Report

Requirement for CD4 T Cell Help in Generating Functional CD8 T Cell Memory

See allHide authors and affiliations

Science  11 Apr 2003:
Vol. 300, Issue 5617, pp. 337-339
DOI: 10.1126/science.1082305

Abstract

Although primary CD8 responses to acute infections are independent of CD4 help, it is unknown whether a similar situation applies to secondary responses. We show that depletion of CD4 cells during the recall response has minimal effect, whereas depletion during the priming phase leads to reduced responses by memory CD8 cells to reinfection. Memory CD8 cells generated in CD4+/+ mice responded normally when transferred into CD4−/− hosts, whereas memory CD8 cells generated in CD4−/− mice mounted defective recall responses in CD4+/+ adoptive hosts. These results demonstrate a previously undescribed role for CD4 help in the development of functional CD8 memory.

The ability to mount an accelerated response to reinfection is the hallmark of immunological memory and the basis of vaccination (1). This ability is mediated by the presence of antigen-specific memory T cells that are capable of responding more rapidly and in a greater magnitude than naı̈ve T lymphocytes (1). Primary responses by naı̈ve CD8 cells to noninflammatory immunogens are known to require CD4 T cell help (TH), and this help is thought to occur through activation of antigen-presenting cells (APCs) (2–5) or to involve direct CD40-CD40L interaction between CD4 and CD8 cells (6). However, the primary CD8 response to infectious agents is often independent of TH (7–10), and it is hypothesized that recognition of microbial products by Toll-like receptors directly activates APCs and thus bypasses the need for CD4 help (11). Relative to the requirements for priming of naı̈ve CD8 responses, little is known about the contribution of TH to the recall response mediated by memory CD8 cells.

To investigate the role of TH in the generation of memory CD8 cells, we used a heterologous prime/boost regimen that is particularly effective in boosting antigen-specific memory responses, thus allowing us to quantitatively measure secondary expansion by memory CD8 cells (12). CD4+/+ and CD4−/− mice were immunized with a recombinant vaccinia virus (rVV33) expressing the GP33-41 epitope from lymphocytic choriomeningitis virus (LCMV) (13–15). At least 30 days later, immunized mice were boosted with a recombinantListeria monocytogenes (rLm33) expressing the same epitope (16). In accordance with previous reports (9, 10), rVV33 immunization induced comparable primary CD8 responses and a similar number of GP33-specific CD8 memory cells in CD4+/+ and CD4−/− mice (6.7 × 103 and 5.8 × 103 per spleen, respectively) (Fig. 1, A and B). Upon rLm33 boost, GP33-specific memory CD8 cells in CD4 +/+ mice mounted a strong recall response, expanding to 7.3 × 106/spleen by day 7 after rLm33 infection. CD4−/− mice also mounted a recall response, but the population of GP33-specific CD8 cells expanded to only 1.1 × 106/spleen, significantly less than in CD4+/+ mice. This difference was not due to a lower initial level of memory in CD4−/− mice after rVV33-prime but rather resulted from a greatly reduced expansion of GP33-specific memory CD8 cells during the recall response in CD4−/−mice (184-fold compared to 1088-fold in CD4+/+ mice) (Fig. 1C). These results indicate that CD4 cells play an important role in supporting a robust recall response by memory CD8 cells upon antigenic restimulation.

Figure 1

Diminished recall response by memory CD8 cells in CD4−/− mice. CD4+/+ (+/+) and CD4−/− (−/−) mice were primed with rVV33 alone (rVV33) or 30 days later were boosted with rLm33 (rVV33 → rLm33). (A) Detection of GP33-specific cells by intracellular IFN-γ (IFNγ+/GP33) and MHC/peptide tetramer (Kb/GP34 and Db/GP33) staining on day 7 after rLm33-boost (15). Numbers indicate the frequencies as percentages of splenic CD8 cells. (B) Total numbers of GP33-specific CD8 cells per spleen. (C) Expansion of GP33-specific CD8 cells during the recall response, calculated as fold increase in rVV33 → rLm33 mice over rVV33 mice.

To investigate the role of TH during the recall response, we depleted CD4 cells in vivo either during rLm33-boost or during rVV33-prime (14). Unexpectedly, depletion of CD4 cells during rLm33-boost had minimal impact on the recall response of GP33-specific memory cells (Fig. 2). Instead, depletion of CD4 cells during the rVV33-prime adversely affected the proliferative potential of memory CD8 cells during subsequent boost with rLm33. GP33-specific cells in these mice expanded only ∼200-fold by day 7 after rLm33-boost, compared to >1000-fold expansion in control mice and mice depleted of CD4+ cells during rLm33-boost. These results show that CD4 cells play an essential role during the priming phase for the generation of memory CD8 cells capable of efficient recall response, whereas TH is not required for secondary expansion in the boosting phase.

Figure 2

Depletion of CD4 cells during priming results in a defective recall response by memory CD8 cells. CD4+/+mice were immunized with rVV33 alone (I and II) or immunized with rVV33 and 33 days later boosted with rLm33 (III to V). CD4 cells were depleted with monoclonal antibodies to CD4 either during rVV33 priming (II and IV) or during the rLm33 boost (V). (A) Detection of GP33-specific cells by intracellular IFN-γ (IFNγ/GP33), and MHC/peptide tetramer (Kb/GP34 and Db/GP33) staining on day 7 after rLm33-boost. Numbers indicate the frequencies as percentages of splenic CD8 cells. (B) Total numbers of GP33-specific CD8 cells per spleen.

To more rigorously test the above conclusions, we examined whether memory CD8 cells primed in CD4+/+ mice could mount a normal recall response in adoptive CD4−/− hosts. LCMV immunization was used to generate a large population of GP33-specific memory CD8 cells (17). Purified CD8 cells from LCMV-immune CD4+/+ mice were transferred into CD4+/+ and CD4−/− mice, which were then infected with rLm33 (14). Donor CD8 cells mounted a vigorous GP33-specific recall response that was similar in magnitude in both CD4+/+ and CD4−/− recipients (Fig. 3, A and B). These results further support the conclusion that TH is not required during the recall response for the rapid expansion of memory CD8 cells.

Figure 3

Memory CD8 cells primed in CD4+/+ mice mount a normal recall response in CD4−/− adoptive hosts, whereas memory CD8 cells generated in CD4−/− mice mount a defective recall response in CD4+/+ adoptive hosts. (A and B) Purified CD8+ cells from LCMV-immune (>60 days) Thy1.1 CD4+/+ (+/+) mice were transferred into congenic Thy1.2 CD4+/+ (+/+ → +/+) or CD4−/− (+/+ → −/−) mice, which were then infected with rLm33. (C toF) CD8+ cells were purified from Ly5.2 CD4+/+ (+/+) or CD4−/− (−/−) mice that were previously immunized with LCMV (> 60 days). Purified CD8+cells, normalized to contain an equal number of GP33-specific cells, were transferred into congenic Ly5.1 CD4+/+ (+/+) mice, which were then infected with rLm33. (A and C) Intracellular IFN-γ staining of GP33-specific cells on day 0 and 7 after rLM33 infection. Numbers indicate the frequencies as percentages of total splenic CD8 cells (gated on CD8+ cells). (B and D) Total numbers of donor GP33-specific cells per spleen at day 0, 4, and 7 after infection (D.P.I.), based on intracellular IFN-γ (IFNγ+/GP33) or MHC/peptide tetramer (Kb/GP34+ or Db/GP33+) staining. Triangles: (+/+ → +/+); squares: (+/+ → −/−); and circles: (−/− → +/+). (E) Total numbers of host GP33-specific CD8 cells per spleen at 0, 4, and 7 D.P.I., as determined by intracellular IFN-γ staining. (F) Numbers of rLm33 on day 4 in spleen of mice that received memory CD8 cells from CD4+/+ (+/+) or CD4−/− (−/−) mice.

We next examined the ability of memory CD8 cells from CD4−/− mice to mount a recall response in adoptive CD4+/+ hosts. Purified CD8 cells from LCMV-immune CD4+/+ and CD4−/− mice were transferred into CD4+/+ recipients. An equal number of GP33-specific CD8 cells were transferred, and these cells engrafted similarly, as was evident from comparable levels of GP33-specific CD8 cells detected in recipient mice before rLm33 infection (Fig. 3C). By day 7 after rLm33 infection, GP33-specific memory cells from CD4+/+ mice had mounted a strong recall response, expanding to a frequency of ∼36.6% of splenic CD8 cells and a total number of 8.8 × 106/spleen (Fig. 3, C and D). In contrast, GP33-specific memory CD8 cells from CD4−/− mice mounted a much weaker recall response, reaching only ∼1.0% of splenic CD8 cells and a total number of 1.3 × 105/spleen. These mice displayed a higher primary GP33-specific response by the host CD8 lymphocytes (Fig. 3, C and E), presumably compensating for the reduced expansion by donor memory cells. Furthermore, as a consequence of the reduced recall response, a lower level of protection against rLm33 infection was observed in these mice, which had 10-fold higher numbers of rLm33 bacteria on day 4 after infection than mice that received donor cells from CD4+/+ mice (Fig. 3F). These results demonstrate that memory CD8 cells generated in the absence of TH are qualitatively different and are less fit to mount a vigorous recall response against reinfection, even in the presence of CD4 cells.

In support of our in vivo results, memory CD8 cells from CD4+/+ mice proliferated more extensively and produced a higher level of cytokines than those from CD4−/− mice upon ex vivo antigenic stimulation (fig. S1). Functional impairment of memory CD8 cells in CD4−/− mice was not due to the inability of these mice to clear infection by rLm33, rVV33, or LCMV (7, 9, 14). Furthermore, memory T cells in these mice showed no evidence of recent T cell receptor activation (fig. S2), indicating that they were not stimulated by persisting antigens and thus were bona fide resting memory T cells. Absence of CD4 help has been shown to lead to defective CD8 responses during chronic viral infection (8,18–20). Previous studies have also reported reduced secondary expansion by memory CD8 cells in CD4-deficient mice following acute viral infections (9, 21), although the precise reason for this defect has not been examined. Recent studies have shown that the generation of efficient CD8 memory requires the presence of CD4 cells during immunization with classically TH-dependent antigens (6, 22). In addition, it has been shown that memory CD8 cells primed by viral infection in the absence of CD4 help mount a defective secondary response in vitro upon restimulation (22).

Our in vivo data from acute viral and bacterial infections show that the presence of CD4 cells during the priming phase is critical for generating functional CD8 memory, whereas TH is not required during the recall response for secondary expansion. These results demonstrate a previously undescribed role for CD4 cells in the CD8 response to infection that had been thought to be THindependent. It remains to be determined if help in this case involves direct CD40-CD40L interaction between CD4 and CD8 cells, as suggested for TH-dependent antigens (6). Elucidation of the mechanism will help us better understand the development of CD8 memory and optimize vaccine strategies for inducing protective immunity.

Supporting Online Material

www.sciencemag.org/cgi/content/full/300/5617/337/DC1

Materials and Methods

References

Figs. S1 and S2

  • * To whom correspondence should be addressed at the Department of Microbiology, University of Pennsylvania School of Medicine, 3610 Hamilton Walk, Philadelphia, PA 19104–6076, USA. E-mail: hshen{at}mail.med.upenn.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article