Bcl-X_L down-regulation suppresses the tumorigenic potential of NPM/ALK in vitro and in vivo

The broad category of non-Hodgkin lymphomas (NHLs) includes distinct neoplasms arising from clonal expansion of B or T lymphocytes at different stages of maturation and differentiation.¹  The CD30/Ki-1⁺anaplastic large cell lymphomas (ALCLs) account for about 20% of high-grade NHLs presenting unique immunophenotypical, morphologic, and genetic features.²  Approximately 70% of ALCLs express an 80-kDa hybrid oncoprotein, named NPM/ALK, arising from the chromosomal translocation t(2;5)(p23;q35).³  The NPM/ALK chimera contains the N-terminal portion of the ubiquitously expressed nucleolar protein nucleophosmin (NPM) fused to the kinase domain of anaplastic lymphoma kinase (ALK).⁴  ALK is a member of the insulin receptor tyrosine kinase superfamily, whose expression is restricted to the central nervous system mainly during the neonatal period.⁵ 

As a consequence of the t(2;5) rearrangement, the strong NPM gene promoter drives the high-level ectopic expression of ALK in lymphoid ALCL cells.^6,7  In addition, the oligomerization domain contained in the NPM portion promotes the constitutive trans-phosphorylation and activation of the ALK kinase domain in NPM/ALK.⁸  Thus, NPM/ALK can exert its oncogenic potential via abnormal stimulation of multiple cellular signaling cascades, enhancing growth factor-independent proliferation and prolonging viability. Several groups described a direct recruitment and activation by NPM/ALK of numerous intracellular signal transduction effectors such as phospholipase C-γ (PLC-γ),⁹  pp60 c-Src kinase,¹⁰  phosphatidylinositol-3-kinase (PI-3kinase),^11,12 Janus kinases (Jaks),^13-15  and signal transducers and activators of transcription (Stats),^13,16,17  but the precise biologic implication of their activation in ALK⁺ ALCL cells remains unclear.

The Bcl-2 family members are important modulators of mitochondrially initiated apoptosis.¹⁸  This still-expanding family contains both prosurvival (Bcl-2, Bcl-X_L, myeloid cell-leukemia-1 [Mcl-1]) and proapoptotic (Bad, Bak, Bax) factors that share the presence of 1 to 4 Bcl-2 homology (BH) domains.¹⁹  Of these domains, BH3 represents a key element in regulating their homo-/heterodimerization and subcellular localization. Bcl-2 and Bcl-X_L exhibit both structural and functional similarity to prokaryotic pore-forming proteins, acting as ionic channels in the outer mitochondrial membrane for the control of mitochondrial swelling and loss of transmembrane potential.^20,21  The “BH3-only” proapoptotic Bad, which lacks the mitochondrial-membrane insertion C-terminal signal, is reversibly regulated by phosphorylation at multiple serine (Ser132, Ser136, and Ser155) residues.^22-24  In particular, phosphorylation at Ser136 is known to create a binding site for the phosphospecific binding of 14-3-3 proteins, which retain Bad in the cytoplasm preventing its translocation to the mitochondrial membrane.^23,25  In fact, upon a death signal, such as withdrawal of interleukin-3 (IL-3)²⁶  or insulin-like growth factor-1,²⁷  dephosphorylated Bad uses its BH3 death domain to bind the hydrophobic groove in the nonmembrane inserted portion of Bcl-X_L, which is then inactivated.^28,29  The collapse of mitochondrial transmembrane potential caused by the Bcl-X_L/Bad association occurs early in the apoptotic process, before other changes develop, such as nuclear DNA fragmentation and the exposure of phosphatidylserine (PS) on the outer surface of the cell membrane.^18,19  This early event in the commitment to apoptosis leads to the subsequent release of cytocrome c from mitochondria and activation of the intracellular apoptogenic effectors caspases.^28,30 

Recently, it has been demonstrated that NPM/ALK blocks the activation of caspase-3 by preventing the cytosolic accumulation of cytocrome c.^31,32  In addition, transcriptional induction of Bcl-X_L is elicited by NPM/ALK through the constitutive activation of the Jak3/Stat3 pathway, protecting cells from drug-induced apoptosis.^13,15 

In this report, the expression and biologic relevance of Bcl-X_L as a potential therapeutic target were evaluated in vitro and in vivo in NPM/ALK-transformed lymphoma cells.

The murine pro-B Ba/F3 cells stably transformed by NPM/ALK and the SUDHL-1 and SUP-M2 cell lines established from human T-cell ALCL (T-ALCL) carrying the t(2;5) were a kind gift from Dr Stephan W. Morris (St Jude Research Hospital, Memphis, TN). The t(2;5)-positive KARPAS-299 (K-299) ALCL cells were purchased from DSMZ (Berlin, Germany). Cells were cultured in RPMI-1640 (Life Technologies, Gaithersburg, MD) supplemented with 10% fetal bovine serum (FBS; GibcoBRL, Paisley, United Kingdom). Parental Ba/F3 cells were maintained in RPMI-1640 supplemented with 10% FBS and 0.2% WEHI-3B-conditioned medium as a source of IL-3 as described previously.¹⁴  All culture media contained 100 Units/mL penicillin, 100 μg/mL streptomycin, and 2 mM l-glutamine (GibcoBRL), and the cells were incubated at 37°C with 5% CO₂ atmosphere.

LY294002 (PI-3kinase pathway inhibitor) and herbimycin A (HA; nonspecific tyrosine kinase inhibitor) were purchased from Calbiochem (La Jolla, CA). Etoposide was from Sigma Chemical (St Louis, MO). All reagents were resuspended in dimethyl-sulfoxide (DMSO; Sigma Chemical) used as a solvent control in some experiments.

The pcDNA3.1-neo expression vector (Invitrogen, Groningen, The Netherlands) bearing the human cDNA for a kinase-defective mutant of NPM/ALK, named NPM/ALK.K210R (kindly provided by Dr Stephan W. Morris),⁹  was transiently transfected into Ba/F3-NPM/ALK⁺ cells by electroporation with a Biorad Gene Pulser apparatus (Biorad Laboratories, Hercules, CA) at 270 V, 1060 μF. The human Bcl-X_L expression vector pRSV-Bcl-X_L was a kind gift from Dr Fabrice Gouilleux (ICGM-Institut National de la Santé et de la Recherche Médicale [INSERM], France). The full-length Bcl-X_L cDNA was then subcloned into a plasmid vector enhanced green fluorescence protein (pEGFP) vector (Clontech, Palo Alto, CA) in a reverse 3′ to 5′ orientation. The resulting Bcl-X_L antisense expression construct was checked by sequencing and restriction enzyme digestion and transiently transfected into K-299 cells using Lipofectamine2000 Transfection kit (Invitrogen) according to the manufacturer's instructions. In order to enhance pEGFP-cytomegalovirus (CMV) promoter activity and antisense transgene expression, K-299 cells were incubated with phytohemagglutinin L (PHA-L) and phorbol myristate acetate (PMA) at final concentration of 1 μg/mL and 50 ng/mL, respectively, for 12 hours following transfection. A sequence of 496 base pairs derived from Bcl-X_L cDNA was ligated in an antisense orientation into pTRE vector (Clontech), containing a doxycycline (Dox)-responsive promoter (pTRE.ASBcl-X_L). To obtain an inducible expression of the antisense Bcl-X_L transgene, Ba/F3-NPM/ALK⁺ cells stably expressing a reverse Tet-repressor (rTetR) encoded by the pTet-ON vector (Clontech) were electroporated with 30 μg pTRE empty vector (pEV) or pTRE.ASBcl-X_L (pAS), and 5 μg pTK-Hygro vector (Clontech). Double-stable Tet-cell lines (Ba/F3-N/A-pEV or Ba/F3-N/A-pAS) were selected for 3 weeks with hygromycin-B (Boehringer Mannheim, Indianapolis, IN), and individual clones were obtained by limiting dilution. Selective antisense Bcl-X_L transcription was confirmed by reverse-transcriptase-polymerase chain reaction (RT-PCR) after addition of 1 μg/mL Dox in culture.

The monoclonal anti-ALK1 antibody was kindly provided by Dr Karen Pulford (John Radcliffe Hospital, Oxford, United Kingdom). The mouse monoclonal antiphosphotyrosine 4G10 antibody and the rabbit polyclonal total anti-Bad or anti-phosphoserine-136-Bad antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). Polyclonal antibody against Bcl-X_L (H-62) and goat polyclonal antibody anti-β-actin were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Bcl-2 (Ab-2) rabbit polyclonal was from Oncogene Research Products (Boston, MA). Protein lysates, prepared as previously reported,¹⁴  were separated by 8% to 16% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; Biorad Laboratories) and blotted onto Immobilon-P nitrocellulose membrane (Millipore, Bedford, MA). Membranes were blocked for one hour at room temperature in Tris (tris(hydroxymethyl)aminomethane)--buffered saline/Tween 20 (TBS-T) buffer containing 5% dried milk (for anti-ALK, anti-Bcl-X_L, anti-Bad, anti-phophoserine-136-Bad blots) or with 2% bovine serum albumin (for 4G10 and anti-Bcl-2 blots). Primary antibody incubation times were 1 to 16 hours. Secondary incubations were for one hour and antibodies used were horseradish peroxidase (HRP)-conjugated antimouse or antirabbit (Amersham, Arlington, Heights, IL). Proteins were visualized by chemiluminescence as recommended by the manufacturer (Super Signal; Pierce, Rockford, IL). Relative protein levels of Bcl-X_L were quantified by immunoblotting, and densitometric analysis of films was carried out on an Eagle Eye II Photodensitometer (Stratagene, La Jolla, CA).

The isolation of mitochondria was adapted from a previously described method.³³  Briefly, cells washed twice with ice-cold phosphate-buffered saline (PBS) were resuspended in a cold hypotonic buffer (10 mM KCl; 1 mM MgCl₂; 10% glycerol; 0.5 mM dithiothreitol [DTT]; 1 mM phenylmethylsulfonyl-fluoride [PMSF]; 15 mg/mL aprotinin; 3 mg/mL leupeptin/pepstatin; 2 mM Na₃V0₄; 0.2% nonidet P-40 [NP-40] with 20 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, pH 7.9]) and homogenized by using a Dounce homogenizer. The homogenate was mixed with an iso-osmotic MSE buffer (210 mM mannitol; 70 mM sucrose; 5 mM Tris-HCl, and 1 mM ethylenediaminetetraacetic acid [EDTA] pH 7.5) and centrifuged at 750g for 20 minutes at 4°C to pellet nuclei and unbroken cells. The pellet containing the mitochondria was resuspended in the MSE buffer and stored at -80°C. Protein concentration was determined using the bicinchoninic acid (BCA) protein assay (Pierce), and 100 μg protein was immunoprecipitated for 16 hours with 4 μg anti-Bcl-X_L polyclonal antibody (H-62). Bcl-X_L immunocomplexes were captured with 60 μg protein A-Sepharose beads (Pharmacia Biotech, St Alban, United Kingdom) for 4 hours at 4°C, separated on 16% SDS-PAGE, and analyzed by Western blotting as described above.

Changes in the mitochondrial transmembrane potential (MTP) were detected using a fluorescent cationic dye, 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-benzamidazolocarbocyanin iodide, commonly known as JC-1-fluorochrome (BIOMOL Research Laboratories, PA). Inside healthy cells, the lipophilic positively charged JC-1 dye aggregates in mitochondria and under fluorescent light emits a red signal (J-aggregates emit at 590 nm). Upon reduction in MTP, the JC-1 fluorochrome converts to a monomeric form that fluoresces green (monomeric dye emits at 490 nm). Cells were incubated with 2 μg/mL JC-1 in media for 30 minutes at 37°C. All stained samples were analyzed using a FACScalibur and Cell-quest software (Becton Dickinson, San José, CA).

The optimal conditions for Bcl-X_L protein depletion were established previously.³⁴  NPM/ALK-positive Ba/F3 cells were incubated for 3 days in the presence of either sense (S), antisense (AS), or nonsense (NS) oligodeoxynucleotides (ODNs) targeting the coding region 5′ to the translation initiation codon site for the Bcl-X_L gene (100% sequence identity between mouse and human) and extending 3′ downstream for a total of 18 bases. The DNA oligonucleotides with a phosphorothioated backbone were synthesized and purified by high-performance liquid chromatography (HPLC; MWG-Biotech, Ebersberg, Germany). The antisense AS-ODN sequence used was 5′-CCGGTTGCTCTGAGACAT-3′. There were 2 additional control ODNs used: (1) a complementary sequence to the antisense in the reverse 5′ to 3′ order, (S: 5′-ATGTCTCAGAGCAACCGG-3′), and (2) a missense sequence containing a 4-base mismatched compared with the antisense sequence (NS: 5′-CCGGTTGCTGTCACAGAT-3′). The lyophilized ODNs were reconstituted in sterile distilled water to 100 mM, filter-sterilized, and stored in aliquots at -20°C as stock solutions. Fresh oligonucleotides were added at 0, 16, 24, 32, 48, and 60 hours.

Apoptosis was assessed by several criteria. The percentage of apoptotic cells was determined by enumerating cells undergoing a reduction of cell volume with chromatin condensation or micronuclear fragmentation in a total of at least 200 cells using the intercalating fluorescent DNA binding dye, acridine orange (AO; Sigma Chemical). Cells were stained with 5 μM AO dissolved in PBS and examined using a Zeiss fluorescence microscope (Carl Zeiss, Oberkochen, Germany).

DNA degradation was further assessed by measuring the DNA content of individual cells by flow cytometry. Cells were washed with ice-cold PBS and fixed in 70% cold ethanol. Samples were then treated with Dnase-free Rnase (Boehringer Mannheim) and stained with 10 μg/mL propidium iodide (Sigma Chemical). Apoptosis was also detected by annexin V-fluorescein isothiocyanate (FITC) binding using the Apoptosis Detection Kit purchased from Bender MedSystems Diagnostic (Vienna, Austria). All samples were prepared following the manufacturer's instructions and analyzed using FACScalibur and Cell-quest software.

Female CD-1 nu/nu mice (7-9 weeks old) were supplied by Charles River (Calco, Como, Italy) and kept under standard laboratory conditions according to the guidelines of our institute. Animal studies were approved by the Ethics Committee for Animal Experimentation of the Istituto Nazionale Tumori, Milan, Italy. Mice were implanted with Ba/F3 cells transformed by NPM/ALK and transfected with the inducible antisense Bcl-X_L transgene (Ba/F3-N/A-pAS) or bearing an empty vector as control (Ba/F3-N/A-pEV) either intravenously (10 × 10⁶ cells/mouse) or subcutaneously (20 × 10⁶ cells/mouse). Therapeutic efficiency of the pAS vector was assessed by the addition of doxycycline hydrochloride in the drinking water (5% sucrose, 0.2 mg Dox/mL), which was changed twice a week. Mice were monitored every other day for the appearance of tumors and for signs of disease (weight loss, adenopathies). To analyze in vivo Bcl-X_L protein levels, subcutaneous tumor-bearing mice were killed and nonnecrotic tumor tissue was extracted and homogenized in a 5-fold volume of sodium dodecyl sulfate (SDS) loading buffer (50 mM Tris-HCl [pH 6.8], 2% SDS, and 5% β-mercaptoethanol). After sonication for 3 minutes, samples were centrifugated at 15 000g for 15 minutes, heated at 95°C for 10 minutes, and analyzed in Western blot as described above.

Statistical analysis was performed with Student test by use of the GraphPad software analysis program (Prism, San Diego, CA). P values of less .05 were considered to be statistically significant and were derived from 2-sided statistical tests. All data are presented as the mean (95% confidence interval [CI]). CIs are displayed when they exceed 10% of the respective mean. Correlation coefficient (CC) was calculated using the Pearson method.

Bcl-2 family members are well-known cell-death regulators.¹⁸  Bcl-2 and Bcl-X_L protect cells from apoptosis, whereas Bad increases susceptibility to cell death.^20,23 

The expression levels of Bcl-X_L and of Bcl-2 were evaluated in human NPM/ALK⁺ ALCL-derived cells carrying the t(2;5) translocation and in pro-B murine Ba/F3 transformed by NPM/ALK (Ba/F3-N/A).^9,12  Analysis on total lysates from Ba/F3 cells showed that NPM/ALK promotes an IL-3-independent Bcl-X_L protein induction of approximately 2.2-fold (as measured by band densitometry; Figure 1, lane c), compared with parental cells starved of IL-3 for 18 hours (Figure 1, lane a). Bcl-X_L expression is induced in parental Ba/F3 cells after stimulation with the growth factor IL-3, as previously reported by Dumon et al³⁵  (Figure 1, lane b compared with lane a). In 3 ALK⁺ ALCL-derived cell lines, Bcl-X_L protein is detected at levels higher than in Ba/F3-N/A⁺ cells. A direct relationship between Bcl-X_L expression levels and the intensity of NPM/ALK bands appears to be present (Figure 1, lanes c-f). The basal level of Bcl-2 in parental Ba/F3 cells remains low and unchanged after ectopic overexpression of NPM/ALK or IL-3 stimulation. These Bcl-2 levels are comparable with those detected in ALCL ALK⁺ cells, suggesting that NPM/ALK does not markedly affect Bcl-2 levels in these cells. The hybridization of the same filter with an anti-total Bad rabbit polyclonal antibody reveals that Bad migrates as a doublet of approximately 26 kDa, as previously described.^24,25  It has been proposed that these 2 distinct forms of Bad could reflect a hyperphosphorylated (pp-Bad) and hypophosphorylated (p-Bad) status of the protein that can be modified at multiple serine residues.²⁵  In Ba/F3 parental cells, 2 hours following withdrawal of IL-3, the hyperphosphorylated Bad was eliminated (Figure 1B, lane a), whereas 30-minute readdition of IL-3 restored the slower-migrating form of Bad (Figure 1B, lane b). The serine 136 has been identified as the major site of Bad phosphorylation by AKT.^23,26  Treatment with an irreversible inhibitor of PI-3kinase, LY294002, in presence of IL-3, impaired AKT activation at serine 473 and phosphorylation of Bad at serine 136, as indicated by using 2 phospho-site specific antibodies (Figure 1B, lane c). Likewise, treatment of Ba/F3-N/A⁺ cells and ALK⁺ ALCL K-299 with LY294002 inhibited AKT and Bad phosphorylation and decreased, but not completely abolished, the hyperphosphorylated form of Bad.

It can be concluded that NPM/ALK, when expressed in ALCL cells as a result of the t(2;5) chromosomal translocation or after transfection of NPM/ALK cDNA into Ba/F3 cells, is able to up-regulate Bcl-X_L (but not Bcl-2) and to promote a hyperphosphorylation of Bad that can be partially modulated by PI-3kinase/AKT signaling.

The tyrosine kinase activity of NPM/ALK is essential to elicit growth factor-independent proliferation and resistance to apoptosis in Ba/F3 cells. In fact, it is known that a kinase-defective mutant of NPM/ALK, named N/A.K210R, characterized by a point mutation in the critical adenosine triphosphate (ATP) binding site (K210R) fails to induce a transformed phenotype in Ba/F3 cells after IL-3 withdrawal.⁹  We used this construct as a dominant-negative inhibitor of ALK kinase activity and determined its effect on Bcl-X_L expression and survival in NPM/ALK-transformed cells. The N/A.K210R mutant was transiently overexpressed in Ba/F3-N/A cells. At 48 hours after transfection, anti-ALK (ALK) and antiphosphotyrosine (PY) immunoblots were performed to evaluate the expression and activation status of NPM/ALK (Figure 2A, upper panel). Decreased tyrosine phosphorylation of NPM/ALK, as well as of several other proteins, was observed in N/A.K210R transfectants, confirming the inhibitory action of N/A.K210R on wild-type NPM/ALK. The transient transfection of N/A.K210R also caused a substantial down-regulation of Bcl-X_L expression but did not change the Bcl-2 protein levels (Figure 2A, lower panel). In addition, N/A.K210R failed to promote Bad phosphorylation on serine 136 without affecting its total expression levels (Figure 2A, lower panel). Upon Bad dephosphorylation, Bcl-X_L/Bad dimerization plays a significant role in promoting apoptosis by inactivating mitochondrial Bcl-X_L.^28,29  Correlating with the inactivation of Bad and down-regulation of Bcl-X_L in Ba/F3-N/A.K210R-expressing cells, an increase in the sub-G₁ peak, indicative of apoptotic cells with subdiploid DNA (42.4%), was detected in these cells compared with control vector Ba/F3-N/A.EV transfectants (5.2%; Figure 2B). In addition, morphologic changes related to the exposure of phosphatidylserine (PS) on the outer surface of Ba/F3-N/A.K210R-transfected cells undergoing apoptosis (7.8% + 29.2%) were detected using an annexin V-propidium iodide staining (Figure 2C).

Treatment of ALK⁺ ALCLs with LY294002^11,12  as well as the nonspecific tyrosine kinase inhibitor herbimycin A (HA)^32,36  has been shown to promote induction of apoptosis. Therefore, we asked whether Bad dephosphorylation on serine 136, upon PI-3kinase/AKT pathway inactivation, was important to promote Bcl-X_L/Bad heterodimerization altering the mitochondria homeostasis (Figure 3). Mitochondrial extracts from K-299 cells untreated (NT) or incubated for 8 hours with vehicle alone (DMSO), 10 μM HA, or 25 μM LY294002 were immunoprecipitated using the anti-Bcl-X_L antibody and immunoblotted with anti-Bad and anti-Bcl-X_L antibodies. Both proteins were detected in samples treated with HA, and to a higher extent, with LY294002, whereas in untreated and vehicle-alone samples only Bcl-X_L was observed (Figure 3A). Western analysis on total lysates confirmed that Bad, mainly in its hypophosphorylated form, was coimmunoprecipitated with an anti-Bcl-X_L antibody upon LY294002 treatment, while in DMSO-treated cells Bad remained detectable as a doublet in the supernatant depleted of Bcl-X_L complexes (Figure 3A). This finding confirmed that a specific inhibition of PI-3kinase activity appears to dictate subcellular localization of Bad and its capacity to associate with Bcl-X_L. As a consequence of Bcl-X_L/Bad heterodimerization, an early event in the commitment to cell death is mitochondrial membrane depolarization, which occurs as a result of the opening of permeability transition pores.^21,28,37  We therefore analyzed the collapse of the mitochondrial transmembrane potential (ΔΦm) in K-299 cells after exposure to LY294002 for 8 hours using the JC-1 dye (Figure 3B). Blocking the PI-3kinase/AKT pathway caused cells to fall into the R5 region, indicating a loss of mitochondrial homeostasis as JC-1 converts from the red-aggregated form to the green-monomeric form.

These data indicate that NPM/ALK blocks the onset of apoptosis by preventing Bcl-X_L/Bad heterodimerization and associated preapoptotic mitochondrial perturbations. Therefore, NPM/ALK can deliver its antiapoptotic signal through modulation of the expression levels of Bcl-X_L and of its active/inactive conformational states via PI-3kinase/AKT/Bad signaling pathway.

The above reported results provide evidence for the hypothesis that NPM/ALK expression protects cells from cell death through the Bcl-X_L pathway.

Since antisense ODNs represent a potent approach to downmodulate the expression of specific genes,³⁸  we used antisense strategies targeted to Bcl-X_L in order to biologically validate this death regulator as a potential therapeutic target in NPM/ALK⁺ malignancies. Ba/F3-NPM/ALK⁺ cells were treated for 72 hours with increasing concentrations of antisense Bcl-X_L ODNs corresponding to the AUG-containing region of Bcl-X_L mRNA. A dose-dependent inhibition of Bcl-X_L protein levels was observed with an almost complete inhibition at 100 μM AS-ODN (Figure 4A). There was no effect of control ODNs representing sense (S) or nonsense scrambled (NS, 4 base mismatched) ODNs on Bcl-X_L expression. The 2 control oligonucleotides had some nonspecific effects on cell survival at 50 μM and 100 μM (data not shown). To avoid such nonspecific toxic effects, all subsequent experiments have been performed with 30 μM ODN. A time course experiment showed down-regulation of Bcl-X_L protein expression after 60 hours of incubation with Bcl-X_L antisense ODNs (Figure 4B). This slightly preceded the massive cell death induction in Ba/F3-NPM/ALK⁺ cells at 96 hours (Figure 4C). A considerable fraction of NPM/ALK-expressing cells incubated with Bcl-X_L antisense ODNs was found in either early (17.6%) or late (54.7%) stages of apoptosis, whereas control ODNs failed to modify the baseline proportion of apoptotic cells compared with the untreated cultures (NT). Figure 4D shows that treatment with Bcl-X_L antisense makes cells more sensitive to killing by etoposide, an antineoplastic drug currently used in the treatment of ALCL. When Ba/F3-N/A⁺ cells incubated with a nonsense oligonucleotide or no oligonucleotide were treated with etoposide, a dose-dependent induction of apoptosis occurred with 25% dead cells at 10 μM drug concentration. In contrast, the combination of the Bcl-X_L antisense oligonucleotide and 10 μM etoposide resulted in more than 40% cell death.

To determine the effect of Bcl-X_L down-regulation also in ALK⁺ ALCL cells, K-299 cells were transiently transfected with a plasmid encoding full-length Bcl-X_L antisense cDNA (ASBcl-X_L). Figure 5A shows a time-dependent inhibition of Bcl-X_L expression with a maximal effect at 96 hours after transfection. The percentage of apoptotic cells was analyzed by annexin V-propidium iodide staining. Figure 5B shows that down-regulation of Bcl-X_L caused approximately 40% of the cell population (13.2% + 26.6%) to undergo cell death (Figure 5B, left panels).

Furthermore, Bcl-X_L down-regulation was also associated with an increased susceptibility of cells to the genotoxic compound etoposide (Figure 5B, right panels), while vehicle alone had little effect (Figure 5B, middle panels), suggesting a synergistic effect of DNA damage and Bcl-X_L down-regulation on cell death induction.

The doxycycline (Dox)-inducible reverse tetracycline transactivator (rtTA) is frequently used for manipulations of transcription levels in a temporally regulated fashion in vivo.³⁹  Using a “pTet-ON” antisense system, murine Ba/F3-NPM/ALK⁺ cells were stably transfected with a plasmid encoding a Bcl-X_L antisense sequence under the control of a doxycycline (Dox)-responsive promoter (Ba/F3-N/A.pAS). After 96 hours of Dox treatment (1 μg/mL) in vitro, Ba/F3-N/A.pAS (N/A-pAS) cells were found to have less than 10% of the initial steady-state level of Bcl-X_L (Figure 6A), while the NPM/ALK expression levels were unaffected by the presence of Dox. At the same time, cell viability was significantly decreased (approximately 90% apoptotic cells) as indicated by acridine orange staining, while Ba/F3-NPM/ALK⁺ cells transfected with an empty vector control (N/A-pEV) did not show any cell death induction in the presence of Dox (Figure 6B).

Based on these in vitro results, we examined the ability of Dox-driven Bcl-X_L antisense transcription to inhibit the tumorigenic potential of NPM/ALK-transformed Ba/F3 cells in vivo. Ba/F3-NPM/ALK cells (10 × 10⁶) bearing a Dox-inducible empty vector (N/A-pEV) or a Bcl-X_L antisense (N/A-pAS) transcript were inoculated intravenously into nude mice (Figure 7A). Half of the mice from each group was maintained under continuous administration of Dox (0.2 mg/mL) in the drinking water. The N/A-pEV and N/A-pAS cell lines induced a fatal disease in 7 of 8 (Figure 7, white squares) and in 15 of 15 (Figure 7, black squares) control animals not receiving Dox, respectively, within a similar time frame (4-8 weeks). The animals showed weight loss, poor fur status, and inguinal/axillary adenopathies. Macroscopic examination of dead animals showed massive para-aortic adenopathies and splenomegaly (data not shown). Doxycycline administration prevented the occurrence of this wasting syndrome in 15 of 15 mice injected with N/A-pAS cells (Figure 7, black circles), while it did not affect the appearance and growth of NPM/ALK⁺ tumors in the control group (N/A-pEV) (Figure 7, white circles). The inhibitory effect of Dox lasted throughout the 2 months of its administration and was also stable after withdrawal of Dox for a further 2 months. Cells (N/A-pEV and N/A-pAS) were also injected subcutaneously, producing a local growth within 8 days (data not shown). In Table 1, Dox administration for 21 days (starting 24 hours after tumor cell injections) prevented the appearance of tumors in 9 of 12 mice inoculated with N/A-pAS cells, with no effect on tumor growth in mice injected with N/A-pEV cells. To further document the in vivo down-regulation of Bcl-X_L by Dox, animals bearing measurable subcutaneous N/A-pEV and N/A-pAS tumors were also treated with Dox. Figure 7B shows that Dox administration for 3 days reduced the expression of Bcl-X_L in 2 N/A-pAS regressing tumors without effects in control (N/A-pEV)-treated animals. NPM/ALK protein levels are used as loading controls and did not show substantial variations.

These findings reveal a crucial role for Bcl-X_L in mediating the in vivo transforming potential of NPM/ALK.

Protection from cell death after cytokine withdrawal or exposure to conventional anticancer drugs has been reported in human NPM/ALK⁺ ALCLs.^9,31  In particular, NPM/ALK exerts its antiapoptotic effect delaying cytosolic accumulation of cytocrome c and activation of caspase-3.^31,32  These findings support a mitochondrial control of the apoptogenic proteolytic cascade, although the involvement of Bcl-2 family members in ALK⁺ ALCL remains unclear.

In the present report we provide direct evidence of a functional link between Bcl-X_L and the survival potential of NPM/ALK. Several antisense strategies have been used to directly validate Bcl-X_L as a crucial cell fate regulator associated to the transforming ability of NPM/ALK in vitro and in vivo. Tyrosine kinase activity of NPM/ALK is required to up-regulate Bcl-X_L protein levels in pro-B murine Ba/F3 cells as well as in human ALK⁺ ALCL-derived cells, apparently without effects on Bcl-2 expression. The immunohistochemically undetectable Bcl-2 expression in human ALK-positive ALCLs has been recently correlated with the favorable clinical outcome of these lymphomas compared with ALK-negative ALCLs.^40,41  However, the expression of Bcl-X_L^13,15  and Mcl-1 antiapoptotic protein,⁴²  another Bcl-2-like member, could represent important prognostic factors for those ALK⁺ ALCL patients refractory to the current highly toxic conventional therapy.⁴³  Our findings in human ALK⁺ ALCL-derived cell lines indicate that Bcl-X_L may provide alternative mechanisms resulting in prolonged tumor cell survival. The involvement of Bcl-X_L in protecting human K-299 cells from apoptosis-inducing agents was investigated in vitro showing that Bcl-X_L down-regulation significantly increased their susceptibility to etoposide. In addition, up-regulation of Bcl-X_L in murine Ba/F3 cells transformed by several hybrid oncogenes, including NPM/ALK, has been recently proposed to work in concert with the overexpression of RAD51 to delay the G₂/M phase transition of the cell cycle, thereby promoting drug resistance.⁴⁴  Therefore, these findings support the possibility that inappropriate Bcl-X_L expression or its functional modulation may represent an important indicator of the outcome of disease treatment and also has important therapeutic implications in NPM/ALK-mediated oncogenesis. In fact, a few resistant NPM/ALK-expressing cells may be selected during remission by a cytotoxic drug, since a third of “ALKoma” patients experience multiple relapses.^45,46 

Bcl-X_L has been identified to be involved in the neoplastic transformation caused by several oncogenes such as BCR/ABL.^47,48  Increased expression of Bcl-X_L has been reported in acute myeloid leukemia (AML) and multiple myeloma patients with poor clinical response.^49,50  The Jak/Stat and PI-3kinase/AKT activation pathways have been involved in Bcl-X_L regulation in multiple hematopoietic cell lineages.^51,52  There are 2 recent reports using cell lines derived from ALK⁺ ALCLs that have shown that Bcl-X_L, but not Mcl-1 or survivin, can be downmodulated by treatment with the Jak/Stat inhibitor AG490.^13,15  In particular, the Bcl-X_L expression rate seems to be strictly related to the NPM/ALK-induced high level of Stat3 protein as well as its constitutive activation in ALK⁺ ALCLs.¹³ 

We also investigated the possibility that posttranslational mechanisms, other than transcriptional induction of Bcl-X_L, might influence the balance of competing death/life of ALK⁺ ALCLs toward survival.

In our experiments, we were unable to detect any coimmunoprecipitation with NPM/ALK and phosphorylated form of Bcl-X_L (data not shown), as instead previously described for Bcl-2.⁵³  In contrast, the Bcl-X_L antagonist Bad has been found to be primarily regulated by phosphorylation. A PI-3kinase/AKT-mediated phosphorylation of Bad on serine 136 is required for coupling NPM/ALK-delivered survival signals to the mitochondrial death machinery. In fact, LY294002-induced Bad dephosphorylation was able to promote association of Bcl-X_L/Bad heterodimers triggering the onset of commitment to apoptosis as indicated by the sudden mitochondrial membrane depolarization. Other than AKT serine/threonine kinase,^23,26  the reversible phosphorylation of Bad can be modulated by survival-promoting kinases including Ras-induced mitogen-activated protein (MAP) kinases,²²  p21-activated kinase 1,⁵⁴  protein kinase A (PKA)⁵⁵  as well as phosphatases, such as protein phosphatase 1 (PP1)⁵⁶  and a PP2a-like enzyme.⁵⁷  This could explain the slower mobility form of Bad, which does not appear to be influenced by inhibition of PI-3/AKT kinases in ALK⁺ ALCL cells. However, further studies are warranted to clarify the biologic implications of these pathways in terms of NPM/ALK-enhanced survival and regulation of Bcl-2 family proteins.

The significant in vitro and in vivo apoptotic effects of Bcl-X_L down-regulation in different cell types including human ALK⁺ ALCL-derived cell lines highlight another important point: the selection of therapeutic targets downstream of known oncogenes. It is known that NPM/ALK activates several signal transduction pathways, therefore it could be hypothesized that the down-regulation of a single protein (or its inhibition) could be easily bypassed by activation of parallel pathways. The data presented here indicate that in the context of NPM/ALK-mediated oncogenicity the presence of Bcl-X_L remains crucial and accounts for most of the survival potential of neoplastic cells. This point is of general importance because primary oncogenes are not always easy targets for therapies, especially when they do not code for an enzyme.

It is important to note in this regard that there is presently considerable interest in discovery of small-molecule inhibitors targeted at the BH3-binding hydrophobic groove of Bcl-X_L, which can antagonize its pore-forming activity^58,59  and protein-protein interactions.⁶⁰  These attempts are based on the hypothesis that Bcl-X_L down-regulation in normal cells will not result in significant toxicity, an important aspect that was not investigated here and will require further analysis.

In conclusion, our results provide evidence that Bcl-X_L is positively regulated by NPM/ALK in human lymphoid cells and could represent a relevant target for a new generation of therapeutics for the treatment of ALK⁺ ALCLs. By altering the ratio of prosurvival and proapoptotic proteins through Bcl-X_L down-regulation or inhibition, it may still be possible to antagonize the pleiotropic antiapoptotic cascades activated by NPM/ALK in lymphoma cells.

We are grateful to Dr Stephan W. Morris (St Jude Research Hospital, Memphis, TN) for the NPM/ALK-transformed Ba/F3 cells as well as for NPM/ALK kinase-defective expression construct. We thank Dr Karen Pulford (John Radcliffe Hospital, Oxford, United Kingdom) for the monoclonal anti-ALK1 antibody and Dr Fabrice Gouilleux (ICGM-INSERM, Aimes, France) for kindly providing the pRSV-Bcl-X_L cDNA expression vector.