Cyclic Adenosine 5′-Monophosphate Response Element Binding Protein Plays a Central Role in Mediating Proliferation and Differentiation Downstream of the Pre-TCR Complex in Developing Thymocytes¹

The developmental processes that shape the peripheral pool of mature αβ T lymphocytes normally occur within the thymus. Developing thymocytes must pass a series of rigorous checkpoints that ensure the emerging cells are programmed to recognize foreign peptides in the context of self MHC (1). The primary checkpoint encountered by CD25⁺CD117⁻ (DN3) thymocytes ensures that only cells that express a functionally rearranged TCR-β chain in the context of the pre-TCR complex can proceed to the next stage of development, termed β-selection (2). The key events that define the completion of this developmental checkpoint are: rescue from apoptosis; differentiation (down-regulation of CD25; up-regulation of the coreceptors CD4 and CD8 to form CD4⁺CD8⁺ double-positive (DP)³ cells); proliferation; allelic exclusion at the TCR-β gene locus; and initiation of TCR-α rearrangement. The signaling pathways that mediate this developmental program remain largely undefined. However, transgenic and knockout mouse models, as well as detailed biochemical analyses of primary thymocytes and pre-T cell lines, have revealed that the signaling mechanisms that control transition through β-selection, downstream of pre-TCR formation, are similar to the signaling pathways that are activated downstream of mature TCR ligation. Therefore, upon pre-TCR formation, Lck phosphorylates the TCR-ζ complex, resulting in the recruitment of ZAP-70 to the pre-TCR complex; ZAP-70 phosphorylates the adapter molecules, linker for activation of T cells and Src homology 2 domain-containing leukocyte protein of 76 kDa, which enables the recruitment of further signaling molecules that enable the activation of the ERK-MAPK- and protein kinase C (PKC)-mediated signaling pathways (2, 3). Although PKC and ERK-MAPK signaling cascades are important for proliferation and differentiation of pre-T cells during β-selection, our recent findings revealed that PKC is additionally responsible for the inhibition of rearrangement at the TCR-β gene locus, resulting in allelic exclusion (4, 5, 6, 7).

Although studies have focused on elucidating the proximal signals that control β-selection, little is known about how these signals are transmitted into the nucleus of pre-T cells to drive this developmental program. As our previous studies established PKC activation as an essential event for the completion of β-selection (5), we were interested in defining which signaling components could mediate PKC activation. Therefore, we have investigated the role of a downstream target of PKC (8), the ubiquitous transcription factor CREB, during this stage of development. CREB has been implicated in the control of differentiation and proliferation in numerous cell types, including endocrine cells and mature T and B lymphocytes (9, 10, 11). Several key genes up-regulated during β-selection have cAMP response element (CRE) sites on their enhancer-promoter region that are regulated by CREB activation, such as TCR-α, TCR-β, Bcl-2, CD3-δ, and CD8 (12, 13, 14, 15). In addition, T cell development in the CREB^null mouse model system, which lacked expression of all functional CREB isoforms (CREBα, β, and δ), was impaired with an overall reduced thymic cellularity, a decrease in DP thymocyte development, and a concurrent increase in the percentage of double-negative (DN) thymocytes, suggesting a defect during β-selection (16). Therefore, we hypothesized that CREB would be a likely candidate to be activated downstream of pre-TCR formation and involved in regulating key events during β-selection.

We now show that CREB is constitutively phosphorylated within developing thymocytes and that the generation of pre-TCR-derived signals leads to an up-regulation of CREB phosphorylation. Our studies reveal that CREB activation downstream of the pre-TCR is entirely mediated by PKCα, and partially dependent on the activation of the ERK-MAPK signaling cascade. To investigate the importance of this CREB activation, we enforced expression of inducible cAMP early repressor (ICER), a naturally occurring inhibitor of CREB function (17) in fetal liver (FL)-derived lymphoid progenitor cells and addressed their developmental potential toward the T lymphocyte lineage by placing them in fetal thymic organ culture (FTOC) (18). Flow cytometric analysis of FTOCs reconstituted with ICER-expressing FL revealed a delay in the generation of CD4⁺CD8⁺ DP thymocytes compared with control FTOCs. In addition, there was a decrease in cellularity of FTOCs reconstituted with ICER-expressing FL cells, indicating that CREB plays an important role during β-selection. Collectively, these studies establish CREB as a key component of the signaling network that controls the proliferation and differentiation downstream of the pre-TCR complex in developing thymocytes.

ICR mice were purchased from Harlan-Olac (Oxon, U.K.) and maintained at the University of Glasgow Central Research Facilities (Glasgow, U.K.). RAG° mice were bred and maintained in-house (19). Time-pregnant CD1 mice were purchased from Charles River Laboratories (Wilmington, MA). Time-pregnant mice were generated, and liver and thymus were extracted from fetuses at day 14 of gestation. All animals were maintained in accordance with local and home office regulations. The SL-12β.12 cell line is a pre-T cell line derived from a spontaneous SCID mouse-derived thymoma, which stably expresses functionally rearranged TCR-β chain at the cell surface with endogenous pTα to form the pre-TCR (20). The cell line was maintained in complete medium (high glucose DMEM containing 10% FBS (BioWhittaker, Wokingham, U.K.), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 50 μM 2-ME, and 10 μg/ml gentamicin (Invitrogen Life Technologies, Paisley, U.K.)) supplemented with 0.5 mg/ml geneticin (Invitrogen Life Technologies). Cells were kept under a humidified atmosphere of 5% CO₂, at 37°C.

SL-12β.12 cells were incubated at 4°C with 10 μg/ml biotinylated anti-TCR-β mAb (H57-597; BD Biosciences, Oxford, U.K.) for 30 min in HBSS (Invitrogen Life Technologies) supplemented with 0.1% BSA (Sigma-Aldrich, Poole, U.K.). The cells were then washed and resuspended at a concentration of 5 × 10⁶ cells. Cells were stimulated by adding avidin (final concentration 25 μg/ml; Sigma-Aldrich) and incubated at 37°C for the indicated time points (see figure legends for details) (6). Stimulation was stopped by the addition of ice-cold PBS containing 1 mM Na₃VO₄, and cells were prepared for Western blotting. Control cells (unstimulated) were treated similarly. RAG° fetal thymuses (FT) were incubated at 37°C in Terisaki plates with 10 μg/ml anti-CD3 mAb (145-2C11; BD Biosciences) for the indicated time points (see figure legends for details). Stimulation was stopped by washing the FTs in ice-cold PBS containing 1 mM Na₃VO₄ (Sigma-Aldrich), and thymocytes were prepared for Western blotting (see below).

Prepared cell samples were immediately lysed in buffer containing 50 mM Tris, pH 7.4, 0.5% Triton X-100 (Roche Diagnostic Systems, Lewes, U.K.), 150 mM NaCl, 1 mM EDTA, 1 mM Na₃VO₄, and complete inhibitor mixture (Roche Diagnostic Systems). Lysates were resolved on NuPAGE 4–12% Bis-Tris gels (Invitrogen Life Technologies) under reducing conditions, and transferred onto polyvinylidene difluoride (PVDF) membranes (Amersham Biosciences, Little Chalfont, U.K.). Phosphorylated CREB proteins were detected using anti-phospho-CREB (Ser¹³³) mAb (Cell Signaling Technology, Hitchin, U.K.), and protein-loading controls were conducted using either anti-CREB Ab or anti-ERK Ab (Cell Signaling Technology), followed by HRP-conjugated goat anti-rabbit IgG (Cell Signaling Technology). Blots were revealed using Supersignal West Pico chemiluminescent substrate (Perbio, Tattenhall, U.K.) (6).

All electroporations were conducted using a Bio-Rad Gene Pulser II (Bio-Rad, Hemel Hempstead, U.K.). SL-12.β12 cells were electroporated at 960 μF, 260 V, attaining a time constant of 50–55 μs. Up to 3 × 10⁷ 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, Amsterdam, The Netherlands), as indicated in figure legends, and 2 μg of 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 (see below) (6). Cells were washed in electroporation medium (RPMI 1640 containing 20% FCS) and resuspended at 12 × 10⁷ cells/ml (250 μl/transfection). DNA and cells were combined in 4-mm sterile cuvettes (Bio-Rad) and incubated on ice for 10 min. The cells were electroporated with the conditions noted above and incubated on ice for an additional 10 min. Transfected cells were put into fresh complete medium and incubated for 24 h at 37°C, with or without the addition of exogenous stimuli or pharmacological inhibitors, as indicated in the figure legends. The cells were then lysed and analyzed for luciferase and β-galactosidase activity.

SL-12β.12 cells transfected with the PathDetect reporter plasmids were assayed for luciferase and β-galactosidase activities using substrates that generate light emission. The cells were lysed in lysis buffer (40 mM tricine, pH 7.8, 50 mM NaCl, 2 mM EDTA, 1 mM MgSO₄, 5 mM DTT, and 1% Triton X-100). Supernatant 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 DTT), and after addition of 1 mM luciferin (BD Biosciences), the sample was immediately assayed for luciferase activity with a Lumat LB 9507 Luminometer (Berthold Technologies, Bad Wildbad, Germany). To assay for β-galactosidase activity, Galacton-Star (substrate for β-galactosidase; Applied Biosystems, Bedford, U.K.) was diluted 1/50 with Galacton-Star reaction buffer diluent (Applied Biosystems) and added to each tube after the luciferase assay was completed. The tubes were incubated for 30–60 min at room temperature, and then the samples were 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.

Total RNA was isolated from the thymuses of RAG° mice up to 10 days after i.p. injection with 10 μg of anti-CD3 mAb/neonatal mouse using TRIzol RNA isolation reagent (Bio-Rad). cDNA from each sample was prepared from 1 μg of RNA using the first strand cDNA synthesis kit (Roche Diagnostic Systems). Semiquantitative PCR were performed using the same serially diluted cDNA batches shown for β-actin. Samples were amplified using a Proteus model thermal cycler (Helena BioSciences, Sunderland, U.K.). Products were separated by agarose gel electrophoresis and visualized using ethidium bromide staining. All PCR products correspond to the expected molecular sizes. Gene-specific primers used for PCR are as follows: ICER 5′, ATGGCTGTAACTGGAGATGA; ICER 3′, AAGGTCCAAATCAAACACAG; β-actin 5′, GATGACGATATCGCTGCGCTG; and β-actin 3′, GTACGACCAGAGGCA TACAGG.

Retroviral constructs were engineered by subcloning the gene of interest (ICER) into the retroviral backbone (MIEV), 5′ of the internal-ribosomal entry site, allowing the bicistronic expression of the gene of interest and GFP (21). The retroviral vectors were stably expressed in the retroviral packaging line, GP+E.86 (5, 22). The GP+E.86 cell lines were seeded at 2 × 10⁴ cells/well and mitomycin C (Sigma-Aldrich) treated for 2–3 h at 37°C (10 μg/ml) 1 day before retroviral infection of primary cells. Hemopoietic stem cell (HSC)-enriched FL cells were prepared by carrying out a CD24 (heat-stable Ag; BD Biosciences) Ab complement-mediated cell lysis on a FL single cell suspension for 30 min at 37°C. The red/dead cells were removed by density centrifugation over Lymphocyte-Mammal (VH Bio, Newcastle, U.K.) at 600 × g for 15 min at room temperature. The resultant HSC-enriched population was cocultured with the mitomycin C-treated retroviral packaging line in FTOC medium (complete medium supplemented with 3% FBS), 500 μg/ml polybrene (Sigma-Aldrich), and 10 ng/ml cytokines (IL-6, IL-7, and stem cell factor; Caltag-MedSystems, Towcester, U.K.) for 24 h before flow cytometric cell sorting and culture in FTOC.

R-PE-, allophycocyanin-, and biotin-conjugated anti-mouse Abs were used for flow cytometric analysis (BD Biosciences). To prepare cells for flow cytometric cell sorting, the retrovirally infected HSC-enriched cells were removed from the GP+E.86 cell line by gentle pipetting and washed in cell-sorting buffer (HBSS (Invitrogen Life Technologies) supplemented with 1% BSA (Sigma-Aldrich)). Cells were stained, then incubated with the appropriate Abs (diluted 1/300), as indicated in the figure legends, for 30 min at 4°C. The cells were then washed twice in cell-sorting buffer. When biotinylated conjugated Abs were used, the cells were incubated with streptavidin-allophycocyanin (diluted 1/300) for 20 min at 4°C. The cells were then washed twice in cell-sorting buffer, resuspended in cell-sorting buffer, and filtered through a 70-μm nylon mesh (BioDesign, Carmel, NY) just before cell sorting. GFP⁺ HSC populations were separated using a Coulter Elite (Beckman Coulter, Fullerton, CA) cytometer. Sorted cells were ∼98% pure, as determined by postsort analysis. For flow cytometric analysis, cells were stained, as described above, using the Abs described in the figure legends. However, the cells were prepared in FACS buffer (cell-sorting buffer supplemented with 0.1% sodium azide (Sigma-Aldrich)). The cells were acquired on a FACSCalibur flow cytometer using the CellQuest software package (BD Biosciences) to acquire and the FlowJo software package (Tree Star, Stanford, CA) to analyze the data. All data shown are live gated by size and lack of propidium iodide uptake.

To analyze the developmental potential of retrovirally infected HSCs, the sorted cells were washed twice in FTOC medium and then cultured with deoxyguanosine (dGuo)-treated FT (23). dGuo-treated FT were generated by incubating d14 FT with 1.1 mM dGuo for 5 days in FTOC configuration (FT are placed on a nucleopore filter (Merck, West Point, PA) at the air-liquid interphase in a humidified incubator, 5% CO₂ at 37°C). After 5 days in culture, the dGuo was removed from the medium and the cells were cultured for a maximum of 2 days before reconstitution with retrovirally infected HSCs. dGuo-treated FTs were seeded with 2000 GFP⁺CD117⁺ cells in Terisaki plates (Fisher Scientific, Loughborough, U.K.) and incubated in hanging drop for 24 h, then placed in FTOC for 7 or 9 days. After this time, a single cell suspension was generated, the cellularity was determined by trypan blue (Sigma-Aldrich) exclusion, and the cells were analyzed by flow cytometry, as indicated in the figure legends.

To address whether CREB is activated downstream of the pre-TCR complex, we first chose to analyze the pre-T cell line, SL-12β.12, which has been successfully used previously to demonstrate the activation of the ERK-MAPK signaling pathway downstream of pre-TCR complex ligation (6). Therefore, we cross-linked the pre-TCR complex on the surface of SL-12β.12 cells using anti-TCR-β (H57-597) mAb and determined the phosphorylation/activation status of CREB by carrying out Western blot analyses using an anti-phospho-CREB mAb (Ser¹³³). Upon stimulation, we observed a slow sustained increase in CREB phosphorylation above background levels, first detectable at 5 min after pre-TCR complex cross-linking, which was sustained for at least 60 min (Fig. 1,A and data not shown). Treatment of SL-12β.12 cells with forskolin (an adenylyl cyclase activator; 50 μM) provided a positive control to reveal maximal phosphorylation of CREB. To control for protein loading, the membranes were probed with an anti-ERK Ab (Fig. 1 A). These results establish that CREB is phosphorylated downstream of the pre-TCR engagement in SL-12β.12 cells.

To address whether CREB was activated downstream of the pre-TCR in developing thymocytes, we made use of the RAG° mouse model that displays an arrest in T cell maturation at the DN3 stage, before expression of the pre-TCR complex, due to an inability to initiate TCR-β gene rearrangement (19). Treatment of RAG°-derived thymocytes with anti-CD3 mAb results in a pre-TCR-like signal that enables DN3 cells to bypass the developmental block and undergo hallmark β-selection events, such as proliferation and differentiation to the DP stage (24, 25). Therefore, we incubated RAG°-derived FT with anti-CD3 mAb for up to 5 h (as indicated in the figure legends) and determined the phosphorylation status of CREB in the RAG° FT lysates by carrying out Western blot analyses using an anti-phospho-CREB mAb. Notably, CREB was phosphorylated in unstimulated RAG° FT (Fig. 1,B), indicative of CREB being constitutively active in developing DN thymocytes. In support of the above findings (Fig. 1,A), the level of CREB phosphorylation increased upon incubation of the RAG° FT with anti-CD3 mAb relative to the protein-loading controls (Fig. 1 B). Thus, our results indicate that CREB is phosphorylated downstream of pre-TCR-like signals in developing thymocytes.

To further investigate CREB activation/phosphorylation in developing thymocytes, we generated protein lysates of thymic tissue samples derived from wild-type mice (composed mainly of DP thymocytes) and RAG° mice (composed mainly of DN3 thymocytes (19)) and analyzed them by Western blotting using an anti-phospho-CREB mAb. Interestingly, phospho-CREB was detected in thymic tissue generated from both wild-type mice and RAG° mice, but not in lysates derived from wild-type spleen or fetal liver (Fig. 1,C and data not shown). These data indicate that CREB is constitutively active specifically in the thymus at both the DN and the DP stages of development. Additionally, the level of CREB phosphorylation observed in RAG° thymocytes was significantly lower than that observed in wild-type thymocytes (Fig. 1,C). These results imply that there is less CREB activity in the absence of a pre-TCR-derived signal in developing thymocytes. Indeed, this finding is supported by the above data showing that the level of CREB phosphorylation increases when RAG° FT are treated with anti-CD3 Ab (Fig. 1 B). Taken together, our studies establish that CREB is constitutively phosphorylated in developing thymocytes, and is activated downstream of the pre-TCR complex.

CREB phosphorylation can be mediated by a number of upstream signaling pathways in different cell types, such as protein kinase A (PKA)-, PKC-, and ERK-MAPK-mediated signaling cascades (8). PKA-mediated signaling pathways have been shown to induce apoptosis in developing thymocytes (26, 27), and result in a developmental block in developing thymocytes at the DN3 stage (28), suggesting that PKA is not responsible for activating CREB-mediated proliferative and differentiative signals downstream of the pre-TCR complex. These findings, coupled with the large body of evidence implicating the activation of ERK-MAPK- and PKC-mediated signaling cascades downstream of the pre-TCR complex (4, 5, 6, 7, 29, 30), led us to focus our efforts on determining whether these signaling pathways are responsible for activating CREB.

To achieve this, we used a reporter plasmid system that would allow us to determine which upstream kinase is responsible for activating CREB downstream of pre-TCR complex ligation within SL-12β.12 cells. This method of detection has been successfully used previously to show that PKC and ERK-MAPK signaling cascades are activated upon formation of the pre-TCR complex in developing thymocytes (5, 6, 31, 32). This system uses two plasmids: a fusion-activator plasmid (pFA2-CREB), which encodes for the trans activation domain of CREB (residues 1–280), 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 CREB-fusion protein by upstream kinases can be readout in the form of luciferase activity.

To address whether the reporter plasmid system was sensitive enough to detect the activation of CREB downstream of the pre-TCR complex, as observed in Fig. 1, SL-12β.12 cells were transfected with pFR-Luc, either alone or together with pFA2-CREB, as indicated in the figure legends. The transfected cells were then stimulated with immobilized anti-TCR-β mAb. Cells transfected with pFR-Luc alone displayed background luciferase activity (Fig. 2,A). Luciferase activity increased slightly above background levels in cells transfected with pFR-Luc and pFA2-CREB (Fig. 2,A), which may reflect the low constitutive CREB activity present within SL-12β.12 cells, as observed in Western blotting analyses (Fig. 1). Importantly, stimulation of pFR-Luc/pFA2-CREB-transfected SL-12β.12 cells with anti-TCR-β mAb resulted in at least a 15-fold stimulation in luciferase activity compared with unstimulated cells (Fig. 2,A). Collectively, our data indicate that engagement of the pre-TCR complex results in the phosphorylation and activation of CREB (Figs. 1 and 2 A).

To test whether the observed CREB activation downstream of the pre-TCR was dependent on PKC activity, we incubated the pFR-Luc/pFA2-CREB-transfected SL-12β.12 cells with specific pharmacological inhibitors during anti-TCR-β mAb cross-linking. Incubation of the TCR-β-stimulated cells with the pan PKC inhibitor, Gö 6983 (1 μM) (33), resulted in almost a complete abrogation in the elevation in luciferase activity observed in the absence of inhibitor (reduced to 26.3 ± 5.67%; Fig. 2 A), indicating that PKC is responsible for activating CREB downstream of the pre-TCR complex.

We have previously shown that PKCα plays a critical role in mediating signaling events during β-selection (5); therefore, to investigate whether classical PKC isoforms, and more specifically PKCα, were responsible for the activation of CREB, we conducted two sets of experiments. First, we incubated pFR-Luc/pFA2-CREB-transfected SL-12β.12 cells with the classical PKC inhibitor Gö 6976 (1 μM) (34) during anti-TCR-β mAb cross-linking. Our analyses revealed that there was almost a complete abrogation in the elevation in luciferase activity observed in the absence of inhibitor (reduced to 33.7 ± 7.97%; Fig. 2,A), indicating that a classical PKC isoform is responsible for activating CREB downstream of the pre-TCR complex. In the second set of experiments, we transfected SL-12β.12 cells with pFR-Luc and pFA2-CREB reporter plasmids either alone or in the presence of a plasmid encoding a dominant-negative form of PKCα (PKCα-KR). As described previously, anti-TCR-β-stimulated pFR-Luc/pFA2-CREB-transfected SL-12β.12 cells exhibited a 20-fold stimulation in luciferase activity (Fig. 2,B). However, upon cotransfecting SL-12β.12 cells with a plasmid encoding PKCα-KR and stimulating the cells with anti-TCR-β mAb, there was a complete abrogation of luciferase stimulation observed in the absence of the PKCα-KR-encoding plasmid (Fig. 2,B). To demonstrate that CREB activation can be efficiently detected with these reporter plasmids, pFR-Luc/pFA2-CREB-transfected cells were also transfected with a plasmid encoding a constitutively active PKCα (PKCα-CAT). These cells showed maximal luciferase activity irrespective of anti-TCR-β mAb stimulation (Fig. 2 B), indicating that this reporter system provides a highly sensitive method for detecting CREB activation. Taken together, these experiments establish that the PKCα isoform is responsible for the activation of CREB downstream of pre-TCR complex ligation.

Our finding that PKCα is responsible for activating CREB (Fig. 2) together with previous reports that PKCα can activate the ERK-MAPK signaling cascade, which plays an important role in mediating signals downstream of the pre-TCR (4, 6, 7, 29, 30, 35), led us to hypothesize that the activation of CREB may be, at least in part, mediated by ERK-MAPK. In support of this, when SL-12β.12 cells were transfected with pFR-Luc/pFA-ELK, to measure ERK activity (6), and stimulated with immobilized anti-TCR-β mAb in the presence of either Gö 6983 or Gö 6976, luciferase activity was partially inhibited (reduced by 34 and 46%, respectively), compared with cells stimulated in the absence of PKC inhibitors, indicating that PKC is partly responsible for the activation of ERK-MAPK downstream of the pre-TCR complex (6).

Therefore, to address whether the ERK-MAPK signaling cascade was responsible for mediating CREB activation, we transfected SL-12β.12 cells with pFR-Luc, either alone or together with pFA2-CREB, as indicated in the figure legends. The transfected cells were then stimulated with immobilized anti-TCR-β mAb, as above. As noted in Fig. 2, stimulation of pFR-Luc/pFA2-CREB-transfected SL-12β.12 cells with anti-TCR-β mAb resulted in almost a 16-fold elevation in luciferase activity compared with unstimulated cells (Fig. 3), supporting the data displayed in Figs. 1 and 2 that CREB is phosphorylated and activated downstream of pre-TCR complex ligation. Upon stimulation of the transfected cells in the presence of the MEK1 inhibitor, PD98059 (10 μM) (36), we noted a significant reduction in the amount of luciferase activity by 43% of that observed in cells stimulated in the absence of PD98059 (Fig. 3). Of note, the elevation in luciferase activity observed in cells transfected with pFR-Luc/pFA-ELK reporter plasmids upon cross-linking the pre-TCR complex was almost completely abrogated in the presence of PD98059, indicating that this concentration of inhibitor is sufficient to inhibit ERK-MAPK activity (reduced by 85.1%) (6). These results establish that the ERK-MAPK signaling cascade is partially responsible for mediating CREB activation in a PKC-dependent manner downstream of pre-TCR complex ligation.

To directly address whether CREB plays a central role during β-selection, we took advantage of the fact that the CREB superfamily contains several naturally occurring inhibitors of CREB function, which are members of the CRE-modulator family of proteins, including the ICER family of proteins that behave as classic dominant-negative molecules (17, 37). Endogenous ICER expression was detected in wild-type and RAG° mouse-derived thymus samples, indicating that this product is naturally expressed in developing thymocytes (Fig. 4 and data not shown). To investigate the expression levels of ICER transcripts during β-selection, RAG° neonatal mice were treated with anti-CD3 mAb, which results in the differentiation of DN thymocytes toward the DP stage of development (38). Thymocytes isolated from untreated and treated RAG° mice (36 h or 10 days (90% DP thymocytes)) were analyzed by semiquantitative RT-PCR for expression levels of ICER. These analyses revealed that ICER expression is not altered upon the generation of a pre-TCR-like signal (36 h; Fig. 4), but is elevated in DP thymocytes (10 days; Fig. 4). These findings suggest that ICER expression is not directly up-regulated downstream of the pre-TCR complex in response to CREB activation (Fig. 1 B), suggesting that enforced overexpression of ICER at this stage of development may result in a reduction of pre-TCR-activated CREB signaling.

Therefore, to examine the effect of diminished CREB activity on T cell development during β-selection, we used a previously characterized retroviral vector, MIEV, to express constitutively overexpressed ICER in developing thymocytes (5, 21). This vector allows for the transcription of a bicistronic message encoding ICER and GFP, thus enabling retrovirally infected cells to be identified by flow cytometry by virtue of their ability to express GFP. The functionality of the ICER construct was demonstrated by its ability to abrogate the activation of CRE-promoter activity upon addition of forskolin to pCRE reporter plasmid-transfected EL4 cells (data not shown). Freshly prepared FL cells were retrovirally infected with either vector alone (MIEV) or ICER-encoding vector (MIEV.ICER). Retrovirally infected lymphoid progenitor cells (CD117⁺GFP⁺) were then separated by flow cytometric cell sorting and placed into FTOC for 7 or 9 days. At these time points, FTOCs were analyzed by flow cytometry for the ability of ICER-expressing progenitor cells to progress through T cell development. Figs. 5 and 6 show that similar percentages of GFP⁺ cells were evident in both vector alone (MIEV) and ICER-expressing (MIEV.ICER) FTOCs at days 7 (Fig. 5) and 9 (Fig. 6), indicating that ICER expression does not impact on thymocyte survival. Flow cytometric analysis of FTOCs at days 7 and 9 revealed normal T cell development in the uninfected (GFP⁻) endogenous thymocyte populations in both MIEV and MIEV.ICER FTOCs (CD4 vs CD8; R1 gated; Figs. 5 and 6), as evidenced by the presence of CD4⁺CD8⁺ DP thymocytes, as well as CD4⁺ and CD8⁺ single-positive cells. These results indicate that the FTOCs provided an appropriate environment for normal T cell development.

Analysis of GFP⁺ cells from FTOCs day 7 revealed a delay in the generation of DN3 cells and a subsequent delay in the differentiation of DN3 cells toward the DP stage of development within MIEV.ICER FTOCs compared with MIEV FTOCs (R2-gated (GFP⁺) cells; Fig. 5). This finding suggests that the expression of ICER in developing thymocytes renders them unable to efficiently transit from the DN to the DP stage of development.

Analysis of GFP⁺ cells in FTOCs at day 9 revealed that there was a delay in the generation of DP thymocytes, as noted by the substantial reduction in the percentage of DP thymocytes present in the MIEV.ICER FTOCs compared with MIEV FTOCs (33.1 vs 61.6% (data not shown)). Indeed, this reduction was more pronounced when the percentages of DP thymocytes were compared in the GFP^low (R2-gated) or GFP^high (R3-gated) populations of MIEV.ICER FTOCS and MIEV FTOCs (R2 gated, 37 vs 62%; R3 gated, 26 vs 58%; Fig. 6,A). These data indicate that the differentiation of DN thymocytes toward the DP stage of development is obstructed by the expression of ICER in a dose-dependent manner. Furthermore, we noted that the cellularity of the MIEV.ICER FTOCs was reduced nearly 5-fold compared with MIEV FTOCs, suggesting that CREB plays a role in the proliferation of DN thymocytes (4.81 ± 0.63-fold increase in cellularity in MIEV FTOCs compared with MIEV.ICER FTOCs; n = 3). Indeed, the number of DP thymocytes in MIEV.ICER FTOCs was reduced by 84.3% in GFP^low population and 86.4% in the GFP^high population compared with MIEV FTOCs (Fig. 6 B). These data suggest that the proliferative capacity of DN thymocytes is reduced by the expression of ICER. Collectively, these results demonstrate that diminished CREB activity in early thymocytes results in an inefficient execution of proliferation and differentiation toward the DP stage of development, thus implicating CREB as a key mediator of the signaling events that control β-selection.

In this study, we have investigated the role of the transcription factor CREB during T cell development. We demonstrate that CREB is constitutively phosphorylated in primary thymocytes, but not in primary splenic lymphocytes. We also show that CREB is phosphorylated and activated downstream of pre-TCR complex ligation and that this activation is mediated by PKCα- and ERK-MAPK-signaling pathways. In addition, we show that constitutive expression of ICER in developing thymocytes hampers the ability of DN thymocytes to differentiate and proliferate toward the DP stage of development. Collectively, our results establish that CREB plays a central role in mediating signals generated downstream of the pre-TCR and ultimately controlling hallmark events during β-selection.

Our finding that high levels of phosphorylated CREB were detected specifically in thymus extracts, but not in spleen, was complemented by previous studies using gel mobility-shift analyses, which established that decamer CRE sequences, conserved in TCR gene promoters, were only bound (i.e., active) when incubated with thymus extracts, and remained inactive in the presence of spleen or nonlymphoid organ extracts (14, 39, 40, 41). These studies suggest that the thymus is a unique site for TCR gene rearrangement due to the conservation of specific regulatory sequences in the promoters of TCR genes coupled with constitutive CREB activity to ensure high transcriptional activity of TCR gene loci (14, 41).

Interestingly, the level of phosphorylated CREB detected in DN thymocyte-enriched (RAG° mouse-derived) lysates was considerably lower than that observed in DP thymocyte-enriched (wild-type mouse-derived) lysates, suggesting that pre-TCR-derived signals may be responsible for the elevation in CREB phosphorylation. However, we observed an elevation in ICER expression in DP thymocytes that did not appear to be a direct result of pre-TCR-mediated signals. This suggests that while ICER can modulate CREB activity in a developmentally regulated manner, it occurs independently of the pre-TCR complex. Indeed, ICER was recently shown to be differentially regulated during T cell development (42). TCR-β and TCR-α gene recombination, while limited to the thymus, is temporally separated with TCR-β gene loci rearrangements occurring at the DN stage of development, while TCR-α gene loci rearrangements occur at the DP stage (43). One of the outcomes of β-selection is the induction of TCR-α gene transcription (44), which is regulated, at least in part, by the occupation of CRE sites in the promoter of the TCR-α gene (15, 40). Therefore, the up-regulation of CREB activity downstream of pre-TCR complex formation may contribute to the induction of TCR-α gene transcription. Collectively, these findings identify the thymus as the unique site for TCR gene rearrangement due to the conservation of specific regulatory sequences in the promoters of TCR genes coupled with constitutive CREB activity to ensure high transcriptional activity of the TCR gene loci (14, 41).

The disparity between the activation status of CREB in SL-12β.12 cells, which are derived from SCID thymoma, and primary thymocytes may be explained by the findings from Lanier et al. (41), who described that while decamer-binding activity was evident in thymus extracts from wild-type mice, no decamer-binding activity was detected in extracts prepared from the thymuses of SCID mice, which fail to rearrange these genes. However, we detected CREB activity in thymuses derived from RAG° mice, which are also unable to rearrange genes (19). Lanier et al. (41) also indicated that no decamer activity was detectable in extracts from two T cell lines that express a functionally rearranged TCR-β chain; this suggests that the presence of decamer activity is not absolutely required for transcription of the TCR-β locus, but may be specifically required for the rearrangement of the TCR gene locus. Unstimulated SL-12β.12 cells exhibit lower levels of phospho-CREB activity than primary thymocytes, possibly because they express a functionally rearranged, exogenously added TCR-β chain (20). Although this does represent a difference between primary thymocytes and SL-12β.12 cells, our finding that the level of CREB phosphorylation is increased upon mimicking a pre-TCR-derived signal, either by stimulating RAG°-derived thymocytes with anti-CD3 mAb or cross-linking SL-12β.12 cells with anti-TCR-β mAb, suggests that the SL-12β.12 cellular model provides an excellent model to ascertain how CREB activity is regulated downstream of the pre-TCR.

The functional redundancy observed within the CREB family of transcription factors has made it difficult to ascertain their function in vivo (45). Initial studies using a transgenic mouse model carrying a mutant CREB^A119 (nonphosphorylatable), with dominant-negative function, under the control of the CD2 promoter/enhancer, did not reveal any alterations in thymocyte development (46). Nevertheless, in response to activation signals, thymocytes and T cells from these animals displayed a profound proliferation defect characterized by a marked decrease in IL-2 production and an increase in apoptosis (46). The lack of a clear phenotype during early thymocyte differentiation in these mice may be attributable to the use of the CD2 promoter/enhancer to drive the expression of the dominant-negative CREB molecule (46, 47) as this promoter only becomes fully active in DN thymocytes at a late, post-β-selection stage (48); indeed, thymocytes from RAG° mice are blocked before the full induction of CD2 expression.

However, a clue that CREB may play an essential role during early T cell development was revealed in the CREB^null mouse model system that lacked expression of all functional CREB isoforms (CREBα, β, and δ). T cell development in these mice was impaired with an overall reduced thymic cellularity, a decrease in DP thymocyte development, and a concurrent increase in the percentage of DN thymocytes, suggesting a defect during β-selection (16). Our data confirm a role for CREB during β-selection, as we have established that the ability of DN thymocytes to proliferate and differentiate is severely compromised with decreasing CREB function. The absence of a complete block in the generation of DP thymocytes, in the presence of ICER, may be due to an inability of ICER to totally extinguish CREB signals, i.e., ICER expression levels are not high enough. Nevertheless, this retroviral system has been successfully used previously to decipher the dose-dependent roles played by Notch and PKC during T cell development, as an increase in the level of GFP expression can be directly linked to an increase in the level of expression of the gene of interest (5, 49). In support of this, we observed a decrease in the percentage of DP thymocytes generated in the GFP^high population, compared with the GFP^low population, as these cells contain higher expression levels of ICER and therefore reduced CREB activity.

Previous studies have indicated that the PKC and Ras/Raf/ERK-MAPK signaling cascades are important for mediating the proliferation and differentiation of pre-T cells during β-selection (4, 5, 7, 29, 30). We now extend these findings, identifying CREB as a central mediator of proliferation and differentiation downstream of the PKC and ERK-MAPK in developing thymocytes. Our studies demonstrate that PKCα activity is entirely responsible for the activation of CREB. Indeed, the role of PKCα in pre-T cell differentiation and proliferation is supported by the finding that constitutive expression of a dominant-negative form of PKCα in developing thymocytes results in a differentiative and proliferative block in T cell development at the DN stage (5). PKCα has been shown to activate ERK-MAPK signaling cascade (35); however, our results indicate that the ERK-MAPK signaling cascade is partially responsible for CREB activation, possibly through an ribosomal S6-kinase-2-dependent mechanism (50, 51, 52), suggesting that the remainder of CREB activity is induced in an ERK-MAPK-independent, PKC-dependent manner.

The expression of constitutively active forms of PKC, Ras, or Raf-1 into developing thymocytes of RAG° mice results in the differentiation and proliferation of DN cells to the DP stage (4, 5, 7, 29). However, evidence that the activation of ERK-MAPK is not sufficient for full differentiation and proliferation of DN thymocytes comes from studies showing that expression of a plasmid encoding constitutively active MEK1 or overexpression of a downstream target of ERK-MAPK, Egr-1, in developing thymocytes derived from RAG° mice, while alleviating the block in T cell development at the DN3 stage, only enabled cells to mature to the CD8⁺ immature single-positive stage of development (30, 53, 54). These studies indicate that additional signals upstream of MEK1, possibly at the level of PKC, are required to elicit full induction of proliferation and differentiation downstream of the pre-TCR activation (Fig. 7). In this study, we identify CREB as a key transducer of both ERK-MAPK-dependent and -independent signaling pathways downstream of the pre-TCR complex, and establish a central role for this ubiquitous transcription factor in the regulation of proliferation and differentiation during β-selection.

We thank Cheryl Smith for excellent technical assistance with cell sorting, and Dr. Margaret M. Harnett and Maria Ciofani for their valuable discussion of the manuscript.