Pten mediates Myc oncogene dependence in a conditional zebrafish model of T cell acute lymphoblastic leukemia

We previously generated a zebrafish model of T-ALL, using rag2 promoter-driven expression of a mouse EGFP-Myc fusion transgene to induce T-ALL which closely resembles the human disease pathologically and by gene expression (Langenau et al., 2003, 2005a). To generate a zebrafish model of T-ALL in which MYC activity could be modulated in established tumor cells, we generated stable transgenic zebrafish expressing rag2:MYC-ER, in which the zebrafish rag2 promoter drives expression of a fusion transgene consisting of human MYC fused to the ligand-binding domain of a modified estrogen receptor that is posttranslationally induced by 4-hydroxytamoxifen (4HT) treatment but not by endogenous estrogens (Littlewood et al., 1995). When treated with vehicle control (ethanol), rag2:MYC-ER heterozygous fish did not develop T-ALL over a 16-wk follow-up period. Thymic fluorescence in 8-wk-old rag2:MYC-ER transgenic fish coexpressing lck:EGFP and rag2:dsRed2 transgenes that were treated with vehicle control was indistinguishable from that of their rag2:MYC-ER negative siblings (Fig. 1, A–D). However, when MYC-ER was induced by treatment with 50 µg/liter (129 nM) 4HT beginning at 5 d postfertilization, rag2:MYC-ER transgenic zebrafish developed fully penetrant T-ALL starting at 5 wk of age (Fig. 1, E and F). Histological analysis of T-ALL tumors induced by rag2:MYC-ER alone revealed that these tumors consisted of sheets of monomorphic lymphoblasts characterized by regular nuclear contours, scant cytoplasm, homogeneous chromatin, and inconspicuous nucleoli (Fig. S1), morphological features which closely resemble those of human T-ALL lymphoblasts. Note that no phenotypic differences were observed between these tumor cells and those arising in zebrafish of other genotypes studied throughout this manuscript (Fig. S1).

To test whether these tumors were dependent on ongoing MYC transgene activation, zebrafish with T-ALL were removed from 4HT to down-regulate MYC activity, and tumors were monitored weekly for 8 wk by fluorescence imaging. Tumor phenotypes were then classified into four categories based on the degree of change in tumor size by the end of the 8-wk monitoring period: (1) complete regression (disappearance of all abnormal extrathymic fluorescence), (2) partial regression (reduction in tumor diameter to less than half its original size), (3) stable disease (tumors that did not meet criteria for regression or progression), and (4) progression (increase in tumor diameter to more than twice its original size). After T-ALL development, removal from 4HT induced complete tumor regression in 75% of zebrafish examined, whereas tumor regression was not seen in any of the zebrafish kept in 4HT (Fig. 1 I; P = 0.0001).

In most experimental models of MYC-induced tumorigenesis, inactivation of MYC leads to loss of the malignant phenotype as a result of apoptotic cell death or differentiation/senescence (Arvanitis and Felsher, 2006). To examine the fate of T-ALL cells in rag2:MYC-ER zebrafish after 4HT withdrawal, we monitored EGFP expression driven by the lck promoter, which is expressed in both immature and mature T cells, and dsRed2 expression driven by the rag2 promoter, which is expressed in immature but not mature T cells (Langenau et al., 2004, 2005a). The loss of tumor cells after removal from 4HT did not appear to predominantly represent T cell differentiation because in each fish with tumor regression we noted the simultaneous disappearance of disseminated cells expressing both EGFP and dsRed2 fluorescence, with no visible residual extrathymic EGFP-positive dsRed2-negative mature T cells (Fig. 1, E–H). Furthermore, examination of the senescence marker acidic β-galactosidase activity revealed no increase above background levels in T-ALL cells after MYC inactivation (Fig. S2). In some settings, collapse of the tumor-associated vasculature is involved in tumor regression after MYC inactivation (Giuriato et al., 2006; Sodir et al., 2011); thus, we also examined whether MYC inactivation affected the tumor vasculature, taking advantage of fli1:EGFP transgenic zebrafish in which EGFP is robustly expressed in vascular endothelium (Lawson and Weinstein, 2002). Immunohistochemistry for EGFP in rag2:MYC-ER;rag2:dsRed2;fli1:EGFP triple-transgenic zebrafish revealed no significant differences in tumor-associated microvascular density 4 d after MYC inactivation (Fig. S3).

Immunohistochemical analysis of T-ALL cells from rag2:MYC-ER transgenic zebrafish revealed increased staining for the apoptotic marker activated caspase 3 after removal from 4HT (Fig. 2 A), implicating the induction of apoptotic cell death in T-ALL cells after MYC down-regulation. We thus reasoned that mitochondrial (intrinsic) apoptosis might be the key mediator of T-ALL regression after MYC down-regulation, a hypothesis which we tested using a rag2:EGFP-bcl2 transgene, which blocks intrinsic apoptosis in thymocytes by binding and sequestering proapoptotic BCL2 family proteins (Langenau et al., 2005b). Indeed, we found that T-ALL tumors induced in rag2:MYC-ER zebrafish that also expressed rag2:EGFP-bcl2 failed to regress after MYC down-regulation (Fig. 2, B and C; P < 0.0001). Collectively, these findings implicate mitochondrial apoptosis as a central mediator of T-ALL regression after MYC down-regulation in this model.

We have recently identified genetic alterations of PTEN and the PI3K-AKT pathway in 48% of primary T-ALL human tumor specimens, most of which (86%) were also characterized by MYC overexpression (Gutierrez et al., 2009). We thus wanted to use our model to test the hypothesis that such alterations would render T-ALLs independent of MYC oncogenic activity for tumor maintenance. Because of a partial genome duplication event during teleost evolution, zebrafish have duplicated pten genes, known as ptenA and ptenB, both of which are highly conserved with mammalian PTEN genes, demonstrate phosphatase activity toward phosphatidylinositol(3,4,5)triphosphate, and have redundant roles during zebrafish development (Faucherre et al., 2008). Complete deficiency of all zebrafish pten alleles leads to embryonic lethality, whereas compound haploinsufficiency is compatible with normal development. To test whether pten is a modulator of MYC oncogene dependence, double-transgenic rag2:MYC-ER;rag2:GFP zebrafish were mated to fish harboring loss-of-function mutations in their ptenA and/or ptenB genes (Fig. 3 A). Offspring from these matings were raised in 50 µg/liter (129 nM) 4HT beginning at 5 d postfertilization, the time at which thymic fluorescence is first visible in rag2:GFP transgenic fish, and T-ALL onset was monitored by weekly GFP fluorescence imaging beginning at 4 wk of age. After development of disseminated T-ALL, fish were removed from 4HT to down-regulate MYC activity, and tumor phenotypes after 4HT removal were assessed based on the degree of change in tumor size by the end of the 8-wk monitoring period, as described in the previous section.

Loss of up to two of the four zebrafish pten alleles did not measurably accelerate the onset of MYC-induced T-ALL (Fig. 3 B); however, pten haploinsufficiency significantly promoted loss of tumor dependence on MYC transgene activation (Fig. 4). The rates of complete or partial regression were 69% in pten wild-type zebrafish, 58% in fish harboring heterozygous ptenB mutations, 18% in fish harboring heterozygous ptenA mutations, and 10% in pten double-heterozygous zebrafish (Fig. 4 E). Conversely, deficiency of pten also significantly promoted tumor progression after MYC-ER inhibition, with progression rates of 13% in pten wild-type zebrafish, 33 and 36% in fish harboring a single heterozygous mutation of ptenB or ptenA, respectively, and 60% in zebrafish harboring heterozygous mutations in both ptenA and ptenB (Fig. 4 E; P = 0.001 by 4 × 4 Fisher’s exact test). Thus, pten haploinsufficiency strongly promoted the loss of MYC transgene dependence. Furthermore, Western blot analysis of the minority of T-ALL tumors from rag2:MYC-ER pten–wild-type zebrafish that progressed despite MYC down-regulation revealed that a subset of these tumors spontaneously acquired activation of the AKT pathway, as indicated by robust phosphorylation of S6 ribosomal protein (Fig. 4 F).

Although PTEN is best known as a negative regulator of oncogenic signaling through the PI3K–AKT pathway, evidence also suggests that PTEN has AKT-independent functions, including regulation of p53 activity, chromosomal stability, and activity of the anaphase-promoting complex (Freeman et al., 2003; Shen et al., 2007; Song et al., 2011). To test whether pten deficiency modulates MYC dependence via Akt pathway activation, we tested whether expression of rag2:myr-mAkt2, encoding a myristoylated constitutively active mouse Akt2 transgene (Tan et al., 2008), would phenocopy the effect of pten inactivation on MYC transgene dependence. Heterozygous rag2:MYC-ER zebrafish that did not harbor a thymic fluorescent transgene were out-crossed to wild-type animals, and all resultant embryos were injected at the one-cell stage with DNA expression constructs encoding rag2:GFP alone or both rag2:GFP and rag2:myr-mAkt2. Injected fish were raised in 4HT to induce T-ALL, after which 4HT was removed and tumor regression was monitored as in the previous section. Expression of rag2:myr-mAkt2 significantly accelerated the onset of T-ALL in rag2:MYC-ER transgenic zebrafish, with a mean age at T-ALL onset of 4.3 wk in rag2:myr-mAkt2–injected fish, compared with 12.1 wk in control MYC-ER fish injected with rag2:GFP alone (Fig. 5 A; P = 0.017). Expression of rag2:myr-mAkt2 in MYC-ER–negative zebrafish also induced T-ALL development in 17% of these fish by 20 wk (Fig. 5 A), indicating that Akt activation is sufficient to induce T-ALL in the zebrafish, albeit at a relatively low penetrance.

Analysis of T-ALL regression after 4HT removal demonstrated that expression of the constitutively active myr-mAkt2 transgene was strongly associated with loss of MYC transgene dependence (Fig. 5, B and C), suggesting that the effect of pten inactivation is mediated by Akt. T-ALLs arising in rag2:MYC-ER transgenic zebrafish expressing both rag2:myr-mAkt2 and rag2:GFP showing a complete or partial regression rate of 14% after MYC-ER inhibition, compared with a regression rate of 75% in rag2:MYC-ER controls injected with rag2:GFP alone. Concomitantly, tumor progression rates were 86% in rag2:myr-mAkt2–injected zebrafish, compared with 17% of rag2:GFP-injected controls (Fig. 5 D; P = 0.0002 by 2 × 4 Fisher’s exact test).

Activated Akt directly phosphorylates and inhibits GSK3 (Cross et al., 1995), which is a key regulator of Myc protein stability (Sears et al., 2000), thus we wanted to test whether Akt pathway activation might substitute for MYC transgene activation by inducing the transcriptional activity of endogenous Myc. The Ddx18 RNA helicase is a well established direct target of Myc in human and mouse cells (Grandori et al., 1996; O’Hagan et al., 2000), and we used injection of Myc mRNA into zebrafish embryos followed by ddx18 RNA in situ hybridization to confirm that the regulation of ddx18 by Myc is conserved in the zebrafish (Fig. 6 A). We then performed quantitative (Q) RT-PCR analysis of ddx18 mRNA levels in sorted T-ALL cells from rag2:MYC-ER;rag2:EGFP-bcl2 zebrafish that also expressed either a rag2:GFP or a rag2:myr-mAkt2 transgene. T-ALL cells were sorted from zebrafish in 4HT and 4 d after 4HT removal. Expression of bcl2 was included in all conditions to avoid the confounding effects of apoptosis induction after MYC down-regulation in the absence of myr-mAkt2 expression. Analysis of the transcriptional activity of Myc, as assessed by ddx18 Q-RT-PCR, revealed that it is regulated by 4HT treatment in rag2:MYC-ER–positive T-ALL cells, whereas the overexpression of constitutively active Akt had no effect on Myc transcriptional activity in the presence or absence of 4HT (Fig. 6 B). These findings were confirmed by Q-RT-PCR analysis for npm1a, a zebrafish orthologue of NPM1, another direct target of MYC (Zeller et al., 2001; Fig. S4). Collectively, these data indicate that the effect of Akt activation on Myc oncogene dependence is not mediated by increasing the transcriptional activity of endogenous Myc.

The striking effect of pten haploinsufficiency on T-ALL regression after MYC down-regulation contrasted sharply with the absence of a detectable phenotype of pten haploinsufficiency on onset of MYC-induced T-ALL. We wondered whether T-ALLs arising in a pten-haploinsufficient background might have an increased propensity to loss of wild-type pten alleles, but sequencing of ptenA and ptenB cDNAs generated from T-ALL cells sorted at the time of tumor onset in pten wild-type and pten double-heterozygous fish (n = 6 tumors from each genotype) revealed no mutational inactivation of wild-type alleles in any of the tumors analyzed.

We reasoned that MYC-induced repression of pten might explain the lack of a phenotype induced by pten haploinsufficiency in the setting of high MYC activity, whereas a striking phenotype might be expected after MYC inactivation if this led to the relief of pten repression. We thus measured ptenA and ptenB mRNA levels by Q-RT-PCR in sorted T-ALL cells from rag2:MYC-ER;rag2:EGFP-bcl2 transgenic zebrafish before and after 4HT removal. Q-RT-PCR analysis revealed that MYC down-regulation leads to a significant increase in transcript levels for both ptenA and ptenB, indicating that MYC represses pten, which is consistent with recent findings in mammalian cells (Mu et al., 2009; Olive et al., 2009; Kim et al., 2010; Mavrakis et al., 2010).

We have demonstrated for the first time that pten is an important mediator of MYC oncogene dependence in T-ALL lymphoblasts, an effect which appears to be mediated by its role as an inhibitor of oncogenic signaling through the Akt pathway. Furthermore, we also find that Myc represses Pten mRNA levels, an effect which is likely mediated by Myc-induced micro-RNAs which have recently been shown to down-regulate Pten (Mu et al., 2009; Olive et al., 2009; Kim et al., 2010; Mavrakis et al., 2010). Our data thus support a model in which Myc dependence is mediated in part by the up-regulation of Pten that follows Myc inactivation, which then acts to inhibit AKT pathway signaling which would otherwise be sufficient to maintain the malignant phenotype, even in the absence of oncogenic signals mediated by Myc (Fig. 7 B).

Our findings implicate apoptotic cell death through the mitochondrial pathway as a critical mediator of Myc oncogene dependence in this model, thus indicating that inhibition of apoptosis downstream of AKT is the likely mechanism through which Pten haploinsufficiency induces loss of Myc oncogene dependence. Akt has previously been shown to inhibit mitochondrial apoptosis via multiple pathways (Manning and Cantley, 2007), including inactivation of BAD via phosphorylation (Datta et al., 1997; del Peso et al., 1997), transcriptional down-regulation of BIM downstream of its effect on FOXO proteins (Dijkers et al., 2002), inhibition of p53 via phosphorylation and activation of MDM2 (Mayo and Donner, 2001; Zhou et al., 2001), and up-regulation of the antiapoptotic MCL1 downstream of both mTOR (Wei et al., 2006; Hsieh et al., 2010) and GSK3 (Maurer et al., 2006). Our results indicate that AKT activation is one mechanism that can overcome MYC oncogene dependence in T-ALL, suggesting that PI3K-AKT inhibitors may have therapeutic utility in combination with inhibitors of MYC itself in the design of new therapeutic strategies for T-ALL.

The rag2:MYC-ER construct was generated by placing a human MYC transgene fused to the ligand-binding domain of the estrogen receptor, which is posttranslationally induced by 4HT but not endogenous estrogens (Littlewood et al., 1995), downstream of the zebrafish rag2 promoter in a modified pBluescript vector in which the rag2:MYC-ER construct is flanked by recognition sequences for I-SceI meganuclease. This circular rag2:MYC-ER was co-injected together with a wild-type mitfa construct under its own promoter, also flanked by I-SceI recognition sequences, into one-cell stage embryos from the nacre line using a previously described I-SceI co-injection strategy (Grabher and Wittbrodt, 2008). Nacre zebrafish harbor a pigmentation defect caused by homozygous germline mitfa loss-of-function mutations, thus allowing rescue of the pigmentation defect by the co-injected mitfa transgene to act as a trans-genesis marker. A line derived from one rag2:MYC-ER;mitf founder zebrafish was identified that developed 4HT-dependent T-ALL and was used for all further studies. Co-segregation of the rag2:MYC-ER transgene together with the pigment-rescue phenotype was confirmed by PCR in all subsequent generations using the PCR primers described in the next section.

The lck:EGFP, rag2:GFP, rag2:EGFP-bcl2, and fli1:EGFP transgenic zebrafish lines have previously been described (Lawson and Weinstein, 2002; Langenau et al., 2003, 2004, 2005b). The transgenic line referred to here as rag2:dsRed2 harbors a rag2:LoxP-dsRed2-STOP-LoxP-EGFP-mMyc transgene, which in the absence of Cre demonstrates dsRed2 expression driven by the rag2 promoter (Langenau et al., 2005a). It is worth noting that zebrafish expressing this construct demonstrate no detectable EGFP-Myc translation in the absence of Cre, as indicated by the absence of detectable EGFP fluorescence or spontaneous T-ALL development and by the lack of modification of T-ALL onset or regression in the rag2:MYC-ER line. Zebrafish harboring loss-of-function mutations in ptenA and ptenB genes were previously described (Faucherre et al., 2008).

The rag2:myr-mAkt2 construct was generated by placing a myristoylated mouse Akt2 transgene (Tan et al., 2008), which is constitutively activated as a result of constitutive membrane localization, downstream of the rag2 promoter in the I-SceI–modified pBluescript vector described in the first paragraph in this section. Microinjections of the rag2:myr-mAkt2 expression construct together with rag2:GFP (15 pg per embryo of each), or of rag2:GFP alone as a control (30 pg per embryo), were performed as previously described (Grabher and Wittbrodt, 2008). Zebrafish demonstrating thymic GFP fluorescence were examined for T-ALL development, and successful rag2:myr-mAkt2 transgene integration was verified by PCR at the end of each experiment.

Treatment with 4HT was begun starting at 5 d postfertilization, with each clutch placed in 750 ml of water containing 50 µg/liter (129 nM) 4HT (Sigma-Aldrich) in ethanol. Treatment with 4HT was performed offline, with weekly water changes beginning at 4 wk of age. Zebrafish raised in 4HT were screened for tumors by fluorescence microscopy weekly beginning at 4 wk of age. At the time of T-ALL onset and weekly during tumor regression monitoring, fish were imaged in both brightfield and GFP or dsRed2 fluorescent channels, as appropriate. Fluorescence microscopy was performed with a microscope (SMZ1500; Nikon) and an X-Cite 120 Fluorescence Illumination System (EXFO). Images were captured using a camera (DS2MBWc; Nikon) and NIS-Elements F Package (version 3.00; Nikon). All images shown represent merged fluorescence (shown in green or red, depending on the fluorophore) and brightfield (shown in grayscale) images. Fluorescence and brightfield images were merged using Photoshop (7.0; Adobe). Onset of T-ALL was defined as the development of a fluorescent mass arising from the thymus that was more than twice the size of normal thymus, together with invasion of adjacent tissues and structures. Note that we use T-ALL herein to refer to localized or disseminated T cell lymphoblastic tumors.

To monitor tumor regression, zebrafish with T-ALL were removed from 4HT and placed in individual tanks online, and tumors were imaged weekly for a total of 8 wk. Tumor phenotypes after 4HT removal were categorized into four classes based on the change in the diameter of the largest contiguous tumor mass by the end of the 8-wk monitoring period, as described in the Results section. Fish that became moribund with leukemia <8 wk after tamoxifen were euthanized and classified into the progression category. Genotyping for rag2:MYC-ER, as well as ptenA and ptenB or rag2:myr-mAkt2, was performed by genomic PCR after the completion of each experiment. PCR genotyping primers were as follows: rag2:MYC-ER forward, 5′-AGTCCTGAGACAGATCAGCA-3′, and reverse, 5′-TCATCATGCGGAACCGACTT-3′; and rag2:myr-mAkt2 forward, 5′-ATGCTAATTTGAAGCACTAGCA-3′, and reverse, 5′-TTGGGTCCTTCTTCAGCAGT-3′. Genotyping for ptenA and ptenB was performed as previously described (Faucherre et al., 2008).

Zebrafish were euthanized in tricaine anesthetic, fixed in 4% paraformaldehyde at 4°C for 2 d, decalcified with 0.25 M EDTA, pH 8.0, for 2 d, dehydrated in alcohol, cleared in xylene, and embedded in paraffin. Slides were deparaffinized and pretreated with 10 mM citrate, pH 6.0, in a steam pressure cooker (Decloaking Chamber; BioCare Medical) according to the manufacturer’s instructions, followed by washing in distilled water. Slides were pretreated with Peroxidase Block (Dako) for 5 min, followed by serum-free protein block (Dako) for 20 min to block nonspecific binding sites. Slides were then incubated with primary rabbit anticleaved caspase-3 antibody (1:50; Cell Signaling Technology) or mouse anti-GFP antibody (1:1500; Takara Bio Inc.). Secondary detection was performed by adding horseradish peroxidase-labeled polymer (Dako) for 30 min. Immunoperoxidase staining was developed with diaminobenzidine + chromogen kit (Dako), according to the manufacturer’s instructions.

Zebrafish were euthanized in tricaine anesthetic, embedded in a 1:1 mixture of 30% sucrose and Tissue-Tek O.C.T. compound (Sakura), and snap-frozen in isobutane precooled in dry ice. 14-µm frozen sections were cut onto slides using a CM3050 cryostat (Leica). Staining for activity of senescence-associated β-galactosidase was performed using the Senescence β-Galactosidase Staining kit (Cell Signaling Technology) according to the manufacturer’s instructions.

DsRed2-sorted T-ALL cells were lysed in RIPA buffer (Millipore) supplemented with Complete Mini protease inhibitor (1 tablet/10 ml; Roche) and PhosSTOP phosphatase inhibitor (1 tablet/10 ml; Roche). 20 µg protein lysate was mixed with Laemmli 4× SDS sample buffer (Boston Bioproducts) and β-mercaptoethanol (Sigma-Aldrich) in appropriate proportions and incubated at 100°C for 3 min before being run on a 4–12% Novex bis-tris polyacrylamide gel (Invitrogen) at 160V for 1 h. Blots were then transferred to a nitrocellulose membrane (Millipore) at 350 mA for 1 h, blocked with 5% milk in tris-buffered saline with 0.1% tween, and probed with anti–phospho-S6 ribosomal protein (Ser240/244) antibody (1:1,000; Cell Signaling Technology), anti-S6 ribosomal protein antibody (1:1,000; Cell Signaling Technology), and anti–α-tubulin antibody (1:1000; Cell Signaling Technology). Secondary detection was with HRP-linked antibodies (Cell Signaling Technology), HRP substrate was SuperSignal West Dura (Thermo Fisher Scientific), and Western blot images were captured using an ImageQuant LAS4000 chemiluminescent image analyzer (GE Healthcare).

Total RNA was extracted from fluorophore-sorted T-ALL cells using Trizol reagent according to the manufacturer’s instructions (Invitrogen). cDNA was synthesized from total RNA using the SuperScript III First-Strand Synthesis System for RT-PCR according to the manufacturer’s instructions (Invitrogen). 3% of the cDNA reaction volume was then taken for each Q-RT-PCR reaction, performed with the SYBR green PCR Core Reagents kit (Applied Biosystems) and a 7300 Real Time PCR System instrument (Applied Biosystems) according to the manufacturer’s instructions. All Q-RT-PCR were performed in triplicate, and β-actin1 was used as the control for all Q-RT-PCR reactions. Q-RT-PCR primer sequences were as follows: ddx18 forward, 5′-CGGCTGAGGAGGATAGTGATG-3′, and reverse, 5′-ACGCACCTGTCAGACCAGAAG-3′; npm1a forward, 5′-ACAAAAGGGCAAGAAGGAACAG-3′, and reverse, 5′-CGGCTAAAGTCCGTGGTGTTT-3′; ptenA forward, 5′-ATCCCCAGGGCTTCAAGAA-3′, and reverse, 5′-TGTGGATCTTTACCTTCAACTGATACA-3′; ptenB forward, 5′-GGGTGTCATGATTTGTGCGTACT-3′, and reverse, 5′-TCGCCTCTGACTGGGAATAGTC-3′; β-actin1 forward, 5′-TACAATGAGCTCCGTGTTGC-3′, and reverse, 5′-ACATACATGGCAGGGGTGTT-3′.

Mouse Myc was cloned into the pCS2⁺ vector and linearized by Not1 digestion. Myc mRNA was then transcribed in vitro using the Ambion SP6 mMessage mMachine kit (Applied Biosystems). Zebrafish embryos were injected at the one-cell stage with 100 pg of Myc or control mCherry mRNA. At the 24-h postfertilization developmental stage, injected embryos were fixed overnight in 4% paraformaldehyde, and ddx18 RNA in situ hybridization was performed as previously described (Payne et al., 2011).

Differences in tumor-free survival were assessed by the log-rank test, and time-to-event distributions were estimated via the Kaplan-Meier method. Differences in categorical data were assessed via Fisher’s exact test, and differences in continuous data were assessed via the Mann-Whitney (rank-sum) test.

Fig. S1 shows histological morphology of T-ALL lymphoblasts of various genotypes analyzed in this study. Fig. S2 shows analysis of the senescence marker acidic β-galactosidase activity in sectioned rag2:MYC-ER; rag2:dsRed2 double-transgenic zebrafish with T-ALL before and after 4HT withdrawal. Fig. S3 shows analysis of microvascular density in T-ALL tumors before and after 4HT withdrawal. Fig. S4 shows the effect of 4HT treatment versus withdrawal, and of expression of activated Akt2 versus control, on mRNA expression of the Myc target npm1a.

We thank Greg Molind and Lu Zhang for zebrafish husbandry.

This work was supported by National Institutes of Health grants NCI 5P01CA68484 and NCI 1K08CA133103 and by the William Lawrence Foundation. A. Gutierrez is a Scholar of the American Society of Hematology-Harold Amos Medical Faculty Development Program. H. Feng is supported by an award from the National Cancer Institute (NCI)/National Institutes of Health (K99CA134743). This work was also supported in part by grant P30 CA-06927 from the National Cancer Institute and by an appropriation from the Commonwealth of Pennsylvania.

The authors have no competing financial interests.