Notch Target Gene Deregulation and Maintenance of the Leukemogenic Phenotype Do Not Require RBP-Jκ in Ikaros Null Mice

Abstract

Ikaros and Notch are transcriptional regulators essential for normal T cell development. Aberrant activation of Notch target genes is observed in Ikaros-deficient thymocytes as well as leukemia cell lines. However, it is not known whether Notch deregulation plays a preferential or obligatory role in the leukemia that arise in Ikaros null (Ik^−/−) mice. To answer this question, the expression of the DNA-binding Notch target gene activator RBP-Jκ was abrogated in Ik^−/− double-positive thymocytes. This was accomplished through conditional inactivation using CD4-Cre transgenic mice containing floxed RBP-Jκ alleles (RBPJ^fl/fl). Ik^−/− × RBPJ^fl/fl × CD4-Cre⁺ transgenic mice develop clonal T cell populations in the thymus that escape to the periphery, with similar kinetics and penetrance as their CD4-Cre⁻ counterparts. The clonal populations do not display increased RBP-Jκ expression compared with nontransformed thymocytes, suggesting there is no selection for clones that have not fully deleted RBP-Jκ. However, RBPJ-deficient clonal populations do not expand as aggressively as their RBPJ-sufficient counterparts, suggesting a qualitative role for deregulated Notch target gene activation in the leukemogenic process. Finally, these studies show that RBP-Jκ plays no role in Notch target gene repression in double-positive thymocytes but rather that it is Ikaros that is required for the repression of these genes at this critical stage of T cell development.

Regulatory factors that are essential for proper T cell development are often common targets for mutation in T cell leukemias. Two such regulators of thymocyte differentiation are Ikaros and Notch, which have both been implicated as important targets for genetic mutation in human acute lymphoblastic leukemia (ALL). High levels of non-DNA binding (dominant-negative) Ikaros isoforms have been observed in leukemic cells from children with T- and B-ALL (1, 2). More recently, mutations in IKZF1, the gene encoding Ikaros, were identified as a strong predictor of increased likelihood of relapse in children with BALL (3, 4) and as a common genetic lesion in Philadelphia chromosome-positive adult ALL (5). In addition, activating mutations in Notch have been observed in >50% of human T-ALLs that have been surveyed (6). Interestingly, activating mutations in Notch have been found in the majority of T leukemia cell lines generated from the naturally occurring leukemia that develops in genetically engineered Ikaros-deficient mice (7–9).

The Notch receptor is a transmembrane protein that undergoes γ-secretase–dependent cleavage upon ligand recognition. This cleavage frees the intracellular domain, which travels to the nucleus. In the nucleus, intracellular Notch (ICN) activates transcription of Notch target genes through association with the DNA-binding factor recombination signal binding protein of Ig Jκ, RBP-Jκ (RBPJ). The activation complex has been purified and shown to minimally consist of a ternary structure containing ICN, RBPJ, and proteins of the Mastermind-like family (10, 11). The mechanism of Notch target gene repression, as potentially mediated by RBPJ, however, is less well understood. In vivo experiments have revealed that deletion of RBPJ in thymocytes does not result in derepression of the Notch target genes Hes1, Hes5, or Pta (12), indicating that RBPJ expression is not required for Notch target gene repression in these cells. However, deletion of the corepressor Mint or Ikaros results in derepression of Notch target genes in double-negative (DN) and double-positive (DP) thymocytes, respectively (8, 13, 14). Furthermore, Ikaros and RBPJ share a similar DNA-binding consensus sequence (15), and Ikaros and RBPJ bind to the Hes1 promoter in a coordinated fashion, which correlates with downregulation of Hes1 expression in a thymocyte cell line (16). However, the relative contributions and requirements for Ikaros versus RBPJ in repression of Notch pathway genes have not yet been addressed.

When Ikaros levels are reduced or completely ablated in mice, thymocytes undergo leukemogenesis with 100% penetrance (17, 18). Leukemia in Ik^−/− mice was defined by identification of clonally expanded T cell populations in the thymus arising as early as 4 wk after birth, which eventually can also be identified in the spleen (18). Overexpression of ICN in bone marrow progenitors also induces T cell transformation upon transplantation into recipient mice (19, 20). In mice, in a study to define cooperative mutations in ICN-induced leukemia, loss-of-function Ikaros mutations were identified as the most frequently occurring genetic lesion (15). However, we have recently demonstrated that deregulated expression of Notch target genes occurs in Ik^−/− thymocytes prior to their transformation, and this aberrant expression is observed in the absence of ICN. Accordingly, this led us to hypothesize that deregulated RBPJ-dependent Notch target gene activation, through aberrant ICN generation/stabilization, is not required for leukemogenesis in Ik^−/− mice.

To test this hypothesis, as well as to investigate the role of Ikaros and RBPJ in Notch target gene repression, we significantly decreased levels of RBPJ-dependent Notch target gene activation in Ik^−/− DP thymocytes through conditional inactivation of RBPJ. First, using this model, we show that Ikaros, but not RBPJ, is required for repression of Notch target genes in DP thymocytes. Second, we show that deregulated RBPJ-dependent Notch target gene activation is not required to maintain the leukemogenic phenotype in Ik^−/− mice, although it does appear to contribute to the aggressiveness of the leukemia.

Materials and Methods

Mice

Ikaros null mice (C57BL/6 × SV129) were generated by intercrossing of Ikaros null heterozygotes. RBP-Jκ floxed mice have been described previously (12) (a gift from T. Honjo, Kyoto University Graduate School of Medicine, Kyoto, Japan). Genotypes were assessed by PCR analysis of tail DNA as described previously (18). CD4-Cre transgenic (Tg) mice (C57BL/6) were obtained from Taconic Farms (Germantown, NY). All animal procedures were approved by the Northwestern University Animal Care and Use Committee.

Cell lines and cell culture

JE131 and D510 cell lines have been described previously (21, 22). All cell lines were maintained in RPMI 1640 medium (Life Technologies, Rockville, MD) supplemented with 10% FBS (Sigma-Aldrich, St. Louis, MO), 50 μM 2-ME, and 100 U/ml penicillin-streptomycin.

c-myc inhibitor cultures and MTT assay

JE131 and D510 cells were plated at 4 × 10⁴ cells/well and treated with increasing concentrations of the c-Myc inhibitor 10058-F4 (Calbiochem, San Diego, CA). At 72 h, 20 μl MTT (Sigma-Aldrich) at 5 mg/ml was added to each well. After incubation at 37°C for 3 h, the MTT medium was removed, and 100 μl DMSO was added. Absorbance was measured using the 540-nm wavelength on a spectrophotometer.

Protein preparation and immunoblotting

Protein extracts were prepared by whole-cell lysis with 420 mM NaCl lysis buffer (20 mM Tris [pH 7.5], 0.1% BSA, 1 mM EDTA, and 1% Nonidet P-40) supplemented with protease inhibitors. Abs against cleaved Notch1-val1744 (Cell Signaling Technology, Beverly, MA) or c-Myc (N-262) (Santa Cruz Biotechnology, Santa Cruz, CA) were diluted 1/500 in TBS-Tween-5% milk and incubated with the membrane overnight at 4°C. Quantitative densitometric analysis of select gel band intensities was performed using ImageJ software (version 1.43r; National Institutes of Health, Bethesda, MD).

Abs and flow cytometry analysis

For flow cytometric analysis, the following Abs were used: anti-CD4 (GK1.5) and anti-CD8 (53-6.7) (eBioscience, San Diego, CA). For analyses of Vβ-chain usage, a Mouse Vβ-TCR Screening Panel kit (BD Biosciences, San Jose, CA) was used. Abs were allophycocyanin, FITC, or PE conjugates. Cells were analyzed on a FACSCanto II (BD Biosciences) flow cytometer using Flow Jo software. Clonal populations were sorted on a Dako MoFlo cell sorter (Dako, Carpinteria, CA) at the Lurie Comprehensive Cancer Flow Cytometry Facility (Northwestern University, Chicago, IL).

PCR analysis of TCR gene structure

DNA was prepared from thymocytes and subjected to PCR to amplify Dβ2-Jβ2 rearrangements as described previously (17) Primer sequences used were as follows: Dβ2.1, GTA GGC ACC TGT GGG GAA GAA ACT; and Jβ2.7, TGA GAG CTG TCT CCT ACT ATC GAT T.

RT-PCR

mRNA was prepared from primary thymocytes using the SV Total RNA Isolation System (Promega, Madison, WI). cDNA was generated with a Superscript III kit or Superscript VILO kit (Invitrogen, Carlsbad, CA). Quantitative PCR was performed on cDNA (Bio-Rad MyiQ Real-Time PCR machine; Bio-Rad, Hercules, CA) and analyzed using the Pfaffl method and are shown as ratios ([E_target]^{−CT target}/[E_reference]^{−CT reference}, where E is efficiency of PCR; target is gene of interest; reference is HPRT). Primers used for real-time PCR were generated by the Beacon Design program and synthesized by IDT DNA Technologies. Primer sequences are available upon request.

Results

Conditional intrathymic deletion of RBPJ

To address the roles of RBPJ and Ikaros in the repression of Notch target genes, we analyzed mice with floxed RBPJ alleles (RBPJ^fl/fl) that have been bred onto the Ikaros null (Ik^−/−) genetic background. Conditional inactivation of RBPJ was achieved by placing the expression of the Cre recombinase under the control of the CD4 promoter/enhancer/silencer (23). Our rationale was that this strategy would result in mice that lack Ikaros and also lack the ability to activate Notch target genes beginning at the late DN to early DP stage of thymocyte differentiation, because this timing for deletion has been demonstrated in many publications (12, 24, 25). Inactivation of RBPJ at earlier stages of thymic T cell development, using an lck-Cre transgene, for example, results in significant impairment of T cell development (12), making this strategy unsuitable for these studies. Four genotypes of mice were analyzed: 1) RBPJ^fl/flCD4-Cre⁻Ik^+/+ (RBPJ⁺Ik⁺), 2) RBPJ^fl/flCD4-Cre⁺Ik^+/+ (RBPJ⁻Ik⁺), 3) RBPJ^fl/fl CD4-Cre⁻Ik^−/− (RBPJ⁺Ik⁻), and 4) RBPJ^fl/flCD4-Cre⁺Ik^−/− (RBPJ⁻Ik⁻). To confirm abrogated RBPJ expression, we purified RNA from sorted thymocyte subsets of 3-wk-old mice and assessed RBPJ expression levels using quantitative RT-PCR (qRT-PCR) (Fig. 1A, 1B). At the DN stage, in Ik^+/+ and Ik^−/− mice, there was a 1.5 and 41.3% decrease, respectively, in RBPJ expression. The increased abrogation of RBPJ expression observed in Ik^−/− DNs as compared with their Ikaros wild-type counterparts is not surprising because Ikaros has been shown to contribute to CD4 silencer activity specifically as thymocytes transit from DN3 to DN4 (26). RBPJ expression was more dramatically downregulated at the DP stage of development, where in Ik^+/+ and Ik^−/− mice, there was an 85 and 88% respective decrease in RBPJ expression.

FIGURE 1.

Abrogated expression of RBPJ occurs in DP thymocytes of RBPJ^fl/fl × CD4-Cre Tg mice. A, qRT-PCR analyses were performed using cDNA prepared from sorted DN and DP thymocyte subsets from young mice (3 wk of age) of each genotype as shown. Y-axis shows RBPJ expression normalized to expression of HPRT. Error bars represent SEM. Experiments were performed from at least two sets of pooled mice, and qRT-PCR was performed at least three times in duplicate. Values of p were calculated using Student t tests (*p ≤ 0.01). B, Surface expression of CD4 and CD8 on thymocytes from mice of the designated genotypes is depicted with percentages of DN and DP subsets in the corresponding gate. Circles indicate the sorted populations used for gene expression analyses shown in A.

Ikaros, but not RBPJ, is required for Notch target gene repression in DP thymocytes

RBPJ has long been postulated to play an integral role as a Notch target gene repressor. However, previous studies have reported that Ik^−/− DP thymocytes display deregulated expression of Notch target genes (8, 14) suggesting that Ikaros is also important for repression of these genes. Several groups have hypothesized a functional interaction between Ikaros and RBPJ in the regulation of Notch target genes using data generated in cells lines and in vitro assays (7, 15). However, the contributory roles of Ikaros and RBPJ in Notch target gene repression have not been assessed in vivo in a cell type in which Notch target genes are actively repressed, such as in DP thymocytes. Therefore, using the RBPJ conditional knockout system, the role of RBPJ in the Ikaros-dependent repression of Notch target genes in DP thymocytes was assessed. DP thymocytes from RBPJ⁻Ik⁺ mice do not exhibit derepression of the Notch target genes Hes1, Deltex1, pTa, Runx1, or Notch1 (Fig. 2A). In contrast, mice whose DP thymocytes lack expression of Ikaros alone (RBPJ⁺Ik⁻) or both Ikaros and RBPJ (RBPJ⁻Ik⁻) exhibit significantly increased Notch target gene expression (Fig. 2A). Levels of Notch target gene expression are equivalent in RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ DP thymocytes, proving that abrogation of RBPJ expression does not further deregulate levels of expression. Taken together, these data clearly demonstrate that Notch target gene repression in DP thymocytes does not require RBPJ and that Ikaros is a required repressor in this subset. Importantly, they also indicate that the increased expression of Notch target genes observed in the absence of Ikaros does not require RBPJ-dependent Notch target gene activation and therefore that the observed deregulated expression is most likely because of derepression.

FIGURE 2.

Ikaros, but not RBPJ, is required for Notch target gene repression in DP thymocytes. A and B, qRT-PCR analyses were performed using cDNA prepared from sorted DP thymocyte subsets. Y-axis shows expression normalized to expression of HPRT. Error bars represent SEM. Experiments were performed from at least two sets of pooled mice, and qRT-PCR was performed at least three times in duplicate. Values of p were calculated using Student t tests (*p ≤ 0.05; **p ≤ 0.001).

It has been demonstrated that Ikaros binds directly to regulatory regions of the Notch target genes, Hes1 and Deltex1, suggesting that Ikaros' role in Notch target gene repression is direct (14, 16). However, lack of Ikaros could also have an indirect role if it results in deregulated expression of other known coactivators or corepressors of Notch target genes in thymocytes. Therefore, expression levels of the genes encoding the corepressors MINT (13) and NKAP (27) and the coactivator Mastermind-like 1 (28) were compared in Ikaros null DP thymocytes and their wild-type counterparts. No significant difference in expression of these genes was observed (Fig. 2B).

Clonally expanded T cell populations are detected in thymuses and spleens of RBPJ⁻Ik⁻ mice

We have previously demonstrated that the proliferation of an Ik^−/− T leukemia cell line occurs independently of aberrant Notch activation (21). However, it has not been previously addressed whether aberrant RBPJ-dependent Notch target gene activation is required for the maintenance/survival of clonally expanded T cell populations in vivo in Ik^−/− mice. Therefore, we examined whether abrogation of RBPJ expression in Ik^−/− DP thymocytes could alter the characteristic Ik^−/− leukemic phenotype.

Clonally expanded thymocyte populations appear in Ik^−/− mice as early as 4–5 wk after birth (18). To determine whether clonal populations arise in RBPJ⁻Ik⁻ thymuses, TCRβ-chain rearrangements were examined by analyzing the repertoire of Dβ₂-Jβ₂ rearrangements in the thymus. Young Ik^−/− mice (<4 wk of age) have a polyclonal repertoire similar to what is observed in wild-type mice. However, older RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice (5–10 wk) display the appearance of clonal populations as determined by the preferential amplification of a single band corresponding to a single Dβ₂-Jβ₂ rearrangement (Fig. 3). The presence of clonally expanded populations can also be detected at the protein level by examining the composition of the Vβ repertoire expressed by the TCRβ-chain of the TCR within the population. As observed with the TCRβ-chain rearrangement analyses, young Ik^−/− mice (<4 wk of age) have a polyclonal repertoire similar to what is observed in wild-type mice (Fig. 4A). However, clonally expanded Vβ populations were detected in thymuses of both RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice as early as 5 wk of age (Fig. 4B, 4C).

FIGURE 3.

Loss of polyclonality is observed in thymuses of RBPJ⁻Ik⁻ mice. A, PCR analyses of TCRβ rearrangements in thymocytes from individual mice are shown. Molecular weight marker is 1 kB-plus ladder (Invitrogen). Individual RBPJ⁻Ik⁻ mice are specified by letters, and individual RBPJ⁺Ik⁻ mice are specified by numbers.

FIGURE 4.

Predominant Vβ populations suggest the presence of clonal populations within the thymuses of RBPJ⁻Ik⁻ mice. TCR Vβ staining of thymocytes from indicated mice is shown. Y-axis represents percentage of total cells that express a specific Vβ from each thymus. A, RBPJ⁺Ik⁺ and RBPJ⁻Ik⁺ plots were compiled from staining of multiple 6-wk-old mice (n = 3). RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ plots were compiled from staining of 3-wk-old mice (n = 3). RBPJ⁺Ik⁻ (B) and RBPJ⁻Ik⁻ (C) graphs represent the staining profiles of individual mice.

Another characteristic of leukemia is the appearance of the clonally expanded populations in the periphery. Flow cytometric analyses of spleens from individual mice reveal that in each mouse, the same Vβ population that was expanded in the thymus was also expanded in the periphery (Fig. 5A). Importantly, the expanded Vβ population observed in the thymus and spleen could also be identified as an expanded population in the bone marrow and blood in both RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice (Fig. 5B)

FIGURE 5.

Clonal populations that are detected in the thymus are also detected in the periphery of RBPJ⁻Ik⁻ mice. A, Flow cytometric profiles depicting TCRVβ-chain expression in thymuses and spleens of indicated mice are shown. Black line in histogram plots indicates Vβ staining in RBPJ⁺Ik⁻ (5) and RBPJ⁻Ik⁻ (E, O) mice. Gray area represents profile from corresponding age-matched RBPJ⁺Ik⁺ or RBPJ⁻ Ik⁺ control mice. B, Flow cytometric histograms showing the same expanded Vβ population in thymus, spleen, bone marrow, and blood of an 8-wk-old RBPJ⁺ Ik⁻ mouse as well as within the same tissues of an age-matched RBPJ⁻Ik⁻ mouse. C, Bar graph depicting absolute numbers of thymocytes from RBPJ⁺Ik⁺ and RBPJ⁻Ik⁺ mice (average from three mice of each age) and individual RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice. Age of the mice in weeks is indicated below the bar graph.

Taken together, these analyses of TCRβ rearrangements at the genomic and protein levels provide strong evidence that RBPJ-dependent Notch target gene activation is not required for the survival of clonally expanded T cell populations in Ik^−/− mice. It is important to note, however, that qualitative differences in the phenotype of the leukemias that arise in older RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice are observed. In many cases, the clonal expansion that occurs in RBPJ⁻Ik⁻ thymuses does not increase thymic cellularity to the same degree as occurs in RBPJ⁺Ik⁻ mice (Fig. 5C). Therefore, although RBPJ is not required for clonal expansion in the absence of Ikaros, it may contribute to aggressive proliferation of these cells. This selective advantage may account for the overexpression or the expression of mutated Notch that is commonly observed in Ik^−/− T cell leukemia lines (8). Whether this increased aggressiveness contributes to increased mortality of RBPJ⁺Ik⁻ mice could not be tested. Ik^−/− mice display defects in pituitary-adrenocortical function, and this, rather than a hematopoietic defect, has been shown to be the major contributor to their impaired growth and diminished survival (29). Therefore, mortality of Ik^−/− mice cannot be unambiguously attributed to their leukemia.

The maintenance of clonal populations in RBPJ⁻Ik⁻ mice does not require selection for RBPJ expressing cells

Analyses of RBPJ⁻Ik⁻ mice indicates that survival of clonally expanded T cell populations observed in the absence of Ikaros does not require RBPJ-dependent Notch target gene activation. However, the clonal populations observed in RBPJ⁻Ik⁻ mice could be the result of selection for an outgrowth of cells that still express RBPJ, thus still allowing for their dependence on deregulated Notch signaling through RBPJ for their survival. To rule out this possibility, RBPJ expression was analyzed in thymuses of older RBPJ⁻Ik⁻ mice using qRT-PCR. If the clonally expanded cells represent an outgrowth of cells expressing RBPJ, then higher levels of RBPJ expression should be observed in thymuses of mice with a greater percentage of clonally expanded cells as compared with those with lesser contribution from the clonally expanded population. However, all the RBPJ⁻Ik⁻ thymuses analyzed expressed similar levels of RBPJ, whether as many as 90% (mouse N) or as few as 9% (mouse I) of thymocytes represented the clonally expanded population (Fig. 6A). As a more definitive analysis, expression levels of RBPJ were compared in the clonal versus the nonclonal population in an RBPJ⁻Ik⁻ thymus, in which 53.5% of the cells expressed Vβ3 (Fig. 6B). RBPJ expression was equivalent in both populations. Taken together, these analyses provide strong evidence that clonal expansion does not require selection for cells that still express RBPJ.

FIGURE 6.

The maintenance of clonal populations in Ik^−/− mice does not select for increased levels of RBPJ expression. A, qRT-PCR analyses of RBPJ expression were performed using cDNA prepared from whole thymuses of indicated mice. Y-axis shows RBPJ expression normalized to expression of HPRT. Error bars represent SEM. The numbers shown below the mouse designations represent the percentage of clonally expanded cells as defined by expression of a specific Vβ-chain identified through flow cytometry. The Vβ repertoire of the populations is shown in Fig. 4. B, TCR Vβ3 staining of thymocytes from an individual RBPJ⁻Ik⁻ mouse (10 wk of age). qRT-PCR analyses of RBPJ expression was performed using cDNA from sorted Vβ3+ (clonal) and Vβ3− (nonclonal) populations. Y-axis shows RBPJ expression normalized to expression of HPRT. Error bars represent SEM. There was no statistically significant difference in RBPJ expression between these two populations.

Deregulated ICN expression is commonly observed in RBPJ⁻Ik⁻ thymuses with clonal expansions

Because RBPJ deletion does not occur until the late DN to early DP stage of thymocyte development in RBPJ⁻Ik⁻ mice, this model cannot be used to definitively address the requirement for RBPJ in the initiation of transformation. However, to begin to address this question, analyses of ICN expression in the thymuses of older RBPJ⁻Ik⁻ mice were performed. Deregulated ICN expression in Ikaros-deficient leukemia has been shown to be the result of mutations at the DNA level (7, 8), which can yield increased ICN generation and/or protein stability (30). Therefore, we predicted that if mutations causing increased ICN expression occurred, even if it was prior to the deletion of RBPJ, these alterations would still be present in the thymuses of RBPJ⁻Ik⁻ mice containing clonally expanded populations. The well-characterized mutations that occur in Notch heterodimerization domain and proline, glutamine, serine, and threonine-rich domain (PEST) regions are not oncogenic on their own (31). Thus, it is more informative to look at alterations in ICN expression at the protein level rather than assay for mutations because this reflects a functional readout for aberrant Notch receptor activation/stability. All RBPJ⁺Ik⁻ thymuses examined from mice with clonal expansions exhibited truncated or overexpressed ICN (Fig. 7). Interestingly, the majority of RBPJ⁻Ik⁻ thymuses that exhibited significant clonal expansions also exhibited truncated or overexpressed ICN. One notable exception was mouse H, in which 50% of the thymocytes represent a clonally expanded population, but which showed relatively normal ICN expression (Fig. 4C). These data suggest that mutations in Notch leading to its overexpression or truncation exert effects required for initiation of leukemia prior to RBPJ deletion. It is also possible that in RBPJ⁻Ik⁻ mice, mutated Notch has important RBPJ-independent functions in maintenance of leukemia.

FIGURE 7.

Thymuses of RBPJ⁻Ik⁻ mice with clonally expanded populations commonly display deregulated ICN expression. ICN expression levels were analyzed by Western blot analyses using whole-cell lysates prepared from the thymuses of indicated mice. Fifty micrograms of protein was loaded per sample. Anti-actin Abs were used as a loading control. Molecular weight markers are shown to the left. Expected size of ICN ∼110 kDa. The intensity of the ICN band, normalized to actin and relative to levels in wild-type (RBPJ⁺Ik⁺), is shown below the image. *, Truncated ICN protein.

Clonal expansion of thymocytes in RBPJ⁻Ik⁻ mice does not require deregulated c-Myc expression

Several groups have reported connections between Notch and/or Ikaros and c-Myc deregulation in the etiology of T-ALL (32–35). More specifically, molecular characterizations of tumor-specific gene expression in ICN-overexpressing tumor cells have identified higher levels of c-Myc expression and have suggested that c-Myc deregulation is required for ICN-dependent tumor development and growth (35).

Therefore, RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice that contain clonal expansions were analyzed for deregulated c-Myc expression. Analyses of c-Myc expression in thymocytes from RBPJ⁻Ik⁻ mice reveal that the majority of mice with clonal populations express similar or lower levels of c-Myc as their Ik⁺ counterparts or as a RBPJ⁻Ik⁻ mouse without a detectable clonal expansion (mouse J) (Fig. 8A). This pattern is also observed in older RBPJ⁺Ik⁻ mice with clonal populations (Fig. 8A). It is important to note that c-Myc levels are not increased in 5- to 6-wk-old Ik^−/− mice with clonal expansions (mice H, I, 4, and 2) and most 9- to 10-wk-old Ik^−/− mice with clonal expansions (mice O, P, and 5), suggesting that increased c-Myc expression is not commonly observed at early or late stages of leukemic transformation in the absence of Ikaros.

FIGURE 8.

Overexpression of c-Myc is not observed in clonally expanded thymocyte populations. A, c-Myc expression levels were analyzed by Western blot analyses using whole-cell lysates prepared from the thymuses of indicated mice. Fifty micrograms of protein was loaded per sample. Anti-actin Abs were used as a loading control. Molecular weight markers are shown to the left. The intensity of the c-myc band, normalized to actin and relative to levels in wild-type (RBPJ⁺Ik⁺) or levels in an RBPJ⁻Ik⁻ thymus with no detectable clonal expansion (J), is shown below the image. B, Effect of c-Myc inhibitor 10058-F4 exposure on cell growth. A total of 4 × 10⁴ cells were plated in 96-well plates (four replicates per experimental point) with increasing concentrations of inhibitor. At 72 h, the plates were analyzed by MTT assay. Experiment was performed twice. Data from a representative experiment are shown.

Although increased c-Myc expression is not commonly observed in clonally expanded Ikaros null thymocytes, this does not prove that c-Myc activity per se is not important for growth and survival of the leukemic cells. Therefore, the effect of a c-Myc inhibitor on growth of two aggressively growing Ikaros-deficient T leukemia lines was examined. The JE131 cells are an Ik^−/− T cell leukemia line that expresses high levels of ICN (8, 16). The D510 cell line was derived from a Tg mouse that expressed high levels of a non–DNA-binding Ikaros isoform exclusively in T cells (22). In contrast to the JE131 cells, it does not express ICN (8). Inhibition of c-Myc activity has profound effects on growth/survival of both of these leukemia cell lines (Fig. 8B). Therefore, c-Myc activity is important for growth/survival of Ikaros-deficient leukemia cells, but this role is independent of mutations in Notch.

Discussion

Prior to the work initiated in this study, a correlative link between Ikaros and Notch in leukemia had been established (7, 8, 15, 33). However, it remained unclear whether Notch deregulation is essential for the survival of clonally expanded thymocyte populations that arise in the absence of Ikaros. Moreover, recent reports by others and us have demonstrated that Notch target genes are derepressed in nontransformed Ik^−/− DP thymocytes in the absence of ICN. This led us to hypothesize that deregulated Notch target gene activation, through aberrant ICN generation/stabilization, may not be required for leukemogenesis in Ik^−/− mice. To test this hypothesis, expression of the DNA-binding Notch target gene activator RBPJ was abrogated in Ik^−/− DP thymocytes through conditional inactivation. In this model, the potentially inappropriate activation of Notch target genes resulting from aberrant expression of ICN is prevented by removing RBPJ in DP thymocytes while preserving the derepression that occurs in the absence of Ikaros. Importantly, these studies have also allowed investigation of the relative contributions of Ikaros and RBPJ to the active repression of Notch target genes in developing thymocytes.

A model using a CD4-Cre transgene to delete RBPJ in DP thymocytes was selected to avoid the inhibitory effecton T cell development that occurs when RBPJ is deleted at earlier stages of thymocyte differentiation. Use of an Lck-Cre transgene to promote deletion of RBPJ could inhibit leukemogenesis indirectly by inducing a developmental block prior to the developmental stage that is susceptible to transformation. In Ikaros-deficient leukemias, this stage is most likely the DP stage because clonal outgrowths are first identified within this population. In addition, in Ik^−/− mice, leukemogenesis can be prevented by blocking T cell development in the thymus prior to the DP stage (36). The DN-DP transition is also considered the target of leukemogenesis in thymocytes when Notch is constitutively activated in an ICN-transduced BM model of T-ALL (35).

Importantly, in this report, we demonstrate that similar clonally expanded T cell populations were observed in the thymus and periphery of Ik^−/− mice both in the presence (RBPJ⁺Ik⁻) or absence (RBPJ⁻Ik⁻) of RBPJ with similar kinetics, showing that RBPJ-mediated Notch target gene activation is not required for the survival of the clonally expanded populations that arise during Ik^−/− leukemogenesis. The clonally expanded populations also do not display selection for increased RBPJ expression, supporting the hypothesis that RBPJ-dependent Notch target gene activation is not required for their survival. These data are concordant with our published leukemia cell line study demonstrating that the ICN-expressing Ik^−/− T cell leukemia line JE131 is refractory to the growth inhibitory effects of γ-secretase inhibitor (16), which inhibits Notch signaling. However, it is important to note that deregulated ICN expression is observed in the majority of analyzed RBPJ⁻Ik⁻ thymuses with clonally expanded populations, highlighting a potentially important role for this event either in the initiation of leukemia, which could occur prior to RBPJ deletion, or in maintenance of leukemia in an RBP-J–independent manner.

Qualitative differences in the phenotypes of the leukemias that arise in RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice were observed, with clonal populations expanding more aggressively in some RBPJ⁺Ik⁻ mice, as opposed to most of their RBPJ⁻Ik⁻ counterparts. These data indicate that deregulated Notch target gene activation through RBP-J may act to increase the aggressiveness of Ik^−/− leukemias. Notch mediates proliferation and survival in normal thymocytes; therefore, these mutations could confer a selective growth advantage to transformed thymocytes. This selective advantage of deregulated ICN was still observed in RBPJ⁻Ik⁻ clonal expansions, even though ICN can only promote RBPJ-independent Notch target gene activation in these cells.

Significantly, in this paper, an important difference was uncovered in Ik^−/− leukemia that differentiate it from leukemias that arise as a result of deregulation of Notch in other mouse models. In many cases, leukemias induced by deregulation of Notch are dependent on high level of expression of c-Myc, which is a direct Notch target gene (32, 35, 37). However, even in RBPJ⁺Ik⁻ and RBPJ⁻Ik⁻ mice with clonal expansions that contained deregulated ICN expression, high levels of c-Myc expression were not observed. These analyses suggest that the mechanism of leukemogenesis in Ik^−/− thymocytes is different from that required in development of Notch-dependent leukemias.

The paradigm of Notch target gene regulation includes RBPJ as an integral component of both the activator and the repressor complexes. In this study, however, we definitively demonstrate that Ikaros is a required component of the repressor complex for Notch target genes in DP thymocytes. In addition, we show Ikaros does not require RBPJ expression for its repressive function. In support of this model, we demonstrate that unlike lack of Ikaros, RBPJ deletion does not induce derepression of Notch target genes. Moreover, the additional deletion of RBPJ in the absence of Ikaros does not exacerbate the level of Notch target gene derepression, thus eliminating a role for cooperative repression between Ikaros and RBPJ. It is important to note that lack of Notch target gene derepression has been observed when RBPJ is inactivated earlier in thymocyte development by Tg expression of lck (proximal promoter)-Cre, but these results are confounded by the heterogeneity of Notch activation/repression in the four DN subsets (12). Thus, we demonstrate for the first time that RBPJ expression is also dispensable for Notch target gene repression in DPs. These data add to the mounting evidence that the switch theory of canonical Notch signaling is not relevant in all cell types in which Notch signaling plays a role.

Abbreviations used in this paper

ALL

acute lymphoblastic leukemia

DN

double-negative

DP

double-positive

ICN

intracellular Notch

qRT-PCR

quantitative RT-PCR

Tg

transgenic