Vascular Damage and Anti-angiogenic Effects of Tumor Vessel-Targeted Liposomal Chemotherapy

Liposomes Preparation.

Nontargeted SLs and peptide-targeted liposomes (NGR-SL and ARA-SL) were synthesized from HSPC:CHOL:DSPE-PEG₂₀₀₀, 2:1:0.1 molar ratio, and HSPC:CHOL:DSPE-PEG₂₀₀₀:DSPE-PEG₂₀₀₀-MAL, 2:1:0.08:0.02 molar ratio, respectively (28) . In some preparations, [³H]CHE was added as a nonexchangeable, nonmetabolizable lipid tracer. After evaporation under nitrogen, dried lipid films were hydrated in 25 mm HEPES and 140 mm NaCl buffer (pH 7.4). The hydrated liposomes were sequentially extruded (LiposoFast-basic extruder; Avestin, Inc., Leiden, the Netherlands) through a series of polycarbonate filters of pore size ranging from 0.2 μm down to 0.08 μm to produce primarily unilamellar vesicles. Liposomal size was characterized by dynamic light scattering using a Brookhaven BI90 submicron particle size analyzer (Brookhaven Instruments Corp., Holtsville, NY).

DXR was loaded into liposomes via an ammonium sulfate gradient, as previously reported (28) . The loading efficiency of DXR was >95% and liposomes routinely contained DXR at a concentration of 150–180 μg DXR/μmol PL.

To enhance the accessibility of NGR peptide when bound to liposomes and also to permit coupling via a thiol to a maleimido moiety on the liposomes, additional residues were added to the peptide NH₂ terminus to provide the peptide GNGRGGVRSSSRTPSDKYC with a NH₂-terminal Cys (29) . The peptide GARAGGVRSSSRTPSDKYC was used as control. Peptide was conjugated to the liposomes by mixing freshly prepared liposomes with an equimolar (with respect to maleimide) quantity of peptide at 4°C for 16 h under argon followed by a 10-fold excess of 2-mercapethanol for 1 h to derivatize remaining maleimido groups. Uncoupled peptides were separated from the liposomes by passing the coupling mixture through a Sepharose CL-4B column in HEPES buffer (pH 7.4). The efficiency of coupling was determined by estimating the amount of liposome-associated peptides using the CBQCA Protein Quantification kit (Molecular Probes Europe, Leiden, the Netherlands).

Cell Lines and Cellular Association Studies.

To broadly cover the phenotypes exhibited by NB cells in vitro, we used five human NB cell lines GI-ME-N, GI-LI-N, HTLA-230, IMR-32, and SH-SY5Y (30) . The cell lines KS1767 (human Kaposi sarcoma) and THP-1 (human acute monocytic leukemia) were used in some experiments as controls. All cell lines were grown in RPMI 1640 supplemented with 10% fetal bovine serum, as described previously (30) .

The cellular association of liposomes targeted via NGR peptides was analyzed by flow cytometry (FACS), using a FACScan instrument for FACS (Becton-Dickinson Immunocytometry Systems). Aliquots of cells (1 × 10⁶/tube) were incubated for 1 h at 4°C with different formulations of liposome-entrapped DXR (SL, NGR-SL, or ARA-SL). The cells were subsequently washed with PBS and enumerated by FACS. Expression of CD13-binding sites by cultured SH-SY-5Y, THP-1, and KS1767 cells was also measured by FACS, using 1 μg/ml WM15 antibody (PharMingen, San Diego, CA), as reported previously (14) .

Orthotopic NB Animal Model.

Five-week-old female SCID mice were purchased from Harlan Laboratories (Harlan Italy-S.Pietro al Natisone, Udine, Italy). Mice (six to eight mice/group) were anesthetized and injected with the different NB cell lines (2.5 × 10⁶ cells in 20 μl of HEPES buffer), after laparatomy, in the capsule of the left adrenal gland. The lethality of the method was 0%. Mice were monitored at least two times weekly for evidence of tumor development, quantification of tumor size, and evidence of tumor-associated morbidity. All experiments involving animals have been reviewed and approved by the licensing and ethical committee of the National Cancer Research Institute and by Italian Ministry of Health.

BD and PK Experiments.

SCID mice carrying orthotopically implanted NB tumors (mean volume ∼100 mm³) were injected via the tail vein with a single dose of liposomes (0.5 μmol PL/mouse), with or without coupled NGR peptides, containing ∼3 × 10⁵ cpm of the lipid tracer [³H]CHE. At selected time points (2, 12, and 24 h) after injection, mice (three mice/group) were anesthetized and sacrificed by cervical dislocation. A blood sample (100 μl) was collected by heart puncture and counted for the [³H] label in a Packard β-counter. Blood correction factors were applied to all samples (28) . BD was determined as described previously (31) . Data were expressed as the percentage of injected dose/g of tissue.

Histological Analysis.

Paraffin-embedded tissue sections (5 μm) were examined after staining with Mayer’s H&E (Sigma Chemical Co., St. Louis, MO). Monoclonal antibodies against the endothelial cell marker, factor VIII (M616; Dako, Glostrup, Denmark) and antihuman NB (NB84a; Dako), were used. Briefly, sections were collected on 3-amino-propyl-triethoxysilane-coated slides, deparaffinized by the xylene-ethanol sequence, rehydrated in a graded ethanol scale and in Tris-buffered saline (pH 7.6), and incubated overnight at 4°C with M616 (1:25 in Tris-buffered saline) or NB84a (1:40) after prior antigen retrieval by enzymatic digestion with Ficin (Sigma Chemical Co.). The immunoreaction was performed with the streptavidin-peroxidase complex (LSAB2; Dako) and Fast Red as a chromogen. TUNEL staining was performed using a commercially available apoptosis detection kit (In situ Cell Death Detection, POD; Roche Molecular Biochemicals, Mannheim, Germany) according to manufacturer’s instructions.

Determination of Microvessel Area.

Two investigators with a computerized image analysis system (Leica Quantimet 5000, Wetzlar, Germany) simultaneously assessed microvessel area. Four to six 250 × magnification fields, covering almost the whole of each three sections (every third section within nine serial sections)/sample, were examined with a 484-intersection point square reticulum (12.5 × 10²/mm²) inserted in the eyepiece. Care was taken to select microvessels, i.e., capillaries and small venules from all of the factor VIII-stained vessels. They were identified as transversally sectioned tubes with a single layer of endothelial cells, without or with a lumen (diameter ranging from 3 to 10 μm). Microvessels were counted by a planimetric point-count method with slight modifications, according to which only microvessels transversally cut occupying the reticulum points were counted. Because the microvessel diameter was smaller than the distance between adjacent points, only one transversally sectioned microvessels could occupy a given point. Microvessels transversally sectioned outside the points and those longitudinally or tangentially sectioned were omitted. It was thus sufficiently certain that a given microvessel was counted only once, even in the presence of several of its section planes. As almost the entire section of each of three nonadjacent sections was analyzed/sample and as transversally sectioned microvessels hit the intersection points randomly, the method allowed objective counts.

In Vivo Therapeutic Studies.

A total of 2.5 × 10⁶ SH-SY-5Y cells were injected orthotopically in the left adrenal gland of mice. Tumors were allowed to grow for 21 days and then i.v. DXR treatment (free or encapsulated in targeted or nontargeted liposomes) was initiated with different doses of DXR/kg every week for 3 weeks (see “Legends” and “Results”). At different time points, mice were sacrificed, and tumors were measured with calipers. Tumor volumes were calculated by the formula π/6 [w₁ × (w₂)² ], where w₁ represented the largest tumor diameter and w₂ represented the smallest tumor diameter. The body weight and general physical status of the animals were recorded daily, and the mice were terminated when tumor reached 1000–1200 mm³. Histological evaluation of microscopic metastases was performed for all tissues. Organs were fixed in neutral buffered 10% formalin, processed by standard methods, embedded in paraffin, sectioned at 5 μm, and stained with H&E and for IHC. Weight because of tumor burden was calculated by subtracting the normal wet organ weight from each tumor-bearing organ. In some experiments, mice were monitored routinely for weight loss, and survival times were used as the main criterion for determining treatment efficacy. All of the in vivo experiments have been performed at least three times with similar results.

Statistical Methods.

The statistical significance of differential findings between experimental groups and controls was determined by Student’s t test and considered significant if two-tailed Ps were <0.05. Peto’s log-rank test determined the significance of the differences between experimental groups in the survival experiments by the use of StatsDirect statistical software (CamCode, Ashwell, United Kingdom).

Characterization of NGR-Targeted Liposomes and Binding to Endothelial Cells in Vitro.

Liposomes were typically 90–115 nm in diameter, with a DXR entrapment efficiency in liposomes of ∼95%. All liposomal formulations showed minimal leakage in PBS, retaining >90% of the encapsulated DXR after a 24-h incubation. As expected from our previous data (28 , 32) , when leakage experiments were conducted in 25% human plasma, liposomes showed a slow DXR leakage (12%) at 24 h. For the targeting domain, we used the NGR motif, which binds a CD13 isoform selectively expressed by tumor-associated vessels (14 , 29) , for which there is evidence of internalization (6 , 16) . An average coupling efficiency for the NGR peptide of 55% (95% confidence interval = 50–60%) was obtained, resulting in a peptide density of 7–8 μg peptide/μmol PL.

To evaluate the specificity of NGR-SL[DXR], we used KS1767 cells, derived from Kaposi sarcoma because they bind the NGR-targeting peptide, as do endothelial cells (6 , 16) . The THP-1 (acute monocytic leukemia) cell line was used as a control because it does not bind the NGR peptide (14) . FACS analysis showed that WM15 (a monoclonal antibody specific for CD13) bound to both THP-1 and KS1767 cells (Fig. 1, A and D ⇓ , respectively). In contrast, NGR-SL[DXR] were able to bind the KS1767 cells (Fig. 1E ⇓ and inset) but not the THP-1 cells (Fig. 1B ⇓ and inset), thus confirming previous results indicating the existence of different isoforms within the CD13 molecule (14) . Noteworthy, both WM-15 and NGR-SL[DXR] did not bind to SH-SY5Y NB cells (Fig. 1, G and H) ⇓ . To further confirm the selectivity of binding of our liposomal formulation, we used the mismatched peptide ARA as a targeting moiety. In this case, ARA-SL[DXR] did not bind to all cell lines (Fig. 1, C, F, and I) ⇓ .

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Cellular association of NGR-targeted liposomes analyzed by flow cytometry. A–C, THP-1 cells; D–F, KS1767 cells; G–I, SH-SY5Y. Cells were incubated with WM15 monoclonal antibody (a, d, and g), NGR-SL[DXR]; (B, E, and H), or ARA-SL[DXR] (C, F, and I). After washing, cells treated with WM15 were incubated with a goat antimouse-FITC secondary antibody and then analyzed by FACS (A, D, and G: thin lines, WM15; bold lines, none). Cells treated with liposomal DXR were washed and directly enumerated by FACS (B, C, E, F, H, and I: thin lines, targeted liposomes; bold lines, nontargeted liposomes). In the insets B, E, and H, cells were evaluated for DXR fluorescence by fluorescence microscopy.

Biologically Relevant Orthotopic NB Xenograft Model.

A more realistic view of the clinical potential of NGR-targeted liposomes could be obtained if a tumor model were available that better reflected the growth of advanced NB in children (i.e., large adrenal gland tumors and multiple small metastatic lesions). All current data support this concept and recommend that orthotopic implantation of tumor cells in recipient animals is mandatory for studies of tumor progression, angiogenesis, invasion, and metastasis (33) .

To provide a well-characterized, relevant, highly reproducible, angiogenic, and metastatic orthotopic model of NB, we initiated studies to define five adrenal NB xenograft models based on previously described human NB cell lines (30) and reported animal models (34 , 35) . From these data (data not shown), we decided to use the intra-adrenal injection of SH-SY5Y cells in mice for additional studies because, 2–3 weeks after injection, adrenal gland tumors were always found in all animals (Fig. 2A) ⇓ . This model best reflected the typical growth pattern of human NB because orthotopic injection of SH-SY5Y cells resulted in solid adrenal tumors that were highly vascular, locally invasive into surrounding tissues, and metastatic to distant sites. Indeed, macroscopic metastases always occurred after 3–4 weeks of injection in the contralateral adrenal (Fig. 2, B and C) ⇓ and liver (Fig. 2D) ⇓ , whereas micrometastases were apparent frequently in the ovary (Fig. 2E) ⇓ , right kidney (Fig. 2F) ⇓ , liver (Fig. 2G) ⇓ , and lung (Fig. 2H) ⇓ .

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Orthotopic NB xenograft model in SCID mice. A and B, adrenal gland tumors (arrows) in mice that were injected orthotopically with SH-SY5Y cells at 14 (A) and 21 (B) days before sacrifice. C and D, representative right adrenal gland (C) and liver (D) samples at 3 and 4 weeks after injection of NB cells, respectively. E–H, histological (H&E) analysis of representative ovary (E), kidney (F), liver (G), and lung (H) samples. Forty days after cell injection, animals were sacrificed, the organs were removed, fixed, paraffin embedded, sectioned at 5 μm, and stained with H&E. Arrows indicate metastatic tumor invasion in the lung. Arrowheads show the normal ovaric follicular structure surrounded by tumor NB cells. Magnification, ×10 (insets, ×63).

PK and BD Profiles of NGR-Targeted Liposomes.

As previously shown (36) , long circulation times are required for SLs to gain access to tumor sites. Thus, the PKs and BD of [³H]CHE-labeled SL[DXR] and NGR-SL[DXR] was evaluated in xenograft models of orthotopic NB in SCID mice. The results of the PK studies are expressed as percentage of the administered dose of lipid remaining in blood. These findings clearly indicate that liposomes coupled to NGR peptide had long-circulating profiles in blood, being almost identical to that obtained with nontargeted SL[DXR] (Fig. 3A) ⇓ . The BD of liposomes was evaluated at 2, 12, and 24 h after injection. At all times, the spleen uptake of targeted liposomes was ∼10–20 times higher than that of SL. No differences between uptake of SL[DXR] versus NGR-SL[DXR] occurred in other tissues, and with the exception of liver, uptake into other organs was very low (Table 1) ⇓ . Finally, we evaluated the tumor accumulation of targeted and nontargeted liposomes. The uptake into tumor by NGR-SL[DXR] was time dependent, being at least 10 times higher than that of SL[DXR] after 24 h (Fig. 3B) ⇓ . No uptake was observed into tumors of mice treated with ARA-SL[DXR]; (Fig. 3, B and F) ⇓ .

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

A, blood clearance kinetics of nontargeted and NGR-targeted DXR-loaded liposomes in SCID mice. Liposomes were labeled with the lipid tracer [³H]CHE and were administered i.v. in a single bolus dose (0.5 μmol PL/mouse). Treatment groups consisted of NGR-SL[DXR]; (•) and SL[DXR]; (○). At different times after injection, blood was collected and counted for [³H] label. Each point represents the average of three mice ± SD. B, tumor accumulation of NGR-targeted DXR-loaded liposomes in SCID mice injected orthotopically with NB cells. [³H]-labeled DXR-loaded liposomes, either nontargeted ( ), ARA-targeted (□), or NGR-targeted ( ) were injected via the tail vein as a single bolus dose. After 2, 12, and 24 h, tumors were collected and counted for [³H] label. Results are expressed as cpm/mg tissue. Each point represents the average of three mice (±SD). C–H, tumor accumulation of NGR-targeted liposomal DXR. Mice were treated with NGR-SL(DXR) as in B. After 2 (C), 12 (D), and 24 (E) h, tumors were collected and DXR visualized by fluorescence microscopy of fixed, paraffin-embedded, tissue sections. Alternatively, mice were injected with control ARA-SL[DXR]; (F), with NGR-SL[DXR] either alone (G) or coinjected with a 50-fold excess of the soluble NGR peptide (H) and tumors collected after 16 h. Magnitude, ×40.

Homing to Tumor Vasculature.

To determine whether the NGR-targeted liposomes could deliver DXR to angiogenic tumor-associated blood vessels, we injected NGR-SL[DXR] into the tail vein of mice bearing established adrenal tumors. In one set of experiments, liposomes were allowed to circulate from 2 to 24 h, followed by perfusion and immediate tissue recovery. There was a clear time-dependent uptake of DXR in the tumor vasculature (Fig. 3, C–E) ⇓ . At 24 h, the staining pattern indicated that the DXR had spread outside the blood vessels and into the tumors. This spreading may be attributable to increased permeability of tumor blood vessels to the intact liposomes (37) or uptake of the targeted liposomes by angiogenic endothelial cells and subsequent transfer to tumor cells, as shown for the uptake of phage (6) . Likely, both mechanisms are working at the same time. In the second set of experiments, tissues were examined 16 h after the injection of NGR-SL[DXR] or ARA-SL[DXR]. Strong DXR staining in tumor vasculature was seen only in animals treated with NGR-liposomes (Fig. 3, G versus F) ⇓ . At this time, minimal DXR staining was observed in the spleen and liver, and no detectable expression was found in the heart, lung, kidney, and brain (data not shown). Tumor-specific DXR uptake was completely blocked when mice were coinjected with a 50-fold molar excess of the soluble NGR peptide (Fig. 3H) ⇓ .

In Vivo Therapeutic Studies.

To determine whether liposomes homing to the tumor vasculature could be used to improve the therapeutic index of the chemotherapeutic agent DXR, we injected SH-SY5Y cells into the left adrenal gland of SCID mice and allowed them to grow for 21 days, at which time, they reached a size of ∼200 mm³. The commonly used dose of DXR in SCID mice with human tumor xenografts is 1–3 mg/kg/week for 3–4 weeks (38 , 39) . Because we expected the liposomal DXR to be more effective and less toxic than the free drug, we initially performed a dose-escalation experiment in which mice with established tumors were treated with different doses of DXR once a week for 3 weeks and then observed for an extended period of time. Tumor-bearing mice treated with 2 mg/kg/week outlived the control mice, all of which died from widespread disease (Fig. 4A ⇓ , log-rank test, P < 0.001 and Fig. 4, B–D ⇓ ). However, higher doses were toxic because all of the mice treated with 4 and 8 mg/kg/week died within 48 h of the third and second drug administration, respectively. Thus, the maximum-tolerated dose of targeted liposomal DXR in SCID mice was between 6 and 12 mg.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Antitumor effects of NGR-targeted liposomal DXR in vivo. A, dose-dependent effects of NGR-SL[DXR] on the survival of tumor-bearing mice. Treatment groups (n = 8/group) consisted of HEPES buffer (control, □), NGR-SL[DXR] at 8 mg/kg (○), 4 mg/kg (▵), 2 mg/kg (×), and 1 mg/kg (+). Mice received injections in the adrenal gland with SH-SY5Y cells on day 0 and received the various treatments 21, 28, and 35 days after inoculation. B–D, mice were inoculated and treated as in a, then photographed on day 36: B, control; C, treated with 1 mg/kg/week; D, treated with 2 mg/kg/week. E–M, IHC analysis. Tumors were harvested on day 36 from control mice (E, H, and K) and mice treated with 1 mg/kg/week (F, I, and L) or 2 mg/kg/week (G, J, and M) as in A. Tissue sections were immunostained for the expression of factor VIII (to show vessels, E, F, and G; magnitude ×10) or for a double label of factor VIII (endothelial cells) and TUNEL (apoptosis; H, I, and J) or for a double label of NB84a (NB cells) and TUNEL (K, L, and M). Red, factor VIII+ endothelial cells or NB84a+ NB cells; green, TUNEL+ cells; yellow, colocalization of TUNEL+-factor VIII+ or –NB84a+ cells. H–M, ×40 magnification.

To assess the impact of NGR-SL[DXR] on tumor cell viability, we stained cryosections taken from tumors at 24 h after the third treatment and examined them at low and medium magnification to evaluate both blood vessels and surrounding tumor parenchyma. Histopathological analysis of excised tumor on day 36 revealed pronounced destruction of the tumor vasculature with a marked decreased in vessel density after treatment of the mice with 2 mg/kg/week × 3 of NGR-SL[DXR]; (Fig. 4G) ⇓ . Double staining of tumors with TUNEL and antifactor VIII antibody or antihuman NB, demonstrated endothelial cell apoptosis in the vasculature (Fig. 4I) ⇓ , as well as increased tumor cell apoptosis (Fig. 4M) ⇓ , respectively. Interestingly, apoptosis induced by NGR-SL[DXR] was prominent only in the tumor tissues; similar treatment resulted in no observable apoptosis in normal tissue such as heart, lung, kidneys, liver, and spleen (data not shown).

To further test the therapeutic efficacy of this treatment, we randomly sorted mice bearing established 200 mm³ of SH-SY5Y adrenal tumors into three groups and treated each group with a weekly tail vein injection of HEPES buffer (untreated controls) or 3 mg/kg DXR entrapped in NGR-SL or in ARA-SL for 3 consecutive weeks. Mice injected with HEPES buffer or ARA-SL[DXR] or NGR-SL without DXR (data not shown) formed large tumors (1100–1200 mm³) and, consequently, were euthanized on day 42 (Fig. 5A) ⇓ . In contrast, mice injected with NGR-SL[DXR] displayed rapid tumor regression (Fig. 5A) ⇓ . One day after the third treatment, four of six mice showed no evidence of tumors, and the two others showed a >80% reduction in tumor mass and a >90% suppression of blood vessel density (Fig. 5A ⇓ inset; P < 0.01). These findings demonstrated that aminopeptidase N-targeting delivery of DXR to blood vessels caused tumor regression because of its ability to promote apoptosis of the angiogenic endothelium.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Delivery of DXR to tumor vessels inhibits angiogenesis, causing regression of established NB tumors. A, SCID mice orthotopically implanted in the left adrenal gland with NB cells were allowed to form tumors of ∼200 mm³ in size and were then injected i.v. with 3 mg DXR/kg/week as in Fig. 4 ⇓ . Tx, start of treatment. (•) HEPES buffer control; (○) ARA-SL[DXR]; (▴) NGR-SL[DXR]. Each point represents the mean ± SD of six replicates. Inset, orthotopic tumors, at day 36 from control and DXR-treated groups, were sectioned and stained with an antibody to factor VIII to count blood vessels. Each bar represents the mean ± SD of five replicates. B and C, mice were treated as above, and on day 36, organs (liver, B; kidney, C) were harvested and weighed. Each bar represents the mean ± SD of six mice. (∗∗,--) P < 0.01). D, effects of different schedules of treatment with NGR-SL[DXR] on life span. Treatment groups (n = 8/group) consisted of HEPES buffer (control, □); free DXR (▵) or NGR-SL[DXR]; (+), treated with 3 mg/kg on days 21, 28, and 35; free DXR (○) or NGR-SL[DXR; ×] treated with 1 mg/kg every other 2 days starting on day 21 for a total of nine injections.

We next examined whether this therapy was also effective against established metastases. Control mice treated with HEPES buffer or ARA-SL[DXR] showed extensive tumor burdens in the liver and right kidney (Fig. 5, B and C) ⇓ . In contrast, mice treated with NGR-SL[DXR] displayed little or no visible tumor metastasis, as demonstrated by a significant reduction in wet liver or kidney weight (Fig. 5, B and C ⇓ ; P < 0.01).

Finally, to test whether continuous low-dose administration of chemotherapy agents or antiangiogenic schedule chemotherapy (19 , 20) could also inhibit orthotopically placed NB tumors, we compared our standard dosing schedule of 3 mg/kg/wk × 3 with a metronomic administration of DXR, either free or encapsulated into NGR-targeted liposomes (1 mg/kg/every other 2 days × 9). Complete tumor eradication was seen in all animals treated with either schedule. However, the observed tumor growth inhibition was maintained beyond 4 months in four of eight mice and in three of eight mice treated with the metronomic or the standard schedule, respectively, showing a significant improvement (P < 0.0001 and P = 0.0002, respectively) in long-term survival compared with control animals or mice treated with free DXR (Fig. 5D) ⇓ .