PKB Rescues Calcineurin/NFAT-Induced Arrest of Rag Expression and Pre-T Cell Differentiation¹ Free

Thymic T cell development is characterized by multiple steps of differentiation and proliferation, which ultimately lead to the emergence of functional self-MHC-restricted CD4⁺ and CD8⁺ T cells (1). Before the CD4⁺CD8⁺ double-positive (DP)³ stage, when thymocytes undergo positive and negative selection, they mature through four distinct CD4⁻CD8⁻ double-negative (DN) developmental stages, which are marked by the sequential gain and loss of the CD44 and CD25 surface molecules (2). DN1 cells (CD44⁺CD25⁻) differentiate to DN2 (CD44⁺CD25⁺) cells and then transit to the DN3 (CD25⁺CD44⁻) stage. Cells of the DN4 (CD25⁻CD44⁻) stage have down-regulated both CD25 and CD44 and undergo extensive expansion giving rise to DP cells.

At the DN3 stage, Rag1 and Rag2 proteins are expressed and initiate VDJ recombination of the TCR-β locus. This leads to expression of the pre-TCR, which consists of a productively rearranged TCR β-chain and the invariant pre-TCR α-chain associated with CD3 molecules (3, 4). Signals initiated via the pre-TCR are essential for survival (a process known as β selection), which triggers multiple signaling cascades leading to differentiation into DN4 cells, expansion and eventually to TCR-α locus rearrangement at the DP stage and replacement of the pre-TCR by a functional αβ TCR.

A considerable number of genetic and biochemical data show that defects in either Rag1 or Rag2 expression, TCR-β gene rearrangement or pre-TCR/CD3 assembly as well as in activation of downstream signaling molecules and transcription factors lead to a partial or total arrest of thymocyte development at the DN3 stage (5, 6, 7, 8). Interestingly, signals evoked from the pre-TCR or αβ TCR are biochemically linked to a similar set of protein kinases, phosphatases, adaptor molecules, and transcription factors (9). One common second messenger of pre-TCR (10) and mature TCR signaling is intracellular calcium, the levels of which are crucial for a number of downstream pathways, most notably activating the ubiquitously expressed serine-threonine-specific phosphatase calcineurin (CN) (11, 12).

Calcineurin, consisting of the large catalytic subunit A and the regulatory subunit B, is a specific target of the immunosuppressive drugs cyclosporin A (CsA) and FK506, which can completely block the calcium-dependent activation of T cells (13). Pharmacological approaches using CsA/FK506 in vivo and in vitro have shown that CN is an important regulator of thymocyte differentiation and selection (14, 15, 16, 17, 18, 19). More recently, genetic ablation of the CN-A β subunit (20) or the CN-B1 regulatory subunit (21) as well as transgenic (tg) expression of a constitutively active form of CN in the thymus (22) have confirmed that CN is required for positive but not for negative selection of DP cells.

Prominent targets of CN are the NFAT transcription factors (11, 23, 24). The four genuine NFAT members (c1–c4) share significant sequence homology, in particular in their DNA binding region. CN dephosphorylates 13 serine residues within the regulatory region of NFAT, resulting in its nuclear translocation and activation. Cessation of CN activity and rephosphorylation of NFAT by (a) nuclear protein kinase(s) results in relocalization to the cytosol. Genetic inactivation of NFATc1, c2, and c3, which are expressed in lymphocytes, has elucidated and confirmed both their common and individual roles in peripheral T cell activation, cytokine production, and Th cell differentiation (11) as well as in control of cell cycle (25) and apoptosis (26). During thymocyte development, individual NFAT family members show distinct expression patterns (27, 28), but single or double deficiency of specific NFAT isoforms has either no or a mild effect on T cell maturation (29, 30, 31).

TCR-induced NFAT activation is associated with the induction of various downstream signaling cascades including that of PI3K. A prominent target of PI3K signaling is the serine-threonine kinase protein kinase B (PKB/Akt), which is an essential regulator of T cell activation and survival, and is critically involved in the development of various types of cancer, including T cell lymphoma (32, 33). The role of serine-threonine kinases in thymocyte development has become obvious more recently, since overexpression of specific isoforms of PKB (34) or mutation of their upstream regulators, including PTEN (35, 36) or phosphoinositide-dependent kinase 1 (37), reveal strong effects on thymocyte survival, selection, and differentiation.

In a previous study, we reported that a constitutively active version of PKB enables T cell proliferation in the presence of CsA and independent of the effective nuclear accumulation of NFAT proteins (38). Here, we studied the interaction of PKB with NFAT signaling in early thymocyte development by analyzing double-tg mice expressing both, a constitutively active form of CN (ΔCam) and a constitutively active form of human myristoylated PKBα (myrPKB). We find that overexpression of CN/NFAT arrests thymocyte development at the DN3 to DN4 transition and identify NFAT proteins as novel regulators of rag expression. Remarkably, the ΔCam-induced defect can be rescued by active PKB, suggesting regulation of NFAT activity by PKB to be crucial for normal β selection of thymocytes.

The truncated murine CN-A mutant (ΔCam) was previously described (39) and provided by R. Kincaid (Department of Pharmacology, Pennsylvania State University College of Medicine, Hershey, PA). The 1.2-kb DNA fragment was cloned into the BamHI sites of the expression cassette harboring the murine proximal lck promoter, a 3′ human growth hormone fragment and the human CD2 locus control region (40). SpeI fragments were injected into oocytes from NMRI mice. Founder mice were crossed to NMRI or C57/BL6 mice for further analysis. Genomic integration of the ΔCam transgene was determined by PCR of genomic tail DNA using the primers forward 5′-CCG AGC CCA AGG CGA TTG ATC C-3′ and reverse 5′-CCC GGT TTC TGA TGA CTT CCT TCC-3′. Tg mice expressing a membrane-targeted myrPKB under control of the human CD2 promoter have been previously described (41). Mice heterozygous for ΔCam were crossed with myrPKB tg or OT1 TCR tg (42) mice (both on C57/BL6 background). In all experiments, F₁ littermates negative for both ΔCam and myrPKB, and positive only for either ΔCam or myrPKB, were used as wild-type (wt) control or control for ΔCam or myrPKB tg mice, respectively. Mice used were 1-day-old or 6- to 10-wk-old unless mentioned otherwise. Identical results were obtained with ΔCam tg mice crossed to the C57BL/6 background for at least five generations and with ΔCam tg mice crossed to a second myrPKB tg line.

Thymus single-cell suspensions were prepared for flow cytometry analysis. For three-color analysis, thymocytes were immunostained with CD4 (GK1.5), CD8 (YTS.169), CD5 (53-7.3), CD24 (HSA, M1/69), CD25 (PC61), CD44 (IM7), CD2 (RM2-5), IL-7Rα (B12-1), CD27 (LG.3A10), TCR-β (H57-597), or TCR-γδ (GL3) mAbs labeled with FITC, PE, or biotin in PBS containing 0.1% BSA and 0.01% sodium azide. Biotin-coupled Abs were revealed with streptavidin-CyChrome. For further analysis of DN3 cells, thymocytes were gated on CD4⁻CD8⁻CD25⁺ cells. Abs were purchased either from BD Pharmingen or were prepared in our laboratory from hybridoma supernatants. For detection of apoptosis in DN3 cells, CD4⁻-, CD8⁻-, and CD25-labeled thymocytes were incubated with Annexin V^FITC (BD Pharmingen) for 15 min at room temperature. Cells were analyzed by three-color flow cytometry using a FACSCalibur and CellQuest software (BD Biosciences). DN thymocytes were enriched by treatment of total thymocytes with Abs against CD4 (GK1.5), CD8 (TiB105), NK1.1 (1D4), CD19 (1D3), and MHC class II (2G9) followed by negative selection using BioMag anti-rat-IgG coupled to magnetic beads (Qiagen). Purity of DN cells was 90–95% as verified by FACS analysis. After depletion of DP cells, DN3 cells were isolated by gating on CD25⁺CD44⁻ cells and sorting on a FACSVantage (BD Biosciences). CD4⁺ T cells from lymph nodes of wt mice and myrPKB homozygous mice with lymphoma were either electronically sorted or purified by depletion of CD4⁻ cells using the CD4⁺ T cell isolation kit from Miltenyi Biotec. CD4⁺ T cells were stimulated with plate-bound CD3 Abs (145.2C11, 5 μg/ml; BD Pharmingen) for 4 h before cells were harvested and RNA or protein extracts were prepared.

Total cellular RNA from isolated DN thymocytes or CD4⁺ T cells was prepared with TRIzol reagent (Invitrogen Life Technologies) and cDNA was synthesized with SuperScript reverse transcriptase (Invitrogen Life Technologies) from 1 μg of RNA as described by the manufacturer’s protocol. Each sample was normalized by PCR for β-actin or hypoxanthine phosphoribosyltransferase (HPRT). Serially diluted cDNA was analyzed for expression of CN (detecting bp 301–909 of its N terminus), ΔCam, myrPKB, Rag1, Rag2, β-actin, or HPRT using the following primers: CN, forward 5′-AAG GAG GGA AGG CTG GAA GA-3′ and reverse 5′-GGC ATC CAT ACA GGC ATC AT-3′; ΔCam, forward 5′-CCG AGC CCA AGG CGA TTG ATC C-3′ and reverse 5′-CCC GGT TTC TGA TGA CTT CCT TCC-3′; myrPKB, forward 5′-AGA TTT CCT GTC CCC TCT CAG G-3′ and reverse 5′-TGT TGG ACC CAG CTT TGC AG-3′; Rag1, forward 5′-TGC AGA CAT TCT AGC ACT CTG-3′ and reverse 5′-ACA TCT GCC TTC ACG TCG AT-3′; Rag2, forward 5′-TCT CTA AAG ATT CCT GCT ACC TC-3′ and reverse 5′-TGG AAT TCA CTG CTG GGG TAC-3′; β-actin, forward 5′-CCA GGT CAT CAC TAT TGG CAA GGA-3′ and reverse 5′-GAG CAG TAA TCT CCT TCT GCA TCC-3′; HPRT, forward 5′-GCT GGT GAA AAG GAC CTC TC-3′ and reverse 5′-CAC AGG ACT AGA ACA CCT GC-3′.

For immunoblotting, cytoplasmic and nuclear protein extracts of purified CD4⁺ T cells or whole cell lysates of DN cells were prepared as described (38). Protein concentration was determined with Bradford’s reagent (Bio-Rad) and 30 μg of whole cell and 10 μg of cytoplasmic or nuclear protein extract was fractionated by 8% SDS-PAGE, transferred to nitrocellulose membranes and analyzed for expression of phospho-PKB (Ser⁴⁷³) and PKB (both Cell Signaling Technology), Rag1 and β-actin (both Santa Cruz Biotechnology), NFATc1 (Axxora), or NFATc2 (a gift from Dr. A. Rao, Center for Blood Research, Harvard Medical School, Boston, MA). Primary Abs were detected by goat anti-rabbit (Santa Cruz Biotechnology), goat anti-mouse, or rabbit anti-goat Abs (Jackson ImmunoResearch Laboratories) coupled with HRP and ECL (Pierce).

For intracellular staining, total thymocytes were first stained for cell surface CD4, CD8, and CD25 molecules. Cells were then fixed with 1% paraformaldehyde (10 min at room temperature) and permeabilized in PBA (1× PBS, 5% BSA, and 0.02% NaN₃) containing 0.5% saponin (10 min at room temperature) (43). Permeabilized cells were incubated with FITC-labeled mAbs against TCR-β, TCR-γδ, or Bcl-2 (BD Pharmingen) diluted in PBA containing 0.5% saponin for 30 min at 4°C. Cells were washed twice in 1 ml of PBA/0.5% saponin, once with PBS/0.1% BSA/0.01% azide, and analyzed by FACS.

Thymic lobes from neonatal (day 1) wt and ΔCam tg mice were placed on nucleopore membranes (Whatman) and cultured on 2 ml of complete RPMI 1640 medium containing 10% FCS in the absence or presence of 100 ng/ml CsA (Sigma-Aldrich) for 4 days. Subsequently, single-cell suspensions were prepared and thymocyte subpopulations were determined by flow cytometry. Intracellular expression of the TCR β-chain was analyzed gating on the CD4⁻CD8⁻ population.

EL4 lymphoma cells were grown in RPMI 1640 medium containing 10% FCS, 0.352 μM 2-ME, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C and 5% CO₂. Transfection of cells was done using FuGene 6 reagent (Roche) according to the manufacturer’s protocol. Briefly, 1 μg of total DNA solution containing 50 ng of murine Rag2-luciferase reporter gene, 50 ng of ΔCam expression vector (39), 400–800 ng of human NFATc1 (26), NFATc2 expression vector, or empty vector DNA, was mixed with 4 μl of Fugene and evenly spread over 1.5–3.0 × 10⁵ cells/well. At 26 h after transfection, cells were induced with PMA (Sigma-Aldrich) and ionomycin (Calbiochem), each 100 ng/ml, for 16–20 h when cells were harvested and luciferase reporter activity was determined. The luciferase reporter construct contained the TCR β-chain enhancer and murine Rag2 promoter region from nt −251 to nt +147 (44).

ChIP analysis was conducted essentially following the protocol of Abcam (〈www.abcam.com/technical〉). Briefly, 4–6 × 10⁶ freshly isolated thymocytes from wt or ΔCam tg mice were cross-linked using 1% formaldehyde. Isolated nuclei were sonicated so that the average length of chromosomal DNA became 500-1000 bp. Chromatin solution was precleared with Sepharose G, and chromatin was immunoprecipitated by incubating with 6–8 μg of NFATc1 Ab (Axxora) or an isotype-matched polyclonal mouse IgG Ab (Jackson ImmunoReserach Laboratories) overnight at 4°C. Immune complexes were collected on protein G-Sepharose beads and immunoprecipitates were eluted with 1% SDS, 50 mM NaHCO₃. After reversal of cross-links and deproteination, the presence of Rag1 and Rag2 promoter sequences was analyzed by PCR using the following oligonucleotide primers: Rag1 promoter forward 5′-CAT TCT CAG GAG ATG AAA TGA CAG C-3′ and reverse 5′-TTG TCC CAT AGT GCA CAA TGC-3′; and Rag2 promoter forward 5′-CTT AAG ACA GTC ATT TTT TGT GGG-3′ and reverse 5′-AAG CCA GGA ATT AAA CTC GGC TC-3′. The sequence for the Rag enhancer element 5′ of the Rag2 promoter was taken from GenBank (accession number AY215076) as previously published (45). Primers used to detect NFATc1 binding within this sequence were: Rag enhancer (4) nt 1171–1193 forward 5′-ATC AGC AAG TGT TGG TAT TC-3′ and nt 1630–1650 reverse 5′-CTT TAT TCA GCA TTC AGA AGG-3′; Rag enhancer (7) nt 18716–18741 forward 5′-CCC ATG TTC AGA TCA ACT AAT AAT CC-3′ and nt 19402–19423 reverse 5′-AAA CTC CAA GCT CTG AGT CAA C-3′. Primers to detect NFATc1 binding in the described Erag enhancer (46) were nt 14–36 forward 5′-CAC CAT TCA ATA TTC AGG AGG G-3′ and nt 346–366 reverse 5′-CTG CTG TAG AAA AAT ACC ACA GAG-3′.

Single-cell suspensions of thymocytes were fixed in 0.5% formaldehyde for 10 min at room temperature and cytospinned onto poly-lysine-coated glass slides. Slides were incubated for 15 s in chilled acetone (−20°C) and cells were permeabilized by incubating in chilled methanol (−20°C) for 3 min. Subsequently, cells were washed thrice in PBS and incubated with goat anti-mouse Fc block (BD Pharmingen) for 15 min in a humid chamber. Afterward, 40 μl of primary Ab solution containing CD4 (Alexa Fluor 647; Caltag Laboratories) and CD8α (Alexa Fluor 488; Caltag Laboratories) and NFATc1 or NFATc3 Abs (Santa Cruz Biotechnology) appropriately diluted in PBS/1% BSA was added to the cells for 45 min. Cells were washed thrice in PBS and were further incubated with donkey anti-rabbit Alexa Fluor 555 Abs (Molecular Probes) for 45 min in the dark. Cells were washed thrice in PBS and twice in dH₂O and then counterstained with DAPI (4′,6′-diamidino-2-phenylindole; Sigma-Aldrich) for 5 min at room temperature for nuclear staining. Stained thymocytes were analyzed for NFAT expression in CD4⁻CD8⁻ cells using a TCS SP2 Leica confocal microscope and Leica confocal software.

To determine a role of PKB in regulating CN signaling during early intrathymic T cell development, we generated mice double-tg for active PKB and a constitutively active version of the large CN subunit A (ΔCam) (39). ΔCam, lacking CN′s calmodulin binding and autoinhibitory domains, was expressed in early thymocytes under control of the proximal p56^lck promoter (Fig. 1,A). Compared with endogenous expression of CN in wt mice, expression of ΔCam was elevated >10-fold in the ΔCam tg mouse line used for all analyses (Fig. 1,B). ΔCam tg mice were crossed with mice expressing myrPKB under control of the human CD2 promoter (41). ΔCam and myrPKB RNA expression were detected in sorted DN3 cells by RT-PCR (Fig. 1,C), and expression of active PKB in ΔCam/myrPKB DN thymocytes was confirmed by Western blot analysis (Fig. 1 D).

Compared with wt or myrPKB tg littermate mice, on average ΔCam tg mice exhibited a 18-fold reduction in thymus cellularity as total thymocyte number from 6- to 8-wk-old wt or myrPKB tg mice on average was 128 × 10⁶ but only 7 × 10⁶ for ΔCam tg mice. Analysis of thymocyte subpopulations in ΔCam tg mice revealed a drastic loss of DP cells, with only 5% DP cells in ΔCam tg vs 86% in wt mice, and a marked increase in the percentage of DN cells (Fig. 2,A, top, and B). The impact of ΔCam on thymic cellularity and composition of cell subsets was already obvious in newborn mice (Fig. 2,A, bottom), suggesting the loss of DP thymocytes is linked to a block in early thymocyte differentiation. This idea was supported by further characterization of the DN population into DN1-DN4 subsets, which revealed that ∼80–90% of ΔCam DN thymocytes are CD25⁺CD44⁻ DN3 cells (Fig. 2 C). Therefore, the loss of DP cells in ΔCam tg mice is due to a developmental arrest at the transition from the DN3 to DN4 stage of early thymocyte differentiation.

Strikingly, simultaneous expression of myrPKB significantly alleviated the DN3 arrest in ΔCam tg mice and allowed efficient T cell maturation such that the thymic phenotype of ΔCam tg mice closely resembled that of littermate wt mice (Fig. 2, A–C). Thymocyte numbers in ΔCam/myrPKB double-tg mice were increased on average by 10-fold ranging from 30 to 140 × 10⁶. The percentage of DP cells was restored to an average of 70% and the percentage of DN cells declined to 13%, closely approaching values for wt mice. Thus, myrPKB releases the DN3-DN4 transitional block evoked by ΔCam, presumably by counteracting ΔCam-induced downstream signaling events, including activation of NFAT factors. The results from this study were also observed with a second myrPKB tg mouse line (data not shown), indicating that the effective rescue of DN3 cells in ΔCam tg mice is mediated via active PKB.

To understand the molecular mechanisms whereby myrPKB can antagonize the differentiation block imposed by active CN, we first investigated several surface Ags that are modulated during progression from the DN3 to DP stage. Expression of CD2 and CD5 molecules is up-regulated by pre-TCR signals as cells transit from the DN3 to DN4 stage (47, 48). Analysis of CD4⁻CD8⁻CD25⁺ DN3 cells from ΔCam tg mice showed that ∼47% of cells exhibited enhanced expression of CD2 compared with ∼25% in DN3 cells from myrPKB tg or wt littermate mice (Fig. 3 A). High level CD5 expression was detected in 87% of ΔCam DN3 cells but on average in only 23% DN3 cells of the other three tg mouse lines. The majority of ΔCam DN3 cells also showed lower HSA (CD24) expression, whereas IL-7Rα level was similar to control cells. Notably, myrPKB reverted the expression levels of the analyzed surface Ags in ΔCam DN3 cells toward those detected in wt cells, indicating that ΔCam DN3 cells underwent normal differentiation in the presence of myrPKB.

To explore whether ΔCam affects the survival of DN3 cells, we examined levels of the antiapoptotic protein Bcl-2 and the degree of apoptosis by annexin V staining. Bcl-2 expression has been described to be high in DN cells and low in most DP cells (49). In DN3 cells of wt and myrPKB tg mice we detected two populations expressing either high or lower levels of Bcl-2 (Fig. 3,B). ΔCam DN3 cells instead comprised a homogeneous population expressing very high levels of Bcl-2, which were even above levels detected in control mice. In the presence of myrPKB, Bcl-2 expression was intermediate between ΔCam and wt DN3 cells. In addition, a population with lower Bcl-2 levels was detectable within the DN3 compartment as found in wt cells, presumably representing those cells that have progressed in the differentiation process toward the DN4 stage. Measuring apoptosis by annexin V staining revealed an about 2-fold increase in the apoptosis of DN3 cells in ΔCam tg mice (Fig. 3 C). Thus, inhibition of DN3 progression by ΔCam is accompanied by an enhanced induction of Bcl-2 and the absence of overt death of DN3 cells.

The block in DN3-DN4 transition of ΔCam cells suggests defects in expression of one or more of the pre-TCR components. CD3ε expression could be excluded as all CD25⁺CD44⁻ DN3 cells of ΔCam tg mice expressed levels of CD3ε comparable to DN3 cells from wt mice (data not shown). However, FACS analysis showed that only 0.1–0.2% of ΔCam DN3 cells contained higher amounts of intracellular TCR-β protein compared with 13–15% in wt or myrPKB DN3 cells (Fig. 4 A). Within the DN4 population, which mainly harbors β-selected cells, only 14% of cells were found positive for intracellular TCR β-chain in ΔCam tg mice. Likewise, although all DP cells of wt and myrPKB tg mice expressed intracellular TCR β-chain, only 43% of the few DP cells arising in ΔCam tg mice had a high level of intracellular TCR-β protein. Therefore, ΔCam suppresses TCR-β formation but some ΔCam DN3 cells can aberrantly develop into DP cells in the absence of a functional pre-TCR. Strikingly, simultaneous expression of myrPKB increased the percentage of ΔCam DN3 cells positive for intracellular TCR-β by 20-fold such that ∼45% of DN4 and virtually all DP cells showed normal expression of intracellular TCR-β protein. Thus, active PKB reverts the defects in intracellular TCR β-chain formation imposed by hyperactive CN.

For further characterization, we analyzed cell size (50) and observed that ΔCam DN3 cells had slightly larger mean fluorescence intensity (MFI 414) than the corresponding wt cells (MFI 387), but within the DN4 population large β-selected cells were missing (Fig. 4,B, left). Likewise, up-regulation of CD27 expression has recently been identified as a marker for β- or γδ-selected DN3 cells (51), and ΔCam DN3 cells showed significantly higher expression of CD27 (MFI 693) than their wt counterparts (MFI 206) (Fig. 4,B, upper right). Furthermore, we could exclude the possibility that ΔCam signaling diverts T cell precursors mainly into the γδ T cell lineage (52, 53, 54) because the percentage of DN4 cells staining positive for intracellular γδ TCR was also greatly reduced (Fig. 4 B, lower right). Overall, these analyses show that although ΔCam signaling blocks DN3-DN4 transition, it induces a phenotype in DN3 cells, at least for the analyzed parameters, resembling that of cells that have passed the β selection checkpoint.

To further dissect the defect in pre-TCR formation, we analyzed the expression of Rag proteins, which are essential for the recombinatorial rearrangements at the TCR β-chain locus (55). In ΔCam DN cells, Rag1 protein expression was extremely low; however, it was fully restored when myrPKB was coexpressed (Fig. 4,C). RT-PCR revealed strongly reduced Rag1 as well as Rag2 mRNA levels in ΔCam DN cells (Fig. 4 D), suggesting ΔCam signaling interferes with Rag expression at the transcriptional or posttranscriptional level. Notably, in ΔCam/myrPKB DN cells, Rag1 and Rag2 mRNA expression was rescued. We therefore concluded ΔCam induces a block in early thymocyte development by diminishing Rag1 and Rag2 expression, which would obstruct pre-TCR formation and further differentiation of DN3 cells. Active PKB, acting directly on CN or its targets, overcomes these negative regulatory effects, thereby enabling ΔCam DN3 cells to proceed in differentiation to DP cells.

To analyze whether providing a functional fully rearranged αβ TCR transgene would overcome the ΔCam-induced differentiation block, we crossed ΔCam tg mice to OT1-TCR tg animals (42). Expression of the OT1-TCR increased the percentage of DP cells in adult ΔCam tg mice to maximally 30% (± 19) vs 5.4% (± 1.5) in ΔCam only and 79.2% (± 0.9) in OT1 tg mice. The total thymocyte number in ΔCam/OT1 double-tg mice was highly variable and increased ∼3-fold to 14.0 × 10⁶ (± 20) compared with 3.3 × 10⁶ (± 2.5) in ΔCam and 157.5 × 10⁶ (± 14.6, n = 4) in OT1 tg mice. Similar results were found for newborn mice as shown in Fig. 4,E. OT1-TCR expression increased the total thymocyte number and percentage of DP cells by about 3-fold compared with ΔCam tg mice. However, full maturation to CD4^highCD8^high-expressing DP cells was hampered as the majority of ΔCam/OT1 DP cells had a CD4^lowCD8^low profile. Analysis of OT1-TCR specific Vα2 and Vβ5 chains (Fig. 4 E, bottom) revealed that the OT1-TCR was properly expressed and that ΔCam actually enhanced the percentage of DN3 cells expressing the tg TCR. These data indicate that expression of a rearranged αβ TCR promotes transition of ΔCam DN cells to the DP stage but neither full maturation of DP cells nor thymic cellularity is rescued as effectively as was found for myrPKB. MyrPKB therefore evokes regulatory mechanisms in addition to enhancing Rag and intracellular TCR β-chain expression that are required for an efficient DN3 progression and expansion of ΔCam tg cells.

Because our previous studies suggested a direct interaction between PKB and NFAT (38), we decided to focus on the regulation of NFAT factors in ΔCam tg mice. First, we investigated whether inhibition of CN, and therefore inhibition of NFAT activity, would restore T cell differentiation in ΔCam tg mice. For this purpose, we cultured newborn thymic lobes from ΔCam tg and wt mice in the presence or absence of CsA, an efficient inhibitor of CN-driven NFAT activation. As shown in Fig. 5,A, compared with untreated ΔCam lobes, CsA treatment led within 4 days to a 10-fold increase in thymus cellularity and induced efficient differentiation of ΔCam DN cells to the DP stage (42% vs 0.7% DP cells in untreated lobes). Likewise, CsA treatment restored expression of intracellular TCR-β protein in ΔCam DN cells to normal level (Fig. 5 B). Thus, down-modulation of CN/NFAT activity allows normal expansion and differentiation of otherwise arrested ΔCam DN3 cells.

In accordance to the role of CN in regulating NFAT proteins and myrPKB’s pivotal effects in ΔCam tg mice, we next asked whether myrPKB counteracts the effects induced by ΔCam by inhibition of NFAT activation. Using confocal microscopy, we studied the nuclear expression of NFATc1 and NFATc3 in DN cells from ΔCam tg and ΔCam/myrPKB double-tg mice. Thymocytes were stained for CD4 and CD8, and NFAT expression was evaluated in CD4⁻CD8⁻ cells. As evident from Fig. 6, in ΔCam tg mice nuclear NFATc1 expression was very high in the majority (∼80%) of DN cells and only ∼20% of cells showed lower levels of NFATc1. In high expressing cells, NFATc1 was totally nuclear as judged by staining of nuclei with DAPI (data not shown). Intriguingly, in ΔCam/myrPKB double-tg mice ∼70% of DN cells showed lower levels of NFATc1 and only 30% of cells had nuclear NFATc1 levels as high as found in ΔCam DN cells. Importantly, this inverted ratio of NFATc1^high vs NFATc1^low cells corresponded to the expression level of NFATc1 in DN cells from wt and myrPKB tg mice.

Similar to NFATc1, the nuclear level of NFATc3 was very high in almost all DN cells of ΔCam tg mice. In contrast, in the presence of myrPKB NFATc3 expression was low in the majority of DN cells of ΔCam tg mice (Fig. 6 A, bottom), corresponding to NFAT activity detected in DN cells of wt or myrPKB tg mice (data not shown). From these experiments it is obvious that one pivotal action of myrPKB is the inhibition of NFATc1 and NFATc3 activation in ΔCam DN cells.

From many studies it has become evident that hyperactivation of PKB is an essential factor in the induction and progression of tumorigenesis, including T cell lymphoma (56). In conjunction with our finding that myrPKB negatively regulates NFAT activation, we suspected a profound effect on NFAT activation in T cell lymphoma that arises in our myrPKB homozygous (our unpublished data) and other tg mice overexpressing active PKB (57, 58). We therefore analyzed CD4⁺ T cells from wt mice and myrPKB induced T cell lymphoma for NFAT activation. In wt CD4⁺ T cells activation of NFATc1 and NFATc2 was clearly detectable 4 h after TCR-CD3 engagement. In contrast, in CD4⁺ lymphoma cells nuclear levels of NFATc1 and NFATc2 were only marginal (Fig. 7A, top). Intriguingly, in the lymphoma cells we simultaneously detected expression of Rag1 and Rag2 mRNA (Fig. 7 A, bottom), supporting the observed link between NFAT activity and Rag transcription in ΔCam tg mice.

Next, we tested the ability of ectopically expressed NFAT to influence Rag promoter activity. EL4 lymphoma cells were transiently transfected with the ΔCam-expressing vector and a luciferase reporter construct harboring nt −251 to nt +147 of the murine Rag2 promoter, which previously was shown to confer lymphocyte-specific promoter activity (44). The Rag2 construct alone produced reproducible and strong luciferase reporter activity, whereas coexpression of NFATc1 or NFATc2 suppressed the Rag2 promoter (Fig. 7 B). NFAT factors therefore negatively interfere with Rag2 promoter activity in EL4 cells, results that are in line with the observations in ΔCam tg mice.

Regulation of Rag1 and Rag2 expression is highly complex and depends on the interaction of the promoters with several cis-acting DNA segments (59, 60). To assess the in vivo association of NFAT transcription factors with the Rag locus, we performed ChIP experiments using freshly isolated thymocytes from wt and ΔCam tg mice. As shown in Fig. 7,C, NFATc1 binding was detected for the Rag1 and Rag2 promoter but also for selected regions from the described cis-acting element 5′ of the Rag2 promoter (45) and the Erag enhancer (46). Binding of NFATc1 to these elements was reduced when ΔCam tg thymocytes were treated with CsA (Fig. 7 C, bottom), which inhibits NFAT activation. These results show that NFATc1 binds to the Rag locus, consistent with a regulatory function.

The targets of pre-TCR-induced calcium flux and NFAT activity (10) are still ill defined. We correlate increased NFAT activity in ΔCam DN cells with defective Rag1, Rag2, and intracellular TCR-β and γδ TCR expression suggesting that the spatial or temporal and quantitative expression of the individual NFAT members has to be tightly controlled at the DN3-DN4 transition to proceed with normal T cell development. Furthermore, our data highlight PKB as a critical regulator of CN/NFAT-mediated differentiation of T cell precursors as simultaneous expression of myrPKB reverted the profound developmental arrest of ΔCam DN3 cells. Although myrPKB seems to play multiple roles, one essential outcome of myrPKB action is the suppression of NFAT activity in DN3 cells of ΔCam tg mice. Therefore, sustained NFAT activation blocks TCR β-chain rearrangement and hence pre-TCR formation. In this scenario, suppression of ΔCam-induced NFAT activity to appropriate levels by myrPKB seems to be a prerequisite to proceed with Rag expression and with recombination processes similar to those observed when ΔCam DN cells were treated with the CN inhibitor CsA. Aifantis et al. (10) described a block in maturation of DP cells when day 14.5 fetal lobes were treated with CsA, whereas we found a rescue of DN3 to DP transition in CSA-treated lobes from newborn ΔCam tg mice. These findings reflect that in view of the massive overexpression of NFAT in ΔCam DN3 cells, CsA most likely did not completely block all NFAT activity but rather reduced it to more physiological levels and that NFAT activity could be critical for the differentiation processes before the DN3 stage.

CN/NFAT signaling is known to have positive but also negative effects on gene transcription. For instance, negative regulation is found for the cdk4 gene, which is repressed under basal conditions by NFAT (61). Although ΔCam represses Rag synthesis and intracellular TCR β-chain synthesis, it simultaneously, at least partially, signals in differentiation. This response is obvious from our observations that despite defective pre-TCR formation, ΔCam DN3 cells showed increased cell size and up-regulated CD2, CD5, and CD27 surface Ags to levels found in cells progressing in differentiation to the DP stage. In contrast, ΔCam DN3 cells failed to decrease Bcl-2 expression. Maintenance of Bcl-2 expression has been described to be a feature of emerging γδ T cells whereas pre-TCR-mediated signaling seems to inhibit Bcl-2 activity (51), but induces the antiapoptotic protein A1, which acts as regulator of pre-T cell survival (62). However, we found no evidence that ΔCam diverts the precursor cells mainly into the γδ lineage because production of intracellular γδ TCR was also reduced. The Bcl-2 promoter contains multiple NFAT consensus sequences, and NFAT involvement in transactivation of the bcl2 gene has been reported for thymocytes (31) and cardiac myocytes (63). Our data therefore strongly suggest that one parameter responsible for heightened Bcl-2 levels in ΔCam DN3 cells is the sustained NFAT activity, which modifies bcl2 transcription.

Besides negatively affecting Rag and intracellular TCR β-chain expression, sustained CN/NFAT activation seems to evoke inhibitory effects at later stages of DN4 to DP transition because simultaneous expression of the fully rearranged OT1 TCR in ΔCam tg mice only partially restored development of DP cells. MyrPKB, in contrast, rescued differentiation in ΔCam tg mice to almost full extent and one major function of myrPKB in ΔCam DN cells seems to be suppression of NFAT activity. We speculate that besides NFAT, other critical regulators of the DN3-DN4-DP transition such as Notch (64, 65, 66, 67), GATA3 (68), Egr (69), NF-κB (70), or Wnt signaling (71) among others (9), could be involved. Because Notch signals required for β selection can be replaced by myrPKB (72) further experimentation concerning this interesting issue is required.

It has become evident that signals from the pre-TCR bifurcate into those that regulate pre-T cell survival and proliferation and those that induce allelic exclusion, a process in which pre-TCR-induced signals terminate further Vβ-Dβ-Jβ rearrangements (50, 73). To date only a few molecules are known that are essential or mimic the ability of the pre-TCR to terminate TCR β-chain rearrangement. Among these few are the adaptor molecule SLP76 (74), Notch1 (64, 66), and cytosolic protein kinase D (75). Seen from this perspective, sustained levels of active NFAT in ΔCam tg mice could mimic those parts of pre-TCR signaling that initiate allelic exclusion and in ΔCam tg mice, therefore, would lead to premature repression of Rag expression and thus inhibition of pre-TCR formation. MyrPKB instead would down-regulate NFAT activity to levels appropriate for the developmental processes to occur.

Rag expression is regulated at the transcriptional and posttranscriptional level and various cis-regulatory elements controlling Rag transcription have been characterized, including promoter, enhancer, silencer, and anti-silencer elements (59, 60). Known transcription factors involved in T cell-specific Rag expression include c-myb (76), GATA3 (44), and Runx (45), and we have identified NFAT factors as part of the Rag transcriptional machinery. In view of the high number of potential NFAT binding sites within the regulatory elements, which may assist in keeping the Rag locus in an open accessible configuration, it appears that extreme fine tuning in the temporal and spatial activity of the individual NFAT members and their transcriptional partners such as NF-κB, which is an important regulator of Rag expression in B cells (77), would be necessary to guarantee tissue and developmental stage-specific expression of the Rag proteins.

Although Rag expression in the T cell lineage is thought to be restricted to developing thymocytes, it has also been reported for certain peripheral T cell subsets (78, 79), autoaggressive T cells (80), and lymphoid tumors (81). Furthermore, Rag tg mice show aberrant thymic development and rapidly develop lymphadenopathy and splenomegaly (82). Thus, our finding that transcriptional regulation of the Rag genes involves NFAT factors, which can be controlled by PKB, may also constitute an important pathway in the generation of autoaggressive T cells or malignant transformation in thymus and periphery. In this line, the link between PKB, NFAT, and Rag activity is supported by an almost complete ablation of NFAT activation in myrPKB-induced T cell lymphoma, which intriguingly express Rag1 and Rag2. PKB-mediated down-modulation of NFAT activity may therefore be an essential parameter in the development of T cell malignancies, and targeting NFAT signaling via PKB as such may underscore new therapeutic strategies.

We thank S. Roth and T. Rüdiger for valuable help with confocal microscopy, R. Kincaid for the ΔCam construct, A. Rao for NFATc2 Ab, the Serfling laboratory for Abs and discussion, M. Klein for Rag1 knockout mice, and T. Twardzik for collaboration at an early stage of the project. We also thank G. Weitz for technical assistance and H. Wolff for animal maintenance.

The authors have no financial conflict of interest.