RNAi-mediated silencing of TEL/AML1 reveals a heat-shock protein– and survivin-dependent mechanism for survival

The TEL/AML1 fusion gene is generated by t(12;21)(p13;q22), the most frequent chromosomal translocation in childhood acute lymphoblastic leukemia (ALL).¹  There is convincing evidence that the gene fusion already takes place in utero in the majority of children with TEL/AML1-positive ALL, and this suggests, together with studies in twins, that it represents an early, or even initiating, event in leukemia development.^2,3  This translocation, however, is not sufficient for the manifestation of a clinical disease, but seems rather to induce a protracted preleukemic state in which additional postnatal genetic events may accumulate, eventually leading to malignant transformation.^4,5 

The role of TEL/AML1 expression in the affected cell was recently addressed in several studies and indicates an interference with differentiation in the B lineage as well as enhanced self-renewal of B cells.^6–8  Of importance, such mutations that impair hematopoietic differentiation are assumed to interfere also with apoptosis.⁹  The mechanisms by which the fusion protein breaches this crucial anticancer shield have, so far, not been identified.

The function of genes, in particular of fusion genes, has been studied in recent years by gene silencing using RNA interference (RNAi), a powerful strategy to specifically target the gene of interest.¹⁰  In this study, we investigated the functional contribution of TEL/AML1 to the malignant phenotype focusing on the evasion of programmed cell death by silencing the fusion gene in the t(12;21)-positive cell line REH.

The B-cell precursor (BCP) leukemic cell line REH (having the TEL/AML1 fusion gene, the second TEL allele deleted and AML1 retained) and the mouse IL3-dependent BCP Ba/F3 cells were grown in 24-well plates (Nunc, Roskilde, Denmark) at 1 × 10⁶ cells/mL.

siRNAs targeting the fusion region of TEL/AML1 were designed¹¹  and synthesized by Dharmacon (Lafayette, CO). siRNA duplexes were handled essentially as described.¹²  The transfection of REH cells with siRNA duplexes at a concentration of 230 nM was performed using Lipofectamin 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions.

Apoptotic cell death was evaluated using annexin V–FITC/PI staining (BD Biosciences Pharmingen, San Jose, CA) and fluorescence-activated cell sorting (FACS) analysis according to the manufacturer's recommendations.

RNA from siRNA-treated and untreated REH cells was isolated using the RNeasy Mini Kit RNA isolation kit (QIAGEN, Valencia, CA). Further processing of RNA, hybridization to Affymetrix Human Genome U133A 2.0 chip gene arrays (Affymetrix, Santa Clara, CA), and data processing were performed by VBC Genomics (Vienna, Austria).

Preparation of cell lysates, Western blot analysis, metabolic labeling, staining of cytoplasmic proteins, and separation by 2D gel electrophoresis were performed as described previously.¹³ 

The TEL/AML1 cDNA (kindly provided by O. Williams, London, United Kingdom⁸ ) was cloned into pCDNA 3.1/V5-His (Invitrogen) and transfected into Ba/F3 cells. The empty vector and a GFP-carrying pCDNA 3.1/V5-His vector were used as controls. Transfected cells were subcloned and stably expressing clones established.

Significant differences between values obtained in cell lines cultured under different experimental conditions were determined using the Student t test.

Several siRNAs targeting the fusion site of the TEL/AML1 mRNA were tested (Figures 1A and S1S1). Based on their influence on the fusion gene expression, C1 was chosen as the functional and S as the control siRNA and used in all further experiments. To obtain efficient silencing of TEL/AML1 in REH cells, siRNAs were transfected twice according to the time course depicted in Figure 1B, and protein depletion was confirmed at day 6 of the experiment (Figure 1C). At this time, cell viability was largely unaffected. However, 2 days later, cultures that were devoid of the fusion protein had a reduction of the cell count by about 60%, while those treated with the control siRNA S were reduced by only 20% compared with untreated REH cells. Moreover, assessment of early and late apoptosis indicated that the majority of TEL/AML1-silenced cells had started the apoptotic program, whereas controls contained only 10% to 15% mostly early apoptotic cells (Figure 1D). These data suggest that TEL/AML1 interferes with the delicate balance between life and death probably by engaging a potent antiapoptotic mechanism.

Having demonstrated the impact of TEL/AML1 protein depletion on cell survival, we set out to unravel the corresponding molecular mechanism. We used gene expression analysis in TEL/AML1-silenced cells and searched for affected genes that are involved in the control of cell survival and apoptosis (Table S1Table S1). Of interest, HSP90 and HSP70 were consistently found to be down-regulated after TEL/AML1 depletion in 2 independent experiments. The nonfunctional siRNA S was used as a control to prove that the results were obtained from specific silencing of the fusion gene. For this purpose, the expression of interferon-responsive genes¹⁰  was evaluated and found not to differ between these 2 populations (data not shown). To further confirm our findings on the protein level, we used high-resolution 2D gel electrophoresis of cytoplasmic proteins that were either metabolically labeled or stained, allowing quantification of synthesis rate and the amount of individual proteins, respectively. In TEL/AML1-silenced cells (day 6 of the experiment, Figure 1B), the synthesis rate and quantity of HSP90 had decreased, whereas HSP70 showed only a reduced synthesis rate (Figure 2A-B). Strikingly, both proteins were reduced 2 days later (Figure 2C). Further scrutiny of our gene expression data revealed a differential regulation of survivin upon silencing, a finding that was substantiated by reduced amounts of survivin protein 2 days later (Figure 2C) and correlated well with the high rate of apoptosis in these cells. To further corroborate the relationship between the fusion protein and HSP90 expression, TEL/AML1 was ectopically expressed in Ba/F3 cells, which was followed by up-regulation of HSP90 and survivin (Figure 2D). The suggested antiapoptotic effect of TEL/AML1 was supported in this model system since Ba/F3 cells expressing the fusion gene were resistant against growth factor withdrawal as well as VP16 treatment (Figure 2E). Further studies need to clarify the mode by which TEL/AML1 influences HSP expression and in particular its possible interference with heat-shock factors, dominant regulators of HSP.¹⁴ 

In this study, we present in 2 model systems a clear functional relationship between TEL/AML1 and the apoptotic network with a major impact on the members of the heat-shock protein family HSP90 and HSP70 as well as on survivin, a potent member of the inhibitor of the apoptosis IAP family.^15–17  The occurrence of TEL/AML1 and its interference with the apoptosis control enables the affected cell to survive, thereby providing the opportunity to acquire further mutations and eventually turn into a fully malignant phenotype. The propensity of leukemic REH cells to undergo apoptosis after TEL/AML1 depletion provides further evidence that this survival advantage is also indispensable for maintaining the leukemic phenotype, a feature that is shared by other leukemia-associated fusion genes.^18–20  And finally, our data (1) provide the rational for the potential clinical use of inhibitors against HSP90 or survivin that are frequently up-regulated during cancer development^15,17  and currently undergoing phase 1 clinical trials also for pediatric patients with recurrent or refractory ALL,^21,22  and (2) support a future siRNA-based therapy²³  in this subtype of leukemia. Such targeted treatment approaches would ideally affect all the fusion gene–carrying cells, be it the chemotherapy-sensitive bulk of leukemic cells, or, more importantly, also the leukemic stem cell as well as the preleukemic clone, the definite and proposed sources of relapses, respectively.^24–26 

Contribution: All authors have substantially contributed to the content of the paper and have agreed to the submission in its current version; C.D. designed and performed research, analyzed data, and wrote the first draft of the paper; C.G. performed and analyzed the 2D PAGE data; G.K., A.I., C.L., J.B., and A.M.D. performed and analyzed experiments and contributed to the experimental design; E.R.P.-G. is responsible for the design of the study, the integrity of the data, and writing the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: E. Renate Panzer-Grümayer, CCRI and St Anna Kinderspital, Kinderspitalgasse 6, A-1090 Vienna, Austria; e-mail: renate.panzer@ccri.at.

This work was supported in part by a grant from the Jubiläumsfonds der ÖNB 10720, the FWF P17551-B14, and the GENAU-CHILD GZ 200.136/1-VI/1/2005 to E.R.P.-G.

We would like to thank Idriss M. Bennani-Baiti for reading the paper and stimulating discussions.