Fenretinide-induced caspase-8 activation and apoptosis in an established model of metastatic neuroblastoma

Background Resistance of high-risk metastatic neuroblastoma (HR-NB) to high dose chemotherapy (HD-CT) raises a major therapeutic challenge in pediatric oncology. Patients are treated by maintenance CT. For some patients, an adjuvant retinoid therapy is proposed, such as the synthetic retinoid fenretinide (4-HPR), an apoptotic inducer. Recent studies demonstrated that NB metastasis process is enhanced by the loss of caspase-8 involved in the Integrin-Mediated Death (IMD) process. As the role of caspase-8 appears to be critical in preventing metastasis, we aimed at studying the effect of 4-HPR on caspase-8 expression in metastatic neuroblasts. Methods We used the human IGR-N-91 MYCN-amplified NB experimental model, able to disseminate in vivo from the primary nude mouse tumor xenograft (PTX) into myocardium (Myoc) and bone marrow (BM) of the animal. NB cell lines, i.e., IGR-N-91 and SH-EP, were treated with various doses of Fenretinide (4-HPR), then cytotoxicity was analyzed by MTS proliferation assay, apoptosis by the propidium staining method, gene or protein expressions by RT-PCR and immunoblotting and caspases activity by colorimetric protease assays. Results The IGR-N-91 parental cells do not express detectable caspase-8. However the PTX cells established from the primary tumor in the mouse, are caspase-8 positive. In contrast, metastatic BM and Myoc cells show a clear down-regulation of the caspase-8 expression. In parallel, the caspases -3, -9, -10, Bcl-2, or Bax expressions were unchanged. Our data show that in BM, compared to PTX cells, 4-HPR up-regulates caspase-8 expression that parallels a higher sensitivity to apoptotic cell death. Stable caspase-8-silenced SH-EP cells appear more resistant to 4-HPR-induced cell death compared to control SH-EP cells. Moreover, 4-HPR synergizes with drugs since apoptosis is restored in VP16- or TRAIL-resistant-BM cells. These results demonstrate that 4-HPR in up-regulating caspase-8 expression, restores and induces apoptotic cell death in metastatic neuroblasts through caspase-8 activation. Conclusion This study provides basic clues for using fenretinide in clinical treatment of HR-NB patients. Moreover, since 4-HPR induces cell death in caspase-8 negative NB, it also challenges the concept of including 4-HPR in the induction of CT of these patients.


Background
Neuroblastoma (NB), the second most common solid tumour in children, is an embryonic malignancy which originates from sympathetic neurons. Whereas localized tumors in young infants often spontaneously regress or mature in response to treatments, the outcome of highrisk NB, incurable in 60% of cases, remains poor and poses a major therapeutic challenge in pediatric oncology [1]. Apoptosis or programmed cell death is believed to represent an essential mechanism of drug-mediated toxicity [2]. Recent studies highlight the role of apoptosis in the metastatic process showing that defects in the apoptosis program contributed not only to malignancy or drug resistance but also to metastasis. Indeed, apoptosis is nowadays considered as a central mechanism in the development of metastases [3].
Apoptosis is an extremely well-organized process mediated by the caspases protein family. Upon activation by proteolytic cleavage, effector caspases cleave their substrates and inactivate proteins essential for survival, leading to the disintegration of cells [4]. The caspase-8 is a key enzyme at the top of the apoptotic cascade, both involved in the extrinsic or death receptors pathway and in the intrinsic mitochondrial pathway. Several studies have reported that the caspase-8 gene was frequently inactivated by hypermethylation in NB cell lines [5][6][7], alterations mainly described in aggressive and amplified MYCN high-stage tumors. However, no correlation has been established between caspase-8 expression and MYCN amplification or aggressive disease criteria [8]. Despite caspase-8 status is not predictive of aggressive NB, recent findings suggest that caspase-8 loss contributes to a metastatic phenotype, thus defining caspase-8 as a metastasis suppressor gene for NB [9]. The authors demonstrated that NB metastasis process is enhanced by the simultaneous loss of caspase-8 and of α3β1 integrin which subsequently impairs integrin-mediated death (IMD) to occur [10]. Therefore, caspase-8 appears as an attractive target to reduce the formation of NB metastases [11]. Among the numerous approaches manipulating caspase-8 levels, therapeutic strategies have used agents such as 5-Azacytidine [6], IFN-γ [12,13], IFN-γ associated with 5-Azacytidine [14], or the synthetic retinoid Fenretinide N-(4-hydroxyphenyl) retinamide/4-HPR [15]. Interestingly, 4-HPR has been reported to inhibit the invasion of several cancer cell types, including breast cancer, ovarian, prostate cancer and Kaposi's sarcoma, and NB [16]. Identified to induce apoptosis in various cell lines, the reagent 4-HPR-, -unlike retinoic acid (RA)-, -is more effective to induce apoptosis than differentiation in NB and proved a good efficacy against NB cell lines which are resistant to RA [17,18]. Moreover, 4-HPR is clinically well tolerated and used for clinical trials in neuroblastoma [19,20].
Our previous studies, performed in the experimental IGR-N-91 human MYCN-amplified model, reported the main involvement of P-gp/MDR in the chemoresistance of metastatic neuroblasts [21,22]. In this NB xenograft model, the IGR-N-91 neuroblasts have disseminated in vivo from the primary nude mouse tumor xenograft (PTX) into myocardium (Myoc) and bone marrow (BM) of the animal. Given the recent findings of i) potentiation of NB metastasis by loss of caspase-8 in a murine model and ii) halting NB metastasis by controlling IMD in the same model, we have investigated the expression of caspase-8 in metastatic neuroblasts of this experimental model. As we have observed a specific loss of caspase-8 expression in metastases, we therefore decided to restore caspase-8 activity with 4-HPR and to study its apoptotic-inducing effect in metastatic neuroblasts. Moreover, we have investigated the putative synergic effect of 4-HPR with cytotoxic drug-or TRAIL-mediated apoptosis in chemoresistant metastatic neuroblasts. Our results provide evidence that 4-HPR enhances cell death in caspase-8 deficient metastatic malignant neuroblasts and thus may be combined with toxic agents to an increased efficacy in the treatment of high-risk metastatic NB.

Cell culture and reagents
NB cell lines include the N-type NB IGR-N-91 cells and the S-type NB cells SH-EP. The IGR-N-91 cell line was established from a stage-4 NB infiltrated bone marrow [21]. Human IGR-N-91 NB cells were injected subcutaneously into nude mice: a primary tumor xenograft cell line (PTX), blood cells (BC), myocardium (Myoc) and bone marrow (BM) sub-lines were obtained from mechanically dissociated tumor samples and were grown as previously described [21,22]. Each cell line was cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 2 mM L-glutamine, 10 μg/ml gentamycin and 10% fetal calf serum, at 37°C in a 5% CO2 humidified atmosphere. 4-HPR (from Sigma-Aldrich, Germany) was dissolved in ETOH. Throughout this study, we used 1-10 μM of 4-HPR for the N-type NB cells and 1-20 μM for the S-type NB cells. The concentrations were sufficient to induce cell death in NB cells as determined previously [23,24]. Cells were treated with 20-40 μM etoposide (VP16) (Sigma, St Louis, MO), both dissolved in DMSO, for 24 or 48 hours. An equal volume of vehicle control (DMSO) was used to treat control cells. Cells were incubated with indicated amount of soluble recombinant TRAIL (a gift from J. Tschopp) and cross-linking anti-Flag Ab M2 (Sigma) with a constant ratio of 1/5 of TRAIL/M2 as previously described [25]. Cells were pretreated with caspase inhibitors zIETD-fmk, zDEVD-fmk, zVDAVD-fmk and zLEHD-fmk (50 μM, R&D systems SA, Switzerland) before drug treatment.

Tumor growth
Athymic locally bred Swiss nude mice were engrafted with the IGR-N-91 cell line, in accordance to the European Community guidelines (authorization #86/609/CEE). Mice (n = 4) were anaesthetized using ketamine 50 mg/kg and xylazine 35 mg/kg intraperitoneally (I.P.). Heterotopic tumors were obtained by subcutaneous injection in the back of the animal of 1 × 10 6 NB cells suspended in 200 ml of Matrigel ® (BD Biosciences, San Diego, CA) using a 22 G needle connected to a 1-ml syringe. After recovery, the animals were kept in a germ-free protected area, with food ad libitum. Mice were sacrificed after 2 months. Tumors were resected and dissociated as described previously [26]. Proteins were isolated using lysis buffer as described above.

Retroviral infection experiments
IGR-N-91 cells were infected with a biscistronic retroviral vector encoding for the GFP and the caspase-8a genes (IGR-N91-C8) or with the GFP control vector (IGR-N91-M) as described previously [25]. One clone was respectively chosen among the GFP positive IGR-N91-M clones and among the IGR-N91-C8 clones, then assessed to control the expression level of caspase-8 by immunoblotting. Stable downregulation of caspase-8 was achieved by RNA interference using short hairpin RNAs [27]. Different splice variants of caspases-8 mRNA were identified leading to the production of diverse C8 isoforms [28]. To efficiently silence all caspase isoforms, the shRNAs have been localized in common exons and correspond to the sequence 5'-AAGTTCCT-GAGCCTGGACTAC-3' as proposed previously [29]. The sequence of oligonucleotides encoding the control random shRNA (shRNAC) is 5'-CGATTGCGATCGTCAATTCCC-3' [30]. The sequences were checked by sequencing a PCRamplified region containing the oligonucleotide and shR-NAs were prepared and used to infect the SH-EP cell line as already described [30].

Cell viability and apoptosis assays
Cell viability was measured using the MTS/PMS cell proliferation kit from Promega (France) according the manufacturer's instructions. 10 5 cells were seeded 24 hours into a 96-well plate containing 100 μl of medium, then incubated with 4-HRP or drugs for 24 or 48 hours. Assays were performed in quadruplicates. Percentage of cell viability/ cell death was determined as compared to untreated controls. Apoptosis was evaluated using the standard PI-staining protocol. Cells were trypsinized and neutralized in complete medium, washed twice with 3 ml PBS. Cell pellets were resuspended in 1 ml of PBS on ice then fixed in 3 ml of 100% ethanol. After fixation, pellets were kept at -20°C. Cells were stained in PI solution (50 μg/ml PI, 100 μg/ml RNAseA in PBS) for 30 min at room temperature.
The stained cells were analyzed using a FACScan flow cytometer (Becton Dickinson, France SAS).
Caspase-8 expression in the IGR-N-91 experimental metastatic model

Caspases activities
Caspases-3, -8, and -9 protease activities were measured using the caspase colorimetric protease assay kits from MBL Co., Ltd, USA as previously described [25]. Cytosolic lysates were prepared 24 or 48 hrs after 4-HPR treatment according to manufacturer's instructions. 100 μg protein extracts were incubated with 200 μM of DEVD-pNA, IETD-pNA, Ac-LEDH-pNA colorimetric substrates for 3 hrs at 37C. Cell lysates were incubated with 10 μM of caspase inhibitor (zDEVD-fmk, zIETD-fmk, zLEDH-fmk) for 30 min before addition of the respective caspase substrates for control experiments. Hydrolyzed pNA was detected using a microtiter plate reader at 405 nm. Background absorbance from cell lysates and buffers were subtracted from the absorbance of stimulated and unstimulated samples before calculation of relative caspases activities.

Statistical Analysis
Results are shown as the mean ± SD of at least three experiments each. Data were analyzed using the Student's t test with P values of <0.05 (*) or <0.01 (**) considered to represent significance.

Loss of caspase-8 expression in metastatic neuroblasts
We have first analyzed caspase-8 expression level in PTX, BM and Myoc cell lines of the IGR-N-91 experimental model ( Figure 1A). We observed that the IGR-N-91 cell line is caspase-8-negative. Surprisingly, caspase-8 expression is restored in PTX compared to parental IGR-N-91 neuroblasts, then downregulated in BM or Myoc metastatic neuroblasts. RNA analysis revealed caspase-8 mRNA transcripts in parental IGR-N-91 cells as well as in Myoc neuroblasts while totally absent in BM cells ( Figure 1A). In contrast, no change was observed in the expression levels of caspases-3, -9 as well as Bcl-2 and Bax. Therefore, we examined by immunoblotting caspase-8 expression level in four primary tumors xenografts of mouse engrafted with IGR-N-91 neuroblasts ( Figure 1B). The SH-EP cell line was used as a caspase-8-positive control cell line. As for PTX neuroblasts, caspase-8 expression is fully restored in the four primary tumors.

4-HPR induced re-expression of caspase-8 and apoptotic cell death in BM metastatic neuroblasts
As the role of caspase-8 appears critical in preventing metastasis, we aimed at studying the effect of the retinoid 4-HPR on caspase-8 expression. Caspase-10 treated with 5 or 10 μM 4-HPR for 48 hours and showed that PTX cells are more resistant to cell death than BM cells ( Figure 3A). Indeed, 4-HPR-induced cell death is doseand time-dependent ( Figure 3A, B). As 4-HPR has been previously described to induce apoptotic cell death in neuroblastoma cell lines [23], we have compared the extent of 4-HPR-mediated apoptosis in PTX and BM neuroblasts. FACS analysis shows that apoptosis was significantly increased in BM neuroblasts, whereas no apoptosis occurred in PTX cells ( Figure 3C). Thus, the increase of the caspase-8 expression appears correlated with a higher sensitivity of apoptotic cell death in BM compared with PTX neuroblasts.

Caspase-8 activation by 4-HPR in BM metastatic neuroblasts
In order to analyze the involvement of caspase-8 in celldeath execution, we conducted studies using specific inhibitors. The 4HPR-induced cell death is not significantly blocked in PTX cells by zDEVD-fmk (a caspase-3 inhibitor), zIETD-fmk (a caspase-8 inhibitor), or zLEHD-fmk (a caspase-9 inhibitor) ( Figure 4A). In contrast, we observed that 4HPR-induced cell death can be blocked by zIETD-fmk (P < 0.01), more partially by zLEDH-fmk (P < 0.01) and slightly by zVDAVD-fmk (a caspase-2 inhibitor) (P < 0.05) in BM cells. However, no inhibition of cell death was noted with the caspase-3 inhibitor zDEVD-fmk ( Figure 4B). The A slight increase of caspase-8 activity is noted within 24 hrs, appears more sustained after 48 hours, and is reduced by the caspase-8 inhibitor zIETD-fmk ( Figure 4C). However, no significant variation of caspases-3 or -9 activities was evidenced in these conditions (data not shown).

Caspase-8 negative NB cell lines are resistant to 4-HPRinduced cell death
To corroborate the specific contribution of caspase-8 in 4-HPR-induced apoptosis, we have used the caspase-8-deficient NB cell line (IGR-N91-M) and the stably restored caspase-8 cell line (IGR-N91-C8) previously described [25]. We observed that the treatment of 2.5 or 5 μM 4-HPR strongly up-regulated caspase-8 expression in the IGR-N91-M neuroblasts, while no change was observed in the IGR-N91-C8 cells ( Figure 5A). In parallel, significant cell death may be noted both in 4HPR-treated IGR91-M and IGR91-C8 neuroblasts ( Figure 5B). Thus, the increase of caspase-8 expression may be correlated with the higher cell death sensitivity noted in the 4HPR-incubated IGR91-M cells compared with control.
To extend the observation that caspase-8 activation plays a major role in 4-HPR-induced cell death, we examined the response of SH-EP cell line to 4-HPR, after silencing of caspase-8 expression. SH-EP cells were stably transfected with shRNAs against caspase-8 (SHEP-shC8) and nonsilencing control (SHEP-nsc). As expected, western blot analysis shows that caspase-8 is silenced in SHEP-shC8 ( Figure 6A). The two cell lines were then treated with increasing doses of 5, 10 or 15 μM 4-HPR for 48 hours. The 4HPR treatment did not change caspase-8 expression neither in SHEP-nsc or SHEP-shC8 cells ( Figure 4B). Cell death was measured using the MTS test ( Figure 6B). Inter-

4-HPR enhances chemotherapeutic drug-or TRAILmediated toxicity in metastatic neuroblasts
Next, we have investigated if 4-HPR could enhance drugmediated apoptosis in BM metastatic neuroblasts, which, contrary to what occurs with 4-HPR, are more resistant than PTX neuroblasts to toxicity by drugs. Indeed, we had previously shown that BM metastatic neuroblasts are resistant to etoposide/VP16 toxicity [22]. Figure 7A confirms that BM are more resistant than PTX cells to drug toxicity after incubation of cells with 20 or 40 μM VP16 for 24 hours. PTX and BM neuroblasts were pretreated with 2.5 μM 4-HPR for 96 hours and data show that 4-HPR enhances VP16-mediated toxicity in BM cells ( Figure  7B). These data are in accordance with a previous study showing a synergistic apoptotic response of 4-HPR with chemotherapeutic agents, including cisplatin, carboplatin or etoposide in NB [31]. Caspase-8-positive NB cells have been previously demonstrated to be weakly sensitive to TRAIL [25]. Moreover, studies reported silencing of caspase-8 expression in malignant IGR-N-91 cells as a possible mechanism of resistance to TRAIL-induced apoptosis. Indeed, sensitivity to TRAIL was fully restored in the caspase-8-complemented NB cell line [32]. Here we analyzed the cell death response of BM cells to TRAIL-mediated toxicity compared with PTX cells ( Figure 8A). The caspase-8silenced BM neuroblasts appear more resistant than PTX neuroblasts to TRAIL-toxicity. Moreover, we observed a more important enhancement of TRAIL-mediated toxicity in BM cells than in PTX after incubation with 2.5 μM 4-HPR for 96 hours ( Figure 8B). These data corroborate previous studies showing that 4-HPR enhances TRAIL-mediated apoptosis in ovarian or colon cancer cells [33,34].

Discussion
Numerous studies have reported that the loss of caspase-8 expression occurs in medulloblastoma [35], colon carcinoma [36], more frequently in small cell lung carcinoma, glioma and malignant neuroblastoma [37,38,6]. In highrisk neuroblastoma, resistance to apoptosis may involve several defects of the apoptotic pathway, such as overexpression of the anti-apoptotic proteins Bcl-2, Bcl-x L , and survivin [39], but also deregulation of the death receptor pathway with the inactivation of caspase-8 gene expression mostly by gene silencing [5,6,40].
The present study shows that the downregulation of caspase-8 plays a role in resistance of metastatic neuroblastoma to apoptosis, which could be reversed by 4-HPR treatment. Since 4-HPR upregulates caspase-8 expression, apoptotic resistance induced by 4-HPR appears improbable. We have investigated the expression of caspase-8 in the IGR-N-91 NB experimental model which offers the opportunity to compare NB cell lines established from a primary tumor and matched metastatic neuroblasts [21,22]. In this model, the IGR-N-91 parental cell line is negative for caspase-8 protein despite normal level of mRNA. This suggests the occurrence of a post-transcriptional mechanism to shut off caspase-8 protein expression in parental IGR-N-91 cells. We observed that caspase-8 protein is re-expressed in the primary tumor cell line (PTX) and both caspases-8 mRNA and protein expressions are lost in metastatic neuroblasts (BM and Myoc). As known and previously described [6], the role of methylation in the caspase-8 silencing process is supported by the up-regulation of caspase-8 observed in NB cell lines treated with demethylating agents such as 5-azacytidine. Thus, changes in methylation and/or chromatin structure might explain the caspase-8 silencing in NB metastatic cells, and inversely the re-expression of caspase-8 in tumors with the contribution of others mechanisms such as interactions with the cellular microenvironment. Moreover, caspase-8 seems to not impact primary tumor growth as the cells interact with each other and with the IGR-N91 neuroblasts sensitivity to 4-HPR-induced cell death remodeling tumor ECM, caspase-8 being able to regulate tumor survival during tissue invasion [41]. The loss of caspase-8 is specific since the expression of other caspases or apoptotic partners was unchanged. Interestingly, our data are consistent with recent findings which first, correlate the decrease of apoptosis induction due to the loss of caspase-8 with the increase of metastatic potential in NB [9], and secondly, with the loss of caspase-8 or integrins allowing malignant disseminated cells to acquire a greater potential to survive [10].
As the role of caspase-8 appears to be crucial in preventing metastasis, we therefore decided to study the effect of the synthetic retinoid fenretinide/4-HPR on caspase-8 expres- In vitro studies in NB have shown that 4-HPR induced apoptosis through RARdependent or -independent pathways [23,24]. The activation of c-Jun N-terminal kinase as well as the activation of mitochondrial pathway via generation of reactive oxygen species (ROS) or the induction of increased ceramide production have been implicated in 4HPR-mediated apoptosis [42,43]. Here, we report that 4-HPR induces apoptosis both in PTX and BM malignant neuroblasts. Interestingly, 4-HPR up-regulates caspase-8 expression in caspase-8 negative IGR-N91 cells, i.e, BM neuroblasts or retrovirally-transduced IGR-N91-control cells. Moreover, this up-regulation is correlated with a significant increase of apoptotic cell death in BM neuroblasts. Data confirm that BM metastatic neuroblasts are more sensitive to 4-HPR-induced cell death than PTX cells and that caspase-8 activation occurs in BM neuroblasts treated by 4-HPR. This higher sensitivity of BM cells to 4-HPR may be explained by two mechanisms: first, the involvement of IMD impairment in the apoptotic resistance of metastatic BM cells. Interestingly, we have observed a decrease of the expression levels of the α3 and β1 integrins in BM or Myoc metastatic cells compared with parental IGR or PTX cells (data not shown). As 4-HPR induces an increase of caspase-8 expression, it would be informative to accurately investigate if 4-HPR is able to restore IMD, i.e not only caspase-8 but also α3β1 integrin expressions. Thus, the higher sensitivity of BM cells to 4-HPR might be imputed to a synergistic effect of both intrinsic 4-HPR toxicity and restored IMD; secondly, studies using specific inhibitors show that 4HPR-induced cell death is almost completely inhibited by the addition of zIETD-fmk, a caspase-8 inhibitor and more partially with zLEHD-fmk, a caspase-9 inhibitor. In contrast, these inhibitors did not protect PTX cells for 4-HPR toxicity, suggesting that cell death induced by 4-HPR occurred by distinct pathways in PTX cells compared to BM cells.
Previous findings reported that 4-HPR stimulates the generation of intracellular free radicals which has a causative role in the induction of apoptosis and that caspase-3 activation is required in this oxidative stress-mediated apoptosis [24,44]. However, in our study, the caspase-3 inhibitor does not inhibit 4-HPR-induced apoptosis in BM neuroblasts, suggesting that other pathways may be involved. If 4-HPR-induced BM cell death is partially inhibited by the caspase-2 inhibitor, further studies are clearly needed to elucidate the mechanism of 4HPR/caspases interactions involved in this experimental model. Nevertheless, the crucial role of caspase-8 in 4-HPR mediated death in BM cells is confirmed by data performed with the transfected and silenced pair of IGR-N91-M/C8 and SH-EP-nsc/shC8 cells. As observed in the PTX/BM pair of cells, caspase-8 expression was increased by 4-HPR in the caspase-8 deficient cells IGR-N91-M cells, whereas in SH-EP-shC8, caspase-8 over-expression could not be detected because C8 mRNA was degraded by the shRNA.
Interestingly, the stable restoration of caspase-8 expression in IGR-N91-C8 cells increased their sensitivity to 4-HPR compared to parental IGR-N91-M cells, while the reversed experiment, performed by down-regulating caspase-8 expression in SH-EP-shC8, increased their resistance to 4-HPR compared to parental SH-EP-nsc.
Similar caspase-8 activation has been described in 4-HPRinduced cell lines such as ovarian cancer cell lines or Fasdefective hepatoma cells [33,45]. More recently, different retinoic acid derivatives, including 4-HPR, were demonstrated to increase caspase-8 transcription via induction of phosphorylated CREB in NB [46]. Moreover, the authors demonstrate that the binding of this transcription factor to sequences within the caspase-8 gene sensitizes cells to TNFα or drug-induced apoptosis [46]. Our study corroborates that caspase-8 appears as a key mediator of the synergistic effect observed between RA derivatives/4HPR and drugs/etoposide or TRAIL. Indeed, 4-HPR represents a potential clinical candidate to enhance the effect of chemotherapy in NB [31]. However, this effect was imputed to the generation of free radicals by 4-HPR, considered as the key property of this retinoid leading to synergistic response with chemotherapy [31]. Resistance to TRAIL-mediated cell death has been attributed to deregulation of a number of signalling molecules [47], including inactivation of caspase-8 [5][6][7]. Indeed, restored caspase-8 expression results in an increase of sensitivity of NB cells in TRAIL-induced apoptosis [25]. Here, we report that the caspase-8-negative BM neuroblasts are more resistant to TRAIL-mediated cell death than PTX cells. Moreover, a more important enhancement of TRAIL-induced cell death is observed in BM neuroblasts pre-treated with 4HPR. Recent studies performed in colon cancer cell lines described such a synergistic effect of the combination of TRAIL with 4-HPR leading to the activation of multiple caspases [34].
Both previous reports and our study propose the restoration of caspase-8 function as essential to sensitize resistant NB to apoptosis. However, NB malignant cells are characterized by a great heterogeneity and show various resistance or sensitivity to TRAIL or drug toxicities. Thus, studies have always to be performed to develop new and various combination regimens for high-risk neuroblastoma. Indeed, previous findings reported the caspase-8 induction by IFN-γ in selected NB cells [12], able to sensitize NB to TRAIL apoptosis [48]. More recently, the association of the demethylating agent 5-Aza-2'-deoxycytidine (5-dAzaC) with IFN-γ was described to cooperate using lower individual drug doses in order to increase the therapeutic efficacy and decrease the toxicity of TRAIL-apoptosis in NB [14].

Conclusion
Our study confirms that 4-HPR/Fenretinide is a very interesting therapeutic drug, reported to exert a benefit effect in the treatment of prostate cancer metastasis or to suppress the invasiveness of bone metastatic breast cancer cells [49,50]. Of note, fenretinide is currently studied for a phase II trial in HR-NB patients following a successful phase I trial and recent pharmacokinetics studies conducted in NB patients indicated that a continuous oral administration of 4-HPR is preferred to maintain a relevant and efficient plasma concentration of 0.7 -10 μM range [19,51)]. Moreover, active metabolites of 4-HPR such as 4-oxo-4-HPR are currently under investigation and may provide a promising therapy for HR-NB [51].