Extracellular Signal–Regulated Kinase (Erk) Activation by the Pre-T Cell Receptor in Developing Thymocytes in Vivo

RAG-2^−/− mice were bred and maintained in our animal facility 34. Timed-pregnant RAG-2^−/− mice were used on day 14 of gestation.

The SL-12β.12 cell line is a pre-T cell line derived from a spontaneous SCID mouse–derived thymoma that stably expresses functionally rearranged TCR β chain at the cell surface with endogenous pTα to form a pre-TCR 48,49. The cell line was maintained in complete media (HG-DME media containing 10% FBS, 1% penicillin/streptomycin, 1% glutamine, 1% sodium pyruvate, 1% Hepes, 0.1% 2-ME, and 0.1% gentamicin [GIBCO BRL] supplemented with 0.5 mg/ml geneticin [GIBCO BRL]). Cells were kept in a humidified atmosphere of 5% CO₂ at 37°C.

SL-12β.12 cells (5 × 10⁶) were incubated at 4°C with 10 μg/ml biotinylated anti–TCR-β mAb (H57-597) for 30 min in DMEM supplemented with 0.1% BSA. Cells were then washed, resuspended in 100 μl DMEM/0.1% BSA containing 25 μg/ml avidin (Sigma Chemical Co.), and incubated at 37°C. Stimulation was stopped by the addition of 3 ml ice cold PBS followed by lysis in buffer containing 50 mM Tris, pH 7.4 (Boehringer Mannheim), 0.5% Triton X-100 (Boehringer Mannheim), 150 mM NaCl (ICN Biomedicals Inc.), 1 mM EDTA (BDH Chemicals Ltd.), 1 mM sodium orthovanadate (ICN Biomedicals Inc.), and Complete Inhibitor Cocktail (Boehringer Mannheim). Control cells (unstimulated) were treated similarly; however, avidin was not added before lysis. Lysates were resolved on 10% SDS-PAGE under reducing conditions and transferred onto polyvinylidene difluoride membranes (Immobilon-P; Amersham Pharmacia Biotech). Phosphorylated ERK-1 (p44 MAPK) and ERK-2 (p42 MAPK) and total ERK-2 proteins were detected using phospho-p44/22 MAPK (Thr202/Tyr204) mAb (New England Biolabs Inc.) and total p42 MAPK antibody (New England Biolabs Inc.), followed by horseradish peroxidase–conjugated goat anti–rabbit IgG (Santa Cruz Biotechnology). Blots were revealed using Supersignal West Pico chemiluminescent substrate (Pierce Chemical Co.).

All electroporations were carried out using a BTX Electro Cell Manipulator 600 apparatus. SL-12.β12 cells were electroporated at 450 μF, 186 Ω, 300 V, attaining a time constant of ∼30 ms. 3 × 10⁶ cells were transfected with desired plasmid DNA (up to 40 μg), as indicated in the figure legends. Each sample was transfected with PathDetect reporter plasmids (Stratagene, Inc.; as indicated in figure legends) and 1 μg plasmid encoding β-galactosidase (pCMV-β-gal). The addition of a fixed amount of β-galactosidase plasmid allowed for the control of transfection efficiency during the experiment. The β-galactosidase activity was used to index the luciferase signal detected, as we were able to assay for luciferase and β-galactosidase activity within the same sample 50. In brief, cells were washed in electroporation media (RPMI 1640 containing 20% FCS) and resuspended at 1.2 × 10⁷ cells/ml (250 μl per transfection). DNA and cells were combined in 4-mm sterile cuvettes (Bio-Rad Labs.) and incubated on ice for 10 min. The cells were electroporated with the conditions noted above and incubated on ice for a further 10 min. Transfected cells were put into fresh complete media and incubated for 24 h at 37°C, with or without the addition of exogenous stimuli, as indicated in the figure legends. The cells were then lysed and analyzed for luciferase and β-galactosidase activity.

Fetal thymic lobes from timed-pregnant RAG-2^−/− mice at day 14 of gestation were placed on nucleopore filters resting on Gelfoam rafts (Upjohn) soaked in fetal thymic organ culture (FTOC) media (complete media supplemented with a further 5% FCS; five to six lobes per filter) 51. The thymic lobes were incubated at 37°C for 6 h in a humidified incubator with 5% CO₂ to allow them to adhere to the filters. After this time, the filters were briefly removed from the Gelfoam rafts, and thymi were immediately transfected by gene gun bombardment with a Helios gene gun set at 200 psi (Bio-Rad Labs.), with the desired DNA bound to gold particles (1 μg gold) (as indicated in the figure legends; one cartridge per set of thymic lobes). The amount of DNA loaded per microgram of microcarriers is referred to as the DNA loading ratio (DLR; see figure legends). The filters carrying the transfected thymic lobes were then placed back on the Gelfoam rafts and incubated at 37°C for the indicated time (see figure legends). For short-term biochemical readout assays, the thymic lobes were incubated for a further 16–20 h, and then a single-cell suspension was prepared and the cells were lysed and assayed for luciferase and β-galactosidase activities as described below. For long-term developmental progression studies, the transfected thymic lobes were incubated for 7 d, with the FTOC media being changed once during the incubation. After this time, a single-cell suspension of the lobes was prepared, and the thymocytes were analyzed by flow cytometry as described below.

FITC-, PE–, and APC-conjugated anti–mouse CD4, CD25, and CD8 were used for flow cytometric analysis (purchased from PharMingen). Staining of cells was carried out as described previously 51. In brief, thymic cell suspensions were washed in FACS buffer (HBSS containing 1% BSA and 0.1% sodium azide) and incubated with antibodies (1:300 dilution) for 30 min on ice. The cells were then washed in FACS buffer before analysis. Stained cells were analyzed with a FACSCalibur™ flow cytometer (Becton Dickinson). Analysis was performed using CELLQuest software (Becton Dickinson); the data were live gated by size and lack of propidium iodide uptake.

Cells transfected with the PathDetect reporter plasmids (Stratagene Inc.) were assayed for luciferase and β-galactosidase activities using the Dual-Light reporter gene assay system (Tropix; Perkin Elmer-Applied Biosystems) as previously described 50. In brief, the cells were lysed in lysis buffer (40 mM tricine, pH 7.8, 50 mM NaCl, 2 mM EDTA, 1 mM MgSO₄, 5 mM dithiothreitol, and 1% Triton X-100). Supernatant (25 μl) was combined with an equal volume of luciferase reaction buffer (30 mM tricine, pH 7.8, 3 mM ATP, 15 mM MgSO₄, 1 mM coenzyme A, and 10 mM dithiothreitol), and after addition of 1 mM luciferin, the sample was immediately assayed for luciferase activity with a Lumat LB 9507 Luminometer (Fisher Scientific Co.). To assay for β-galactosidase activity, Galacton-Plus (substrate for β-galactosidase; Tropix) was added to each tube after the luciferase assay was completed. The tubes were incubated for 30–60 min at room temperature, and then Accelerator-II (Tropix) was added to each tube and the samples were immediately assayed for β-galactosidase activity, measured as light emission with the Lumat LB 9507 Luminometer. Results represent the average luciferase activity indexed for β-galactosidase activity.

To aid in deciphering the signaling pathways that lead to the differentiative and proliferative events of β-selection, the phosphorylation status of ERK-1/2 was investigated by performing Western blot analyses using an antiphospho-ERK-1/2 mAb. A SCID mouse–derived pre-T cell line, SL-12.β12, preincubated with biotinylated anti–TCR-β (H57-597) mAb, was stimulated for 1–30 min upon cross-linking with avidin. Engagement of the pre-TCR complex resulted in a rapid and transient phosphorylation of ERK-1/2 (Fig. 1 A). Phosphorylation was observed as early as 1 min but declined almost to background levels 5 min after stimulation; subsequently, a secondary, weaker level of phosphorylation was evident between 10 and 30 min (Fig. 1 A). To control for protein loading, the total amount of ERK-2 was assessed by Western blot with an anti–ERK-2 antibody (Fig. 1 B).

To study whether the decline in phosphorylation of ERK-1/2 at 5 min was specific to pre-TCR activation, we performed a similar experiment with variants of the SL12 cell line expressing comparable levels of TCR β chain with either a TCR α chain (mature TCR) or a pTα chain (pre-TCR), as determined by flow cytometry (data not shown). The TCR β chain on both cell lines was cross-linked for 1–30 min, and the phosphorylation status of ERK-1/2 was determined. In both cases, there was an attenuation of ERK-1/2 phosphorylation levels at 5 min, indicating that this phosphorylation pattern is characteristic of the SL-12 cell line (data not shown). These results indicate that engagement of the pre-TCR on SL-12β.12 cells leads to the phosphorylation of ERK-1/2, demonstrating the activation of the Ras/Raf/MEK/ERK signaling cascade.

We sought to determine if the phosphorylation of ERK-1/2 observed upon cross-linking of the pre-TCR (Fig. 1 A) leads to their activation. To this end, we employed a reporter plasmid system (described in Materials and Methods) that allows detection of ERK-1/2 activation within the cell. This system utilizes two plasmids: a fusion activator plasmid (pFA-Elk), which encodes for the transactivation domain of the ERK substrate, Elk-1 (residues 307–427), fused with the DNA-binding domain of GAL4; and a luciferase reporter plasmid (pFR-Luc), which encodes for the luciferase gene under the control of five GAL4 binding elements. Hence, in transfected cells, phosphorylation of the Elk-1 fusion protein can be read in the form of luciferase activity. Accordingly, we transfected SL-12β.12 cells with pFR-Luc, either alone or together with pFA-Elk. The transfected cells were then stimulated with immobilized anti-TCR-β mAb or with PMA and ionomycin.

Cells transfected with pFR-Luc alone displayed background luciferase activity (31 ± 18 relative light units [RLUs]). However, luciferase activity increased almost 850-fold above background levels in cells transfected with pFR-Luc and pFA-Elk (26,221 ± 1,119 RLUs; Fig. 2). This elevation in luciferase activity may reflect low-level constitutive ERK activity within SL-12β.12 cells (Fig. 1 A), as it was blocked by the addition of the MEK1 inhibitor, PD98059 (data not shown). Importantly, anti–TCR-β mAb stimulation of pFR-Luc/pFA-Elk–transfected SL-12β.12 cells resulted in a fivefold increase in luciferase activity compared with unstimulated cells (Fig. 2 A). This elevation in luciferase activity was abrogated by the addition of PD98059 (129% inhibition). These data indicate that engagement of the pre-TCR complex results in the activation of ERK and subsequent phosphorylation of the ERK-1/2 substrate, Elk-1 (Fig. 2 A).

Activation of pFR-Luc/pFA-Elk–transfected cells by the addition of PMA/ionomycin induced a 20-fold stimulation in luciferase activity over unstimulated cells (Fig. 2 B); again, addition of PD98059 curtailed this activity (data not shown). As a control to demonstrate that ERK-1/2 activation can be efficiently detected with these reporter plasmids, pFR-Luc/pFA-Elk–transfected cells were also transfected with a plasmid encoding a constitutively active MEK1, pFC-MEK1. This serine/threonine kinase is responsible for the phosphorylation and activation of ERK-1/2. These cells showed maximal luciferase activity irrespective of PMA/ionomycin treatment (Fig. 2 B), indicating that this reporter system provides a highly sensitive method for detecting ERK activation.

Taken together, these results confirm that the ERK signaling cascade is activated upon engagement of the pre-TCR, as shown by the phosphorylation of ERK-1/2 (Fig. 1 A) and the phosphorylation/activation of the ERK-1/2 substrate, Elk-1 (Fig. 2 A), in the SL-12β.12 cell line.

An important application of the above reporter plasmid system would be to detect specific pre-TCR–mediated signals as they occur within an intact thymus. Therefore, an accelerated DNA/particle bombardment delivery system (gene gun) was adapted to transfect thymocytes in FTOC 52. This approach allows transfected cells to respond to normal microenvironmental stimuli encountered during differentiation within the thymic milieu 52,53. We have previously used this method to study transcriptional regulation events in the developing fetal thymus 52.

The preservation of structural and functional integrity of gene gun–transfected thymi was tested by allowing FTOCs to complete a program of β-selection. Thus, RAG-2^−/− mouse–derived FTOCs were gene gun transfected with control DNA and cultured for 7 d, either alone (Fig. 3 A, Control) or in the presence of anti-CD3 mAb (Fig. 3 A, anti-CD3). Transfected FTOCs were then analyzed by flow cytometry for evidence of differentiation to the DP stage 17,18,19. This analysis revealed that the thymocytes from anti-CD3–treated RAG-2^−/− gene gun–transfected FTOCs progressed to the DP stage, as indicated by upregulation of CD8/CD4 (Fig. 3 A, anti-CD3). This differentiative event involved the loss of CD25 cell surface expression and proceeded through a CD8⁺ immature SP stage (Fig. 3 A, anti-CD3). These results indicated that thymocytes and stromal elements within gene gun–transfected thymic lobes remain viable.

To determine the long-term developmental capacity of transfected thymocytes, RAG-2^−/− thymic lobes were gene gun transfected with plasmids encoding either constitutively active Lck (Lck[F505]) or a productively rearranged TCR β chain. It has been previously shown in RAG^−/− mice that introduction of transgenes encoding a kinase-active Lck or a functionally rearranged TCR β chain induced thymocyte proliferation and differentiation toward the DP stage of development 39,41. Indeed, these events were also observed in gene gun–transfected RAG-2^−/− FTOCs (Fig. 3 B). Flow cytometric analysis of Lck(F505)- and TCR-β–transfected FTOCs after 7 d revealed that thymocytes differentiated from the DN stage to the DP stage, as evidenced by upregulation of CD4 and CD8 surface expression and downregulation of CD25 surface expression (Fig. 3 B). The apparent difference in the levels of CD8 surface expression in active Lck-transfected FTOCs was due to the fact that the experiment shown was analyzed separately and a different set of fluorochrome-labeled antibodies was used. These phenotypic changes are consistent with differentiation events that occur after completion of β-selection (Fig. 3 B; reference 5, 10–13).

Our finding that gene gun transfection of RAG-2^−/− FTOCs with a functional TCR β chain resulted in the normal developmental progression of DN thymocytes indicated that a functional pre-TCR complex was generated 39. These data imply that after gene gun transfection, the formation of the pre-TCR results in signaling events that occur during normal β-selection in vivo.

Because the cellular signaling processes that underlie β-selection are not well defined, we combined gene gun–mediated transfection of FTOCs with the reporter plasmid system. In this way, the intracellular signaling cascades that occur after pre-TCR formation can be monitored in a biological time frame and within a relevant in vivo setting. We hypothesized that gene gun–mediated transfection of RAG-2^−/− mouse–derived FTOCs with a plasmid encoding a functionally rearranged TCR β chain, together with reporter plasmids, would allow us to capture de novo β-selection signaling events that occur after the generation of the pre-TCR.

Our results show that compared with FTOCs transfected with pFR-Luc alone (Fig. 4), there was a slight elevation in luciferase activity in FTOCs transfected with pFR-Luc/pFA-Elk (6 ± 1.8 vs. 59 ± 17 RLUs; Fig. 4). This level of luciferase activity may represent endogenous ERK-1/2 activity present in transfected FTOCs.

To directly test our hypothesis, RAG-2^−/− FTOCs were gene gun transfected with reporter plasmids (pFR-Luc/pFA-Elk) together with a plasmid encoding a functional TCR β chain. Our analysis revealed an almost sixfold increase in luciferase activity in FTOCs transfected with a TCR-β plasmid as compared with FTOCs transfected with the reporter plasmids alone (333 ± 56 RLUs vs. 59 ± 17; Fig. 4, TCR-β). A substantial luciferase activity was also observed in FTOCs transfected with a constitutively active MEK1 plasmid together with the reporter plasmids (9,919 ± 1,933 RLUs; Fig. 4, MEK1). The level of stimulation observed in MEK1-transfected FTOCs demonstrates the sensitive nature of this assay.

Several reports have suggested that the Ras/Raf/MEK/ERK signaling cascade is involved at the β-selection stage 13,42,43,54. However, our data provides the first direct evidence for coupling of the pre-TCR complex to the ERK signaling cascade in vivo.

Lck is thought to play an essential role as a mediator of β-selection events 15,17,24,41,55. Therefore, we investigated whether Lck could activate the ERK signaling cascade in FTOCs. To this end, FTOCs were gene gun transfected with a kinase-active Lck together with the reporter plasmids. Fig. 5 shows that a sixfold increase in luciferase activity was observed in FTOCs transfected with constitutively active Lck (387 ± 75 RLUs; Fig. 5). This level of luciferase activity is comparable to the stimulation observed in TCR-β–transfected FTOCs (Fig. 5, TCR-β).

To address whether the observed ERK activation after the pre-TCR complex formation was dependent on Lck activity, we took advantage of a plasmid encoding a dominant negative Lck (Lck[R273]). This mutant form of Lck has been shown to cause a dramatic block in T cell development at the DN stage 15. Fig. 5 shows that transfection of RAG-2^−/− FTOCs with the reporter plasmids together with plasmids encoding a TCR β chain and a kinase-negative Lck led to an attenuation of luciferase activity, as compared with FTOCs transfected with TCR-β alone (84 ± 56 vs. 333 ± 56 RLUs; Fig. 5).

These results indicate that expression of the pre-TCR complex by developing thymocytes activates the ERK signaling cascade and that this activation is directly dependent on Lck during thymocyte differentiation within the thymus.

In this study, we demonstrate that formation and/or engagement of the pre-TCR complex results in the activation of the ERK signaling cascade. We have shown this in two distinct systems: (a) in a SCID mouse–derived pre-T cell line, SL-12β.12 48, in which engagement of the pre-TCR leads to phosphorylation and activation of ERK-1/2; and (b) within the thymus, where formation of the pre-TCR in developing thymocytes induces ERK activation and phosphorylation of the downstream substrate, Elk-1. Furthermore, our findings confirm the central role of the nonreceptor tyrosine kinase Lck in β-selection by demonstrating that ERK activation is strictly dependent on Lck function in vivo.

The ability to introduce reporter plasmids into developing thymocytes in vivo provides an important diagnostic tool for the detection of signaling events during T cell development in the thymus (Fig. 3 and Fig. 4). The significance of this method is highlighted by the fact that, in the absence of the thymic microenvironment (i.e., cell suspension), thymocytes fail to respond to normal developmental cues and undergo rapid programmed cell death 56,57. Indeed, transfection of thymocytes from RAG^−/− mice with reporter plasmids revealed that cells in suspension fail to respond when cotransfected with a plasmid encoding a TCR β chain (data not shown). Thus, thymocytes in cell suspension appear to be unresponsive to pre-TCR–mediated signals, possibly due to the lack of stromal elements to assist in this signaling. These results are consistent with previous data showing that anti-CD3 mAb–mediated differentiation of RAG^−/− thymocytes to the DP stage in FTOC does not occur when these thymocytes are treated in cell suspension 58.

The notion that the Ras/Raf/MEK/ERK signaling cascade plays a central role during β-selection is supported by data obtained from mice bearing targeted genetic deficiencies of key signaling molecules such as Lck/Fyn, ZAP-70/Syk, LAT, or SLP-76 24,25,26,28,29,30,31. These mice displayed a block in T cell development at the DN stage, suggesting that these molecules are essential for differentiation to the DP stage of development 24,25,26,28,29,30,31. However, several contradicting reports using transgenic mice carrying either constitutively active or dominant negative members of these signaling molecules have left the involvement of this pathway unresolved. Studies introducing constitutively active Lck, Ras, and Raf into RAG^−/− mice also indicate that the Ras/Raf pathway is crucial at this stage of development 41,42,43. In contrast, studies using transgenic mice carrying dominant negative forms of Ras (Ha-Ras^N17) 45, Raf 46, and MEK-1 47 failed to support a role for these signaling molecules during β-selection. In this regard, it is important to note that mice carrying a dominant negative Lck transgene displayed varying degrees of DN to DP thymocyte development arrest directly correlating with the levels of transgene expression 15. Thus, the conflicting results obtained using dominant negative molecules may reflect their ability to be sufficiently expressed at the appropriate stage of development. In addition, studies using retroviral infection of thymocytes with dominant negative MEK-1 showed a perturbation in the transition of thymocytes from the DN to DP stage, suggesting that the inhibitory effectiveness of dominant negative molecules appears to be dependent on the experimental system used 54.

Although several studies have indirectly supported the importance of the Ras/Raf/MEK/ERK-mediated signals downstream of the pre-TCR, in our system, the fact that the detection of ERK activation is predicated by the expression of the pre-TCR within individually transfected developing thymocytes directly links the pre-TCR to the activation of these kinases. Thus, this study provides the first evidence for the pre-TCR being responsible for the generation of specific downstream signals within the thymus (Fig. 4) and shows that Lck acts proximally to the nascent pre-TCR complex, leading to the subsequent activation of ERK (Fig. 5).

The transcription factor Elk-1 has been shown to be phosphorylated by ERK-1/2, although more recently it has been demonstrated that JNK (c-Jun NH₂-terminal kinase) also has the ability to phosphorylate and activate Elk-1 59,60,61. However, mouse models with gene-targeted disruptions of key signaling components of the JNK signaling pathway (JNK2 and stress-activated protein kinase kinase 1) did not reveal any abnormalities in β-selection events 62,63,64. In this regard, our data with the SCID mouse–derived pre-T cell line SL-12β.12 show that engagement of the pre-TCR results in ERK phosphorylation and activation of the ERK substrate, Elk-1 (Fig. 1 and Fig. 2), and that this can be blocked by addition of the MEK1 inhibitor, PD98059. Together, these data indicate that, in this setting, Elk-1 is specifically activated by ERK.

Pre-TCR signals can be experimentally bypassed by introducing constitutively active mutants of certain signaling components, inducing β-selection–mediated events. Interestingly, the introduction of an activated form of Ras (Ha-Ras^V12) or Raf-1 (Raf-CAAX), although allowing the differentiation of RAG^−/− thymocytes to the DP stage, failed to establish allelic exclusion at the TCR-β locus when introduced into wild-type mice 43,65. This indicates that although certain β-selection outcomes are mediated by the Ras/Raf pathway, distinct pre-TCR–derived signals upstream of Raf are responsible for the control of allelic exclusion. Therefore, it is interesting to note that RAG^−/− FTOCs retrovirally infected with constitutively active MEK1 do not differentiate to the DP stage of development 54. This finding suggests that Raf may activate additional signaling pathways enabling developing thymocytes to undergo full differentiation to the DP stage. Therefore, the intrathymic signal detection system developed here will allow for detailed studies of distal signaling events activated by the pre-TCR, facilitating the elucidation of potential signaling branchpoints.

We would like to thank Drs. Pamela Ohashi, Michael Julius, and James Carlyle for their critical review of the manuscript, P. Ohashi and Sanjeev Mariathasan for the gift of antibodies and advice with Western blotting, André Veillette for his kind gift of the plasmids encoding constitutively active and dominant negative Lck, and Michael Lenardo for his gift of the plasmid encoding a functionally rearranged TCR β chain.

A.M. Michie is supported by a postdoctoral fellowship from the Cancer Research Institute. S. Trop is supported by a Doctoral Research Award from the Medical Research Council of Canada (MRC). J.C. Zúñiga-Pflücker is supported by a scholarship from the MRC. This work was funded by grants from the Leukemia Research Fund of Canada and the MRC.