 44 days postinfection, no CD8 T cell response could be detected to any of the LCMV epitopes (NP396–404, GP33–41, GP276–296, NP205–212, and GP92–101) in clone t1b-infected CD40L-deficient mice. In contrast, CD8 T cells specific to GP33–41 and GP276–296 were readily detectable in +/+ mice that had controlled (but not fully eliminated) the infection. An example of the GP33–41-specific CD8 T cell responses in t1b-infected +/+ and CD40L-deficient mice is shown in Fig. 7⇓. In +/+ mice, 9% of CD8 T cells were specific to GP33–41 as measured by intracellular cytokine staining for IFN-γ. In contrast, CD40L-deficient mice did not show any staining above the background level. Taken together, the results presented in Figs. 6⇑ and 7⇓ show that CD40L-deficient mice are unable to control a chronic LCMV infection and that under these conditions of persistent antigenic stimulation, there is also a functional loss of virus-specific CD8 T cell responses.
Absence of functional virus-specific CD8 T cells in CD40L-deficient mice after chronic infection. CD8 responses in mice infected 6 wk earlier with LCMV t1b were quantitated by intracellular staining for IFN-γ. Numbers shown are the percentages of CD8 T cells that stained positive for IFN-γ and are representative of 12 mice/group analyzed from two separate experiments.
Discussion
Acute LCMV infection of mice leads to a strong primary virus-specific CD8 T cell response that is relatively independent of CD4 Th cells (49, 50, 51, 58). This makes it possible to dissect the rules that affect CD8 responses from those that affect CD4 responses. This study shows that CD40-CD40L interaction is dispensable for generating virus-specific CD8 T cell responses, but is required for virus-specific CD4 T cell activation and expansion.
Previous studies have shown that CD40-CD40L interaction is not required for CTL generation as measured by CTL killing of virally infected target cells, but these CTL assays primarily measured the CD8 response to the dominant MHC class I binding epitopes of LCMV (16, 23, 24). Here, we extend these studies and show that CD8 responses to subdominant epitopes are also normal. This is important, as it has been shown that vaccination with subdominant epitopes can confer protection against viral challenge (55, 57, 59), and these results indicate that CD40 costimulation is not necessary to activate these cells. In contrast to CD8 T cells, virus-specific CD4 T cell responses were severely compromised (∼90% inhibition) in CD40L-deficient mice. This requirement for CD40-CD40L costimulation was seen for both Th1 (IFN-γ, IL-2) and Th2 (IL-4) CD4 responses.
Why are CD4 T cells but not CD8 T cells affected by the absence of CD40-CD40L interaction? There are several possibilities. 1) One possibility is that the amount of Ag that CD8 T cells encounter is sufficient to activate them, whereas CD4 T cells encounter less Ag during the viral infection. One study that examined the number of TCRs required to activate T cells found that about 8000 TCR molecules had to be engaged to achieve activation, but if the T cells were also stimulated through CD28, then the number of TCRs that had to be engaged was reduced to approximately 1500/cell (60). It is possible that a large number of MHC class I molecules contain viral peptides, so a sufficient number of TCR molecules may be engaged for CD8 T cell activation. Also, most cells express MHC class I molecules. Fewer cells (namely B cells, macrophages, and dendritic cells) express MHC class II, and it is possible that Ag density per cell (i.e., MHC class II molecules presenting viral peptides) is not sufficient to activate CD4 T cells in the absence of CD40-CD40L costimulation. Hence, CD4 T cells may not reach the activation threshold without additional costimulatory signals. 2) The strength of the TCR signal for CD8 T cells could obviate the need for costimulation (61, 62). The TCR signal delivered by the LCMV peptides in association with MHC class I along with any contribution by the CD8 coreceptor may be strong enough to activate CD8 T cells. In contrast, the MHC class II-restricted peptides of LCMV may not deliver as strong a TCR signal or the CD4 coreceptor is too weak, so that costimulation is needed to activate CD4 T cells. In particular, CD4 T cells may require more CD40L signaling, and circumstantial evidence for this is the fact that CD4 T cells contained higher levels of CD40L than CD8 T cells (Fig. 5⇑). 3) Lymphoid architecture could be more important for activating CD4 T cells than CD8 T cells. Histological analysis of spleens from CD40L-deficient mice show that the architecture of these spleens is abnormal, and there is little or no migration of B cells and B7.1+ cells into the follicles after infection with LCMV (our unpublished observation). B7 signaling may play a role in priming CD4 T cells (41, 45, 63), and since APC and CD4 T cells are not parked next to each other, CD28-B7 interactions would not occur. In contrast, CD8 T cells are localized outside of the follicles in the marginal zone and red pulp, and therefore are less affected by follicular architecture. 4) Another possibility is that there may be additional molecules used by CD8 T cells. 41BB-41BBL interaction has been shown to be important for alloreactive CD8 T cells (64), and it may also be important for antiviral CD8 T cells (J. T. Tan, J. K. Whitmire, R. Ahmed, T. C. Pearson, and C. P. Larsen, manuscript in preparation).
Why is CD40-CD40L interaction required for clearance of chronic viral infection? Initially after chronic infection, CD40L-deficient mice generate normal anti-viral CD8 responses (J. K. Whitmire and R. Ahmed, manuscript in preparation), but with time there is a loss of these virus-specific responses until none can be detected (Fig. 7⇑ and unpublished observations). These CD8 T cells become nonresponsive with regard to IFN-γ production or cytotoxic activity, or they are deleted (J. K. Whitmire and R. Ahmed, manuscript in preparation). This most likely reflects the requirement for sufficient numbers of primed CD4 T cells. When mice have a deficiency in CD4 responses due to an absence of CD40-CD40L interaction (this study), an absence of B7-CD28 interaction (unpublished observation), or a deletion of CD4 T cells (58, 65), then they will be impaired in their ability to resolve chronic viral infections even though their CD8 response to acute viral infection is normal. CD4 T cells provide IL-2 and other growth factors that prevent T cell death (66, 67). Exposure to high levels of Ag over an extended period of time induces CD8 T cells to die by apoptosis, and this can be prevented in vitro by the addition of IL-2 to cultures. IL-2 may play a similar role in vivo during a persisting viral infection, and it remains to be seen how administration of IL-2 will affect virus-specific CTL during these infections. CD4 Th cells could also affect APC. Since CD8 and CD4 T cells may interact with a common APC that expresses MHC class I and MHC class II, CD4 T cells could activate this shared APC so that it, in turn, maintains effector CTL (68). One effect that T cells could have through the CD40-CD40L pathway is the induction of IL-12 by dendritic cells, which could be important in activating T cells (40). Recently, it has been shown that CD40L-CD40 interaction between activated CD4 T cells and Ag-bearing dendritic cells is important for priming CD8 CTL (35, 36, 37). These reports show that signaling through CD40 “licenses” APC in some way to prime Th-dependent CTL responses, perhaps by increasing MHC expression or costimulatory molecules. While the CTL response against LCMV can be generated in the absence of CD4 T cells or CD40-CD40L interaction, an analogous role for APC licensing may exist during chronic viral infection when CTL must be maintained for extended periods of time so that they do not undergo activation-induced cell death or become functionally exhausted (65). This licensing could also manifest itself in the preservation of APC that might otherwise be deleted during chronic LCMV infection. Some strains of LCMV, such as those mentioned in these studies, result in immunopathology with altered splenic architecture and a loss of some cell types, including APC (69, 70). CD40-CD40L interaction may preserve these APC or maintain the proper microenvironment so that they can interact with and maintain CTL.
Protracted HIV infection leads to a loss of virus-specific CD4 T cells, followed by the loss of anti-viral CD8 T cell responses over time (71). The results reported here suggest a role for CD40-CD40L interaction and CD4 T cells in preventing chronic viral infections. Help from CD4 Th cells likely plays a pivotal role in driving and maintaining the CTL responses under conditions of a protracted viral infection (58, 65, 72). Future investigations should resolve how CD4 T cells assist CTL, whether it is through production of growth factors such as IL-2 or through modification of APC. Dissecting the mechanisms that are important for generating and maintaining CD8 and CD4 T cell responses will lead to improved strategies for the prevention and treatment of chronic viral infections.
Acknowledgments
We thank Morry Hsu, Mary Kathryn Large, and Kaja Madhavi-Krishna for excellent technical assistance.
Footnotes
-
↵1 This work was supported by National Institutes of Health Grants AI30048 and NS21496 (to R.A.) and AI/DK40519–01 (to C.P.L. and T.C.P.). R.A.F. is an investigator of the Howard Hughes Medical Institute.
-
↵2 Current address: Department of Immunology, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080-4990.
-
↵3 Address correspondence and reprint requests to Dr. Rafi Ahmed, Emory Vaccine Center, Emory University School of Medicine, G211 Rollins Research Building, 1510 Clifton Road, Atlanta, GA 30322. E-mail address: ra{at}microbio.emory.edu
-
↵4 Abbreviations used in this paper: LCMV, lymphocytic choriomeningitis virus; CD40L, CD40 ligand; ELISPOT, enzyme-linked immunospot assay.
- Received December 23, 1998.
- Accepted July 13, 1999.
- Copyright © 1999 by The American Association of Immunologists