Cutting Edge: Long-Lived CD8 Memory and Protective Immunity in the Absence of CD40 Expression on CD8 T Cells¹

Recent work has demonstrated that the absence of CD4 T cells results in defective CD8 T cell memory (1, 2, 3, 4). The “unhelped” memory CD8 T cells were characterized by an impairment in their ability to proliferate and secrete effector cytokines in response to secondary challenge. The precise role CD4 T cells play in generating or maintaining memory CD8 T cells has not been elucidated. One mechanism that has been proposed suggests that CD40 expression on activated CD8 T cells is required for proper proliferation and differentiation into functional memory cells (4). In this model, CD4 T cells directly provide signals to CD8 T cells via CD40 ligand (CD40L)³-CD40 interactions, similar to helper interactions involving B cells and CD4 T cells (5).

In certain systems, ligation of CD40 on APCs or other cells has been shown to replace the requirement for CD4 T cell help in the priming of CD8 T cell responses (6, 7, 8). Similarly, blockade of CD40L using mAbs results in the inhibition of helper-dependent CD8 T cell responses, presumably by interfering with CD4-APC interactions (8). In light of new suggestions of a role for CD40 expression on CD8 T cells, one may ask whether the activating and blocking Abs influence CD8 responses exclusively at the level of the APC, or also directly affect the CD8 T cell.

Although CD40 expression on APC and CD8 T cells during priming with cell-associated Ag in a noninflammatory environment is important (4, 7, 8), an acute bacterial infection, such as with Listeria monocytogenes, can overcome CD40L blockade and allow for priming of CD8 T cell responses (9). No studies have been done to directly examine a role for CD40 on CD8 T cells in the generation of long-lived CD8 T cell memory in the context of infection and inflammation. In this study, we investigate what role CD40 expression on CD8 T cells may play in the generation of effector responses and subsequent long-term, protective CD8 memory.

Age-matched C57BL/6 (B6) and MHC class II^−/− mice were purchased from Taconic Farms (Germantown, NY). Age-matched B6, B6.PL (Thy1.1), B6.SJL (Ly5.1), CD4^−/−, CD40^−/−, and CD40L^−/− mice were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were housed in specific pathogen-free facilities at the University of Washington (Seattle, WA). Experiments began when mice were 6–8 wk of age; all experiments were done according to institutional guidelines.

To generate chimeric mice, 6- to 8-wk-old recipient B6.SJL (Ly5.1) or B6 (Ly5.1 × Ly5.2) mice were irradiated with 1000 rad from a ¹³⁷Cs source, and i.v. injected 1 day later with a 1:1 mixture of T cell-depleted bone marrow cells isolated from wild-type congenic B6 and CD40^−/− mice. Bone marrow cells were purified by negative selection with biotin-labeled anti-CD3 (BD PharMingen, San Diego, CA) followed by anti-biotin MACS beads (Miltenyi Biotec, Auburn, CA). The resulting population contained <1% T cells. Chimeric mice were maintained on antibiotic water containing neomycin sulfate and polymyxin B sulfate for 3 wk following irradiation. PBL from chimeric mice were analyzed to confirm equal wild-type and mutant lymphocyte reconstitution and the animals were infected 8 wk following bone marrow reconstitution.

Wild-type and MHC class II^−/− mice were injected i.v. with 50 μg of purified anti-CD40 mAb (IC10) or PBS at the time of immunization. To determine the effectiveness of the mAb treatment, we showed that CD8 T cell responses to soluble OVA could be induced in MHC class II^−/− mice treated with anti-CD40 mAb, but not in untreated mice (Table I).

Wild-type mice were injected i.p. with 500 μg of purified anti-CD40L mAb (MR1) or PBS at days −2, 0, and 2 following immunization. MR1 was kindly provided by C. Larsen (Emory University School of Medicine, Atlanta, GA). As with the anti-CD40 mAb, we tested the efficacy of the anti-CD40L mAb during CD8 T cell priming to cell-associated OVA, showing that only untreated mice, and not anti-CD40L-treated mice, were able to mount CD8 responses (Table I).

The recombinant L. monocytogenes strains engineered to secrete either chicken OVA (rLmOva) or the glycoprotein of lymphocytic choriomeningitis virus (LCMV; rLmGP) were kindly provided by H. Shen (University of Pennsylvania School of Medicine, Philadelphia, PA). Frozen stocks of the rLmOva or rLmGP were grown in brain-heart infusion broth. Bacteria culture samples were grown to mid-log phase, measured by OD (A₆₀₀), and diluted in PBS for injection. Injected bacteria numbers, or CFU, were more accurately determined by spreading bacterial samples on brain-heart infusion plates and incubating them overnight at 37°C. Mice were infected with priming doses equivalent to 2000–5000 CFU of the recombinant Listeria and challenge doses equivalent to 1–2 × 10⁵ CFU by tail vein injection. CFU per spleen in the infected mice were determined by plating serial 1/10 dilutions of disassociated spleen suspensions and counting colonies following overnight incubation at 37°C.

LCMV Armstrong 53b was grown on BHK cells and titered on Vero cells. Mice were infected i.p. with 10⁵ PFU of virus.

Intracellular IFN-γ staining was performed in accordance with the manufacturer’s protocol (BD PharMingen). In 96-well plates, 1–2 × 10⁶ cells/well were stimulated with medium alone or 10⁻⁸ M OVA (SIINFEKL) or GP (KAVYNFATC) peptide for 5 h in the presence of 1 μg/ml brefeldin A. Cells were then washed, stained with anti-CD8, anti-Thy1.1,anti-Ly5.1, anti-Ly5.2, or anti-CD40 mAb (BD PharMingen), resuspended in permeabilization-fixation buffer, and stained with anti-IFN-γ Ab. Labeled cells were washed in permeabilization buffer, resuspended in fix buffer, and analyzed on a FACSCalibur (BD Biosciences, Mountain View, CA). In all experiments, cells incubated in medium without peptide gave <0.1% positive cells in the CD8 population.

If CD40L-CD40 interactions between CD4 and CD8 T cells are essential for the proper differentiation of effector and memory CD8 T cells, an activating mAb to CD40 that directly engages CD8 T cells would be sufficient to provide the required “help” signals lacking in CD4-deficient environments. To test this, we immunized wild-type C57BL/6 (B6) and MHC class II-deficient mice with recombinant L. monocytogenes which secretes OVA (rLmOva), along with an activating anti-CD40 mAb. A control group of mice was infected without the anti-CD40 mAb. As expected, at day 7 postinfection (PI), all mice showed strong primary OVA-specific CD8 responses, measured by intracellular IFN-γ staining (Fig. 1,A). Mice were given a secondary challenge dose of rLmOva at 28 days PI and 3 days following challenge, secondary CD8 responses were measured along with pathogen clearance. In wild-type mice, the recall OVA-specific CD8 T cell response was not affected by treatment with anti-CD40 mAb (Fig. 1,A). As in previous studies, we observed a drastic deficiency in the recall CD8 response of class II^−/− mice and this was not reversed in class II^−/− mice that had been treated with the anti-CD40 mAb during priming (Fig. 1,A). In addition, whereas both sets of wild-type mice cleared the challenge dose of Listeria, untreated and αCD40-treated class II^−/− mice were unable to control the infection (Fig. 1 B).

We also used a mAb that binds to CD40L on CD4 T cells to inhibit interaction with APCs or CD8 T cells. It has been previously shown that Listeria infection can overcome a CD40L blockade for the primary CD8 response, presumably working to replace CD4 help by directly activating APCs (9). In line with this, we observed that primary CD8 T cell responses in anti-CD40L-treated wild-type mice were only slightly diminished compared with untreated controls (Fig. 2). Both sets of mice cleared the infection within 7 days (data not shown). Mice were given a secondary challenge of rLmOva at 60 days PI. The recall CD8 T cell response was comparable between both groups (Fig. 2), and no impairment in the ability to clear bacteria challenge was observed in the mice treated with anti-CD40L compared with untreated controls (data not shown).

CD40^−/− mice were used to examine primary and recall CD8 T cell responses in the absence of CD40. Upon immunization of wild-type, CD4^−/−, and CD40^−/− mice, all strains cleared the primary rLmOva infection within 7 days. Furthermore, the primary CD8 response was comparable among all three strains of mice, with the CD4^−/− mice showing somewhat reduced CD8 T cell numbers compared with wild-type mice (Fig. 3,A). We have evidence that the reduction in the MHC class I-restricted response in CD4^−/− mice can be accounted for by the presence of a large pool of MHC class II-restricted T cells in the CD8 population, which dilutes the conventional response (10). We challenged the mice 60 days PI and, as previously described (1, 2), CD4^−/− mice showed severely diminished protection compared with wild-type mice (Fig. 3,B). However, CD40^−/− mice were able to completely clear the challenge dose of bacteria within 3 days (Fig. 3,B). In line with this, secondary CD8 T cell responses were diminished in CD4^−/− mice, but comparable between wild-type and CD40^−/− mice (Fig. 3 A).

We went on to measure the primary and secondary memory CD8 T cell responses in CD40L-deficient mice. We immunized wild-type and CD40L^−/− mice with a recombinant Listeria that secretes a portion of the LCMV glycoprotein (rLmGP). Both strains cleared the infection within 7 days and generated comparable primary CD8 responses (Fig. 4). All mice challenged at 60 days PI were able to clear the infection within 3 days (data not shown) and generated comparable secondary CD8 responses (Fig. 4). Thus, ablating CD40-CD40L interactions in CD40^−/− and CD40L^−/− mice does not result in defective CD8 T cell immunity against acute bacterial infection.

In mAb treatment and knockout studies, we cannot rule out global effects that may influence cells other than CD8 T cells. Nor can we be certain that blocking one receptor-ligand interaction is not compensated by other factors. To examine directly the effects of CD40 expression on CD8 T cells, we generated radiation bone marrow chimeric mice in which wild-type B6 (Ly5.1) and CD40^−/− (Ly5.2) CD8 T cells develop together in irradiated F₁(Ly5.1 × Ly5.2) B6 hosts. In these chimeric mice, we can distinguish both donor T cell populations from each other (Ly5.1 or Ly5.2 only) and from the host (Fig. 5,A). Within the same host, we compared the frequency of donor-derived GP-specific wild-type and CD40^−/− CD8 T cells at day 8 following immunization with LCMV. The primary CD8 response in the spleen was similar in the CD40^−/− population compared with wild type (Fig. 5,B). This pattern was recapitulated in the CD8 response in the liver. We measured GP-specific CD8 memory levels at 60 days PI, and found similar levels between mutant and wild-type cells (Fig. 5 B). Thus, we conclude that CD40 expression on CD8 T cells does not confer any advantage in terms of contributing to the long-lived memory pool.

To examine the effect of CD40 expression on CD8 T cells during primary and secondary expansion, we generated another set of bone marrow chimeric mice in which wild-type B6 (Thy1.1) and CD40^−/− (Thy1.2) CD8 T cells develop in irradiated Ly5.1 B6 hosts. As in the previous experiment, we can distinguish both donor T cell populations from each other and from the host. Within the same host, we compared the frequency of GP-specific wild-type and CD40^−/− CD8 T cells at day 7 following primary immunization with rLmGP. The primary response measured by IFN-γ secretion was similar for wild-type and CD40^−/− CD8 T cells (Fig. 6). Again, this pattern was recapitulated in the CD8 response in the liver. When these chimeric mice were challenged at 60 days PI with either rLmGP or LCMV, we observed similar secondary GP-specific CTL responses 3–6 days following challenge in the CD40^+/+ and CD40^−/− cells (Fig. 6). There was no reproducible difference in the recall response in any tissue examined. These data strongly suggest that in the context of an acute infection, CD40 on the surface of CD8 T cells is neither required nor contributes to the generation of functional, long-term CD8 memory.

In this report, we demonstrate that protective immunity against acute bacterial infection is maintained in the absence of CD40 on CD8 T cells. Both generation of primary effector CTL and secondary responses against bacterial challenge were unhindered by the lack of CD40 signals. Similarly, we show that the number of memory CD8 T cells generated in response to LCMV infection is not affected by the lack of CD40 expression on the CD8 T cells themselves.

Although experiments using activating mAbs to CD40 can also activate APCs such as dendritic cells, macrophages, and B cells, this Ab did not have adverse affects on protective immunity or the formation of secondary effector CTL during a bacterial challenge in wild-type mice. However, in mice lacking CD4 T cells, treatment with anti-CD40 during primary immunization did not enhance protective immunity or CD8 T cell responses against secondary challenge. Therefore, triggering CD40 cannot replace CD4 help in the generation of protective immunity and rapidly responding memory CD8 T cells.

In addition to disrupting cross-talk between CD4 and CD8 T cells, blocking mAbs that bind to CD40L will also hinder CD40L-CD40 interactions between CD4 T cells and APCs. This block in the CD40L-CD40 interaction between CD4s and APCs may be manifested in our experiments by the slightly diminished primary CD8 T cell responses seen in anti-CD40L-treated relative to untreated mice. Our data confirm previous findings showing that although Listeria infection can overcome CD40L blockade, the primary CD8 T cell response is reduced compared with control mice (9). However, slightly reduced primary responses did not result in defects in secondary expansion of CTL, suggesting that CD40L does not contribute to CD8 T cell programming during immunization with L. monocytogenes.

Previous studies have examined the requirement for CD40L-CD40 interaction between CD4 T cells and APCs in the generation of CD8 T cell responses to viral and bacterial infection in CD40^−/− and CD40L^−/− mice (11, 12, 13, 14, 15). These analyses demonstrated that signals other than those mediated through CD40 ligation could lead to the efficient priming of CD8 T cell responses. Our current study corroborates the earlier findings and now extends the sum of these findings by demonstrating the lack of requirement for CD40L-CD40 interactions in the generation of protective immunity and CD8 T cell memory. We show in CD40^−/− mice that protective memory and CD8 T cell secondary responses are the same whether or not CD40 is present on either APCs or CD8 T cells during acute bacterial infection. We further show in CD40L^−/− mice that whether or not CD4 interacts directly with APCs or CD8s via CD40L during immunization, protective memory and CD8 T cell secondary responses remain unchanged.

Our most stringent test of a role for CD40 expression on CD8 T cells was done in bone marrow chimeric mice, in which wild-type and CD40-deficient CD8 T cells develop and encounter Ag in the same environment. In this study, in a direct comparison of CD8 T cells from the two different sources, we found that expression of CD40 on the surface of CD8 T cells afforded no marked advantage or disadvantage in their response to Listeria or LCMV. CD40 expression by T cells themselves did not affect trafficking or expansion in nonlymphoid tissues. We conclude that CD40 on the surface of CD8 T cells plays no role in the generation of functional memory or recall CD8 responses following priming with an acute bacterial or viral infection. A recent article by Lee et al. (16) similarly concluded that CD40 signaling directly to CD8 T cells is not relevant in the primary or secondary response to influenza virus.

Our conclusions are in sharp contrast to those of a previous study (4) showing a role for CD40 expression on CD8 T cells during immunization against a cell-associated Ag (male cells expressing the HY Ag). Several differences in the two systems could account for the discrepancies between the findings. In a helper-dependent system, CD40 on CD8 T cells could be playing a role that is bypassed in our immunization schemes. Inflammation during priming could play a critical role in determining whether or not CD8 T cells require direct CD40 signaling for robust memory generation. Furthermore, immunization of adoptively transferred TCR-transgenic T cells in a lymphopenic host, as performed in the HY studies, contrasts with our studies looking at a polyclonal T cell population in a normal environment.

Our current findings suggest that CD40L-CD40 interactions between CD4 and CD8 T cells are not important for the generation of a protective memory CD8 T cell response in the context of acute infection. Cases in which this receptor-ligand interaction have been reported to be essential for antiviral protection may be explained either by its impact on the Ab response or effects on APC activation. Future studies will examine the precise role CD4 T cells provide in the generation of protective CD8 T cell memory. Determining the signals given by CD4 T cells, whether via a direct or indirect mechanism, will aid vaccine development.