Stem Cells, Vol. 14, No. 5, 490-500,
September 1996
© 1996 AlphaMed Press
Cross-Talk between the T Cell Antigen Receptor and the Glucocorticoid Receptor Regulates Thymocyte Development
Jonathan D. Ashwella,
Leslie B. Kingb,
Melanie S. Vacchiob
a Laboratory of Immune Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA;
b Laboratory of Immunology, Division of Hematologic Products, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, Maryland, USA
Key Words. Steroids • Thymocyte development • Thymocyte selection • T cell receptor • Glucocorticoids • Glucocorticoid receptor • Apoptosis
Correspondence: Dr. Jonathan D. Ashwell, Room 1B-40, Building 10, National Institutes of Health, Bethesda, MD 20852, USA.
 |
Abstract
|
|---|
The fate of an immature thymocyte, life or death, is largely determined by the ligand-specificity of its T cell antigen receptor (TCR). The default pathway for thymocytes bearing TCRs with subthreshold avidity for self-antigens is death (death by neglect). Thymocytes bearing TCRs with high avidity for self also undergo apoptosis (negative selection). Those thymocytes with intermediate avidities, or that perhaps recognize self-peptides that have partial agonist or antagonist properties, survive and differentiate into mature immunocompetent T cells (positive selection). How TCR avidity is interpreted as a "rescue" signal or a death signal is unknown. Based upon a T cell hybridoma model, our laboratory proposed that glucocorticoids, which themselves are potent inducers of thymocyte apoptosis, antagonize TCR-mediated thymocyte deletion and allow positive selection to occur. In fact, epithelial cells in the thymus proved to be a source of steroid production, and interference with steroid synthesis in fetal thymic organ culture resulted in a greatly enhanced sensitivity of thymocytes to TCR-mediated apoptosis. Transgenic mice with reduced glucocorticoid receptor (GR) levels were produced by tissue-specific expression of GR antisense. Thymocytes in these mice had high levels of spontaneous apoptosis, and were exquisitely sensitive to deletion induced by cross-linking the TCR. Moreover, there was a very large (
90%) loss of CD4+CD8+ thymocytes, signifying a block at the CD4–CD8– to CD4+CD8+ transition, perhaps due to apoptosis of cells upon engagement of the pre-TCR in the absence of an antagonizing glucocorticoid stimulus. The molecular mechanism of the antagonism is currently being investigated. These data indicate that there is cross-talk in thymocytes between the TCR and glucocorticoid signaling pathways resulting in apoptosis, and that locally produced steroids, in a paracrine fashion, participate in setting the TCR avidity thresholds that determine whether developing thymocytes survive or die, and therefore help to mold the antigen-specific T cell repertoire.
|
Thymocyte Development
|
|---|
The fate of the vast majority of thymocytes is death. The teleological reason for this is simple: mature T cells (differentiated cells that have emigrated from the thymus) are only of use if they bear antigen-specific receptors, T cell antigen receptors (TCRs), that can recognize antigen bound to major histocompatibility complex (MHC)-encoded molecules. Since thymocytes cannot "know" what combination of MHC molecules and antigens they might encounter prior to expression of an antigen-specific TCR, a developmental strategy evolved in which TCRs with an enormous number of possible antigen-specificities are randomly produced by rearrangement of germline TCR-
and -ß loci (the theoretical limit for the number of unique ß combinations, based upon primary amino acid diversity, has been estimated to be ~1015 [1]). Thymocytes that fail to express a TCR (for example, because of nonproductive rearrangement of germline or ß loci) or that express TCRs with little or no avidity for self-antigen/MHC die in the thymus. The generation of an enormous number of cells to obtain the (relatively) small number required to maintain an immune response is one of the fundamental features of T cell development. Although the exact percentage of cells that fall into this category is not known, it is thought to comprise the large majority of thymocytes (>90%) [2]. Calculations based upon the steady-state levels of thymocyte subpopulations and the rate of thymocyte generation have led to the estimate that "neglected" thymocytes live approximately 3.5 days [3, 4]. In mice that express no classic MHC-encoded molecules, and that therefore have no classic ligands for ß TCR-bearing cells, dying thymocytes can be identified in situ by the presence of "nicked" or cleaved DNA and are found largely as phagocytosed bodies in intrathymic macrophages [5]. While it is not clear why neglected thymocytes are fated to undergo apoptotic death, the observation (first made in 1924 [6]) that adrenalectomy results in a rapid (within five days [7]) increase in thymocyte number has been taken to mean that circulating steroids, most likely glucocorticoids, are at least in part responsible. It is unlikely, however, that glucocorticoid-induced death is the sole mechanism for limiting the viability of "useless thymocytes," since thymocytes from Bcl-2-overexpressing transgenic (Tg) mice are resistant to glucocorticoid-induced apoptosis, but the thymus is not hypertrophic and there is no increase in the number of CD4+CD8+ (steroid-sensitive) cells [8, 9].
During thymocyte maturation, a small proportion of thymocytes receives a signal that rescues them from the default death pathway. Although some minimal TCR avidity for self-antigen/MHC appears to be required to avoid death by neglect, other variables participate in deciding the ultimate fate of the cell. One commonly held hypothesis [10] is that thymocytes bearing TCRs recognizing self-antigen/MHC with so-called "low-to-moderate" avidity survive and differentiate, a process known as positive selection. Potentially autoreactive thymocytes bearing TCRs with high avidity for self-antigen/MHC are eliminated (die) in the thymus, a process known as clonal deletion or negative selection. An alternative model is that while TCR occupancy by all peptide/MHC ligands can cause negative selection of thymocytes, positive selection occurs only when TCRs encounter a peptide with partial agonist or antagonist properties, as defined by their effect on activation of antigen-specific peripheral T cells [11]. In this case, a (presumably) qualitatively unique signal generated by the TCR results in rescue from death. The nature of this putative signal is unknown, but recent studies with "altered peptides" (peptides that differ from purely agonistic peptides by limited amino acid substitutions) with partial agonistic or antagonistic properties have suggested that engagement with these ligands may result in phosphorylation of the TCR 
chain without activation of the ZAP-70 tyrosine kinase [12, 13]. However, since partial agonists/antagonists induce negative selection when efficiently presented or at high concentrations [14], the avidity of the TCR-ligand interaction must still determine the thymocyte's fate in this model. This review describes one mechanism by which the TCR avidity threshold that separates positive and negative selection is set: cross-talk between glucocorticoid receptor (GR) and TCR signaling pathways.
|
Glucocorticoids
|
|---|
Glucocorticoids are small lipophilic molecules with potent effects on a variety of metabolic and secretory pathways, having a profound influence on the physiologic function of virtually all tissues [15]. Glucocorticoids are derived from cholesterol by a series of enzymatic reactions that generate intermediates such as pregnenolone, progesterone and 11-deoxycorticosterone or 11-deoxycortisol (Fig. 1
). Cortisol is the primary glucocorticoid in man, while in rodents, which lack the enzyme 17-hydroxylase (P450c17) [16], the principal corticosteroid is corticosterone; the deoxy forms of both are biologically inactive [17]. Glucocorticoids are produced at high levels primarily in the zona fasciculata of the adrenal cortex and circulate in the blood largely in a protein-bound state. Corticoid-binding globulin is a high-specificity, low-capacity carrier, while albumin is a low-specificity, high-capacity carrier. The free (unbound) form is the biologically active fraction, and constitutes approximately 10% of the total circulating amount.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 1. Simplified scheme of some of the major products and enzymatic activities in steroid biosynthesis. 3 ßHSD = 3 ß-hydroxysteroid dehydrogenase and isomerase. Pharmacologic inhibitors of enzymes in the pathway are underlined.
|
|
The physiological effects of corticosteroids are mediated by the GR, an ~100 kDa molecule that resides in the cytoplasm in a complex with other proteins, notably heat shock proteins such as hsp90 and the immunophilins hsp56 and CyP-40 [18]. Like other steroid hormone receptor superfamily members, the GR has (from C- to N-terminus) a ligand-binding domain, a DNA-binding domain with two Zn fingers and a transactivation domain [19]. When occupied by ligand, the GR translocates to the nucleus where it binds as a homodimer to specific DNA sequences called glucocorticoid responsive elements (GREs). The GR regulates transcription of the corresponding steroid-responsive genes, thereby functioning as a ligand-regulated transcription factor. One of the more intriguing mechanisms by which the occupied GR affects cell function is by its ability to interact with other transcription factors, such as AP-1, NF-
B, and the retinoic acid receptor (RAR), leading to transcriptional antagonism or, in some cases, synergy [19, 20]. A number of mechanisms have been proposed for this phenomenon, including direct physical interaction between transcription factors in solution or at nearby DNA-binding sites or, in the case of NF-B, the induction of IB by glucocorticoids [21, 22]. Cross-talk between the GR and other transcription factors provides a mechanism for fine control of biological responses to external stimuli.
One of the most prominent biological effects of glucocorticoids on lymphocytes is the induction of apoptosis. This is most pronounced in the thymus, where CD4+CD8+ (double-positive; DP) cells are particularly sensitive. Although the mechanism by which glucocorticoids cause apoptosis is not clear, it requires the presence of the GR and in thymocytes is prevented by reagents that block mRNA or protein synthesis [23]. It has therefore generally been assumed that one or more corticosteroid-responsive genes is responsible for causing thymocyte death. While this hypothesis was supported by early experiments showing that the transactivating domain of the GR was required for steroid-induced apoptosis [24], others have shown that the ligand-binding and transactivation domains of the GR are dispensable for the induction of apoptosis, and that only a portion of the DNA-binding domain was sufficient to cause cell death [25]. More recently it was shown that a transactivation-defective GR mutant that still binds glucocorticoids and represses AP-1 activity is capable of inducing apoptosis in a T cell line [26], leading to the proposal that it is in fact a nontranscriptional GR activity, such as binding to and/or interference with other nuclear transcription factors, that is the means by which glucocorticoids exert their lethal effect.
Glucocorticoid-induced T cell apoptosis is regulated in vivo by Bcl-2. Bcl-2 expression is high in very immature CD4–CD8– thymocytes (double-negative; DN), but goes down as the transition to DP cells occurs [27]. Expression increases once again in cells undergoing positive selection, and remains at a high level in CD4+CD8– and CD4–CD8+ peripheral T cells [28]. Complementary genetic experiments have confirmed that this pattern of Bcl-2 expression is responsible for susceptibility to glucocorticoid-induced apoptosis. First, in Tg mice that overexpress Bcl-2 in the thymus, DP thymocytes are resistant to glucocorticoid-induced death (as well as most other apoptotic stimuli) [8, 9, 29]. Second, mature T cells from Bcl-2 "knockout" (KO) mice are just as susceptible to glucocorticoid-induced apoptosis as are DP thymocytes [30]. Thus, regulated expression of Bcl-2 appears to be a key determinant in the cellular response to glucocorticoids.
|
Mutual Antagonism between Glucocorticoid- and T Cell Activation-Induced Apoptosis
|
|---|
T cell hybridomas have been used since the early 1980s as an in vitro model for the study of T cell activation. In the course of analyzing the effect of responding cell number on antigen-induced lymphokine production, our laboratory made the observation that stimulation of T cell hybridomas with antigen results in a G1/S cell cycle block and cell death [31, 32]. Subsequent studies revealed that the death was apoptotic, sensitive to cyclosporin A, Ca2+dependent, and required de novo mRNA and protein synthesis [33]. With the knowledge that glucocorticoids were also potent inducers of T cell hybridoma apoptosis, we wondered whether the cell death caused by these two stimuli was mediated by similar or distinct pathways. A simple experiment was devised to address this question: is the amount of apoptosis caused by the combination of T cell activation and glucocorticoids additive (implying the same pathway) or synergistic (implying different pathways) compared to the amount of death caused by either stimulus alone? Surprisingly, neither turned out to be the case. Rather, cells that were exposed to glucocorticoids in the presence of TCR occupancy survived, a phenomenon termed "mutual antagonism" [34]. It is now known that activation-induced T cell hybridoma death is due to the upregulation of Fas ligand and its interaction with Fas, and in fact glucocorticoids inhibit Fas ligand upregulation [35]. The molecular mechanism for this striking phenomenon, and how activation interferes with glucocorticoid-induced apoptosis, is as yet unknown, but presumably reflects transactivational interference between the occupied GR and other nuclear transcription factors, perhaps AP-1.
This observation led to the speculation that perhaps cross-talk between TCR and GR signaling pathways might regulate the survival of thymocytes undergoing selection [34, 36, 37]. In this model (Fig. 2), the death of thymocytes bearing TCRs with subthreshold avidity for self-antigen/MHC is due, at least in part, to glucocorticoids. Thymocytes with low-to-moderate avidity for self-antigen/MHC would also be stimulated to die by glucocorticoids, as they would by engagement of the TCR, but because these stimuli are mutually antagonistic, apoptosis is prevented and the thymocytes survive and differentiate to CD4+CD8– or CD4–CD8+ (single positive; SP) cells (positive selection). Finally, for thymocytes bearing TCRs with high avidity for self-antigen/MHC, the signal for apoptosis is too potent to be antagonized by the available glucocorticoids, resulting in antigen-specific clonal deletion (negative selection). Implicit in this model is that signaling via the TCR simultaneously drives the cell toward two distinct outcomes: differentiation as well as apoptosis. Obviously, death is "dominant," so that in the absence of a stimulus that counters TCR-induced apoptosis (i.e., glucocorticoids), the only measurable outcome is cell death. Furthermore, this model does not assume that there is a qualitative difference between the signals generated by a TCR occupied with a low-to-moderate avidity versus a high avidity ligand, although it is compatible with this possibility. The reason for the dramatic difference in the results obtained under each condition, life or death, is because of cross-talk from the TCR and the GR signaling pathways. Therefore, as shown in Figure 2, glucocorticoids set the avidity threshold that separates positive from negative selection.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 2. Fetal thymic organ cultures were performed as described [40] using fetal day 17 BALB/c thymi. Thymi were cultured in serum-free medium in the absence or presence of 0.5 mM aminoglutethimide (an inhibitor of P450scc), 20 mg/ml H57 (anti-TCR- ß), or 10–9 M corticosterone for three days at 37°C in 5% CO2. At that time, the cells were harvested, stained with antibodies against CD4 and CD8, and analyzed by flow cytometry. The error bars represent SEM of results of triplicate cultures per group.
|
|
|
Steroid Production in the Thymus
|
|---|
The mutual antagonism model, in its simplest form, cannot readily account for thymocyte development and selection because circulating glucocorticoid levels are extremely low during neonatal life [38, 39]. Circulating glucocorticoids do not reach adult levels until approximately four weeks after birth in rodents, and one year in humans. This is a problem for the mutual antagonism model, because the thymus is actively producing mature, antigen-specific T cells during fetal and neonatal life. Therefore, for the model to be valid, it must be true that the thymus itself is a site of glucocorticoid biosynthesis. A variety of methods were used to determine if, in fact, the thymus synthesizes steroids. First, the conversion of cholesterol to pregnenolone, the first step in the commitment to steroid biosynthesis (Fig. 1
), was determined in vitro. Pregnenolone was undetectable in the supernatants of dispersed, collagenase-treated thymus cells cultured for four days in the presence of cholesterol [40]. However, pregnenolone was readily detected in the supernatants of cells cultured in the presence of trilostane, an inhibitor of the enzyme 3ß-hydroxysteroid dehydrogenase (3ßHSD), which converts pregnenolone to progesterone. Besides establishing that the thymus synthesizes pregnenolone, the trilostane effect implied that the thymus contains 3ßHSD, and that thymic pregnenolone is normally converted to other steroids. In fact, small amounts of deoxycorticosterone were also detected in the in vitro cultures. To determine if the classical enzymes required for steroid biosynthesis are present in the thymus, paraffin-embedded sections of thymus were stained with antibodies to either P450scc or P450c11 (which converts deoxycorticosterone to corticosterone [Fig. 1]). Both antisera specifically stained large, radioresistant cells that had abundant cytoplasm (i.e., not thymocytes). Analysis of serial sections revealed that the two antisera stained the same cells, and at least some of these cells were identified as epithelial because they coexpressed the epithelial cell product cytokeratin.
Of particular importance is the ontogeny of steroid production by the thymus. As noted above, steroid production by the adrenal cortex is very low in fetal and neonatal life. The reason for this is unclear, but has been attributed to decreased responsiveness to adrenocorticotropic hormone (ACTH), decreased activity and/or presence of steroidogenic enzymes, or a combination of both [41]. To determine if this is true for the thymus as well, thymi from mice of different ages, from late fetal life to 12 weeks old, were collected and assayed for their ability to produce pregnenolone when cultured in vitro. Interestingly, the amount of pregnenolone produced on a per cell basis was highest for fetal thymi, declining with age until it reached a nadir at four weeks of age. It is notable that it is at this age in the rodent that adrenal production reaches adult levels. Therefore, the thymus is capable of producing steroids, and can do so during the earliest time of active thymocyte selection: fetal and neonatal life.
|
Development of Glucocorticoid-Insensitive Thymocytes
|
|---|
As one means of exploring the role of glucocorticoids in thymocyte development, an attempt was made to generate Tg mice in which the thymocytes were insensitive to the effects of glucocorticoids. Targeted disruption of the GR gene was avoided because we reasoned that this would likely be lethal, which ultimately proved to be the case [42]. Instead, we created Tg mice expressing mRNA antisense to the 3' untranslated region of the GR under the control of the proximal lck promoter. This promoter allows gene expression in a tissue-specific and developmentally coordinated fashion: its activity is high in early thymocyte development and decreases once differentiation to mature T cells and emigration from the thymus has occurred [43]. This strategy proved successful, and in animals homozygous for the GR antisense (called GR-TKO mice), an approximately twofold reduction in GR mRNA and GR protein was found [44]. This reduction is similar to that found in the brains of mice Tg for a GR antisense construct driven by the neurofilament promoter [45]. As expected, expression of the transgene was limited to the thymus, and GR levels were unaffected in a peripheral lymphoid organ, the spleen. The reduction of GR levels was found to be biologically significant as judged by a reduction in the response to glucocorticoids in two different assays: transactivation of a GRE-regulated luciferase reporter, and induction of apoptosis.
The most obvious effect of the antisense GR transgene was a dramatic reduction in the number of thymocytes. Total cellularity was reduced two- to threefold in hemizygous mice, and 10-fold or more in homozygous animals [44]. A phenotypic analysis was undertaken to determine at what stage of thymocyte development steroids exerted their effects. The most immature thymocytes express neither CD4 nor CD8, and are TCR– because they have not rearranged the TCR- or -ß genes. The TCR-ß chain is the first to rearrange, and if this event is successful, the TCR-ß chain is expressed on the cell surface as a heterodimer with a nonpolymorphic 33 kDa chain, pre-TCR- and in a complex with CD3 [46]. This pre-TCR is thought to interact with an as yet unidentified ligand, and the signals thus generated drive the thymocyte to differentiate to the CD4+CD8+ (DP) stage. This transition is accompanied by the rearrangement of TCR- chain genes and thus the appearance of CD4+CD8+TCR+ cells. It is these triple-positive cells that are susceptible to selection, either positive (in which case they mature into SP cells, CD4+CD8– and CD4–CD8+) or negative (in which case they die), if they encounter ligand in the thymus. In the GR-TKO mice, it was the DP cells and their SP progeny that were primarily affected. Importantly, DN cells were found in normal numbers, implying that colonization of the thymus by pro-thymocytes is normal in these mice, consistent with the notion that the effect(s) of glucocorticoids on T cell development are dependent on signaling via the TCR.
The data suggest that loss of GR leads to a "block" in the transition from DN to DP cells. This could be because glucocorticoids are required for progression to the DP stage or because differentiation to DP cells occurs normally, but the cells rapidly undergo apoptosis in the absence of glucocorticoid signaling. To test the latter possibility, apoptotic fresh thymocytes from normal mice and GR-TKO animals were identified by an assay in which nicked DNA is fluorescently labeled. While <3% of normal thymocytes were detected in this assay, almost 10% of thymocytes from GR-TKO mice were positive. The latter number probably underestimates the real fraction of cells undergoing apoptotic death, since apoptotic thymocytes are rapidly phagocytosed [5]. If the high frequency of apoptosis in vivo was due to enhanced susceptibility to TCR-mediated signals, one would predict that thymocytes from GR-TKO mice would be very sensitive to stimuli that mimic TCR occupancy. This was tested by adding an anti-TCR antibody that recognizes all ß TCRs to fetal thymic organ culture. With normal thymi, DP thymocyte deletion was detected only at antibody concentrations of 2 µg/ml. However, with GR-TKO thymi, statistically significant levels of deletion were detected with as little as 0.02 µg/ml of antibody. This 100-fold shift in the dose-response curve confirms that thymocytes hyporesponsive to glucocorticoids are in fact highly sensitive to TCR-mediated deletional signals.
Although these data are consistent with a role for glucocorticoids in maintaining the viability of antigen-stimulated thymocytes, this is unlikely to fully account for the phenotype of the GR-TKO mice because the loss of thymocytes is too great. If, as estimated, only a few percent of TCR+ thymocytes recognize self with a biologically significant avidity, why should a decrease in steroid-responsiveness result in loss of >90% of DP cells? This issue was explored by carefully examining the phenotype of the thymus during fetal ontogeny. Thymocytes are TCR– until approximately day 15 of fetal development, at which time they begin to rearrange TCR-ß. These thymocytes, which have not yet rearranged TCR-, enter a proliferative stage and begin to move from the DN to the DP stage [47-49]. Once the thymocytes are DP, TCR- rearrangement proceeds and significant numbers of TCR-ß+ cells begin to appear on fetal day 17. Studies with the GR-TKO mice revealed that the thymic phenotype was indistinguishable from that of nontransgenic controls through fetal day 15. On day 16, however, there was a significant (twofold) reduction in cell recovery, and by day 18, there were 10-fold fewer thymocytes (due to a reduction in the number of DP cells) in the GR-TKO thymi. Since there is little TCR-ß expressed at this time, this effect of the GR antisense transgene must not result from loss of antagonism of antigen-specific receptor occupancy.
The GR-TKO mice, therefore, have two defects in thymocyte development: a block in the transition from the DN to DP stage of differentiation, and increased apoptosis of those DP cells that manage to negotiate this transition. While it is certainly possible that the block in transition is because there are glucocorticoid-regulated genes necessary for biological processes involved in differentiation, there is one simplifying possibility that would account for both effects of the transgene. As outlined above, DN cells express a pre-TCR at approximately day 15 of fetal development, and expression of this receptor is necessary for progression from the DN to the DP stage of development. Given that signaling via the pre-TCR appears to be similar to signaling via the ß TCR [50], it is possible that occupancy of either receptor leads to apoptosis, an outcome that is prevented by glucocorticoids. If so, all cells that successfully rearrange TCR-ß and express a pre-TCR must pass this checkpoint, and in the GR-TKO mice many fail (and die) because of their insensitivity to glucocorticoids. This hypothesis predicts that the GR antisense transgene would have little or no effect if the progression from DN to DP could be achieved in the absence of pre-TCR occupancy. This does, in fact, occur in mice that overexpress a constitutively active form of the Lck tyrosine kinase [51], and we are currently undertaking the breeding necessary to perform this experiment.
|
Fetal Thymocyte Development in the Absence of Corticosterone
|
|---|
Although the GR antisense mice demonstrated that responsiveness to glucocorticoids is essential for normal T cell development, they do not allow one to determine if thymus-derived, as opposed to circulating, glucocorticoids have an important role in this process. To test this, development of thymocytes from normal mice was followed in fetal thymic organ culture (FTOC). Fetal thymi taken at day 17 of gestation, can be cultured on a sponge in vitro for 7-10 days, during which time the thymocytes differentiate as they would in vivo. One clear prediction of the mutual antagonism model is that removing glucocorticoids in these cultures should enhance the susceptibility of DP thymocytes to deletion by a TCR-mediated stimulus. To achieve this, we have used the drug aminoglutethimide, which inhibits P450scc, the enzyme responsible for the conversion of cholesterol to pregnenolone (Fig. 1). As a TCR-mediated stimulus, an anti-TCR-ß antibody (H57) was added to the FTOC. By cross-linking the TCR on immature thymocytes, this antibody mimics receptor occupancy and causes deletion of DP cells [52]. In fact, as shown in a representative experiment (Fig. 2), in a three-day FTOC performed with serum-free medium, the anti-TCR antibody resulted in a modest loss of DP cells (from 76% to 49% in this experiment). While aminoglutethimide alone at the concentration used had little effect on thymocyte development (data not shown), it markedly enhanced TCR-mediated deletion (from 49% to 26% DP cells in this experiment). To ensure that this effect was related to the inhibitory activity of aminoglutethimide on P450scc function, cultures in which corticosterone was added back at physiological levels (nanomolar for protein nonbound, or free, steroid) were analyzed: the aminoglutethimide effect was reversed in the presence of corticosterone. Similar results were obtained with two reagents that are more specific for GR-mediated effects: RU-486, a competitive glucocorticoid antagonist, and metyrapone, which inhibits the enzyme responsible for the conversion of deoxycorticosterone (inactive) to corticosterone (active), P450c11 [40]. These results strongly suggest that endogenous glucocorticoids do in fact determine the susceptibility of thymocytes to TCR-mediated deletion.
Among the predictions that follow from the mutual antagonism model of thymocyte selection is that inhibition of glucocorticoid production should cause low-to-moderate avidity TCR interactions, which normally result in rescue from the default death pathway, to induce apoptosis (negative selection). To test this, use was made of Tg mice that express a TCR-ß recognizing the male H-Y antigen in association with the H-2Db-encoded class I molecule (Db) [53]. A key feature of these mice is that while thymocytes in male Tg mice are deleted because of the high avidity interaction of the clonotypic TCR with H-Y/Db, thymocytes from female mice, which do not express the H-Y antigen, are not deleted. Moreover, there appears to be a low-to-moderate avidity ligand for this Tg TCR expressed in the thymus of female mice, in H-2b females, but not females of other haplotypes, thymocytes are positively selected, as detected by the appearance of TCR-ß-transgene-bearing CD4–CD8+ thymocytes. We have found that the addition of metyrapone to FTOC of day 18 thymi results in a much greater decrease in DP thymocyte number in female H-2b TCR Tg mice (~70%) than in non-Tg control animals (~%20) (MSV and JDA, manuscript submitted). In contrast there was no difference in the effect of metyrapone on thymocyte recovery when non-H-2b thymi from TCR transgenic and nontransgenic mice were studied. Therefore, the inhibition of glucocorticoid production specifically caused clonal deletion when the TCR had low-to-moderate avidity for self (i.e., an avidity that normally results in positive selection).
|
Implications and Speculations
|
|---|
One of the central issues remaining in cellular immunology is how the antigen-specific repertoire is formed. We know that the life-or-death decision made by each developing thymocyte is based largely upon the avidity with which its TCR recognizes self-antigen/MHC, but how this information is translated into a biological outcome is a mystery. Our data suggest that the biological outcome is not simply determined by TCR signaling, but rather by interactions, or cross-talk, between TCR and GR signaling pathways (Fig. 3). Although the details of this cross-talk are not yet known, there are numerous clues that point to a likely molecular mechanism. The occupied GR can interact with other transcription factors, such as AP-1 (which is activated by TCR signaling events) and RARs, and cause inhibition of nuclear gene activation [20]. In the case of T cell hybridomas, activation-induced apoptosis is the result of upregulation of Fas ligand (FasL), which then binds to Fas [54, 55], and glucocorticoids inhibit activation-induced death by preventing the upregulation of FasL [35]. Whether this is due to interference with AP-1, for which there are several potential binding sites in the FasL 5' promoter/enhancer region [56], remains to be determined. Similarly, GRE-regulated gene transcription in T cell hybridomas is inhibited by TCR-mediated activation, presumably because of AP-1/GR interactions (C. M. Zacharchuk and JDA, unpublished observation). It is unlikely that the genes encoding Fas and FasL are the targets for the GR/TCR cross-talk in thymocytes, however, as this receptor-ligand pair does not appear to play an important role in thymocyte selection [57]. A recent study indicates that the FasL-related ligand for CD30 may be essential for negative selection [58]. It is an intriguing possibility that this molecule may be the relevant target for receptor cross-talk.
View larger version (21K):
[in this window]
[in a new window]
|
Figure 3. Schematic of the mutual antagonism model. As shown, thymic epithelial cells as well as the adrenals produce glucocorticoids. Circulating corticosteroids are implicated in death by neglect because adrenalectomy results in thymic hyperplasia; their role in antagonizing TCR-mediated apoptosis is unknown. Exposure to thymic-derived corticosteroids in the presence of a low-to-moderate avidity TCR-ligand interaction results in mutual antagonism, survival and differentiation. High avidity TCR-ligand interactions are too potent to be antagonized by glucocorticoids.
|
|
One possible relationship between avidity of the TCR for self-antigen/MHC and cell fate is shown in Figure 4. The available evidence suggests that the large majority of thymocytes have little or no avidity, and are neither positively nor negatively selected. The few percent that have significant avidity for self are either rescued from death or are induced to undergo apoptosis. Our results suggest that the threshold separating positive from negative selection is adjustable, and that interfering with the response to locally produced glucocorticoids can shift it towards a lower avidity. This finding has implications for both physiologic and pathologic immune function, and provides at least a theoretical potential for pharmacologic manipulation of the antigen-specific T cell repertoire. One important implication of this model is that, as glucocorticoids (or responsiveness to glucocorticoids) decrease, the positive selection "window" will shrink. As a result, thymocytes that would normally be positively selected and contribute to the peripheral immune response are instead deleted. In the extreme, the loss of these cells will result in "holes in the repertoire," which is manifested as a failure to respond to a particular peptide because the T cells with the proper specificity have been deleted. We are currently testing this possibility by taking advantage of the fact that despite the very small size of the thymus, the GR-TKO mice have relatively high levels of peripheral T cells (~50% of normal). These mice will be immunized with defined peptides to which only a restricted array of TCRs normally respond. Evidence of a shift in the antigen-specific repertoire would include hyporesponsiveness to a peptide/MHC combination that normally generates a good response, or perhaps a change in the TCR-ß combinations used by responding cells. Another possible consequence of deletion of thymocytes with relatively high avidity for self-antigen/MHC (but not high enough to cause deletion under normal circumstances) is that the cells that are most likely to be self-reactive in a pathologic manner (i.e., autoimmune) might be deleted. To test this, we are crossing the GR-TKO mice with autoimmune strains, and will monitor the onset and severity of antiself responses. Since the transgene is not expressed in mature T cells [44], any abnormalities would presumably reflect the selection of cells that emigrate from the thymus. Ultimately, proof of the hypothesis that glucocorticoids regulate antigen-specific T cell selection rests on the demonstration of their effect on the antigen-specific T cell repertoire.
View larger version (16K):
[in this window]
[in a new window]
|
Figure 4. Model for the relationship between thymocyte TCR avidity for self-antigen/MHC and cell fate. Glucocorticoids, possibly only those that are produced in the thymus, regulate the setting of the threshold that separates positive from negative selection.
|
|
|
References
|
|---|
-
Davis MM, Bjorkman PJ. T-cell antigen receptor genes and T-cell recognition. Nature 1988;334:395-402.[Medline]
-
Sprent J, Lo D, Gao E-K et al. T cell selection in the thymus. Immunol Rev 1988;101:172-190.
-
Shortman K, Egerton M, Spangrude GJ et al. The generation and fate of thymocytes. Semin Immunol 1990;2:3-12.[Medline]
-
Huesmann M, Scott B, Kisielow P et al. Kinetics and efficacy of positive selection in the thymus of normal and T cell receptor transgenic mice. Cell 1991;66:533-540.[Medline]
-
Surh CD, Sprent J. T-cell apoptosis detected in situ during positive and negative selection in the thymus. Nature 1994;372:100-103.[Medline]
-
Jaffe HL. The influence of the suprarenal gland on the thymus. III. Stimulation of the growth of the thymus gland following double supradrenalectomy in young rats. J Exp Med 1924;40:753-759.[Abstract]
-
Inomata T, Nakamura T. Influence of adrenalectomy on the development of the neonatal thymus in the rat. Biol Neonate 1989;55:238-243.[Medline]
-
Sentman CL, Shutter JR, Hockenbery D et al. bcl-2 inhibits multiple forms of apoptosis but not negative selection in thymocytes. Cell 1991;67:879-888.[Medline]
-
Strasser A, Harris AW, Cory S. bcl-2 transgene inhibits T cell death and perturbs thymic self-censorship. Cell 1991;67:889-899.[Medline]
-
Ashton-Rickardt PG, Tonegawa S. A differential-avidity model for T-cell selection. Immunol Today 1994;15:362-366.
-
Hogquist KA, Jameson SC, Heath WR et al. T cell receptor antagonist peptides induce positive selection. Cell 1994;76:17-27.[Medline]
-
Sloan-Lancaster J, Shaw AS, Rothbard JB et al. Partial T cell signaling: altered phospho- and lack of zap70 recruitment in APL-induced T cell anergy. Cell 1994;79:913-922.[Medline]
-
Madrenas J, Wange RL, Wang JL et al. Zeta phosphorylation without ZAP-70 activation induced by TCR antagonists or partial agonists. Science 1995;267:515-518.[Abstract/Free Full Text]
-
Page DM, Alexander J, Snoke K et al. Negative selection of CD4+CD8+ thymocytes by T-cell receptor peptide antagonists. Proc Natl Acad Sci USA 1994;91:4057-4061.[Abstract/Free Full Text]
-
Goodman HM. Basic Medical Endocrinology, Second Edition. New York: Raven Press, 1994.
-
LaPlante C, Giroud CJP, Stachenko J. Lack of appreciable 17-hydroxylase activity in the normal and regenerated rat adrenal cortex. Endocrinology 1964;75:825-827.
-
Milgrom E. Steroid hormones. In: Baulieu E-E, Kelly PA, eds. Hormones: from Molecules to Disease. New York: Hermann Chapman and Hall, 1990:386-437.
-
Owens-Grillo JK, Hoffmann K, Hutchison KA et al. The cyclosporin A-binding immunophilin CyP-40 and the FK506-binding immunophilin hsp56 bind to a common site on hsp90 and exist in independent cytosolic heterocomplexes with the untransformed glucocorticoid receptor. J Biol Chem 1995;270:20479-20484.[Abstract/Free Full Text]
-
Beato M, Herrlich P, Schutz G. Steroid hormone receptors: many actors in search of a plot. Cell 1995;83:851-857.[Medline]
-
Zacharchuk CM, Ashwell JD. Fruitful outcomes of intracellular cross-talk. Curr Biol 1992;2:246-248.[Medline]
-
Auphan N, Didonata JA, Rosette C et al. Immunosuppression by glucocorticoids: inhibition of NF-B activity through induction of IB synthesis. Science 1995;270:286-290.[Abstract/Free Full Text]
-
Scheinman RI, Cogswell PC, Lofquist AK et al. Role of transcriptional activation of IB in mediation of immunosuppression by glucocorticoids. Science 1995;270:283-286.[Abstract/Free Full Text]
-
Schwartzman RA, Cidlowski JA. Apoptosis: the biochemistry and molecular biology of programmed cell death. Endocr Rev 1993;14:133-151.[Medline]
-
Dieken ES, Miesfeld RL. Transcriptional transactivation functions localized to the glucocorticoid receptor N terminus are necessary for steroid induction of lymphocyte apoptosis. Mol Cell Biol 1992;12:589-597.[Abstract/Free Full Text]
-
Thompson EB, Nazareth LV, Thulasi R et al. Glucocorticoids in malignant lymphoid cells: gene regulation and the minimum receptor fragment for lysis. J Steroid Biochem Mol Biol 1992;41:273-282.[Medline]
-
Helmberg A, Auphan N, Caelles C et al. Glucocorticoid-induced apoptosis of human leukemic cells is caused by the repressive function of the glucocorticoid receptor. EMBO J 1995;14:452-460.[Medline]
-
Hockenbery DM, Zutter M, Hickey W et al. Bcl-2 protein is topographically restricted in tissues characterized by apoptotic cell death. Proc Natl Acad Sci USA 1991;88:6961-6965.[Abstract/Free Full Text]
-
Linette GP, Grusby MJ, Hedrick SM et al. Bcl-2 is upregulated at the CD4+CD8+ stage during positive selection and promotes thymocyte differentiation at several control points. Immunity 1994;1:197-205.[Medline]
-
Siegel RM, Katsumata M, Miyashita T et al. Inhibition of thymocyte apoptosis and negative antigenic selection in bcl-2 transgenic mice. Proc Natl Acad Sci USA 1992;89:7003-7007.[Abstract/Free Full Text]
-
Nakayama K, Nakayama K, Negishi I et al. Disappearance of the lymphoid system in Bcl-2 homozygous mutant chimeric mice. Science 1993;261:1584-1588.[Abstract/Free Full Text]
-
Ashwell JD, Cunningham RE, Noguchi PD et al. Cell growth cycle block of T cell hybridomas upon activation with antigen. J Exp Med 1987;165:173-194.[Abstract/Free Full Text]
-
Mercep M, Bluestone JA, Noguchi PD et al. Inhibition of transformed T cell growth in vitro by monoclonal antibodies directed against distinct activating molecules. J Immunol 1988;140:324-335.[Abstract]
-
Ucker DS, Ashwell JD, Nickas G. Activation driven T cell death. I. Requirements for de novo transcription and translation and association with genome fragmentation. J Immunol 1989;143:3461-3469.[Abstract]
-
Zacharchuk CM, Mercep M, Chakraborti P et al. Programmed T lymphocyte death: cell activation- and steroid-induced pathways are mutually antagonistic. J Immunol 1990;145:4037-4045.[Abstract]
-
Yang Y, Mercep M, Ware CF et al. Fas and activation-induced Fas ligand mediate apoptosis of T cell hybridomas: inhibition of Fas ligand expression by retinoic acid and glucocorticoids. J Exp Med 1995;181:1673-1682.[Abstract/Free Full Text]
-
Zacharchuk CM, Mercep M, Ashwell JD. Thymocyte activation and death: a mechanism for molding the T cell repertoire. In: Edelson RL, ed. Antigen and Clone-Specific Immunoregulation. New York: Annals of the New York Academy of Science, 1991:52-70.
-
Iwata M, Hanaoka S, Sato K. Rescue of thymocytes and T cell hybridomas from glucocorticoid-induced apoptosis by stimulation via the T cell receptor/CD3 complex: a possible in vitro model for positive selection of the T cell repertoire. Eur J Immunol 1991;21:643-648.[Medline]
-
Henning SJ. Plasma concentrations of total and free corticosterone during development in the rat. Am J Physiol 1978;235:E451-E456.[Abstract/Free Full Text]
-
Savu L, Zouaghi H, Nunez EA. Serum inflammatory responses of transcortin binding activities and of total and free corticosterone and progesterone levels in developing rats: a kinetic approach. Int J Tissue React 1985;7:443-448.[Medline]
-
Vacchio MS, Papadopoulos V, Ashwell JD. Steroid production in the thymus: implications for thymocyte selection. J Exp Med 1994;179:1835-1846.[Abstract/Free Full Text]
-
Arai M, Widmaier EP. Steroidogenesis in isolated adrenocortical cells during development in rats. Mol Cell Endocrinol 1993;92:91-107.[Medline]
-
Cole TJ, Blendy JA, Monaghan AP et al. Targeted disruption of the glucocorticoid receptor gene blocks adrenergic chromaffin cell development and severely retards lung maturation. Genes Dev 1995;9:1608-1621.[Abstract/Free Full Text]
-
Garvin AM, Abraham KM, Forbush KA et al. Disruption of thymocyte development and lymphomagenesis induced by SV40 T-antigen. Int Immunol 1990;2:173-180.[Abstract/Free Full Text]
-
King LB, Vacchio MS, Dixon K et al. A targeted glucocorticoid receptor antisense transgene increases thymocyte apoptosis and alters thymocyte development. Immunity 1995;3:647-656.[Medline]
-
Pepin MC, Pothier F, Barden N. Impaired type II glucocorticoid-receptor function in mice bearing antisense RNA transgene. Nature 1992;355:725-728.[Medline]
-
Saint-Ruf C, Ungewiss K, Groettrup M et al. Analysis and expression of a cloned pre-T cell receptor gene. Science 1994;266:1208-1212.[Abstract/Free Full Text]
-
Rothenberg EV. The development of functionally responsive T cells. Adv Immunol 1992;51:85-214.[Medline]
-
Andjelic S, Jain N, Nikolic-Zugic J. Ontogeny of fetal CD8lo4lo thymocytes: expression of CD44, CD25 and early expression of TcR mRNA. Eur J Immunol 1993;23:2109-2115.[Medline]
-
Groettrup M, von Boehmer H. T cell receptor ß chain dimers on immature thymocytes from normal mice. Eur J Immunol 1993;23:1393-1396.[Medline]
-
van Oers NS, von Boehmer H, Weiss A. The pre-T cell receptor (TCR) complex is functionally coupled to the TCR- subunit. J Exp Med 1995;182:1585-1590.[Abstract/Free Full Text]
-
Mombaerts P, Anderson SJ, Perlmutter RM et al. An activated lck transgene promotes thymocyte development in RAG-1 mutant mice. Immunity 1994;1:261-267.[Medline]
-
Finkel TH, Cambier JC, Kubo RT et al. The thymus has two functionally distinct populations of immature ß+ T cells: one population is deleted by ligation of ß TCR. Cell 1989;58:1047-1054.[Medline]
-
Kisielow P, Blüthmann H, Staerz U et al. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4+8+ thymocytes. Nature 1988;333:742-746.[Medline]
-
Brunner T, Modil RJ, LaFace D et al. Cell-autonomous Fas (CD95)/Fas-ligand interaction mediates activation-induced apoptosis in T-cell hybridomas. Nature 1995;373:441-444.[Medline]
-
Ju S-T, Panka DJ, Cui H et al. Fas(CD95)/FasL interactions required for programmed cell death after T-cell activation. Nature 1995;373:444-448.[Medline]
-
Takahashi T, Tanaka M, Inazawa J et al. Human Fas ligand: gene structure, chromosomal location, and species specificity. Int Immunol 1994;6:1567-1574.[Abstract/Free Full Text]
-
Singer GG, Abbas AK. The Fas antigen is involved in peripheral but not thymic deletion of T lymphocytes in T cell receptor transgenic mice. Immunity 1994;1:365-371.[Medline]
-
Amakawa R, Hakem A, Kundig TM et al. Impaired negative selection of T cells in Hodgkin's disease antigen CD30-deficient mice. Cell 1996;84:551-562.[Medline]
accepted for publication June 7, 1996.
Top
Abstract
Thymocyte Development
