|Home | About | Journals | Submit | Contact Us | Français|
Memory T cells exhibit low activation thresholds and rapid effector responses following antigen stimulation, contrasting naive T cells with high activation thresholds and no effector responses. Signaling mechanisms for the distinct properties of naive and memory T cells remain poorly understood. Here, I will discuss new results on signal transduction in naive and memory T cells that suggest proximal control of activation threshold and a distinct biochemical pathway to rapid recall. The signaling and transcriptional pathways controlling immediate effector function in memory T cells closely resemble pathways for rapid effector cytokine production in innate immune cells, suggesting memory T cells use innate pathways for efficacious responses.
The development of memory T cells following antigen activation of naive T cells involves profound cellular and molecular changes. The memory differentiation process begins when naive CD4 T cells become activated by coordinate engagement of the T-cell receptor (TCR) and co-stimulatory receptor CD28 on the T cell surface by MHC-peptide antigen complexes and co-stimulatory ligands (B7-1/B7-2), respectively, on the surface of antigen-presenting cells (APC). Through integration of these signals, naive T cells are stimulated to proliferate and differentiate into different types of effector T cells producing effector cytokines. Most of these activated effector T cells die after a brief lifespan in vivo, yet a subset of primed, antigen-specific T cells persists as memory T cells via mechanisms that are not clearly defined. The resultant memory T cells can be maintained up to the lifetime of an individual via perpetual turnover and homeostasis, and when reactivated, coordinate the faster, stronger, and more prolonged memory immune response [1, 2]. The ability of memory T cells to mediate efficacious responses is the cornerstone of the anamnestic immune response, although the pathways and mechanisms for the enhanced responses of memory T cells remain undefined.
Memory T cells acquire distinct phenotypic and functional properties that enable them to mediate enhanced secondary responses. Phenotypically, both naive and memory T cells are small resting cells with low level expression of the IL-2Rα chain (CD25) and expression of the IL-7Rα chain (CD127) [3, 4]. However, memory T cells express elevated levels of the adhesion markers CD44  and CD11a , compared to naive T cells, which may facilitate their interactions with APC. Functionally, memory T cells exhibit rapid production of multiple effector molecules such as IFN-γ and TNF-α  or IL-4 and IL-5 within hours of stimulation whereas naive T cells require days of sustained activation to differentiate into effector cytokine producers [8, 9]. The immediate response kinetics of memory T cells is similar to the cytokine response of innate immune cells, such as NK cells. In the case of memory CD8 T cells, increased expression of the cytotoxic molecules perforin and granzyme B also occurs at rapid times [10, 11]. In addition, memory T cells have less stringent activation requirements compared to naive T cells, including a reduced activation threshold for low antigen doses , and an ability to be fully activated by diverse APC such as resting B cells, macrophages and endothelial cells [13, 14] and DC  which are the primary APC for naive T cell activation. This rapid recall response and reduced activation threshold are the defining functional attributes of memory CD4 and CD8 T cells, although the cell-intrinsic mechanisms controlling these processes are undefined.
The study of TCR-coupled intracellular signaling pathways has revealed important insights into biochemical mechanisms controlling T cell activation, regulation and anergy. Analysis of signal transduction pathways in memory T cells has posed challenges due to the difficulty of obtaining sufficient numbers of memory T cells for biochemical analysis, and the fact that genetic deletion of key TCR-coupled signaling intermediates in vivo tends to disrupt T cell development and preclude any analysis of differentiated T cell populations. However, recent advances in single-cell analysis of signaling pathways and new results on signaling in diverse systems have provided new perspectives on biochemical control of memory T cell recall. Here, I will discuss new findings on memory T cell signal transduction that suggest proximal signaling control of activation threshold, as well as the acquisition of a separate biochemical pathway to rapid recall. In addition I will discuss aspects of memory T cell function, signaling and activation properties that bear striking similarities to innate cells such as NK and NKT cells, and suggest a common pathway to rapid cytokine responses in memory and innate immune cells. The embodiment of innate and adaptive properties in memory T cells through signaling pathways has important implications for understanding the evolution and regulation of immunological memory.
The kinetics and magnitude of cytokine production by memory T cells following antigen stimulation could be due to enhanced interactions with APC, altered intracellular signaling, or a combination of both. There is now substantial evidence that rapid memory responses are due to cell intrinsic properties of memory T cells, independent of TCR specificity or APC-mediated effects. Notably, naive and memory CD4 or CD8 T cells derived from TCR-transgenic mice bearing the identical TCR exhibit the same functional and kinetics differences as polyclonal counterparts-- with TCR-transgenic memory T cells producing effector cytokines within 3-6 hours of TCR engagement by antigen, and naive T cells not producing measurable cytokines at these early timepoints [11, 16]. It has also been suggested that the enhanced expression of adhesion markers and integrins on the surface of memory T cells may facilitate contact with APC and lead to more efficient activation [17, 18]. However, rapid recall of memory T cells can result from TCR stimulation in the absence of APC, using anti-CD3 crosslinking, or PMA/ionomycin [16, 19]. Moreover, in vivo image analysis using confocal laser scanning microscopy revealed no differences in the ability of naive or memory T cells to form conjugates with DC, and these contacts did not depend on expression of the integrin LFA-1 . When taken together, the functional and imaging results indicate that memory T cells possess an intrinsic TCR-coupled pathway to rapid recall. Thus, distinct memory cell functions could derive from altered intracellular signaling coupled to the TCR, altered mobilization of transcription factors in the nucleus, epigenetic changes in the loci of cytokine genes, and/or a combination of all three factors, as discussed below.
The procession of TCR-coupled signaling events leading to IL-2 gene transcription in primary naive T cells and T cell lines has been elucidated (see  for a detailed review on the subject), revealing crucial processes necessary for transducing TCR engagement to nuclear events. TCR-coupled signal transduction can be divided into proximal, linker-adapter and distal processes. Events proximal to TCR engagement begin with phosphorylation of TCR-coupled CD3 subunits, CD3ζ and CD3ε, by the Src-family tyrosine kinases, Lck and Fyn , leading to the mobilization, phosphorylation and subsequent activation of the T cell-specific tyrosine kinase ZAP-70 (zeta-associated protein of 70kDa) [23, 24] . The ZAP-70 kinase subsequently phosphorylates the key linker/adapter molecules SLP-76 (SH2-containing protein of 76kDa)  and LAT (linker for activated T-cells) , which molecularly couple early phosphorylation signals to downstream signaling events [28, 29]. Phosphorylated linker-adapter molecules form molecular complexes that trigger activation of the ras pathway through interactions with vav and promote release of calcium from the intracellular stores by interactions with PLCγ1 [30, 31]. Subsequent distal signaling events ensue, including protein kinase C activation, and activation of the MAP kinases Erk1/2 and p38 via the GDP-binding proteins (for reviews, see [32, 33]). Together, activated MAP-kinases and Calcium flux promote nuclear translocation and mobilization of the nuclear factor for activated T cells (NFAT) for IL-2 gene transcription [34, 35]. TCR-signaling also leads to activation of the transcription factor NFκB, which is required for transcription of multiple genes involved in T cell activation, including CD25 and IL-2 [36-38].
Whether a similar orderly procession of signaling events as described above occurs in memory T cells has not been fully elucidated. Several laboratories, including my own, have applied conventional biochemical analysis to determine whether certain critical TCR-coupled signaling events occur in memory T cells. Analysis of TCR-coupled phosphorylation of signaling intermediates in phenotypic subsets of naive and memory T cells revealed decreased phosphorylation of proximal and distal molecules in mouse and human memory compared to naive CD4 T cells [39-43]. For CD8 T cells, there are no measurable phosphorylation differences between naive and memory subsets . Interestingly, mouse memory CD4 T cells also exhibited decreased signaling at the linker/adapter level, with decreased phosphorylation of SLP-76 and association to downstream molecules . These findings of reduced TCR-coupled phosphorylation and linker/adapter signaling in memory CD4 T cells were paradoxical considering their robust and rapid functional responses. Based on these results, we hypothesized that for memory CD4 T cells, fewer signaling events transduced from TCR engagement to functional output may represent a signaling “short-circuit” and a mechanism to promote rapid recall by memory T cells . However, biochemical analysis of signaling pathways in lysates of naive and memory T cells did not indicate how specific signaling events were coupled to functional output, and new approaches were needed.
In recent years, the availability of antibodies that recognize signaling intermediates in the native form, has enabled visualization of signaling intermediates or their phosphorylated forms on the single cell level [45-47]. We applied this flow cytometry technique to analyze the expression and phosphorylation of signaling intermediates in resting and antigen-activated naive and memory CD4 T cells, and also used this approach for coordinate analysis of signaling events in conjunction with cytokine production. This combined method enabled direct assessment of the functional coupling of specific signaling events and whether effector cytokine production was associated with, or preceded by, a particular signaling event.
Single cell analysis of signaling intermediates enabled a thorough evaluation of the resting state signaling configuration in resting memory T cells. One hypothesis to account for the ability of memory T cells to signal for rapid recall is that signaling molecules are in a response-ready configuration in resting memory T cells. This potential could be achieved by the constitutive phosphorylation of key signaling intermediates, and/or pre-formed associations of specific signaling complexes. Biochemical analysis of lysates of resting naive and memory T cells indicated comparable low phosphorylation levels in both subsets [41, 43, 48, 49]. Similarly, analysis of signaling intermediate expression by intracellular flow cytometry, confirmed low basal phosphorylation of key intermediates including ZAP-70, PLC-γ, Erk1/2, and p38 . These results indicate that the ability of memory T cell to mediate rapid responses is not due to the presence of constitutively phosphorylated signaling intermediates. Another way in which signaling molecules could be poised for efficient biochemical responses could be via their association to membrane microdomains, called lipid rafts, which are cholesterol-rich membrane regions where signaling intermediates congregate [50, 51]. Indeed, an increase in the association of signaling intermediates with membrane raft microdomains has been identified in antigen-specific memory CD4 and CD8 T cells [44, 52]. Whether these observed differences in raft-associated signaling molecules had a functional consequence has not been demonstrated, and the role of lipid rafts in controlling T cell activation remains controversial .
Although we did not detect differences in basal tyrosine phosphroylation in naive and memory T cells, we found that protein expression of certain signaling intermediates was remarkably different between these subsets. In particular, protein expression of the proximal ZAP-70 kinase was consistently 3-5-fold higher in antigen-specific memory compared to naive CD4 T cells, while expression of other signaling molecules such as CD3ε and Erk was comparable ( and Fig. 1A). This augmentation in protein expression of a signaling intermediate in memory compared to naive CD4 T cells was apparent in antigen-specific and polyclonal populations of mouse memory CD4 T cells [16, 54], and in human memory versus naive CD4 T cells (Fig. 1B). Interestingly, we found that new naive CD4 T cells from umbilical cord blood expressed the lowest level of ZAP-70; CD45RA+ CD4 T cells from adult peripheral blood which contain naive and non-naIve cells  expressed higher levels of ZAP-70, and adult CD45RO+ memory CD4 T cells expressed the highest level of ZAP-70 (Fig. 1B). These results suggest developmentally-dependent increases in ZAP-70 expression in peripheral CD4 T cells.
In addition to resting state differences in ZAP-70 expression in CD4 T cell subsets, we also found that ZAP-70 expression increased progressively during antigen-stimulation and effector cell differentiation, and was comparably high in effector and memory CD4 T cells (Fig. 1C and ). Moreover, ZAP-70 co-localized to CD3ε in effector and memory CD4 T cells (Fig. 1C), suggesting that an association of ZAP-70 with the TCR/CD3 complex marked cells with effector function. Consistent with this idea, we found that IFN-γ production from differentiated effector cells occurred only from the ZAP-70hi population, and conversely, downmodulation of ZAP-70 expression by siRNA in memory CD4 T cells reduced rapid IFN-γ production . These results suggested that a high basal level of ZAP-70 kinase expression and its association to CD3ε in memory CD4 T cells facilitated propagation of downstream signals leading to rapid recall.
Our findings of elevated ZAP-70 expression in memory compared to naive CD4 T cells and its requirements for effector function by memory CD4 T cells, suggested a new mechanism for controlling T cell functional capacity by varying protein expression of a signaling intermediate. A recent study also reported a link between protein expression levels of signaling molecules and T cell functional capacity using a combination of computer modeling and single cell experimental analysis . In the case of memory T cells, increased ZAP-70 itself is probably not sufficient for propagating downstream signals leading to rapid effector function, for two reasons. First, our preliminary experiments in human T cells indicate that overexpression of ZAP-70 in newly activated naive T cells was not sufficient to promote rapid IFN-γ production (F.O. Okoye and D.L. Farber, unpublished data). Second, the increased level of ZAP-70 protein does not appear to promote enhanced levels of downstream signaling ([16, 41] and see below), suggesting that either a different pathway is operative or additional downstream alterations are necessary to transduce rapid recall function in memory T cells. We propose that increased expression of ZAP-70 and its association to the TCR/CD3 complex may act to lower the activation threshold in memory T cells  and promote functional responses with reduced TCR triggering. In this way, increased expression of a proximal signaling initiator provides a biochemical basis for enhanced memory responses at the level of TCR engagement with its ligand.
The coupling of specific signaling events to functional output in naïve versus memory T cells has been dissected using several approaches examining distal signaling events, costimulatory requirements and transcription factor utilization. From these multiple approaches, a view is emerging of TCR-coupled signaling in memory T cells that suggests two distinct pathways controlling recall responses. The different results that support this model are discussed below.
To trace the signaling events associated with rapid recall in memory T cells from the proximal ZAP-70 kinase, we simultaneously analyzed the phosphorylation of downstream mediators in conjunction with IFN-γ production in naive versus memory CD4 T cells following antigen activation over a broad kinetic window. Directly downstream of ZAP-70, we found that SLP-76 signaling was required for memory CD4 T cell recall (manuscript submitted). In addition, we found that naive and memory CD4 T cells exhibited similar kinetics for phosphorylation of signaling intermediates distal of ZAP-70 and SLP-76 including PLC-γ, p38, and p42. While resting and short-term-stimulated naive and memory T cells did not exhibit distal phosphorylation, we found that prolonged antigen activation for >24hrs resulted in the accumulation of phospho-PLC-γ, -p38, and -p42 in both subsets, enabling us to analyze whether IFN-γ production was associated with accumulated downstream phosphorylation. In naive CD4 T cells, IFN-γ production only occurred only after 24-48hrs of antigen stimulation from a population with extensive distal phosphorylation (p-PLC-γhi, p-p42hi. p-p38hi) (Fig. 2 and ). By contrast, IFN-γ production from memory T cells at early and later times post-antigen stimulation occurred from cells with different signaling signatures; after 6hrs of stimulation, IFN-γ producing T cells exhibited low levels of distal phosphorylation; and at later times (>24hr), IFN-γ production occurred from cells with high levels of distal phosphorylation (Fig. 2). This multiparameter analysis provided a key insight – that early memory recall may be triggered by a signaling pathway distinct from that triggering late effector function in either naive and memory CD4 T cells.
The existence of two pathways for memory recall is further supported by their differential requirements for CD28 signaling. Signaling through the CD28 costimulatory receptor was initially found to be dispensable for reactivation of memory T cells based on in vitro studies and results using CD28-deficient mice. These earlier studies found that memory CD4 T cells could be activated to proliferate and produce multiple cytokines in vitro, by antigen presented by B7-deficient APC [13, 57]. Similarly, LCMV infection of CD28-deficient mice led to wildtype levels of memory CD8 T cells with intact recall function . The notion of memory T cells being independent of the CD28 pathway was further consistent with their lower activation threshold through the TCR compared to naive T cells. However, further examination of memory T cell activation in vivo using specific inhibitors of the CD28 pathway and CD28-deficient mice revealed that wildtype memory CD4 and CD8 T cells do require CD28 costimulation for optimal IL-2 production and proliferative expansion in vivo [58-61]. In our studies, we found that inhibiting CD28 costimulation using the fusion protein CTLA4Ig which binds the B7 ligands, impaired IL-2 production and in vivo proliferation of antigen-stimulated memory CD4 T cells, but did not appreciably affect rapid IFN-γ production . These findings indicated that CD28 costimulation was required for certain memory T cell functions, but not others.
We investigated whether the differential use of CD28 costimulation for IL-2 versus IFN-γ production could be attributed to different subsets of memory CD4 T cells, such as central and effector memory subsets that differ in homing receptor expression and can exhibit functional differences [19, 62]. However, fractionated central (CD62Lhi) and effector (CD62Llo) memory CD4 T cells from mice both produced IFN-γ and IL-2 in response to antigenic stimulation, and IL-2 production from both memory subsets was inhibited by interfering with CD28 costimulation . These findings therefore suggest two pathways operative in memory CD4 T cells: an early pathway to effector cytokine production that does not require CD28 costimulation, and a later pathway leading to proliferation and IL-2 production that is CD28 dependent.
The different signaling pathways in memory T cells are marked by distinct kinetics, distal signaling and CD28 requirements; however, it is not known whether these pathways culminate in specific transcriptional activities in the nucleus. Thus far, there is some evidence for differential activation of transcription factors in naive and memory CD4 T cells. Polyclonal mouse memory CD4 T cells exhibited increased accumulation of NFAT following 36hrs of stimulation ; although the role of NFAT in rapid cytokine production was not examined. Expression of the transcription factor T-bet is essential for IFN-γ production by CD4 T cells and differentiation of the T-helper 1 (Th1) effector lineage ; however, IFN-γ production by CD8 T cells can be mediated either by T-bet or by the transcription factor Eomesodermin (Eomes) . Both T-bet and eomesodermin have been shown to regulate memory CD8 T cell survival [65, 66], but it is not known whether they are required for rapid recall of memory CD8 or CD4 T cells. The differential engagement of transcription factors on the promoters of cytokine genes in antigen stimulated memory compared to naive CD4 T cells has not been addressed.
We have examined T-bet expression and upregulation in antigen-specific memory compared to naive CD4 T cells, based on the hypothesis that enhanced transcriptional activation of the IFN-γ gene may be driven by elevated or constitutive expression of T-bet. By flow cytometry, we observed that T-bet expression was comparably low in resting naive and memory CD4 T cells, and was upregulated after antigen stimulation, with peak expression after 24hrs stimulation of memory CD4 T cells, and 48hrs stimulation of naïve CD4 T cells (W. Lai and D.L. Farber, unpulished data). However, we identified a differential association of IFN-γ production with T-bet expression in naive and memory CD4 T cells. In naive CD4 T cells, IFN-γ production occurred following sustained (24-48hrs) antigen activation only from the T-bet+ population, consistent with a T-bet requirement for IFN-γ production by primary CD4 T cells . By contrast, in memory CD4 T cells, early (≤6hrs) IFN-γ production after antigen stimulation occurred from a T-betlo population that did not exhibit substantial T-bet upregulation, whereas late IFN-γ production occurred from a T-bet+ populations (W. Lai and D.L. Farber, unpublished results). These results are consistent with the existence of two distinct signaling pathways to IFN-γ production in memory CD4 T cells—a pathway for early IFN-γ production that is T-bet independent, and a pathway to late IFN-γ production that requires T-bet, similar to naive T cells.
A putative T-bet-independent pathway to IFN-γ production in memory CD4 T cells us certainly consistent with the known effects of T-bet in inducing stable epigenetic changes in the IFN-γ promoter during activation of naive T cells [67, 68]. Reiner and colleagues showed that interaction of T-bet with a nuclear protein Hlx, during activation of naive T cells induced modifications in the IFN-γ locus resulting in increased promoter accessibility that were maintained during the effector stage [67, 68]. These results suggested that T-bet was important for the induction but not maintenance of the capacity to produce IFN-γ . For memory CD8 T cells, epigenetic changes involving increases in acetylated histones in the promoter regions of perforin, Eomes, and other effector molecule connote regions of active transcription, suggesting increased accessibility of promoter regions to transcription factor engagement in memory T cells [10, 70]. These structural genetic changes may further obviate the need for certain transcription factors. A similar analysis of chromatin structure of memory CD4 T cells has not yet been reported.
If T-bet expression is not associated with early memory T cell IFN-γ production, than what is the critical transcription factor important for this rapid effector production? A likely candidate is NFκB, as NFκB has been shown to control IFN-γ transcription in Th1 effector populations [71, 72], and controls many other genes important for T cell activation, including IL-2 and CD25, which are also more rapidly upregulated following TCR stimulation of memory T cells [48, 73]. Activation of NFκB requires phosphorylation and inactivation of the NFκB inhibitor, IκB, by the IκB kinase. This pathway for TCR-mediated activation of NFκB has been recently elucidated and involves the coupling of specific downstream mediators with PKCθ (for reviews, see [37, 74]). Specifically, phosphorylation of Carma1 and mobilization of the Carma1/Bcl10/MALT1 (CBM) complex leads to activation of the IκB kinase (IKK) for phosphorylation of IκB, and enables nuclear translocation of NFκB, [36-38]. It is not known how the TCR-coupled signaling axis to NFκB activation is oriented in memory T cells, and this pathway will be important to investigate. Another possibility is that early engagement of NFκB or T-bet on the IFN-γ promoter may be driving the rapid recall and early and late pathways. Determining the temporal association of specific transcription factors on cytokine gene promoters following antigen activation will be crucial for dissecting mechanisms for rapid recall.
The evidence discussed above implicates at least two distinct signaling pathways, designated “early” and “late” in memory CD4 T cells that are coupled to the production of the effector cytokine, IFN-γ (Fig. 3). Both of these pathways share certain properties in that they are marked by increased ZAP-70 expression and both require ZAP-70 and SLP-76 signaling (Fig. 3). Following this common proximal step, the two pathways appear to diverge. The late pathway for IFN-γ production in memory T cells is similar to that in naive T cells and arises following sustained TCR stimulation, leading to augmented distal phosphorylation and upregulation of T-bet expression which then translocates to the nucleus to initiate IFN-γ transcription (Fig. 3, right pathway). This late pathway for effector cytokine production requires CD28 signaling, and is coincident with IL-2 production and cell cycle entry. By contrast, the rapid pathway to early effector cytokine production is triggered within hours of TCR stimulation in memory but not naive T cells, and is not associated with downstream phosphorylation, T-bet upregulation or CD28 signaling (Fig. 3, left pathway indicated by a dotted line). These two pathways could potentially be operative within the same cell, as individual memory T cells can be multifunctional and produce multiple cytokines and/or undergo proliferative turnover [7, 76].
What is the nature of this rapid pathway for memory recall? We propose that the pathway coupled to TCR-mediated rapid effector function resembles an innate-like signaling pathway similar to NK cells and “innate-like” T cells, such as NKT cells, in terms of functional kinetics, activation requirements, signaling and transcriptional regulation. Importantly, memory T cells exhibit both the kinetics of effector cytokine production and the activation requirements of innate-type cells. Kinetically, IFN-γ production by antigen-specific memory CD4 and CD8 T cells occurs within hours of antigen stimulation [11, 16, 77], similar to the kinetics of IFN-γ production by NK cells in response to NK-activating receptors , and by innate-type NKT cells  in response to engagement of their invariant TCR by α–Gal-cer . In addition, IFN-γ production by memory T cells can also occur via antigen-independent stimulation, similar to innate-type cells. For example, it has been shown by several groups that the cytokines IL-12 and IL-18 can stimulate memory CD8 T cells to produce rapid IFN-γ [81, 82]. Recently, memory-phenotype CD4 T cells were also found to share this innate function, producing IFN-γ in response to IL-12 and IL-18 stimulation . This ability of memory T cells to produce recall cytokines in response to inflammatory cytokines is shared by NKT cells, which can be stimulated to produce IFN-γ directly by exogenous IL-12 itself or as produced by TLR-stimulated dendritic cells (DC) [84, 85]. When taken together, these findings suggest that memory T cells can be triggered early in an immune response by engagement of the TCR or cytokine receptors.
While the kinetic and functional data suggest innate-like functions to memory T cells, it is not known whether the signaling requirements and pathways for rapid IFN-γ production in memory T cells are similar to those in innate type cells. However, evidence from multiple studies suggest similar transcriptional profiles for rapid IFN-γ production in memory T cells and NK cells. Wilson and colleagues demonstrated the lack of T-bet expression did not impair IFN-γ production by NK cells in vivo following Listeria infection , consistent with our findings of lack of T-bet upregulation accompanying early recall of memory CD4 T cells. In addition, there is increasing evidence that NFκB is integrally involved in signaling for IFN-γ production in NK cells, similar to its role in controlling IFN-γ production in previously activated effector CD4 T cells [71, 72]. In NK cells, early IFN-γ requires NFκB activation  and inhibition of NFκB in vivo impairs IFN-γ production by both NK cell and T cells . Importantly, two groups recently demonstrated that NK cells, like T cells, require the Carma1/Bcl10/MALT1 complex to signal for IKK activation and subsequent activation of NFκB [89, 90]. These results demonstrate that signaling intermediates upstream of NFκB activation are common to both NK and T cells [89, 90], and could be a shared mechanisms for rapid recall.
There is also evidence that non-cognate stimulation of memory T cells by IL-12 stimulation can mobilize NFκB activity. IL-12 has been shown to promote nuclear translocation of NFκB in human tumor-infiltrating memory CD4 T cells , and can promote IL-2 production and CD25 upregulation in CD8 T cells  which are both targets for NFκB transcriptional activity. Given its central role as a transcriptional activator in both innate and adaptive pathways, NFκB is a common link between early IFN-γ responses in memory T cells and innate cells. A thorough examination of NFκB transcriptional activity and its engagement on the IFN-γ promoter in memory T cells, NKT cells and NK cells can establish whether NFκB is the common denominator for rapid cytokine signaling in all of these cells types.
While NFκB may represent the signal transducer common to rapid function in memory and innate cells, the Itk tyrosine kinase appears to be the signaling intermediate that distinguishes between innate (or early) and adaptive (or late) type signaling. Itk is an IL-2 inducible Tec-family tyrosine kinase that regulates TCR signaling . Several groups showed that mice deficient in the Itk kinase develop peripheral CD8 and CD4 T cells with a predominant memory (CD44hi) phenotype [94-96]. CD8+CD44hi cells from Itk-deficient mice have been the most extensively studied from these mice, and were found to possess innate function, with rapid production of IFN-γ as a result of TCR engagement or IL-12 stimulation [94-96]. These cells have been referred to as “innate T cells” , although their expression of a conventional αβ TCR, memory phenotype and function suggest that lack of Itk during development triggers T cells to convert directly to memory T cells, rather than generating naive T cell signaling configurations. Thus, the biochemical coupling of signaling pathways through the TCR in memory T cells may derive from the more primordial innate signaling, and that later in evolution, the TCR signaling configuration in naive T cells developed to uncouple rapid and innate function from TCR engagement.
In addition to similar kinetics, signaling and activation requirements for rapid IFN-γ production by memory T cells and innate cells, both memory T and NK cells also exhibit commonalities in the epigenetic changes in IFN-γ locus, that may also be important for promoting rapid cytokine responses. Methylation of histones is known to regulate promoter accessibility, with an “open” chromatin configuration characterized by hypomethylation at a specific promoter locus (for reviews, see ). Epigenetic modification defined by hypomethylation have been identified in the IFN-γ locus in polyclonal CD44hi memory CD8 T cells [99, 100], and also in Th1 effector cells , and have been shown to regulate the efficiency of IFN-γ expression on the level of gene accessibility . Interestingly, in NK cells, the IFN-γ promoter was found to be constitutively hypomethylated at regions similar to those found in Th1 cells [102, 103]. These results, together with commonalities in signaling and activation pathways, suggest that a combination of intracellular signaling and promoter accessibility contribute to the rapidity of IFN-γ production following receptor engagement in both memory T cells and innate NK cells. Memory T cells must undergo prior activation with antigen to acquire the functional, signaling and epigenetic changes already present within NK cells, further suggesting that memory T cells use a primordial pathway for their recall responses.
The analysis of intracellular signaling and its functional coupling in memory T cells has provided new insights into the mechanisms for rapid recall that endows memory T cells with their ability to respond efficaciously and coordinate a rapid anamnestic response. The emerging knowledge of the participants and intermediates in this memory recall pathway bear striking similarities to pathways in innate immune cells, including NK cells and NKT cells. Such similarities suggest a common innate pathway to rapid recall function in memory T cells that has important implications for the development and evolution of immune memory. Dissecting the intricacies of this pathway in future studies will reveal crucial insights into how the immune system uncouples rapid innate from delayed function, for proper immune regulation and homeostasis.
This work was supported by NIH AI42092 awarded to D.L.F. The author wishes to thank Martin Flajnik for critical reading of this manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.