Functional Signatures of Antiviral T Cell Immunity as A Measure of Virus Replication Activity and of Virus-Associated Disease

Background: Antiviral T-cell responses are functionally and phenotypically heterogeneous. Despite the large number of studies in this field, it has been difficult to correlate different patterns of antiviral T cell responses with virus-associated disease activity. Methods: PBMC from 25 subjects were stained with CD3, CD8, CCR7, CD45RA, CD27, CD28, CD7, CD127, CD57 or MHC Class I tetramers specific for HIV-1, CMV, EBV and Influenza (Flu). Furthermore, intracellular IFN-γ and IL-2 secretion, proliferation and cytotoxicity measured by the degranulation assay were assessed following stimulation with virus-derived peptides. Flu-infection was considered as a model of virus elimination, CMV and EBV as models of chronic controlled infection and HIV-1 as a model of uncontrolled virus infection. Results: A large number of CD8 T cell populations, i.e. up to 24 populations, were identified using CCR7 and CD45RA, CD27, CD28, CD7, CD127 or CD57. More importantly, no correlation was found between phenotypically defined CD8 T cell populations and virus replication and virus-associated disease activity. In contrast, when cytokines secretion, proliferation and degranulation/cytotoxicity were evaluated, three functionally distinct CD4 and CD8 T cell populations were identified. In particular, based on IL-2 and IFN-γ secretion, virus elimination was associated with a dominant IL-2 functional signature, virus persistence with control replication with a polyfunctional (IL-2 plus IFN-γ) signature, and virus persistence with non-controlled virus replication with a dominant IFN-γ functional signature. Proliferation capacity was substantially impaired with non-controlled virus replication while the extent of degranulation in CD8 T cells was similar in both controlled and non-controlled virus infections. Conclusions: These results clearly demonstrate the lack of overlap between the different surface markers commonly used and more importantly that functional and not phenotypic signatures can be a mirror of virus replication activity and a novel tool for monitoring virus-associated disease. INTRODUCTION AND AIMS There is an urgent need to identify correlates of protection to evaluate the effectiveness of antiviral agents and, more importantly, of vaccine-induced T cell responses. Despite extensive investigations in this field, very few correlates of protection have been identified. T cell responses can be characterized according to phenotypic markers, i.e. the expression of surface markers, or according to their functional activities. A phenotypic heterogeneity of antigen-specific T cell responses has already been shown and many markers are used to identify memory T cells. Many of these markers are used indiscriminately to identify and characterize naïve, effector and memory T cell subsets although the degree of overlap between the different markers has never been extensively investigated. Along the same line, a functional heterogeneity has also been shown and new markers have recently been identified. Of note, the function of virus-specific CD8 T cells against Cytomegalovirus (CMV), Epstein Barr virus (EBV), Influenza (Flu) and HIV-1 was recently analyzed on the basis of the ability of CD8 T cells to secrete IFN-γ and IL-2 and to proliferate (Zimmerli et al, PNAS 2005 and poster #404P). A selective defect of IL-2 secreting CD8 T cells was only found in subjects with progressive HIV-1 infection while the frequency of IFN-γ secreting CD8 T cells was similar within the different virus-specific responses. In this study, we have investigated: q The degree of overlap between different surface markers of CD8 T cell subsets; q The degree of overlap between different surface markers of virus-specific CD8 T cell subsets; q The relevance of the analysis of functional patterns rather than phenotypic patterns. RESULTS AND DISCUSSION Degree of overlap between different surface markers of CD8 T cell subsets Previous studies have shown that several subsets of CD8 T cells can be identified using combinations of different surface markers, basically CD45RA and CCR7 or CD27 and CD28. As shown in Figure 1, four distinct subsets of CD8 T cell were identified using either CD45RA and CCR7 or CD27 and CD28. Based on previous studies, CD45RA+CCR7+ CD8 T cells are naïve cells, CD45RA-CCR7+ CD8 T cells are T central memory cells, CD45RA-CCR7CD8 T cells are effector memory cells and CD45RA+CCR7CD8 T cells are terminally differentiated T cells. Along the same line, CD28+CD27+CD8 T cells are naïve cells or antigen experienced cells in an early stage of differentiation, CD28-CD27+CD8 T cells are intermediate-differentiated and CD28-CD27-CD8 T cells are late-differentiated cells. Of note, CD28+CD27CD8 T cells, which represent 10-15% of total CD8 T cells (Figure 1), were poorly investigated. We then combined these four markers (i.e. CD45RA, CCR7, CD28 and CD27) to investigate the degree of overlap between each of them on CD8 T cells. As shown in Figure 1, we could identify more than 10 distinct CD8 T cell populations. We then further characterized the different T cell subsets by combining with an additional marker, such as CD7, CD127 or CD57, and observed an even more complex sub-division. Actually, up to 16-24 phenotypically distinct populations were identified (Figure 2). These data suggest that there is a very poor redundancy between each of these markers although they theoretically identify similar CD8 T cell subsets. Degree of overlap between different surface markers of virus-specific CD8 T cell subsets We then performed the same type of analysis on virus-specific CD8 T cells identified by peptide-MHC class I Tetramer complexes. Tetramers were generated for Influenza (Flu), Epstein Barr virus (EBV and Cytomegalovirus (CMV) peptides. Staining of Tetramer+ CD8 T cells with CD45RA and CCR7 showed that CD45RA-CCR7and CD45RA+CCR7were the dominant populations while CD45RA+CCR7+ and CD45RA-CCR7+ were poorly represented (Figure 3). Of note, CD45RA+CCR7were almost absent in Flu-specific CD8 T cells while it represented the majority of CMV-specific CD8 T cells. These data suggest that the phenotypic heterogeneity is reduced in the context of virus-specific CD8 T cells as compared to total CD8 T cells. Of interest, when we combined these stainings with CD27 and CD28, we observed the same phenomenon as in total CD8 T cell population: most of the subsets, included the ones which were very homogeneous (i.e. the CD45RA-CCR7-), were further sub-divided and we could identify up to 10 phenotypically distinct populations of virus-specific CD8 T cells. These data suggest that even on antigen-experienced CD8 T cells, there is a very poor degree of overlap between each of the above mentioned markers. Poster # 441P Gated on CD8 T cells