Thymocyte-positive selection is believed to result from low affinity/avidity interactions of TCR and MHC proteins, but how these weak interactions translate into downstream biochemical signals and how such signals modulate gene expression is unknown. Using a culture system where isolated immature thymocytes can be induced to differentiate along the CD4 lineage pathway, we show that sustained low level mitogen-activated protein/extracellular regulated kinase activity is required for cell differentiation and survival. Furthermore, induction of mitogen-activated protein/extracellular regulated kinase activity is surprisingly slow under conditions that induce differentiation. This pattern of kinase activity not only selects which genes are expressed but also regulates the temporal pattern of expression of transcription factor genes that are induced during double-positive thymocyte differentiation.
The TCR-mediated recognition of MHC is required for continued development of immature thymocytes past the CD4+8+ (double-positive, DP)4 stage; a process called positive selection (1). Because the elimination of potentially autodestructive thymocytes, termed negative selection, is also TCR mediated, it has been proposed for many years (2), and recent evidence supports (3, 4), that positive selection results from low affinity or low avidity TCR recognition, whereas higher affinity/avidity TCR interactions induce thymocyte death. Thus, the presumption has been that positive selection results in “weak” signals that can be distinguished from the signals that induce negative selection or mature T cell activation. How this is accomplished biochemically and how such signals regulate gene expression are unknown. We focus here on understanding how mitogen-activated protein/extracellular regulated (MAP/ERK) kinase activation, which is obligatory for thymocyte positive selection (5, 6), is regulated under conditions that induce thymocyte differentiation, and how the activation of this signaling pathway influences gene expression.
A combination of phorbol ester and calcium ionophore can mimic many aspects of the biochemical signals required for mature T cell activation (7). Recently, it has been demonstrated that these same pharmacological agents can also induce the differentiation and survival of isolated DP thymocytes (8, 9, 10). In this instance, however, a narrow range of concentrations of ionomycin and PMA, as well as limited exposure, are required. This system allows one to directly measure intracellular changes as a function of time after initiation of the activating signal. Using this system, we have directly measured the activation of MAP/ERK kinases and changes in expression of downstream gene targets that accompany the very earliest stages of differentiation of isolated thymocytes. We have made the unexpected observation that the strength of stimulus plays a critical role in the kinetic regulation of MAP/ERK kinase activation. In addition, we show that the temporal pattern of MAP kinase activation has important consequences for downstream gene regulation, and presumably thymocyte differentiation. Moreover, our results point to a novel mechanism, in addition to strength of signal, by which modulation of a single downstream signaling pathway can differentially affect gene expression. Thus, the MAP kinase signaling pathway can transmit more information than a simple on-off switch that couples cell surface signals to changes in gene expression.
Materials and Methods
TCR α-chain knockout mice (11) were purchased from The Jackson Laboratory (Bar Harbor, ME) and were housed under specific-pathogen-free conditions.
Thymocytes derived from TCR α-chain knockout mice were induced to differentiate over 44 h by PMA and ionomycin treatment in a two-step culture system as previously described (8, 9, 10). Briefly, 8 × 106 cells/ml were cultured with 0.2 ng/ml PMA and 0.2 μg/ml ionomycin in the presence or absence of selective MEK inhibitors. After 20 h, cells were washed and cultured for an additional 24 h in the absence of stimulation. A coreceptor reexpression assay was used to determine the lineage commitment of thymocytes, as previously described (15). In this assay, cultured thymocytes are treated with pronase to remove cell surface coreceptor. Thymocytes are then cultured for an additional 20 h to allow reexpression of newly synthesized cell surface CD4 and CD8, a reflection of gene expression and lineage commitment.
Thymocytes were stained with anti-CD4-PE (Life Technologies) and anti-CD8α (conjugated to Cychrome or APC) mAbs (PharMingen, San Diego, CA) or single-color stained with anti-CD69-biotin (PharMingen) and streptavidin-APC (Biomedia, Foster City, CA). Stained cells were analyzed using CellQuest software on a FACSort upgraded to a FACSCaliber (Becton Dickinson, San Jose, CA). Shown is the log fluorescence of 5,000–10,000 viable cells, gated according to their sideways and forward light scatter.
To determine the distribution of early growth response 1 (Egr-1) expression in thymocyte subsets, thymocytes were surface stained with anti-CD4-PE and anti-CD8-APC mAbs and then internally stained for Egr-1 as previously described (16) using an anti-Egr-1 antiserum (sc-189) (Santa Cruz Biotechnology, Santa Cruz, CA). The same protocol was also used for internal staining with an anti-Nur77 antiserum (sc-990) (Santa Cruz Biotechnology). This particular antiserum was generated against Nurr1 but cross-reacts with Nur77. Nurr1 is not expressed in the thymus (17). A second specific anti-Nur77 antiserum (sc-7014) (Santa Cruz Biotechnology) gave similar results, although the intensity of staining with this particular reagent and corresponding secondary Ab was markedly lower (data not shown).
MAP kinase assay
Thymocytes were stimulated with 0.2 μg/ml ionomycin and various concentrations of PMA at 37°C in the presence or absence of 25 μM PD98059. MAP/ERK kinase activity in total thymocyte cell lysates (1.5 × 105 cell equivalents) was measured in triplicate by the 32P-phosphorylation of a specific peptide substrate as previously described (18).
Total RNA was isolated from thymocytes using the RNeasy RNA kit (Qiagen, Chatsworth, CA) and reverse transcribed using the SuperScript Preamplification System (Life Technologies, Gaithersburg, MD) according to the manufacturer’s instructions. Reactions contained 10 pmol of each primer, 1× PCR buffer (Life Technologies), 0.2 mM dNTPs, 2 mM MgCl2, 1 U Taq polymerase (Life Technologies), and 5 μl appropriately diluted cDNA in a total reaction volume of 25 μl. Cycle conditions were 94°C for 4 min followed by 30 cycles of 94°C for 30 s, 60°C for 40 s, and 72°C for 1 min. Primer sequences have been previously reported (16, 18) or are available on request. Comparisons for a given gene are made between identical cell equivalents of cDNA.
Electrophoretic mobility shift assay
Nuclear extracts derived from 1 × 106 cultured thymocytes were assayed for Egr-1 DNA binding activity using an [α-32P]dCTP-labeled oligonucleotide probe containing an Egr-1 consensus binding site as previously described (16).
Sustained MEK activation is required for initiation of immature thymocyte differentiation
DP thymocytes derived from TCR α-chain knockout mice (typically >90% DP) cultured for 20 h with low concentrations of PMA and ionomycin down-regulate CD4 and CD8 (Fig. 1,A). After an additional 24 h of culture in the absence of stimulation, the majority of thymocytes transit to a CD4+8low phenotype (Fig. 1,A). Although both CD4 and CD8 lineage thymocytes can pass through a phenotypically similar stage (15, 19), the majority of the CD4+8low thymocytes present after 44 h of culture belong to the CD4 lineage (8, 9), as assessed using a coreceptor reexpression assay (15) (Fig. 1,A). In contrast, thymocytes that do not receive stimulation in primary culture remain DP (Fig. 1 A). CD4 single-positive thymocytes generated under these conditions express other markers indicative of thymocyte maturation and have been demonstrated to produce cytokines after stimulation through CD3/CD28 (Refs. 8, 9, 10 and data not shown).
Activation of both p21ras and downstream MEK has been demonstrated to be obligatory for positive selection (5, 6, 16, 20). Activation of protein kinase C, the target of PMA, activates this pathway (21), and thus it seemed likely that PMA may induce thymocyte differentiation by activation of this MAP kinase-signaling cascade. This was investigated by analyzing the effects of the selective MEK inhibitors PD98059 (12, 13) and U0126 (14) on thymocyte differentiation in the two-step culture system described. PD98059 has previously been reported to block positive selection of thymocytes in fetal thymic organ culture at concentrations of 20–30 μM (22, 23). In accordance with these previous findings, the presence of the PD98059 during the initial 20-h culture inhibited the production of CD4+ thymocytes (Fig. 1, B and C). Similarly, the presence of U0126, shown previously to block the ERK pathway in T cells at 10 μM (24), resulted in an almost complete inhibition of thymocyte differentiation at concentrations as low as 2 μM (Fig. 1, D and E). In contrast to these results, when the MEK inhibitor PD98059 was present only during the final 24 h of culture, no inhibition was observed (Fig. 1, B and C).
To investigate how long a MEK signal was required for the induction of thymocyte maturation, PD98059 was added at various times after addition of PMA + ionomycin. If PD98059 was added as late as 8 h, but before 18 h, after initiation of culture, there was also a marked inhibition of the production of CD4 lineage thymocytes (Fig. 1, B and C). Thus, differentiation of DP thymocytes to a more mature phenotype requires sustained MEK activity during the initial activation phase of the culture, although subsequent cellular changes are MEK independent.
The activation of TCR-α− thymocytes in the presence of MEK inhibitors not only inhibited thymocyte differentiation but also decreased the proportion of DP thymocytes at the end of culture, suggesting an increase in cell death under these conditions (Fig. 1, B and D). Thus, the effects of such inhibitors on cell survival in culture were investigated further. As shown in Fig. 2, there is increased recovery of thymocytes that have been stimulated with PMA + ionomycin in the absence of MEK inhibitors as compared with unstimulated controls. This is consistent with previous results using this system (8). At 20 h of culture, the presence or absence of MEK inhibitors has little effect on cell recovery (Fig. 2,B). In contrast, at 44 h, the recovery of cells activated with PMA + ionomycin in the presence of PD98059 (Fig. 2,A) or U0126 (Fig. 2,B) is significantly reduced in comparison to cells triggered in the absence of inhibitors, and below medium controls. In the absence of activation, however, the equivalent concentrations of PD98059 and U0126 had no detrimental effects on cell recovery, indicating that the effect of such inhibitors on the survival of DP thymocytes is unlikely to be the result of nonspecific toxicity. This is further supported by the lack of any effect of PD98059 if added late to culture (Fig. 1, B and C). Together, the data indicate that there is poor survival, and possibly enhanced death, of thymocytes that have been activated in the absence of MEK-mediated signals.
Thymocyte differentiation is associated with slow accumulation of low level MAP/ERK kinase
In vitro kinase assays were performed to directly measure the activation of MAP/ERK kinases during DP thymocyte differentiation. Because differentiation and survival of thymocytes in this system is dependent on the concentration of PMA, we were also able to directly compare kinase activation under conditions that induce cell differentiation or cell death. A high (20 ng/ml) concentration of PMA (in the presence of ionomycin), which is optimal for mature T cell activation but induces thymocyte death (Refs. 7 , 9 , and 25 and data not shown), elicits a rapid rise in MAP/ERK kinase activity (Fig. 3,A). The peak of enzymatic activity is transient and rapidly declines within 30 min, followed by a sustained lower plateau of activity. Reducing the concentration of PMA to 0.5 ng/ml eliminates the early rise of MAP/ERK kinase activity, although by 2 h kinase activity increases to a level comparable to that seen with 20 ng/ml PMA (Fig. 3,A). Surprisingly, at the concentration of PMA required to induce optimal thymocyte differentiation (0.2 ng/ml), there is a gradual increase in MAP/ERK kinase activity that even by 5 h does not reach the level obtained with higher concentrations of PMA (Fig. 3,A). As expected for bona fide MAP/ERK kinase activity, this low level activity could be inhibited by PD98059 (Fig. 3 B). Although we do not detect induced enzymatic activity in cultures treated with 0.2 ng/ml PMA at early time points, there is clearly functional MAP kinase activation at least by 2 h as assessed by changes in gene expression (see below).
Temporal regulation of gene expression during thymocyte differentiation
To determine how these distinct patterns of MAP/ERK kinase activation would impact gene expression, RT-PCR analysis was performed on thymocytes activated with high (20 ng/ml) or low (0.2 ng/ml) concentrations of PMA, all in the presence of ionomycin (Fig. 4,A). CD4 gene expression was down-regulated after 5 h of culture with high or low concentrations of PMA. CD8α and CD8β genes displayed a similar pattern of expression, although the higher concentration of PMA had a more profound effect. These expression patterns are consistent with the observed initial down-regulation of CD4, and to a lesser extent CD8α, on the cell surface following activation with PMA (Fig. 1).
We next examined expression of transcription factors that are induced during thymic selection. Egr-1, Egr-2, and Egr-3 are immediate early genes encoding related zinc finger transcription factors that are up-regulated as a consequence of TCR activation during positive selection (16). We have previously shown that expression of these transcription factors in a thymocyte cell line is dependent on Ras activation (16). As expected, Egr-1 mRNA (Fig. 4,A) and DNA binding activity (Fig. 4,C) are also up-regulated during differentiation of thymocytes induced by PMA + ionomycin, and expression is inhibitable by PD98059 (Fig. 4,C). These data confirm the requirement for MAP/ERK kinase activation in Egr-1 up-regulation during thymocyte differentiation in this system. In thymocytes treated with 20 ng/ml PMA + ionomycin, the Egr-1 gene was expressed within 30 min, and expression was maintained throughout the 5 h of the experiment (Fig. 4,A). This is similar to the expression pattern observed in mature lymphocytes, where Egr-1 mRNA expression peaks at 30 min after activation (26). In contrast, stimulation with 0.2 ng/ml PMA + ionomycin resulted in delayed expression of Egr-1, as well as Egr-2 and Egr-3 (Fig. 4,A). However, unlike Egr-1 and Egr-2, Egr-3 expression was transient, and the lower concentration of PMA was unable to fully up-regulate this gene. The induction of CD69, often used as an early marker of positive selection and expression of which is Ras dependent (15, 17, 27, 28), was also delayed under conditions that elicit differentiation (Fig. 4 A).
In contrast to these results, the Nur77 gene, encoding a transcription factor implicated in T cell death (29) and thymocyte-negative selection (30), is up-regulated by stimulation of thymocytes with ionomycin alone (Fig. 4,A), consistent with a previous report (31). This up-regulation is transient, however, and Nur77 gene expression returns to basal levels by 2 h in culture. At low concentrations of PMA, there is only a modest increase in expression, whereas in the presence of 20 ng/ml PMA, high level Nur77 expression is sustained (Fig. 4,A). The maintenance of Nur77 expression by PMA is inhibitable by PD98059 (Fig. 4 B), suggesting that the duration, but not initiation, of expression of Nur77 may be influenced by the MAP/ERK kinase signaling pathway. Together, these results demonstrate that both the kinetics and magnitude of MAP/ERK kinase activation play a role in regulating downstream transcription factor gene expression.
Slow accumulation of active MAP kinase on a per cell basis
The results presented in this study indicate that PMA concentration can regulate the level and kinetics of induction of MAP kinase activity, resulting in alterations in the temporal pattern of gene expression. To support the interpretation that the observed slow accumulation of MAP kinase activity is on a per cell basis as opposed to a gradual increase in the number of cells with activated MAP kinase, we stained thymocytes for CD69 at various times after activation. Rapid induction of CD69 in this system is sensitive to PD98059 (not shown).
As shown in Fig. 5, low level cell surface CD69 is first detected at 1 h in thymocytes activated with 20 ng/ml PMA. By 3 h after activation, all the thymocytes express high levels of CD69, and expression is maintained at the 5-h time point. In contrast, thymocytes triggered by 0.2 ng/ml PMA have no detectable CD69 at 1 h and only low level expression at 3 h after activation. By 5 h after initiation of culture, there is only a slight difference in expression of CD69 between thymocytes triggered with the different concentrations of PMA. These results show that all thymocytes are responsive to stimulation and that the expression of cell surface CD69 increases, on a per cell basis with time. The kinetics of CD69 induction is dependent on the concentration of PMA. These results support the view that during thymocyte differentiation there is slow accumulation of active MAP kinase on a per cell basis.
Expression of Egr-1 and Nur77 in the thymus is consistent with the expression pattern observed in cultured thymocytes
The pattern of gene expression obtained with cultured thymocytes would suggest that in the very earliest stages of positive selection, before Egr-1 and CD69 up-regulation, there might be a rapid and transient up-regulation of Nur77. There are no markers that allow one to isolate thymocytes that have received positive selection signals in vivo within the time frame that we have analyzed in vitro. However, to test whether the predicted pattern of gene expression was consistent with that observed in the thymus, we stained thymocytes for cell surface CD4 and CD8 in conjunction with internal staining for Egr-1 or Nur77. We focused on the DP thymocyte subpopulation, containing cells that have most recently received positive selection signals. Consistent with our previous finding (16), ∼5–10% of DP thymocytes express Egr-1 (Fig. 6). Up-regulation of Egr-1 in the thymus is a consequence of selection events (16), as further evidenced here by the reduced frequency of Egr-1+ DP thymocytes in TCR α-chain-deficient mice (Fig. 6). Nur77 is detectable in significantly fewer DP thymocytes than is Egr-1 (Fig. 6). However, like Egr-1, the frequency of Nur77+ thymocytes is reduced in TCR α-chain-deficient mice (Fig. 6). These results are consistent with the proposed transient expression of Nur77 in cells that have just initiated positive selection.
Previous studies using transgenic mice or fetal thymic organ culture have implicated the Ras/MAP kinase pathway in positive selection based upon inhibition of single positive thymocyte production under conditions in which this signaling pathway was inhibited (5, 6, 22, 23). The process of positive selection encompasses both the induction of survival and differentiation of immature thymocytes. Whether the Ras/MAP kinase pathway is obligatory for cell differentiation and/or cell survival was not clear from these former studies. We have previously shown that expression of a dominant negative mutant of Ras in a thymocyte cell line inhibits early cellular changes associated with positive selection, implicating the Ras/MAP kinase pathway in the differentiation aspect of positive selection (18). The results presented in this study have also shown that MEK activation, and by inference MAP/ERK kinase activation, is required for differentiation of isolated thymocytes as assessed by early changes in gene expression and cell surface changes in coreceptors. Moreover, we have shown that MEK activation must be sustained for at least 8 h for efficient generation of CD4 lineage thymocytes. In total, the data strongly implicate the Ras/MAP kinase pathway in regulating thymocyte differentiation.
Our data also suggest that in addition to cell differentiation, the MAP kinase signaling pathway may be required for regulating the survival of DP thymocytes undergoing selection. Thus, there may be significant overlap in the signals involved in these two facets of positive selection. Bcl-2 has been reported to be up-regulated in thymocytes that have been stimulated with concentrations of PMA and ionomycin that induce differentiation (8), and this is a possible mechanism by which signaling via the Ras/MAP kinase pathway also regulates survival. It remains to be determined whether this pathway also regulates thymocyte survival in vivo.
Sustained MAP kinase activation is required for nuclear translocation and subsequent induction of cellular differentiation in other systems (32) and may be a common regulatory mechanism. It is also possible that extended activation of this pathway is required to trigger successive waves of expression of transcriptional regulators and downstream targets. The delayed and transient down-regulation of coreceptor genes may fall into this category. However, we have also shown that conditions that are optimal for differentiation result in surprisingly slow accumulation of MAP kinase activity. This gradual increase in enzymatic activity may partly explain the necessity for sustained activation of MEK to induce cell differentiation in this instance. In contrast, conditions that induce cell death result in rapid induction of high level MAP kinase activity. The distinct pattern of activation of this downstream signaling pathway during differentiation may be one way in which TCR affinity is ultimately coupled to cell fate. Moreover, if different genes have different thresholds for activation, then low level MAP kinase activity induced during positive selection may only activate a specific subset of potential target genes. Egr-3 is a case in point, where under conditions that elicit differentiation of thymocytes, there is low level expression of this gene.
Recently, it has been demonstrated that the MAP kinase pathway functions as an on-off switch during oocyte maturation (33). In this system, increasing the concentration of agonist results in an apparently graded response that is actually a result of an increasing number of responding cells and not changes in the magnitude of the response of individual cells. In contrast to this, the data presented in this study point to slow accumulation of MAP kinase activity on a per cell basis. To verify that the changes in gene expression observed by RT-PCR actually reflect changes in the response of individual cells, the induction of cell surface CD69 expression, which is Ras dependent (27, 28), was studied. The results show that all thymocytes are responsive to stimulation as assessed by induction of CD69, regardless of the strength of the stimulus. Importantly, the data also show that the kinetics of CD69 induction is dependent on the strength of the signal. These results are entirely consistent with the observed effects of PMA concentration on the temporal pattern of CD69 gene expression, and they support the interpretation that during thymocyte differentiation there is slow accumulation of active MAP kinase on a per cell basis. Thus, rather than an on-off switch, we favor a rheostat model of downstream signaling that allows the weak signals of positive selection to engage a very specific program of gene expression via “low and slow” MAP kinase activation. Certainly, it remains to be determined whether the temporal regulation of the MAP kinase pathway by PKC activation in this model system is a close mimic of TCR-mediated activation during positive selection in vivo. However, given the stringent requirements for inducing the differentiation of isolated thymocytes, it is reasonable to expect that the pattern generated under these conditions is similar to that induced during selection.
Our results also suggest a novel mechanism, in addition to strength of signal, by which modulation of a single downstream signaling pathway can differentially affect gene expression. Thus, the kinetics of MAP kinase activation may control the expression pattern, and thus temporal overlap, of transcription factors such as Nur77 and Egr-1 that play critical roles in regulating cell survival and cell differentiation. We have found that conditions that favor thymocyte death in culture induce a rapid up-regulation of Nur77 and Egr-1, and expression of the former is sustained by a MEK-dependent mechanism. In sharp contrast, differentiation is associated with transient expression of Nur77 and delayed, but interestingly not reduced, up-regulation of Egr-1. In vivo, a significantly smaller number of DP thymocytes express Nur77 than Egr-1, consistent with a transient up-regulation of Nur77 and a more sustained up-regulation of Egr-1 during the initial phase of positive selection as well. The cell fate of DP thymocytes that express Nur77 as a consequence of TCR activation is not known, and we cannot rule out the possibility that that these cells are in the process of negative selection. However, we also detect a similar pattern of up-regulation of Nur77 in AND TCR transgenic DP thymocytes (34) on an H-2b selecting background (data not shown). In addition, data suggest that the thymic cortex, where DP thymocytes reside, is not the major site of negative selection (35). Thus, the data are consistent with expression of Nur77 in DP thymocytes that have just initiated positive selection. Although the downstream targets of Egr-1 in the thymus remain to be identified, overexpression of this transcription factor can influence the threshold of activation of DP thymocytes during positive selection (36).
The importance of regulating the temporal patterns of expression of transcription factors may be enhanced by the potential for direct or indirect interactions between them. In this regard, Egr-1 has been implicated in delayed early expression of Nur77 (37). Low affinity/avidity interactions that are essential for positive selection may therefore promote a particular ordered sequence of gene expression as a consequence of both the level of activity and the temporal pattern of activation of downstream signaling pathways.
We thank Elyssa Rubin for additional technical assistance.
This research was supported by National Institutes of Health Grants AI-33219 and AI-31231 to J.K. This is manuscript 11806-IMM from The Scripps Research Institute.
Abbreviations used in this paper: DP, CD4+CD8+ double positive; Egr, early growth response; ERK, extracellular regulated kinase; MAP kinase, mitogen-activated protein kinase; MEK, mitogen-activated ERK-activating kinase.