Neuroblastoma, the most common solid tumor of infancy derived from the sympathetic nervous system, continues to present a formidable clinical challenge. Sterically stabilized immunoliposomes (SIL) have been shown to enhance the selective localization of entrapped drugs to solid tumors, with improvements in therapeutic indices. We showed that SIL loaded with doxorubicin (DXR) and targeted to the disialoganglioside receptor GD2 [aGD2-SIL(DXR)] led to a selective inhibition of the metastatic growth of experimental models of human neuroblastoma. By coupling NGR peptides that target the angiogenic endothelial cell marker aminopeptidase N to the surface of DXR-loaded liposomes [NGR-SL(DXR)], we obtained tumor regression, pronounced destruction of the tumor vasculature, and prolonged survival of orthotopic neuroblastoma xenografts. Here, we showed good liposome stability, long circulation times, and enhanced time-dependent tumor accumulation of both the carrier and the drug. Antivascular effects against animal models of lung and ovarian cancer were shown for formulations of NGR-SL(DXR). In the chick embryo chorioallantoic assay, NGR-SL(DXR) substantially reduced the angiogenic potential of various neuroblastoma xenografts, with synergistic inhibition observed for the combination of NGR-SL(DXR) with aGD2-SIL(DXR). A significant improvement in antitumor effects was seen in neuroblastoma-bearing animal models when treated with the combined formulations compared with control mice or mice treated with either tumor- or vascular-targeted liposomal formulations, administered separately. The combined treatment resulted in a dramatic inhibition of tumor endothelial cell density. Long-term survivors were obtained only in animals treated with the combined tumor- and vascular-targeted formulations, confirming the pivotal role of combination therapies in treating aggressive metastatic neuroblastoma. (Cancer Res 2006; 66(20): 10073-82)

The dependence of tumor growth on the development of a neovasculature is a well-established concept in cancer biology. In 1971, Judah Folkman proposed that the progression of cancer might be inhibited by preventing tumors from recruiting new blood vessels (1). This process, termed angiogenesis, is important for supplying oxygen, nutrients, growth factors, hormones, and proteolytic enzymes which control the coagulation and fibrinolytic systems, as well as the dissemination of tumor cells to distal sites (2). In the last several years, much effort has been directed towards the development of antiangiogenic agents that block this process. In this regard, intimate contact with the blood makes the tumor vasculature endothelial cell a uniquely accessible target within the tumor. Decades of research investigating the molecular basis of angiogenesis has identified several proteins in solid tumor–associated angiogenic vessels that are absent or barely detectable in established blood vessels (3, 4), including the αv integrins, receptors for angiogenic growth factors, and other types of membrane-spanning molecules, such as aminopeptidase N (CD13) and A (57). Moreover, in vivo screening of peptide-phage libraries has proven to be useful for the discovery of additional ligands that selectively home to tumor vessels (8). One of the tumor-targeting ligands thus far identified is a peptide containing the NGR motif, which recognizes a tumor-specific isoform of CD13. This peptide, when coupled to anticancer compounds, was capable of delivering the therapeutic molecule specifically to tumor vessels (810). We recently showed that liposomal formulations of DXR, bearing the NGR peptide at the outer surface, led to dramatic vascular damage and antitumor effects against an orthotopic model of human neuroblastoma (11).

Unlike antiangiogenic agents, which prevent new blood vessel formation from existing vessels, vascular targeting agents, such as ligand-targeted liposomes and drug conjugates, work by destroying established tumor vasculature, thereby starving the tumor cells of oxygen and nutrients. It has been shown that the combination of vascular targeting agents with antitumor agents or angiogenesis inhibitors can lead to additive or synergistic activity in experimental solid tumors (12, 13). However, traditional chemotherapeutic antitumor agents kill not only cancer cells, but all rapidly growing cells in the body, such as blood and hair cells and cells lining the intestine. This leads to the distressing side effects of chemotherapy, and imposes practical limits on the drug dose and dosing frequency. In the last few years, liposome-encapsulated therapeutic agents have been used to increase the selective toxicity of chemotherapeutics in cancer, resulting in improved therapeutic outcome and/or minimized damage to normal tissues such as heart or bone marrow (14, 15). Moreover, ligand-targeted liposomal drugs (immunoliposomes) that promote selective binding and internalization of the liposomes into target cells have shown enhanced antitumor effects relative to nontargeted liposomal drugs (15). Among tumor-associated antigens, the disialoganglioside GD2 is widely expressed on cancer cells of neuronal origin, including neuroblastoma, and at low levels, in cerebellum and peripheral nerves (16). Thus, it has been considered an excellent targeting moiety in applications involving liposome-entrapped drugs (1720).

Treatment of neuroblastoma, the second most common solid tumor in childhood, is successful in less than half of patients with high-risk disease (21). The presence of circulating neuroblastoma cells in the blood and micrometastases in the bone marrow post-primary surgery is a strong predictor of relapse (22). Furthermore, a high vascular index in neuroblastoma correlates with poor prognosis (23), suggesting the dependence of aggressive tumor growth on active angiogenesis. The matrix-degrading enzymes, matrix metalloproteinase-2 and matrix metalloproteinase-9, are elevated in high-risk neuroblastoma compared with low-risk disease (24) and are expressed in neuroblastoma cell lines, which themselves induce the proliferation of endothelial cells (25), supporting the angiogenic phenotype of this tumor (23). Thus, the use of vascular targeting approaches in combination with tumor targeting therapies could potentially improve outcome in children with neuroblastoma and may warrant further study.

The primary goal of this study was to achieve proof-of-principle for the hypothesis that combined administration of liposomal formulations of DXR targeted against tumor cells via anti-GD2 monoclonal antibodies (mAb) and against the tumor vasculature via the NGR peptide will have improved therapeutic effects relative to each therapy used individually. This hypothesis was tested in two highly aggressive pseudometastatic neuroblastoma animal models. Our results clearly show that liposomes administered in a sequential manner were statistically more effective in inhibiting neuroblastoma tumor proliferation in mice compared with formulations given alone. Additive or synergistic antitumor and antivascular treatments could be an excellent novel strategy and may deserve clinical evaluation as an adjuvant therapy for advanced stage neuroblastoma disease, for the treatment of disease resulting from incomplete surgery, or for the eradication of early micrometastic lesions.

Liposome preparation and DXR loading. Tumor- and vascular-targeted liposomes were prepared as previously reported (11, 20, 26).

Cell lines and animal models. Three neuroblastoma cell lines, HTLA-230, SH-SY5Y, and NXS2, positive for GD2 antigen expression (17, 20, 27), the ovarian cancer cell line, OVCAR-3 (28), and the lung cancer cell line, Colo-699N (29), were grown in complete RPMI 1640, as previously described (30).

All animals were purchased from Harlan Laboratories (Harlan Italy, San Pietro al Natisone, Udine, Italy) and experiments involving animals were reviewed and approved by the licensing and ethical committee of the National Cancer Research Institute and by the Italian Ministry of Health. For therapeutic studies, pseudometastatic neuroblastoma animal models were done by injecting 5 × 104 NXS2 or 4 × 106 HTLA-230 cells, in 200 μL of HEPES-buffered saline, into the tail vein (i.v.) of 4-week-old athymic nude (nu/nu) mice.

For pharmacokinetic and biodistribution studies, an orthotopic neuroblastoma animal model was used. Five-week-old severe combined immunodeficiency (SCID) mice were anaesthetized and injected, after laparatomy, with the SH-SY5Y cell line (2.5 × 106 cells in 20 μL of HEPES-buffered saline) in the capsule of the left adrenal gland, as previously described (11).

Pharmacokinetic and biodistribution experiments. SCID mice bearing orthotopically implanted neuroblastoma tumor (mean tumor volume of ∼100 mm3), were injected via the tail vein with a single dose of dual-labeled liposomes (0.5 μmol PL/mouse), with or without coupled NGR or ARA peptides, containing ∼3 × 105 cpm of the lipid tracer [3H]CHE and 3 × 105 of the drug tracer [14C]DXR. At selected time points (2, 12, 24, and 48 hours) post-injection, mice (three mice/group) were anaesthetized and sacrificed by cervical dislocation. A blood sample (100 μL) was collected by heart puncture and counted for the [3H]- and [14C]-labels in a Packard beta-counter. Blood correction factors were applied to all samples (11). Pharmacokinetic data are expressed as the percentage of the injected dose remaining in blood at the various time points post-injection. Biodistribution was determined as previously described (11). Data are expressed as the percentage of injected dose/g of tissue.

In vivo therapeutic studies. For tumor- and vascular-targeted therapeutic studies, two highly aggressive pseudometastatic neuroblastoma animal models were established by injecting tumor cells in the tail vein of mice. In the first model, athymic nude mice (eight mice/group) were injected i.v. with NXS2 cells on day 0 and treated once at day 1 after cell inoculation with 8 mg/kg of DXR, as a free drug, encapsulated in nontargeted liposomes [SL(DXR)] or encapsulated in anti-GD2-targeted liposomes [aGD2-SIL(DXR)]. At 8 days after cell inoculation, a different group of tumor-bearing mice was treated with 6 mg/kg of DXR, once a week for a total of three injections, with DXR encapsulated in vascular-targeted liposomes [NGR-SL(DXR)]. As a control, some mice were treated with DXR encapsulated in liposomes coupled with a nonspecific peptide [ARA-SL(DXR)]. Another group of animals was injected at 8 days after cell inoculation with a lower DXR concentration of NGR-SL(DXR) for a longer period of time (1.5 mg/kg twice weekly for 4 weeks total). This schedule of drug administration will be addressed as “low dose-long time” (LDLT). In these animals, treatment was delayed until 8 days post-inoculation to allow the tumor vasculature to form. Combined experiments were done via the injection of both tumor- and vascular-targeted liposomal formulations, following the same time schedules mentioned above, and at half the DXR dose per administration, so that the total DXR dose was the same in all treatment groups.

On the basis of our previous results (20), in the second pseudometastatic neuroblastoma model, HTLA-230 cells were injected i.v. on day 0 and mice (eight mice/group) treated with 8 mg/kg of DXR encapsulated in Fab′ fragments of GD2-targeted liposomes [Fab′-SIL(DXR)], on days 1, 5, or 10 after tumor cell inoculation, once a week for 2 weeks. This experiment was done to study the effectiveness of the tumor-targeted liposomes as a function of time/tumor volume. Specifically, one group of tumor-bearing mice (12 mice/group) was treated once a week for 2 weeks (at 5 days after tumor inoculation) with 8 mg/kg of DXR encapsulated in Fab′ fragments of GD2-targeted liposomes. A second group was treated at 19 days after cell inoculation, once a week for 2 weeks, with 8 mg/kg of DXR encapsulated in vascular-targeted liposomes [NGR-SL(DXR)]. Combined experiments were done in a third group of animals, using tumor- and vascular-targeted liposomal formulations, administered in a sequential manner, following the time schedules mentioned above, at half the dose of DXR for each formulation. In all experiments, control mice were treated with a weekly injection of HEPES-buffered saline.

In all experiments, mice were monitored routinely for weight loss and sacrificed when signs of poor health became evident. Survival times were used as the main criterion for determining treatment efficacy.

Chorioallantoic membrane assay. Chorioallantoic membrane (CAM) experiments have been done according to Ribatti et al. (31). Biopsy fragments, 1 to 2 mm3, from xenografts derived from SH-SY5Y cells injected in SCID mice were then grafted onto the CAM either alone or together with 1 to 2 μg of DXR encapsulated in SL or NGR-SL or ARA-SL. Alternatively, biopsy fragments, 1 to 2 mm3, from xenografts derived from HTLA-230 cells injected in nude mice were grafted onto the CAM either alone or together with 1 to 2 μg of DXR encapsulated in SL or Fab′-SIL or NGR-SL. CAMs were examined daily until day 12 and photographed in ovo with a stereomicroscope equipped with a camera and image analyzer system (Olympus Italia, Italy).

At day 12, the angiogenic response was evaluated by the image analyzer system as the number of vessels converging toward the sponges or the grafts. CAMs were then processed for light microscopy, as previously reported (31). Microvessel density was expressed as the percentage of the total number of intersection points occupied by vessels cut transversely (diameter, 3-10 μmol/L). Mean values ± SD were determined for each analysis.

Histologic analysis. Histologic evaluation of cryopreserved metastatic tissues was done at 30 or 50 days from the beginning of the treatment of mice bearing NXS2 or HTLA-230, respectively. Briefly, tumor-bearing mice were anesthetized with halothane and euthanized by cervical dislocation, and tumor samples were fixed in 4% paraformaldehyde for 5 hours at room temperature, washed twice in PBS, dehydrated in 30% sucrose for 12 hours at 4°C, embedded in Tissue-Tec optimum cutting temperature compound, snap-frozen, and stored at −80°C until use. Cryopreserved tissue sections (5 μmol/L) were examined after staining with Mayer's H&E (Sigma, St. Louis, MO).

MAb against neuroblastoma cells (mouse anti-human neuroblastoma, clone NB84a; Dako, Glostrup, Denmark), the Ki-67 proliferation antigen (mouse anti-human Ki-67, clone Ki-55; Dako), and the CD31 (goat anti-mouse, clone SC-1506; Santa Cruz Biotechnology, Santa Cruz, CA), and terminal deoxynucleotidyl transferase–mediated end labeling (TUNEL) staining were used as previously described (11, 32).

Determination of microvessel area. Two investigators with a computerized image analysis system (Quantimet 5000; Leica, Wetzlar, Germany) simultaneously assessed microvessel area, as previously described (11, 32).

Statistical methods. All the in vivo experiments were done at least twice with similar results. Results are expressed as the mean ± 95% confidence intervals (95% CI). The statistical significance of differential findings between experimental groups and controls was determined by Student's t test with Welch's correction or by ANOVA with Tukey's multiple comparison test in GraphPad Prism. The significance of the differences between experimental groups (n = 6-12 mice/group) in the survival experiments was determined by Kaplan-Meier curves by the use of Peto's log-rank test (Stats Direct statistical software; CamCode, Ashwell, United Kingdom). These findings were considered significant at P < 0.05.

Pharmacokinetic and biodistribution profiles of dual labeled NGR vascular-targeted liposomes. As previously shown, long circulation times are required for small liposomes to gain access to tumor sites (14). In order to quantify the fate of both the liposome and the drug, pharmacokinetic and biodistribution studies were done using dual-labeled liposomes. Specifically, [3H]-CHE-labeled Stealth liposomes [3H-SL(DXR)], vascular-targeted liposomes [NGR-3H-SL(DXR)], or mismatched peptide-targeted liposomes [ARA-3H-SL(DXR)], loaded with [14C]-labeled DXR, were evaluated in a biologically relevant human orthotopic neuroblastoma (SH-SY5Y cell line) xenograft model (11). The results of the pharmacokinetic studies are expressed as a percentage of the administered dose of lipid (Fig. 1A) and DXR (Fig. 1B) remaining in the blood. These findings clearly indicate that liposomes coupled to NGR peptide have good stability and long circulation times. They are removed only slightly more rapidly than the nontargeted formulation, with ∼30% of both drug and carrier remaining in the blood 24 hours after liposome inoculation (Fig. 1A and B).

Figure 1.

Pharmacokinetics and biodistribution of liposomal DXR in a murine model of neuroblastoma. SCID mice were orthotopically implanted with the SH-SY5Y cell line. Liposomes were dual-labeled with the lipid tracer [3H]CHE and the drug tracer [14C]DXR and injected i.v. in a single bolus dose (0.5 μmol PL/mouse). Liposomes were nontargeted (SL), vascular-targeted (NGR-SL), or targeted via a mismatched peptide (ARA-SL). Treatment groups consisted of 3H-SL[14C-DXR] (▪ for pharmacokinetics; gray columns for biodistribution), ARA-3H-SL[14C-DXR] (▴ for pharmacokinetics; white columns for biodistribution) and NGR-3H-SL[14C-DXR] (▾ for pharmacokinetics, black columns for biodistribution). At selected time points (2, 12, 24, and 48 hours) post-injection, blood and organs were counted for 3H (A and C-G, respectively) and 14C (B and H-L, respectively) labels in a Packard beta-counter. Points, average of three mice; bars, ±SD.

Figure 1.

Pharmacokinetics and biodistribution of liposomal DXR in a murine model of neuroblastoma. SCID mice were orthotopically implanted with the SH-SY5Y cell line. Liposomes were dual-labeled with the lipid tracer [3H]CHE and the drug tracer [14C]DXR and injected i.v. in a single bolus dose (0.5 μmol PL/mouse). Liposomes were nontargeted (SL), vascular-targeted (NGR-SL), or targeted via a mismatched peptide (ARA-SL). Treatment groups consisted of 3H-SL[14C-DXR] (▪ for pharmacokinetics; gray columns for biodistribution), ARA-3H-SL[14C-DXR] (▴ for pharmacokinetics; white columns for biodistribution) and NGR-3H-SL[14C-DXR] (▾ for pharmacokinetics, black columns for biodistribution). At selected time points (2, 12, 24, and 48 hours) post-injection, blood and organs were counted for 3H (A and C-G, respectively) and 14C (B and H-L, respectively) labels in a Packard beta-counter. Points, average of three mice; bars, ±SD.

Close modal

The results of the biodistribution studies are expressed as a percentage of the injected doses of both lipid and DXR. Tumor uptake of NGR-SL(DXR) was time-dependent, and was ∼2- to 3-fold higher than that of SL(DXR) after 24 and 48 hours, respectively (Fig. 1C and H). Five percent of both the liposomes and the drug had localized to the tumor by 24 hours post-injection and this had increased to 13% and 8% of the liposomes and the drug, respectively, by 48 hours post-injection. The difference between the two may be due to drug release over the 48-hour time point. The nonspecific ARA-targeted formulation showed negligible accumulation within the tumor (Fig. 1C and H). The biodistributions of DXR-encapsulated liposomes were evaluated in the liver, spleen, kidney, and heart, with the latter being an important indicator of doxorubicin-mediated cardiotoxicity (33). At all time points, the accumulation of DXR or liposomes in the heart was low (Fig. 1D and I), confirming that the NGR-targeted liposomes minimized nonspecific heart uptake.

NGR-SL(DXR) inhibits neuroblastoma-induced angiogenic responses. The antiangiogenic activity of NGR-targeted liposomes in vivo using the CAM assay is shown in Fig. 2. Tumor xenografts derived from SCID mice injected orthotopically with SH-SY5Y cells were grafted onto the CAM. CAMs treated with PBS (A) or with 1 to 2 μg of DXR, encapsulated in nontargeted (B), or in ARA-targeted (C) liposomes, showed numerous allantoic vessels radiating in a “spoked wheel” pattern towards the neuroblastoma xenografts. However, incubation of the CAM with NGR-SL(DXR) at the same drug concentration, substantially reduced the number of radiating vessels that invaded the implant (D).

Figure 2.

Effect of NGR-SL(DXR) on angiogenesis in vivo in the CAM model. Macroscopic observations of the angiogenic response induced by tumor xenografts derived from SCID mice injected orthotopically with SH-SY5Y cells and treated with PBS, as control (A). Numerous allantoic vessels converge radially in a spoked wheel pattern toward the implant. Similar features are recognizable when CAM are treated with 1 to 2 μg of DXR encapsulated in nontargeted (SL) (B) or in mismatched peptide-targeted (ARA-SL) liposomes (C). Following treatment with the same concentration of DXR loaded into vascular-targeted liposomes [NGR-SL(DXR)], fewer vessels surrounding the sponges invaded the xenograft (D). Original magnification, ×50.

Figure 2.

Effect of NGR-SL(DXR) on angiogenesis in vivo in the CAM model. Macroscopic observations of the angiogenic response induced by tumor xenografts derived from SCID mice injected orthotopically with SH-SY5Y cells and treated with PBS, as control (A). Numerous allantoic vessels converge radially in a spoked wheel pattern toward the implant. Similar features are recognizable when CAM are treated with 1 to 2 μg of DXR encapsulated in nontargeted (SL) (B) or in mismatched peptide-targeted (ARA-SL) liposomes (C). Following treatment with the same concentration of DXR loaded into vascular-targeted liposomes [NGR-SL(DXR)], fewer vessels surrounding the sponges invaded the xenograft (D). Original magnification, ×50.

Close modal

Therapeutic effects of combinations of tumor- and vascular-targeted liposomal doxorubicin. In order to evaluate if the combination of tumor- and vascular-targeted therapies would further inhibit neuroblastoma cell survival in vivo, we used two recently developed pseudometastatic experimental neuroblastoma animal models (20, 27). The neuroblastoma cell lines, NXS2 and HTLA-230, were chosen for our investigation into the therapeutic efficacy of combined therapies because, when injected i.v. in nude mice, these xenograft models mimic the metastatic spread observed in patients with advanced stage neuroblastoma (17, 20).

The NXS2 pseudometastatic neuroblastoma model resulted in a very aggressive behavior, with control mice dying from metastatic disease within 35 days from tumor cell inoculation. Nude mice were treated 24 hours post-inoculation, with 8 mg DXR/kg × 1, as free drug (free-DXR), or encapsulated in nontargeted liposomes [SL(DXR)], or in tumor-targeted liposomes [aGD2-SIL(DXR)]. aGD2-SIL(DXR) inhibited the growth of NXS2 tumor (Fig. 3A) with a significant increase in life span compared with control mice (P = 0.0047) or mice treated with DXR encapsulated in nontargeted liposomes (P = 0.0054). In the same set of experiments, different, randomly chosen mice were treated at 8 days post-inoculation with 6 mg/kg × 3 (total dose 18 mg DXR/kg) of liposomal DXR targeted to the vasculature via the NGR peptide. Mice receiving LDLT treatment received 1.5 mg/kg twice weekly for 4 weeks (total dose 12 mg DXR/kg). Delayed treatment at 8 days was with the intention of allowing the cells to establish metastatic foci in various organs and to start angiogenesis. NGR-SL(DXR), at both the higher dose and the LDLT drug concentration, resulted in a significant inhibition of tumor growth, compared with control mice (P = 0.0157 and P = 0.0112, respectively) and mice treated with a nonspecific peptide-targeted formulation (ARA-SL(DXR); Fig. 3B; P = 0.024 and P = 0.0166, respectively). There was no significant difference between the higher dose and the LDLT treatments. Remarkably, giving the tumor-targeted therapy at 24 hours, and then the vascular-targeted formulation at 8 days, further improved the life span in NXS2 tumor–bearing mice, compared with those treated with either liposomal formulation administered alone (Fig. 3C; aGD2-SIL(DXR) + NGR-SL(DXR) versus aGD2-SIL(DXR), P = 0,047; aGD2-SIL(DXR) + NGR-SL(DXR) versus NGR-SL(DXR), P = 0,0118; aGD2-SIL(DXR) + NGR-SL(DXR), LDLT versus aGD2-SIL(DXR), P = 0,0167; aGD2-SIL(DXR) + NGR-SL(DXR), LDLT versus NGR-SL(DXR), P = 0,0056). The higher dose combined therapy was not significantly different from the combined LDLT therapy. Of note, the Kaplan-Meier plots of the therapeutic results obtained in this neuroblastoma model derive from the same survival experiment, but for a complete understanding of the results, survival graphs have been presented in three separate panels (Fig. 3A-C).

Figure 3.

Effect of in vivo combination treatment of neuroblastoma-bearing mice with tumor- and vascular-targeted therapies. A to C, nude mice were inoculated i.v. with NXS2 neuroblastoma cells, and randomly assigned to groups of eight animals. A, at 24 hours post-inoculation, mice received HEPES-buffered saline, as control, or 8 mg DXR/kg/wk, as free drug or encapsulated in nontargeted [SL(DXR)] or aGD2-targeted [aGD2-SIL(DXR)] liposomes. B, in the same experiment, a second group of animals was treated at 8 days post-inoculation with 6 mg DXR/kg, once a week for 3 weeks, encapsulated in NGR- or ARA-targeted liposomes, respectively [NGR-SL(DXR), ARA-SL(DXR)]. A third group of mice was inoculated with 1.5 mg/kg of NGR-SL(DXR), twice a week for 4 weeks (LDLT dosing schedule). C, tumor-bearing mice were treated with tumor-targeted liposomal DXR at 24 hours and subsequently with vascular-targeted liposomal DXR at 8 days. Mice receiving combined treatments were administered half the dose of DXR for each type of treatment. D and E, changes in CD31-immunoreactive vessels on day 30 after treatment of NXS2 liver metastases. Tissue sections were immunostained for CD31. D, bar, 200 μm. E, columns, mean percentage of control level of CD31-positive vessels; bars, 95% confidence intervals. P values were calculated using ANOVA, with a Tukey post-test. **, P < 0.005 versus control; ***, P < 0.001 versus control and all other groups.

Figure 3.

Effect of in vivo combination treatment of neuroblastoma-bearing mice with tumor- and vascular-targeted therapies. A to C, nude mice were inoculated i.v. with NXS2 neuroblastoma cells, and randomly assigned to groups of eight animals. A, at 24 hours post-inoculation, mice received HEPES-buffered saline, as control, or 8 mg DXR/kg/wk, as free drug or encapsulated in nontargeted [SL(DXR)] or aGD2-targeted [aGD2-SIL(DXR)] liposomes. B, in the same experiment, a second group of animals was treated at 8 days post-inoculation with 6 mg DXR/kg, once a week for 3 weeks, encapsulated in NGR- or ARA-targeted liposomes, respectively [NGR-SL(DXR), ARA-SL(DXR)]. A third group of mice was inoculated with 1.5 mg/kg of NGR-SL(DXR), twice a week for 4 weeks (LDLT dosing schedule). C, tumor-bearing mice were treated with tumor-targeted liposomal DXR at 24 hours and subsequently with vascular-targeted liposomal DXR at 8 days. Mice receiving combined treatments were administered half the dose of DXR for each type of treatment. D and E, changes in CD31-immunoreactive vessels on day 30 after treatment of NXS2 liver metastases. Tissue sections were immunostained for CD31. D, bar, 200 μm. E, columns, mean percentage of control level of CD31-positive vessels; bars, 95% confidence intervals. P values were calculated using ANOVA, with a Tukey post-test. **, P < 0.005 versus control; ***, P < 0.001 versus control and all other groups.

Close modal

To elucidate the additivity or synergy of antitumor responses, we next decided to carry out angiogenesis-specific studies on liver metastases derived from NXS2 cells injected i.v. in nude mice. Histologic evaluation of cryopreserved metastatic tissues, done 30 days after the beginning of the treatment, clearly indicated a strong reduction of CD31-positive endothelial cells (Fig. 3D) and ∼90% suppression of blood vessel density (Fig. 3E; ***, P < 0.001 versus all other groups) in mice treated with the combination therapy.

Because no differences were observed in terms of increased life span, between high-dose and LDLT administration of the vascular-targeted formulations, mice implanted with the second pseudometastatic neuroblastoma model (HTLA-230 cell line) were treated only with the DXR higher dose administration. Our previous in vivo experiments in this animal model showed that an immunoliposome (SIL) formulation directed to the disialoganglioside GD2 antigen, expressed on the surface of neuroblastoma cells, produced a substantial inhibition of tumor growth and dissemination of human neuroblastoma in nude mice (20). Because of its immunogenic potential against HTLA-230 cells (34), the whole mAb recognizing the GD2 epitope was deprived of its Fc portion prior to coupling with the external surface of liposomes and mice were treated with aGD2 Fab′-targeted liposomal DXR. Almost 90% of animals, treated 1 day after tumor cell inoculation, outlived the control mice by more than 3 months (Fig. 4A; ref. 20).

Figure 4.

Survival of neuroblastoma (HTLA-230)-bearing mice after various treatments. A, mice (eight mice/group) were inoculated i.v. with 4 × 106 HTLA-230 cells on day 0 and were treated on days 1, 5, or 10 with HEPES-buffered saline (Control) or with 8 mg DXR/kg/wk for 2 weeks as aGD2 Fab′-targeted liposomes [Fab′-SIL(DXR), t1, t5, t10, respectively]. B, mice (12 mice/group) were inoculated i.v. with 4 × 106 HTLA-230 cells 0 and 5 days post-inoculation, mice received 8 mg of DXR/kg/wk for 2 weeks of Fab′-SIL(DXR). Another group of animals was treated at 19 days post-inoculation, once a week for 2 weeks with 8 mg DXR/kg as NGR-targeted liposomes [NGR-SL(DXR)]. The combination therapy group was treated with tumor-targeted liposomal DXR at 5 days and subsequently with vascular-targeted liposomal DXR at 19 days. Mice receiving combined treatments were administered half the dose of DXR for each type of treatment. Inset, body weight changes in grams as a function of time posttreatment.

Figure 4.

Survival of neuroblastoma (HTLA-230)-bearing mice after various treatments. A, mice (eight mice/group) were inoculated i.v. with 4 × 106 HTLA-230 cells on day 0 and were treated on days 1, 5, or 10 with HEPES-buffered saline (Control) or with 8 mg DXR/kg/wk for 2 weeks as aGD2 Fab′-targeted liposomes [Fab′-SIL(DXR), t1, t5, t10, respectively]. B, mice (12 mice/group) were inoculated i.v. with 4 × 106 HTLA-230 cells 0 and 5 days post-inoculation, mice received 8 mg of DXR/kg/wk for 2 weeks of Fab′-SIL(DXR). Another group of animals was treated at 19 days post-inoculation, once a week for 2 weeks with 8 mg DXR/kg as NGR-targeted liposomes [NGR-SL(DXR)]. The combination therapy group was treated with tumor-targeted liposomal DXR at 5 days and subsequently with vascular-targeted liposomal DXR at 19 days. Mice receiving combined treatments were administered half the dose of DXR for each type of treatment. Inset, body weight changes in grams as a function of time posttreatment.

Close modal

Here, the aim of our work was to verify whether these antitumor effects were maintained in more established tumors or if the therapeutic efficacy declined when treatment was delayed. Indeed, a longer period of time between inoculation of cells and the administration of treatment would allow the tumor cells to establish metastases that might escape treatment. Thus, HTLA-230 cells were injected i.v. into nude mice on day 0. Animals were then treated, once a week for 2 weeks, with 8 mg/kg of DXR encapsulated in aGD2 Fab′-targeted liposomes (Fab′-SIL(DXR) at 1, 5, or 10 days post-inoculation). As expected (Fig. 4A), a delay in the start of treatment substantially reduced the therapeutic activity of the tumor-targeted formulation. Confirming our previous findings (20), seven of eight mice treated 24 hours post-inoculation with Fab′-SIL(DXR) were still alive at 120 days, showing a significant improvement (P = 0.0002) in long-term survival compared with control mice, which died from widespread metastatic disease. However, only one of eight animals treated at 5 days post-inoculation was still alive after 120 days (P = 0.0013). Finally, mice treated 10 days post-inoculation had no long-term survivors, but still had a statistically significant increase in life span (P = 0.0045) compared with control mice.

Once we had shown the time dependence of the antitumor activity of Fab′-SIL(DXR), in a second set of experiments, mice were treated with a sequential combination of tumor- and vascular-targeted liposomes. Nude mice were inoculated i.v. with HTLA-230 cells on day 0 and treated at 5 days post-inoculation with tumor-targeting Fab′-SIL(DXR) liposomes (8 mg DXR/kg weekly × 2) and then, after an additional 14 days, with vascular-targeting NGR-SL(DXR) liposomes (8 mg DXR/kg weekly for 2 weeks). The results obtained clearly show that the combination of tumor- and vascular-targeted liposomal therapies enhanced the cytotoxic effects against this aggressive human neuroblastoma animal model. In particular, combined therapy (half doses of each single liposomal formulation, but equal total doses of DXR) showed a significantly additive activity compared with tumor-targeted (P = 0.006) and vascular-targeted (P = 0.0009) formulations, administered alone (Fig. 4B). Interestingly, the average body weight of mice receiving the combined treatment was not reduced (Fig. 4B , inset), indicating that side effects of the combination therapy were minimal.

Combination therapies increased antivascular effects, tumor growth inhibition, and apoptosis in vivo. To elucidate possible additive or synergistic antitumor responses, angiogenesis-specific studies on metastases derived from HTLA-230 xenograft were done. Histologic evaluation of cryopreserved kidney metastases, done 50 days after the beginning of the treatment, clearly indicated a reduction of >90% of CD31-positive endothelial cells (Fig. 5A) and ∼90% suppression of blood vessel density in mice treated with the combination therapy (control, 35 ± 5 CD31-positive vessels; Fab′-SIL(DXR)+NGR-SL(DXR), 3 ± 2; P < 0.001). Moreover, the antiangiogenic activity of combined targeted liposomes was seen using the CAM assay. CAMs treated with positive buffered control showed numerous allantoic vessels radiating in a spoked wheel pattern towards HTLA-230 xenografts grafted onto CAM (Fig. 5B). Incubation with Fab′-SIL(DXR) and with NGR-SL(DXR) at the same drug concentration led to a 50% and 60% reduction of the number of radiating vessels that invaded the implant, respectively. The combination of Fab′-SIL(DXR) plus NGR-SL(DXR) exerted its angiostatic activity leading to an ∼90% reduction of radiating vessel number, once again indicating the additivity of the individual treatments (Fig. 5B). These observation were confirmed by morphometric assessment of microvessel area [CAMs treated with tumor xenografts alone: 2.95 × 10−2 mm2 (95% CI, 2.5-3.4 × 10−2 mm2); CAMs treated with tumor xenografts with Fab′-SIL(DXR): 1.5 × 10−2 mm2 (95% CI, 1.25-1.75 × 10−2 mm2); P < 0.001 versus control; CAMs treated with tumor xenografts with NGR-SL(DXR): 1.25 × 10−2 mm2 (95% CI, 1.05-1.45 × 10−2 mm2); P < 0.001 versus control; CAMs treated with tumor xenografts with Fab′-SIL(DXR) + NGR-SL(DXR): 0.35 × 10−2 mm2 (95% CI, 0.28-0.42 × 10−2 mm2); P < 0.001 versus control, P < 0.01 versus Fab′-SIL(DXR), and P < 0.05 versus NGR-SL(DXR)].

Figure 5.

Effect of combination therapies on angiogenesis, tumor cell proliferation, and apoptosis in vivo. A, C, and D, immunohistochemical analysis of neuroblastoma metastases removed from untreated mice or mice treated with individual or combined liposomal formulations. Tumors were harvested on day 50 and tissue sections were immunostained for (A) CD31 to show endothelial cells, (C) Ki-67 to show tumor proliferating cells, (D) NB84a to show neuroblastoma cells, or else double-labeled for NB84a and TUNEL, to detect tumor apoptosis. Red, NB84a+ neuroblastoma cells; yellow, merging of red NB84a signal and green TUNEL signal. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole. Bars, 200 μm (A); 100 μm (C and D). B, angiogenesis inhibition in vivo in the CAM model for combinations of Fab′-SIL(DXR) and NGR-SL(DXR). Original magnification, ×50.

Figure 5.

Effect of combination therapies on angiogenesis, tumor cell proliferation, and apoptosis in vivo. A, C, and D, immunohistochemical analysis of neuroblastoma metastases removed from untreated mice or mice treated with individual or combined liposomal formulations. Tumors were harvested on day 50 and tissue sections were immunostained for (A) CD31 to show endothelial cells, (C) Ki-67 to show tumor proliferating cells, (D) NB84a to show neuroblastoma cells, or else double-labeled for NB84a and TUNEL, to detect tumor apoptosis. Red, NB84a+ neuroblastoma cells; yellow, merging of red NB84a signal and green TUNEL signal. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole. Bars, 200 μm (A); 100 μm (C and D). B, angiogenesis inhibition in vivo in the CAM model for combinations of Fab′-SIL(DXR) and NGR-SL(DXR). Original magnification, ×50.

Close modal

Finally, to assess the effect of combined therapies on tumor cell proliferation, viability, and apoptosis in vivo, we stained cryosections taken from HTLA-230 kidney metastases and evaluated the possible tumor parenchyma changes. Histopathologic analysis of excised tumor revealed a pronounced inhibition of tumor cell proliferation as assessed by the drastic decrease in Ki-67-positive cells (Fig. 5C). The percentage of Ki-67-positive cells were quantified in four random fields from two independent experiments: control, 95 ± 4% Ki-67-positive cells; Fab′-SIL(DXR), 40 ± 6% (P < 0.001 versus control); NGR-SL(DXR), 57 ± 7.5% (P < 0.001 versus control); Fab′-SIL(DXR) + NGR-SL(DXR), 15 ± 2.5% [P < 0.001 versus control, P < 0.01 versus Fab′-SIL(DXR), P < 0.001 versus NGR-SL(DXR)]. Notably, only few cells stained for Ki-67 in the sections from animals receiving combined treatments, confirming the higher antitumor activity of this treatment with respect to the single treatment groups. Double staining of tumors with TUNEL and anti-human neuroblastoma showed an increased level of tumor cell apoptosis in the treated mice (Fig. 5D). Interestingly, apoptosis induced by the combined therapy was much prominent respect to the single treatment. Indeed, the percentages of TUNEL+-NB84a+ cells were quantified in four random fields from two independent experiments: control, 3 ± 2% TUNEL+-NB84a+ cells; Fab′-SIL(DXR), 29 ± 4.5% (P < 0.01 versus control); NGR-SL(DXR), 21 ± 3.5% (P < 0.05 versus control); Fab′-SIL(DXR) + NGR-SL(DXR), 65 ± 9% (P < 0.001 versus all the other groups). No observable apoptosis in normal tissue such as heart, lung, kidneys, liver, and spleen was observed in all treated animals (data not shown).

In this work, we have explored two types of antitumor therapies, both of which have attracted considerable attention: the targeting of anticancer therapeutics to tumor cells via ligands against tumor-associated antigens, and the targeting of anticancer therapeutics to the tumor vasculature via ligands directed selectively against tumor endothelial cells. We have now shown that these two therapies are complementary and, when used in combination in a temporal sequence, have additive effects compared with either therapy used alone.

To perform our experiments, we chose murine models of neuroblastoma cells that involved i.v. injection of the cells because this method produces a pseudometastatic animal model that reflects the metastatic advance of disease observed in patients with advanced stage neuroblastoma (17, 20, 27). Indeed, a number of studies conducted on large cohorts of patients have shown that the presence of circulating neuroblastoma cells in the blood and micrometastases in the bone marrow at the time of primary surgery is a strong predictor of relapse (22). Because bone marrow micrometastases are a direct measurement of the ability of tumor cells to spread systemically, the establishment of a model that closely mimics the clinical situation allows a more realistic evaluation of antitumor therapies (17, 20, 27). Thus, the experimental metastatic models employed in these studies provided a reasonable test for the potential of combined antitumor and antivascular therapies in neuroblastoma.

In previous studies, almost complete tumor growth inhibition was observed when treatment of animals with GD2-targeted immunoliposomes (SIL) was started 1 day after tumor cell inoculation (20). Here, we show that this tumor-targeted treatment loses effectiveness if the treatment is delayed for several days. Delayed treatment increases the chances of cell lodging in different organs and establishing metastases. With increasing time between inoculation and treatment, the metastatic cells would become less and less accessible from the vasculature and the tumor-targeted liposomes become less effective as their accessibility to the tumor cells becomes compromised. However, with increasing time, the new lesions begin to recruit blood vessels to support their growth and the lesions will have increased sensitivity to antivascular therapy with time. Hence, the two therapies are complementary. Circulating tumor cells, very early metastases and, arguably, the dividing rim of mature tumors will be most sensitive to tumor-targeted therapies, whereas angiogenic metastases and more mature tumors will be most sensitive to antivascular therapies.

This pseudometastatic model gave us the opportunity to test the well-established concept that tumor growth and progression depends on vasculature development, and to test the hypothesis that both tumor-targeted and vascular-targeted therapies will be time dependent, but in contrasting ways. To this end, we combined the temporal targeting, in vivo, of tumor-targeted liposomal DXR, followed several days later by vascular-targeted liposomal DXR, in the treatment of metastatic neuroblastoma. It was reported that the cytotoxic effects of liposomes encapsulating antineoplastic drugs were enhanced when administered in combination with antitumor and/or antiangiogenic compounds (3537). Straubinger et al. clearly showed that the combination of multiple carrier-based therapies, involving an initial permeability-enhancing sequence of treatments and followed by treatment with antiangiogenic or cytotoxic agents, may represent an important novel strategy for enhancing the efficacy of antitumor treatments (38). In our work, the combination of dual targeting therapy, using two nonimmunogenic ligands (20, 39) coupled to the external surface of liposomes, enhanced the therapeutic effects of the encapsulated DXR, with no obvious side effects. Indeed, the combination therapy was well tolerated by animals, and resulted in a substantial decrease in neuroblastoma progression, leading to pronounced destruction of the tumor vasculature and prolonged survival in mice.

Based on the fact that the CD13 isoform, recognized by NGR-containing peptides, is expressed in the endothelial cells within most (and perhaps all) solid tumors, in this work, we validated the potential of vascular-targeted strategy using NGR-targeted liposomal DXR in in vivo DXR-resistant human lung and ovarian carcinoma experimental models (see Supplemental Data).

NGR-targeted formulations of liposomal DXR administered less frequently, at higher drug concentrations, led to an increased life span in mice compared with those treated more frequently with a lower dose of liposomal DXR over a longer period of time (LDLT administration). The results, which apparently contradict the so-called metronomic chemotherapy hypothesized by Kerbel and Kamen (40), might be explained by the fact that liposomes, themselves, behave as a “metronomic dosing system,” because they are long-circulating and have sustained release properties (Fig. 1). The half-life for release of DXR for liposomes of the composition used in these experiments is 315 hours (41). Moreover, no differences were observed, in terms of increased life span, between high-dose and LDLT administration of the vascular-targeted formulation against two aggressive, pseudometastatic neuroblastoma animal models, in spite of a lower total dose (12 mg DXR/kg for LDLT versus 18 mg DXR/kg for the higher dose). Therapeutic effects are related to the concentration of bioavailable (released) drug as a function of time at the target cells, i.e., tumor area under the curve for the bioavailable drug (42). The relationships between dose and dosing schedule and therapeutic activity for sustained release carriers such as liposomes are complex, and the results with high-dose and LDLT dosing regimens suggest the need for further experimental studies for a better understanding of the crucial balance between the two.

Because other neuroectoderma-derived tumors, such as melanoma, express abundant amounts of the GD2 epitope (43, 44), we speculate that the sequential administration of combinations of GD2-targeted and, subsequently, NGR-targeted liposomes, may provide an effective and more specific therapeutic strategy for additional GD2-positive human tumors. Moreover, because all solid tumors present several unique antigens on their surfaces, several different liposomal drugs targeted against a variety of antigens might become an appealing tool in cancer treatments. One of the authors has shown proof-of-principle for this hypothesis in the treatment of human B lymphoma with combinations of anti-CD19 and anti-CD20-targeted liposomal vincristine (45). Producing tumor-targeted liposomes in a conventional manner (20) or with the post-insertion approach (46), ligand-targeted liposomes could be specifically designed for several types of cancers. In particular, referring to the lung and the ovarian cancer models used in this work, anti–epidermal growth factor receptor and anti-CA125 antibodies, recognizing specific tumor-associated antigens (47, 48), respectively, could be coupled to the external surface of liposomes. The use of tumor-targeted immunoliposomes, in combination with the NGR vascular-targeted formulations, could represent the basis of a new pharmacologic approach for the treatment of malignancies, by taking advantage of formulations that deliver cytotoxic agents to both blood vessels located at sites of disease and to the tumor cells themselves.

Recently, endothelial progenitor cells in the adult bone marrow have shown their ability to circulate in the peripheral blood and incorporate into new blood vessels, contributing to tumor angiogenesis. Among these cells, a subset of tumor-infiltrating cells expressing the angiopoietin receptor, TIE2, have been hypothesized as key effectors of the angiogenic switch during tumor angiogenesis (49). A therapeutic strategy based on targeting TIE2-expressing tumor cells, singularly or in combination with other tumor- and vascular-targeted therapies, is worth further investigation.

Although the tumor endothelial cell is increasingly accepted as a valid target for cancer therapies, another vascular cell type, the pericyte, has been recently recognized as a potentially important, complementary, target for cancer treatment. Pericytes have begun to attract attention due to their apparent involvement in a variety of diseases, including cancer (50). It has been clearly shown that the membrane-associated protease, aminopeptidase A, is up-regulated and enzymatically active in perivascular cells (pericytes) of tumor blood vessels (51). The in vivo screening of peptide-phage libraries allowed the identification of a peptide that selectively recognizes the aminopeptidase A epitope (7). This peptide showed its ability to home to tumor vasculature and inhibit tumor growth in vivo. Thus, the availability of ligands binding to additional tumor-associated antigens, and to additional targets on both endothelial and perivascular tumor cells would allow one to design more sophisticated cancer treatment strategies that exhibit high levels of selective toxicity for the cancer cells. This would improve therapeutic outcome, decrease dose-limiting side effects and improve patient quality of life.

In conclusion, our novel approach relies on the combination of two different targeting strategies and the current article provides the first proof-of-principle for this approach. NGR peptide-targeted liposomal doxorubicin binds to and kills angiogenic blood vessels and, indirectly, the tumor cells that these vessels support, mainly in the tumor core. The anti-GD2-targeted liposomes will result in direct cell kill, including cytotoxicity against cells that are at the tumor periphery and are independent of the tumor vasculature. We believe that this combination approach is a promising one in the search for more effective and less toxic cancer treatments.

Note: Supplementary data for this article are available at Cancer Research Online (http://clincancerres.aacrjournals.org).

F. Pastorino and G. Pagnan are recipients of a Fondazione Italiana per la Lotta al Neuroblastoma fellowship. C. Brignole is a recipient of a Fondazione Italiana Ricerca Cancro fellowship.

Grant support: Fondazione Italiana per la lotta al Neuroblastoma, Associazione Italiana Ricerca Cancro, Ministry of Health (Finalizzata GPT) and the Canadian Institutes for Health Research (MOP-9127).

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.

We thank M. Cioni for expert technical assistance, V. Pistoia for criticism and helpful discussions, and C. Bernardini for editing.

1
Folkman J. Transplacental carcinogenesis by stilbestrol.
N Engl J Med
1971
;
285
:
404
–5.
2
Fidler IJ, Ellis LM. The implications of angiogenesis for the biology and therapy of cancer metastasis.
Cell
1994
;
79
:
185
–8.
3
Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis.
Cell
1996
;
86
:
353
–64.
4
Hicklin DJ, Ellis LM. Role of the vascular endothelial growth factor pathway in tumor growth and angiogenesis.
J Clin Oncol
2005
;
23
:
1011
–27.
5
Hood JD, Bednarski M, Frausto R, et al. Tumor regression by targeted gene delivery to the neovasculature.
Science
2002
;
296
:
2404
–7.
6
Pasqualini R, Koivunen E, Kain R, et al. Aminopeptidase N is a receptor for tumor-homing peptides and a target for inhibiting angiogenesis.
Cancer Res
2000
;
60
:
722
–7.
7
Marchio S, Lahdenranta J, Schlingemann RO, et al. Aminopeptidase A is a functional target in angiogenic blood vessels.
Cancer Cell
2004
;
5
:
151
–62.
8
Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model.
Science
1998
;
279
:
377
–80.
9
Curnis F, Sacchi A, Borgna L, et al. Enhancement of tumor necrosis factor α antitumor immunotherapeutic properties by targeted delivery to aminopeptidase N (CD13).
Nat Biotechnol
2000
;
18
:
1185
–90.
10
Ellerby HM, Arap W, Ellerby LM, et al. Anti-cancer activity of targeted pro-apoptotic peptides.
Nat Med
1999
;
5
:
1032
–8.
11
Pastorino F, Brignole C, Marimpietri D, et al. Vascular damage and anti-angiogenic effects of tumor vessel-targeted liposomal chemotherapy.
Cancer Res
2003
;
63
:
7400
–9.
12
Siemann DW, Mercer E, Lepler S, et al. Vascular targeting agents enhance chemotherapeutic agent activities in solid tumor therapy.
Int J Cancer
2002
;
99
:
1
–6.
13
Siim BG, Lee AE, Shalal-Zwain S, et al. Marked potentiation of the antitumour activity of chemotherapeutic drugs by the antivascular agent 5,6-dimethylxanthenone-4-acetic acid (DMXAA).
Cancer Chemother Pharmacol
2003
;
51
:
43
–52.
14
Gabizon A, Catane R, Uziely B, et al. Prolonged circulation time and enhanced accumulation in malignant exudates of doxorubicin encapsulated in polyethylene-glycol coated liposomes.
Cancer Res
1994
;
54
:
987
–92.
15
Allen TM. Ligand-targeted therapeutics in anticancer therapy.
Nat Rev Cancer
2002
;
2
:
750
–63.
16
Schulz G, Cheresh DA, Varki NM, et al. Detection of ganglioside GD2 in tumor tissues and sera of neuroblastoma patients.
Cancer Res
1984
;
44
:
5914
–20.
17
Raffaghello L, Pagnan G, Pastorino F, et al. In vitro and in vivo antitumor activity of liposomal Fenretinide targeted to human neuroblastoma.
Int J Cancer
2003
;
104
:
559
–67.
18
Pagnan G, Stuart DD, Pastorino F, et al. Delivery of c-myb antisense oligodeoxynucleotides to human neuroblastoma cells via disialoganglioside GD(2)-targeted immunoliposomes: antitumor effects.
J Natl Cancer Inst
2000
;
92
:
253
–61.
19
Brignole C, Pastorino F, Marimpietri D, et al. Immune cell-mediated antitumor activities of GD2-targeted liposomal c-myb antisense oligonucleotides containing CpG motifs.
J Natl Cancer Inst
2004
;
96
:
1171
–80.
20
Pastorino F, Brignole C, Marimpietri D, et al. Doxorubicin-loaded Fab' fragments of anti-disialoganglioside immunoliposomes selectively inhibit the growth and dissemination of human neuroblastoma in nude mice.
Cancer Res
2003
;
63
:
86
–92.
21
Maris JM, Matthay KK. Molecular biology of neuroblastoma.
J Clin Oncol
1999
;
17
:
2264
–79.
22
Moss TJ, Reynolds CP, Sather HN, et al. Prognostic value of immunocytologic detection of bone marrow metastases in neuroblastoma.
N Engl J Med
1991
;
324
:
219
–26.
23
Meitar D, Crawford SE, Rademaker AW, et al. Tumor angiogenesis correlates with metastatic disease, N-myc amplification, and poor outcome in human neuroblastoma.
J Clin Oncol
1996
;
14
:
405
–14.
24
Sugiura Y, Shimada H, Seeger RC, et al. Matrix metalloproteinases-2 and -9 are expressed in human neuroblastoma: contribution of stromal cells to their production and correlation with metastasis.
Cancer Res
1998
;
58
:
2209
–16.
25
Ribatti D, Alessandri G, Vacca A, et al. Human neuroblastoma cells produce extracellular matrix-degrading enzymes, induce endothelial cell proliferation and are angiogenic in vivo.
Int J Cancer
1998
;
77
:
449
–54.
26
Hansen CB, Kao GY, Moase EH, et al. Attachment of antibodies to sterically stabilized liposomes: evaluation, comparison and optimization of coupling procedures.
Biochim Biophys Acta
1995
;
1239
:
133
–44.
27
Lode HN, Xiang R, Varki NM, et al. Targeted interleukin-2 therapy for spontaneous neuroblastoma metastases to bone marrow.
J Natl Cancer Inst
1997
;
89
:
1586
–94.
28
Burbridge MF, Kraus-Berthier L, Naze M, et al. Biological and pharmacological characterisation of three models of human ovarian carcinoma established in nude mice: use of the CA125 tumour marker to predict antitumour activity.
Int J Oncol
1999
;
15
:
1155
–62.
29
Boehle AS, Dohrmann P, Leuschner I, et al. An improved orthotopic xenotransplant procedure for human lung cancer in SCID bg mice.
Ann Thorac Surg
2000
;
69
:
1010
–5.
30
Ponzoni M, Bocca P, Chiesa V, et al. Differential effects of N-(4-hydroxyphenyl)retinamide and retinoic acid on neuroblastoma cells: apoptosis versus differentiation.
Cancer Res
1995
;
55
:
853
–61.
31
Ribatti D, Gualandris A, Bastaki M, et al. New model for the study of angiogenesis and antiangiogenesis in the chick embryo chorioallantoic membrane: the gelatin sponge/chorioallantoic membrane assay.
J Vasc Res
1997
;
34
:
455
–63.
32
Brignole C, Marimpietri D, Pastorino F, et al. Effect of bortezomib on human neuroblastoma cell growth, apoptosis, and angiogenesis.
J Natl Cancer Inst
2006
;
98
:
1142
–57.
33
Richardson DS, Johnson SA. Anthracyclines in haematology: preclinical studies, toxicity and delivery systems.
Blood Rev
1997
;
11
:
201
–23.
34
Raffaghello L, Marimpietri D, Pagnan G, et al. Anti-GD2 monoclonal antibody immunotherapy: a promising strategy in the prevention of neuroblastoma relapse.
Cancer Lett
2003
;
197
:
205
–9.
35
Lopes de Menezes DE, Hudon N, McIntosh N, et al. Molecular and pharmacokinetic properties associated with the therapeutics of bcl-2 antisense oligonucleotide G3139 combined with free and liposomal doxorubicin.
Clin Cancer Res
2000
;
6
:
2891
–902.
36
Li M, Ye C, Feng C, et al. Enhanced antiangiogenic therapy of squamous cell carcinoma by combined endostatin and epidermal growth factor receptor-antisense therapy.
Clin Cancer Res
2002
;
8
:
3570
–8.
37
Brouckaert P, Takahashi N, van Tiel ST, et al. Tumor necrosis factor-α augmented tumor response in B16BL6 melanoma-bearing mice treated with stealth liposomal doxorubicin (Doxil) correlates with altered Doxil pharmacokinetics.
Int J Cancer
2004
;
109
:
442
–8.
38
Straubinger RM, Arnold RD, Zhou R, et al. Antivascular and antitumor activities of liposome-associated drugs.
Anticancer Res
2004
;
24
:
397
–404.
39
Di Matteo P, Curnis F, Longhi R, et al. Immunogenic and structural properties of the Asn-Gly-Arg (NGR) tumor neovasculature-homing motif.
Mol Immunol
2006
;
43
:
1509
–18.
40
Kerbel RS, Kamen BA. The anti-angiogenic basis of metronomic chemotherapy.
Nat Rev Cancer
2004
;
4
:
423
–36.
41
Allen TM, Mumbengegwi DR, Charrois GJ. Anti-CD19-targeted liposomal doxorubicin improves the therapeutic efficacy in murine B-cell lymphoma and ameliorates the toxicity of liposomes with varying drug release rates.
Clin Cancer Res
2005
;
11
:
3567
–73.
42
Laginha KM, Verwoert S, Charrois GJ, et al. Determination of doxorubicin levels in whole tumor and tumor nuclei in murine breast cancer tumors.
Clin Cancer Res
2005
;
11
:
6944
–9.
43
Pastorino F, Stuart D, Ponzoni M, et al. Targeted delivery of antisense oligonucleotides in cancer.
J Control Release
2001
;
74
:
69
–75.
44
Pagnan G, Montaldo PG, Pastorino F, et al. GD2-mediated melanoma cell targeting and cytotoxicity of liposome-entrapped fenretinide.
Int J Cancer
1999
;
81
:
268
–74.
45
Sapra P, Allen TM. Improved outcome when B-cell lymphoma is treated with combinations of immunoliposomal anticancer drugs targeted to both the CD19 and CD20 epitopes.
Clin Cancer Res
2004
;
10
:
2530
–7.
46
Iden DL, Allen TM. In vitro and in vivo comparison of immunoliposomes made by conventional coupling techniques with those made by a new post-insertion approach.
Biochim Biophys Acta
2001
;
1513
:
207
–16.
47
Giaccone G. Epidermal growth factor receptor inhibitors in the treatment of non-small-cell lung cancer.
J Clin Oncol
2005
;
23
:
3235
–42.
48
Cannistra SA. Cancer of the ovary.
N Engl J Med
2004
;
351
:
2519
–29.
49
De Palma M, Venneri MA, Galli R, et al. Tie2 identifies a hematopoietic lineage of proangiogenic monocytes required for tumor vessel formation and a mesenchymal population of pericyte progenitors.
Cancer Cell
2005
;
8
:
211
–26.
50
Pietras K, Hanahan D. A multitargeted, metronomic, and maximum-tolerated dose “chemo-switch” regimen is antiangiogenic, producing objective responses and survival benefit in a mouse model of cancer.
J Clin Oncol
2005
;
23
:
939
–52.
51
Schlingemann RO, Oosterwijk E, Wesseling P, et al. Aminopeptidase A is a constituent of activated pericytes in angiogenesis.
J Pathol
1996
;
179
:
436
–42.

Supplementary data