Ikaros is absolutely required for pre-B cell differentiation by attenuating IL-7 signals

We first analyzed Ikaros expression in BM B cells. WT B220⁺ cells were purified into fraction A (pre/pro-B; CD43⁺CD24⁻BP-1⁻), B (early pro-B; CD43⁺CD24⁺BP-1⁻), C (late pro-B; CD43⁺CD24⁺BP-1⁺), C′ (large pre-B; CD43⁺CD24^hiBP-1⁺), D (small pre-B; CD43⁻IgM⁻), and E (B220⁺IgM⁺) cells. Ikaros mRNA levels were high in fractions A and B, reduced in C, and increased in C′, D, and E cells (Fig. 1 A). This pattern suggests an early and late role for Ikaros in B cell development.

To delete Ikzf1 in B cell progenitors, we engineered a floxed Ikaros allele (Ik^f/f) in which exon 8 was flanked by loxP sites. Deletion by the Cre recombinase results in a null allele similar to that described by Wang et al. (1996). Ik^f/f mice were crossed with Mb1-Cre animals, which express Cre under the control of the Cd79a promoter (Hobeika et al., 2006). The resulting Ik^f/f Mb1-Cre⁺ (conditional KO [cKO]) mice showed deletion of the Ik^f/f allele that began in fraction A and was completed in fraction B cells; Ikaros proteins were likewise detected in fraction A but not in fraction B cells (Fig. S1). In the BM, adult cKO mice exhibited an almost complete block in B cell differentiation at the B220⁺CD43⁺ stage (Fig. 1 B). Furthermore, fraction C′ cells accumulated in cKO mice compared with WT (Ik^f/f Mb1-Cre⁻). No differences were found in intracellular μ expression and the proliferative state of WT and cKO fraction C and C′ cells (not depicted). In the periphery, B220⁺CD19⁺ splenic B cells, as well as peritoneal B2, B1a, and B1b cells, were almost completely lost in cKO animals.

A few B cells were observed in the cKO BM and periphery. To determine whether they had deleted the Ik^f/f alleles, we evaluated them for intracellular Ikaros (Fig. 1 C). Although Ikaros was not detected in most fraction C and C′ cells, it was expressed in cKO BM B220⁺CD43⁻ cells. Furthermore, Ikaros^hi and Ikaros^med B cells were often detected in cKO splenic CD19⁺ B cells, indicating that these cells had retained at least one functional Ikzf1 allele, probably as the result of incomplete deletion. We therefore conclude that Ikaros is absolutely required for B cell differentiation from the fraction C′ stage.

As IKZF1 deletions are associated with human B-ALL, we asked whether Ikaros deficiency in B cells leads to the development of B cell leukemias. cKO (n = 7) and WT (n = 7) mice were followed over 18 mo. Old cKO mice showed a similar block at the pre-B cell stage but were otherwise healthy (Fig. 1 D). No apparent signs of leukemia were detected (not depicted). Thus, Ikaros deletion in fraction B cells does not initiate transformation.

To determine whether Ikaros reexpression rescues pre-B cell development, we set up an in vitro gain-of-function system using cKO and WT BM cells. As pro/pre-B cells can be expanded in IL-7 and induced to differentiate upon IL-7 withdrawal, multiple IL-7–dependent cultures were set up for both genotypes. WT cultures were B220⁺CD43⁺CD24⁺BP-1^−/+; they resembled fraction B and C cells and expressed λ5 but not surface μ (Fig. S2 A). Very few fraction C′ cells were present, which mimicked the in vivo phenotype. These cultures grew poorly and could not be maintained over time (not depicted). In contrast, cKO cultures grew robustly and were easily maintained as cell lines (BH1-3). The results presented here were obtained with BH1 cells; they were reproducibly repeated with BH2 and BH3 cells. BH1 cells showed deletion of the Ik^f/f allele and lacked Ikaros proteins (Fig. S2, B and C). They resembled fraction C′ cells (B220⁺CD43⁺CD24^hiBP-1⁺, λ5⁺, surface μ⁺; Fig. S2 A). Stimulation with anti-μ antibodies induced an intracellular Ca²⁺ response (Fig. S2 D), indicating the presence of a functional pre-BCR. Conversely, WT cultures did not respond to anti-μ, although they displayed a Ca²⁺ flux to ionomycin. BH1 cultures therefore recapitulated the in vivo accumulation of fraction C′ cells in cKO animals. BH1 and WT cultures were subjected to IL-7 withdrawal, and the expression of κ/λLC and μHC was monitored (Fig. 2 A). Although WT cultures differentiated into BCR⁺ cells, BH1 cells did not, indicating that Ikaros deficiency blocks pre-B cell differentiation in vitro.

BH1 cells were transduced to express the Ikaros-1 isoform fused to the ligand-binding domain of the estrogen receptor (Ik1-ER) in conjunction with GFP. Ik1-ER translocates from the cytoplasm to the nucleus upon 4-hydroxytamoxifen (4OHT) treatment. However, GFP⁺ BH1–Ik1-ER cells were very sensitive to Ikaros activation and died within 24 h, even in the presence of IL-7, although they expressed similar levels of Ikaros mRNA compared with WT fraction C′ and D cells (not depicted). We therefore transduced BH1–Ik1-ER cells to express the antiapoptotic protein BCL-2 and a DsRed reporter, which efficiently prevented apoptosis in 4OHT-treated GFP⁺DsRed⁺ BH1–Ik1-ER–Bcl2 cells (not depicted). As negative controls, BH1–NLS-ER–Bcl2 cells, which expressed a nuclear translocation signal sequence fused to ER (to control for nonspecific ER- and Bcl-2–related effects), and BH1–IkΔDBD-ER–Bcl2 cells, which expressed Ik1 without the DNA binding domain, were generated. Substantial Ik1-ER and IkΔDBD-ER proteins were translocated to the nucleus upon 4OHT treatment, indicating that these proteins were stable and functional (Fig. S2 E). When BH1–Ik1-ER–Bcl2 cells were subjected to IL-7 withdrawal and 4OHT, pre-B cell differentiation was rescued, as measured by surface μ and κ/λLC expression (Fig. 2 B). Interestingly, Ik1 expression also induced differentiation in the presence of IL-7, although to a lesser degree. This effect was dose dependent, as GFP^hi cells showed more μ⁺κ⁺ cells than GFP^med cells (Fig. 2 C). No differentiation was observed in NLS-ER or IkΔDBD-ER cells (Fig. 2 D), demonstrating the importance of DNA binding for Ikaros function.

To determine whether Ikaros rescues pre-B cell differentiation in vivo, BH1–Ik1-ER–Bcl2 cells were adoptively transferred into Rag1^−/− mice. These mice were treated with tamoxifen (TAM) for 5 d and evaluated 2 d later (Fig. 2 E). Significantly more B cells expressing κLC and IgD were detected in the BM and spleens of TAM- versus vehicle-treated mice. The small number of differentiated B cells in the control mice was puzzling and may reflect an in vivo leakiness of the Ik1-ER or improved pre-B cell survival because of the BCL2 protein. These results demonstrate that Ikaros alone is sufficient to induce pre-B cell differentiation in vitro and in vivo.

We first analyzed the impact of Ikaros on proliferation. Ikaros activation in the presence of IL-7 reduced the proliferation of BH1–Ik1-ER–Bcl2 cells (Fig. 3 A), an effect not seen in BH1–IkΔDBD-ER–Bcl2 cells (not depicted). As expected, IL-7 withdrawal alone led to an almost complete proliferation arrest. The combination of Ikaros and IL-7 withdrawal further reduced proliferation in a small but significant manner. The reduction in proliferation was caused by a decrease in cycling cells and not an increase in cell death (not depicted). Similar results were observed with Myc expression. Both Ikaros and IL-7 withdrawal alone reduced Myc mRNA levels, and this effect was cumulative (Fig. 3 B). Thus Ikaros and IL-7 withdrawal cooperate to promote cell cycle exit.

We then evaluated pre-BCR expression, Ig LC germline transcription, and LC rearrangement. 4OHT-treated BH1–Ik1-ER–Bcl2 cells decreased λ5 expression (Fig. 3 C), indicating loss of pre-BCR expression. This was Ikaros specific as it was not seen with IL-7 withdrawal alone. However, the combination of Ikaros and IL-7 withdrawal enhanced λ5 down-regulation. Strikingly, Ikaros induced the appearance of Igκ germline transcripts as early as 4 h after 4OHT treatment (Fig. 3 D); at 6 h, this expression reached that of fraction D cells. Igκ germline transcripts were detected in the presence of IL-7 and increased upon IL-7 withdrawal. IL-7 withdrawal had no effect on its own without Ikaros. We then asked whether Ikaros affects LC rearrangement. In the absence of Ikaros, BH1–Ik1-ER–Bcl2 cells did not express detectable Rag1,2 mRNA (Fig. 3 E). However, Ikaros induced Rag1,2 mRNA levels in the presence of IL-7, which was enhanced more than fivefold upon IL-7 withdrawal. In addition, Ikaros alone induced κ/λLC rearrangement, and IL-7 withdrawal further enhanced rearranged κ/λLC mRNA levels >10-fold (Fig. 3 F). Thus, Ikaros cooperates with IL-7 withdrawal to allow cell cycle exit, pre-BCR down-regulation, and LC expression.

To determine whether a functional BCR is sufficient to overcome the loss of Ikaros, we crossed cKO mice with MD4 tg animals, which express a BCR specific for hen egg lysozyme (Goodnow et al., 1988). cKO MD4⁺ cells remained blocked at fraction C′ (Fig. 4 A), and their surface marker expression was similar to that of cKO MD4⁻ cells, even though the tg BCR was expressed at the proper fraction B and C′ stages (Fig. 4 B). Thus, Ikaros is required for other events in addition to promoting BCR expression.

To gain insight into the gene expression changes induced by Ikaros, we analyzed the effects of Ikaros reexpression in BH1–Ik1-ER–Bcl2 cells at 6 and 24 h by transcriptome profiling, in the presence of IL-7 to focus on Ikaros-associated events. 4OHT treatment for 24 h resulted in the activation and repression (>1.4-fold in both experiments) of 536 and 399 genes, respectively (Fig. 5 A). Most genes were already altered after 6 h (Fig. 5 B), including most of the down-regulated (363/399 genes or 91%; cluster 2) and half of the up-regulated genes (285/536 genes or 53%; cluster 1). We then compared our data with the published genome-wide Ikaros binding profiles in primary pre-B cells to identify direct Ikaros target genes (Ferreirós-Vidal et al., 2013). Ikaros binding was found in >80% of the genes up- or down-regulated at both 6 and 24 h, suggesting that many of the Ikaros-regulated genes may be direct targets (see Fig. S3 for representative genes).

To determine whether the genes regulated by Ikaros in BH1 cells correspond to genes normally regulated during pre-B cell differentiation, we compared our data with those from the ImmGen Consortium for fractions A–F (Fig. 5 C). Comparisons of genes changed after 6 h of 4OHT (clusters 1 and 2) revealed that most of the Ikaros-activated genes were up-regulated in fractions C′ and D (190 genes; left), whereas those repressed by Ikaros were down-regulated in fractions B/C (234 genes), C′ (24 genes), and D (54 genes). Conversely, when we compared the genes up-regulated between fractions C′ and D (651 genes with >1.5-fold change) with those activated by Ikaros in BH1 cells (Fig. 5 D), 400 genes (60%) were common to both. In contrast, only 76 genes (9%) down-regulated between fractions C′ and D were repressed by Ikaros in BH1 cells (not depicted). These results suggest that Ikaros functions mainly as an activator in fraction C′.

Our results suggest that Ikaros cooperates with IL-7 withdrawal for multiple events. We therefore evaluated the gene expression changes in BH1–Ik1-ER–Bcl2 cells in response to Ikaros and/or IL-7 withdrawal (Fig. 5 E). 4,592 genes were up- or down-regulated >1.4-fold between conditions in two independent experiments. Of these, Ikaros cooperated with IL-7 withdrawal to up-regulate the expression of 1,832 genes. Most (n = 1,411) were separately induced by Ikaros or IL-7 withdrawal, and their expression was further increased under both conditions. They included genes important for pre-BCR and BCR signaling (e.g., Blnk, Syk, Lyn, Btk, and Pten) and B cell regulators like Pax5, Foxo1, and Ebf1. A second group (n = 421) was induced only by Ikaros, and their expression increased upon IL-7 withdrawal (e.g., Cdkn1b, Spib, and Bcl2l1). In contrast, a smaller group was induced only by IL-7 withdrawal, but was less activated upon Ikaros reexpression (e.g., Ikzf3, Irf8, and Bcl6).

Ikaros also cooperated with IL-7 withdrawal to down-regulate 1,860 genes. These genes were primarily down-regulated by either Ikaros or loss of IL-7 or similarly affected by either condition. In all cases, the combination of Ikaros and IL-7 withdrawal led to a cumulative down-regulation of gene expression. Affected genes included Myc and genes implicated in cell growth and proliferation (i.e., tRNA biosynthesis, nucleotide biosynthesis, glycolysis, mitochondrial function, and DNA replication), which are broadly influenced by MYC (Eilers and Eisenman, 2008). Only a few genes were singularly regulated by Ikaros or IL-7 withdrawal and not further affected by the combination of both (Fig. 5 E, bottom).

Importantly, Ikaros activated 155/210, and repressed 116/139, genes between fractions C′ and D that were similarly regulated by IL-7 withdrawal. This corresponds to 75% of the genes regulated by Ikaros at these stages (Fig. 5 F). These data highlight a strong synergy between Ikaros and IL-7 withdrawal during pre-B cell differentiation.

Our results demonstrate that Ikaros is a major, nonredundant, regulator of B cell development. The Ikaros mutant phenotype closely resembles that of Irf4,8^−/− double mutant mice, which also show a block at the fraction C′ stage (Lu et al., 2003), strongly suggesting that Ikaros and IRF4,8 share a common pathway. Interestingly, Irf4,8^−/− pre-B cells do not express Ikaros, and either IRF4 or IRF8 can activate Ikaros transcription (Ma et al., 2008). Thus, Ikaros may be a critical downstream target of IRF4,8 in pre-B cells.

Ikaros is required for pre-BCR down-regulation and Ig LC expression. In addition to its known effect on λ5 repression (Sabbattini et al., 2001; Thompson et al., 2007; Ma et al., 2008), Ikaros rapidly activates Igκ germline transcription. Interestingly, Ikaros binds the Ig Eκ3′ and Eκi enhancers in pre-B cells (Fig. S3). Ikaros appears to override the repressive effect of IL-7–STAT5 on Igκ transcription (Malin et al., 2010; Mandal et al., 2011), as Igκ transcription is induced by Ikaros in the presence of IL-7, and Ikaros is required for Igκ transcription upon IL-7 withdrawal. Ikaros also activates Rag1,2 transcription to promote LC recombination and binds the Erag enhancer and both Rag1,2 promoters (Fig. S3). As Ikaros also induces the expression of Foxp1, Foxo1, and Foxo4, as well as Gadd45a and Pten (Fig. 5, D and E), which promote FOXO function (Amin and Schlissel, 2008; Herzog et al., 2008), Ikaros may increase the expression of critical regulators of Rag1,2 transcription in a feed forward mechanism.

These events are essential for BCR expression but not sufficient for differentiation. We found that Ikaros regulates a large cohort of fraction C′ genes. These include genes involved in lymphocyte trafficking (S1pr1), actin filament organization, and cell migration (Coro1a and Asb2), suggesting that Ikaros may influence pre-B cell migration from IL-7–rich pockets. Ikaros up-regulates the expression of the genes encoding folliculin (Flcn) and folliculin-interacting protein 1 (Fnip1), important for energy homeostasis, which were recently identified as essential checkpoints in pre-B cells (Baba et al., 2012; Park et al., 2012). Ikaros also down-regulates genes involved in the mitochondrial creatine pathway (ckmt1 and gamt), implicated in ATP synthesis. Furthermore, Ikaros affects genes linked to redox homeostasis, like Slc7a11 (repressed) or Ero1b and Mgst1 (activated). Lastly, Ikaros activates the expression of prosurvival genes like Bcl2l1, which encodes BCL-xL, and Gadd45a,b (Fig. 5, D and E; Engelmann et al., 2008) physiologically induced in fraction C′ cells. As BCL-xL is an essential survival factor in pre-B cells (Motoyama et al., 1995), its low expression in cKO pre-B cells may explain why the abrupt reintroduction of Ikaros induces cell death in BH1 cells. Thus, Ikaros is important for the regulation of many developmentally regulated genes, but future studies will be needed to distinguish direct from indirect targets.

An important question is how Ikaros is regulated during pre-B cell differentiation. Ikaros transcripts and protein are down-regulated in fraction C and up-regulated in fraction C′ cells, suggesting that Ikaros function may dip when the pre-BCR is first expressed, and that pre-BCR signaling may up-regulate Ikaros expression. Furthermore, FOXO1 may regulate Ikaros through correct mRNA splicing (Alkhatib et al., 2012), and SYK may influence Ikaros localization and binding activity (Uckun et al., 2012). Additional research will be needed to determine how Ikaros activity is modulated during differentiation.

Finally, our results suggest that Ikaros deficiency may contribute to B-ALL development by antagonizing STAT5 function. Indeed, IKZF1 mutations are strongly associated with events leading to STAT5 activation (BCR-ABL expression or CRLF2 amplifications; Mullighan et al., 2008; Harvey et al., 2010). IKZF1 haploinsufficiency may enhance STAT5 signaling in tumor cells where STAT activity is already up-regulated. This may explain why IKZF1 deletions are frequently secondary hits in B-ALL (Kastner et al., 2013) and why Ik^f/f Mb1-Cre⁺ mice do not develop disease.

The key to Ikaros function might thus lie in its ability to turn off the IL-7–dependent gene expression program. As IL-7 functions through STAT5A,B, Ikaros may target the machinery that activates STAT5 or interact with essential STAT5-responsive genes to modulate their transcription. Interestingly, only a small fraction of Ikaros-regulated genes appear to be critical for its function in early B cell development and as a tumor suppressor (Schjerven et al., 2013). These mechanisms will be important to elucidate in the future.

The Ik^f/f mouse line was engineered by inserting loxP sites around exon 8 via homologous recombination in ES cells (performed at the Institut Clinique de la Souris), similar to the mutation described by Wang et al. (1996). To generate Ik^f/f Mb1-Cre⁺ mice, Ik^f/f mice were bred with Mb1-Cre tg mice (gift of M. Reth, Max Planck Institute of Immunobiology and Epigenetics, Freiburg, Germany; Hobeika et al., 2006) and progeny were backcrossed to obtain Ik^f/f Mb1-Cre⁺ (cKO) and control mice (Ik^f/f Mb1-Cre⁻). The MD4 and Rag-1^−/− mouse lines were obtained from the Jackson Laboratory (Tg(IghelMD₄)₄Ccg and Rag1^tm1Mom strains). All animal experiments were approved by the Institut de Génétique et de Biologie Moléculaire et Cellulaire (IGBMC) ethical committee (Com’Eth #2012-092).

The cKO and WT pro/pre-B cell lines were established from BM cells from Ik^f/f Mb1-Cre⁺ or Ik^f/f Mb1-Cre⁻ mice, respectively. These lines were cultured in Iscove’s medium supplemented with 10% FCS, 100 U/ml penicillin-streptomycin, 2 mM glutamine, 50 µM 2-mercaptoethanol, and 8% conditioned medium from a J558L cell line stably overexpressing mouse IL-7 (gift of M. Reth; Winkler et al., 1995). Retroviruses (pMCSV-mBcl2-DsRed [gift of J. Ghysdael, Institut Curie, Orsay, France; Gachet et al., 2013], pMIG-Ik1-ER [generated by A.S. Geimer Le Lay, IGBMC, Illkirch, France], pMIG-NLS-ER, and pMIG-Ik1ΔDBD-ER [lacking Ikaros amino acids 119–223; generated by A. Apostolov, IGBMC]) were generated by CaCl₂ transfection of the EcoPhoenix cell line (gift of G. Nolan, Stanford University, Stanford, CA). The Mig vector was a gift of W. Pear (University of Pennsylvania, Philadelphia, PA). Retrovirus-containing supernatants were harvested after 36 h. Viral infections of 2 × 10⁵ BH1 cells were performed by the addition of 500 µl supernatant followed by 3-h centrifugation at 2,000 rpm. Cells were expanded and sorted based on the expression of GFP and DsRed.

Cells from total BM, spleen, or peritoneal cavity (PEC) were analyzed (LSRII; BD) or purified (FACSAria SORP; BD) as fraction A (B220⁺CD19⁻IgM⁻CD43⁺CD24⁻BP-1⁻), B (B220⁺CD19⁺IgM⁻CD43⁺CD24⁺BP-1⁻), C (B220⁺CD19⁺IgM⁻CD43⁺CD24⁺BP-1⁺), C′ (B220⁺CD19⁺IgM⁻CD43⁺CD24^hiBP-1⁺), D (B220⁺CD19⁺IgM⁻CD43⁻), E (B220⁺CD19⁺IgM⁺CD43⁻), splenic (B220⁺CD19⁺), PEC B2 (CD19⁺CD11b⁻CD5⁻), B1a (CD19⁺CD11b⁺CD5⁺), and B1b (CD19⁺CD11b⁺CD5⁻) B cells. Purity was >98%. Help for cell sorting was provided by C. Ebel (IGBMC). The following antibodies or reagents were used: anti–B220-PECy7 (BioLegend), anti–CD43 (S7)-PE, anti–BP-1–FITC, anti–CD24-bio, anti–λ5 (LM34)-bio, anti–Igκ-bio, anti–CD19-PerCPCy5.5 (BD), anti–IgD (11-26c)-eFluor450N, anti–CD5 (53-7.3)-bio, anti–CD11b (M1/70)-APC (eBioscience), anti–mouse IgG-Cy5, anti–IgM-Cy5 (Jackson ImmunoResearch Laboratories, Inc.), anti–Igλ-bio (SouthernBiotech), streptavidin–Alexa Fluor 405 (Invitrogen), anti-Ikaros (A3; in-house rabbit polyclonal antibody against the Ikaros C terminus), and anti-HEL (gift of A.S. Korganow, Institut de Biologie Moléculaire et Cellulaire, Strasbourg, France). Intracellular staining was performed according to the Foxp3 staining protocol (eBioscience). Dead cells were excluded with 10 µg/ml propidium iodide. Cell cycle analysis was performed with DAPI. For staining of the HEL-specific BCR, cells were saturated with 200 ng/ml HEL, followed by incubation with the anti-HEL antibody. Data were analyzed with FlowJo software (Tree Star).

For in vitro differentiation, 10⁶ cells were plated in Iscove’s medium supplemented or not with 8% IL-7 or 100 nM 4OHT (Sigma-Aldrich). For in vivo differentiation, sublethally irradiated (5 Gy) Rag-1^−/− mice were injected i.v. with 2 × 10⁷ cells. From day 2 after transplantation, mice were injected daily i.p. with 500 µg TAM (Sigma-Aldrich) in sunflower oil for 5 d, and B cells were assessed 2 d after the last injection.

Genomic PCR amplification of Ikaros was performed with the following primers: Ikzf1 flox, 5′-GAGGACCAGATATAAGGCAGCTGG-3′ and 5′-GGCCATCAACGGCATGGAAACGATAA-3′; and Ikzf1 deletion, 5′-AGCACAGGTTGGACAATACCTGAAA-3′ and 5′-GGCCATCAACGGCATGGAAACGATAA-3′.

For RT-qPCR, RNA was extracted using RNeasy kit (QIAGEN). 350 ng RNA was reverse transcribed with Superscript Reverse Transcription (Invitrogen). RT-qPCR was performed with the SYBR Green PCR Master Mix (Sigma-Aldrich, Roche, or QIAGEN) on an LC480 light cycler (Roche). Oligonucleotides for Igκ germline transcript and rearranged Igκ and Igλ were used as described previously (Fuxa et al., 2004; Mandal et al., 2011).

Other primers used were as follows: Hprt, 5′-GTTGGATACAGGCCAGACTTTGTTG-3′ and 5′-GATTCAACTTGCGCTCATCTTAGGC-3′; Rag1, QT00243621 (QIAGEN); Rag2, QT00253414 (QIAGEN); Ikzf1, QT01060689 (QIAGEN); and Myc, Mm.PT.56.28494642 (Integrated DNA Technologies).

3 × 10⁶ cells were incubated with 5 µg/ml Indo-1 (Invitrogen) and 0.5 µg/ml Pluronic acid (Invitrogen) in Iscove’s medium containing 1% FCS for 45 min at 37°C. The cells were resuspended in Iscove’s medium containing 1% FCS, and the Ca²⁺ response was induced by addition of 20 µg/ml anti-µ (Jackson ImmunoResearch Laboratories, Inc.). As a positive control, cells were stimulated with 1 µM ionomycin.

10⁷ cells were treated for 6 or 24 h with or without 4OHT in the presence or absence of IL-7, as described above. RNA was extracted with the RNeasy kit. Biotinylated single-strand cDNA targets were prepared, starting from 150 ng of total RNA, using the Ambion WT Expression kit and the Affymetrix GeneChip WT Terminal Labeling kit. After fragmentation and end-labeling, 1.9 µg cDNA was hybridized for 16 h at 45°C on GeneChip Mouse Gene 1.0 ST arrays (Affymetrix). Data were normalized with the Robust Multi-array Average (RMA) software using default settings and analyzed with Excel (Microsoft) and Cluster 3. Probe sets that did not correspond to an identified gene were not considered for analysis. Clusters were visualized with Treeview. Help with analyses were provided by C. Thibault, D. Dembélé, and A. Oravecz (IGBMC). The data are available in the Gene Expression Omnibus database under accession no. GSE51350.

Cytoplasmic and nuclear extracts from 10⁷ cells were prepared using cytoplasmic lysis buffer (10 mM Hepes, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, and protease inhibitor cocktail) and nuclear lysis buffer (20 mM Hepes, pH 7.9, 1.5 mM MgCl₂, 420 mM NaCl, 0.5 mM EDTA, 0.5 mM DTT, 25% glycerol, and protease inhibitor cocktail), respectively. Lysates were separated by SDS-PAGE on a 10% gel, and Western blotting was performed using an enhanced chemiluminescence system. The following antibodies were used for detection: anti-Ikaros, anti-ER, anti-TBP (all in-house), and anti–β-actin (BD).

Fig. S1 shows an analysis of Ikaros expression in cKO mice.

We thank M.C. Birling, M. Reth, J. Ghysdael, A.S. Korganow, W. Pear, G. Nolan, A.S. Geimer Le Lay, A. Apostolov, A. Oravecz, C. Thibault, D. Dembélé, P. Marchal, M. Gendron, S. Falcone, and C. Ebel for reagents, help, and discussion.

We received grants from Agence Nationale de la Recherche (ANR-11-BSV3-018), La Ligue Contre le Cancer (Labellisée 2006-2012), and Institut National du Cancer (INCa #2011-144) and funding from Institut National de la Santé et de la Recherche Médicale, Centre National de la Recherche Scientifique, and Université de Strasbourg. B. Heizmann received postdoctoral funding from Fondation ARC.

The authors have no conflicting financial interests.