Reactive Oxygen Species Mediates the Synergistic Activity of Fenretinide Combined with the Microtubule Inhibitor ABT-751 against Multidrug-Resistant Recurrent Neuroblastoma Xenografts

Retinoids are active against neuroblastoma both as differentiation inducers and as cytotoxic agents. The differentiation inducer 13-cis-retinoic acid (13-cis-RA) achieved complete responses in a high-risk neuroblastoma phase I study of high-dose (2 weeks on, 2 weeks off) after myeloablative therapy (1), and a randomized phase III study demonstrated that maintenance therapy of high-risk neuroblastoma with 13-cis-RA after completion of cytotoxic consolidation therapy significantly enhanced event-free survival (EFS; ref. 2). Outcome in a subsequent non-randomized study employing 13-cis-RA as maintenance therapy was consistent with a benefit from 13-cis-RA maintenance therapy (3). A phase III trial demonstrated that maintenance therapy that combined 13-cis-RA with ch14.18 antibody + cytokines further improved outcome (4).

The synthetic retinoid N-(4-hydroxyphenyl)retinamide (fenretinide; 4-HPR) is cytotoxic for neuroblastoma in vitro by p53-independent and caspase-dependent and -independent mechanisms that involve increases of reactive oxygen species (ROS) and dihydroceramides (5, 6). Clinical activity in recurrent neuroblastoma of 4-HPR when given as suboptimally bioavailable capsules was modest in early-phase studies (7–9). Incorporation of 4-HPR into a lipid matrix (LYM-X-SORB) and formulation as an oral powder (4-HPR/LXS) increased bioavailability and antineuroblastoma activity in mouse xenografts (10). A pediatric phase I trial of 4-HPR/LXS demonstrated significantly higher 4-HPR exposures than previously achieved with the capsule formulation and documented four complete responses in recurrent neuroblastoma (11).

ABT-751 is a sulfonamide microtubule inhibitor that binds to the colchicine-binding site on β-tubulin, inhibiting microtubule polymerization (12, 13). In cell lines, exposure to ABT-751 leads to a cell-cycle block at G₂–M and induces apoptosis (13). As ABT-751 is not a substrate for P-glycoprotein, it is active against tumor models resistant to other microtubule inhibitors, such as vincristine and paclitaxel, and ABT-751 has been shown to be active against neuroblastoma xenografts (14). Phase I clinical trials with ABT-751 achieved stable disease but few objective responses in recurrent neuroblastoma (15–17). However, a Children's Oncology Group phase II study of ABT-751 failed to demonstrate an increase in progression-free survival when compared with historical controls (18).

Although there has been a consistent improvement in outcome for neuroblastoma patients over the past 2 decades, many patients with high-risk neuroblastoma develop disease progression that is refractory to further therapy (19). New drugs and drug combinations active against recurrent, multidrug-resistant neuroblastoma are needed to improve survival. Moreover, well-tolerated agents that demonstrate clinical activity against recurrent neuroblastoma may be used to further improve outcome by employing them in post-consolidation maintenance therapy (2, 4, 20) Both 4-HPR/LXS and ABT-751 are well-tolerated, orally available drugs with differing mechanisms of action and nonoverlapping systemic toxicities that have shown single-agent activity in recurrent neuroblastoma preclinical models (5, 10, 14, 17). Thus, we investigated whether combining 4-HPR with ABT-751 could enhance antineuroblastoma activity against neuroblastoma cell lines and multidrug-resistant neuroblastoma xenografts established from patients with progressive disease.

4-HPR and 4-HPR/LXS (the LYM-X-SORB powder formulation of 4-HPR; ref. 10) were provided by the Developmental Therapeutics Program of the National Cancer Institute (NCI; Bethesda, MD). ABT-751 was supplied by Abbot Laboratories. Vincristine sulfate salt, vinorelbine ditartrate, paclitaxel, colchicine, ascorbic acid (vitamin C; Vit C), α-tocopherol (vitamin E; Vit E), n-acetylcysteine (NAC), sodium thiosulfate (STS), fluorescein diacetate (FDA), eosin Y, DMSO, and ethanol were purchased from Sigma-Aldrich. For in vivo administration, vincristine was manufactured by Mayne Pharma Inc., colchine was manufactured by West-Ward Pharmaceutical Corp., and paclitaxel (Taxol) was manufactured by Bristol Myers Squibb Co. Paraformaldehyde was from USB Corporation. MitoProbe JC-1 and 5-(and-6)-carboxy-2′,7′-dichlorodihydrofluorescein diacetate (carboxy-H₂DCFDA) were from Invitrogen. 4-OH-fenretinide was from Toronto Research Chemicals, Inc. References to chemical structures are 4-HPR (21), ABT-751 (22), vincristine, vinorelbine, and colchicine (23).

Human neuroblastoma cell lines, SMS-KCNR, SMS-SAN, CHLA-15, CHLA-20, CHLA-90, CHLA-119, CHLA-136, and CHLA-140 have been described previously (24–28). SMS-SAN and CHLA-15 were established at diagnosis before therapy, SMS-KCNR at progressive disease after dual-agent induction chemotherapy. CHLA-20, CHLA-119, and CHLA-140, are multidrug-resistant cell lines established at time of progressive disease; CHLA-119 was established from a rapidly fatal marrow relapse after a single course of 13-cis-RA after complete response to CCG-3891 consolidation chemotherapy (2). FU-NB-2006 was established from post-mortem blood at time of progressive disease after intensive multiagent chemotherapy and single-agent oral fenretinide. CHLA-90 and CHLA-136 are multidrug-resistant cell lines established at relapse after myeloablative chemotherapy and autologous bone marrow transplant. CHLA-90 and CHLA-119 bear TP53 loss-of-function mutations (26); CHLA-90, CHLA-119, CHLA-136, and CHLA-140 overexpress the MDR1 gene relative to neuroblastoma cell lines established at diagnosis (Reynolds, unpublished data).

SMS-SAN and SMS-KCNR were cultured in RPMI-1640 (Mediatech Inc.) supplemented with 10% heat-inactivated FBS. CHLA-15, CHLA-20, CHLA-90, CHLA-119, CHLA-136, CHLA-140, and FU-NB-2006 were cultured in Iscove's modified Dulbecco medium (Cambrex) supplemented with 3 mmol/L l-glutamine, 5 μg/mL insulin, 5 μg/mL transferrin, 5 ng/mL selenous acid, and 20% heat-inactivated FBS. Cell lines used in this study were tested to be Mycoplasma-free and maintained at 37°C in a humidified incubator containing 95% room air and 5% CO₂ atmosphere. Cell line identities were validated at time of experimentation by short tandem repeat (STR) profiling (29) as compared with the Children's Oncology Group cell line and xenograft STR database (www.COGcell.org).

Cytotoxicity dose–response curves (4 days after start of drug exposure) were determined using the DIMSCAN digital imaging microscopy assay system as previously described (30, 31). Concentration ranges tested in vitro (modeled on clinically obtainable plasma levels of active agents and taking into account protein-binding considerations) were: 4-HPR, 0–10 μmol/L; ABT-751, 0–500 ng/mL colchicine, 0–10 μmol/L; vincristine, 0–10 μmol/L; vinorelbine ditartrate, 0–10 μmol/L; and paclitaxel; vitamin C, 0–250 μmol/L; vitamin E, 0–250 μmol/L; n-acetylcysteine 0–500 μmol/L; sodium thiosulfate 0–500 μmol/L. Cells (4,000 in 100 μL/well) were seeded in 96-well plates 16 to 24 hours before 100 μL of drugs (stock solutions: 10 mmol/L 4-HPR in 99.5% ethanol; 1 mg/mL ABT-751 in DMSO) were added to each well (n = 12 replicates). For antioxidant studies, cells were pretreated for 3 hours with antioxidants before addition of 4-HPR and ABT-751.

CHLA-119 cells (1 × 10⁶) were treated with 0.62 μmol/L 4-HPR, 62.5 ng/mL ABT-751, and 0.62 μmol/L 4-HPR + 62.5 ng/mL ABT-751, alone or in combination with antioxidants [250 μmol/L vitamin C, 250 μmol/L vitamin E, 500 μmol/L n-acetylcysteine (NAC), or 500 μmol/L sodium thiosulfate (STS)] for 3 hours. For some experiments, concentrations of NAC to 10 mmol/L and of STS to 5 mmol/L were employed. Cells were incubated in 1 mL of medium containing 50 μmol/L of the ROS-sensitive probe carboxy-H₂DCFDA (5) for 25 minutes at 37°C. For a positive control, H₂O₂ was added to the cells at a final concentration of 100 μmol/L for 15 minutes. Cells were centrifuged, resuspended in 500 μL of medium, and analyzed by a BD LSR II flow cytometer.

TUNEL (terminal deoxynucleotidyltransferase dUTP nick end labeling; APO-DIRECT kit; BD Biosciences) was used to assess apoptosis. CHLA-119 cells (1 × 10⁶) exposed for 20 hours to 0.62 μmol/L 4-HPR, 62 ng/mL ABT-751, or 0.62 μmol/L 4-HPR + 62 ng/mL ABT-751 (minimal drug concentrations demonstrating strong synergy) with or without a 3-hour pretreatment with (and continued exposure to) antioxidants (250 μmol/L vitamin C, 250 μmol/L vitamin E, 500 μmol/L n-acetylcysteine, or 500 μmol/L sodium thiosulfate). Cells were fixed with 1% (w/v) paraformaldehyde in PBS, stored in 70% ethanol at −20°C, washed and incubated in 50 μL of the TUNEL staining solution (2 hours, 37°C), washed twice, and 500 μL of propidium iodide/RNase added before flow cytometry using bandpass filters: 525 ± 25 nm (FITC), 610 ± 10 nm (propidium iodide).

Cells were incubated in 1 mL of medium containing 1 μmol/L JC-1 mitochondrial probe (Invitrogen) for 30 minutes at 37°C, washed once with PBS and analyzed by flow cytometry. Mitochondrial membrane depolarization was indicated by a decrease in the red (590 ± 10 nm) to green (525 ± 25 nm) fluorescence intensity ratio.

CHLA-119 cells (10 × 10⁶ cells) were treated with 0.62 μmol/L 4-HPR, 62.5 ng/mL ABT-751, and 0.62 μmol/L 4-HPR + 62.5 ng/mL ABT-751 ± 250 μmol/L vitamin E for 24 hours and subjected to subcellular fractionation (BioVision Mitochondria/Cytosol Fractionation Kit) to generate cytosolic (supernatant) and mitochondrial (pellet) fractions. For cytochrome c detection, 35 μg of the cytosolic fraction was separated by a 4% to 12% Bis-Tris precast gel (Invitrogen), transferred to a nitrocellulose membrane (Protran), and incubated with 1:500 dilution of mouse monoclonal anti-cytochrome c (BD Biosciences) followed by 1:2,000 dilution of horseradish peroxidase (HRP)–conjugated anti-mouse IgG (Santa Cruz Biotechnology). Antibody binding was visualized with chemiluminescent substrate (Pierce) and autoradiography film (Denville Scientific, Inc.), imaged with the VersaDoc 5000 MP Imaging System (Bio-Rad Laboratories) linked with Quantity One (version 4.6.6, Bio-Rad Laboratories) software. Cytochrome c release relative to vehicle control was normalized to β-actin expression. For proapoptotic protein expression cells were cells were lysed in radioimmunoprecipitation (RIPA) buffer (Upstate) containing 1 mmol/L phenylmethanesulphonylfluoride (PMSF, Sigma), and 1 μg/mL of protease inhibitor cocktail (Sigma) consisting of aprotinin, bestatin, leupeptin, and pepstatin. Lysates were incubated on ice for 30 minutes and sonicated briefly before centrifugation at 12,000 × g for 30 minutes. Protein quantification of the supernatants employed the BCA Protein Assay Kit (Pierce). Equal amounts of protein were resolved on 4% to 12% Bis-Tris precast gels (Invitrogen), transferred to a nitrocellulose membrane, and incubated with primary antibodies: rabbit polyclonal anti-cleaved caspase-9 at 1:1,000 dilution (Cell Signaling Technology), rabbit polyclonal anti-cleaved caspase-3 at 1:1,000 dilution (Cell Signaling Technology), rabbit polyclonal anti-PARP at 1:1,000 dilution (Santa Cruz Biotechnology), and mouse monoclonal anti–β-actin at 1:2,000 dilution (Santa Cruz Biotechnology), followed by 1:2,000 dilution of HRP-conjugated anti-mouse and anti-rabbit IgGs (Santa Cruz Biotechnology). Antibody binding was visualized with chemiluminescent substrate and autoradiography. Densitometry was used to evaluate changes in protein expression.

Six- to 8-week-old athymic(nu/nu) mice (The Jackson Laboratory) were injected subcutaneously between the shoulder blades with 12 to 15 million human neuroblastoma cells (SMS-KCNR, CHLA-90, CHLA-136, and CHLA-140; refs. 25–28) mixed in Matrigel Matrix HC (BD Biosciences). Xenografted mice were randomized into four treatment groups (5–6 mice/group): vehicle, 4-HPR/LXS, ABT-751, and 4-HPR/LXS + ABT-751. Drug treatment was begun when progressively growing tumors measured 100 to 200 mm³. 4-HPR/LXS (240 mg 4-HPR/kg/d; slurried in water) and ABT-751 (75 mg/kg/d; dissolved in 95% D5W, 0.6% HCL, and 3.9% ethanol) were administered by gavage, 4-HPR/LXS twice daily divided doses and ABT-751 as a single daily dose, five days per week. Control animals received powderized LYM-X-SORB matrix slurried in 96% D5W, 0.6% HCL, and 3.4% ethanol. Dosing schedules for other microtubule inhibitors were as follows: colchicine, 0.05 mg/kg/d in divided daily doses 5 days a week by oral gavage; vincristine, 0.75 mg/kg/day once a week by intraperitoneal injection; and paclitaxel, 10 mg/kg/d once a week by intraperitoneal injection. Tumor growth and mouse weight was assessed twice weekly by caliper measurement and tumor volumes were calculated as 0.5 × height × width × length (32). Mice were sacrificed by CO₂ narcosis when tumor volumes exceeded 1,500 mm³ or serious morbidity was observed. EFS was time from initial xenografting until death from any cause. Animals were housed and treated according to protocols approved by the Institutional Animal Care and Use Committee.

When xenograft tumors measured 300 to 500 mm³, 96-hour treatment of the mice was begun as described for xenograft testing. Mice were sacrificed 4 hours after the last treatment, tumors excised and fixed in formalin. Paraffin sections (5 μm) were stained for apoptosis via TUNEL. Ten high-power fields for each treatment group were scored for apoptotic cells.

To examine the effect of antioxidants on the in vivo apoptosis of 4-HPR/LXS + ABT-751, nu/nu mice (3/group) bearing 300 to 500 mm³ SMS-KCNR xenografts were randomized into four treatment groups: vehicle, vitamins C+E, 4-HPR/LXS + ABT-751, and 4-HPR/LXS + ABT-751 + vitamins C+E. Mice were pretreated for 2 days with 3 mg/d of both vitamins C and E that continued during 48-hours of 180 mg/kg/d 4-HPR/LXS + 75 mg/kg/d ABT-751. Mice were sacrificed 4 hours after the last treatment and tumors assessed for apoptosis by TUNEL.

Combination index (CIN) determined by CalcuSyn software (version 2.1, Biosoft), was used to assess drug synergism (33), based on the CIN values (calculated for each concentration level): CIN > 1.10, antagonism; CIN of 0.9 to 1.10, additive; CIN < 0.9, synergism. Statistical significance of differences in means was determined by the one-way ANOVA test using SigmaPlot software (version 11.0, Systat Software Inc.). Xenograft EFS used Kaplan–Meier log-rank analysis. All P values were two-sided and statistical significance was defined as P < 0.05.

We determined the cytotoxicity of 4-HPR, ABT-751, and fixed-ratio concentration combinations of both drugs, in nine human neuroblastoma cell lines using the DIMSCAN cytotoxicity assay. Dose–response curves to 4-HPR, ABT-751, and 4-HPR + ABT-751 in the neuroblastoma cell lines are shown in Fig. 1. ABT-751 was more active than 4-HPR in some lines (CHLA-15, CHLA-20, and FU-NB-2006), whereas 4-HPR was more active than ABT-751 in others (SMS-SAN, SMS-KCNR, CHLA140, and CHLA-119), but the combination was active (achieving at least 2 logs of cell kill) in all the lines, though often activity of the combination was not significantly higher than of 4-HPR as a single agent. Cell lines established at time of progressive disease after chemotherapy were as sensitive as CHLA-15 and SMS-SAN (established before therapy), and the activity of ABT-751 ± 4-HPR was more active in CHLA-20 (postchemotherapy) than in CHLA-15 (pre-therapy line from the same patient as CHLA-20). Combining 4-HPR and ABT-751 showed a modest additive effect in most of the cell lines with drug synergy apparent in only one cell line, CHLA-119. 4-HPR + ABT-751 exhibited synergistic cytotoxicity in CHLA-119 at two fixed-ratio concentrations (4-HPR 0.62 μmol/L, ABT-751 62.5 ng/mL; and 4-HPR 1.25 μmol/L, ABT-751 125 ng/mL) with CIN values of 0.73 and 0.76, respectively, achieving 2 to 4 logs of tumor cell kill. Combination indices for the nine cell lines tested with 4-HPR + ABT-751 are shown in Supplementary Table S1.

The in vivo activity of 4-HPR/LXS oral powder + ABT-751 was assessed in four subcutaneous human neuroblastoma xenograft models (CHLA-90, CHLA-136, CHLA-140, and SMS-KCNR), all established at time of progressive disease after chemotherapy or myeloablative chemoradiotherapy. Mice received 4-HPR/LXS and ABT-751, alone or in combination, by oral gavage 5 days/week. In contrast with the modest combination effect observed for most cell lines in vitro, activity of 4-HPR/LXS + ABT-751 was substantially greater than for either single agent in all four xenograft models (Fig. 2; individual mouse curves in Supplementary Fig. S1). In CHLA-90, SMS-KCNR, and CHLA-140 there were occasional mice treated with ABT-751 that maintained complete responses (MCR) >60 days. However, multiple MCR were observed in SMS-KCNR, CHLA-136, and CHLA-140 with 4-HPR/LXS + ABT-751. Mouse EFS for 4-HPR/LXS + ABT-751 was significantly greater than for either single agent for CHLA-90, SMS-KCNR, and CHLA-136 (P = 0.01), but not for CHLA-140 (Fig. 3). Median survival for all 4 xenograft models combined (n = 21 per cohort) was 28 days for the control cohort, 49 days for the 4-HPR–treated cohort, 77 days for the ABT-751–treated cohort, and >150 days for the 4-HPR/LXS + ABT-751–treated cohort (Fig. 3E), with the EFS of the combination significantly greater than either single agent (P < 0.001).

The enhanced cytotoxicity observed in vitro from combining 4-HPR + ABT-751 was less than the enhanced antitumor effect of the combination in vivo. A possible explanation for this could be that a metabolite of 4-HPR generated only in vivo was more effective at promoting cytotoxicity of ABT-751 than 4-HPR, and a 4-OH-fenretinide metabolite has been reported to have antimicrotubule activity (34). However, we tested this hypothesis by comparing the ability of 4-HPR and 4-oxo-fenretinide to synergize with ABT-751 in vitro (fixed ratio dose–response curves in both 20% and 2% oxygen using DIMSCAN) in the CHLA-119 and FU-NB-2006 cell lines and found no significant difference between 4-HPR and 4-OH-fenretinide in cytotoxic synergy with ABT-751 in vitro (data not shown).

As 4-HPR single-agent cytotoxicity is partially mediated through increased ROS levels (5, 35, 36), we examined the effects of antioxidants on the cytotoxicity of 4-HPR + ABT-751 in the CHLA-119 cell line, which showed the highest synergy in vitro between 4-HPR + ABT-751. Cells were pretreated for 3 hours with vitamin C, vitamin E, n-acetylcysteine (NAC), or sodium thiosulfate (STS) before addition of 4-HPR and ABT-751 (Fig. 4A). Vitamin C and vitamin E, but not NAC or STS, significantly diminished the cytotoxicity of 4-HPR and 4-HPR + ABT-751 (P < 0.001); antioxidants did not alter cytotoxicity of ABT-751. Dose–response curves of 4-HPR + ABT-751 ± antioxidants are shown in Fig. 4B. We repeated these experiments using higher concentrations of NAC (0.5–10 mmol/L) and STS (0.5–5 mmol/L), which (unlike vitamins C and E) failed to decrease the cytotoxicity of 4-HPR + ABT-751 (Supplementary Fig. S2).

CHLA-119 cells were treated with 4-HPR, ABT-751, or 4-HPR + ABT-751 for 20 hours (± antioxidants added 3 hours before 4-HPR) and then analyzed for apoptosis by TUNEL (flow cytometry; Fig. 4C). Apoptosis was seen in 11.5% of control cells (data not shown), 68.5% with 4-HPR, 37.7% with ABT-751, and 76.5% with 4-HPR + ABT-751. 4-HPR + ABT-751 somewhat increased apoptosis relative to single agent–treated cells, suggesting the enhanced cytotoxicity observed with the drug combination in vivo could be due to enhanced apoptosis. Vitamins C and E, but not NAC or STS, significantly decreased the percentage of apoptotic cells in 4-HPR- and combination-treated cells (Fig. 4C; P < 0.01). Of note, in cells treated with 4-HPR + ABT-751, vitamins C and E reduced apoptosis to the level of cells treated with ABT-751 alone. Antioxidants did not statistically alter the percentage of apoptotic cells in ABT-751–treated cells.

We determined whether combining 4-HPR with ABT-751 increased ROS relative to single agents. CHLA-119 cells were treated with 4-HPR ± ABT-751 for 3 hours and ROS levels were measured via flow cytometry using DCFDA. Mean DCFDA intensity values were: vehicle = 1,821; 4-HPR = 3,847; ABT-751 = 1,847; 4-HPR + ABT-751 = 4,432 for (Fig. 4D, left). DCFDA intensity in cells treated with 4-HPR + ABT-751 was approximately 15% greater than 4-HPR alone (P = 0.04) while mean DCFDA intensity of ABT-751-treated cells was not statistically different from controls. Vitamins C, E, and NAC significantly reduced the mean DCFDA intensity of 4-HPR + ABT-751–treated cells (Fig. 4D, right, P < 0.01). STS did not statistically alter the ROS level of 4-HPR + ABT-751–treated cells (P = 0.10). These results are roughly concordant with the observed effects of antioxidants on apoptosis induced in vitro (Fig. 4C).

We examined the effects of 4-HPR + ABT-751 (20-hour exposure; ± antioxidants) on mitochondrial membrane depolarization (JC-1 staining) and cytochrome c release in CHLA-119 cells. Mitochondrial depolarization was approximately 22% in vehicle control cells, approximately 46% with 4-HPR, approximately 39% with ABT-751, and approximately 57% with 4-HPR + ABT-751 (Supplementary Fig. S3A). Vitamins C and E significantly decreased depolarization in 4-HPR–treated cells to approximately 25% and 26% respectively, and in 4-HPR + ABT-751–treated cells to approximately 35% and 31%, respectively (P < 0.01). NAC or STS did not statistically alter mitochondrial membrane depolarization.

Cytochrome c release from mitochondria into the cytosol (immunoblotting) was assessed in CHLA-119 cells treated with 4-HPR, ABT-751, and 4-HPR + ABT-751 for 24 hours ± vitamin E (Supplementary Fig. S3B). Densitometry values (normalized to β-actin, vehicle control set to 1.00) for cells treated with vitamin E = 0.9, 4-HPR = 2.0, ABT-751 = 3.5, 4-HPR + ABT-751 = 7.9, and 4-HPR + ABT-751 + vitamin E = 3.7. Cytochrome c release was increased in 4-HPR– and ABT-751–treated cells, and to a greater degree in cells treated with both drugs. The addition of vitamin E to 4-HPR + ABT-751–treated cells decreased cytochrome c release into the cytosol.

Caspases-9 and -3 are major executors of the mitochondrial apoptotic cascade (Supplementary Fig. S3B; ref. 37). Densitometry values (normalized to β-actin, vehicle control set to 1.00) for cleaved caspase-9 for cells treated with vitamin E = 0.7, 4-HPR = 1.8, ABT-751 = 2.9, 4-HPR + ABT-751 = 4.8, 4-HPR + ABT-751 + vitamin E = 1.4. For cleaved caspase-3: vitamin E = 1.2, 4-HPR = 4.8, ABT-751 = 14.0, 4-HPR + ABT-751 = 23.3, 4-HPR + ABT-751 + vitamin E = 6.5. For cleaved-PARP: vitamin E = 0.8, 4-HPR = 2, ABT-751 = 4.2, 4-HPR + ABT-751 = 5.7, 4-HPR + ABT-751 + vitamin E = 2.7. Expression of cleaved-caspase-9, cleaved-caspase-3, and cleaved-PARP (at 85 kDa) was increased in 4-HPR- and ABT-751–treated cells and to a greater extent in cells treated with 4-HPR + ABT-751. Similar to the cytochrome c release data, cotreatment of the drug combination with vitamin E reduced expression of cleaved caspases and cleaved-PARP.

We assessed apoptosis in SMS-KCNR xenografts from mice treated for 96 hours with control vehicle, 4-HPR/LXS, ABT-751, or 4-HPR/LXS + ABT-751. The mice were sacrificed and tumors stained for apoptosis via TUNEL. The average numbers of apoptotic cells in 10 high power fields of tumors were: control = approximately 8, 4-HPR = approximately 27, ABT-751 = approximately 30 and 4-HPR + ABT-751 = approximately 74 (Fig. 5A). The enhanced apoptosis from combining 4-HPR + ABT-751 was observed in neuroblastoma tumor cells and not in tumor stroma. Apoptosis induced by 4-HPR/LXS + ABT-751 was significantly higher than in control and single-agent–treated tumors (P < 0.01). To assess the involvement of ROS in the synergy of 4-HPR/LXS + ABT-751, mice bearing SMS-KCNR xenografts were pretreated for 2 days with vitamins C + E and throughout 48-hour treatment with 4-HPR/LXS + ABT-751. The average number of apoptotic cells in 10 high-power fields of tumors were: vehicle = approximately 8, vehicle + vitamins C + E = approximately 12, 4-HPR + ABT-751 = approximately 64, and 4-HPR + ABT-751 + vitamins C + E = approximately 27 (Fig. 5B). Vitamin C + E treatment significantly reduced apoptosis from 4-HPR/LXS + ABT-751 (P < 0.01).

We compared drug synergy between 4-HPR and five microtubule inhibitors (ABT-751, colchicine, vincristine, vinorelbine ditartrate, and paclitaxel) in SMS-KCNR, CHLA-119, and CHLA-90 (Supplementary Fig. S4). Drug synergy was also observed in SMS-KCNR when 4-HPR was combined with colchicine (95% CI, 0.69) or vincristine (95% CI, 0.71) and in CHLA-140 when 4-HPR was combined with colchicine (95% CI, 0.76), vincristine (95% CI, 0.57), or paclitaxel (95% CI, 0.75). Both vitamin E and vitamin C decreased the synergistic toxicity observed in vitro between 4-HPR and all 5 tested microtubule inhibitors (Supplementary Fig. S5). The IC₉₀ values and combination indices for the 5 tested microtubule inhibitors combined with 4-HPR in 7 neuroblastoma cell lines are shown in Supplementary Table S2.

The in vivo efficacy of 4-HPR/LYM-X-SORB oral powder combined with four microtubule inhibitors (ABT-751, vincristine, paclitaxel, and colchicine) was assessed in SMS-KCNR xenograft models (Fig. 6). The xenografts received the indicated doses of 4-HPR/LXS and microtubule inhibitors, alone or in combination, until treatment endpoint or morbidity. 4-HPR/LXS + ABT-751 was highly active relative to vehicle or single-agent treatments (Fig. 6A). Combining 4-HPR/LXS with three other microtubule inhibitors, vincristine, paclitaxel, or colchicine, was less active than what we observed with ABT-751. Treatment with vincristine alone was only slightly better than vehicle and 4-HPR/LXS treatment (Fig. 6B); single-agent paclitaxel (Fig. 6C) or colchicine (Fig. 6D) displayed no activity over vehicle-only treatment. Combining 4-HPR/LXS with vincristine and paclitaxel was more effective in suppressing tumor growth than the single-agent treatments; however, the combinations did not achieve the level of activity observed with the 4-HPR/LXS + ABT-751. Colchicine + 4-HPR/LXS was not more active than 4-HPR/LXS alone.

Through a series of early-phase and then randomized trials, intensive induction chemotherapy, local control with surgery and radiation, consolidation with myeloablative therapy, and post-consolidation maintenance therapy with 13-cis-retinoic acid + ch14.18 antibody + cytokines has been defined as the current optimal therapy for patients with high-risk neuroblastoma (2, 4, 19). However, even with optimal therapy about one half of patients will develop progressive disease that is almost always less responsive to therapy than disease before progression and progressive disease on or after therapy is often fatal for high-risk neuroblastoma. Thus, there is a critical need for new drugs and novel drug combinations that are active against recurrent and refractory neuroblastoma.

In preclinical and clinical studies, the orally available colchicine-binding site microtubule inhibitor, ABT-751, was well tolerated and showed signals of activity against recurrent neuroblastoma (14–17), but in a neuroblastoma phase II clinical trial, ABT-751 achieved few objective responses and did not increase time-to-progression over historical controls (18). Fenretinide also showed activity against recurrent neuroblastoma in preclinical studies (5, 6, 10), signals of activity in early-phase trials of a suboptimal capsule formulation (7–9), and multiple complete responses in a phase I trial of a novel lipid matrix oral powder formulation (4-HPR/LXS) that increased 4-HPR exposures (11). On the basis of known mechanisms of action, we postulated that 4-HPR and ABT-751 would demonstrate additive antitumor activity in vivo, but it was not anticipated that 4-HPR + ABT-751 would be synergistic, per se. Indeed, our in vitro testing performed after pilot xenograft experiments evidenced a high activity of the 4-HPR + ABT-751 combination indicated only modest additive to synergistic activity in most lines and high synergy in only a single cell line. However, testing in multiple neuroblastoma xenograft models confirmed that the two drugs together achieved a striking activity (with multiple maintained complete responses) not observed with either single agent. Combining 4-HPR/LXS + ABT-751 significantly prolonged mouse survival relative to control and single-agent treatments with the majority of mice treated with the combination surviving progression-free ≥100 days. The precise mechanism(s) of the discrepancy in results between these in vitro and in vivo models awaits future elucidation, but our data excluded metabolism to 4-OH-fenretinide (34) and cytotoxicity for tumor stroma as potential mechanisms.

It has been well documented that mechanisms of 4-HPR cytotoxicity in cell lines from multiple cancer types in vitro can involve the induction of ROS (5, 36, 38). Thus, we examined the effect of antioxidants on the synergistic cytotoxicity between 4-HPR and ABT-751, and we found that the addition of vitamins C or E abrogated the enhanced cytotoxicity seen with 4-HPR, 4-HPR + ABT-751, but not with ABT-751 alone. Interestingly, thiol antioxidants (sodium thiosulfate and n-acetylcysteine) did not antagonize the cytotoxicity of 4-HPR + ABT-751, even though n-acetylcysteine significantly decreased ROS in cells treated with 4-HPR + ABT-751. We speculate that the differential effects of these antioxidants may be due to their differing cellular compartmentalization. Vitamins C and E, two naturally occurring antioxidants, are capable of entering the mitochondria (39–41), where the mitochondrial respiratory chain is a major source of intracellular ROS generation, as well as, an important target of ROS damage (42). However, N-acetylcysteine and sodium thiosulfate, two thiol antioxidants that aid in the replenishment of glutathione (43, 44), have not been shown to enter mitochondria.

As single agents, both 4-HPR and ABT-751 induced apoptosis, and we observed enhanced apoptosis in vitro and in vivo when combining the two agents. In cell lines in vitro, 4-HPR + ABT-751 resulted in increased ROS, activation of the mitochondrial apoptotic pathway, and greater mitochondrial membrane depolarization, cytochrome c release, caspase activation, and apoptosis than either agent alone. These enhanced effects of 4-HPR + ABT-751 were blocked by mitochondrial-penetrant antioxidants in vitro and in vivo. Thus, our data suggest that 4-HPR-induced an increase of ROS, and the activity of 4-HPR ± ABT-751 that promotes apoptosis, may be occurring in mitochondria. As 4-HPR/LXS + ABT-751 was very well tolerated in mice where complete responses were achieved in tumor xenografts, the effect of the combination on enhancing cell death mechanisms is largely restricted to malignant cells. Our data indicate a novel mechanism of action for fenretinide, enhancing activity of antimicrotubule via the generation of ROS in selective cellular compartments of malignant cells.

We compared activity of 4-HPR in combination with other microtubule disruptors (vincristine, vinorelbine, and colchicine) and the microtubule-stabilizing taxane paclitaxel. Like ABT-751, the other 4 microtubule inhibitors showed additive or synergistic cytotoxicity in vitro that was mediated by 4-HPR–generated ROS. However, in xenograft testing activity of vincristine or paclitaxel was much less than observed with ABT-751, consistent with previously reported data for the single-agent activity of vincristine or taxanes against neuroblastoma xenografts (45, 46). Colchicine + 4-HPR/LXS showed no increase in activity against xenografts compared with 4-HPR alone, likely due to the limited doses required in mice for colchicine due to systemic toxicity.

In summary, these data demonstrate that: (i) combining 4-HPR with ABT-751 increased activation of the mitochondrial apoptotic cascade in vitro; (ii) the increases in cytotoxicity and apoptosis of the 4-HPR + ABT-751 combination were mediated by increased ROS, likely of mitochondrial origin, as activity was suppressed by vitamin C and vitamin E and not thiol antioxidants; and (iii) the combination of 4-HPR/LXS + ABT-751 was highly active against several multidrug-resistant, recurrent neuroblastoma xenograft models in immunodeficient mice. Early-phase trials of 4-HPR/LXS and ABT-751 as single agents have demonstrated drug exposures as high or higher than those obtained in mice (8, 9, 11, 47), that both 4-HPR/LXS and ABT-751 are well tolerated as single agents (8, 9, 11, 15, 16, 18), and that multiple complete responses in recurrent high-risk neuroblastoma have been observed with 4-HPR/LXS (11). Thus, together, the preclinical and clinical data support that the combination of 4-HPR/LXS and ABT-751 warrants clinical investigation in recurrent and refractory high-risk neuroblastoma.

B.J. Maurer is co-founder and CMO of, reports receiving a commercial research grant from, has ownership interest (including patents) in, and is a consultant/advisory board member for CerRx Inc. C.P. Reynolds is Chief Scientific Officer and has ownership interest (including patents) in CerRx Inc. and patents owned by CHLA. No potential conflicts of interest were disclosed by the other authors.

Conception and design: C.P. Reynolds

Development of methodology: B.J. Maurer, C.P. Reynolds

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): V. Khankaldyyan, H. Shimada, M.M. Song, C.P. Reynolds

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.M. Song, B.J. Maurer, C.P. Reynolds

Writing, review, and/or revision of the manuscript: H. Shimada, C.P. Reynolds

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N.E. Chen, V. Maldonado, V. Khankaldyyan, C.P. Reynolds

Study supervision: C.P. Reynolds

The authors thank the following for providing materials used in this article: The Developmental Therapeutics Program of the National Cancer Institute (NCI, Bethesda, MD) for providing fenretinide drug substance and the formulated fenretinide LYM-X-SORB powder; Abbot Laboratories for providing ABT-751; and The Children's Oncology Group Cell Line and Xenograft Repository (www.COGcell.org) for providing cell lines.

This study was supported by NCI grants CA82830, CA81403, and Cancer Prevention and Research Institute of Texas grant RP100762 (to C.P. Reynolds).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.