The Quassinoid Derivative NBT-272 Targets Both the AKT and ERK Signaling Pathways in Embryonal Tumors

Embryonal tumors (ET) represent an important group of childhood malignancies arising from different tissues of fetal origin (1, 2). Medulloblastoma (MB) is a tumor of the central nervous system, originating from granular progenitor cells of the cerebellum (3) and accounting for about 20% of all childhood brain tumors (4). Neuroblastoma (NB) is the most common extracranial solid cancer in children originating from the sympathetic nervous system (5, 6). Constant progress has been achieved in clinical practice by increasing the overall survival rate of MB and NB patients, although novel therapeutic strategies are still needed, particularly to treat the high-risk patients characterized by a more aggressive disease and poor prognosis. In both MB and NB, high-grade malignancy is often associated with overexpression or amplification of MYC oncogenes (7–11) playing a role in embryogenesis and, if aberrantly expressed, in tumor development as well (12).

There is a long history of therapeutic benefits produced by natural compounds and some of these have formed the backbone for further targeted chemical modifications improving efficacy and selectivity. Quassinoids have been investigated for their antineoplastic properties (13), which has led to the identification of interesting candidates such as bruceantin, with significant inhibitory effects on leukemia, lymphoma, and myeloma cell growth, both in vitro and in vivo (14). Bruceantin was reported to induce apoptosis and cell differentiation, thus justifying efforts to develop quassinoid analogues with improved efficacy.

Among the derivatives of bruceantin, NBT-272 displayed a significantly stronger toxicity than the parent molecule, as tested in different models of hematological and solid tumors (15, 16). Moreover, NBT-272 was reported to induce downregulation of c-MYC in MB-derived cells (16), thus raising the question of the mechanisms underlying NBT-272's cytotoxicity and its selectivity for MYC in ET.

Besides activating programmed cell death, quassinoids were shown to arrest cell growth by impairing protein synthesis mediated by the eukaryotic translation initiation factor 4E (eIF4E; ref. 17). The phosphorylation status of eIF4E plays an important biological role in modulation of normal cell growth in response to extracellular stimuli and during development (18). However, eIF4E was also reported to be an oncogene (19) and its transformation properties were attributed, at least in part, to induction of cell cycle progression through cyclin D1 (CCND1) and MYC (20). Activation of eIF4E itself is mainly regulated by 2 pathways, namely the mitogen-activated protein kinase pathway (21) and the AKT/mammalian target of rapamycin (mTOR) pathway via the 4E binding protein (4EBP-1; ref. 22).

In this study, we have investigated the effect of NBT-272 in different cellular models of ET, both in vitro and in vivo. The compound triggered multiple cellular responses, ranging from cell cycle arrest to induction of autophagy, which are downstream effector mechanisms of the AKT/mTOR and extracellular signal-regulated kinase (ERK) pathways. In turn, interference with these signaling events could also explain the inhibitory effect of NBT-272 on protein synthesis and on expression of c-MYC/MYCN in different ET entities.

The following human cell lines were used: the NB SK-N-BE, SH-SY5Y, and WAC-2 (23; G. M. Brodeur, Children's Hospital of Philadelphia, Philadelphia, PA); retinoblastoma (RB) Y79 and WERI (A. Eggert, University Children's Hospital, Essen, Germany); Ewing's sarcoma family of tumors (ESFT) A673 and TC71 (H. Kovar, St. Anna Children's Cancer Research Institute, Vienna, Austria); and malignant rhabdoid tumor (MRT) LP and MON (O. Delattre, Institut Curie, Paris, France). MB DAOY and D341 were purchased from American Type Culture Collection. Cells that were not purchased from the American Type Culture Collection were authenticated by comparative genomic hybridization (in the laboratories mentioned above or as described below). The clone DAOY M2 (overexpressing c-MYC) was kindly provided by D. Stearns (Johns Hopkins University, Baltimore, MD; ref. 11). NB, ESFT, and MRT cell lines were grown in RPMI medium (Life Technologies/Invitrogen) supplemented with 10% (v/v) fetal calf serum (FCS) and penicillin/streptomycin (PS)/l-glutamine, whereas RB cells were grown in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 15% (v/v) FCS Gold (Invitrogen), 10 μg/mL insulin (Sigma-Aldrich), 50 μmol/L 2-β-mercaptoethanol (Sigma-Aldrich), and PS/l-glutamine. MB cells were cultured in Richter's zinc option medium (Invitrogen) supplemented with 10% (v/v) FCS and PS/l-glutamine.

The ET-derived cell lines used in the study were profiled by array comparative genomic hybridization (aCGH) using a custom 60K Embryonal Tumor array CGH-chip (ET-aCGH chip; Agilent Technologies), as described by Kumps et al. (24).

The 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS)-based CellTiter 96 AQueous One assay (Promega) was used to quantify cell viability. Colony formation in agar was tested by seeding single cells (4,000 per well) in agar-containing medium [0.35% (w/v) final concentration] on top of a 0.5% (w/v) agar layer. Apoptosis was investigated by quantifying the activation of caspases, by using the Caspase-Glo 3/7 Assay (Promega) according to manufacturer's instructions. The In Situ Cell Death Detection Kit (Roche) was also employed to evaluate the proapoptotic response of ET cells. The terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) reaction was done in the presence of fluorescein-labeled nucleotides and followed by detection under a fluorescent microscope (Axioskop 2; Zeiss).

Cells [(5–10) × 10⁵] were treated for the indicated time intervals with different doses of NBT-272 or were mock treated (control). After washing with PBS, the cells were washed and harvested in a 2 mmol/L EDTA/PBS solution. Cell pellets were fixed with ice-cold 70% ethanol on ice for 40 to 60 minutes and, after centrifugation at 2,500 g for 5 minutes, resuspended with 50 μg/mL propidium iodide/PBS containing 100 U/mL RNAase (Sigma-Aldrich). The DNA content was estimated by using a BD FACS Canto II flow cytometry system for acquisition and the FlowJo software for the data analysis.

Total RNA from DAOY, D341, SK-N-BE, and SH-SY-5Y cells, treated with 20 nmol/L NBT-272 for 24 hours or left untreated, was subjected to reverse transcriptase (RT; RT² First Strand Kit; SABiosciences). The resulting cDNA was used as a template for RT² Profiler PCR Array (SABiosciences) for pathway-focused gene expression analysis (using sets of arrayed TaqMan probes for cell cycle and apoptosis genes). Gene expression was expressed as log₂ ratio (where ratio is 2^−ΔΔCt), relative to normalization to untreated control samples and to the average expression of five housekeeping genes (B2M, HPRT1, RPL13A, GAPDH, and ACTB). Only statistically significant gene expression changes over 3 independent experiments were considered.

Total cell lysates were prepared in a radioimmunoprecipitation assay buffer buffer [50 mmol/L Tris-HCl (pH 7.5), 0.1% sodium laurilsulfate, 0.1% sodium deoxycholate, 150 mmol/L NaCl, 1% NP40, 1 mmol/L EDTA, and 1 mmol/L EGTA], supplemented with protease inhibitors (Roche) and phosphatase inhibitors (10 mmol/L sodium fluoride, 10 mmol/L sodium orthovanadate, and 10 mmol/L sodium β-glycerophosphate). The amount of protein was determined by BCA Protein Assay (Pierce). Cell lysates in SDS-sample buffer were boiled for 8 minutes and equal protein amounts were resolved by 4% to 12% gradient SDS-PAGE and immunoblotting using polyvinylidene difluoride membranes. The antibodies reacting against the following proteins were employed: CCND1, caspase-3, AKT, mTOR, phospho-ERK 1/2(Tyr202/204; Santa Cruz Biotechnology); ERK, phospho-AKT (Ser473), phospho-MEK1/2 (Ser217/221), phospho-RB (Ser795), RB, phospho-S6 (Ser235/236), S6, phospho-mTOR (Ser2248), phospho-MNK1 (Thr197/202), MNK1, phospho-eIF4E (Ser209), eIF4E, 4EBP-1, phospho-4EBP-1 (Thr37/46), cyclin E, CDK2, CDK4, PARP, LC-3, c-MYC, MYCN, and ATG5 (Cell Signaling); β-tubulin and β-actin antibodies were purchased from Sigma-Aldrich, whereas the anti-SQSTM1/p62 was purchased from BIOMOL International. To inhibit proteasome-dependent protein degradation, cells were treated with 50 μmol/L MG-132 (Sigma-Aldrich). The PI3K/mTOR inhibitor BEZ-235 was purchased by Axon Medchem.

After treatment (with NBT-272, rapamycin, or vehicle) for different time intervals, DAOY and SK-N-BE cells were subjected to fixation at room temperature for 30 minutes with 2% glutaraldehyde and 0.8% formaldehyde in 50 mmol/L sodium cacodylate (pH 7.3), followed by 30-minute exposure to 1% OsO₄ and 1.5% K₄Fe(CN)₆ in 50 mmol/L sodium cacodylate (pH 7.3). Fixed samples were contrasted using 2% uranyl acetate in water for 2 hours, progressively dehydrated in increasing concentrations of ethanol, and embedded into epon (Catalys). Ultrathin sections of 50 nm were prepared and contrasted with uranyl acetate and lead citrate. Samples were analyzed by using a CM 100 transmission electron microscope (TEM; Philips).

Human NB SH-SY5Y cells were stably transfected with a pLentiV5-Luciferase-expressing vector (Invitrogen) as described elsewhere (25). SH-SY5Y-LUC clones were grown under standard conditions in DMEM containing 10% (v/v) FCS, l-glutamine, and blasticidine for selection.

Eight-week-old female athymic nude mice from Harlan Laboratories were inoculated subcutaneously in the right rear flank with 2 × 10⁶ SHSY5Y-LUC cells, injected in 100 μL PBS solution. Tumor cell growth was monitored weekly by measuring luminescence emission, referred to as bioluminescence imaging (BLI) levels, using the IVIS 3D Illumina Imaging System (Xenogen Corp.). A detailed procedure can be obtained from Caliper. The BLI/tumor size ratio was quantified by encircling the luciferase-emitting areas using the Living Images Software 3.2 (Xenogen). Three weeks after implantation, the mice were divided into 2 homogeneous groups based on their BLI level (photon/s/cm²). Each group received either 10 μmol/L per week of NBT-272 or vehicle (EtOH 0.003%/PBS), administered as 100 μL solution each time, 3 times a week. The analysis of the tumor growth was done 3, 4, and 6 weeks after the first NBT-272 administration. The dose was adjusted by taking into account preliminary data provided by NaPro BioTherapeutics, Inc. (maximum tolerated dose = 0.9 mg/kg/wk; ref. 15) and the in vitro results. The treatment was done for 6 weeks before the animals were sacrificed, and tumor samples were analyzed by immunohistochemistry and Western blotting. Quantitative data analysis of the tumor size was done by ANOVA statistical tests, using the Statview program.

All animal experiments were conducted according to national and international guidelines and following approval by the Institutional Animal Care and Ethical Committee of CEINGE-University of Naples Federico II (Protocol 29, September 30, 2009) and of the Italian Ministry of Health (Dipartimento Sanità Pubblica Veterinaria D.L. 116/92).

Paraffin-embedded tumor samples were treated with xylene, heated for 45 minutes (in citrate buffer, pH 6.0), and finally treated for 15 minutes with 1% H₂O₂ (Carlo Erba Reagents). Tissues for immunohistochemistry were blocked with 3% bovine serum albumin, 5% total goat serum, and 0.3% Tween 20/PBS (1 hour, room temperature). Primary antibodies reactive against human Ki67 (M7240; Dako, Glostrup, Denmark) and human cleaved caspase-3 (9661; Cell Signaling) were used (1 hour, room temperature, diluted 1:100 in blocking solution). After incubation with a biotin-conjugated secondary antibody (1 hour, room temperature), the detection was done in the presence of 3,3′-diaminobenzidin (Dako) according to the manufacturers' protocol. The sections were counterstained by hematoxylin/eosin (Bio Optica), washing 3 times with 0.2% Triton X-100/PBS after every incubation steps. As negative control, blocking solution was used instead of the primary antibody.

NBT-272 (NaPro BioTherapeutics, Inc.; Fig. 1A) is a semisynthetic analogue of bruceantin (14, 17), with 10-fold higher toxicity than the parent compound in vitro (15). Previously published data showed the ability of NBT-272 to drastically reduce the protein expression level of the oncogene c-MYC in MB-derived cell lines (16), although the exact mechanism of action remained unexplored. In this study, the biological responses induced by NBT-272 were investigated in a panel of 13 cell lines, representative of 6 different ET entities (Fig. 1). Expression of c-MYC and MYCN was evaluated in each cell line both at protein (Supplementary Fig. S1A) and mRNA levels (Supplementary Fig. S1B); these varied significantly, reflecting the gene expression ranging from single copy to gene amplification.

The ET cell lines were profiled for copy number changes (gain/loss) using aCGH. In Supplementary Fig. S2, the aCGH data of 8 cell lines are reported. Some cell lines displayed very few genomic aberrations (1q gain in the MRT cells LP, and 1q gain and 8q gain in MB D341), whereas others have very complex profiles, such as WERI cells with a 1q gain region that appeared very discontinuous.

The effect of NBT-272 was first evaluated in terms of cell viability, clearly showing a very potent effect of the compound in all selected ET-derived cell lines (Fig. 1B; Table 1). The IC₅₀ values, calculated from the dose–response curves, varied in the nmol/L range of concentration, that is, about 2 orders of magnitude lower compared with many small-molecule inhibitors investigated as anticancer drugs. Additionally, 5 representative ET cell lines (WAC-2, SK-N-BE, A673, MON, Wit-49) were tested for anchorage-independent growth in soft agar in the presence of NBT-272 (Fig. 1C). The compound reduced colony formation by 50% (compared with mock-treated controls) at doses that were similar or lower than the IC₅₀ obtained in cell viability assays (Table 1).

To identify which molecular characteristics were responsible for NBT-272's toxicity, we sought to investigate the diverse biological responses triggered in ET cell lines by the compound. The effect on cellular distribution at different stages of the cell cycle was evaluated by DNA staining in MB and NB cells. In dose–response experiments at different time points, 5 nmol/L NBT-272 was already sufficient to almost halve the population of replicating DAOY cells after 24 hours, with a concomitant increase in G₁ population (Supplementary Fig. S3). This observation was found to extend to NB cell lines (WAC-2, SK-N-BE, and SH-SY5Y; Fig. 2A). Here, NBT-272 induced a similar effect after 16-hour treatment, thus suggesting that blocking the cell cycle progression at the G₁ to S transition occurs as one of the earliest mechanisms leading to proliferative arrest.

Consistent with the data obtained by cytofluorimetry, NBT-272 induced loss of CCND1 and of its interacting partner the cyclin-dependent kinase (CDK)4 (Fig. 2B). These proteins are both important regulators of the cell cycle progression through the G₁ phase (26), along with the phosphorylation status of the retinoblastoma protein, which changes throughout the cell cycle with highest levels when cells enter the S phase (27). Consistent with an arrest in the G₁/S transition, RB was hypophosphorylated on treatment with NBT-272 (Fig. 2B). On the contrary, the effect on CCNE1/CDK2 was transient.

Two NB cell lines (SH-SY5Y and SK-N-BE) and the MB cells DAOY and D341 were treated for 24 hours with 20 nmol/L NBT-272, followed by expression profiling of 84 cell cycle–related genes by qRT-PCR. Supplementary Fig. S4A illustrates the statistically significant gene expression changes induced by NBT-272. The 23 transcripts fell into 3 functional categories: (1) G₁/S transition of the cell cycle; (2) DNA damage (e.g., growth arrest and DNA damage-inducible alpha protein, GADD45A); and (3) regulation of transcription elongation. The most relevant changes concerned genes involved in the G₁ arrest, in support of the data from our biological assays (Fig. 2). For instance, consistent with a G₁/S block of the cell cycle, the mRNA level of CDK2 decreased, whereas the inhibitor molecules CDKN1A/p21^Cip1 and CDKN1B/p27^Kip were induced on treatment (Supplementary Fig. S4A). Whereas CCND1 and CCNE1 were regulated solely at the protein level (Fig. 2B), other cyclins and CDK molecules involved in the RNA polymerase II–mediated gene transcription (CCNC/CDK8, CCNH/CDK7, and CCNT/CDK9) were found to be upregulated.

Induction of apoptosis was assessed by quantifying the activation of caspase-3/7 induced by NBT-272 (Fig. 3A). The proapoptotic response to NBT-272 varied significantly in a cell type– and tumor-dependent way. This observation was also confirmed by Cell Death ELISA assays (Roche; data not shown). Whereas NBT-272 concentrations as small as 10 nmol/L triggered apoptosis in MB-, ESFT-, and MRT-derived cell lines after 24 hours, the same doses had a very modest proapoptotic effect in other cellular models, particularly in NB (Fig. 3A). The limited proapoptotic effect of NBT-272 in MB and NB cells was also shown by immunoblotting (Fig. 3B), where caspase-3 could be only detected as full length protein and PARP cleavage was only induced in MB cell lines (e.g., in D341; Fig. 3B). This observation was confirmed by TUNEL assay, where NBT-272 was able to induce a breakage of the genomic DNA in DAOY cells (Supplementary Fig. S5).

Using a similar approach to that employed for cell cycle–related genes, we investigated the transcriptional changes related to apoptosis triggered by NBT-272 (Supplementary Fig. S4B). Among the differentially expressed genes, we found key components of both the extrinsic and the intrinsic pathways of apoptosis, mainly upregulated on treatment with the compound. This was the case with proapoptotic members of the Bcl-2 family (BIK, BAD, and BAK) and with proteins bearing the caspase recruitment domain (CARD), such as Bcl-10, PYCARD and RIPK2, implicated in the recruitment and binding of caspases. Another major group, namely genes encoding for ligands and receptors of the tumor necrosis factor (TNF) superfamily, was induced (Supplementary Fig. S4B). It included TNF-α, TNF-β/LTA, TNFSF10/TRAIL, and their corresponding receptors TNFRSF10A and TNFRSF10B. Also the expression of TNFRSF5, TNFRSF9, and TNFRSF6/FAS increased on NBT-272 treatment, altogether indicating enhanced cell responsiveness to death receptor–mediated signaling.

Because NBT-272 induced apoptosis with differing efficiency in the ET cell lines tested, we sought additional mechanisms that could lead to cell death. Formation of the distinctive traits of autophagy was investigated at the ultrastructure cellular level by TEM, thus revealing induction of a high number of double-membrane structures containing cellular material (i.e., cytosolic cargo molecules and organelles), described as autophagosomes (28).

In Fig. 3C, TEM images of DAOY (MB) and SK-N-BE (NB) cells are reproduced. Compared with mock-treated controls (Fig. 3C, a and b), treated cells (20 nmol/L NBT-272, 16 hours) displayed all the characteristic signs of autophagy induction, in particular, formation of large autophagosomes (arrow heads in Fig. 3C, e, f, and g) and increased lysosomes and vacuoles. Moreover, the cytosol appeared looser than in control cells, the nuclear envelope became irregular, the mitochondria acquired swollen crests indicative of dysfunctional activity, and myelin figures could be visualized as electron-dense structures. As positive control, DAOY cells were treated separately with rapamycin (200 nmol/L, 48 hours; Fig. 3C, h), inducing autophagy as an event associated with inhibition of mTOR (29). On comparing the effect of the 2 compounds in DAOY cells, NBT-272 proved to be a much stronger inducer of autophagy.

Formation of autophagosomes requires a rather complex rearrangement of Atg proteins, including the protein conjugation system generating the complex Atg5-Atg12 (30), which was detected by immunoblotting, on treatment of DAOY cells with 20 nmol/L NBT-272 (Fig. 3D). Concomitantly, we were able to show the degradation of sequestosome 1 (SQSTM1/p62), that is, a multifunctional adaptor molecule driving the degradation of polyubiquitinated proteins such as those tagged for autophagic clearance (31). Further, we documented the conversion of the microtubule-associated protein light chain 3 (LC3-I) into its membrane-interacting counterpart (LC3-II), which is a well-documented marker of autophagy induction.

Because NBT-272 was originally described as a MYC inhibitor, we evaluated the correlation of the expression of c-MYC and MYCN in the ET cell lines (Supplementary Fig. S1) with cell sensitivity to the compound. However, only the MB cell line D341 (bearing c-MYC amplification) was significantly more sensitive to NBT-272 than DAOY cells expressing c-MYC as single copy gene (Fig. 1; ref. 16), whereas cells from other ET entities displayed a sensitivity similar to NBT-272 (Fig. 1).

Moreover, again regardless of the MYC isoform expressed and the gene copy number of c-MYC/MYCN, NBT-272 induced depletion of MYC in several ET cell lines, although exclusively at the protein level (Fig. 4A). In contrast, mRNA expression was either left almost unchanged (e.g., MYCN in SK-N-BE cells) or was enhanced (e.g., c-MYC in D341 cells) by NBT-272 in a dose-dependent manner (Fig. 4A).

The MYC family of transcription factors is known to have short half-lives (32). Therefore, we researched whether NBT-272 could have an effect on protein stability, thus helping to explain MYC depletion. In fact, by using the proteasome inhibitor MG132 (33) in combination with NBT-272, c-MYC was already almost completely rescued after 4 hours (Fig. 4B).

Additional lines of evidence supported an indirect effect of NBT-272 on MYC. Two small-molecule inhibitors specifically targeting the MYC/Max interaction (34, 35) did not show the same profile of activity as NBT-272 (Fig. 4C). In fact, significant cell viability could still be measured at high doses of both Mycro3 and 10058-F4. In particular, WAC-2 cells were almost unaffected by either of the compounds, whereas 10³ times less concentrated NBT-272 completely arrested cellular growth.

Finally, only a limited number of gene expression changes induced by NBT-272 matched with the transcriptional effects of siRNA-mediated silencing of c-MYC in DAOY cells (Supplementary Fig. S6; Fig. 4D). In fact, only a restricted number of genes were up-/downregulated in both sets of experiments, that is, 11 apoptosis-related and 11 cell cycle–related transcripts (out of 84 in each analysis). Among them, even fewer genes were known direct targets of c-MYC (i.e., GADD45A, BIRC3, DFFA, CDK6, CDKN1A/p21, and PCNA). On the other hand, genes reported to be MYC targets exhibited a different expression profile, that is, they were mostly downregulated in c-MYC–depleted cells, but induced or left unchanged in cells treated with NBT-272 (e.g., the cyclins CCNB1, CCNB2, CCND2, CCNH, and the DNA replication regulators MCM4 and MCM5).

Altogether, these results led us to conclude that a direct inhibition of MYC is unlikely to explain the effects triggered by NBT-272 in ET cells, which are more likely to be the result of more complex mechanisms and of interference with upstream pathways.

To improve our understanding of the regulatory mechanisms leading to arrest of the cell cycle, apoptosis, and autophagy, we analyzed the activation status of 2 signaling pathways, that is, the AKT/mTOR and the MEK/ERK pathways, regulating these cellular responses (36). Furthermore, both pathways have been documented to be relevant also in the development of ET.

Figure 5A shows the effect of NBT-272 on the activation status of AKT varying in different cell lines, with phosphorylation at Ser473 significantly decreasing in WAC-2, SK-N-BE, and SH-SY5Y cell lines, but slightly enhanced in DAOY cells. On the other hand, the level of phosphorylated ERK 1/2 (p44/p42) diminished in all cell lines tested, although with different kinetics (Fig. 5A), suggesting a different sensitivity of the 2 pathways to NBT-272. To test whether NBT-272 had a direct effect on the enzymatic activity of either AKT or ERK and could thus be regarded as a kinase inhibitor, we quantified the in vitro kinase activity of 271 purified recombinant kinases (Supplementary Fig. S7A). However, none of them were significantly impaired by NBT-272 (20 nmol/L). Additionally, different subunits and isoforms of phosphoinositide-3 kinase (PI3K) were tested (Supplementary Fig. S7B), again showing no effect from NBT-272 (20 nmol/L).

Although the exact mechanism of action of NBT-272 remained unsolved, the compound clearly affected AKT- and ERK-dependent cellular functions. The downstream target of ERK, MNK1 (37), was also deactivated on treatment with NBT-272, whereas the upstream kinase MEK1/2 was only slightly affected (Fig. 5A). MNK1 is known to phosphorylate both eIF4E and its repressor, 4EBP-1 (38). 4EBP-1 is also a direct target of mTOR, playing a key role as regulator of protein translation downstream of AKT (22). Interestingly, regardless of the AKT activation status, phosphorylation of both eIF4E and 4EBP-1 was impaired by NBT-272 in MB and NB cells (Fig. 5B), thus preventing the cap-dependent protein translation machinery from being completely functional. Additionally, the phosphorylation status of the ribosomal S6 protein, target of mTOR via S6 kinase (S6K; ref. 39), was proven to be unaffected, thus reinforcing the idea of a 4EBP-1–dependent block of protein synthesis.

Moreover, we asked whether the cellular sensitivity to NBT-272-mediated G₁/S block of the cell cycle, apoptosis, or autophagy would vary if 1 of the pathways had been affected by using specific inhibitors. To test the relevance of the mTOR/AKT pathway, BEZ235 was used (40). A pretreatment with the PI3K/mTOR inhibitor increased both depletion of CCND1 and degradation of SQSTM1/p62 when compared with NBT-272 alone (Supplementary Fig. S8A), whereas no significant enhancement of the proapoptotic functions of NBT-272 was observed (Supplementary Fig. S8B). Although BEZ235 alone had apparently no effect, under these conditions, on the expression of the cell cycle and autophagy markers in DAOY cells, the inhibitor was functional, as shown by the strong deactivation of AKT (Supplementary Fig. S8A) and the induction of apoptosis (Supplementary Fig. S8B).

NB cells bearing a transgene encoding luciferase (SH-SY5Y-LUC) were injected in the flank of 10 athymic nude mice and were allowed to develop for 3 weeks. The effect of the intraperitoneal administration of NBT-272 (10 μmol/L per week) and of the vehicle (0.003% EtOH/PBS) on the tumor growth was monitored by quantifying the luminescent signal from engrafted SH-SY5Y-LUC cells. After 6 weeks of treatment, the tumors almost stopped growing in the mice receiving NBT-272, whereas the human NB SH-SY5Y cells expanded into large tumors in the control animals (Fig. 6A and B; Supplementary Fig. S9). In fact, a 6.3-fold increase of tumor size could be detected in the control group compared with the NBT-272–treated group (P = 0.037 according to Student's t test). The result clearly indicated that NBT-272 has a significant antiproliferative effect on the tumor development in vivo. The hemotoxylin/eosin staining of some representative tumor samples showed a compact and rounded shape of the tumors, which were characterized by a very dense cell distribution with some necrotic areas (Fig. 6C, a–d). Tumors from NBT-272-treated mice showed a clear reduction of Ki67-positive cells (Fig. 6C, f and h), whereas a prominent cellular perinuclear staining in vehicle-treated mice was observed (Fig. 6C, e and g), thus confirming the significant antiproliferative effect of NBT-272 also in vivo. Activation of caspase-3 was also evaluated (Fig. 6C, i and j), which shows that the treatment with NBT-272 did not induce a significantly higher induction of apoptosis than with control tumors. This result confirmed our findings in vitro (Fig. 3A and B) by referring to the same cellular model, that is, SH-SY5Y cells, which were used to induce the tumors. Finally, NBT-272 treatment had an effect on MYC stability/expression in the tumors, in which ERK activation was also impaired (Fig. 6D).

A fine regulation of protein turnover by a tunable modulation of protein stability requires specific post-translation modifications (such as phosphorylation) of key molecules. Many of these molecules are significant components of prosurvival pathways and often play a crucial role in tumor development and progression, in conditions of constitutive and growth factor–independent activation. In past decades, great efforts have been made to interfere with those signaling pathways by employing different strategies of targeted therapy, ranging from antibodies against receptor tyrosine kinases (RTK) to small-molecule inhibitors blocking downstream signaling effectors. Downstream of RTK activation, the induction of MYC transcription factors and of their pleiotropic effects on cell growth has been frequently associated with neoplastic transformation. Therefore, MYC itself clearly represents an attractive target for anticancer therapy (41), also due of its correlation with poor prognosis and high-grade malignancy in many types of tumors, including pediatric malignancies (10, 11, 42, 43).

In the context of MYC-overexpressing ET tumors, we wanted to investigate the biological responses of a group of ET-derived cell lines to NBT-272, a small molecule of about 550 Da, previously described as inhibitor of MYC (15, 16). Its toxicity and ability to prevent colony formation in vitro indicated a very strong effect of NBT-272 at low concentrations.

In the search for an association between toxicity of NBT-272 and MYC expression levels in different ET cell lines, no significant correlation could be found. This observation led us to question any direct effect of NBT-272 on the oncogene, which would more likely be the result of different mechanisms, such as interference with protein synthesis and stability, affecting also the MYC family members known to have short half-lives (32). Indeed, we could observe that inhibition of the proteasome-dependent protein degradation pathway restored the depleting effect of NBT-272 and almost completely stabilized c-MYC. Moreover, the effect induced by 2 specific MYC/Max inhibitors (34, 35) and by c-MYC silencing in ET cells did not match the cellular response to NBT-272, again detrimental to the hypothesis of a direct effect on MYC. This conclusion, however, certainly does not exclude that MYC inhibition contributed to the panel of cellular responses observed on NBT-272 treatment.

Even though the compound is not a direct kinase inhibitor, our study indicated the ability of NBT-272 to interfere with the activation status of different components of the AKT/mTOR and MEK/ERK signaling pathways. We therefore attributed the effect of NBT-272 on apoptosis, cell cycle, and autophagy to its ability to interfere with these pathways. Indeed, the basal level phosphorylation of ERK1/2, MNK1, and AKT in MB and NB cells was impaired by NBT-272, under the same conditions that induced efficient cell cycle arrest (G₁/S) and autophagosome formation. It is also relevant to emphasize that the 2 pathways converge to regulate the activation status of eIF4E and 4EBP-1, both dephosphorylated (i.e., inactive) in the presence of NBT-272, thus leading us to speculate that targeting these pathways would prevent full activation of the cap-dependent protein translation machinery. Furthermore, MYC depletion could possibly be explained, not only as an effect of protein synthesis inhibition, but also as a consequence of decreased phosphorylation at its Ser62 residue, which has a stabilizing effect and is catalyzed by ERK (44).

In total, our observations strongly suggest that both the AKT and MEK/ERK signaling pathways participate in orchestrating the cellular responses to NBT-272, at least in vitro. Moreover, combination of NBT-272 with a specific inhibitor of PI3K/mTOR, BEZ235 (40), seemed to further increase the NBT-272–dependent effect on cell cycle and autophagy, but not apoptosis. Numerous components of the AKT/mTOR and MEK/ERK pathways have been found deregulated in cancer, including ET. High basal activation levels of AKT and of the S6 ribosomal protein were documented in MB (45, 46) and a relevant role in NB pathogenesis was attributed to PI3K, found to be overexpressed in NB tumors (47). Phosphorylation of AKT, S6, and ERK has also been reported as a frequent event in NB, in which AKT activation in particular was correlated with diverse indicators of aggressive disease and poor prognosis (48). These 2 signaling pathways were also negatively associated with malignancy in childhood rhabdomyosarcoma (49) and in Ewing's sarcoma (50), thus rendering it certainly worthwhile to promote the idea of developing common therapeutic strategies effective in different ET entities.

Finally, NBT-272 arrested tumor growth in a xenograft model of human NB, showing that the potent effect of this compound could also be reproduced in vivo, with concomitant reduction of MYC expression and of ERK activation in the treated tumors. These very encouraging results warrant the extension of investigations to more ET models.

In summary, NBT-272 induced a relatively complex pattern of cellular responses, including inhibition of crucial regulators of protein synthesis. All these events could be related functionally to interference with key cell survival pathways (AKT and MEK/ERK) playing a role in the pathogenesis of several ETs. Our findings increase the present level of understanding of the mechanisms of action of NBT-272, explaining its potent antitumor effect in in vitro and in vivo models of ET. This study certainly justifies further efforts to define more clearly the potential benefits of using NBT-272 in novel therapeutic strategies for pediatric tumors.

No potential conflicts of interest were disclosed.

We thank Dr. Brian Carter for his precious contribution in proofreading the manuscript and Dr. Beat Bornhauser and Prof. Shida Yousefi for helpful discussion. We are grateful to Lawrence Helson and James McChesney for providing NBT-272, developed by Tapestry Pharmaceuticals (Boulder, CO). We also thank Thererse Bruggmann and Gery Barmettler for their excellent technical support during the preparation and analysis of samples for electron microscopy.

Deborah Castelletti and Giulio Fiaschetti were supported by the European Community FP6, project STREP (EET-pipeline, number: 037260). Valeria Di Dato and Daniela De Martino were supported by FP6-EET pipeline LSH-CT-2006-037260 (M. Zollo), FP7-Tumic HEALTH-F2-2008-201662 (M. Zollo), Associazione italiana contro la lotta al Neuroblastoma “Progetto Pensiero” (M. Zollo), and AIRC Tumori Pediatrici 2007-2010 (M. Zollo).