M. A. Williams, Department of Pathology, University of Utah, Salt Lake City, UT 84112, USA. Email: email@example.com Senior author: Matthew A. Williams
The initiation of a T-cell response begins with the interaction of an individual T-cell clone with its cognate antigen presented by MHC. Although the strength of the T-cell receptor (TCR) –antigen–MHC (TCR-pMHC) interaction plays an important and obvious role in the recruitment of T cells into the immune response, evidence in recent years has suggested that the strength of this initial interaction can influence various other aspects of the fate of an individual T-cell clone and its daughter cells. In this review, we will describe differences in the way CD4+ and CD8+ T cells incorporate antigen-driven differentiation and survival signals during the response to acute infection. Furthermore, we will discuss increasing evidence that the quality and/or quantity of the initial TCR-pMHC interaction can drive the differentiation and long-term survival of T helper type 1 memory populations.
Following exposure to an intracellular pathogen, T cells undergo numerous rounds of cell division, expanding up to 50 000-fold in the course of about a week.1,2 During this period of priming and expansion, they also differentiate into cytokine-producing and cytolytic effector cells and make fate decisions that govern their entry into the long-lived memory pool. While most effector T cells die in the weeks following antigen clearance, a small proportion survives and becomes long-lived memory T cells.3,4 Because cells fated for either memory differentiation or death can be identified at the peak of the primary response,5 numerous studies have attempted to identify signals delivered during the primary response that drive the effector/memory fate decision. Generally, these studies have either focused on the context, duration, quantity or quality of the antigen-specific signal delivered through the T-cell receptor (TCR) or the role of non-antigen-specific signals, such as growth or inflammatory factors, in driving differentiation of effector and memory T cells. The ability of any individual naive T cell to recognize peptide–MHC complexes (pMHC) via its TCR regulates its recruitment into the immune response but the subsequent role of TCR signals in driving expansion, effector differentiation and survival remains unresolved. We suggest that disparate results regarding TCR-driven T-cell differentiation can be largely attributed to differences in the way CD4+ and CD8+ T cells incorporate and respond to antigen-specific signals during the primary response to acute infection, and this review will focus particularly on the role of TCR signals in driving not only recruitment and expansion of CD4+ T cells, but also their acquisition of effector functions and ability to populate the long-lived memory compartment.
Extrinsic and intrinsic differentiation cues
Following exposure to antigen, T cells undergo a period of massive expansion, sometimes exceeding 50 000-fold, and acquire effector functions that enable them to co-ordinate pathogen clearance. After clearance, the majority of effector T cells die, leaving behind a population (5–10%) of long-lived memory T cells that provide enhanced protection from re-exposure to the same or a related pathogen.3,4,6,7 Memory T-cell populations maintain the ability to both survive independently of cognate antigen8 and self-renew in response to external homeostatic signals such interleukin-15 (IL-15) and IL-7,9,10 hence maintaining themselves at stable levels for many years. However, it has become increasingly clear in recent years that crucial T-cell fate determination events occur during the primary response, as short-lived effector T cells can be distinguished from memory precursor effector T cells during its later stages. For example, memory precursor T cells can be identified on the basis of expression of IL-7 receptor α,5,11 although IL-7 signals themselves may not be absolutely required in the memory fate decision.12–14 Furthermore, in the CD8+ effector T-cell compartment expression of a variety of natural killer cell receptors, most particularly KLRG1, has been associated with short-lived effector/effector memory differentiation and inversely correlates to entry into the long-lived effector/central memory pool.15
In addition, the effector and memory differentiation of T cells can be remarkably heterogeneous. Memory T cells can be broadly divided into central and effector memory subsets based on cell surface phenotype and tissue localization; the relative roles of these subsets in protection from secondary pathogen exposures has only been understood in recent years.16,17 For T helper type 1 (Th1) memory T cells, for example, one of the best predictors of protective capacity is the ability to make multiple cytokines [interferon-γ (IFN-γ), tumour necrosis factor-α and IL-2) immediately upon re-exposure to antigen.18,19 Whereas central memory cytotoxic T lymphocytes (CTL) are thought to have the best proliferative and protective capacity,20 in some models of acute infection and re-challenge, the activation phenotype of memory CD8+ T cells has proved to be the best predictor of protection from secondary challenges, independent of central or effector memory phenotype.21 These findings do not rule out a role for tissue-residing effector memory T cells in protection. Although they appear to be less stable as a population and have less proliferative capacity compared with central memory populations, they probably provide an important line of defence at the site of infection. Indeed, we would argue that the community dynamics of pathogen transmission would dictate that the most likely period of re-exposure to a pathogen are the months immediately following the initial exposure. The long-lived central memory compartment can ‘remember’ infectious history over the course of many years whereas the shorter-lived effector memory compartment provides a short-term memory of not only the nature of the pathogen but the site of the initial pathogenic challenge. Although we simplistically refer to effector/memory fate decisions in the current review, a full and accurate understanding of T-cell fate decisions will require a better understanding of the nature and function of memory T cells across the spectrum of their differentiation states. A primary goal in efforts to understand the biological processes that drive the simultaneous but asymmetric and heterogeneous differentiation of effector and memory T cells during the primary response should be to identify the nature of the fate determination signals, with obvious implications for a wide range of vaccination and immunotherapy strategies.
Efforts to understand why a few effector T cells go on to populate the memory pool while the majority die following pathogen clearance have focused on the role of external environmental cues (e.g. co-stimulation, inflammatory mediators, growth factors) as well as intrinsic differences in the ability of individual T-cell clones to respond (e.g. TCR avidity). In this, the differentiation of pathogen-specific CD4+ and CD8+ T cells differs in certain key respects. Once CD8+ T cells are recruited into the immune response against a foreign pathogen, all aspects of differentiation are enabled. Clonal progeny from initially recruited CD8+ T-cell responders differentiate into cytotoxic effector cells, cytokine producers and long-lived memory cells after only a brief period of selection22–24 over a wide range of TCR avidities for antigen.25 The repertoires of CD4+ primary and secondary effector T-cell populations also appear to be dominated by the preferential expansion of clones with high avidity TCRs following peptide or protein immunization.26–29 CD4+ T cells, however, also appear to undergo multiple stages of differentiation, with progressively stronger activation signals promoting not only their recruitment and activation, but also differentiation and survival.30,31
Environmental cues and memory T-cell differentiation
External signals play a key role in the recruitment, expansion and differentiation of antigen-specific effector and memory T cells. One recent study found that clonal recruitment of CD4+ T cells following peptide or protein immunization was acutely dependent on the presence and nature of an accompanying adjuvant.32 Inflammatory adjuvants promoted the recruitment of a T-cell repertoire with a wider range of individual TCR avidities for antigen following peptide immunization.32 Even though only a short period of antigenic stimulation (6–12 hr) is required in vivo for recruitment of antigen-specific CD8+ T cells, their optimal expansion requires the presence of a robust, infection-induced inflammatory response.23 Furthermore, in the context of a Listeria monocytogenes infection antigen-specific T cells were efficiently recruited into the immune response across a 700-fold range of TCR avidities. Interestingly, however, the extent and kinetics of expansion were proportional to the avidity of the TCR-pMHC interaction,25 suggesting that whereas low-avidity TCR signals can mediate recruitment, stronger and/or prolonged TCR signals probably influence the final clonal composition of the effector CTL pool.
The inflammatory environment also profoundly impacts the differentiation of effector function and memory potential. For example, high levels of IL-12 or type I IFN signalling have been implicated in the preferential differentiation of effector CTL,15,33,34 at least in part through the graded up-regulation of the T-box transcription factor T-bet.15 Interleukin-2 is another cytokine implicated in driving effector and memory CTL differentiation.35,36 In particular, activation of T cells in the absence of IL-2 results in a decrease in primary effector function in response to acute and chronic infection and a severe impairment in the ability of resulting memory T cells to generate secondary responses upon re-challenge.37,38 A related cytokine, IL-21, has also been found to drive effector T-cell differentiation during chronic infection,39–41 though whether it plays a similar role to IL-2 in promoting primary and secondary effector T-cell differentiation during acute infection remains to be determined.
Recent attempts to understand the fate decisions that T cells make during the primary response have revolved around the function of several transcription factors whose expression has been linked to exposure to growth and inflammatory cytokines. The T-box transcription factors T-bet and Eomes have been shown to promote CD8+ effector T-cell differentiation,15,42 and high levels of T-bet activity are associated with terminal differentiation into short-lived effector T cells.15 Blimp-1, a zinc finger transcription factor required for the differentiation of antibody-secreting plasma cells,43 has recently been found to play a role in the effector differentiation of T cells responding to either acute or chronic viral infection.44–46 Conversely, the transcriptional repressor Bcl-6 has been reported to promote lymphocyte differentiation and survival of memory T cells.47–49
The role of these transcription factors is complicated by the observation that relatively small differences in activity can have large consequences for differentiation outcomes. For example, high levels of inflammatory cytokines such as IL-12 or type I IFNs have been shown to induce strong T-bet activity in CD8+ T cells and promote entry into a terminal differentiation pathway. On the other hand, activation in the complete absence of T-bet results in dysfunctional memory T cells, perhaps partly because of poor expression of IL-15Rβ (CD122).15 Similarly, whereas either strong or prolonged IL-2 signals preferentially promote CD8+ T-cell effector differentiation,35,36 the complete absence of IL-2 results in dysfunctional memory cells that are unable to re-enter the effector differentiation pathway upon reactivation.37,38 In these and other cases, CD8+ memory T-cell differentiation seems to follow the ‘Goldilocks’ principle in that they require just the right amount of effector differentiation stimuli. Certain activation stimuli are required for CD8+ memory T-cell differentiation, whereas an over-abundance of these stimuli leads to committed effector CTL differentiation, and a complete absence of these stimuli leads to defects in both effector and memory T-cell differentiation. The full nature and timing of the stimuli remain controversial. One recent study proposed that CD8+ memory T-cell differentiation might result from a first asymmetric division directly subsequent to priming events.50 Another recent report demonstrated that CD8+ memory T cells capable of robust secondary replicative function develop from precursors that have undergone some level of effector differentiation.51,52 In all, a pressing question is when during the response do CD8+ T cells become committed to either a memory or effector differentiation pathway?
Remarkably, far less is known regarding the general nature of fate decisions that CD4+ T cells make during the primary response to acute infection, and even less is known regarding the role of the accompanying inflammatory environment. T-bet has a well-described IL-12-dependent function in driving polarization of IFN-γ-producing Th1 responders.53 However, remarkably little is known regarding the role of T-bet in the differentiation and survival of long-lived CD4+ memory T-cell populations, or whether CD4+ T cells themselves even undergo the same sort of terminal differentiation pathway that has been observed for plasma cells and now CD8+ effector T cells. Blimp-1 plays a role in repressing genes required for Th1 differentiation, including T-bet,54 and Bcl-6 may be important for Th1 memory differentiation,55 but little is known regarding how the activity of these transcription factors in Th1 cells is modulated in response to their external environment and how they shape effector and memory Th1 cell differentiation.
TCR-driven differentiation: CD8+ T cells
The relative roles of antigen-mediated signals through the TCR and non-specific signals such as inflammatory cytokines and growth factors in driving T-cell differentiation are still being defined, and compelling evidence exists for both scenarios. Two broad models have been proposed to describe the required duration of antigen presentation in driving the recruitment, expansion and differentiation of T cells during their response to intracellular pathogens. The first, termed progressive differentiation, posits that sequential encounters with antigen progressively promote cell division, enhanced survival and functional differentiation.31 An extended period of antigen presentation would be required for the full differentiation. The second, termed programmed differentiation, proposes that the differentiation of T cells is programmed upon a brief initial period of antigen encounter.56 Although there is substantial evidence for each model, neither comprehensively describes what has been observed for in vivo CD4+ T-cell responses.
For CD8+ T cells, a short encounter with dendritic cells presenting antigen is sufficient to induce downstream stages of differentiation. As little as 6–24 hr is sufficient to recruit CD8+ T cells into the immune response directed toward a foreign pathogen. Additionally, once recruited, CD8+ T cells are capable of undergoing dividing, developing effector functions and differentiating into memory cells independent of further antigen encounter.22–24 One recent study found that while the recruitment of antigen-specific CD8+ T cells into the immune response can be accomplished with a short exposure to antigen, optimal expansion is subsequently driven primarily by the inflammatory environment.23
Although TCR signals, as regulated by TCR avidity and the abundance of cognate antigen, drive recruitment and expansion of primary CTL, the precise role of TCR signals in promoting memory differentiation remains controversial. Studies in multiple model systems have suggested that the TCR repertoire of effector CTL is widely dispersed in tissues and similar to that of both memory and secondary effector CTL,57–61 suggesting that once a clonal population is represented in the expanded effector CTL pool, it maintains no intrinsic competitive advantage in proceeding to the memory pool. Likewise, another recent study found that whereas changes in CTL TCR avidity for antigen across a 700-fold range impacted the kinetics and magnitude of clonal expansion, the size of the resulting memory population remained proportional to the peak of the effector pool.25 Similarly, while altering antigen dose impacted CTL expansion, it did not alter the clonal composition of the effector pool.62 Indeed, a single naive CD8+ T-cell precursor can give rise to heterogeneous effector and memory CTL differentiation that largely mimics that of polyclonal responders,63,64 again suggesting that TCR signals do not play a dominant role in differentiation outcome. Conversely, one recent report found that disruption of TCR signalling could have differential effects on T-cell fate. In this study, mutations to the TCR-β transmembrane domain that prevented efficient synapse formation, and attenuated TCR signalling allowed efficient effector CTL differentiation but poor memory CTL differentiation.65
One question arising from these studies is whether the extent of expansion is influenced by a short period of high avidity interactions between the TCR and pMHC or a prolonged period of TCR-pMHC interaction throughout the primary expansion phase. At least two pieces of evidence suggest that the former possibility is likely. First, direct visualization of the first few days of the in vivo T-cell response suggests that stable T-cell–antigen-presenting cell (APC) interactions are largely confined to the first couple of days following antigen challenge.66 Second, experiments in which antigen is presented to T cells for variable amounts of time in the presence or absence of inflammation indicate that the determining factor for robust clonal expansion is the persistence of the inflammatory environment following recruitment of T cells into the response.23 Therefore, it seems likely that the ability to undergo robust clonal expansion is dictated by the quality or quantity of the TCR-pMHC interaction at the initiation of the response, whereas the infectious inflammatory environment plays a central role in driving that expansion. Conversely, CD8+ T cells recruited in a limited or non-inflammatory environment undergo less initial expansion but retain or quickly recover the ability to undergo secondary expansion upon immediate re-challenge.67–69
TCR-driven differentiation: CD4+ T cells
For CD4+ T-cell responses the role of antigenic signal strength is more complex. The selection of Th1 precursors into the response is influenced by TCR avidity for peptide/MHC class II, and certainly at some level this influences the repertoire of the effector pool.26 CD4+ T cells, in contrast to CD8+ T cells, also require extended or repeated contacts with antigen during the first few days of the response for full expansion.70,71 Evidence also suggests that secondary stimulation of CD4+ memory T cells continues to skew the TCR repertoire towards higher avidity responders, an observation not seen for CD8+ T cells.29 These observations suggest that CD4+ T cells are faced with a longer period of selection, during both primary and secondary responses, on the basis of their ability to bind antigen. Other studies have also supported a model in which antigen signals drive progressive differentiation of CD4 responders, with increasing signals driving first expansion, then effector function and survival.30,31,72–74 Polyclonal CD8+ T cells, on the other hand, require a short window of antigenic stimulation in vivo (12–24 hr).75 The magnitude of expansion is impacted by the duration or strength of the initial stimulus but T cells that are recruited into the response readily go on to form memory populations capable of robust secondary responses.22 One interpretation of these results is that once CD8+ T cells reach a certain activation threshold, all phases of differentiation are enabled, with non-antigen-specific signals promoting differentiation of end-stage effector and memory precursor populations. CD4+ T cells, on the other hand, with their extended requirement for exposure to antigen, demonstrate a consistent skewing towards high-avidity TCRs throughout the effector response and following subsequent re-challenges.
Analysis of TCR repertoires has also suggested a role for the TCR in driving CD4+ T-cell differentiation. The TCR repertoire of antigen-specific T cells narrows to progressively higher avidity throughout the primary response and during subsequent re-challenges,27,29 suggesting that CD4+ T-cell responses undergo selection for high-avidity clones in the presence of antigen. Furthermore, the avidity of CD4+ effector and memory T cells is dependent on the initial antigen dose.28 On the other hand, the role of antigen in driving T-cell differentiation is not solely dependent on high avidity TCR. Clonal populations of CD4+ T cells can undergo functional avidity maturation throughout the primary response, resulting in the emergence of T cells with heightened sensitivity to antigen even as the TCR itself remains fixed.2 One possible explanation for these results is that functional avidity is hardwired into the response on the basis of the quality of the initial interaction with APC. In this scenario, one could envisage heterogeneity in the abundance of antigen, the surrounding microenvironment, the type of APC and the activation status of that APC leading to a spectrum of functional avidity in a monoclonal T-cell population. Conversely, non-TCR-specific mechanisms may promote the acquisition of heightened TCR signal sensitivity throughout the primary response, leading to the selection of effector clones in a largely TCR-independent manner.
Recent work has demonstrated a role for TCR–antigen interactions in driving not only effector differentiation but also survival into the memory pool. The initial clues that this might be the case were derived from clonal competition experiments. Increasing the frequency of antigen-specific CD4+ T cells at the initiation of the response inversely corresponded to memory differentiation potential. At very high clonal frequencies of TCR transgenic T cells, presumably accompanied by fierce competition for available antigen, the differentiation of virus-specific Th1 memory was almost entirely impaired.76 Clonal competition also impacted the long-term maintenance of Th1 memory cells.77 Competition during the primary CD4+ T-cell response for factors other than TCR, such as IFN-γ, has also been shown to impact the quantity and quality of ensuing memory T-cell populations.78
Our recent findings have also indicated that not all CD4+ T-cell clones that undergo massive expansion and effector differentiation in response to acute infection are capable of populating the memory pool.79 Small numbers of adoptively transferred SMARTA TCR transgenic T cells, which are specific for the lymphocytic choriomeningitis virus (LCMV) glycoprotein-derived immunodominant CD4+ T-cell epitope GP61–80, effectively mimic endogenous CD4+ T-cell responses in the same host following LCMV infection. SMARTA cells respond quite differently, however, following challenge with recombinant Listeria monocytogenes expressing the same epitope (Lm-gp61). Although SMARTA cells expand robustly initially, the resulting Th1 effector cells demonstrate partially impaired cytokine-producing capabilities. Furthermore, they are unable to progress to the memory pool and disappear entirely in the weeks following pathogen clearance, despite the efficient development of polyclonal endogenous CD4+ memory T cells directed toward the same epitope in the same host.79
We found that memory differentiation potential corresponded to both structural and functional TCR avidity at the peak of the effector response. In LCMV-infected hosts, SMARTA effector cells demonstrated similar avidity to polyclonal responders in the same host. Conversely, in Lm-gp61-infected hosts SMARTA effector cells displayed lower TCR avidity than endogenous responders. One possibility, therefore, is that their failure to enter the memory pool was reflected in their poor sensitivity to antigen during the primary response. In support of this, polyclonal CD4+ memory T-cell populations skewed to a higher functional avidity in the transition from effector cells (1 week post-infection) to memory cells (6 weeks post-infection). Furthermore, in the months following infection, CD4+ memory T cells consistently skewed to a higher functional avidity.79 It has been reported that CD4+ memory T cells decline over time,80 an observation that we make in our own studies. However, we found that the rate of decline decreases over time (6 months to 1 year), corresponding to the emergence of CD4+ memory T cells with high functional avidity.79 One possibility, therefore, is that CD4+ memory T-cell populations eventually stabilize following acute pathogenic challenge, resulting in highly functional high-avidity memory cells capable of long-term protection. This would be consistent with observations of human populations in which vaccine-induced CD4+ memory T-cell populations can persist for up to 75 years with a half-life of 8–15 years.81
While much of the focus regarding CD4 memory differentiation has been on the signals during the primary response that promote memory differentiation, another line of research has sought to identify the factors that promote massive cell death during the contraction phase. In our studies, we found that the pro-apoptotic molecule Bim was highly up-regulated in ‘doomed to die’ SMARTA Th1 cells following Lm-gp61 infection, providing a possible link between weak antigenic signals and effector cell elimination following pathogen clearance.79 Bim and Bcl-2 play opposing roles in promoting T-cell survival during various stages of differentiation.82 In particular, Bim activity has been shown to mediate the death of end-stage effector CD4+ and CD8+ T cells during the contraction phase of the response.83–85 However, less is known about the factors controlling the activity of Bim. The decision to enter a Bim-mediated death pathway is probably made before the contraction phase, as IL-7 receptor αhi memory precursor cells at the peak of the response are largely spared Bim-mediated apoptosis.85 Our finding that Bim expression is associated with SMARTA responders that have low functional avidity compared with the endogenous CD4+ T-cell response suggests that TCR signals may influence the ability of the Th1 effector cells to survive into the memory phase of the response through regulation of Bim activity. We further found that ‘doomed-to-die’ SMARTA effectors induced by Lm-gp61 expressed higher levels of the transcription factor FoxO3a, compared with their LCMV-activated counterparts. FoxO3a has been shown to regulate the survival of CD4+ memory T cells and promotes apoptosis in T cells.86–88 One target of FoxO3a is Bim, leading us to hypothesize that FoxO3a may also play a role in regulating the survival of CD4+ T cells in the transition from the effector to the memory pool in a Bim-dependent manner.
Hierarchical CD4+ T-cell differentiation
The above-described differences in the role of antigen-driven TCR signals in the expansion, differentiation and survival of CD8+ and CD4+ T cells responding to acute infection show that while all aspects of CD8+ T-cell differentiation are enabled, or ‘programmed’, following the initial activation/recruitment event, CD4+ T-cell differentiation is hierarchical, with increasing antigenic signals enabling progressively enhanced expansion, effector differentiation and survival. We hypothesize a model in which CD4+ memory T-cell differentiation and the ability of CD4+ memory T cells to survive long-term during the memory maintenance phase are enabled, at least in part, by the strength of the antigenic signal during the primary response (Fig. 1). Several studies have analysed the evolution of the TCR repertoire during the CD8 response to acute infection and found that that effector repertoire and the memory repertoire were similar.57,61 However, even though a strong rationale exists for a higher dependence on antigen for the differentiation of CD4+ memory T cells, no analogous studies of in vivo infection-induced CD4+ memory T-cell repertoires as they develop and then decline over time have been performed. Furthermore, previous studies have largely used broad methods for analysing clonotype distribution, such as analysis of V-region or J-region subsets or spectratyping analysis of CDR3 length distribution, rather than analysis of the distribution of single clonal populations through individual CDR3 sequences, as measured by deep sequencing analysis. A primary focus of current research efforts should be to determine if, on a clonal level, high-avidity TCR-pMHC interactions during the primary response result in preferential differentiation of long-lived CD4+ memory T cells.
Several other possibilities exist. First, strong TCR-pMHC interactions may be important during the memory maintenance phase after antigen has been cleared rather than during the primary response. It is possible, for example, that clonotypes with strong avidity for cognate antigen presented by MHC class II also have higher avidity for self MHC class II molecules in general. Recent support for a role for MHC class II interactions in promoting CD4+ memory T-cell survival was provided by the finding that the half-life of CD4+ memory T cells was inversely proportional to their frequency, although how these findings apply to diverse polyclonal memory populations with presumably varying specificities for endogenous peptides remains unclear.77 A second possibility is that the high functional avidity of CD4+ memory T cells compared with the effector populations that they are derived from is unrelated to actual TCR avidity but is instead a property acquired upon receipt of non-antigen-specific memory differentiation/maintenance signals. Monoclonal CD4+ and CD8+ T cells have been shown to acquire higher functional avidity throughout the effector response, and a cardinal feature of memory T cells is their lower activation threshold.2,89 Finally, it is possible that antigen retained after pathogen clearance could continue to shape the TCR repertoire in ways similar to that seen during chronic antigen exposure. In support of this, it has been shown that antigen retained in germinal centres can continue to shape the CD4+ memory T-cell compartment for several weeks after clearance.72 A careful and thorough analysis of Th1 TCR repertoire evolution during the effector response, in the transition from effector to memory and during the long-term maintenance of memory in the absence of continued antigen stimulation, is needed to ascertain whether and to what extent the strength of the initial TCR-pMHC interaction influences the long-term fate of individual T-cell clones and their daughters.
The authors declare no financial or conflict of interest.