Ablation of oncogenic ALK is a viable therapeutic approach for anaplastic large-cell lymphomas

Anaplastic large-cell lymphomas (ALCLs) represent a subset of neoplasms with distinctive genetic defects that are immunophenotypically characterized by the sustained expression of CD30.¹  In the late 1980s several groups described the unique association of the t(2;5)(p23;q35) chromosome translocation with ALCL. However, the corresponding genes were discovered several years later by Morris and colleagues.²  In the t(2;5)(p23;q35) translocation the ALK (anaplastic lymphoma kinase) gene on chromosome 2 is fused to the NPM1 (nucleophosmin) gene on chromosome 5. Recently, several groups have successfully cloned many additional ALK translocations.³  Translocations of ALK have also been discovered in inflammatory myofibroblastic tumors (IMTs),⁴  and deregulated expression of ALK has been documented in a subgroup of diffuse large-cell lymphomas (DLCL).^5-7  Notably, the constitutive expression of the wild-type ALK receptor (ALK-R) has been shown in several cancer cell lines⁸  and in primary neuroblastoma and rhabdomyosarcoma.^9,10  The ALK gene encodes a tyrosine kinase receptor whose physiologic expression in mammals is largely limited to neuronal cells.^11,12  At present, the physiologic role of ALK is still largely unclear. Nevertheless, ALK and its ligand play an important role in the formation of visceral muscle in Drosophila,^13,14  and in the regulation of neurologic synapses in Caenorhabditis elegans.¹⁵  All ALK fusion proteins maintain the intracytoplasmic moiety of the ALK-R at their C terminus. This region, which contains the catalytic domain, is required for cellular transformation,¹⁶  whereas the N-terminal fusion partners act as dimerizing domains.^17,18  Activated ALK chimeras bind multiple adaptor proteins such as Grb2, Shc, IRS-1, and p130Cas, which link them to diverse pathways regulating cell proliferation, survival, and cell transformation. Among them, PLC-γ, PI3K, and Jak3 lead to the activation of numerous downstream molecules including cyclin D, Erk1/2, STAT3, and AKT.^19,20  Using cell lineage–specific conditional knock-out models, we have recently demonstrated that the genetic ablation of STAT3 in ALK⁺ cells leads to T- and B-cell death, and prevents the generation of B-cell neoplasms.²¹  Previously, PLC-γ and AKT have been shown to play an essential role in ALK-mediated transformation, in vitro.^16,22-24  Thus, it is conceivable that the inhibition of ALK could induce biologic changes capable of inhibiting cell growth and/or promoting cell death. This hypothesis is supported by a study in which a nonselective, but active drug against ALK was shown to successfully lead to tumor cell death.²⁵  Nonetheless, the specific down-modulation of NPM-ALK mRNA via siRNA does not induce apoptosis, despite a significant reduction in NPM-ALK protein.²⁶  Thus, the formal demonstration that the specific loss of ALK kinase activity has relevant biologic effects is still missing.

Several kinase inhibitors were recently developed and successfully used in the treatment of human tumors. Systematic analyses in vitro and in vivo have shown that some of these inhibitors can efficiently block not only the kinase that was designed to be targeted, but neutralize other enzymes as well. In this context, imatinib, originally developed for Bcr-Abl,²⁷  also efficiently hinders the activity of c-kit and PDGFR molecules. No specific inhibitor(s) against ALK are available, and thus alternative strategies designed to inhibit and/or neutralize ALK may have great relevance. Several approaches to block ALK functional activity including small peptides²⁸  and antisense molecules²⁹  might be envisioned. More recently, silencing the expression of a specific gene has become possible using RNA interference (RNAi).^30,31  Targeting of Bcr-Abl via RNAi, for example, has proven feasible in inhibiting growth of leukemic tumor cells^27,32  acting similarly to imatinib, the “gold-standard” drug for the treatment of Bcr-Abl⁺ leukemias.²⁷  Taking advantage of this new approach, novel avenues for the treatment of human diseases might be opened by pivotal trials under investigation.³³ 

To further dissect the role of ALK in the pathogenesis of ALK⁺ ALCL, we have developed stable small hairpin RNAs (shRNAs) targeting sequences coding for the intracytoplamic region of ALK. This approach efficiently and specifically down-regulates ALK protein expression, resulting in the reversion of ALK-mediated tumor phenotype in vitro and in the impairment of cell growth of ALK-transformed mouse embryonal fibroblasts (MEFs) in vivo. More importantly, ALK shRNA specifically induced cell death of human ALK⁺ ALCL cell lines, and inhibited their growth in vivo. These data demonstrate that ALK is required for the growth and survival of ALK⁺ ALCLs. Thus, the shRNA approach is not only a powerful tool to characterize signals that mediate tumorigenesis of ALK⁺ ALCLs, but also represents a potential strategy for the treatment of ALK⁺ neoplasms.

We designed 14 short hairpin oligonucleotides (Invitrogen, Carlsbad, CA) directed against the cytoplasmic domain of the ALK-R. shALK sequences were identified following the recommendations of Brummelkamp et al³⁴  and/or a dedicated program (Ambion, Austin, TX). The sense strand of the ALK-A5 shRNA sequence is: gatccccGGGCGAGCTACTATAGAAAttcaagagaTTTCTATAGTAGCTCGCCC tttttggaaa. The 19–nucleotide (nt) ALK target sequences are indicated in uppercase letters, whereas the hairpin and the sequences necessary for the directional cloning are depicted in lowercase letters. Restriction endonuclease sequences (5′-BlgII and 3′-HindIII) for directional cloning into the pSuper vector (Oligoengine, Seattle, WA) were included. Single-strand ALK oligonucleotides were first annealed and then cloned into the BglII-HindIII sites of expression vectors as described.³⁴  The fidelity of cloned double-strand DNA was confirmed by DNA sequencing of both DNA strands using specific oligoprimers.

Cassettes containing the H1 promoter and shALK sequences were subcloned into a pSuperRetro vector (Oligoengine) previously modified by the insertion of a cytomegalovirus–enhanced green-fluorescent protein (CMV-EGFP) (XhoI-blunt) reporter to yield the pSuperRetro-GFP (pSRG).

LV-shALK vectors were constructed by subcloning the H1 promoter-shALK cassette into the EcoRV-XhoI sites of the vector pCCL.sin.PPT. hPGK.GFPWpre, kindly provided by Dr Luigi Naldini (San Raffaele Scientific Institute, Milan, Italy).

NPM-ALK, ATIC-ALK, ALK-R, or TPR-MET expression was achieved via the CMV promoter of the episomal retroviral vector Pallino,³⁵  carrying EGFP under the transcriptional activity of the Moloney leukemia long terminal repeat region (LTR).³⁶ 

The inducible expression of NPM-ALK was obtained using a Tet-Off bidirectional pBI-EGFP vector (Clontech, Palo Alto, CA) in which NPM-ALK (HindIII-XhoI), after blunting, was subcloned into the MluI cloning site. NPM-ALK expression was repressed by culturing NPM-ALK–transfected MEF-3T3 Tet-Off cells (Clontech) in the presence of doxycycline (1 μg/mL).

HEK-293T (ATCC, Manassas, VA), GP-293 (Clontech) packaging, and MEF Tet-Off cells were cultured under standard conditions (37°C in humidified atmosphere, with 5% CO₂) in Dulbecco modified Eagle medium supplemented with 10% fetal calf serum (FCS). Human ALK⁺ (TS [a subclone of SUP-M2], Karpas 299, SU-DHL-1 [kindly provided by Dr A. Lorenzana, Department of Pediatrics, University of Toronto, ON, Canada]),³⁷  murine ALK (CD4-43), and ALK^– (Jurkat and CCRF-CEM) cells were grown in RPMI 1640 medium with 10% FCS.

Transfections of HEK-293T, GP-293, and MEF Tet-Off cells were performed with Effectene reagent (Qiagen, Valencia, CA), according to the manufacturer's instructions. The NPM-ALK MEF Tet-Off cell line was generated by transfecting pBI-EGFP-NPM-ALK vector in Tet-Off MEF cells. Clones were selected based on their lowest basal NPM-ALK protein expression in presence of doxycycline and their highest expression in absence of the drug.

Retroviral supernatants were produced by cotransfection of GP-293 packaging cells with pSRG vectors and pVSV-G vector (Clontech). NPM-ALK Tet-Off MEF cells infected with pSRG retroviruses were enriched by selection with puromycin (1 μg/mL, for 7 days).

High titer lentiviral vector stock was produced in HEK-293T cells by calcium phosphate–mediated transfection of the modified transfer vector and the packaging vectors pMDLg/pRRE, pRSV-Rev, and pMD2.VSVG.³⁸  Supernatants were harvested over 36 to 60 hours, filtrated (0.22-μm pore), and used directly or after concentration by ultracentrifugation (50 000g for 2 hours). Virus titers were assessed by transducing HeLa cells with serial dilutions of viral stocks. Aliquots of virus, plus 8 μg/mL of polybrene, were used to infect exponentially growing cells (1 × 10⁵/mL). Fresh medium was supplemented at 24 hours after the infection. The infectivity was determined (after 72 hours) by FACS analysis of EGFP⁺ cells.

To study the growth patterns of ALCL cells infected with shRNAs, TS cells were first infected with ALK-A5, ALK-A6, or empty retroviral supernatants. Three days after infection cells were extensively washed with phosphate-buffered saline (PBS) and seeded (1 × 10⁴/well) in different concentration of FCS or in presence of decreasing concentration of bleomycin or cyclophosphamide with 10% FCS. Cells were appropriately fed every 48 hours. Percentages of EGFP⁺ cells within viable cells were calculated by flow cytometry.

MEF Tet-Off and TS cells were grown in complete cell-culture media onto glass cover-slips, rinsed in PBS, and fixed with ice-cold methanol for 20 minutes at –20°C, followed by permeabilization for 10 minutes with 0.25% Triton X-100 in PBS at room temperature. Immunofluorescence stainings were performed with mouse monoclonal anti-ALK (4C5B8) antibody (Zymed, South San Francisco, CA), and detected with antimouse biotinylated antibodies (Sigma Aldrich, St Louis, MO), followed by streptavidin-Cy3 (Sigma Aldrich). Samples were counterstained with Hoechst 33258 (Sigma Aldrich). Coverslips were mounted in antifading solution and viewed using a Leica TCS SP2 laser-scanning confocal microscope driven by Leica confocal software; the images were acquired at room temperature by means of a 63 × PL APO objective (numeric aperture [NA] 1.32) (Leica, Heidelberg, Germany). Bright-field images were acquired on a Leica DM IRE2 microscope using 40 × (NA 0.55) and 20 × (NA 0.40) objectives, and a DC300F camera, and were analyzed with IM 50 software.

For Western Blot analysis, cells were lysed (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Triton X-100, 5 mM EDTA, 1 mM Na₃VO₄, 1 mM phenylmethyl sulfonyl fluoride, and protease inhibitors). Total protein lysates (20 μg) after sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) were transferred onto nitrocellulose membranes. Filters were first blocked (5% lowfat milk in PBS with 0.1% Tween 20; 1 hour at room temperature [RT]), and then incubated with primary antibodies for 1 hour at RT. After 3 washes, filters were incubated with horseradish peroxidase–conjugated goat antimouse or antirabbit antibodies (1:5000; Amersham, Piscataway, NJ) for 1 hour at RT. The detection of immunocomplexes was performed with an enhanced chemiluminescence system (ECL; Amersham).

The following primary antibodies were used: mouse anti-ALK (4C5B8), anti-SKP2 (2C8D9), and anti-STAT3 (5G7) from Zymed; mouse anti– α-tubulin (B-5-1-1) from Sigma Aldrich; mouse anti-PCNA (PC10) and cyclin D1 (DCS6) from DAKO (Fremont, CA); rabbit anti–phospho STAT3 Tyr705, phospho AKT Ser473, phospho ERK1/2 Thr202/Thy204, AKT, cleaved caspase 3, cleaved caspase 6, cleaved caspase 7, and cleaved caspase 9 from Cell Signaling Technology (Beverly, MA); mouse anti–cyclin E (HE12) and p53 (DO1), rabbit anti–cyclin E (M20), cyclin A (H-432), cyclin B1 (H-433), Jun B (N-17), ERK1 (C-16), ERK2 (C-14), PARP (H-250), BclX (S-18), and eIF2α (FL-315) from Santa Cruz Biotechnology (Santa Cruz, CA); rabbit anti–phospho eIF2α (Ab1) and survivin (Ab1) from Oncogene Research (San Diego, CA); mouse anti-p21/waf1 (70), p27/kip1 (57), and RB (G3-245) from BD Pharmingen Biosciences (San Diego, CA); and mouse anti–cyclin D3 (DCS-22) from Neomarkers (Fremont, CA).

NPM-ALK MEF Tet-Off cells were infected with pSRG or pSRG-A5 supernatants and puromycin-selected in the presence of doxycycline for 7 days; NPM-ALK expression was achieved by the removal of doxycycline and confirmed by Western blotting. Enriched EGFP⁺ cells (1 × 10⁶; > 90%) were cultured in the absence of doxycycline (48 hours) and then injected subcutaneously into athymic Nu/Nu mice (Charles River Laboratories, Wilmington, MA). Tumor growth was scored weekly over a period of 4 weeks by determining the volumes of tumor masses. Lentivirus-infected TS cells (100% EGFP⁺) were injected (2 × 10⁶) into FoxChase (C.B-17) severe combined immunodeficient (SCID) mice (Charles River Laboratories) and tumor growth was determined over time. Animals carrying tumor masses of 1 cm³ were injected intratumorally with 50 μL (1 × 10⁸ transduction mits [TU]) of ALK-A5 or control concentrated lentivirus preparations every alternate day. Mice were ultimately killed 1 day following the third viral injection. Animals were housed in the animal facility of the New York University School of Medicine and treated under the NIH animal guidelines. Animal studies were approved by the NYU animal committee.

For DNA content studies, 5 × 10⁵ cells were fixed for 1 hour in 70% ethanol at 4°C. After washing, cells were treated with RNase (0.25 mg/mL) and stained with propidium iodide (50 μg/mL) or with 7-aminoactinomycin D (7-AAD; BD Pharmingen Biosciences). The G₁/G₀-phase fraction was calculated using the CellQuest program (BD Pharmingen Biosciences).

For BrdU labeling, 5 × 10⁵ cells were incubated in complete cell-culture media in the presence of BrdU (10 mM) for 45 minutes at 37°C in 5% CO₂. BrdU incorporation was determined by flow cytometry, using the APC BrdU flow kit (BD Pharmingen Biosciences) according to the manufacturer's instructions.

Apoptosis was measured by flow cytometry after staining with annexin V or with the mitochondrion-permeable voltage-sensitive dye tetrametylrodamine methyl ester (TMRM; Molecular Probes, Eugene, OR).³⁹  Cells (5 × 10⁵) were washed once in PBS, incubated for 15 minutes at 37°C in HEPES buffer solution (10mM HEPES pH 7.4, 140mM NaCl, 2.5mM CaCl) with 2 μL biotin-conjugated annexin V (BD Pharmingen Biosciences), or with 200 nM TMRM. annexin V binding was revealed by additional incubation with streptavidin-PE (BD Pharmingen Biosciences). Cells were analyzed by FACScan using CellQuest Program. TUNEL (terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling) assay was performed following manufacturer's instructions (Roche, Indianapolis, IN).

We designed shRNA spanning the cytoplasmic domain of ALK-R since this region is conserved in all oncogenic fusion proteins.³  We reasoned that a functional shRNA specifically recognizing the cytoplasmic moiety would result in broad-spectrum inhibition of ALK expression, targeting all ALK chimeras and ALK-R. This strategy appears advantageous in view of the fact that the expression of ALK-R is also deregulated in nonlymphoid neoplasms. Fourteen shRNA sequences were generated. ALK shRNA pSuper (pS) and NPM-ALK cassettes were first cotransfected into HEK-293T cells. We identified a single construct (ALK-A5) that potently down modulated (> 90%) the ectopic expression of NPM-ALK (Figure 1A). Partial abrogation was observed with a second shRNA, also recognizing a sequence within the intracytoplasmic domain (A7). Expression of a p53 shRNA pS cassette³⁴  did not result in changes of ALK expression, despite robust down-modulation of the p53 (data not shown). Similar findings were obtained using a pS carrying a shRNA specific for the break point of TPR-MET fusion protein (Figure 1B).³⁸  To validate the efficacy of ALK-A5 against other ALK fusion proteins and ALK-R, we ectopically expressed ATIC-ALK and ALK-R in HEK-293T cells transfected with several ALK shRNAs. In this set of experiments, ALK-A5 efficiently abrogated the expression of both ATIC-ALK fusion and of the full-length ALK-R proteins (Figure 1C-D).

We next determined whether ALK-A5 shRNA could abrogate or revert the transforming properties of NPM-ALK, ectopically expressed in MEF cells. We employed an inducible expression system in which ALK-mediated transformation is tightly regulated after doxycycline withdrawal. MEFs carrying the NPM-ALK Tet-Off expression system were infected with the ALK-A5 pSRG virus in the presence of doxycycline. The percentages of infected cells before and after puromycin selection were monitored using EGFP as a reporter. Figure 2A shows a significant down-regulation of NPM-ALK protein expression in the absence of doxycycline when the cells were transduced with ALK-A5 shRNA. The degree of protein ablation was proportional to the percentages of EGFP-positive cells (percentages of infected cells). Moreover, the loss of NPM-ALK expression lead to decreased levels of phosphorylation of Stat3, a known ALK substrate,⁴⁰  and to the down-modulation of JunB, a protein highly expressed in ALCLs,⁴¹  whose transcription is positively modulated via NPM-ALK (R. C. and G. I., manuscript in preparation).

Since NPM-ALK expression in MEF cells also leads to reproducible morphologic changes, including cell elongation and bipolarization, and loose adhesion to the tissue culture plate (Figure 2B, top panels), we next studied the phenotypic changes associated with ablation of NPM-ALK. NPM-ALK⁺ MEFs infected with ALK-A5 shRNA (pSRG-ALK-A5 NPM-ALK MEFs) expressed undetectable levels of NPM-ALK by immunofluorescence and became large and oval, with rounded nuclei, and strongly adhered to the plate in the absence of doxycycline (Figure 2B, bottom right panel). These features were virtually identical to those seen in NPM-ALK MEF cells grown in the presence of doxycycline and lacking ALK expression. ALK-A5 pSRG NPM-ALK MEFs not only lack the transformed phenotype, but also the ability to grow in soft agar (data not shown). Finally, these cells (> 95% EGFP⁺) when injected into athymic Nu/Nu mice (1 × 10⁶ cells/mouse), generated significantly smaller tumors than those derived form cells infected with the control pSRG vector (Figure 2C).

ALK shRNA was subsequently tested in multiple human ALK-positive ALCL cell lines. To determine if the down-regulation of NPM-ALK expression would lead to inhibition of cell growth, we transduced human ALCL cells with pSRG, pSRG-A5, pSRG-A6, or pSRG-TM2 retroviral supernatants. Jurkat and CCRF-CEM cell lines were used as ALK^– controls. Percentages of the EGFP⁺ cells ranged from 15% to 70% in the different cell lines, with TS (ALCL, ALK⁺) being the most efficiently transduced cell line (Figure 3A). After infection, the overall percentages of EGFP ALK-A5⁺ cells decreased overtime in all ALK⁺ cell lines. Similar findings were obtained with several murine NPM-ALK–positive cell lines derived from primary T lymphoblastic lymphomas of NPM-ALK Tg mice (Figure 3E).⁴²  On the other hand, EGFP⁺ CCRF-CEM and Jurkat cells (ALK^–, Figure 3B-F; and data not shown) remained unchanged following transduction with controls or pSRG-A5 viruses. Likewise, the percentages of EGFP⁺ ALCL cells infected with TM2, ALK-A6, or empty control vectors remained unvaried over time (data not shown). BrdU incorporation assays showed that pSRG-A5 ALCL cells displayed a lower percentage of proliferating cells compared with uninfected cells within the same cultures or with control infected cells (Figure 3G-H).

To study whether ALK-A5 shRNA could sensitize infected cells to stress conditions, we first cultured recently transduced (3 days) ALCL cells in suboptimal concentrations of FCS. As shown in Figure 4A, the percentages EGFP⁺ ALK-A5–transduced TS cells cultured with low concentrations of FCS decreased rapidly as a percentage of total. In contrast, the loss of EGFP was less evident and required longer periods of time when the same cells were cultured in 10% FCS. Moreover, under the same conditions, EGFP^– cells displayed a higher percentage of BrdU incorporation (data not shown). Similar findings were observed when ALK-A5⁺ cells, but not control cells, were grown in the presence of low doses of chemotherapeutic agents such as bleocin or cytoxan (Figure 4B-C).

The amount of time required to achieve detectable biologic effects after retroviral infection of ALK-A5 might be due to the long half-life of the NPM-ALK protein, and/or to low-level shRNA expression. In fact, the protein levels of NPM-ALK in these cells were only slightly decreased (data not shown). Since concentrated VSV-G lentiviruses can efficiently infect the majority of the target cells, and allow a more potent expression of transgenes, we generated a lentivirus cassette carrying the ALK-A5 shRNA under the H1 promoter.⁴³  An empty EGFP⁺ lentivector or a lentivector carrying shRNA with mutated ALK-A5 (ALK-A5M) or specific for TPR-MET (TM2) sequences were used as controls. Lentivirus infection led to more than 99% transduction (based on EGFP expression) of all human ALK⁺ cells and the relative intensity of EGFP⁺ cells increased proportionally with higher viral titers (data not shown). TS cells transduced with ALK-A5 lentivirus showed an increased expression of EGFP over time, and this was associated with a down-modulation of ALK protein expression, detectable as early as 72 hours after infection (Figure 5A-B). Known targets of NPM-ALK signaling, such as the phosphorylation of STAT3 and AKT, and the expression of JunB, were concomitantly down-modulated in ALK-A5⁺ cells (Figure 5C).

Moreover, transduction with ALK-A5 lentivirus induced G₀/G₁ arrest by 96 hours after infection (Figure 6A). Cell-cycle arrest was sustained by the concomitant down-modulation of cyclin A, cyclin B, and Skp2, by the up-regulation of p27 and p21, which resulted in hypophosphorylation of RB (Figure 6B). In parallel with the growth arrest, NPM-ALK silencing was associated with an increased number of apoptotic cells at 4 and 5 days after infection (Figure 6C). Accordingly, activation of effector caspases known to precede overt apoptosis, together with the down-regulation of the antiapoptotic protein survivin, were distinctively detected in ALK⁺ ALCL cells 4 days after transduction with ALK-A5 (Figure 6D).

Finally, to verify that “off-target” silencing effects were not responsible for the observed decrease in cell growth and viability, we performed a functional rescue experiment. We designed an NPM-ALK expression construct carrying 4 conservative nucleotide mutations, without any changes in the corresponding translated amino acids, within the ALK-A5 target sequence. TS cells ectopically expressing the mutated NPM-ALK construct were resistant to the ALK-A5 silencing, remained viable, and continued to grow normally (data not shown). The specificity of the shRNA ALK-A5 findings was further validated, demonstrating that the protein expression of known genes, whose regulation is modulated in response to interferon-mediated responses, was not significantly changed (Figure 6D, P-eIF2α and eIF2α).

To study whether the loss of NPM-ALK could modulate the growth of human ALCL cells in vivo, we used 2 different strategies. In the first approach, ALCL cells were infected with ALK-A5 or control virus and then, after 48 hours of culture (> 95% EGFP- and < 10% annexin V⁺ cells), were injected into FoxChase (C.B-17) SCID mice. Tumor growth was monitored over time. TS cells in which NPM-ALK expression was silenced grew far less efficiently than controls (Figure 7A). In a second series of experiments, we treated large (1-cm) xenograft ALCL tumor masses with multiple injections (3 injections on alternate days) of concentrated ALK-A5 or control lentiviral preparations (0.5 to 1 × 10⁸ TU) and showed that the tumor growth of neoplastic cells was significantly inhibited by lentiviruses carrying the ALK-A5 shRNA construct compared with the control ALK-A6 vector (Figure 7B) (P < .001). Flow cytometric analysis of EGFP⁺ tumor cells showed that approximately 50% of transplanted cells were efficiently transduced in vivo. Notably, the very large majority of EGFP ALK-A5⁺ cells were also annexin V positive, in contrast to a smaller fraction of control (ALK-A6) cells (Figure 7C, left quadrants). Finally the histologic sections of the tumors (obtained 1 day after the last injection) demonstrated that tumor masses treated with ALK-A5 viral preparations were largely necrotic, and in more viable areas, many apoptotic cells were interspersed among viable cells. At this time, no significant tissue remodeling or fibrosis could be observed. On the contrary, control samples were largely viable (Figure 7C, middle panels). TUNEL analyses further confirmed these findings, demonstrating that the very large majority of the ALK-A5⁺ cells display DNA genomic breaks (Figure 7C, right panels).

Small interfering RNAs have proven to be a useful tool for gene silencing and represent a promising approach for studying the tumorigenic role of many genes. To validate ALK as a target for therapeutic intervention, we designed a specific ALK shRNA construct which effectively down-modulated ALK protein expression. NPM-ALK ablation led to cell growth inhibition, loss of the tumor phenotype of transformed MEFs, and cell death of human ALK⁺ ALCL lymphocytes in vitro and in vivo. These findings clearly demonstrate that ALK fusion proteins are suitable targets for therapeutic intervention in ALK⁺ ALCLs and possibly for other human tumors in which ALK-R could play a pathogenetic role. Accordingly, silencing ALK tyrosine kinase activity via RNAi technologies or alternative approaches, such as antisenseoligonucleotides, tyrosine kinase inhibitors, or peptidomimetics should be actively pursued for the treatment of ALK⁺ tumors.

Several studies have demonstrated that RNAi targeting oncogenic chimeric RNA sequences can be successfully used to down-modulate the expression of oncoproteins.^27,32,44  Since ALK-R expression is a feature of some nonlymphoid tumors including neuroblastomas, rhabdomyosarcomas, and other mesenchymal tumors, we designed a series of shRNA constructs directed against sequences encoding the cytoplasmic domain of the ALK oncogene. Based on the physiologic expression pattern of ALK-R,^11,12  and the absence of a striking phenotype in knockout mice,⁴⁵  the inhibition of ALK-R in normal cells is expected to produce only minor to nil effects. In this perspective, we should consider that ALK-R mRNA transcripts have been recently detected in many human tumors.^9,10  If these findings are confirmed and a pathogenetic role of ALK in these tumors is demonstrated, the availability of ALK inhibitors and or RNAi will have considerable clinical implications.

RNAi approaches have also been proven to be specific in vivo^46,47  and recently a novel clinical trial using chemically modified siRNA was initiated.⁴⁸  In our study, we confirmed that shRNA leads to the specific loss of ALK. No changes in ALK protein expression were seen when unrelated RNA interfering sequences were transduced in ALK⁺ cells and no detectable biochemical and/or biologic effects were observed when ALK-A5 shRNA was expressed in several ALK^– cells. More importantly, the absence of significant interferon-mediated responses and the functional rescue experiments excluded the possibility that unintentional silencing of nontarget genes was responsible for the observed reduction in cell viability and cell growth.

This study demonstrates that the sustained ablation of NPM-ALK protein expression promotes human lymphoid cells carrying the ALK fusion protein to undergo apoptosis. These findings strongly suggest that ALK is absolutely necessary for the survival of ALK⁺ lymphoma cells and essential to maintain their overall growth. These findings are in concordance with the data obtained using a pharmacologic approach,²⁵  which, however, has a relatively large range of action.

The precise mechanisms leading to NPM-ALK transformation and the requirements of ALK for tumor growth are still unclear. ALK is known to trigger multiple signaling pathways, which lead to the activation of known antiapoptotic molecules, such as AKT,^22,23  Bcl-xL,^40,49  and survivin.^40,50  Thus, the loss of positive antiapoptotic signals alone may be sufficient to drive cells into apoptosis. It is also conceivable that more complex scenarios are at work. For example, we have recently demonstrated that ALK efficiently activates Ras and its downstream effectors, and through Stat3, up-regulates cyclin D1 and c-myc. Moreover, loss of Stat3 via antisense oligonucleotides or shRNA leads invariably to cell death and growth inhibition in ALK⁺ tumors.²¹ 

Overall, our findings justify the search for specific inhibitors of ALK and/or the development of novel strategies, such as oligonucleotide-based antisense targeting, for the clinical treatment of patients carrying ALK⁺ neoplasms. Interestingly, a moderate down-regulation of ALK protein expression appears to be sufficient to produce measurable biologic effects, even when it is not sufficient to directly induce cell death. While ALK-A5 shRNA sequences delivered using retroviruses resulted only in a moderate reduction of NPM-ALK protein levels, the infected cells displayed a lower rate of proliferation and impaired biologic features. Cell growth disadvantage was further apparent when cells were challenged with suboptimal doses of chemotherapeutic drugs or cultured in stress conditions. Both conditions specifically promote the cell death of ALCL cells infected with ALK-A5 shRNA, but at appropriate concentrations, had very little effect on control cells. These findings suggest that an effective treatment of ALCL can be achieved using a combination approach: avoiding unnecessary toxicity accompanied by moderate- to high-level dosing with the same reagents. The association of low doses of chemotherapeutic drugs in combination with ALK RNAi might also result in a lower incidence of tumoral drug resistance.^51,52  This is an important consideration particularly in view of the fact that RNAi resistance can occur.⁵³  Notably, cells infected with a low copy number of ALK-A5 were not resistant to ALK shRNA, since upon reinfection with a high titer of ALK-A5 lentivirus, they underwent apoptosis (data not shown). Even though our in vitro model might not fully reproduce an in vivo scenario, our findings are encouraging and support further investigation. Finally, the data for intratumoral delivery of a high viral titer of shRNA lentivirus demonstrate the possible therapeutic feasibility of RNAi anti-ALK in the treatment of ALCL. In these experiments we have shown that we could induce massive cell death even though we did not observe, immediately after the treatment, the complete disappearance of the tumor masses. The significance of this approach should be further evaluated by studying whether this method alone could improve the overall survival of the animals or, alternatively, if combined treatment with conventional drugs might be required. Nonetheless, many challenges still lay ahead prior to the clinical use of the RNAi technology in cancer treatment. Despite the fact that intratumoral delivery of shRNA cassettes may be feasible in some instances, the targeting of shRNAs to the proper cells remains a challenge. Various research fronts are working together to overcome the obstacles efficiently, and safely target viral shuttle vectors to limited cellular targets.⁵⁴ 

In conclusion, we have shown that stable shRNA targeting ALK is a new a powerful approach to investigate and dissect the pathogenetic role of ALK in ALCL lymphomagenesis. Based on these findings, the ablation of ALK should be actively pursued as a new and efficient strategy for the clinical treatment of patients with ALK⁺ tumors.

Note added in proof. While this paper was in press, a new study⁵⁵  confirmed that ALK is a molecular target for ALCLs by using specific kinase inhibitors.

We thank Dr L. Naldini (San Raffaele Scientific Institute, Milan, Italy) for reagents. Sequences of the shRNA oligonucleotides are available on request.