Vezys et al. reply

Abstract

Replying to: R. M. Welsh and L. K. Selin Nature 459, 10.1038/nature08091 (2009)

We reported that it is possible to increase the total number of memory CD8 T cells within an organism, and to establish preternatural numbers of vaccine-specific effector memory CD8 T cells while preserving naive CD8 T cells and most pre-existing memory CD8 T cells specific for a previously encountered infectionhttps://www.nature.com/articles/nature08091¹. These findings raise new questions regarding the regulation and limits of generating CD8 T cell immunity. Our discussion highlighted three points related to the issue of attrition. First, that it is possible to over-estimate perceived attrition by only examining percentages (see Fig. 1 of ref. 1)². Second, our vaccine regimen resulted predominantly in the generation of effector memory CD8 T cells located outside of lymph nodes. It remains possible that the number of lymph node central memory T cells remains tightly regulated. Third, we noted that our data did not refute that attrition could happen under a variety of circumstances. However, our data demonstrate that attrition is not an axiomatic property of immunization, mandated by stringent regulation of the size of the total memory CD8 T-cell compartment. Indeed, we saw no evidence of attrition after single infections with a virus (vaccinia), an intracellular bacteria (Listeria monocytogenes) and a parasite that induces massive splenomegaly (Plasmodium yoelii), and observed comparatively little attrition after a heterologous prime–boost regimen involving successive immunization with three viruses¹.

Main

The issue at hand is whether our results represent rare exceptions that could be exploited for vaccination, or whether preservation of immunological memory is a normal feature of the mammalian immune system. Welsh and Selin² argue that memory T-cell maintenance is regulated by the quantity of type 1 IFN produced in response to subsequent infections, and that most infections invariably lead to profound loss of pre-existing memory CD8 T cells. Data supporting this claim demonstrate that up to 25–90% of pre-existing memory CD8 T cells are lost after single infections, and that attrition accumulates further with each infection^3,4. Surprisingly, infection-specific naive and memory T cells were not spared⁵. The current model of Welsh and Selin² predicts that immunological memory must be short-lived in the face of occasional infections. Consequently, they question the evidence that CD8 T-cell memory specific for vaccinia has a long half-life in humans⁶. Memory CD8 T cells have also been examined after a single exposure to both measles and hantavirus, and robust populations could be detected >10 years later^7,8. Protective immunity against measles virus (for which T-cell memory may be important) persists >65 years (ref. 9). Memory B cells persist indefinitely after smallpox immunization, and the half-life of serum antibody, which must be continuously produced by differentiated antigen-experienced B cells, ranges from 92 to 3,014 years after exposure to measles, mumps, rubella or vaccinia virus^10,11. These data indicate that immunological memory does not necessarily undergo rapid erosion, although further studies regarding the longevity of CD8 T memory are needed.

The concept advanced by our paper is that attrition is not an axiomatic property of immunization, and that the memory CD8 T-cell compartment is capable of expansion. In support of these findings, it was recently reported that primary human cytomegalovirus infection resulted in a long-lived increase in the total number of antigen-experienced CD8 T cells, and induced a reduction in the frequency, but not number, of pre-existing memory CD8 T cells specific for influenza or Epstein–Barr viruses in blood¹². It should be stressed that immunity to certain agents (for example, after vaccination with non-replicating agents) may be intrinsically prone to attrition, and certain infections may significantly erode immunological memory. These may include pathogens that infect lymphocytes, destroy lymphoid tissue or are associated with other pathologies, as indicated by Welsh and Selin². Although our results raise the bar for what levels of CD8 T-cell memory might be achievable through vaccination, Welsh and Selin² highlight the important fact that each vaccine vector should be evaluated empirically for safety and the preservation of pre-existing immunological memory.