Targeting Aberrant PI3K/Akt Activation by PI103 Restores Sensitivity to TRAIL-Induced Apoptosis in Neuroblastoma

Neuroblastoma represents the most common extracranial solid tumor outside the central nervous system in childhood (1, 2). The prognosis of children older than 1 year with stage 4 disease is still very poor with long-term survival rates of only 40% despite intensive and multimodal therapy (3). Therefore, novel strategies are required for the treatment of advanced stage neuroblastoma.

Programmed cell death or apoptosis is essential to maintain tissue homeostasis (4). In addition, induction of apoptosis in malignant cells presents a key principle of cancer therapy (5). There are 2 major pathways that initiate apoptosis and eventually lead to activation of caspases: one is initiated by death receptors on the cell surface (death receptor or extrinsic pathway), whereas the other is mediated by the mitochondria (mitochondrial or intrinsic pathway; ref. 5). The death receptor pathway is typically stimulated by ligation of death receptors of the TNF receptor superfamily by their cognate ligands or agonistic antibodies resulting in the formation of the death-inducing signaling complex (DISC; ref. 6). This in turn leads to caspase-8 activation, which can directly cleave downstream caspases (6). Stimulation of the mitochondrial apoptosis pathway culminates in the permeabilization of the outer mitochondrial membrane, the release of cytochrome c into the cytosol, and caspase-3 activation via the apoptosome complex (7). Members of the Bcl-2 family proteins are critical regulators of apoptosis by controlling outer mitochondrial membrane permeabilization (8). Bcl-2 family proteins can be divided into antiapoptotic proteins, such as Bcl-2 and Mcl-1 or proapoptotic proteins like Bax, Bak, and BH3 domain-only molecules (e.g., Bid, Bim, and Noxa; ref. 8). The ratio of pro- versus antiapoptotic Bcl-2 proteins plays a critical role in determining whether a cell will die or survive (8).

TRAIL is considered as a promising candidate for cancer therapy, because it directly engages the death receptor pathway of apoptosis (6). Clinical trials show that TRAIL or agonistic TRAIL receptor antibodies can be safely administered alone and in combination with chemotherapeutics (6, 9). However, many tumors present with primary or acquired resistance to death receptor–induced apoptosis, which may be due to aberrant activation of survival pathways (10).

The phosphoinositide 3′-kinase (PI3K)/Akt/mTOR pathway plays an important role in mediating prosurvival signals (11). Among its various functions, Akt inhibits apoptosis either directly by phosphorylating apoptosis‐signaling molecules or indirectly by modulating the activity of transcription factors (12). In neuroblastoma, we recently identified PI3K/Akt as a new predictor of poor outcome, indicating that this pathway represents a clinically relevant target for therapeutic intervention (13). We therefore investigated in this study whether pharmacological inhibition of PI3K can be exploited as a novel strategy to restore the sensitivity of neuroblastoma toward TRAIL-induced apoptosis.

Neuroblastoma cell lines were obtained from the American Type Culture Collection and maintained in Dulbecco's modified Eagle's medium (SH-EP), RPMI 1640 (LAN-5), or MEM (CHP-212) medium (Life Technologies, Inc.), supplemented with 10% fetal calf serum (Biochrom), 1 mmol/L glutamine (Biochrom), 1% penicillin/streptomycin (Biochrom), and 25 mmol/L HEPES (Biochrom) as described previously (14). N-benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (zVAD.fmk) was purchased from Bachem, PI103 from Alexis Biochemicals, LY294002 from Calbiochem, and TRAIL was obtained from R&D Systems. RAD001 (everolimus) was kindly provided by Novartis Institute for BioMedical Research (Oncology Basel, Novartis Pharma AG, Basel, Switzerland). All other chemicals were purchased by Sigma, unless otherwise indicated.

Apoptosis was determined by fluorescence-activated cell sorting (FACScan; BD Biosciences) analysis of DNA fragmentation of propidium iodide–stained nuclei as described previously (14). The percentage of specific apoptosis was calculated as follows: 100 × [experimental apoptosis (%) − spontaneous apoptosis (%)]/[100% − spontaneous apoptosis (%)].

To determine colony formation, 200 cells were seeded in a 6-well tissue culture plate and allowed to settle overnight. Cells were treated with TRAIL for 24 hours with or without 0.6 μmol/L PI103. Next day, medium was exchanged and colonies were stained after 10 days with 0.75% crystal violet, 50% ethanol, 0.25% NaCl, and 1.57% formaldehyde. Colonies were counted and the percentage of surviving colonies relative to solvent-treated controls was calculated.

For transient gene knockdown cells were seeded, allowed to settle overnight and transfected with 60 pmol of PIK3CA Stealth RNAi (PIK3CAHSS10800 4-6; Invitrogene), PIK3CB Stealth RNAi (PIK3CBHSS10800 7-9; Invitrogene), mTOR Stealth RNAi (FRAP1HSS10382 5-7), Stealth RNAi against Bax (BAXHSS14135 4-6; Invitrogene), Bak (NM001188-875;473;613; Invitrogene), Bid (BIDHSS14147 6-8; Invitrogene), or Stealth RNAi nontargeting control (12935; Invitrogene) by using TransMessenger transfection (301525; Qiagen).

For stable gene knockdown, short-hairpin RNA (shRNA)-targeting Noxa (5′-GATCCCCGTAATTATTGACACATTTCTTCAAGAGAGAAATGTGTCAATAATTACTTTTTGGAAA-3′ (15) or shRNA-targeting Mcl-1 (GGCAGTCGCTGGAGATTAT, GATTGTGACTCTCATTTCT) and a sequence with no corresponding part in the human genome (GATCATGTAGATACGCTCA) that was used as control were cloned into pRETRO-SUPER as previously described (16). Stable clones were generated by selection with 2 μg/mL puromycin (Clontech). For Bcl-2 overexpression, cells were transduced with murine stem-cell virus (pMSCV) vector containing mouse Bcl-2 or empty vector using the packaging cell line PT67 (BD Biosciences). Stable cell lines were selected by 10 μg/mL blasticidin (Invitrogen).

Western blot analysis was carried out as described previously (14) by using the following antibodies: rabbit anti–phospho-Akt (Ser473), rabbit anti–phospho-S6 ribosomal protein (Ser235/236), mouse anti-S6 ribosomal protein, rabbit anti-mTOR, rabbit anti–PI3K-p110α (1:1,000; Cell Signaling), mouse anti-Akt (1:500), mouse anti-XIAP (1:1,000; clone 28) and rabbit anti–Bcl-X_L (1:1,000; BD Transduction Laboratories), rabbit anti-survivin (1:1,000; R&D Systems), mouse anti–caspase-8 (1:1,000; Alexis Biochemicals), mouse anti-cFLIP (1:500; Alexis Biochemicals), rabbit anti–caspase-3, rabbit anti-Bid and rabbit anti-Bim (1:1,000; Cell Signaling), rabbit anti–caspase-9, mouse anti–Bcl-2, rabbit anti-Bak, mouse anti–cytochrome c (1:1,000; BD Pharmingen), mouse anti-Noxa (1:1,000; Alexis), rabbit anti–Mcl-1 (Stressgen), rabbit anti-BaxNT (1:5,000; Upstate Biotechnology), mouse anti-OxPhos Complex IV (1:2,000; Invitrogen), or rabbit anti–PI3K-p110β (1:500; Abcam). Mouse anti–β-actin (1:10,000; Sigma), mouse anti-GAPDH (1:5,000; HyTest), or mouse anti–α-tubulin (1:3,000; Calbiochem) were used as loading controls followed by goat anti-mouse immunoglobulin G (IgG) or goat anti-rabbit IgG conjugated to horseradish peroxidase (1:5,000; Santa Cruz Biotechnology). Enhanced chemiluminescence was used for detection (Amersham Bioscience). All Western blots shown are representative of at least 2 independent experiments. Densitometry analysis was done by ImageJ.

To determine mitochondrial transmembrane potential cells were incubated with CMXRos (1 μmol/L; Molecular Probes) for 30 minutes at 37°C and immediately analyzed by flow cytometry.

Cytochrome c release was assessed by flow cytometry as previously described (17). To analyze cytochrome c and Bid in cytosolic and mitochondrial extracts by Western blot cells were harvested and washed with PBS. Cells were suspended in lysis buffer [2 mmol/L NaH₂PO₄, 16 mmol/L Na₂HPO₄, 150 mmol/L NaCl, 500 mmol/L sucrose, 1 mmol/L Dithiothreitol (DTT), Protease Inhibitor (Roche), and 0.5 mg/mL digitonin] for 3 minutes on ice. Unbroken cells, mitochondria, and nuclei were removed by centrifugation at 14,000 rpm for 1 minute at 4°C. The supernatant was collected as cytosolic fraction and the pellet was resuspended in lysis buffer [30 mmol/L Tris HCL, 150 mmol/L NaCl, 1% Triton X, 10% glycerol (Invitrogen), Protease Inhibitor (Roche), 2 mmol/L DTT, 500 μmol/L phenylmethylsulfonylfluoride] for 2 hours at 4°C, and centrifuged at 14,000 rpm for 20 minutes at 4°C. The supernatant was collected as mitochondrial fraction. Protein expression of cytochrome c or Bid were analyzed by Western blotting. OxPhos Complex IV, α-tubulin, and β-actin were used to control the purity and loading of the mitochondrial and cytosolic fractions.

Bax and Bak activation was determined by immunoprecipitation as previously described (18). Briefly, cells were lysed in CHAPS lysis buffer (10 mmol/L HEPES, pH 7.4; 150 mmol/L NaCl; 1% CHAPS). Protein (1 mg) was incubated with 2 μg mouse anti-Bax antibody (6A7; Sigma) or anti-Bak antibody (AB-1; Calbiochem) overnight at 4°C followed by addition of 10 μL Dynabeads Pan Mouse IgG (Dako), incubated for 2 hours at 4°C, washed with CHAPS lysis buffer, and were analyzed by Western blotting by using rabbit anti-BaxNT or rabbit anti-Bak (BD Pharmingen) antibody.

Chorioallantoic membrane (CAM) assay was done as described previously (17). Briefly, 2 × 10⁶ cells were resuspended in 10 μL serum-free medium and 10 μL Matrigel matrix (BD Biosciences) and implanted on the CAM of fertilized chicken eggs on day 8 of incubation, allowed for 48 hours to form tumors, and treated with 1 μmol/L PI103 and/or 10 ng/mL TRAIL for 2 days. Four days after seeding tumors were sampled with the surrounding CAM, fixed in 4% paraformaldehyde, paraffin embedded, cut in 5 μmol/Lsections, and analyzed by immunohistochemistry by using 1:1 hematoxylin and 0.5% eosin. Images were digitally recorded at a magnification of ×2 with an AX70 microscope (Olympus), tumor areas were analyzed with ImageJ digital imaging software.

Statistical significance was assessed by Student's t Test (2-tailed distribution, 2-sample, unequal variance).

To analyze the role of PI3K signaling in the regulation of TRAIL-induced apoptosis in neuroblastoma, we selected from a panel of human neuroblastoma cell lines SH-EP, CHP-212, and LAN-5 neuroblastoma cells, because they all express caspase-8, a key component of the death receptor pathway (data not shown). Initially, we investigated the optimal conditions at which the recently developed class I PI3K inhibitor PI103 (19) blocks PI3K-mediated signaling in these cells. Dose–response and kinetic analysis revealed that PI103 inhibits PI3K signaling at nanomolar concentrations rapidly and for a prolonged time up to 48 hours (Fig. 1A). Importantly, PI103 significantly enhances TRAIL-induced apoptosis in a dose- and time-dependent manner (Fig. 1B and C). Calculation of combination index revealed a synergistic interaction of PI103 and TRAIL (Supplementary Fig. S1). Also, PI103 significantly reduced colony formation on TRAIL treatment (Fig. 1D), showing an effect on long-term survival. These data show that inhibition of PI3K by PI103 sensitizes neuroblastoma cells to TRAIL-induced apoptosis.

As we previously found that activation of Akt, but not of mTOR correlates with poor prognosis in neuroblastoma, we asked whether mTOR signaling may differentially modulate apoptosis sensitivity compared with PI3K. To address this question, we blocked PI3K/Akt/mTOR signaling selectively at the level of PI3K versus mTOR by using RNA interference (RNAi). Control experiments confirmed that silencing of the PI3K subunits p110α and p110β or of mTOR resulted in selective downregulation of the respective target proteins and to decreased phosphorylation of Akt (in case of p110α/p110β knockdown) or of S6 ribosomal protein (in case of mTOR or p110α/p110β knockdown; Fig. 2A). As silencing of mTOR inhibits both mTORC1 and mTORC2 complexes, the increase in Akt phosphorylation in cells with mTOR knockdown (Fig. 2A) may be due to differential sensitivities of mTORC1 and mTORC2 complexes to incomplete mTOR inhibition. Interestingly, knockdown of p110α/p110β significantly increased TRAIL-induced apoptosis (Fig. 2A). The reduced efficacy of genetic versus pharmacological inhibition of PI3K to confer sensitization to TRAIL-induced apoptosis may be due to incomplete knockdown of p110α/p110β. By comparison, silencing of mTOR did not alter sensitivity toward TRAIL (Fig. 2A). Similarly, pharmacological inhibition of mTOR by everolimus and rapamycin failed to sensitize neuroblastoma cells for TRAIL-induced apoptosis, although both inhibitors suppressed phosphorylation of S6 ribosomal protein at nanomolar concentrations (Fig. 2B and C). Doxorubicin was used as positive control for everolimus-mediated enhancement of apoptosis in neuroblastoma cells (20). These findings show that pharmacological or genetic inhibition of PI3K is superior to mTOR inhibition to sensitize neuroblastoma cells to TRAIL-mediated apoptosis.

To gain insights into the molecular mechanisms mediating apoptosis sensitization by PI3K inhibition, we monitored caspase cleavage by Western blot analysis. Combined treatment of TRAIL and PI103 enhanced the cleavage of caspase-8, caspase-3, and caspase-9 as well as the conversion of Bid into tBid (Fig. 3A, top). To test the requirement of caspases, we used the broad-range caspase inhibitor zVAD.fmk. Apoptosis on treatment with TRAIL and PI103 was completely blocked in the presence of zVAD.fmk (Fig. 3A, bottom), showing that apoptosis was mediated by caspases.

As tBid links the death receptor to the mitochondrial pathway by translocating to mitochondrial membranes, we then analyzed tBid in the mitochondrial fraction. Interestingly, PI103 and TRAIL cooperated to cause accumulation of tBid at mitochondrial membranes (Fig. 3B). Silencing of Bid via siRNA significantly reduced apoptosis triggered by PI103 and TRAIL and also by treatment with TRAIL alone (Fig. 3C), showing the requirement of tBid for apoptosis induction. Furthermore, PI103 significantly enhanced TRAIL-induced loss of mitochondrial membrane potential (MMP) and cytochrome c release (Fig. 3D). Translocation of tBid to mitochondrial membranes, loss of MMP, and cytochrome c release were all blocked in the presence of zVAD.fmk (Fig. 3B, 3D), showing that they occur in a caspase-dependent manner. This set of experiments shows that PI103 sensitizes neuroblastoma cells to TRAIL-induced caspase activation, Bid cleavage, translocation of tBid to mitochondria, and caspase-dependent mitochondrial outer membrane permeabilization.

Next, we monitored protein expression levels of key apoptosis regulators by Western blot analysis. Combination treatment of TRAIL and PI103 reduced protein levels of Mcl-1, XIAP, survivin, cFLIP_L, and cFLIP_S (Fig. 4A; Supplementary Fig. S2). Also, phosphorylation of Bim_EL was suppressed in cells treated with PI103 and TRAIL as indicated by the downward mobility shift of Bim_EL (Fig. 4A). By comparison, treatment with TRAIL increased the expression of Noxa, whereas expression levels of Bax, Bak, Bcl-2, or Bcl-X_L were not substantially altered (Fig. 4A; Supplementary Fig. S3).

As we observed a decline of cFLIP_L and cFLIP_S expression on treatment with TRAIL and PI103 (Fig. 4A), we analyzed formation of the DISC as one of the earliest signaling event on TRAIL receptor ligation and TRAIL receptor surface expression. However, we found no detectable changes in the recruitment of the adaptor molecule Fas-associated protein with Death Domain (FADD) or caspase-8 to stimulated TRAIL receptors within the first 4 hours on addition of TRAIL or in surface expression of agonistic TRAIL receptors (Supplementary Fig. S4).

As Mcl-1 and Noxa are markedly downregulated and upregulated, respectively, on treatment with PI103 and TRAIL (Fig. 4A), we next explored their functional relevance by using RNAi-mediated silencing. Knockdown of Noxa significantly reduced, whereas silencing of Mcl-1 significantly increased apoptosis on combination treatment with TRAIL and PI103 (Fig. 4B). Also, apoptosis induced by treatment with TRAIL alone was enhanced and reduced when Mcl-1 and Noxa were silenced, respectively (Fig. 4B). Also, knockdown of Mcl-1 led to increased PI103-mediated apoptosis (Fig. 4B).

As activation of Bax and Bak presents a central event in mitochondrial outer membrane permeabilization, we next analyzed their activation status by using conformation-specific antibodies. Notably, PI103 enhanced the TRAIL-mediated activation of Bax and Bak (Fig. 4C; Supplementary Fig. S5). Simultaneous knockdown of Bax and Bak significantly reduced TRAIL- and PI103-mediated cell death (Fig. 4C), indicating that Bax and Bak are involved in PI103-conferred sensitization to TRAIL-induced apoptosis. Together, this set of experiments indicates that changes in the expression or activation of pro- and antiapoptotic Bcl-2 proteins present critical molecular events for the PI103-mediated sensitization of neuroblastoma cells toward TRAIL.

To further investigate whether the mitochondrial pathway is required for the synergistic interaction of TRAIL and PI103, we overexpressed Bcl-2 to interfere with mitochondrial perturbations. Ectopic expression of Bcl-2 profoundly reduced apoptosis on combination treatment with PI103 and TRAIL and also with TRAIL alone (Fig. 5A). In addition, Bcl-2 overexpression inhibited TRAIL-induced Bax activation, loss of MMP, as well as cleavage of caspase-8 into p43/p41 and p18 active fragments, cleavage of Bid, and processing of the caspase-3 p20 fragment into the p17/p12 active fragments (Fig. 5B–D). Likewise, inhibition of caspases by zVAD.fmk inhibited Bax activation (Fig. 5B) and Bid cleavage, its translocation to mitochondrial membranes, and loss of MMP (Fig. 3B and D). Together, these findings point to a mitochondria-controlled feedback amplification loop from caspase-3 to caspase-8 activation, Bid cleavage, and mitochondrial outer membrane permeabilization.

Finally, we investigated the antitumor activity of PI103 and TRAIL in vivo, using the CAM assay, an established in vivo tumor model for neuroblastoma (17, 18, 21, 22), for example. Neuroblastoma cells were seeded on the CAM of chicken embryos, allowed to settle and to initiate tumors, and then treated with TRAIL in the presence or absence of PI103. Importantly, PI103 and TRAIL acted in concert to significantly suppress tumor growth of neuroblastoma compared with either agent alone (Fig. 6). This shows that PI103 cooperates with TRAIL to suppress neuroblastoma growth in an in vivo model of neuroblastoma.

We recently identified aberrant PI3K/Akt activation as a novel indicator of poor prognosis in neuroblastoma (13). Therefore, we investigated whether therapeutic targeting of PI3K/Akt by the recently developed PI3K inhibitor PI103 could restore apoptosis sensitivity in neuroblastoma. Here, we report for the first time that inhibition of PI3K is an efficient strategy to prime neuroblastoma cells to TRAIL-induced apoptosis. This conclusion is supported by several independent lines of evidence. First, pharmacological or genetic inhibition of PI3K synergistically enhances TRAIL-induced apoptosis as shown by calculation of combination index. Second, data obtained in several neuroblastoma cell lines underscore the generality of this finding. Third, the combination treatment of PI103 and TRAIL is superior to single agents in suppressing long-term survival in colony forming assays and cooperates to reduce tumor growth in an in vivo model of neuroblastoma, underlining the clinical relevance of our findings.

Mechanistically, our data show for the first time that PI3K inhibition acts in concert with TRAIL to cleave Bid into tBid and to trigger accumulation of tBid at mitochondrial membranes, an event that contributes to apoptosis induction as shown by knockdown experiments (Fig. 6B). Interestingly, PI103 enhances TRAIL-induced cleavage of caspase-8 and Bid in a mitochondrial feedback amplification loop, which likely involves caspase-3–mediated activation of caspase-8 and Bid, because Bcl-2 overexpression blocks the PI103-mediated enhancement of TRAIL-induced cleavage of caspase-3, caspase-8, and Bid. By comparison, PI103 has no effect on the recruitment of caspase-8 to the TRAIL DISC. This caspase-3–driven mitochondrial amplification loop is further facilitated by the PI103-mediated reduction of XIAP expression. To this end, Akt has been reported to protect XIAP from proteasomal degradation via phosphorylation (23).

Furthermore, we found that PI103 reduces phosphorylation of Bim_EL and Mcl-1 expression. The reduced phosphorylation of Bim_EL by PI103 is consistent with the Akt-mediated phosphorylation of Bim_EL, which enhances its proteasomal degradation (24). Mcl-1 downregulation by PI103 may be the result of reduced Mcl-1 transcription on PI3K/Akt inhibition (25). In addition, the rapid downregulation of Mcl-1 points to PI103-mediated posttranslational events, for example, Mcl-1 degradation via inhibition of GSK3β phosphorylation on PI3K inhibition (26). In addition, Noxa exerts its proapoptotic function by antagonizing the antiapoptotic effect of Mcl-1 (27). The key involvement of Mcl-1 and Noxa in this PI103-mediated sensitization to TRAIL-induced apoptosis is supported by data showing that knockdown of Noxa rescues SH-EP neuroblastoma cells from apoptosis, whereas knockdown of Mcl-1 leads to increased cell death. These alterations in the expression or activation of tBid, Noxa, and Mcl-1 shift the balance toward proapoptotic Bcl-2 proteins, which enhances TRAIL-induced Bax activation and mitochondrial outer membrane permeabilization in the presence of PI103. Additionally, PI103 promotes Bax activation by preventing the Akt-mediated inhibitory phosphorylation of Bax (28, 29). The key role of the mitochondrial pathway is underscored by experiments carried out in Bcl-2 overexpressing cells, which are almost completely resistant to PI103- and TRAIL-induced apoptosis.

In contrast to PI3K inhibition, we found that genetic or pharmacological inhibition of mTOR does not confer sensitivity to TRAIL-induced apoptosis. This result is in accordance with our recent study showing that phosphorylation of S6 ribosomal protein does not correlate with poor prognosis in neuroblastoma, whereas phosphorylation of Akt does correlate (13). In glioblastoma, we similarly found that mTOR inhibition does not enhance TRAIL- or doxorubicin-induced apoptosis (20, 30). By comparison, mTOR inhibitors have been reported to suppress tumor growth and angiogenesis in neuroblastoma cells with MYCN amplification (31) and to induce apoptosis together with chemotherapeutic agents (20, 32), implying a context-dependent role of mTOR signaling in neuroblastoma. In addition to neuroblastoma, PI3K inhibition has previously been reported to prime glioblastoma, leukemia, or colon carcinoma cells for TRAIL-induced apoptosis (20, 33–35).

Our findings have several important implications. First, our study translates our recent identification of PI3K/Akt as a potential therapeutic target in neuroblastoma (13) into the development of a novel strategy to prime neuroblastoma for TRAIL-induced apoptosis. Therapeutic intervention of aberrant PI3K/Akt activation in cancers, including neuroblastoma, is currently an area of high interest. So far, PI103 is described as a chemosensitizer in combination with anticancer drugs, for example, in T-cell acute lymphoblastic leukemia (36), glioblastoma (30, 37), or chronic lymphocytic leukemia (38). Our study identifies a novel indication for PI103 by showing that PI103 sensitizes for TRAIL-induced apoptosis. As TRAIL receptor agonists are currently evaluated as single agents in early clinical trials including childhood cancer (www.clintrials.org), TRAIL-based combination therapies, for example, by the addition of a PI3K inhibitor, become relevant to maximize the antitumor activity of TRAIL. This approach will likely be effective in a subpopulation of neuroblastoma, because caspase-8 is frequently silenced by epigenetic mechanisms in neuroblastoma (18, 39–41).

Second, our data provide novel insights into the signal transduction pathways, which are regulated by PI3K inhibitors and which can be exploited to prime cancer cells for TRAIL-induced apoptosis. Thus, beyond neuroblastoma these results will likely have an important impact on the development of combination therapies with PI3K inhibitors also in other types of cancer.

In conclusion, this preclinical evaluation of a rational combination of 2 novel classes of targeted drugs, that is, the PI3K inhibitor PI103 and TRAIL, in relevant preclinical in vitro and in vivo models of neuroblastoma provides the molecular basis for the design of new combination therapies for the treatment of neuroblastoma. This strategy may help to overcome apoptosis resistance of neuroblastoma and other cancers with aberrant activation of PI3K/Akt.

No potential conflicts of interest were disclosed.

We thank C.A. Schmitt (Berlin, Germany) for kindly providing mouse Bcl-2 vector and the Novartis Institute for BioMedical Research for kindly providing everolimus.

This work has been partially supported by grants from the Deutsche Forschungsgemeinschaft, Deutsche Krebshilfe, Novartis Stiftung für therapeutische Forschung, Kind-Philipp-Stiftung, European Community (ApopTrain, APO-SYS), and IAP6/18 (to S. Fulda).