Dr D. L. Farber, Department of Surgery, University of Maryland School of Medicine, MSTF Building, Room 400, 685 W. Baltimore St., Baltimore, MD 21201, USA. Email: firstname.lastname@example.org Senior author: Donna L. Farber
The development of immune memory mediated by T lymphocytes is central to durable, long-lasting protective immunity. A key issue in the field is how to direct the generation and persistence of memory T cells to elicit the appropriate secondary response to provide protection to a specific pathogen. Two prevailing views have emerged; that cellular and molecular regulators control the lineage fate and functional capacities of memory T cells early after priming, or alternatively, that populations of memory T cells are inherently plastic and subject to alterations in function and/or survival at many stages during their long-term maintenance. Here, we will review current findings in CD4 T-cell memory that suggest inherent plasticity in populations of memory CD4 T cells at all stages of their development – originating with their generation from multiple types of primed CD4 T cells, during their persistence and homeostatic turnover in response to T-cell receptor signals, and also following secondary challenge. These multiple aspects of memory CD4 T-cell flexibility contrast the more defined lineages and functions ascribed to memory CD8 T cells, suggesting a dynamic nature to memory CD4 T-cell populations and responses. The flexible attributes of CD4 T-cell memory suggest opportunities and mechanisms for therapeutic manipulation at all phases of immune memory development, maintenance and recall.
The generation and persistence of immunological memory after pathogen encounter provides the basis for an efficient immune response, as previous immune activations are recorded and stored on the cellular and molecular level. Memory T cells direct and co-ordinate efficacious secondary immune responses through their enhanced functional, activation and migration properties compared with naive T cells. These enhanced immune responses ensure that an individual is protected from succumbing to repeated pathogen infections over a lifetime. In mouse models, memory CD4 and CD8 T cells mediate protective immunity to bacterial and viral pathogens, although in humans the ability to stimulate memory T-cell development and modulate their function in vaccines and anti-pathogen immunity remains elusive. For this reason, much of the current research on immune memory is concentrated on understanding mechanisms for the generation and persistence of functional memory T cells.
The life cycle of memory T cells can be broken down into three phases. The first phase comprises the initial generation of memory T cells following priming of naive, antigen-specific T cells with antigen and the appropriate co-stimulatory signals, resulting in their expansion and differentiation to effector cells within the first 7–9 days of infection. Following this acute phase, a large proportion of effector cells die and a population of surviving memory T cells emerges, with a decreased activation threshold and high functional capacity. During the second phase, memory T cells persist via steady-state homeostatic turnover in response to T-cell receptor (TCR) and/or cytokine-mediated signals. The final phase consists of the reactivation of memory T cells, leading to efficacious secondary responses.
A key issue in the field is how to direct the generation and persistence of memory T cells to elicit the appropriate secondary response to provide protection to a specific pathogen. Currently two prevailing views have emerged; that cellular and molecular regulators control the lineage fate and functional capacities of memory T cells early after priming, or alternatively, that populations of memory T cells are inherently plastic and subject to alterations in function and/or survival at many stages during their long-term maintenance.
The majority of studies on memory T cells have focused on CD8 T cells mainly in the lymphocytic choriomeningitis virus infection model. These studies have defined some key molecular regulators of CD8 effector and memory T-cell fate. By contrast, mechanisms for memory CD4 T-cell development and persistence appear to be distinct from CD8 T cells and with less defined pathways and mediators. Moreover, CD4 T cells are required for the generation of memory CD8 T cells and are central regulators of secondary immune responses. Here, we will review the state of current findings in CD4 T-cell memory that suggest inherent plasticity in populations of memory T cells at all stages of their life cycle – originating with their development from primed CD4 T cells, dynamically altering over time during their persistence and homeostasis, and also following secondary challenge. This flexibility in memory CD4 T-cell development and functional maintenance suggests opportunities for therapeutic manipulation at all phases of immune memory development, persistence and secondary responses.
Memory CD4 T-cell generation: multiple precursors
The key issue in dissecting mechanisms for the generation of memory T cells is to define the signals and properties that distinguish short-lived effector T cells from memory T cells, which survive long-term. In studies on anti-viral effector and memory CD8 T cells, it is possible to isolate precursors to relatively short-lived effector cells (SLEC) and memory T cells (memory precursor effector cell; MPEC) based on expression of effector markers (T-bet, CD27), receptors for survival cytokines (interleukin-7 receptor; IL-7R), and markers associated with apoptosis (KLRG-1).1 A number of groups have recently shown that altering inflammation and effector markers,2,3 IL-2 responses,4,5 and cytokine signalling or transcriptional regulation6–9 lead to variations in the proportion of SLEC compared with MPEC. These findings indicate that effector or memory CD8 T-cell fate is determined at priming, which ultimately sets the extent of memory CD8 T-cell persistence and the potency of secondary CD8 T-cell-mediated immune responses. In response to different infections, SLEC and MPEC can both exhibit long-lived persistence,10 suggesting that these populations may also exhibit some flexibility during their maintenance in vivo.
In contrast to CD8 T cells, there are as yet no defined precursors to effector and memory CD4 T cells, and the SLEC and MPEC designations have no direct correlates for CD4 T cells. Importantly, the production of effector cytokines does not seem to be a controlling factor for the development of memory CD4 T cells as shown in a number of studies. Mouse models in which interferon-γ (IFN-γ) -producing cells were genetically marked in vivo or isolated before adoptive transfer revealed memory CD4 T-cell generation from both IFN-γ+ and IFN-γ− populations.11,12 Similarly, antigen-specific CD4 T cells that were primed for only short periods of time and consequently did not acquire effector function, were still able to develop into memory CD4 T cells in vivo.13 These data suggest that the precursors to CD4 T-cell memory consist of a heterogeneous population of cells present within effector cytokine-producing and non-producing populations. The type of effector cells generated during priming may affect subsequent mechanisms for survival as memory T cells. Recently CD44 expression was shown to be required for memory CD4 T-cell generation from T helper type 1 (Th1) effector cells, but not Th2 or Th17 cells, because of its regulation of Fas-mediated apoptosis specifically in the Th1 subset.14 These findings suggest that while multiple activated CD4 T-cell precursors may survive to become long-lived memory CD4 T cells, the specific mechanisms for their survival may differ.
Cytokines of the γc family can also affect memory T-cell generation and survival. Interleukin-7 is a known survival cytokine for T lymphocytes,15,16 and precursors to memory CD8 T cells have been isolated based on up-regulation of IL-7R expression.17 While memory CD4 T cells have been shown to require IL-7 for long-term persistence,18 studies so far do not implicate IL-7R as a marker for memory CD4 T-cell precursors. While IL-7R is down-regulated after CD4 T-cell activation, it also can be rapidly up-regulated from IL-7R-negative activated T cells,13 indicating that IL-7R down-regulation does not limit the potential to respond to IL-7 survival signals as memory T cells. Moreover, memory CD4 T-cell generation is not appreciably altered in vivo by fixed expression of the IL-7R or exogenous IL-7,19,20 suggesting that IL-7 is not a predominant controlling factor for memory CD4 T-cell development.
Interleukin-2 has been shown to differentially drive effector or memory CD8 T-cell development,21 with high IL-2 responses driving differentiation of terminal effector CD8 T cells.4,5 By contrast, for memory CD4 T cells, increased responses to IL-2 quantitatively enhance the generation of both effector and memory CD4 T cells. Increased IL-2 production during priming correlated with increased frequency of memory CD4 T cells22 and inhibited the apoptosis of effector CD4 T cells.23 In addition, precursors to memory CD4 T cells all up-regulate the IL-2R, CD25,13,24 in contrast to CD8 T cells where CD25 expression marks different effector and memory T-cell fates.4 Hence, although IL-2 has quantitative effects on overall memory CD4 T-cell yield, it does not serve as a controller of cell fate.
Other factors that affect priming and development of memory CD4 T cells are the duration/dose of antigenic priming. Unlike CD8 T cells, which require only limited contact with antigen for initiating effector and memory development,25,26 CD4 T cells require more sustained activation with antigen to promote full activation.27 Prolonged antigen exposure during priming has been shown in several studies to increase the proportion of memory CD4 T cells generated,13,28 although short-term activated T cells can also develop into memory populations.13 When taken together, the flexible requirements for priming conditions and differentiation suggest multiple precursors to memory CD4 T cells (Fig. 1) as we have also previously suggested.13,29 While primed/effector CD4 T cells can undergo activation-induced cell death in response to repeated or robust stimulation, below this threshold and above a minimum activation requirement, there appears to be a broad range of cellular activation states that can give rise to memory CD4 T cells (Fig. 1). Memory CD4 T-cell differentiation can therefore be viewed as multiple branched pathways from activated precursors to memory T cells surviving via homeostasis (Fig. 1 and13). The signals for branched memory T-cell development may be set during the initial cell division, leading to asymmetric apportioning of effector molecules to daughter cells,30 or alternately, via the cessation of TCR signals that occurs following removal from antigen, as ‘rested’ effector CD4 T cells were shown to have similar gene expression profiles as in vivo-generated memory CD4 T cells.31 These broadly defined requirements for precursors to memory CD4 T-cell generation can predispose the population to functional diversity.
Memory homeostasis and persistence
Once generated, memory CD4 T cells can persist for up to the lifetime of an individual via long-term homeostasis and turnover. In humans, anti-viral memory CD4 T cells have been shown to persist for longer periods relative to memory CD8 T cells, which decay more quickly.32,33 The very nature of their long-term maintenance suggests the potential for dynamic alteration of memory T cells over time based on their perception of signals and subsequent responses. Whether memory T cells require antigen and/or non-cognate T-cell stimulation has been extensively investigated. Memory CD8 T cells have been shown to persist independent of TCR signalling and major histocompatibility complex (MHC) class I engagement,34,35 and to be more dependent on homeostatic cytokines such as IL-15 and IL-7 for their survival and homeostasis. By contrast, memory CD4 T cells appear to have more stringent requirements for TCR engagement for their persistence/homeostasis. Previous studies showed functional deterioration of memory CD4 T cells in MHC class II-deficient hosts,36,37 and a role for TCR signalling in optimal persistence of memory CD4 T cells.38 We recently showed that ablation of TCR signals in memory CD4 T cells through conditional deletion of the gene encoding SLP-76, a key TCR-coupled linker-adapter signalling molecule, resulted in almost complete inhibition of memory CD4 T-cell homeostatic turnover both in steady-state and lymphopenic conditions, and a corresponding diminution in their persistence.39 The addition of excess levels of IL-7 in vivo could not restore wild-type levels of homeostatic turnover in SLP-76-deficient memory CD4 T cells,39 suggesting that TCR signals are the predominant regulators of memory CD4 T-cell homeostasis. Further studies will be needed to precisely define how cytokines such as IL-7 and IL-15, which are also implicated in memory CD4 T-cell survival and/or homeostasis,18,38,40 contribute to memory CD4 T-cell homeostatic turnover in conjunction with TCR-mediated signals.
The requirement for TCR signalling and MHC class II engagement for memory CD4 T-cell homeostasis has important implications. Continuous TCR engagement implies that signals are being dynamically perceived during memory CD4 T-cell maintenance with potential effects on functional and/or survival capacities. Memory CD4 T-cell turnover driven by TCR engagement can result in selective survival of specific clones as a result of their propensity to receive TCR signals, resulting in an overall narrowing of the TCR repertoire of memory CD4 T cells over time. Indeed, expanded clones of memory CD4 T cells have been detected in aged individuals, supporting this prediction.41 This perception of T-cell signals during memory CD4 T-cell homeostasis can also result in functional alteration of these cells. As lymphopenia-driven homeostatic proliferation of naive T cells results in their conversion to memory phenotypes and functions,42,43 steady-state homeostasis of memory T cells, albeit at a slower rate, may also have functional consequences over time.
Memory CD4 T-cell heterogeneity
Populations of memory T cells persist in diverse subtypes in lymphoid and non-lymphoid tissues. Over 10 years ago, two major memory T-cell subsets were delineated based on expression of the lymph node homing receptors CD62 ligand (CD62L) and CCR7, which enable entry into lymph nodes through high endothelial venules.44–46 The CD62Llo CCR7lo memory subtype is defined as effector-memory (TEM) while the CD62Lhi CCR7hi-expressing subset is central-memory (TCM). The initial finding in human peripheral blood CD4 T cells identified the TEM subset as producing effector cytokines, with the TCM subset predominantly producing IL-2.45,47 However, in subsequent studies of antigen-specific subsets in vivo, both TEM and TCM populations were found to have effector function,48–51 although the TCM subset has a greater proliferative capacity compared with the TEM subset in both humans and mice.47
Whether the TEM/TCM designations represent stable populations or populations that are dynamically altered during their persistence in vivo is not yet resolved. However, there is evidence that homeostatic proliferation of TCM results in both maintenance of TCM subsets and conversion of some progeny to TEM phenotypes, mostly based on results with CD8 T cells.52,53 Conversely, TEM can convert to TCM cells during persistence or following activation for CD8 or CD4 T cells,50,54 indicating that TEM/TCM subset designations are not fixed. Moreover, the phenotypic markers that define TEM/TCM subsets do not encompass all of the multiple phenotypic variations seen in memory CD4 T cells (for a review, see ref. 51). Notably, in humans, memory CD4 T cells were found to differ in a variety of chemokine receptors (distinct from CCR7), associated with different replicative histories as revealed by telomere length.55 The complexities of phenotypic variation in memory T cells could arise either during their initial generation from multiple precursors or through changes and signals perceived during their homeostatic maintenance.
Memory T-cell diversity in homing/chemokine receptor expression reflects their diverse capacities for trafficking and residence in multiple lymphoid and non-lymphoid tissue sites, including spleen, lung, liver, gut and bone marrow.56–58 Although expression of specific chemokine receptors can be associated with T-cell trafficking to certain tissue sites, such as intestine and skin,59–61 mechanisms directing populations of memory T cells into non-lymphoid tissue sites remain largely undefined. In mice, dendritic cells from specific tissue sites can induce expression of tissue-specific homing molecules,60,62–66 suggesting that priming conditions can direct memory T cells to specific tissues. Memory T cells resident in distinct tissues also have specific functional capacities with memory CD4 and CD8 T cells in gut and bone marrow having more effector-like properties,67–71 and lung memory CD4 T cells having distinct homing properties.48 The homing profile of memory cell subsets is also predictive of functional capacities, with specific chemokine receptors appearing coincident with Th1, Th2, Th17 and T follicular helper functional T-cell subsets.57,72 These findings suggest that tissue-specific factors play a significant role in tuning the memory response in diverse anatomical sites.
Whether tissue-homing memory CD4 T cells further differentiate or can migrate back to the lymphoid compartment has not been resolved; however, evidence points to peripheral tissue-resident memory CD4 T cells as end-stage memory subsets. In general, non-lymphoid resident memory T cells have more effector-like properties,68,73 and TCM cells with primary residence in lymphoid tissue convert to TEM phenotype cells in non-lymphoid sites,53 supporting a role for lymphoid memory in seeding non-lymphoid sites through further differentiation. Furthermore, repeated boosting of antigen-specific memory populations results in their preferential residence in non-lymphoid sites,74 indicating that increased differentiation is associated with peripheral resident memory T cells. Given the large fraction of persisting memory CD4 T cells residing in tissue sites, dissecting the mechanisms and specific responses of tissue-resident memory CD4 T cells will be an important focus for future research to target the generation or specific homing of memory T cells to the site of pathogen entry.
Environmental regulation during secondary activation
The idea of T-cell effector function as an indication of irreversible cellular differentiation to defined subsets has been re-evaluated in the context of a large body of evidence demonstrating significant plasticity in previously committed effector T-cell trafficking, cytokine production and transcriptional regulation. Indeed, effector T cells previously thought to represent fully differentiated states have now been clearly shown to be capable of co-expression of alternative effector molecules, or even of full conversion to wholly new phenotypes. Although expression of transcriptional regulators, such as T-bet or GATA-3, clearly directs cells towards Th1 or Th2 functional capacities, respectively,75–77 they do not necessarily result in terminal differentiation as previously thought.78–80
Memory CD4 T-cell responses have been shown to maintain the effector functions from the primary response, based on studies showing conserved Th1 or Th2 cytokine profiles from populations of polarized Th1 or Th2 effector cells, respectively.81,82 For CD8 T cells, expression of T-bet and effector cytokine production is conserved in effector and memory responses.2 For memory CD4 T cells however, this functional capacity does not appear to be irreversibly committed. We initially demonstrated functional plasticity with a population of memory CD4 T cells that produced predominantly IFN-γ in the primary response, but could be differentially activated in the secondary response by altering the nature of the TCR stimulus to produce predominantly IL-4.83 Human memory Th1 and Th2 cells were similarly found to have flexible functional profiles upon recall.84 Inflammation and cytokine environment can likewise alter the outcome of memory CD4 T-cell responses with pro-inflammatory environments promoting Th1-like cytokines from populations of Th2 or uncommitted memory CD4 T cells,12,85 and Th1-like memory CD4 T cells producing Th2 cytokines in a Th2 cytokine environment.86 Continued environmental input was demonstrated to be necessary for induction of Th1 effector functions even following memory induction from a strongly Th1 primary response.85 Together these studies suggest that during activation of secondary CD4 T-cell responses, environmental cues can adjust the functional outcome, and ultimately determine its efficacy.
Memory CD4 T-cell functional plasticity allows for continuous adjustment of immune responses to multiple and sequential pathogens and sites of infection. In Fig. 2, we present a schematic diagram of possible mechanisms for plasticity in secondary responses from a population of memory CD4 T cells either during antigenic recall and/or their persistence by homeostatic mechanisms. Upon antigenic recall, there are three possible outcomes for memory CD4 T-cell-mediated responses (Fig. 2, middle): (i) a conserved response in which antigenic recall of the memory CD4 T-cell cohort would elicit effector responses similar to those observed in the primary response; (ii) an altered response in which the memory CD4 T-cell cohort responds to a change in TCR stimulus or cytokine/inflammatory environment, and generates an effector response distinct from that observed in the primary response; (iii) a recall antigen could elicit a response from only a subset of pre-existing memory CD4 T cells with a subset of the functions. In these ways, a population of memory CD4 T cells could be uniquely poised to respond in a similar way to previous activations or to vary their response in altered environmental conditions.
In addition to plasticity in antigenic recall responses, memory CD4 T cells can also undergo changes during their maintenance and homeostasis. Selective survival of a particular memory population as a result of TCR signals encountered during homeostasis can result in a narrowing of the cohort of surviving memory CD4 T cells responding to recall antigen (Fig. 2, bottom). In addition, homeostatic mechanisms with TCR or cytokine signals can result in further differentiation of a memory population leading to new functions elicited upon antigenic recall (Fig. 2, top arrow). With these mechanisms, a given population is not irreversibly fixed to mediate a specific recall function, but remains receptive to cues in the environment and during persistence to fine-tune responses. The effect of these altered secondary responses on long-term changes in specific memory populations is not known, although successive boosting of memory CD8 T-cell responses resulted in altered phenotypes and tissue distribution,74 suggesting that each antigen exposure may imprint the persisting memory population in distinct ways.
Vaccine development and memory modulation
The plasticity in memory CD4 T-cell generation, persistence and recall function has important implications for vaccine strategies for protective immunity, and for memory modulation in disease. Flexibility in memory CD4 T-cell precursors suggests that the potential for generating memory CD4 T cells to a given antigen immunization is quite high. Whether the resultant memory populations mediate protection upon secondary challenge could be directly related to their migration or tissue-resident properties – therefore targeting memory T cells to the appropriate compartments may be more important than initial priming in vaccine design. In addition, long-term maintenance of specific memory CD4 T-cell populations may be favoured by intermittent boosting as memory CD4 T cells are optimally maintained with TCR signals. Hence, plasticity in memory CD4 T-cell generation and homing suggests that targeting strategies following memory generation may be the best way to optimize memory CD4 T-cell responses.
Memory CD4 T cells can also mediate and perpetuate undesirable immune reactions in autoimmunity and also in rejection of transplanted organs.87 Functional plasticity of memory CD4 T-cell populations suggests that they can be modulated by altering the in vivo environment or by immunotherapy.88 Immunomodulators such as CD28 co-stimulation inhibition, previously thought to only affect naive CD4 T cells,89 have recently been shown to affect memory CD4 and CD8 T-cell responses in vivo.90,91 In addition, T-cell depleting strategies may differentially affect naive and memory CD4 T-cell homeostasis.92 Elucidating mechanisms for modulating memory T-cell responses during the recall phase will have broad use in treating immunopathologies perpetuated by memory T cells.
We present here evidence for memory CD4 T-cell plasticity at each stage in memory T-cell development – beginning with memory CD4 T-cell generation from multiple types of activated precursors, during its persistence by homeostatic turnover in response to TCR-driven signals, and also in secondary responses to altered antigenic recall and cytokine environments. These multiple aspects of memory CD4 T-cell flexibility contrast the more defined lineages and functions ascribed to memory CD8 T cells, suggesting a dynamic nature to memory CD4 T-cell responses. It should be possible to exploit this plasticity for therapeutic use, to target the generation of specific types of memory responses long after initial priming, and for regulating memory responses that cause immunopathologies.
D.L.F. is supported by NIHAI083022 and a grant from the Juvenile Diabetes Research Foundation; J.R.L. is supported by a grant from the National Multiple Sclerosis Society.