Measuring T cell Responses: Apples, Orange Pies & Interferon Gamma

As mentioned in a prior post, one of the key questions that the HIV vaccine field has to wrestle with is: what is the best method for measuring vaccine-induced HIV-specific CD4 and CD8 T cell responses? The failure of the Merck vaccine - the first T cell-based HIV vaccine to be tested for efficacy - has shone a harsh light on this question, because the construct appeared to induce reasonable T cell responses according to the most commonly used tests.

To start at the beginning, there are several important features of the T cell response:

• Specificity: this refers to the specific parts of a pathogen that T cells are targeting. T cells recognize small slices of proteins from pathogens called epitopes via molecular docking bays on their surface called T cell receptors (TCRs).
• Breadth: refers to how many different epitopes from a given pathogen that the T cell response is targeting (narrow=few epitopes, broad=many).
• Function: the ability of responding T cells to perform specific functions e.g. proliferation (the ability of the T cell to copy itself), production of immune system proteins called cytokines (e.g. interferon gamma, IL-2 and TNF-alpha) and chemokines (e.g. MIP-1beta). Other functional measures are also possible such as the measurement of the ability of HIV-specific CD8 T cells to kill HIV-infected cells in a lab dish, some studies use surrogate measures of cell-killing ability (e.g. the expression of a molecule called CD107a).
• Phenotype: A term that refers to the fact that T cells can belong to different subsets which immunologists define based on particular functional properties and expression of cell surface markers. It’s worth noting that these definitions can be a little fuzzy, and that T cells can also transition back and forth between different phenotypes depending on whether they are actively responding to something and where they are in the body (e.g. the blood vs. body tissues). Two broad subsets of memory T cells with differing properties are “central memory” T cells, which circulate in a resting state but have a prodigious ability to proliferate if they encounter the pathogen that they recognize, and “effector memory” T cells which circulate in a more activated state which is associated with rapid production of the cytokine interferon gamma but a poor ability to proliferate. Sometimes immunologists also refer to additional categories such as “transitional effector memory“ T cells (cells transitioning from central memory to effector memory subsequent to activation) and simply “effector” T cells which are so battle weary (or “terminally differentiated” in immunological parlance) that they cannot return to a resting central memory state.

For vaccine researchers, the goal is to try and understand what an ideal HIV-specific T cell response should look like, based on each of these features; e.g. which epitopes should be targeted? How broad should the response be? What are the ideal functional properties? Should the response be biased toward a particular phenotype (e.g. central vs. effector memory)?

In the run-up to the Merck trial, the leading assay for studying T cell responses was the relatively new interferon gamma (IFNg) ELISpot (Enzyme-Linked ImmunoSpot). This test quantitates T cell responses to a given epitope (or mixture of epitopes) based on their ability to make just this one cytokine. Merck and researchers from the HVTN (along with many other immunologists working in the vaccine field) put a great deal of effort into standardizing the IFNg ELISpot in order to try and ensure that results were reproducible and could be compared across trials of different candidates. Several of the large coordinating bodies in HIV vaccine research – including the Global HIV/AIDS Vaccine Enterprise and US government-sponsored Partnership for AIDS Vaccine Evaluation (PAVE) – also sponsored overlapping processes aiming to promote standardization of immunogenicity assays.

However, at the same time as these efforts were underway, researchers using a different assay called intracellular cytokine staining (ICS) showed that IFNg-producing T cells do not reflect the entire memory T cell response generated by vaccination; in fact, in Stephen De Rosa’s initial study using various licensed vaccines, IFNg-producing T cells were often in the minority compared to T cells producing other cytokines and chemokines such as IL-2. De Rosa’s work began a trend toward evaluating multiple parameters involved in T cell function, a trend that was accelerated by data showing that, in people with HIV infection, the presence of so-called “polyfunctional” CD4 and CD8 T cell responses targeting HIV is associated with control of viral load and long-term non-progression. Researchers at the NIH’s Vaccine Research Center, who are blessed to possess one of the largest multi-parameter flow cytometers in the land (a device which allows multiple different facets of a cell’s function to be analyzed simultaneously), pioneered these studies of polyfunctional T cells and thus, unsurprisingly, began to use this approach to analyze the T cell responses induced by their HIV vaccine candidate. The VRC have focused on five main T cell functions: the production of the cytokines IL-2, TNF-alpha, IFNg and the chemokine MIP-1beta, and expression of the CD107a molecule mentioned earlier. The VRC now typically present the spectrum of T cell functions they detect amongst a particular antigen-specific population in the form of pie charts, divided into the proportion of T cells elaborating a particular number of functions (e.g. blue for 1, green for 2, etc.).

The upshot of these developments is that anyone attempting to directly compare the immunogenicity of the Merck HIV vaccine and the VRC’s candidate has their work cut out (as Jerry Sadoff, a member of the AIDS Vaccine Research Subcommittee complained at their December 12 meeting). The Merck data presented and published to date primarily includes information derived from the IFNg ELISpot (with some additional ICS data regarding IL-2), while the VRC have presented ICS data addressing multiple functions as well as some IFNg ELISpot data. In both cases, there are no published data addressing the immunogenicity of the exact vaccine regimens selected for use in the efficacy trials in a systematic and detailed way; Merck has published data from a scattering of individuals who received the prototype Ad5 construct (encoding only gag) at various doses, some also after receiving a DNA priming immunization (which Merck decided not to use). The VRC has published data from the phase I trials of the DNA and Ad5 components of their regimen in separate papers in the same issue of the Journal of Infectious Diseases, but data from individuals who have received the full DNA prime/Ad5 boost immunization series is limited and has only been presented at meetings. In order to inform the HIV vaccine field – and particularly to assist those making decisions about PAVE100 - one immediate priority should be the publication of detailed information regarding the T cell responses induced by the Merck & VRC candidates that addresses their specificity, breadth, function & phenotype.

As if things weren’t complicated enough, CD8 T cell aficionado Otto Yang has recently highlighted another important issue relating to immunogenicity assays in a paper forthcoming in JID. Yang points out that the best test for CD8 T cells may be one that directly measures their ability to kill HIV-infected cells in a lab dish. In his JID paper, Yang uses such a test to show that CD8 T cells that appear to recognize HIV variants from different clades (when evaluated using IFNg ELISpot) may actually not be able to kill cells infected with those same variants. Yang argues that more direct measures of CD8 T cell efficacy should be used in vaccine evaluation, a point echoed by David Watkins in a recent article in the IAVI Report. Watkins suggests that, in some cases, using IFNg ELISpot may be equivalent to measuring non-neutralizing antibodies, while direct measures of CD8 T cell killing are more akin to a neutralization assay. Patricia D’Souza and Marcus Altfeld make a similar point in the JID commentary that accompanies Otto Yang’s report: “The accurate assessment of the antiviral activity of HIV-1-specific CD8 T cells is a crucial step toward identifying the immune correlates of vaccine efficacy. There is a tendency to consider CD8 lymphocyte responses as equivalent, perhaps in the same way that the Env-specific antibodies were once thought to be equally effective in mediating antiviral activity. It is possible that there is a hierarchy of efficacious CD8+ lymphocyte responses in which the most effective CD8 lymphocytes may be an analogue of a potent and broad neutralizing antibody.” However, assays for measuring CD8 T cell killing also vary and so – if the field does follow this path – another effort to identify, select and standardize the best method will be necessary.

Recently published papers, commentaries & opinion pieces:

The Journal of Infectious Diseases 2008;197:000–000
DOI: 10.1086/525288
EDITORIAL COMMENTARY

Measuring HIV‐1–Specific T Cell Immunity: How Valid Are Current Assays?
M. Patricia D’Souza1 and Marcus Altfeld2

1Vaccine Clinical Research Branch, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland; 2Partners AIDS Research Center, Massachusetts General Hospital, Charlestown, and Division of AIDS, Harvard Medical School, Boston

The Journal of Infectious Diseases 2008;197:000–000
DOI: 10.1086/525281
MAJOR ARTICLE

Cross-Clade Detection of HIV-1–Specific Cytotoxic T Lymphocytes Does Not Reflect Cross-Clade Antiviral Activity
Michael S. Bennett,1,3, Hwee L. Ng,2,3, Ayub Ali,2,3 and Otto O. Yang1,2,3

1Department of Microbiology, Immunology, and Molecular Genetics and 2Division of Infectious Diseases, Department of Medicine, David Geffen School of Medicine, and 3UCLA AIDS Institute, University of California, Los Angeles

Presented in part: First Annual Ugandan AIDS Conference, Kampala, Uganda, 7 December 2006.

(See the editorial commentary by D'Souza and Altfeld, on pages XXX–XX.)

The genetic divergence of human immunodeficiency virus (HIV)–1 into distinct clades is a serious consideration for cytotoxic T lymphocyte (CTL)–based vaccine development. Demonstrations that CTLs can cross‐recognize epitope sequences from different clades has been proposed as offering hope for a single vaccine. Cross‐clade CTL data, however, have been generated by assessing recognition of exogenous peptides. The present study compares HIV-1–specific CTL cross-clade epitope recognition of exogenously loaded peptides with suppression of HIV‐1–infected cells. Despite apparently broad cross-clade reactivity of CTLs against the former, CTL suppression of HIV-1 strains with corresponding epitope sequences is significantly impaired. The functional avidity of CTLs for nonautologous clade epitope sequences is diminished, suggesting that CTLs can fail to recognize levels of infected endogenously derived cell-surface epitopes despite recognizing supraphysiologic exogenously added epitopes. These data strongly support clade-specific antiviral activity of CTLs and call into question the validity of standard methods for assessing cross-clade CTL activity or CTL antiviral activity in general.

AIDS:Volume 22(3)30 January 2008p 325-331
[OPINION]

Aiming for successful vaccine-induced HIV-1-specific cytotoxic T lymphocytes
Yang, Otto O

From the Division of Infectious Diseases, Department of Medicine and Department of Microbiology, Immunology, and Molecular Genetics, UCLA AIDS Institute, David Geffen School of Medicine at UCLA, Los Angeles, California, 90095, USA.

JVI Accepts, published online ahead of print on 7 November 2007
J. Virology doi:10.1128/JVI.01634-07
REVIEW

HIV-1 Vaccine Development: Recent Advances in the CTL Platform "Spotty Business"
Kimberly A. Schoenly and David B. Weiner

The Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine

The Journal of Experimental Medicine, Vol. 205, No. 1, 7-12
Published online 14 January 2008
doi:10.1084/jem.20072681
COMMENTARY

The failed HIV Merck vaccine study: a step back or a launching point for future vaccine development?
Rafick-Pierre Sekaly

R.-P. Sekaly is at Université de Montréal, CR-CHUM, Institut National de la Santé et de la Recherche Médicale U743, Montréal, Québec H2X1P1, Canada

The world of human immunodeficiency virus (HIV) vaccines has suffered a baffling setback. The first trial of a vaccine designed to elicit strong cellular immunity has shown no protection against infection. More alarmingly, the vaccine appeared to increase the rate of HIV infection in individuals with prior immunity against the adenovirus vector used in the vaccine. A new study in this issue suggests that a different vaccine approach—using a DNA prime/poxvirus boost strategy—induces polyfunctional immune responses to an HIV immunogen. The disappointing results of the recent vaccine trial suggest that a more thorough assessment of vaccine-induced immune responses is urgently needed, and that more emphasis should be placed on primate models before efficacy trials are undertaken.

Comments

[Ian]

Yang points out that the best test for CD8 T cells may be one that directly measures their ability to kill HIV-infected cells in a lab dish.

I don't think you can make this statement as a general rule, though; it's probably different for different viruses, and there may be no way of knowing if it's true for a specific virus without actually testing. For example, Frank Chisari and others have shown that hepatitis B virus can be suppressed without cell killing in what seemed to be an IFN-dependent manner. There are several other instances of viral suppression without killing.

That said, of course killing seems to be tremendously important in HIV in particular, and since there are strong clinical correlations with CTL killing and HIB suppression it does make sense to focus on that in this particular case.

[Otto Yang]

Kudos on your very nice summary of this topic (and I like the term "aficionado"!). I wanted to respond to the comment by Ian, which covers two important points about assays for CTL.

1. Activity against virus-infected cells probably is central to any good assay for vaccines (as the antibody researchers have taught us). The problem with ELISpot and ICS, as currently practiced, is that the target cells are loaded with excess epitopes compared to virus-infected cells. Thus, the hurdle for giving a signal by these assays is much lower than the hurdle for CTL recognition of infected cells. This means one will see many false positive "responses" as we saw in the cited JID paper. Note also that we were not doing ELISpots in that paper, but rather using chromium release assays, which are far less sensitive than ELISpot, yet we still had a very high false positive rate compared to measurements of activity against actual HIV-infected cells.

2. CTL may have multiple possible mechanisms against viruses, including IFN-gamma (the Chisari work pointed out by Ian is a nice example). Killing may or may not be the key activity against HIV (although my opinion is that it probably is), but a true gold standard for all viruses may be some direct measure of suppression of viral replication by the CTL (whatever the actual mechanism). We have shown both killing and direct assessment of viral inhibition, and found that they correlate well for HIV.

Thus, the ability of CTL to recognize infected cells, and the effect they have once they do recognize the cells, are two separate and important points for assay development.

Finally, I'd like to mention some brewing controversy about the polyfunctional assays (multiparameter ICS) mentioned. CTL in HIV-infected persons with excellent immune containment tend to be highly polyfunctional. The question has been raised whether polyfunctionality is the cause of good immune control (in which case these measurements are highly important for predicting vaccine efficacy), or vice-versa (in which case polyfunctionality is an epiphenomenon of good immune control or simply a healthy immune system). This remains unanswered, but so far most vaccines in HIV-seronegatives appear to have excellent polyfunctionality, including vaccines that are derivatives of strategies that have thus far not shown much promise (e.g. poxvirus-based constructs).

It will be fascinating to see what opinions are proposed at the upcoming HIV vaccine summit meeting called by NIH this Tuesday (3/25/08).