Dynamic tuning of lymphocytes: physiological basis, mechanisms, and function.

IY33CH22-Paul

ARI

30 January 2015

V I E W

N

I N A

Annu. Rev. Immunol. 2015.33. Downloaded from www.annualreviews.org Access provided by Selcuk University on 02/07/15. For personal use only.

Review in Advance first posted online on February 6, 2015. (Changes may still occur before final publication online and in print.)

C E

S

R

E

9:51

D V A

Dynamic Tuning of Lymphocytes: Physiological Basis, Mechanisms, and Function∗ Zvi Grossman1,2 and William E. Paul1 1

Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892; email: [email protected], [email protected]

2

Department of Physiology and Pharmacology, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

Annu. Rev. Immunol. 2015. 33:22.1–22.37

Keywords

The Annual Review of Immunology is online at immunol.annualreviews.org

adaptation, activation threshold, subthreshold interaction, self/nonself discrimination, reciprocal tuning, anergy, immune regulation, dynamic programing, adaptive differentiation

This article’s doi: 10.1146/annurev-immunol-032712-100027 ∗ This is a work of the U.S. Government and is not subject to copyright protection in the United States.

Abstract Dynamic tuning of cellular responsiveness as a result of repeated stimuli improves the ability of cells to distinguish physiologically meaningful signals from each other and from noise. In particular, lymphocyte activation thresholds are subject to tuning, which contributes to maintaining tolerance to self-antigens and persisting foreign antigens, averting autoimmunity and immune pathogenesis, but allowing responses to strong, structured perturbations that are typically associated with acute infection. Such tuning is also implicated in conferring flexibility to positive selection in the thymus, in controlling the magnitude of the immune response, and in generating memory cells. Additional functional properties are dynamically and differentially tuned in parallel via subthreshold contact interactions between developing or mature lymphocytes and self-antigen-presenting cells. These interactions facilitate and regulate lymphocyte viability, maintain their functional integrity, and influence their responses to foreign antigens and accessory signals, qualitatively and quantitatively. Bidirectional tuning of T cells and antigen-presenting cells leads to the definition of homeostatic set points, thus maximizing clonal diversity.

22.1

Changes may still occur before final publication online and in print

IY33CH22-Paul

ARI

30 January 2015

9:51

INTRODUCTION: BACKGROUND AND OVERVIEW General Background


The concept of dynamic tuning of lymphocytes was introduced into the area of immune regulation based on its biological plausibility and apparent explanatory power (1). It has gradually gained substantial experimental support with validation of different aspects of the theory. Tuning has become incorporated into discussions of results in many immunological settings. Specific recognition of antigen by lymphocytes is not necessarily linked to massive protective responses. Immune recognition may result in many functional consequences, including enhanced or diminished lymphocyte viability, self-renewal, differentiation, or selective induction of regulatory or tissue-surveillance functions (2). We and others added the characteristic of dynamic tuning to the list of potential functional consequences. In particular, lymphocyte activation thresholds are tunable, as are the maintenance of a cell’s viability, its functional integrity, and, presumably, its bias as to fate determination. In other cell systems, neurons in particular, tuning helps cells distinguish meaningful signals or signal patterns from basal ambient signals, or noise, even though the same sets of receptors may be utilized. By continuously adjusting their internal controls to counter signal-mediated perturbations, lymphocytes, like neurons, acquire greater tolerance to fluctuations in ambient-signal strength while maintaining or enhancing their capacity to be activated in response to physiologically meaningful signals that are of sufficient strength and proper orchestration. Beyond the capacity of lymphocytes to tolerate ambient signal fluctuations under relatively stationary microenvironmental conditions, the built-in adjustability of intracellular controls is utilized by the immune system to improve its organization and function.

Historical Overview The dynamic tuning hypothesis was first presented in 1992 in a paper entitled: “Adaptive Cellular Interactions in the Immune System: The Tunable Activation Threshold and the Significance of Subthreshold Responses” (1). Most of those who later referred to this paper and to its follow-up publications (3–5) were interested in the activation-threshold tuning part of the hypothesis, and then mainly in regard to self/nonself discrimination. But we equally emphasized the concept of subthreshold responses as physiologically important events. Indeed, tuning of activation thresholds of lymphocytes are themselves subthreshold responses to antigen. Tunable activation thresholds (TATs) facilitate a range of subthreshold responses while preserving activation to superthreshold stimuli. In considering tuning in the immune system, we derived important insights from other physiological systems, especially the neural and hematopoietic systems, allowing not only explanation of known phenomena but also prediction of subsequent empirical observations. We focused our attention on T lymphocytes and continue to do so in this review, although work related to the tuning of B and natural killer (NK) cell responsiveness is also cited. In addition, reciprocal tuning of antigen-presenting cells (APCs) during their interaction with their T cell counterparts is suggested by evidence and is theoretically appealing. Such reciprocal tuning is also discussed in this review. We regret that space does not permit us to review and discuss the several mathematical models that have been developed to explore the implications of the tuning concept. The dynamic tuning hypothesis challenged prevailing ideas about autoreactivity, anergy, and the extent of genetic preprogramming of lymphocyte responses. At that time, death or complete functional paralysis (anergy) were thought to be the mandatory fates of lymphocytes that happened to recognize self-antigens. Benign autoreactivity existing in healthy individuals and the 22.2

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

presence of autoreactive lymphocytes during immune responses to infectious agents indicated that the self/nonself discrimination paradigm might be much more complex than the simple failure of competent lymphocytes to recognize self-antigens. Moreover, anergy as a self-tolerance mechanism was a problematic concept: One might expect that the number of autoreactive cells, potentially the cause of autoimmunity, should be reduced to a minimum and that anergic cells once generated should not continue to persist and burden the system uselessly. This difficulty disappears if one assumes that the anergic state in vivo is functional, adjustable, and reversible. In this view, autoreactivity is physiologically indispensable and is being controlled rather than avoided. The TAT model described cellular tolerance as a dynamic state (1, 3) in which the threshold for the stereotypic mode of response—activation—had been dynamically tuned and elevated to a degree determined by the cell’s recent subthreshold interactions with ambient antigens. Because every T cell can presumably bind some self-peptide MHC (spMHC) in the periphery with significant affinity (6), as it was required to do to survive positive selection in thymus, all or most T cells would be desensitized to some degree—sufficiently to inhibit activation by these peptides but not more than that; that is, their desensitization would not be high enough to block activation by their cognate antigens in an immunogenic context. Therefore, although capable of recognizing with significant affinity at least the spMHC complexes that had driven positive selection in the thymus, naive or memory lymphocytes would not proliferate and differentiate to mediate or help develop destructive immunity in response to what had become subthreshold interactions. The perspective that autoreactivity is a major factor in the facilitation and regulation of immunity, and that subthreshold interactions are involved in these processes, has gained popularity and experimental support. This can be encapsulated in the words of recent reviewers: “It is now understood that the T cell repertoire is in fact broadly self-reactive, even self-centered. The strength with which a T cell reacts to self-ligands and the environmental context in which this reaction occurs influence almost every aspect of T cell biology, from development to differentiation to effector function” (7, p. 815). There is growing evidence that self-MHC recognition in the periphery is critical for several processes including (a) controlling autoimmunity; (b) maintaining tonic T cell receptor (TCR) signaling and T cell antigen sensitivity, which optimize responses to subsequent challenge with cognate antigen; (c) synergism at the time of cognate antigen recognition leading to increased T cell responses; (d ) maintaining cell viability and functional integrity; (e) homeostatic regulation of T cell numbers, clonal diversity and functional heterogeneity. In addition, the function of naturally occurring regulatory T cells nTregs, and presumably of naturally occurring memory-phenotype T cells (see below), is largely mediated through the recognition of spMHC. Tuning may be essential for the control of potential bystander effects of infection, although in some instances pathogen sensing/adjuvant action, through its effects in enhancing antigen-presenting/stimulatory properties of dendritic cells (DCs), may elevate the autoreactivity of lymphocytes from a subthreshold to a superthreshold (activating) response. A unifying outlook, incorporating the integrated function of pathogen sensing, Tregs, and tuning, has been proposed (8, 9).

THE DYNAMIC TUNING HYPOTHESIS Definition and Physiological Basis Our definition of tuning as it applies to cells of the immune system is a reversible change of cellular phenotype that reflects strength, frequency, and structure of external signals measured over time. A cell requires repeated stimulation to implement the phenotypic changes resulting from tuning. These changes are short-lived in the absence of further stimulation and are not heritable. The requirement for continued stimulation and the lack of heritability distinguishes tuning from www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.3

ARI

30 January 2015

9:51

differentiation. Upon differentiation, heritable modifications in gene expression are induced that do not require a continued presence of the extracellular signals that facilitated their induction and that carry over to daughter cells when the cell divides. Unlike differentiation, which involves a qualitative change in the pattern of gene expression and is hard to reverse, tuning is a reversible quantitative process. The traces of previous signaling events are gradually erased, actively and/or passively, in the absence of continued stimulation by the same signals and are dynamically modified if stimulation continues but varies. Therefore, tuning mirrors the cell’s stimulation experience, with more weight given to more recent signaling. Molecular manifestations of tuning may generally include quantitative changes in levels of gene expression; in the expression of viability-promoting factors, receptors, transcription factors, and other regulatory molecules; in the clustering of such molecules on the cell membrane or their polarization across the cell; and in the activation status of enzymes that facilitate signal transduction (e.g., phosphorylation). The functional consequence of a new tuned state is a modified cellular responsiveness to external signals, potentially entailing changes in the cell’s viability, in the relative probabilities of alternative differentiation fates, in the relative probability of proliferation versus differentiation, or in activation thresholds. Nevertheless, under certain circumstances the dynamically induced functional-reprogramming effects of tuning could gradually develop into adaptive differentiation (1, 3), presumably when tuning has been operating at the level of transcription factor expression. Tuning extends genetically encoded signal discrimination and contributes to the system’s functional resilience, as it allows cells to modify their interpretation of external signals continuously based on their recent experience. In particular, lymphocytes classify signal patterns in part by the strength of the perturbations of the cellular state that they induce. Adjustment of the thresholds that are used in this classification by the tuning process allows the cells to optimally distinguish immunologically meaningful stimuli from each other and from background noise. Meaningful signals evoke a response. The internal molecular organization of nonterminally differentiated, mitotic lymphocytes is characterized by limited stability in the face of perturbations and by multiple means of regaining stability. Depending on the existing state and on the perturbation, the cell may ignore the perturbation, undergo self-renewal division (in which both daughter cells maintain the existing pattern of gene expression), make reversible adjustments to the amount and/or activation state of expressed proteins (tuning), or become transiently unstable or activated. When activated, a lymphocyte may proliferate rapidly for several days, with some progeny subsequently dying, others undergoing further differentiation, and still others regaining the original phenotype but not necessarily the same tuned state. The patterns of gene expression of the activated cells and their progeny undergo quantitative and qualitative changes that are largely predictable and that have well-defined functional consequences. Such burst-like proliferation-anddifferentiation events (activation bursts) are characteristic of acute immune responses to infection and of sufficiently strong antigenic stimulation by nonpathogens. Tuning that increases activation thresholds (reducing overall responsiveness) enhances selectivity for activation, while parallel modifications of other cellular characteristics may enhance and/or guide responses once activation is allowed to occur. Autoreactivity is enforced during positive selection of T cells in thymus and controlled by dynamic tuning and Tregs in the periphery. We and others proposed that lymphocytes are selected in this way to be moderately and variably autoreactive so that they can use recognition of selfantigens to sustain and tune their viability and function (1, 4, 5, 10–15); hence, there is a need for concomitant tuning of the activation thresholds. As is discussed below, the dependence of viability and functionality on such interactions is required as a means of ensuring clonal diversity and functional heterogeneity, two essential features of immunity.


IY33CH22-Paul

22.4

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

During subthreshold interactions, patterns of signals are recognized and evoke selective responses, other than the typical explosive response that activated lymphocytes mount in response to acute infection. In inflammatory settings, lymphocytes rendered self-tolerant by tuning may cooperate with tuned or partially tolerized (exhausted) foreign antigen–specific cells in suppressing or confining persistent pathogens or transformed cells (and be involved in other useful bodily functions) through mechanisms that, unlike conventional responses, minimize immunopathogenic consequences (4, 5). In addition, such lymphocytes might be stimulated by their cognate selfinteractions to participate, locally and transiently, in immune responses against pathogens, which would provide meaning to bystander activation (16). Especially important in this function are probably spontaneously arising memory-phenotype T cells, which tend to show higher affinity for spMHC peptides, substantial diversity, and a potential to secrete cytokines upon activation (17). Finally, tuning of individual cells depends on, and in turn affects, communication among cells that are capable of interacting with each other. Regulatory feedback loops among cells of different phenotypes and maturation states tend to stabilize tissue-cell organization. But in the face of recurring perturbations (as in chronic infection), tuning and eventually adaptive differentiation may have far-reaching effects on the parameters that define these loops and on tissue-cell organization.

The Concept of Tunable Activation Threshold TAT is a specific manifestation of tuning, one among several. A basic assumption of TAT is that receptor-mediated stimulation of a lymphocyte that fails to result in activation nonetheless produces an alteration in the activation threshold. The change is reversible but takes some time to be reversed. Tuning is thought to typically occur as lymphocytes recurrently interact with ambient spMHC under normal physiological conditions or when they acquire and then actively maintain tolerance to a long-lived antigen, such as those displayed by a persistent pathogen or tumor cells. The length of the tuning memory (the time that an alteration in the activation threshold lasts in the absence of reinforcing stimuli) would depend on which signal-transduction modules had been affected (see below) and to what extent. In the case of chronic infection, some alterations associated with tuning may later become epigenetically imprinted, and thereby less reversible, or lead to secondary alterations in a process that may be called adaptive differentiation (1, 5). This interpretation has been recently proposed in connection to the phenomenon of T cell exhaustion (18, 19).

THE EXCITATION-DE-EXCITATION PARADIGM Multidimensionality of Activation and Tuning The molecular mechanisms underlying tuning are intimately related to those regulating activation; the latter have been subject to extensive investigation for many years, and great progress has been made, although a full understanding lies in the future. Naive or memory lymphocyte activation is restricted by the requirement that the cellular activation thresholds must be exceeded. Activation is multiphasic (possessing several checkpoints along a given pathway of activation) and multidimensional (possessing several parallel interacting pathways, either synergistic or competing). A schematic, rudimentary discussion of such organization of checkpoints and pathways and of the implications can be found in Reference 5. The multiphasic nature suggests the possibility of incomplete activation, failing to induce significant effector function or proliferation but leading to modifications of certain signal-transduction module or modules (groups of interacting signals) that distribute signals downstream and sideways (20). Such modifications would include desensitization of these modules, implying that tuning of www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.5

IY33CH22-Paul

ARI

30 January 2015

9:51


activation thresholds may involve the TCR as well as additional intracellular modules, depending on the nature and depth of the perturbation that led to the aborted/incomplete activation. This probably accounts for the divergent molecular correlates of naive T cell tuning in response to relatively weak signaling by nonagonist ligands versus those induced by persisting agonists, and of memory cell tuning to persistent antigen in the absence of inflammation versus tolerance to chronic infection (discussed in some detail in the following sections). Multidimensionality implies that when the major pathway of activation is inhibited, alternative pathways may be affected but need not be obstructed. This is how viability and functionality continue to be reinforced and tuned by subthreshold antigenic stimuli and accessory signals (11). Indeed, recent studies have supported the view that T cell signal transduction is a complex branching network, rather than a top-down signaling pathway, and that activation, proliferation, differentiation, and the termination of cellular responses are controlled through interactions between different branches of the signal transduction network (reviewed in 21).

Steady States and Perturbations To explain how lymphocytes discriminate among identical sets of external signals according to differences in their integrated strength, as it is perceived by the cell, a simple generic paradigm was postulated (1, 5). At various steps of the signal transduction pathway, local intracellular decision events depend on the balance between pertinent excitation- and de-excitation-inducing factors that are recruited by external signals (1, 5). Excitation consists of biochemical changes that converge toward gene activation. De-excitation consists of changes that reverse or negate the effects of excitation. Excitation factors include, for example, enzymes that upon their activation (e.g., phosphorylation) can in turn catalyze activation of other signal transduction molecules. De-excitation factors may consist of enzymes that upon activation inhibit or control the activity of excitation factors, either directly or by deactivation (e.g., dephosphorylation) of substrates upon which excitation factors act. With homeostatic or adaptive behavior in mind, de-excitation factors are best conceived as the products of negative-feedback control mechanisms (20). Different excitation and de-excitation factors would be associated with different steps of signal transduction. At the cell surface, for instance, the receptors that bind antigen and accessory signals and enzymes that phosphorylate critical domains of these receptors may be considered excitation factors, whereas molecules that, when activated, reduce the number of these receptors or desensitize them, or interfere with their activity and/or functional organization, would be considered de-excitation factors. Tyrosine kinases and phosphatases were originally proposed in the role of the opposing factors at unspecified phases of the major activation pathway (1). We assumed that under relatively constant, normal conditions T cells maintain a quiescent (or resting) quasi steady state because, in signaling modules that are critical for the control of cell activation (such as the TCR complex), the balance between excitation and de-excitation factors is skewed in favor of the latter. This balance, and the resting state, is dynamically imposed by the interactions between TCRs and spMHC molecules. A change in the presence of external signaling molecules, such as cognate antigen or selecting self-antigen (see below), potentially perturbs the balance among excitation and de-excitation factors in such modules: for example, consider an increase in the level of external signals. As a consequence, the presence and activity of both types of factors will increase, either as independent processes that are elicited in parallel or because de-excitation activity is elicited in response to that of the excitation factors, mediating a negativefeedback control. Thus, signal-induced perturbations translate into a kinetic competition between excitation and de-excitation. 22.6

Grossman

·

Paul


IY33CH22-Paul

ARI

30 January 2015

9:51


The Molecular Basis of Activation Thresholds To explain how quantitative differences between such perturbations are further translated with a remarkable precision into qualitatively distinct responses at the signaling module level, and subsequently into different functional responses at the cellular level, we introduced the excitationde-excitation kinetic competition model (1, 5). We reasoned that the intracellular concentrations of excitation and de-excitation factors trace the changes in external stimulation with inherently different kinetics. The rate of increase in excitation factors is variable and can be either rapid or slow, depending on the quality and intensity of stimulation and on the initial level of de-excitation factors, the latter being determined intrinsically and by tuning. Excitation factors rapidly decline when stimulation is removed. On the other hand, the rates of increase in de-excitation factors can vary within a more limited range. Their decline postexcitation is constitutively slow in comparison to excitation decline. When the level of stimulation rises abruptly in the face of a low de-excitation factor baseline, the latter factors may not be able to keep up with the rapid rise in excitation, and the balance can be temporarily reversed in favor of the excitation factors. We assumed that once the excitation-de-excitation difference or ratio [or some other measure(s) of the relation between these quantities] exceeds a certain critical value, a clearly distinguishable class of intracellular responses will occur that in turn lead to a biological response at the cellular level. Excitation rates would thereby be sharply discriminated. Note that the critical value is an intrinsic property of the cellular state, whereas the excitation rate reflects stimulation strength, and the module’s activation threshold (or rather excitation threshold, as activation threshold more precisely refers to the entire cell) is subject to tuning (5). We offered alternative explanations for the existence of qualitatively different outcomes that depend on whether the level of excitation in relation to de-excitation exceeds a certain threshold or not. Opposing actions of enzymes on the same substrate, such as kinases and phosphatases, could in principle result in ultrasensitivity to the ratio of enzyme concentrations under certain nonstringent conditions (22, 23), such that a small change in that ratio around a critical value results in a transition of the substrate from a state in which it is phosphorylated most of the time to the opposite state, or vice versa. But the most general explanation (5, 24), which has gained experimental support, was that the process of excitation involves a self-enhancing component. This could result from a positive feedback loop or some other form of cooperativity between excitation factors and/or between them and their precursors, amplifying the direct dependency of the excitation rate on stimulation strength. By definition, such a self-enhancing effect has a faster-than-linear dynamics; it is small as long as the excitation factor(s) has not yet reached a sufficiently high concentration. If de-excitation factors control the growth in excitation early enough, a quiescent state will be restored. In contrast, if excitation, driven by strong binding and keeping ahead of de-excitation, is allowed to grow beyond a certain level, the self-enhancing effect takes over, switching the signaling module into an activation state of self-sustained excitation. In turn, this is controlled and eventually terminated by other mechanisms, possibly involving different negative-feedback products (21). In this way, a self-enhancing effect coupled to negative controls facilitates bi-stability—the coexistence of a quiescent baseline state and an activation state, each of which can be reached under appropriate conditions. In mathematical descriptions of such systems, there is a well-defined separation line, or surface, in parameter space between the domains of influence of the different quasi stable states; a perturbation may shift the system from one domain to the other. An antagonist binds to the TCR more weakly than an agonist but can nonetheless block activation of a T cell by an agonist presented to the cell at the same time. Stefanova and Germain, who studied the dynamics of biochemical changes that take place upon TCR engagement by agonists

www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.7

ARI

30 January 2015

9:51

in contrast to partial agonists or nonagonists, revealed the operation of a positive feedback loop in this context and demonstrated its implications when coupled to a negative feedback mechanism (25; see also 26). They demonstrated the importance of a rapid rise in induced positive signal levels in order for excitation signals to outcompete the negative signals, as had been predicted (1, 5). Their results fully complied with a generic description of the emergence of bi-stability (5). Briefly (see References 26–28 for more detail and citations to additional papers), appropriate T cell binding to an agonist presented by an APC resulted in sequential recruitment and phosphorylation of the src-family kinase Lck, the TCR-associated γ, δ, and ζ chains, the kinase ZAP-70, the coreceptor-associated linker for activation of T cells (LAT), and the mitogen-activated protein (MAP) kinase ERK that translate ligand engagement into gene regulatory events controlling T cell differentiation and proliferation. In contrast, when an antagonist was used, this chain of events was interrupted early on due to negative feedback regulation of TCR signaling. Indeed, when Lck is activated, it also partially phosphorylates the protein tyrosine phosphatase SHP-1, which is recruited to Lck associated with the TCR complex and has the capacity to dephosphorylate various signaling molecules including Lck itself, establishing a negative feedback loop. When the phosphorylated phosphatase binds to Lck in the TCR complex and is further phosphorylated, a loss of Lck activity, ζ chain, and ZAP-70 phosphorylation follows, and ERK is not significantly induced. However, strong stimulation with agonists rapidly induces phospho-ERK, raising its level to a point where it starts promoting its own production. The products phospho-ERK and SHP-1 may be considered excitation and de-excitation factors, respectively. The self-enhancing component of excitation, a positive feedback loop, is somewhat complex. It comes about because phospho-ERK phosphorylates Lck, preventing tyrosine phosphorylated SHP-1 from docking and thereby protecting Lck against deactivation by SHP-1, maintaining the activation of ERK and possibly allowing high-gain activation of additional molecules of ERK. This mechanism can thus bestow a switching capacity to the signaling module/subsystem. The competition between excitation and de-excitation or, equivalently, competition for dominance between the positive and the negative regulatory loops, is won within a minute. When excitation had the upper hand, the TCR remained activated for additional 30–45 min, during which binding of TCR-associated Lck by phospho-SHP-1 was not seen. Eventually, such binding did occur, in concert with an attenuation of ζ chain phosphorylation and ZAP-70 activity, terminating the module activation episode. One may conclude that agonists are superior to partial agonists and antagonists because the stronger TCR-pMHC binding they achieve implies a faster local buildup of activated ERK, whereas the rates of recruitment and activation of SHP-1 have a more limited range. In particular, an affinity threshold for activation of the TCR-signaling module would in principle be defined in this model for any fixed set of other relevant parameters as the minimal affinity required for activated ERK to reach a sufficient local concentration, enabling the self-enhancing mechanism to overcome the negative, restoring force of SHP-1, at which point the positive loop ignites. Overcoming the TCR-activation threshold (that is, the local excitation threshold) does not necessarily lead to cell activation, as additional excitation thresholds may arise downstream, where the concentration and activity of excitation molecules reflect the number of activated TCRs and their rate of activation, as well as the amount and quality of signals transduced via other channels (29). In the following, we generally use the terms excitation threshold and activation threshold interchangeably, for simplicity, unless otherwise stated. Interestingly, it appears that the amount of activated ERK following T cell stimulation does not strongly correlate with the rate of TCR activation as it did not change when ligand concentration was varied. Instead, the ERK response was digital: 70–80% of maximum within a few minutes following stimulation with different agonist concentrations, and no significant response to antagonists (30). This is consistent with a high-gain amplification achieving saturation already at the


IY33CH22-Paul

22.8

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

lower ligand concentrations. Increase in ligand concentration resulted in an increasing number of activated cells with a maximal ERK response, likely reflecting the preexisting variation among individual cells, even within an apparently homogeneous antigen-specific population, due both to intrinsic noise and to differences induced stochastically by extrinsic factors, including differences in tuning. A sharp (digital) discrimination of antigenic stimuli at the single-cell level is necessarily translated into a more graded (analog) discrimination at the level of the cell population; the shape of the dose-response curve is typically sigmoidal. To explain the relation between TCR-ligand binding properties and the ligand capacity to activate, we proposed that the local buildup of excitation factors, such as phospho-ERK, depended on the integration of a large number of fast, reversible molecular interactions involving local populations of receptors and the signaling molecules surrounding or attached to them. This proposition requires that the interactions initiated by TCR ligation consist of rapid cycles of phosphorylation and dephosphorylation and of receptor binding and unbinding. It also requires that recruited molecules, particularly phospho-ERK and active SHP-1, should be capable of progressively increasing their effective local concentrations in the TCR-proximal membrane compartment during the process in a way that would affect the frequency of interactions with their target molecules such as Lck. The interactions are stochastic, but their population dynamics lead to robust outcomes. We argued that, when a small population of TCRs collectively interacts, locally, with a small population of ligands undergoing rapid engagement and disengagement events, reengagement of the same TCR by the same ligand following disengagement is not necessary for a TCR to become progressively excited and then activated. What is important in this case is the average rates of engagement and disengagement of participating TCRs, which determine the cumulative binding time of the ligand or ligands to several TCRs, upon which the buildup rate of excitation factors depends. Because SHP-1 and ERK may quickly target receptor complexes in the same domain (31), fluctuations in the signaling by individual receptors would tend to average out. Thus, in our model small populations of microclustered TCRs, rather than an individual TCR, collectively act as the state-switching unit. Indeed, such microclusters have been directly demonstrated (32–34). The sharpness of pMHC ligand discrimination, as we have discussed, is inherent in the nonlinear nature of the interactions, i.e., in the coupling of negative controls to a self-enhancing excitation mechanism. An increase in the number of ligands on the APC does not significantly compensate for a weaker binding; a partial agonist remains a partial agonist even at high concentration. We argued that this is due to the localized nature of the interactions, imposing physical-geometric constraints on the access of APC membrane–associated ligands to the relevant TCR subpopulation; increase in ligand number would result mainly in a larger number of units rather than in a better signaling quality of individual units. In other words, individual TCR complexes gather signals from locally interacting signaling molecules and therefore measure mainly the quality of the ligand rather than its multiplicity. The latter is important, though, for the integration of TCR-generated signals by downstream modules, into which accessory (non-TCR) signals are also channeled. This downstream integration of excitation signals and their negative controls may also be capable of showing bistability and switching (29), although switching can be delayed by hours, during which a competition with induced and preexisting de-excitation factors takes place. An organized network of several signaling modules (21) with threshold-controlled excitation thresholds (35) is envisioned to define, collectively, the T cell activation threshold (or a set of thresholds for the activation of different effector functions). The apparent dominant role of the TCR-ligand dissociation (off ) rate in determining the effectiveness of signaling, with its apparent insensitivity to the association rate, led Germain and Stefanova at the time to adopt McKeithan’s kinetic proofreading hypothesis in regard to the www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.9

ARI

30 January 2015

9:51

efficacy of excitation rather than a molecule-population dynamics approach (reviewed in 26). According to that hypothesis, nonagonist pMHC ligands fail to activate ERK simply because the ligand dissociates from the receptor too quickly. The assumption was that stable, uninterrupted recruitment of the Lck to the occupied TCR for a sufficient period of time was required for activated Lck to become protected against SHP-1-mediated dephosphorylation. This complicated the model, now including two different kinetic barriers to TCR activation: one a requirement for a minimal binding time and the other a requirement for an early onset of a positive feedback process. But an inherently greater dependence of the outcome on the off rate is no longer believed to be a necessary feature of ligand discrimination models. Recent studies have shown that the two-dimensional interface between the T cell and the APC dictates a highly complex dependency on both dissociation and association rates in determining ligand signaling potency for activating the TCR-complex module. Both two-dimensional association and dissociation rates are much faster for agonists than what is measured in three-dimensional assays, and agonists tend to be characterized more by their high association rates than by the rates of dissociation. A long-lasting bond is not essential because “high bond formation frequency also accumulates a large fraction of engagement time” (36, p. 935). The characterization of microclusters (33, 37, 38) and of new forms of TCR cooperativity (reviewed in 39) provided a natural explanation for the existence of small high-density TCR-ligand interaction units and hence for the superior role of binding affinity over antigen dose in facilitating suprathreshold TCR activation (36). Together, these developments support the excitation-de-excitation model in its general, robust form as a basis for the interpretation of postbinding molecular kinetics first reported by Germain and colleagues (25). Not surprisingly, the actual interplay of positive and negative factors observed experimentally is more complex than proposed in the initial models, but the concept that such an interplay plays a crucial role in signal discrimination has been established (20, 21). By extrapolation, the model provides a working hypothesis—a paradigm—that explains the existence of potential multiparametric thresholds for the activation of additional signaling modules and of the T cell itself. With almost no further assumptions, the model also suggests a simple mechanistic interpretation of tuning as it applies to excitation and activation thresholds.


IY33CH22-Paul

The Molecular Basis for Tunable Activation Threshold Antagonist ligands can protect a T cell from activation by cross-reactive agonists due to the presence of de-excitation factors, induced during ongoing interactions with the antagonist pMHC ligand. Such protection was more effective for antagonists with relatively strong subthreshold binding, that is, ligands whose binding to the TCR more closely approximated that of agonists (26). This somewhat unexpected result was explained by the stronger generation of SHP-1 by the more avidly binding ligands in the absence of an overriding ERK response. Indeed, stronger inputs generally trigger stronger counter-responses in feedback-control situations. The antagonist pMHC did not necessarily bind the same TCR that could engage the agonist pMHC presented by the same APC. Rather, antagonism could operate between different receptors (26), and this was best explained by the observation of local sharing and spreading of de-excitation factors. According to the TAT hypothesis, qualitatively similar interpretations apply to multiple T cell–APC encounters: The baseline levels of excitation and de-excitation factors reflect several T cell–APC subthreshold encounters and are important determinants of the kinetic competition occurring during subsequent encounters. Tuning, as it affects a given antigen-binding event, is embodied in the molecular residues of past events. In particular, any self or foreign peptide that can bind to the TCRs displayed by a given T cell but that fails to activate them plays the role of an antagonist peptide in desensitizing these receptors to activation by the same or different peptides. 22.10

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

The proposed basis for such desensitization of activation thresholds, as for the very existence of these thresholds, is the different kinetics with which the intracellular excitation and de-excitation mechanisms trace fluctuations and other changes in the level of external stimulation (1, 5). The postulated slow postexcitation decline of de-excitation factors in comparison to that of excitation factors would allow de-excitation factors to predominate during quasi-stationary stimulation, imposing a resting/quiescent state. Quasi-stationary stimulation means that stimuli are frequent and relatively uniform in strength. Under these conditions, the level of de-excitation (de-excitation index) essentially traces, with a certain safety margin, the moving average of the levels of excitation induced by ambient antigens during transient excitation events (Figure 1). The safety margin is defined as the critical perturbation of the baseline concentrations/activity of signaling molecules required for activation; because of tuning, the critical perturbation is relatively invariant to the baseline stimulation itself, and this is consistent with the notion that cells of the immune system respond to abrupt changes in stimulation levels while adapting to relatively constant stimulation. The actual value of this safety margin, the critical perturbation, can be thought to represent the usual trade-off between selectivity and sensitivity. Although the purpose of tuning is to reject unintended perturbations, it should not significantly compromise the ability of cells to become activated in response to an abrupt increase in the level of excitation in the physiological range, on encountering their cognate foreign antigen or under other physiological circumstances requiring activation. Which perturbations need to be rejected? Significant nominal autoreactivity of many T cells for tissue-specific antigens, and even for peripheral antigens that are present in thymus during positive and negative selection, may require restraint to avoid undue activation (6). The perturbations requiring control arise from the discrete nature of binding events and from unavoidable fluctuations in external signal strength as it is perceived by signal transduction mechanisms. Fluctuations in self-peptide concentrations alone are not sufficient to induce TCR activation, which is sensitive to the quality of binding rather than to ligand concentration. However, quantitative variation of internal signaling molecules may introduce large differences in autoreactivity among cells of the same clone; tuning is required also to counter such variation. Indeed, single-cell measurements revealed that the level of each signaling protein may differ by a factor of three among cells of a clonal population (40). Model calculations showed that the thresholds of a multicomponent signaling system may possess even stronger variability because several signaling protein concentrations jointly determined the signaling threshold. Simulations further demonstrated that tuning of basal de-excitation levels (basal feedback inhibitors) can correct for such basal state variability. In the context of immunity, this is equivalent to saying that tuning provides feedback control in which responsiveness of cells manifesting higher autoreactivity is tuned to a larger degree; sensitive cells receive stronger basal signals and therefore acquire higher basal de-excitation levels. The simulations elegantly validated the two kinetic requirements for signal discrimination that the TAT hypothesis had postulated (5): (a) relative stability of the (transient) stimulation-related increase in the expression and/or distribution of de-excitation molecules, to facilitate memory of subthreshold stimulation events; and (b) a slow and /or delayed induction of additional de-excitation molecules in response to acute stimulation, providing a time window early after stimulation when the steepness of the increase in excitation is only moderately affected by negative feedback, allowing stimulation by agonists to exceed a critical excitation level and to benefit from cooperativity inherent in the signal transduction process, triggering a switch-like response. Phospho-SHP-1 has been implicated in tuning from the outset (5, 41). SHP-1 fulfills requirement (b), as discussed above. Recent studies have shown that circulating autoreactive CD8+ T cells capable of recognizing self/tumor-associated antigens become resistant to activation by these www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.11

IY33CH22-Paul

ARI

30 January 2015

9:51

a

Weak subthreshold signals Activation threshold Signal

Critical perturbation


Signal levels (arbitrary units)

De-excitation index

b

Strong subthreshold signals

c

Strong, infrequent subthreshold signals

d

Varying signals

Activation

Time (arbitrary units) Figure 1 Schematic representation of TAT kinetics. Blue lines represent TCR-mediated signal strength. The lower gray solid line represents the kinetics of the de-excitation index, a measure of the level of persisting de-excitation molecules. De-excitation index (x) reflects a feedback response to subthreshold stimulation during TCR-ligand binding events followed by a slow decay of de-excitation molecules between such events. The feedback leads to a steeper transient increase in response to stronger signals but with limited adjustability. If x(0) denotes the de-excitation index at the beginning of a decay phase and x(t) is its magnitude after a stimulation-free time interval, t, then the linear approximation of the decay rate, (x(t) − (x(0))/t, represents the average slope of a more realistic exponential decay, −x(0)(1 − e−kt )/t, because x(t) = x(0)e−kt , with a constant rate-coefficient k. The upper solid black line represents the parallel kinetics of the activation threshold, which is defined by adding to the de-excitation index an invariant critical perturbation (see text). (a) Adaptation to recurring weak, equal stimuli that are equally spaced. The activation threshold fluctuates around a constant average. (b) Adaptation to recurring stronger stimuli, still subthreshold, spaced as in panel a. De-excitation index and activation threshold are adapted to be higher, such that a steeper negative feedback response is counterbalanced by a faster decline. (c) Adaptation to recurring infrequent stronger stimuli. Signal strength is as in panel b, but with longer time intervals between encounters. The de-excitation index is adapted to assume a lower level so that the negative feedback response (same as in panel b) is counterbalanced by a less steep but longer decline. (d ) Varying signal strengths and interstimulation times. The activation threshold adapts to the changing stimuli, but an increase in signal strength that is too abrupt for the cells to adjust exceeds the activation threshold.

22.12

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

antigens in the tumor environment and that this functional inhibition correlates both with the TCR-spMHC binding affinity and with the expression of SHP-1 (42). Thus, SHP-1 stimulationinduced expression appears also to fulfill the stability requirement (a, above), suggesting that TCR-mediated SHP-1 signaling may be generally involved in dynamic counterregulation of autoreactivity, as proposed. Other direct evidence supporting this proposition is still lacking. The potential analogy for the role of SHP-1 in dynamic tuning with its demonstrated role in the response to antagonists was criticized, based in part on the observation that naive CD4+ T cells freshly isolated from lymph nodes of TCR-transgenic mice showed relatively little SHP-1 accumulation in the region of clustered TCRs compared to cells stimulated for 1 min with APCs prepulsed with antagonist peptide, which showed extensive membrane association of SHP-1 in the region of receptor polarization (43). However, although the natural level of exposure of cells to their cognate self-antigen in vivo is expected to be low, a more recent communication reported that downregulation of SHP-1 could lead to hypersensitivity to a self-like ligand (28). The investigators concluded that, consistent with the TAT model, exposure to spMHCs may induce a small amount of SHP-1 recruitment to the TCR pool, just sufficient to prevent effective responses by poor TCR binders (28). Substantial evidence has accumulated for the role of the cell surface molecule CD5 in tuning; we review this in the following section.

TUNING INHIBITS ACTIVATION OF THYMOCYTES AND NAIVE T CELLS Early Observations Several papers provided data that can now be interpreted in terms of the tuning hypothesis. Jameson et al. (44) presented data suggesting that flexibility in CD8 expression levels is one way for thymocytes to fine-tune the efficacy of their interaction with pMHC ligands present in the thymus and to achieve positive selection while maintaining self-tolerance. Goodnow and associates (45) observed that B and T cells acquired changes in their activation thresholds (set points) as a result of previous antigen encounters. Although the authors used the term tuning for this phenomenon and recognized its considerable flexibility, they did not show that such tuning was a dynamic, reversible process and in fact did not distinguish it from classical anergy. Ohashi and colleagues (46–49) demonstrated tuning of thymocytes bearing a defined transgenic TCR. Thymocytes selected upon a certain peptide lost their ability to proliferate to that peptide as mature T cells while still retaining the ability to proliferate to higher-affinity peptides.

CD5 Mediates Dynamic Tuning Expression levels of CD5 on T cells and thymocytes reflect the strength of antigen receptor signaling, which reciprocally tunes the threshold of the response (50–52, 53, 54). CD5 is a transmembrane protein that is expressed on the surface of T cells and a subset of B cells. The regulatory effect of CD5 involves an immune-receptor tyrosine-based inhibitory motif in the cytoplasmic domain of CD5 and is thought to be mediated by the association of CD5 with SHP-1 (55, 56), implicating both these molecules in the same negative feedback mechanism (41). The extracellular domain of CD5 is not involved in CD5-mediated downregulation of TCR signaling during intrathymic development (57). The molecular mechanism of CD5-dependent downregulation of TCR signaling remains unresolved, and the association of SHP-1 to membrane-proximal tyrosine of CD5 is controversial, but independent studies implicate activation of other negative regulators such as c-Cbl (58) and Cbl-b (59). www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.13

ARI

30 January 2015

9:51

To the extent that CD5 exerts its inhibitory effect on TCR-proximal signaling through its association with SHP-1, it may be considered to be the rate-limiting de-excitation molecule in the negative-feedback pathway, given that expressed CD5 molecules do not quickly dissipate. In addition, because updating of CD5 expression depends on transcription, it would be slow enough to allow for the lag-time window required for the operation of a switch-like decision-making mechanism. CD5 may then be the real facilitator of TCR tuning. Alternatively, SHP-1 by itself may facilitate rapid and relatively short-lived tuning events in response to frequent weak-binding encounters, whereas CD5 expression, via its physical and functional coupling to SHP-1, could be tracing the input of less frequent, and presumably stronger, self-recognition events. Additional receptors and regulatory molecules appear to be involved in the tuning of downstream signaling modules. Azzam et al. (51) demonstrated that CD5 was upregulated at crucial points during thymocyte development by pre-TCR and TCR engagement and that the level of CD5 surface expression was directly related to pre-TCR and TCR signaling intensity. In particular, CD5 surface levels varied considerably among mature single-positive (SP) thymocytes and among T cells that expressed distinct TCRs. The level of CD5 expression paralleled the avidity of the positively selecting TCR–MHC ligand interaction. This and a subsequent study by the same group (52), in which the effect of CD5 overexpression or CD5 deletion on thymocyte selection was assessed in different TCR-transgenic backgrounds, supported the view that activation-threshold tuning endows thymic selection with a degree of flexibility (60). Accordingly, thymocytes are sensitive to a sharp increase in TCR signaling and can undergo deletion at any time. Downtuning of the functional responsiveness of thymocytes to stimulation by autoreactive self-antigens acts to counter the progressive increase in intrinsic responsiveness that is associated with T cell development, providing a dynamic mechanism for protecting a larger portion of thymocytes—all but those possessing the highest affinities—from deletion. Deletion occurs only when tuning cannot keep up with the rate of increase in the intensity of signaling, which is directly related to affinity/avidity (60) (Figure 2). In this way tuning was proposed to facilitate positive selection over a protracted period of time and for a broader range of TCR-ligand affinities than would be possible if the affinity of the TCR for selecting ligand alone dictated the outcome. Thus, the potential for TCR signal modulation by CD5 may be particularly useful for generating the maximum possible diversity in the mature T cell repertoire (51, 60). Rudensky and colleagues (61) analyzed the phenotype and function of TCR transgenic thymocytes developing in the presence of altered repertoires of MHC class II (MHC-II)–bound peptides. They showed that, in addition to the selecting ligand(s), a broad spectrum of self-peptides interact with the TCR of immature thymocytes at the double-positive (DP) stage of T cell development and cause alterations in thymocyte surface phenotype and TCR sensitivity. TCR engagement of these intrathymic pMHC ligands resulted in a signal that induced cells to upregulate not just CD5 but also CD69, CD2, and perhaps other accessory molecules. Greater upregulation of these molecules was seen on more diverse pMHC backgrounds, as one would expect if peptides with a broader range of avidities were available to multiply engage each TCR, influencing the extent of alteration in surface-marker expression levels. It was proposed that “TCR sensitivity can be adjusted via a rheostat-like mechanism whereby self-peptides variably tune the levels of TCR signaling” (61, p. 1185). In terms of the TAT model, the rheostat dynamically integrates molecular responses to multiple TCR-mediated subthreshold stimuli (1, 60). The authors observed that T cell accessory molecules such as CD5, CD2, and CD28 were altered in splenic T cells compared with thymic CD4 SP thymocytes, suggesting that the adjustment of TCR sensitivity continued in mature T cells in the periphery. Furthermore, adoptive transfer of TCR -transgenic CD4+ T cells into hosts expressing altered sets of self-peptides resulted in changes in CD5 expression.


IY33CH22-Paul

22.14

Grossman

·

Paul


IY33CH22-Paul

ARI

30 January 2015

9:51

a Positively selected thymocyte Activation threshold “Smoothened” activation threshold

Signal levels (arbitrary units)


De-excitation index Signal

b Negatively selected thymocyte Deletion

Time (arbitrary units) Figure 2 Schematic representation of positive and negative selection in the thymus. (a) A thymocyte adapts its activation threshold to developmental increase in perceived signals mediated by a cognate self-peptide-MHC ligand. (b) A thymocyte binding a cognate self-ligand with a higher affinity fails to adapt to the faster developmental increase in signal strength and is negatively selected. This is a consequence of the limited adjustability of the negative feedback response to a steeper increase in excitation.

Smith et al. (53) showed that pMHC contact modulated the expression of CD5 by naive CD4+ T cells in a process that required continued expression of p56 lck. Reduced CD5 levels in T cells deprived of pMHC contact for three days were associated with elevated Ca2+ responses to subsequent TCR engagement. The authors concluded that naive CD4+ T cells undergo a process akin to sensory adaptation. The results suggested that pMHC contact in peripheral lymphoid organs fine-tunes T cell responsiveness to restrict the consequences of autoreactivity. Stockinger and colleagues (54) described a double-transgenic model in which monoclonal CD8+ T cells were chronically exposed to self-antigen (nucleoprotein) in the periphery but not during thymic development. Such peripheral exposure rendered these cells unable to proliferate or produce cytokines in response to the antigenic stimulation in vitro, although they still retained some killer function in vivo. Importantly, the anergic state was reversible upon transfer into www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.15

ARI

30 January 2015

9:51

antigen-free recipients, resulting in full recovery of in vitro responsiveness to nucleoprotein. Downtuned CD8+ T cells expressed higher levels of CD5, whereas after transfer and residence in antigen-free hosts, CD5 levels returned to normal. This suggested that upregulation of negative T cell regulators in peripheral T cells exposed to chronic stimulation by antigen prevented full functionality, and thus overt autoimmunity was avoided. Analyzing the basal phosphorylation of the TCRζ chain and of the protein tyrosine kinase ZAP-70, Stefanova et al. (62) showed that the signaling properties of TCRs change during thymocyte maturation, differentially affecting responses to related pMHC molecule complexes. They suggested that a combination of programmed events and tuning was at work. Marquez et al. (63) carried out in-depth investigation of T cell responsiveness tuning as a mechanism to control autoreactivity in vivo. To find out how mature T cells and thymocytes cope with excessive autoreactivity, they increased the excitability of the TCR signaling machinery by using a gain-of-function mutant of ZAP-70. The mutant was expressed in a transgenic mouse model in post-selection CD8+ thymocytes and T cells, resulting in a strong increase in TCR-dependent signaling. The mutant’s cells were not activated, but rather they readjusted their activation threshold dynamically, in part by reducing TCR levels. The physiological relevance of such artificial manipulations is enhanced given the natural variation in the sensitivity of the signaling apparatus, and the findings are in line with the conclusions of in silico experiments (40). Predictably, these adaptive modifications that countered enhanced autostimulation affected the response of tuned cells to external stimulation: They produced lower cytokine amounts and showed faster dephosphorylation of ZAP-70 following its faster transient activation, the latter observation suggesting enhanced feedback induction of phosphatase and/or recruitment of stored phosphatases. Normal TCR levels and cytokine production were restored by culturing cells in the absence of TCR–spMHC interaction, demonstrating that tuning was dynamically geared to the level of stimulation. CD5 levels were also dynamically adjusted, but, somewhat surprisingly, they were lower in the mutant-bearing mice, suggesting that T cells generally exploit different tuning mechanisms in a flexible, nonstereotypic manner. The effect of avidity for self-ligand(s) on this sensory adaptation was directly demonstrated by expressing the ZAP mutant in two different TCR transgenic mice representing prototypes of high and low avidity for spMHC. The investigators concluded that signaling plasticity based on TCR and CD5 level modulation operated during thymic maturation concomitantly with selection processes, endowing flexibility to the latter to facilitate development of T cells that bind their selecting self-antigens with lower or higher affinity (63). A later study used a series of bone marrow chimeras and genetically modified mouse models to further dissect mechanisms of tuning during T cell development (64). MHC-II expression on either medullary thymic epithelial cells or thymic DCs was sufficient to progressively reduce TCR responsiveness to weak ligands. Persistent TCR interactions with pMHC-II on thymic medullary stroma inhibited ERK activation and elevated SHP-1 expression in maturing SP thymocytes. The authors did not study the relation of these effects to CD5 expression. In contrast to the findings reported by several others and reviewed here, the authors suggested that interactions of immature SP thymocytes with pMHC-II resulted in nonreversible changes in the TCR signaling properties, as mature cells transferred into MHC-II-deficient thymi did not alter their responsiveness. Even before the dynamic aspect of developmental tuning in thymus was established, several studies indicated that SHP-1 sets the TCR signaling threshold between positive and negative selection (reviewed in 65). This was challenged in a recent study using Cre recombinase gene under transcriptional control of elements of the Cd4 gene to induce conditional deletion of Shp-1. No obvious differences in positive or negative selection relative to wild-type cells were seen (66). However, as already suggested by the findings of Marquez et al. (63), tuning mechanisms are likely to be


IY33CH22-Paul

22.16

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

redundant and context-dependent rather than stereotypic. There is possible redundancy in protein tyrosine phosphatase function through activity of SHP-2 and potentially other phosphatases, especially those involved in regulation of ERK signaling, such as DUSP5 and DUSP6 (65). Recently, Germain and colleagues (67) found a highly significant linear relationship between the TCR-proximal signal strength measured by phospho-ζ and the expression of CD5. Such positive correlation between CD5 and basal TCR signaling strength appeared unexpected (7, 65, 67), given the inhibitory role of CD5 on TCR signaling, but in fact it is fully consistent even with a simplistic model in which CD5 mediates perfect feedback control of pMHC-induced excitation. Although the regulation of intracellular signaling by CD5 results in a reduced TCR signal strength, as compared to the strength it would have if CD5 regulation were absent, a higher-affinity binding of a ligand to the TCR necessarily resulted in a stronger subthreshold signal, countered but not annulled by a higher level of CD5 expression. Therefore, although feedback control has a normalizing effect, acting to blunt differences in the perturbations to the excitation-de-excitation balance elicited by self-antigens in different clones, such a mechanism is not supposed to equalize subthreshold TCR signal strength among CD4+ T cells. Centering on the role of TCR–MHC interaction on T cell survival, Takada & Jameson (68) showed that depriving naive CD8 cells of contact with MHC-I ligands in nonlymphopenic hosts leads to a shortened life span but upregulation of CD8 expression in the surviving cells. The cells were thereby retuned to give enhanced responses to foreign antigens, though only to weak antigens (68). These results were consistent with work reported by Singer and colleagues (69), who linked CD8-mediated tuning to a dynamic modulation of IL-7 signaling and to homeostatic regulation of the CD8+ T cell compartment (coreceptor signaling; see below).

Tuning Enhances Sensitivity to Agonist Ligands Using three separate approaches, Stefanova et al. (43) tested the functionality of CD4+ T cells that were removed for hours or minutes from environments in which MHC-II was expressed and therefore deprived of interaction with self-antigen. In each case, they observed a loss of TCRζ phosphorylation and reduced responsiveness to antigen. CD4+ T cells interacting with APCs had modestly polarized surface TCR expression, and disrupting the interaction with self-antigen resulted in redistribution of the TCRs. The authors concluded that maintenance of interaction with self-antigen-MHC ligands enables the T cell to better respond to foreign antigen by inducing partial TCRζ phosphorylation and clustering receptors at the APC interface (see also 70, 71). These findings were seemingly at odds with the desensitization model (43). But in fact, sensitizing the sensory end of the T cell’s signal transduction pathways to improve performance in a future immune response is perfectly in line with the broader perspective of the tuning hypothesis, which maintained that the advantages of subthreshold autostimulation include active maintenance of the lymphocyte’s functional integrity. In addition, such sensitization presumably improves the cell’s capacity to efficiently and sensitively participate in surveillance of its antigenic environment for the presence of small amounts of cognate antigen during brief abortive binding events. Importantly, such a sensitization effect need not normally affect the activation threshold significantly, as a parallel tuning of de-excitation factors at the level of TCR-proximal signaling can be assumed to counter this effect, in the same way it was shown (as discussed above) to counter gain-of-function of ZAP-70. The increase in the threshold for proliferation measured by Stefanova et al. (43) upon deprivation from interaction with self-antigen presumably reflected a delay in downregulation of the other factors due to their relative stability or, more likely, a more generalized reduction in performance due to the deprivation. Therefore, these findings do not contradict other studies that demonstrated a reduction in the activation threshold upon longer-period deprivation of T www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.17

IY33CH22-Paul

ARI

30 January 2015

9:51

cells from TCR-mediated stimulation (discussed in the following section), although those results might also be confounded by a nonphysiological loss of functional integrity. A rather similar interpretation was offered by Hogquist et al. (72).

TUNING INHIBITS ACTIVATION OF MEMORY-PHENOTYPE T CELLS MHC Restrains Auto-Aggression


Singer and colleagues (73) assessed the TAT prediction that peripheral MHC-II expression controls T cell reactivity. They transferred mature CD4+ T cells into an environment lacking both MHC-II and T cells or into mice lacking T cells but expressing MHC-II. The cells adopted an activated memory phenotype upon transfer into both recipients. However, only in the MHCII-deficient hosts did the transferred CD4+ T cells mediate rejection of subsequently grafted syngeneic MHC-II-expressing skin, indicating loss of tolerance to self. The authors showed that the CD4+ T cells were hyperresponsive to TCR-mediated stimuli in vitro, implying a reduced level of tuning. Additional evidence that spMHC interactions restrain autoimmunity by tuning was provided by transfer of naive T cells into mice lacking both MHC-I and MHC-II (74). These cells acquired the capacity to reject subsequently transplanted wild-type pancreatic islets or transplantable tumors. In contrast, naive T cells undergoing homeostatic proliferation in an MHC-replete environment did not acquire such capacity. The authors implicated intrinsic mechanisms largely independent of Treg cells. They concluded that unresponsiveness to tumors or to pancreatic islets in nonautoimmune-prone animals is reversed by the absence of ligands that limit self-reactivity, i.e., spMHC.

Adaptive Tolerance Dynamic tuning of memory-phenotype T cells postactivation has been thoroughly studied by Schwartz and colleagues. Tanchot et al. (75) established a model in which transferred TCR transgenic CD4+ T cells were chronically exposed to stimulation by their cognate agonist ligand, transgenically expressed in a lymphopenic environment. Transferred naive cells expanded extensively for several days, acquiring a memory phenotype, and then became hyporesponsive to TCR-mediated stimulation, as revealed by several in vitro assays and by their poor proliferation upon secondary transfer into a similar antigen-bearing lymphopenic environment. The hyporesponsiveness slowly dissipated without proliferation when the cells were transferred into a nonantigen-bearing host and became more pronounced when the second host expressed the antigen. Singh & Schwartz (76) then examined the dependence of this tuning on antigen concentration. The transgenic T cells were transferred to lymphopenic environments expressing fourfold different levels of their cognate antigen, to which they responded differently. T cells expanded and entered a tolerant state with different kinetics in response to the two levels of stimulation. In vivo restimulation revealed a greater impairment in the proliferative capacity of T cells resident in the environment with higher antigen presentation. Differences in TCR signaling and in vitro cytokine production on stimulation were consistent with that different adaptation. The observations were consistent with TAT. Naive T cells preexisted in a self-tuned state, i.e., in equilibrium with the level of spMHC, and their transfer from an antigen-free environment to one expressing cognate antigen abruptly changed the level of the ambient stimulus and elicited a proliferative burst. At the end of this phase, the cells differentially and reversibly adapted their 22.18

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

activation thresholds to the level of chronic exposure to antigen and maintained a quasi steady state. Although signal discrimination during each T cell–APC encounter depends mainly on the quality of binding rather than on ligand concentration, the latter likely affects the frequency of effective encounters contributing to tuning. A higher frequency of effective tuning events would lead to a higher level of TCR-associated de-excitation factors and therefore to a higher activation threshold (see Figure 1). Another reason for the difference is the multiphasic/multidimensional aspect of activation and tuning discussed above. Adaptation to chronic stimulation by a strong agonist, unlike that to common self-antigens, may often generate supercritical perturbations at the TCR level. In that case, downstream modules become tuned as well during recurrent stimulation events, triggering feedback molecular responses and setting the levels of different sets of de-excitation molecules. Chronic stimulation with a higher ligand concentration would result in a more extensive integration of signals from activated TCRs downstream, entailing induction of a higher steady-state concentration of de-excitation molecules that set higher thresholds for proliferation and other functions. Importantly, adaptive tolerance was also induced to a similar degree in the physiologically more relevant T cell replete environment (77). Further investigating the molecular correlates of adaptive tolerance, Schwartz and colleagues (78, 79) showed that TCR tuning resulted in a reversible inhibition of TCR-proximal signaling events required to initiate the cascade of downstream signals leading to cell activation, including impairment of LAT phosphorylation, a failure to recruit VAV proteins into the T cell–APC interface, and impaired formation of the immunological synapse. Programmed death -1 (PD-1) expression has been associated with T cell exhaustion in certain chronic infections. It is also highly expressed by a fraction of short-lived CD4+ T cells generated during acute activation bursts in animal models (80). Whether PD-1 is also involved in tuning is still an open question. An intriguing recent study revealed a role for rapidly inducible PD1 expression in a dynamic TCR-dependent tuning of the sensitivity of effector CD4+ T cells to foreign antigen in inflamed tissues (81). After TCR engagement by antigen-bearing APCs, effector T cells immediately exhibited reduced velocity and high IFN-γ production, but within a few hours PD-1 expression was increased and the cells gradually ceased effector activity and recovered motility. The investigators proposed that this mechanism constrains tissue damage by adapting the magnitude of the effector response to the amounts of ambient antigen within the inflamed site. However, it is not yet clear whether PD-1 expression by individual cells is reversible in this context or whether PD-1-expressing cells are constantly replaced by newly generated effector cells, or both.

Detuning T Cells by Inflammatory Cytokines It would appear that, to be a reliable gatekeeper of signal discrimination, tuning must rely on intrinsic cellular measures of signal strength that are not tunable themselves but rather are invariant, having been determined genetically and epigenetically. This may be an oversimplification. In the local microenvironment of gut-associated lymphoid tissues, inflammatory cytokines elicited by commensal flora dynamically enhanced antigen responsiveness of memory-phenotype CD4+ T cells whose activation thresholds were otherwise tuned to a systemic self-antigen (82). This was demonstrated both by ex vivo functional assays and by documenting exacerbation of Th1 cell–driven autoimmune arthritis for which disease incidence and severity were linked to a specific flora. The molecular basis was not studied in detail, but the relevant cytokine, IL-12, appeared to enhance mammalian target of rapamycin (mTOR) activation in the CD4+ T cells as it was reported to do in CD8+ T cells (83). These results demonstrated that tuned T cells can further adapt www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.19

ARI

30 January 2015

9:51

dynamically to changes in the local nonantigenic microenvironment. Such additional plasticity likely has both a physiological basis and pathogenic implications, depending on circumstances. It has also been reported that cytokines such as IL-7 and IL-21, independently and synergistically, downmodulate CD5 expression in CTLs by upregulating the expression of the E protein family transcription factor E47, a transcriptional repressor of CD5 (84) (reviewed in 58). Such cytokine signaling–dependent downmodulation of CD5 leads to increased responsiveness of CD8+ T cells to antigen and provides a mechanism by which CTLs can develop adaptive ability to increase their antitumor activity. In the context of infection, pathogen-associated inflammatory cytokines were shown to dynamically tune even unrelated memory T cells for a faster and more efficient effector response upon superthreshold TCR stimulation by their cognate antigens. Richer et al. (85) demonstrated that the antigen-dose sensitivity of memory CD8+ T cells is dynamically regulated at the individual cell level by pathogen-induced inflammatory cytokines (see also 86). Antigen sensitivity was measured ex vivo, to stimulation by DCs pulsed with titrated amounts of peptide. The results implied that infection enhances the antigen sensitivity of memory CD8+ T cells regardless of their antigen specificity. The signal transduction capacity of key TCR-proximal molecules was enhanced, which reduced the antigen density required to trigger antimicrobial functions. Antigen sensitivity of circulating memory T cells was blunted rapidly as inflammatory stimuli waned. These findings are consistent with results described by Raue et al. (87) in which memory lymphocytic choriomeningitis virus (LCMV)–specific CD8+ T cells cultured for 5 h with a combination of IL-12 and IL-18 demonstrated enhanced antigen-independent proliferation, cytokine production, and cytotoxicity in vitro and clearance of LCMV upon adoptive transfer to LCMV-infected mice. Although cytokines sensitize T cells nonspecifically in the inflamed microenvironment, enhancing their responsiveness to small amounts of their cognate antigen and preparedness to deploy effector functions, the tuned resistance of the TCRs to activation by spMHC ligands may remain largely intact. This would be the case if cytokine-induced sensitization affects mainly the downstream module(s), where signals from activated TCRs are integrated and further transduced.


IY33CH22-Paul

TUNING, VIABILITY, AND HOMEOSTASIS Reciprocal Tuning of T Cells and APCs Activated DCs change their inductive properties in the course of an immune response (88) and continue to retain behavioral plasticity long after the initial activating stimulus has ended (89). Abdi et al. suggested that “DCs are messengers, rather than controllers, continually integrating signals from the environment and modulating the messages they deliver” (89, p. 5988). Several studies have revealed the ability of T cells to stimulate or inhibit DCs and in this way facilitate or suppress the response of other T cells that recognize antigen on the same DCs (90). This capacity does not seem to be limited to Foxp3+ Tregs, natural or induced. Normally, the tuning interaction of a CD4+ T cell and a DC occurs under conditions in which relatively low levels of MHC-II and costimulatory molecules are expressed on the DC surface. Vendetti et al. (91) showed that contact between mouse T cell clones that had been rendered unresponsive in vitro and nonactivated DCs presenting cognate antigen reduced even further the immunogenicity of the DCs. Expression of MHC-II, CD80, and CD86 (but not CD40) on these DCs was downregulated in such interactions. We proposed that active regulation of the expression of MHC and costimulatory molecules on DCs occurs, dynamically, during recurrent subthreshold interactions with resting T cells that can bind, with sufficient affinity, pMHC ligands presented by these DCs (8). The stimulation capacity of the DCs and the activation thresholds of the T 22.20

Grossman

·

Paul


IY33CH22-Paul

ARI

30 January 2015

9:51

cells would thereby be tuned concomitantly and reciprocally: The level of T cell tuning would be adjusted to the controlled level of subthreshold stimulation by the DCs.


Reciprocal Tuning and Treg Cells Control Naive to Memory-Phenotype Transitions Most memory-phenotype T cells in young adult mice do not appear to have originated in conventional antipathogen immune responses of naive T cells. As a population, they can be distinguished from conventional memory cells by their relatively high resting self-renewal rate (∼10% in S phase each day compared to ∼1% of authentic memory cells) and high clonal diversity (17). [In the following, the term memory-phenotype cells refers to this population of innate-like T cells (16) unless otherwise stated.] We have recently proposed a unifying hypothesis regarding the interplay of tuning, Tregs, and pathogen sensing in the regulation of T cell responses to self and foreign antigens (8). Below, we list our assumptions and mention relevant observations. 1. Because of the stochastic nature of self-interactions in terms of frequency and strength, TCR tuning likely does not prevent occasional activation of those autoreactive naive T cells with moderate but significant binding affinities for their ligands, especially those belonging to the higher-affinity tail of the spectrum. Such events may occur quite often but remain undetectable when not leading to significant proliferation and manifestation of effector functions (92). Indeed, CD5high naive CD8 T cells, in particular, have been described as “constantly poised to lose quiescence, given a change in homeostatic cues” (93, p. 229). 2. To produce an explosive proliferation/ differentiation burst as a result of self-reactivity, a certain “cell-coupling threshold” needs to be exceeded (3), enabling a positive circuit of reciprocal stimulation between the T cell and the APC to be switched on and allowing the partners to overcome intrinsic negative controls and extrinsic negative regulation. Such bursts normally do not develop in a noninflamed environment, except under lymphopenic conditions. 3. Intrinsic controls involve molecules such as Tob, Nfatc2, and Smad3 that actively enforce T cell quiescence (94). Extrinsic regulation includes (a) tonic T cell–mediated suppression of the resting APC’s capacity to stimulate, and (b) Treg-mediated, contact-dependent reversal of DC activation once it has been initiated. The Treg and the conventional T cell need to recognize self-peptides on the same APC. 4. In a lymphopenic environment, there is no preexisting tonic inhibition of APCs, and therefore a small fraction (see 8) of naive T cells whose TCRs can be spontaneously triggered by self-ligands receive strong costimulation, expand extensively, and may express effector functions. Preexistence of a heterogeneous, but not homogeneous, population of memoryphenotype cells of a sufficient size can largely block the expansion of any naive T cell clone, whereas memory-phenotype cell populations of intermediate heterogeneity allow the expansion of a certain fraction of naive T cell clones (reviewed in 8). 5. Due to stochastic variability in intracellular content of various self-proteins, each individual DC may effectively present only a partial set of spMHC (first proposed in Reference 95; indirect evidence is discussed in Reference 8). By “effectively present” we mean that they present spMHC at a high enough frequency to activate Tregs with matching TCRs or to engage resting T cells with such TCRs in reciprocal tuning. The expression of only a subset of spMHCs on any particular DC explains the requirement for heterogeneity of the memory-phenotype population in order to enable it to engage and suppress simultaneously all or most DCs, blocking activation of most naive T cells and restraining the proliferative www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.21

ARI

30 January 2015

9:51

response of those few that become activated. A limited TCR repertoire when the total number of cells is similar implies that, among those cells, many have the same specificity and thus as a population compete less well. 6. We speculate that memory-phenotype cells, like Tregs, predominantly mediate their function in the periphery by recognizing the spMHC molecules that they recognized during their thymic positive selection or related molecules. This provides the evolutionary advantage for their affinity to be relatively high for the selecting molecules because such complexes will be their cognate antigen in the periphery. Memory-phenotype cells are presumably subject to tuning but are activated and participate in immune responses, as do conventional memory cells during secondary responses, either through the binding of relevant self-ligands on APCs that have been preactivated by pathogen-related adjuvants or possibly through direct action of cytokines. 7. Activation by self-ligands in this context, despite an established state of tuning (of the TCR and downstream signaling modules), is facilitated by the abrupt nature of the perturbation. Because of its relatively high binding affinity to its cognate spMHC complex, a memoryphenotype T cell probably relies more on a deep tuning downstream from the receptor rather than on TCR tuning alone; after all, TCR-level tuning did not protect the naive precursor of this memory-phenotype cell from activation in the first place. When DCs are turned on by pathogens (see below), or stimulated by adjuvants/inflammation in the off state, they upregulate MHC expression and other accessory molecules and thus may more potently present more spMHC to a tuned cell, stimulating it to a level exceeding the adjusted activation threshold established by having interacted with DCs expressing relatively low levels of these molecules. Failure to quickly adjust the level of tuning to the new level of stimulation in such cases would result in exceeding the proliferation threshold. This presumably does not apply to most naive and conventional memory bystander T cells possessing lower self-ligand binding affinities, whose tuned TCR-excitation thresholds cannot be overcome just by increased concentration of presented ligand and costimulation; as discussed above, variation of these parameters can affect downstream competition between excitation and de-excitation processes but not the balance of TCR-proximal signaling. 8. Thus, regulation of T cell autoreactivity by DCs is in part specific and in part semispecific. TCR-mediated tuning of the T cell activation threshold is fully specific, mediated by specific TCR-spMHC binding, whereas the regulation of autoreactivity via modification of the seemingly nonspecific ligand presentation and costimulation properties of DCs is actually semispecific, as the limited range of effective ligand presentation by each DC diminishes the interference and competition among clones. This facilitates the buildup and maintenance of a diverse memory-phenotype cell population through selective conversion of naive T cells into memory-phenotype cells. Preferentially blocked will be the spontaneous activation and/or proliferation of naive autoreactive T cells whose cognate self-ligands are already engaged by existing memory-phenotype cells of the same specificity. This is because DCs that preferentially (abundantly) present the self-ligand will be suppressed. Cells specific for the same pMHC complexes may find it harder to coexist. Partial or complete depletion of memory-phenotype cells specific for that ligand or/and of Tregs of the same specificity will tend to increase the rate of activation and proliferation of naive cells, leading them to acquire the memory phenotype. Therefore, the DC population provides a fluid system of effective niches occupied by memory-phenotype T cells. When DCs within such niches are suboptimally engaged, they selectively prompt the activation of naive T cell that recognize pMHC that the DCs express. Accordingly, holes in the memory-phenotype cell repertoire would be regularly filled.


IY33CH22-Paul

22.22

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

9. If reciprocal tuning alone is able to control autoreactivity, as also demonstrated by the above-described experiments related to adaptive tolerance that did not involve Tregs, then what is the role of Tregs? Tregs are essential to counteracting excessive stimulation that DCs provide to spMHC-binding T cells in the absence of reciprocal tuning. Pathogensensing mechanisms enhance DC-provided stimuli even when these DCs do not present pathogen-derived peptides but only self-peptides. Thus, it is likely that Tregs are preferentially recruited to those DCs that are strongly activated by molecular cues associated with such a level of activation. These cues are provided by the DC and/or by the responding T cell. One such example is T cell–derived IL-2. In response to an IL-2 cue, Tregs exert their regulatory functions when they are most needed. 10. Adjuvants or strong exogenous pathogen-sensing reactions overcome the Treg-mediated restraint on DC activation and allow stimulated T cells interacting with stimulated DCs to expand and differentiate in a burst-like fashion. Elaborating previous arguments (8), we propose that when Tregs keep DCs turned off, activated T cells can still proliferate, variably, and even differentiate, but within a limited range of differentiated phenotypes. However, artificial or natural pathogen-derived adjuvants enhance the cooperative, bidirectional exchange of signals between the DC and a pathogen-specific T cell. And given the advantage of a stronger binding affinity, the interacting cells typically overcome Treg-induced negative signaling, turning on a positive feedback loop. This leads to a burst-like response involving rapid proliferation, differentiation, and expression of effector functions. As in the case of the TCR’s excitation threshold, emerging from the coupling of negative controls and a positive feedback loop (see above), a switch-like transition of the DC–T cell interaction from an off to an on state may be interpreted as the breaching of a cell coupling threshold. Thus, Tregmediated opposition to the triggering of full-blown immune responses (96) is not futile but rather enhances the discrimination between pathogens and self.

Lymphocyte-APC Interactions Tune T Cell Viability and Maintain Diversity Tonic TCR signals, in conjunction with IL-7, maintain cell viability by promoting expression of prosurvival molecules such as Bcl-2 (6). A major rationale for a continued dependence of peripheral T cells on peptide-specific, TCR-mediated stimuli in maintaining viability and self-renewal is that it imposes on the population a similar degree of clonal diversity to the one that has been achieved during positive selection (97). As different self-peptides are present in different quantities and different T cells have different cross-reactivities, this strategy by itself would imply high variation in the frequencies of different TCRs. However, combined with receptor tuning, which calibrates TCR sensitivities to self-peptides, more uniform clone frequencies could be maintained (98). T cell parameters such as the expression of the IL-7 receptor are dynamically adjusted, and therefore the process is categorized as dynamic tuning. The size of the naive and memory T cell populations are quite strictly maintained in the face of variations in the inputs and outputs and of other perturbations (95), and, together with a strict maintenance of clonal diversity, this strongly implies the existence of specificity-based niches (99), with well-defined carrying capacities and filling-up mechanisms. The niche is not a unit of space but rather defined at any time by the ensemble of DCs in secondary lymphoid tissues presenting a given spMHC. Without making specific observations or assumptions regarding the homeostatic mechanisms, which are largely out of the scope of this review, it can be safely argued that the antigen-presenting DCs must also be subject to tuning. The only way for niche DCs to count and regulate the number of T cells that belong in the niche (in a sense) is to gear the presented ligand concentration, or other parameters associated with tuning of the T cell’s viability and renewal, www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.23

ARI

30 January 2015

9:51

to the number of existing cells. More specifically, because cells of a given specificity are mobile and sparse, the DC must measure the frequency of encounters over time, and this implies cellular memory of such encounters and integration of molecular traces that are left by them: namely, tuning. TCR-mediated signaling (via a prosurvival pathway distinct from the T cell activation pathway) and cytokines such as IL-7 act synergistically to regulate naive and memory cell viability and to induce renewal division of those cells. TCR-mediated signals upgrade the cell’s responsiveness to the cytokines. Indeed, homeostatic (subthreshold) TCR signaling enhances sensitivity to IL-7 (reviewed in 100); stronger TCR signals promote increased survival (101). We note that niches determine the relative responsiveness and therefore the relative size of clones, whereas availability of cytokines determines the size of the total T cell population. Because of the characteristic degeneracy of the TCR–ligand interaction, especially of weak binding, a TCR may cross-react significantly, in the subthreshold mode, with several different spMHC complexes. The ligand that most significantly tunes the cell’s activation threshold need not be one of those effectively delivering pro-viability signals. T cells are presumably also selected on the basis of their ability to detect low-affinity self-ligands (but multiply expressed on a DC) with high sensitivity and with functional consequences, in addition to their sensitivity for a small number of high-affinity foreign ligands in the midst of such self-ligands. A striking example of uncoupling the tuning of the activation threshold from viability tuning was reported by Singh et al. (102). They showed that the viability of TCR-transgenic CD4+ T cells— which had been rendered tolerant to their cognate transgenically expressed agonist peptide—was inhibited by a smaller clone of T cells cross-reactive for a nonagonist self-peptide. Thus, despite the strongly tuned state of the agonist-adapted clone, implying a high baseline of de-excitation factors along the TCR-mediated activation pathway, viability of the cells could be sustained in the absence of the cross-reactive resident cells via weak interaction with a self-peptide. This example also illustrates the concept that TCRs on the same individual cell can use their cross-reactivity to interact simultaneously with different peptides, either independently (as in this case), cross-purposely (when one set antagonizes the other), or synergistically (e.g., when selfpeptides act as coagonists during recognition of foreign antigen) (103, 104).


IY33CH22-Paul

Coreceptor Tuning The biological purpose of ongoing TCR–spMHC subthreshold signaling is to promote the viability of lymphocytes in such a way that clonal diversity is maintained and functional adaptation facilitated; the need for activation-threshold tuning to maintain self-tolerance is a consequence. Indeed, Singer and colleagues (69) have referred to this ongoing interaction as “homeostatic TCR signaling.” Their work highlighted a surprising mechanistic relation between these two targets of tuning, viability and self-tolerance. They reported that signaling via IL-7 and signaling via TCRs affect each other in naive CD8+ T cells in a process referred to as “coreceptor tuning” (69). TCR-engagement time is controlled by the amount of CD8 that the T cell expresses. In addition to promoting T cell survival and regulation of their own receptors, IL-7 molecules (and other common γ-chain cytokines), which are produced by other cell types, transcriptionally increase CD8 expression and thereby promote TCR engagement of self-ligands. These authors reported that self-antigen-induced TCR signals then impair cytokine signaling and thereby decrease CD8 expression, causing TCR disengagement. Importantly, this interplay between cytokine and TCR signaling quantitatively modulates baseline CD8 expression so that CD8 expression is inversely proportional to the intensity of

22.24

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

TCR signals induced by self-antigens. Because the CD8 synergizes with the TCR to enhance ligand-induced proximal signaling, the net result of stronger TCR-ligand binding is elevation of the TCR excitation threshold, which amounts precisely to activation-threshold tuning. Interestingly, the transient interruption of IL-7 signaling during binding does not result in a decreased baseline expression of the IL-7 receptor; on the contrary, that expression is higher for more avidly binding spMHC ligands, as mentioned above. The critical connection between TCR signaling and survival/viability is found elsewhere. Singer and colleagues (105) reported that intermittent interruption of IL-7 signaling, as described, is paradoxically what is critical for T cell survival. Uninterrupted IL-7 signaling results in apoptosis triggered by IFN-γ, which continuous IL-7 signaling induces CD8+ T cells to produce, an outcome referred to as cytokine-induced cell death (CICD) (105). Singer and colleagues (105) proposed that one biological manifestation of this requirement for interruption of IL-7 signaling by homeostatic TCR signaling would be that CD8+ T cells with TCRs of insufficient affinity to engage peripheral self-ligands would not survive in vivo because they are continuously signaled by IL-7 to express IFN-γ and to preferentially undergo IFN-γtriggered CICD. Only naive CD8+ T cells with TCRs that have sufficient affinity for in vivo self-ligands to interrupt IL-7 signaling are maintained during in vivo homeostasis, again stressing the notion that autoreactivity is a critical feature of a functional immune system. This proposed manner of dependence on TCR signaling for survival bestows the necessary specificity to the homeostatic regulation of CD8+ T cells and, moreover, defines a novel strategy to ensure the maintenance of diversity (that is, coexistence in the periphery of as many clones as positive selection in thymus may have generated). The growth of all of the clones would be similarly restrained because access to their recognized ligands is limited, or, alternatively, because a high encounter rate of a T cell clone with those APCs that collectively define the clone’s interactive niche (see above) will saturate the capacity of the niche APCs to interrupt IL-7 signaling. Specifically, an increasing frequency of encounters could increasingly reduce MHC expression on APCs, which would increasingly fail to interrupt IL-7 signaling often enough to prevent apoptotic T cell death.

Impact of Tuning on the Quality of Subsequent Responses We have raised the possibility that the self-peptide-dependent interaction of CD4+ and CD8+ T cells with DCs gives rise to reciprocal tuning processes, dynamically adjusting the activation thresholds of both parties and the viability and renewal capacity of (at least) the T cell partners. Because the interaction is effectively structured in terms of receptor specificities into many peptideand/or DC-defined niches, the process we have described provides a sound basis for homeostatic regulation, limiting clonal sizes and maintaining diversity. In addition, depriving T cells of contact with MHC for different periods of time demonstrated that this interaction is essential for maintaining the functional integrity of T cells (reviewed in 6), as had been hypothesized (1, 4). Recent studies have started to shed light on the range of functional characteristics that the recurring contact with spMHC complexes regulates. From Listeria monocytogenes–infected mice, Weber et al. (106) generated two CD4+ T cell lines with transgenic expression of similar TCR molecules that differed in only 15 residues in their complementarity-determining regions. These lines had identical sensitivity to the immunodominant listeriolysin epitope of L. monocytogenes but different avidity for spMHC, as indicated by their different surface expression of CD5. Despite this identical sensitivity to a pathogen-derived peptide, cells from these lines responded differently in secondary normal hosts to infection with the pathogen. The CD5high cells underwent more activation-induced cell



22.25

ARI

30 January 2015

9:51

death and expanded to a lesser extent. Recent work by this group confirmed that the two clones indeed had identical affinity for their L. monocytogenes epitope but different affinities for spMHC and correlated the susceptibility to activation-induced cell death of CD5high cells to a greater phosphorylation of Erk and production of IL-2. These differences had been prewired into the cells during their development in the thymus (because the same phenotype was observed in mature thymocytes) but required continued engagement of spMHC molecules in the periphery to be sustained. Indeed, much of the phenotype of the T cells was lost when these cells were deprived of contact with spMHC [15; see also commentary by Martin & Surh (107)]. The observation that the functional differences were quantitatively linked to the strengths of TCR-mediated weak stimuli and were reversible strongly suggests dynamic tuning, both developmental and peripheral. It would be of great interest to find whether such cells can change and even reverse their functional phenotype in different microenvironments with different peptide contents. Other results reported by Persaud et al. (106) are also consistent with the conventional form of tuning, namely, desensitization of the TCRs. They found that although deprivation of contact with spMHC before activation negatively affected the response of each clone, the population expansion of the CD5low cells was largely unaffected, and that of CD5high cells was diminished even further. This is explained by the fact that detuning of the TCRs of the latter cells enhanced their response to the pathogen-derived ligand, which entailed increased IL-2 secretion and accelerated activationinduced cell death. In the commentators’ words (107), the relatively higher avidity of CD5high cells for self-molecules correlated with a stronger activation response in terms of the phosphorylation of Erk and production of IL-2, as well as a greater dependence on contact with spMHC ligands for the clone to mount any response at all. Adjustment of the activation threshold to the strength of ambient signaling presumably occurs in parallel with the functional tuning downstream of the TCR via different pathways. What is the biological significance of the generation of suboptimally responding T cell clones, if any? A conceivable role for such T cells is that they participate in the immune response by producing large amounts of IL-2 and then die (107). Alternatively, such clones might have encountered a weak agonist, in which case they would have mounted a correct response based on their imprinted gain of function. More generally, assuming that the above-described observations can be generalized, we suggest that subthreshold TCR-mediated stimuli (homeostatic TCR signaling) promote not only clonal diversity but functional diversity as well. A broad heterogeneity of baseline propensities, predicting a wide range of differences in the balance of proliferation, differentiation, effector functions, and death in response to identical stimulation, would ensure an optimal response to a variety of agonists. Both clonal diversity and functional heterogeneity endow the immune system with the vital resilience that it evidently shows in the face of an enormous variety of challenges, which cannot be anticipated based on evolutionary or previous life experience. Thus, the structural diversity of peptides not only drives a corresponding structural diversity of TCRs but may also create a broad repertoire of cells differing in their functional properties, which the system can use in a context-dependent manner to optimize its response. Thus, both clonal repertoire and broad functional heterogeneity are generated in the thymus and then actively maintained, and perhaps even regenerated as needed in peripheral tissues. This process is enabled by the plasticity (tunability) of signaling modules and by the cross-reactive and partly stochastic manner in which T cells sort and integrate ambient signals while in a relative resting state. It becomes increasingly clear that maintaining preparedness for potential challenges is a prerequisite for a successful response. An essential element of such preparedness is prediversification of the response capabilities of individual cells at all levels. For diversification by tuning, the strategy is a continuous, strictly regulated interaction with the world of self-peptides.


IY33CH22-Paul

22.26

Grossman

·

Paul


IY33CH22-Paul

ARI

30 January 2015

9:51

POSTACTIVATION TUNING, CHRONIC INFECTION, AND EXHAUSTION


General Considerations Anderton and colleagues (108, 109) investigated the mechanisms that prevent the induction of experimental autoimmune encephalomyelitis (EAE) when myelin basic protein (MBP)-reactive T cells are exposed to a strong, antigenic stimulus. MBP-reactive TCR transgenic T cells that survived stimulation with a cross-reactive agonist peptide had a reduced sensitivity for the selfantigen that correlated with elevated levels of CD5. This phenotype was reversible in culture. The authors concluded that the cells were subject to sensory adaptation (i.e., tuning) to MBP in the host. (Postactivation tuning is broadly discussed in Reference 95.) To what extent conventional memory T cells rely on activation-threshold tuning for selftolerance is not known. Unlike memory-phenotype T cells that may have arisen spontaneously, based in part on a higher reactivity to spMHC, conventional memory cells are activated by their cognate antigen and would typically have relatively weak affinity for self, and thereby would show little or no evidence of tuning. The maintenance of memory CD8+ T cells seems to be independent of spMHC (13), whereas that of memory-phenotype CD4+ T cells may continue to depend on TCR signals (7). A dependence on self-antigens is required to ensure the repertoire diversity of memory-phenotype T cells (and Treg cells). In contrast, the repertoire of conventional memory cells is narrower, reflecting selection by pathogens, and there seems to be no purpose for reshaping the relative sizes of antigen-specific memory cell populations on the basis of autoreactivity. Specific memory cells require smaller antigen concentrations and respond more rapidly to stimulation by their cognate antigen than naive cells, but this does not necessarily mean that their TCR excitation threshold is lower. In fact, that threshold may be higher, so that most conventional memory T cells might essentially ignore weak cross-reactive spMHC. At the same time, their downstream threshold(s) for proliferation and effector functions appear to be inherently lower, requiring less signaling from TCRs once the latter have been triggered individually by the agonist. These considerations probably do not apply to memory-phenotype T cells, which may respond to spMHC as their cognate antigen when the APCs have been activated by natural adjuvants. We speculate that the activation thresholds of memory-phenotype cells are dynamically tuned and that they depend on TCR-mediated signals for survival and for homeostatic regulation. Acquisition and active maintenance by antigen-specific T cells of adaptive tolerance would be required during chronic infection to limit the immunopathogenic damage to tissues associated with a protracted, full-blown immune response. The dynamic nature of tuning and the hierarchy of signaling modules would allow T cells to adjust their functional state to the strength and nature of the persisting stimuli, enabling them to maintain and display a partial set of effector functions. In this way they could control or confine the pathogen, short of eradication, while minimizing the collateral damage.

Tuning Versus Exhaustion According to the prevailing view, the hyporesponsive state manifested by pathogen-specific T cells in several chronic viral infections (or among tumor-infiltrating cells) is a pathologic cellular phenotype acquired progressively in response to excessive stimulation (as the term exhaustion implies). This view is undergoing revision. An influential paper described exhaustion as a process in which antigen levels drive the hierarchical loss of different CD8+ T cell effector functions, leading to distinct stages of functional impairment and eventually to physical deletion of virus-specific T cells (110). Yet the authors www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.27

ARI

30 January 2015

9:51

suggested that, rather than a subversion of the host immunological defense, “the differential loss of effector functions may reflect a balance between minimizing immunopathology and maintaining some level of antiviral capability” (p. 4925). Recent studies suggest that exhausted T cells may actually represent various states of tuning, activation, and differentiation displayed by different cells. Tuning implies reversibility of the functionally impaired phenotype when stimulation is removed, and this indeed has been recently demonstrated. Loss of antiviral functions in T cells during chronic infection is in part due to expression of and signaling through the inhibitory PD-1 receptor. In HIV infection, PD-1 levels decrease upon antiretroviral treatment concomitantly with a decreased viral load to undetectable levels. Indeed, Youngblood et al. (111) reported that HIV-specific CD8+ T cells have reduced or no PD-1 expression and regained proliferative capacity when isolated from donors who have undergone successful treatment. These cells had earlier been exhausted in the chronically infected host, as indicated by a stable feature of their PD-1 epigenetic program acquired during chronic exposure to the antigen: The transcriptional regulatory region was unmethylated in the PD-1high HIV-specific CD8+ T cells during chronic infection, and remained unmethylated after the cells had lost PD-1 expression. Thus, chronic exposure to antigen and infection does induce a differentiation step in CD8+ T cells that makes them poised for transcriptional activation and PD-1 expression upon TCR-mediated stimulation, but from that point onward they manifest reversible functional tuning. We note that PD-1 may be regarded as a de-excitation factor, a marker of activation and molecular correlate of tuning, a role somewhat analogous to that played by CD5 in the tuning of thymocytes and naive T cells to spMHC stimulation. One difference is that the negative effect here is triggered by engagement of both the TCR by its pMHC ligand and PD-1 by its ligand. This role of PD-1 in tuning is confounded by the fact that PD-1 is also an acute activation and differentiation marker, transiently expressed strongly at the peak of burst-like immune responses and on terminally differentiated T cells (112). Indeed, it has been shown for example that the PD-1high phenotype is associated with accelerated in vivo CD8+ T cell turnover in SIV-infected rhesus macaques, in the generally immune activated T cell population, and especially within the SIV-specific CD8+ pool (80). The data suggest that the persistence of PD-1high SIV-specific CD8+ T cells in chronic infection is maintained in vivo by a mechanism involving high production coupled with a high disappearance rate. The role of PD-1 in tuning T cell responsiveness is not limited to chronic infections. It has recently been described in a state-of-the-art review as a rheostat of antigen-specific immunological regulation (113). PD-1-deficient mice are characterized by an autoimmune phenotype. As demonstrated in T cells from HIV-infected individuals, the exhausted PD-1high population in the chronic LCMV infection model was not functionally incompetent but contained memory cells capable of reexpansion and viral clearance after secondary infection with a nonpersisting strain of LCMV. PD-1 inhibits signal transduction after T cells have recognized their cognate antigen and started the activation process. It therefore appears to set an excitation threshold downstream of the TCR, unlike CD5 expressed by thymocytes and naive cells. Moreover, the progressive and antigen-load-dependent deepening of the hyporesponsive phenotype of chronically stimulated T cells over time observed in several models suggests that additional poorly defined signaling modules become progressively tuned in an increasing number of cells.


IY33CH22-Paul

BEYOND T CELLS B Cell Tuning Goodnow and associates (45) were among the first to observe that B and T cells acquired changes in their activation thresholds and phenotype as a result of previous antigen encounters and to 22.28

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

recognize the flexibility of these changes, referring to them as tuning. In parallel to work on T cell tuning reviewed above, investigators later demonstrated in mice that induced tolerance in naturally occurring autoreactive B cells that had escaped clonal deletion and receptor editing is a reversible signaling state (114, 115) (reviewed in 116). A distinct subset of naturally occurring autoreactive B cells with postdevelopmentally induced tolerance (PIT) has been identified in humans (117). A continuum of functional PIT persists in the mature naive B cell repertoire (118). In B cells, but not in T cells, self-antigen exposure tunes the responsiveness of BCR signaling by modifying basal Ca2+ levels (and by downmodulating the expression of surface IgM), which apparently led Zikherman et al. (118) to conclude that endogenous antigen does not tune the responsiveness of naive T cells. Kirchenbaum et al. (119) used a different approach to substantiate the hypothesis that dynamic downmodulation of surface IgM serves as a mechanism for maintaining weakly self-reactive B cells in a responsive state to foreign antigens by decreasing their avidity for self-antigen. They performed functional tests with IgMlow and IgMhigh B cells from mice and showed that they mobilized Ca2+ equivalently when the same number of IgM molecules were engaged. The IgMlow cell population had elevated levels of the Nur77 transcript, a biochemical signature of stimulation induced following BCR engagement and was enriched with autoreactive specificities.

NK Cell Tuning Starting from the notion that adaptation to the microenvironment and context is a central feature of many cell types and organisms (e.g., nerve cells adapting to background levels of light, or bacteria to chemotactic stimuli), we had argued that many features of T cells can be explained in terms of a tuning model. Like other cells, although for different reasons, T cells sense and respond to perturbations of state. The range of perturbations they can interpret is necessarily limited. By adapting their quiescent state to the (properly averaged) ambient signal strength, which is variable, and by sensing relative changes rather than absolute strength of signal, adapting cells maintain the specificity and sensitivity of their response over a wider range of stimuli compared to cells that do not adapt. As these considerations hold for NK cells as well, these cells evolved tuning mechanisms of their own. Like T cells, NK cells show a degree of unresponsiveness that can vary dynamically according to the strength and frequency of recent stimulatory signals. Dynamic tuning can explain in particular many features of NK cell education by MHC-I (120–122). Brodin et al. (120) argued that NK cell responsiveness is tuned along a continuum determined by the strength of the inhibitory input received by the individual NK cell during education. They proposed that NK cell education operates as a rheostat, tuned by a quantitative influence on individual NK cells by the MHC-I alleles present in vivo. In the physiologic setting of a complex receptor-ligand environment, NK cells continually adapt. Analogous to the excitation-de-excitation paradigm, the tuning or rheostat model proposes that the relative strength of the activating and inhibitory signals that an NK cell receives tunes up (arms) or down (disarms) the NK cell responsiveness and varies the activation threshold. The molecular mechanisms underlying these processes are under active study (e.g., 123). In NK cells, excitation and de-excitation signals are delivered to the cells by different sets of surface receptors. Cognate interactions between inhibitory receptors (iNKR) and self-MHC-I ligands regulate the activation threshold and safeguard against autoimmunity, and in parallel increase basal responsiveness in NK cells in an educational process referred to as arming or licensing (reviewed in 124). NK cells that sense self at steady state are more reactive to stimulation and changes in MHC-I expression than otherwise. This is analogous to tonic spMHC-induced sensitization of the T cell www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.29

ARI

30 January 2015

9:51

through partial TCRζ phosphorylation and clustering receptors at the APC interface, occurring presumably in parallel to elevation of the activation threshold downstream. As in the case of T cells, such sensitization allows the NK cell to better detect and respond to changes in MHC-I expression on potential target cells: Encountering a target with a reduced expression in MHC-I, which is often associated with transformed or virus infected cells, perturbs the excitation-de-excitation signaling balance in favor of the first by reducing de-excitation, resulting in NK cell activation. Activation can also be induced directly by ligands that bind to stimulatory receptors avidly enough, perturbing the balance directly by inducing strong excitation. Thus, although unlicensed NK cells are capable effectors, education via iNKR provides NK cells with enhanced activation potential and an added potential for the detection of MHC-I fluctuations and/or aberrations. Importantly, the licensing (i.e., tuning) status of an NK cell is not fixed. Rather, several studies have shown that the responsiveness of NK cells can be reset after their transfer from one MHCI environment to another. These results suggest that the acquisition and control of functional reactivity is a dynamic tuning process that results in continual NK cell education and reeducation (124, 125).


IY33CH22-Paul

Mast Cell Tuning The term tuning has been used, in congruence with our general definition, to describe how (genetic and) environmental factors (including IgE itself ) can position, or tune, mast cells within a broad spectrum of functional responsiveness. In particular, these environmental factors vary the constellation of critical negative regulators of signaling (de-excitation factors) via the highaffinity IgE receptor (FcεRI), such as SHIP or RabGEF1, tuning the cell’s resistance to stimuli of activation. Consequently, such environmental factors dynamically modify secretion of mediators by mast cells when stimulated with IgE and antigen (126).

SUMMARY AND FUTURE DIRECTIONS It was once believed by many that an understanding of immunity and tolerance could be achieved once we had characterized all or most of the relevant genetically encoded hard-wired signaling pathways that cause lymphocytes to respond to a recognized antigen or to be restrained. The limits of this reductionist approach are better recognized today. We need to deal with the fact that signals that change cell behavior are often overlapping and pleiotropic and that their integration into cells and exchange among cells, while subject to genetically imposed constraints, are flexible and dynamic processes. A quest for general rules is important, mainly because it may influence the interpretation of specific experimental results and suggest new lines of investigation. The tuning paradigm emerged from an effort to explain unexpected findings relating to apparently paradoxical manifestations of autoreactivity by lymphocytes. It has developed into efforts by several labs to better understand the dependence of immunity on self-recognition and to explore the role and limits of lymphocyte adaptation, not only in the sense of not responding, but as a means for the cells to diversify and tune their functions in a context-dependent way. The model of TATs reconciles the idea of physiological autoreactivity with the requirement of self-tolerance and, more generally, with the need to limit the destructive potential of the immune response once it has been unleashed. We have proposed that individual T cells continuously integrate antigenic and other signals and respond differently to the rates of change in the level of stimulation, translated intracellularly into perturbations (4). The organization of the immune response at the cell-population level in

22.30

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

space and time is also conductive to discriminating the magnitude of perturbations; selectivity (discrimination) arises from the dynamic tension between activation and suppression at the cellular and multicellular levels (5). In this way, the immune system responds in a characteristically explosive way to episodes of infection but not to the continuous presence of self-antigens or persistent pathogens. Tunable cellular thresholds, Tregs, and desensitized responding cells that act effectively as suppressors in spreading tolerance (1) all serve to enhance discriminatory capacities. Mechanistically, negative regulators adjust their levels to the degree of stimulation with predetermined tempos and may either keep up or not keep up with the rate of positive signaling, resulting in adaptation or activation, respectively. An offshoot of the perturbation concept is discontinuity theory (127), which states that effector immune responses are induced by an antigenic discontinuity, that is, by the sudden modification of molecular motifs with which immune cells interact. This has been inspired by recent developments in the understanding of NK cell activation strategy, which in turn appear to have been inspired by the perturbation concept (120, 121). Can a better molecular understanding of tuning states and mechanisms facilitate therapeutic interventions in autoimmunity, inflammatory diseases, cancer, and transplantation biology? A prerequisite to this is identification of the molecular signature of tuning. Tuning by definition affects the pattern of gene expression, but there are no tuning genes analogous to lineage-defining genes or cell differentiation–defining genes. One potential approach is to apply statistical methods for classification of gene- or protein-expression patterns. For example, Lyubchenko et al. (116) established a statistical index that numerically determines the level of homology between activation patterns of BCR signaling intermediates in B cells that are either tolerized or activated by autoantigen exposure, and thus quantifies the level of peripheral immune tolerance. Other future directions for investigation are also suggested by the data discussed in this review. The concept of reciprocal tuning of T cells and DCs imparts a greater role than presently assumed by most immunologists for these APCs in the resting state, beyond their crucial role in pathogen sensing and in initiating and driving immune responses. This area of DC tuning has been little explored. There are other issues: What are the epigenetic mechanisms that transform a tuned state into a differentiated state (adaptive differentiation)? Are stem cell–like memory T cells a distinctly differentiated set, or is their functional state at least partly a result of reversible imprinting by environmental/tissue factors? Will we be able to affect functional characteristics of naive and memory-phenotype T cells through manipulation of their microenvironments?

DISCLOSURE STATEMENT The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS The research from our laboratory described here was supported by the Division of Intramural Research of the National Institute of Allergy and Infectious Diseases. All animal experimentation in our laboratory was carried in accordance with a protocol approved by the National Institute of Allergy and Infectious Diseases Animal Care and Use Committee. We thank the NIH Office of AIDS Research for their support of Z.G. and of our work. We thank our many colleagues for their continued help and interest in this work.



22.31

IY33CH22-Paul

ARI

30 January 2015

9:51

LITERATURE CITED


1. Grossman Z, Paul WE. 1992. Adaptive cellular interactions in the immune system: the tunable activation threshold and the significance of subthreshold responses. PNAS 89:10365–69 2. Paul WE. 2013. Fundamental Immunology. Philadelphia: Lippincott Williams & Wilkins. 7th ed. 3. Grossman Z. 1993. Cellular tolerance as a dynamic state of the adaptable lymphocyte. Immunol. Rev. 133:45–73 4. Grossman Z, Paul WE. 2000. Self-tolerance: context dependent tuning of T cell antigen recognition. Semin. Immunol. 12:197–203 5. Grossman Z, Paul WE. 2001. Autoreactivity, dynamic tuning and selectivity. Curr. Opin. Immunol. 13:687–98 6. Sprent J, Surh CD. 2011. Normal T cell homeostasis: the conversion of naive cells into memoryphenotype cells. Nat. Immunol. 12:478–84 7. Hogquist KA, Jameson SC. 2014. The self-obsession of T cells: how TCR signaling thresholds affect fate ‘decisions’ and effector function. Nat. Immunol. 15:815–23 8. Paul WE, Milner JD, Grossman Z. 2013. Pathogen-sensing, regulatory T cells, and responsivenesstuning collectively regulate foreign- and self-antigen mediated T-cell responses. Cold Spring Harb. Symp. Quant. Biol. 78:265–76 9. Paul WE, Grossman Z. 2014. Pathogen-sensing and regulatory T cells: integrated regulators of immune responses. Cancer Immunol. Res. 2:503–9 10. Tanchot C, Lemonnier FA, Perarnau B, Freitas AA, Rocha B. 1997. Differential requirements for survival and proliferation of CD8 naive or memory T cells. Science 276:2057–62 11. Ernst B, Lee DS, Chang JM, Sprent J, Surh CD. 1999. The peptide ligands mediating positive selection in the thymus control T cell survival and homeostatic proliferation in the periphery. Immunity 11:173–81 12. Jameson SC. 2002. Maintaining the norm: T-cell homeostasis. Nat. Rev. Immunol. 2:547–56 13. Surh CD, Sprent J. 2008. Homeostasis of naive and memory T cells. Immunity 29:848–62 14. Cho JH, Kim HO, Surh CD, Sprent J. 2010. T cell receptor-dependent regulation of lipid rafts controls naive CD8+ T cell homeostasis. Immunity 32:214–26 15. Persaud SP, Parker CR, Lo WL, Weber KS, Allen PM. 2014. Intrinsic CD4+ T cell sensitivity and response to a pathogen are set and sustained by avidity for thymic and peripheral complexes of self peptide and MHC. Nat. Immunol. 15:266–74 16. Hashiguchi T, Oyamada A, Sakuraba K, Shimoda K, Nakayama KI, et al. 2014. Tyk2-dependent bystander activation of conventional and nonconventional Th1 cell subsets contributes to innate host defense against Listeria monocytogenes infection. J. Immunol. 192:4739–47 17. Younes SA, Punkosdy G, Caucheteux S, Chen T, Grossman Z, Paul WE. 2011. Memory phenotype CD4 T cells undergoing rapid, nonburst-like, cytokine-driven proliferation can be distinguished from antigen-experienced memory cells. PLOS Biol. 9:e1001171 18. Hosking MP, Flynn CT, Botten J, Whitton JL. 2013. CD8+ memory T cells appear exhausted within hours of acute virus infection. J. Immunol. 191:4211–22 19. Han S, Asoyan A, Rabenstein H, Nakano N, Obst R. 2010. Role of antigen persistence and dose for CD4+ T-cell exhaustion and recovery. PNAS 107:20453–58 20. Acuto O, Di Bartolo V, Michel F. 2008. Tailoring T-cell receptor signals by proximal negative feedback mechanisms. Nat. Rev. Immunol. 8:699–712 21. Brownlie RJ, Zamoyska R. 2013. T cell receptor signalling networks: branched, diversified and bounded. Nat. Rev. Immunol. 13:257–69 22. Goldbeter A, Koshland DE Jr. 1981. An amplified sensitivity arising from covalent modification in biological systems. PNAS 78:6840–44 23. Koshland DE. 1987. Switches, thresholds, and ultrasensitivity. Trends Biochem. Sci. 12:225–29 24. Ferrell JE Jr. 1996. Tripping the switch fantastic: how a protein kinase cascade can convert graded inputs into switch-like outputs. Trends Biochem. Sci. 21:460–66 25. Stefanova I, Hemmer B, Vergelli M, Martin R, Biddison WE, Germain RN. 2003. TCR ligand discrimination is enforced by competing ERK positive and SHP-1 negative feedback pathways. Nat. Immunol. 4:248–54 22.32

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

26. Germain RN. 2010. Computational analysis of T cell receptor signaling and ligand discrimination—past, present, and future. FEBS Lett. 584:4814–22 27. Smith-Garvin JE, Koretzky GA, Jordan MS. 2009. T cell activation. Annu. Rev. Immunol. 27:591–619 28. Feinerman O, Germain RN, Altan-Bonnet G. 2008. Quantitative challenges in understanding ligand discrimination by αβ T cells. Mol. Immunol. 45:619–31 29. Au-Yeunga BB, Zikherman J, Mueller JL, Ashouri JF, Matloubian M, et al. 2014. A sharp T-cell antigen receptor signaling threshold for T-cell proliferation. PNAS 11:E3679–88 30. Altan-Bonnet G, Germain RN. 2005. Modeling T cell antigen discrimination based on feedback control of digital ERK responses. PLOS Biol. 3:e356 31. Stefanova I, Dorfman JR, Tsukamoto M, Germain RN. 2003. On the role of self-recognition in T cell responses to foreign antigen. Immunol. Rev. 191:97–106 32. Bunnell SC, Hong DI, Kardon JR, Yamazaki T, McGlade CJ, et al. 2002. T cell receptor ligation induces the formation of dynamically regulated signaling assemblies. J. Cell Biol. 158:1263–75 33. Yokosuka T, Sakata-Sogawa K, Kobayashi W, Hiroshima M, Hashimoto-Tane A, et al. 2005. Newly generated T cell receptor microclusters initiate and sustain T cell activation by recruitment of ZAP-70 and SLP-76. Nat. Immunol. 6:1253–62 34. Campi G, Varma R, Dustin ML. 2005. Actin and agonist MHC-peptide complex-dependent T cell receptor microclusters as scaffolds for signaling. J. Exp. Med. 202:1031–36 35. Li QJ, Chau J, Ebert PJ, Sylvester G, Min H, et al. 2007. miR-181a is an intrinsic modulator of T cell sensitivity and selection. Cell 129:147–61 36. Huang J, Zarnitsyna VI, Liu B, Edwards LJ, Jiang N, et al. 2010. The kinetics of two-dimensional TCR and pMHC interactions determine T-cell responsiveness. Nature 464:932–36 37. Varma A, Huang KC, Young KD. 2008. The Min system as a general cell geometry detection mechanism: Branch lengths in Y-shaped Escherichia coli cells affect Min oscillation patterns and division dynamics. J. Bacteriol. 190:2106–17 38. Fooksman DR, Vardhana S, Vasiliver-Shamis G, Liese J, Blair DA, et al. 2010. Functional anatomy of T cell activation and synapse formation. Annu. Rev. Immunol. 28:79–105 39. Huppa JB, Davis MM. 2013. The interdisciplinary science of T-cell recognition. Adv. Immunol. 119:1–50 40. Cohen-Saidon C, Cohen AA, Sigal A, Liron Y, Alon U. 2009. Dynamics and variability of ERK2 response to EGF in individual living cells. Mol. Cell 36:885–93 41. Mueller DL. 2003. Tuning the immune system: competing positive and negative feedback loops. Nat. Immunol. 4:210–11 42. Hebeisen M, Baitsch L, Presotto D, Baumgaertner P, Romero P, et al. 2013. SHP-1 phosphatase activity counteracts increased T cell receptor affinity. J. Clin. Investig. 123:1044–56 43. Stefanova I, Dorfman JR, Germain RN. 2002. Self-recognition promotes the foreign antigen sensitivity of naive T lymphocytes. Nature 420:429–34 44. Jameson SC, Hogquist KA, Bevan MJ. 1994. Specificity and flexibility in thymic selection. Nature 369:750–52 45. Goodnow CC. 1996. Balancing immunity and tolerance: deleting and tuning lymphocyte repertoires. PNAS 93:2264–71 46. Kawai K, Ohashi PS. 1995. Immunological function of a defined T-cell population tolerized to lowaffinity self antigens. Nature 374:68–69 47. Sebzda E, Kundig TM, Thomson CT, Aoki K, Mak SY, et al. 1996. Mature T cell reactivity altered by peptide agonist that induces positive selection. J. Exp. Med. 183:1093–104 48. Mariathasan S, Bachmann MF, Bouchard D, Ohteki T, Ohashi PS. 1998. Degree of TCR internalization and Ca2+ flux correlates with thymocyte selection. J. Immunol. 161:6030–37 49. Saibil SD, Ohteki T, White FM, Luscher M, Zakarian A, et al. 2003. Weak agonist self-peptides promote selection and tuning of virus-specific T cells. Eur. J. Immunol. 33:685–96 50. Tarakhovsky A, Kanner SB, Hombach J, Ledbetter JA, Muller W, et al. 1995. A role for CD5 in TCRmediated signal transduction and thymocyte selection. Science 269:535–37 51. Azzam HS, Grinberg A, Lui K, Shen H, Shores EW, Love PE. 1998. CD5 expression is developmentally regulated by T cell receptor (TCR) signals and TCR avidity. J. Exp. Med. 188:2301–11 www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.33

ARI

30 January 2015

9:51

52. Azzam HS, DeJarnette JB, Huang K, Emmons R, Park CS, et al. 2001. Fine tuning of TCR signaling by CD5. J. Immunol. 166:5464–72 53. Smith K, Seddon B, Purbhoo MA, Zamoyska R, Fisher AG, Merkenschlager M. 2001. Sensory adaptation in naive peripheral CD4 T cells. J. Exp. Med. 194:1253–61 54. Stamou P, de Jersey J, Carmignac D, Mamalaki C, Kioussis D, Stockinger B. 2003. Chronic exposure to low levels of antigen in the periphery causes reversible functional impairment correlating with changes in CD5 levels in monoclonal CD8 T cells. J. Immunol. 171:1278–84 55. Sen G, Bikah G, Venkataraman C, Bondada S. 1999. Negative regulation of antigen receptor-mediated signaling by constitutive association of CD5 with the SHP-1 protein tyrosine phosphatase in B-1 B cells. Eur. J. Immunol. 29:3319–28 56. Perez-Villar JJ, Whitney GS, Bowen MA, Hewgill DH, Aruffo AA, Kanner SB. 1999. CD5 negatively regulates the T-cell antigen receptor signal transduction pathway: involvement of SH2-containing phosphotyrosine phosphatase SHP-1. Mol. Cell. Biol. 19:2903–12 57. Bhandoola A, Bosselut R, Yu Q, Cowan ML, Feigenbaum L, et al. 2002. CD5-mediated inhibition of TCR signaling during intrathymic selection and development does not require the CD5 extracellular domain. Eur. J. Immunol. 32:1811–17 58. Soldevila G, Raman C, Lozano F. 2011. The immunomodulatory properties of the CD5 lymphocyte receptor in health and disease. Curr. Opin. Immunol. 23:310–18 59. Bachmaier K, Krawczyk C, Kozieradzki I, Kong YY, Sasaki T, et al. 2000. Negative regulation of lymphocyte activation and autoimmunity by the molecular adaptor Cbl-b. Nature 403:211–16 60. Grossman Z, Singer A. 1996. Tuning of activation thresholds explains flexibility in the selection and development of T cells in the thymus. PNAS 93:14747–52 61. Wong P, Barton GM, Forbush KA, Rudensky AY. 2001. Dynamic tuning of T cell reactivity by selfpeptide-major histocompatibility complex ligands. J. Exp. Med. 193:1179–87 62. Lucas B, Stefanova I, Yasutomo K, Dautigny N, Germain RN. 1999. Divergent changes in the sensitivity of maturing T cells to structurally related ligands underlies formation of a useful T cell repertoire. Immunity 10:367–76 63. Marquez ME, Ellmeier W, Sanchez-Guajardo V, Freitas AA, Acuto O, Di Bartolo V. 2005. CD8 T cell sensory adaptation dependent on TCR avidity for self-antigens. J. Immunol. 175:7388–97 64. Stephen TL, Tikhonova A, Riberdy JM, Laufer TM. 2009. The activation threshold of CD4+ T cells is defined by TCR/peptide-MHC class II interactions in the thymic medulla. J. Immunol. 183:5554–62 65. Fu G, Rybakin V, Brzostek J, Paster W, Acuto O, Gascoigne NR. 2014. Fine-tuning T cell receptor signaling to control T cell development. Trends Immunol. 35:311–18 66. Johnson DJ, Pao LI, Dhanji S, Murakami K, Ohashi PS, Neel BG. 2013. Shp1 regulates T cell homeostasis by limiting IL-4 signals. J. Exp. Med. 210:1419–31 67. Mandl JN, Monteiro JP, Vrisekoop N, Germain RN. 2013. T cell-positive selection uses self-ligand binding strength to optimize repertoire recognition of foreign antigens. Immunity 38:263–74 68. Takada K, Jameson SC. 2009. Self-class I MHC molecules support survival of naive CD8 T cells, but depress their functional sensitivity through regulation of CD8 expression levels. J. Exp. Med. 206:2253–69 69. Park JH, Adoro S, Lucas PJ, Sarafova SD, Alag AS, et al. 2007. ‘Coreceptor tuning’: cytokine signals transcriptionally tailor CD8 coreceptor expression to the self-specificity of the TCR. Nat. Immunol. 8:1049–59 70. Hochweller K, Wabnitz GH, Samstag Y, Suffner J, Hammerling GJ, Garbi N. 2010. Dendritic cells control T cell tonic signaling required for responsiveness to foreign antigen. PNAS 107:5931–36 71. Garbi N, Kreutzberg T. 2012. Dendritic cells enhance the antigen sensitivity of T cells. Front. Immunol. 3:389 72. Hogquist KA, Starr TK, Jameson SC. 2003. Receptor sensitivity: when T cells lose their sense of self. Curr. Biol. 13:R239–41 73. Bhandoola A, Tai X, Eckhaus M, Auchincloss H, Mason K, et al. 2002. Peripheral expression of selfMHC-II influences the reactivity and self-tolerance of mature CD4+ T cells: evidence from a lymphopenic T cell model. Immunity 17:425–36 74. Jubala CM, Lamerato-Kozicki AR, Borakove M, Lang J, Gardner LA, et al. 2009. MHC-dependent desensitization of intrinsic anti-self reactivity. Cancer Immunol. Immunother. 58:171–85


IY33CH22-Paul

22.34

Grossman

·

Paul



IY33CH22-Paul

ARI

30 January 2015

9:51

75. Tanchot C, Barber DL, Chiodetti L, Schwartz RH. 2001. Adaptive tolerance of CD4+ T cells in vivo: multiple thresholds in response to a constant level of antigen presentation. J. Immunol. 167:2030–39 76. Singh NJ, Schwartz RH. 2003. The strength of persistent antigenic stimulation modulates adaptive tolerance in peripheral CD4+ T cells. J. Exp. Med. 198:1107–17 77. Singh NJ, Chen C, Schwartz RH. 2006. The impact of T cell intrinsic antigen adaptation on peripheral immune tolerance. PLOS Biol. 4:e340 78. Chiodetti L, Choi S, Barber DL, Schwartz RH. 2006. Adaptive tolerance and clonal anergy are distinct biochemical states. J. Immunol. 176:2279–91 79. Choi S, Schwartz RH. 2011. Impairment of immunological synapse formation in adaptively tolerant T cells. J. Immunol. 187:805–16 80. Petrovas C, Yamamoto T, Price DA, Rao SS, Klatt NR, et al. 2013. High production rates sustain in vivo levels of PD-1high simian immunodeficiency virus-specific CD8 T cells in the face of rapid clearance. J. Virol. 87:9836–44 81. Honda T, Egen JG, Lammermann T, Kastenmuller W, Torabi-Parizi P, Germain RN. 2014. Tuning of antigen sensitivity by T cell receptor-dependent negative feedback controls T cell effector function in inflamed tissues. Immunity 40:235–47 82. Chappert P, Bouladoux N, Naik S, Schwartz RH. 2013. Specific gut commensal flora locally alters T cell tuning to endogenous ligands. Immunity 38:1198–210 83. Rao RR, Li Q, Odunsi K, Shrikant PA. 2010. The mTOR kinase determines effector versus memory CD8+ T cell fate by regulating the expression of transcription factors T-bet and Eomesodermin. Immunity 32:67–78 84. Gagnon J, Chen XL, Forand-Boulerice M, Leblanc C, Raman C, et al. 2010. Increased antigen responsiveness of naive CD8 T cells exposed to IL-7 and IL-21 is associated with decreased CD5 expression. Immunol. Cell Biol. 88:451–60 85. Richer MJ, Nolz JC, Harty JT. 2013. Pathogen-specific inflammatory milieux tune the antigen sensitivity of CD8+ T cells by enhancing T cell receptor signaling. Immunity 38:140–52 86. Wilson EB, Brooks DG. 2013. Inflammation makes T cells sensitive. Immunity 38:5–7 87. Raue HP, Beadling C, Haun J, Slifka MK. 2013. Cytokine-mediated programmed proliferation of virusspecific CD8+ memory T cells. Immunity 38:131–39 88. Langenkamp A, Messi M, Lanzavecchia A, Sallusto F. 2000. Kinetics of dendritic cell activation: impact on priming of TH1, TH2 and nonpolarized T cells. Nat. Immunol. 1:311–16 89. Abdi K, Singh NJ, Matzinger P. 2012. Lipopolysaccharide-activated dendritic cells: “exhausted” or alert and waiting? J. Immunol. 188:5981–89 90. Lanzavecchia A, Sallusto F. 2000. Dynamics of T lymphocyte responses: intermediates, effectors, and memory cells. Science 290:92–97 91. Vendetti S, Chai JG, Dyson J, Simpson E, Lombardi G, Lechler R. 2000. Anergic T cells inhibit the antigen-presenting function of dendritic cells. J. Immunol. 165:1175–81 92. Germain RN. 2012. Maintaining system homeostasis: the third law of Newtonian immunology. Nat. Immunol. 13:902–6 93. Hamilton SE, Jameson SC. 2012. CD8 T cell quiescence revisited. Trends Immunol. 33:224–30 94. Modiano JF, Johnson LD, Bellgrau D. 2008. Negative regulators in homeostasis of naive peripheral T cells. Immunol. Res. 41:137–53 95. Grossman Z, Min B, Meier-Schellersheim M, Paul WE. 2004. Concomitant regulation of T-cell activation and homeostasis. Nat. Rev. Immunol. 4:387–95 96. Pasare C, Medzhitov R. 2003. Toll pathway-dependent blockade of CD4+ CD25+ T cell-mediated suppression by dendritic cells. Science 299:1033–36 97. Goldrath AW, Bevan MJ. 1999. Selecting and maintaining a diverse T-cell repertoire. Nature 402:255–62 98. Johnson PL, Goronzy JJ, Antia R. 2014. A population biological approach to understanding the maintenance and loss of the T-cell repertoire during aging. Immunology 142:167–75 99. Almeida AR, Rocha B, Freitas AA, Tanchot C. 2005. Homeostasis of T cell numbers: from thymus production to peripheral compartmentalization and the indexation of regulatory T cells. Semin. Immunol. 17:239–49 www.annualreviews.org • Dynamic Tuning of Lymphocytes


22.35

ARI

30 January 2015

9:51

100. Carrette F, Surh CD. 2012. IL-7 signaling and CD127 receptor regulation in the control of T cell homeostasis. Semin. Immunol. 24:209–17 101. Kieper WC, Burghardt JT, Surh CD. 2004. A role for TCR affinity in regulating naive T cell homeostasis. J. Immunol. 172:40–44 102. Singh NJ, Bando JK, Schwartz RH. 2012. Subsets of nonclonal neighboring CD4+ T cells specifically regulate the frequency of individual antigen-reactive T cells. Immunity 37:735–46 103. Krogsgaard M, Juang J, Davis MM. 2007. A role for “self ” in T-cell activation. Semin. Immunol. 19:236– 44 104. Garbi N, Hammerling GJ, Probst HC, van den Broek M. 2010. Tonic T cell signalling and T cell tolerance as opposite effects of self-recognition on dendritic cells. Curr. Opin. Immunol. 22:601–8 105. Kimura MY, Pobezinsky LA, Guinter TI, Thomas J, Adams A, et al. 2013. IL-7 signaling must be intermittent, not continuous, during CD8+ T cell homeostasis to promote cell survival instead of cell death. Nat. Immunol. 14:143–51 106. Weber KS, Li QJ, Persaud SP, Campbell JD, Davis MM, Allen PM. 2012. Distinct CD4+ helper T cells involved in primary and secondary responses to infection. PNAS 109:9511–16 107. Martin CE, Surh CD. 2014. Self-gratification yields not-so-naive T cells. Nat. Immunol. 15:217–19 108. Ryan KR, McCue D, Anderton SM. 2005. Fas-mediated death and sensory adaptation limit the pathogenic potential of autoreactive T cells after strong antigenic stimulation. J. Leukoc. Biol. 78:43– 50 109. Ryan KR, Patel SD, Stephens LA, Anderton SM. 2007. Death, adaptation and regulation: the three pillars of immune tolerance restrict the risk of autoimmune disease caused by molecular mimicry. J. Autoimmun. 29:262–71 110. Wherry EJ, Blattman JN, Murali-Krishna K, van der Most R, Ahmed R. 2003. Viral persistence alters CD8 T-cell immunodominance and tissue distribution and results in distinct stages of functional impairment. J. Virol. 77:4911–27 111. Youngblood B, Noto A, Porichis F, Akondy RS, Ndhlovu ZM, et al. 2013. Cutting edge: Prolonged exposure to HIV reinforces a poised epigenetic program for PD-1 expression in virus-specific CD8 T cells. J. Immunol. 191:540–44 112. Legat A, Speiser DE, Pircher H, Zehn D, Fuertes Marraco SA. 2013. Inhibitory receptor expression depends more dominantly on differentiation and activation than “exhaustion” of human CD8 T cells. Front. Immunol. 4:455 113. Okazaki T, Chikuma S, Iwai Y, Fagarasan S, Honjo T. 2013. A rheostat for immune responses: the unique properties of PD-1 and their advantages for clinical application. Nat. Immunol. 14:1212–18 114. Gauld SB, Benschop RJ, Merrell KT, Cambier JC. 2005. Maintenance of B cell anergy requires constant antigen receptor occupancy and signaling. Nat. Immunol. 6:1160–67 115. Kilmon MA, Rutan JA, Clarke SH, Vilen BJ. 2005. Low-affinity, Smith antigen-specific B cells are tolerized by dendritic cells and macrophages. J. Immunol. 175:37–41 116. Lyubchenko T, Zerbe GO. 2014. B cell receptor signaling-based index as a biomarker for the loss of peripheral immune tolerance in autoreactive B cells in rheumatoid arthritis. PLOS ONE 9:e102128 117. Duty JA, Szodoray P, Zheng NY, Koelsch KA, Zhang Q, et al. 2009. Functional anergy in a subpopulation of naive B cells from healthy humans that express autoreactive immunoglobulin receptors. J. Exp. Med. 206:139–51 118. Zikherman J, Parameswaran R, Weiss A. 2012. Endogenous antigen tunes the responsiveness of naive B cells but not T cells. Nature 489:160–64 119. Kirchenbaum GA, St Clair JB, Detanico T, Aviszus K, Wysocki LJ. 2014. Functionally responsive selfreactive B cells of low affinity express reduced levels of surface IgM. Eur. J. Immunol. 44:970–82 120. Brodin P, Kärre K, Hoglund P. 2009. NK cell education: not an on-off switch but a tunable rheostat. ¨ Trends Immunol. 30:143–49 121. Brodin P, Lakshmikanth T, Johansson S, Karre K, Hoglund P. 2009. The strength of inhibitory input during education quantitatively tunes the functional responsiveness of individual natural killer cells. Blood 113:2434–41


IY33CH22-Paul

22.36

Grossman

·

Paul


IY33CH22-Paul

ARI

30 January 2015

9:51


122. Joncker NT, Fernandez NC, Treiner E, Vivier E, Raulet DH. 2009. NK cell responsiveness is tuned commensurate with the number of inhibitory receptors for self-MHC class I: the rheostat model. J. Immunol. 182:4572–80 123. Sullivan RP, Fogel LA, Leong JW, Schneider SE, Wong R, et al. 2013. MicroRNA-155 tunes both the threshold and extent of NK cell activation via targeting of multiple signaling pathways. J. Immunol. 191:5904–13 124. Nash WT, Teoh J, Wei H, Gamache A, Brown MG. 2014. Know thyself: NK-cell inhibitory receptors prompt self-tolerance, education, and viral control. Front. Immunol. 5:175 125. Long EO, Kim HS, Liu D, Peterson ME, Rajagopalan S. 2013. Controlling natural killer cell responses: integration of signals for activation and inhibition. Annu. Rev. Immunol. 31:227–58 126. Galli SJ, Kalesnikoff J, Grimbaldeston MA, Piliponsky AM, Williams CM, Tsai M. 2005. Mast cells as “tunable” effector and immunoregulatory cells: recent advances. Annu. Rev. Immunol. 23:749–86 127. Pradeu T, Jaeger S, Vivier E. 2013. The speed of change: towards a discontinuity theory of immunity? Nat. Rev. Immunol. 13:764–69



22.37

Physiological basis of stress.

Physiological basis of ischemic enteropathy.

Dynamic surrounds of receptive fields in primate striate cortex: a physiological basis for perceptual completion?

Physiological basis for absorptive and renal hypercalciurias.

Sustained and Transient Vestibular Systems: A Physiological Basis for Interpreting Vestibular Function.

Disparity tuning in mechanisms of human stereopsis.

The physiological basis of implosive therapy.

The physiological basis of bird flight.

Biochemical Basis of Sestrin Physiological Activities.

[Physiological basis of the microcirculation: vascular adaptation].

Electrophysiologic basis of dynamic extensor splinting.

Physiological basis of visual acuity and its development in kittens.

Physiological and pharmacological basis for the chemotherapy of enuresis.

The physiological basis and clinical significance of lung volume measurements.

ClC-5: Physiological role and biophysical mechanisms.

Mechanisms of mouse spleen dendritic cell function in the generation of influenza-specific, cytolytic T lymphocytes.

Neuroanatomical basis of cochlear coding mechanisms.

Physical basis of some membrane shaping mechanisms.

Processing, signaling, and physiological function of chemerin.

The physiological function of glucagon.

Pseudoproteases: mechanisms and function.

Systematic tuning of the conduction mechanisms in ferroelectric thin films.

A physiological basis for the treatment of essential hypertension.

The physiological basis of therapies for cerebellar ataxias.