Fundamental to the concept of adaptive immunity is the ability of lymphocytes to recognize and respond to specific foreign Ags. B and T cells that are specific for peptides of an invading pathogen will proliferate and produce Abs, cytokines, and effector molecules and, upon resolution of the infection, form a long-lived memory response to protect against reinfection. However, for many decades, it was not possible for immunologists to track cells specific to a particular Ag over an extended time period. And although we have now delineated molecular markers of the many B and T cell subsets that occupy distinct spatial and functional niches, much of this progress arose from the physical ability to sort, classify, and track Ag-specific cells.
In these Pillars of Immunology articles, we highlight the foundational discoveries made by Murali-Krishna et al. (1) and Zajac et al. (2) in CD8+ T cells, which reshaped our understanding of Ag-specific T cell responses and how they differ between acute and chronic infection. This work began with the pioneering technology of Altman et al. (3) in Mark Davis’s laboratory, who developed fluorescently labeled tetrameric peptide–MHC class I complexes that could bind to TCRs with sufficient affinity to detect epitope-specific T cells by flow cytometry. Critically, “MHC tetramers” enabled precise enumeration and longitudinal monitoring of distinct Ag-specific populations during and after infection, as well as a method for cell purification using FACS.
Through much of the 1990s, the T cell community hotly debated a simple but fundamental question: during an active infection, when the T cell population rapidly expands, how many of those cells are actually specific to the pathogen (4–6)? Estimates at the time were that ∼1–5% of all T cells were Ag specific (4), whereas the remaining cells were believed to be “bystanders,” specific to other targets but spurred to action by inflammatory cues (7). These figures were largely based on the “limiting dilution assay,” which relies on in vitro expansion of T cell clones that recognize specific Ags (8). However, because this assay is dependent on T cell proliferation over several weeks, it is biased for cells whose differentiation states have greater proliferative capacity and therefore underestimates the true numbers of Ag-specific CD8+ T cells.
MHC tetramers proved to be a key technology in resolving the question of Ag-specific versus bystander T cell proliferation. Murali-Krishna et al. (1) used tetramers to show that the majority of T cells (50–70%) proliferating in response to infection are in fact Ag specific, at least 10-fold higher than the previous estimates. The authors corroborated this finding by showing that a similar percentage of CD8+ T cells produced IFN-γ when restimulated with virus-specific peptides. Further, sorting CD8+ T cells on the basis of tetramer staining yielded populations of cells that could specifically produce IFN-γ or kill target cells pulsed with the corresponding peptide in vitro. The use of MHC tetramer staining during viral infection provided the first accurate determination of immunodominance, the hierarchical ordering of epitope-specific T cell populations based on their proportions.
By following various Ag-specific CD8+ T cell populations longitudinally over time, Murali-Krishna et al. (1) clearly showed that the immunodominance hierarchy established during the initial wave of expansion was maintained in the memory pool. In other words, the relative kinetics of expansion and contraction of virus-specific CD8 T cells was highly similar across multiple epitopes. Then, to determine whether the recall response was entirely Ag specific or based on bystander cell contributions, the authors rechallenged lymphocytic choriomeningitis virus (LCMV)–immune mice with LCMV or vaccinia virus and found that reinfection with LCMV induced proliferation of Ag-specific cells, whereas infection with vaccinia did not. That is, a heterologous infection did not drive bystander proliferation in the LCMV-specific cells but, interestingly, the inflammation produced by vaccinia infection increased cytotoxicity among LCMV-specific cells. In a back-to-back paper, Butz and Bevan (9) similarly demonstrated the “massive” expansion of virus-specific CD8+ T cells in response to LCMV infection, specific effector CD8+ T cell responses to LCMV peptides in vitro, and the lack of response of LCMV-experienced cells to a heterologous vaccinia infection. Thus, being able to track and count Ag-specific CD8+ T cells enabled fundamental insights into the generation of effector and memory CD8+ T cells and their recall responses.
MHC tetramers also enabled unprecedented functional, genetic, and phenotypical dissection of pathogen-specific T cells. Follow-up studies used sorted, Ag-specific populations and began to define the phenotypes and gene expression signatures of different effector and memory subsets, including memory cell precursors (10, 11) and the transcriptional machinery that programmed these distinct differentiation states (12). At the same time, this more detailed understanding of Ag-specific CD8+ T cell responses catalyzed new insights into how dendritic cells and CD4+ T cells (13, 14) primed and supported an effective cytotoxic CD8+ T cell response (15).
Shortly after the work by Murali-Krishna et al. (1), another landmark study from the Ahmed laboratory (2) applied tetramers to dissect the mechanisms by which CD8+ T cell responses fail during chronic infection, and they made surprising observations. Previously, it had been reported that deletion of CD8+ cells specific to the LCMV peptide gp33–41 occurred simultaneously with a drop in cytotoxicity in vitro and failure to contain a chronic variant of LCMV virus in vivo (16). A logical interpretation of these observations was that the physical loss of Ag-specific cells in the host, which was originally referred to as T cell “exhaustion,” was the cause of viral persistence (16). However, in addition to CD8+ T cell deletion during chronic LCMV infection, Zajac et al. (2) identified an alternate form of T cell exhaustion wherein the virus-specific CD8+ T cells persisted but in a less functional state. More specifically, using MHC tetramers to track the different epitope-specific CD8+ T cells during viral infection, Zajac et al. (2) identified that the CD8+ cells specific to the immunodominant LCMV peptide NP396–404 were physically deleted during chronic infection. In contrast, the CD8+ T cells specific for gp33–41 epitope could be readily detected by MHC tetramers, but their effector functions in response to peptide restimulation could not. In fact, there was a notable loss in the ability of these gp33–41–specific T cells to produce cytokines such as IFN-γ. Likewise, another study by Gallimore et al. (17) published the same year found a loss of IFN-γ production among gp33–41–specific cells. Following these reports, the more contemporary use of the term CD8+ T cell exhaustion became used to refer to these less functional CD8+ T cells that were “tired out,” rather than the most extreme case of physically deleted cells. Together, these findings showed that the presence of chronic Ag is the primary factor in driving deletion or functional exhaustion and that CD8+ T cells adopt these distinct fates based on their epitope specificity.
Additionally, Zajac et al. (2) showed that CD4+ T cell help was vital for sustaining CD8+ T cell production of IFN-γ during chronic viral infection. Since then, CD4+ T cell depletion has become a mainstay procedure for inducing severe CD8+ T cell exhaustion. Further, the finding that CD4+-based help could reduce CD8+ exhaustion foreshadowed therapeutic opportunities for maintaining CD8+ function in cancer, even when chronic Ag persists.
Critically, Zajac et al. (2) laid the groundwork for future molecular studies on CD8+ T cell exhaustion, which led to a deeper and more accurate understanding that “exhausted” T cells are more than just shells of their former “more functional” selves. The field went on to find numerous hallmarks of T cell exhaustion, including the loss of production of inflammatory cytokines such as IFN-γ, TNF, and IL-2, high expression of inhibitory receptors such as PD-1, CTLA-4, and TIM-3, and rewired transcriptional and epigenetic programs (18, 19). Our current understanding is that T cell exhaustion occurs in a progressive fashion as activated CD8+ T cells transit from a “progenitor” to a “terminal” exhausted state (19–22). These findings refined the understanding of CD8+ T cell exhaustion as an adaptive differentiation process in response to chronic antigenic signaling, which helps to balance pathogen control with CD8+ T cell survival and reduced immunopathology.
From a therapeutic perspective, the identification of exhausted cells paved the way for the functional rejuvenation of “exhausted” cells by immune checkpoint blockade (23–25). Since then, rejuvenating anti-tumor T cells through checkpoint blockade has become a pillar of modern cancer therapy, and ongoing work seeks to further tune T cell effector functions to create superior anti-tumor T cells.
The works by Murali-Krishna et al. (1) and Zajac et al. (2) provided a foundation for profiling Ag-specific CD8+ T cells in the context of many different settings of infection, cancer, and autoimmunity and cancer. However, we face a challenge to identify the Ag-specific T cells in the settings of cancer and autoimmunity because the number of potential Ags across patients is considerably larger than those in specific pathogens. Perhaps, the biggest impact of interrogating T cell exhaustion has been in the area of cancer research and connecting the sequences and abundances of distinct TCR clonotypes with single-cell gene expression profiles. This integrative approach shows promise for understanding the landscape of Ag-specific responses among tumor-infiltrating T cells (26, 27). Indeed, as sequencing-based approaches expand our view of the heterogeneity of T cell responses, the shared features of Ag-specific responses illuminated by Murali-Krishna et al. (1), Zajac et al. (2), and Altman et al. (3) are all the more remarkable. By counting, tracking, and profiling the functional phenotypes of Ag-specific CD8+ T cells, these studies laid the groundwork for our modern understanding of T cell memory and exhaustion.
T.H.M. acknowledges support from the Damon Runyon Cancer Research Foundation (Fund DRG 2358-19). S.M.K. acknowledges support from the NIH (Grant R37AI066232).
Abbreviation used in this article
lymphocytic choriomeningitis virus
The authors have no financial conflicts of interest.