Post-immunization spleen
Spleen 3 days after immunisation.
B cells (red), CTL (green) and dendritic cells (blue

I noted some time ago that, despite the name, it’s clear that cytotoxic T lymphocytes (CTL) are more than just cytotoxic. They can limit or eliminate virus infections by killing infected cells, sure; but as well, they can  produce cytokines, like interferons, that also shut down viruses.

So which of these abilities makes for an effective CTL response? What should we be aiming for when we design vaccines? What should we be measuring, to test whether our vaccines are doing something useful?

The recent failure1 of the STEP anti-HIV vaccine trial makes this an especially timely question:

… the disappointing outcome of this study highlights one particular point: our understanding of T cell-mediated efficacy remains limited. Without such knowledge, vaccine design strategies will remain largely empirical, and further failures are likely. 2

(My emphasis.)

Jenner vaccinatingThe latest Nature Medicine has a review2 and a perspective3 looking at this issue.

In their review, Appay et al. emphasize that not all CTL are equal. Frequency of CTL isn’t a good correlate with control of viruses (at least with HIV) or of tumors. Part of this is undoubtedly because of immune escape –that is, the virus or tumor are likely to have mutated to evade recognition by the most frequent CTL –  but in this review Appay et al. are looking at the other side of the equation: What’s wrong with the CTL, that they can be escaped from, or otherwise can’t control the virus? And how to decide if your vaccine has a chance of working? If raw CTL numbers are not useful, is there some characteristic of CTL that does correlate with control?

They talk about “polyfunctional” CTL (CTL that are able to produce several kinds of cytokines on recognizing their target — a relatively new concept4 that I’ve been a little skeptical about in the past, but my skepticism is slowly waning) as well as “functional avidity”, clonal senescence, and some other factors. Essentially, they conclude, the ideal CTL would be highly sensitive to antigen; would have multiple responses to draw on; and would be able to replicate and maintain itself over a long period.

So how can we drive the development of such highly effective CTL with vaccines? They make several comments:

  • Low antigen concentration is more likely to induce highly effective CTL. “… it should be borne in mind that recurrent immunizations and boosting with high-dose antigen, conducted with the aim of achieving maximum immunogenicity as determined by solely quantitative measures, may have adverse effects; specifically, effective vaccine-induced T cells could become exhausted through the loss of replicative capacity and apoptotic deletion.”
  • Because low levels of antigen may not drive any immune response whatsoever, we need to think about effective co-stimulation to kick-start the immune response. “… the level and type of costimulation may compensate for low antigen abundance and play a considerable part in boosting T cells with high antigen sensitivities. ” This includes working through innate immune receptors, a very active field of research.
  • Dendritic cells in a lymph node
    Dendritic cells (red) in a lymph node

    As part of improved co-stimulation, we need to think about inducing CD4 (helper) T cells along with CTL. CD4 T cells are important in activating and licensing the dendritic cells that then in turn activate CTL. However, inducing CD4 T cells has the risk of stimulating a regulatory (TReg) type response, which might reduce CTL functions. We need to understand this better; it’s another very active field.

  • Route of vaccine administration may be more important than previously thought. Different antigen-presenting cells are present in different tissues, and the likelihood of active versus regulatory responses is strongly affected by these cells.
  • Vaccine antigens may not be optimal. “… for practical reasons, vaccine formulations often contain consensus or optimized antigens … it seems that vaccine-induced T cells generally recognize the native antigen less efficiently and are therefore less effective in the face of their real targets.”

All of these points are the subject of research (some more than others). Hopefully, the outcome of the STEP trial will lead to a better understanding of underlying principles, and so eventually to much more effective vaccine development.

The failure of the Merck STEP trial represents a turning point for the field of vaccination. However, far from embodying the end of the T cell vaccine strategy, it heralds a new era in vaccine research based on comprehensive immunomonitoring.5


  1. Though see Richard Jeffrey‘s comment in Nature (Vaccine failure is not a ‘crisis’ for HIV research. Richard Jefferys.  
Nature 453, 719-720 (5 June 2008) | doi:10.1038/453719d) as to the meaning and implication of “failure” here[]
  2. Appay, V., Douek, D.C., Price, D.A. (2008). CD8+ T cell efficacy in vaccination and disease. Nature Medicine, 14(6), 623-628. DOI: 10.1038/nm.f.1774[][]
  3. Nature Medicine 14, 617 – 621 (2008) doi:10.1038/nm.f.1759. David I Watkins, Dennis R Burton, Esper G Kallas, John P Moore & Wayne C Koff. Nonhuman primate models and the failure of the Merck HIV-1 vaccine in humans[]
  4. Betts MR, Nason MC, West SM, De Rosa SC, Migueles SA, Abraham J, Lederman MM, Benito JM, Goepfert PA, Connors M et al. (2006) HIV nonprogressors preferentially maintain highly functional HIV-specific CD8+ T cells. Blood 107:4781–4789.[]
  5. CD8+ T cell efficacy in vaccination and disease. Victor Appay, Daniel C Douek & David A Price. Nature Medicine 14, 623 – 628 (2008) doi:10.1038/nm.f.1774[]