Systemic Delivery of Liposomal Short-Chain Ceramide Limits Solid Tumor Growth in Murine Models of Breast Adenocarcinoma

In vitro tumor cell culture models have illuminated the potential therapeutic utility of elevating the intracellular concentration of the antimitogenic and proapoptotic sphingolipid, ceramide. However, although cell-permeable, short-chain ceramide is an effective apoptotic agent in vitro, its use as an in vivo, systemically delivered therapeutic is limited by its inherent lipid hydrophobicity and physicochemical properties. Here, we report that the systemic i.v. delivery of C₆-ceramide (C₆) in a pegylated liposomal formulation significantly limited the growth of solid tumors in a syngeneic BALB/c mouse tumor model of breast adenocarcinoma. Over a 3-week treatment period, a well-tolerated dose of 36 mg/kg liposomal-C₆ elicited a >6-fold reduction in tumor size compared with empty ghost liposomes. Histologic analyses of solid tumors from liposomal-C₆-treated mice showed a marked increase in the presence of apoptotic cells, with a coincident decrease in cellular proliferation and in the development of a microvessel network. Liposomal-C₆ accumulated within caveolae and mitochondria, suggesting putative mechanisms by which ceramide induces selective cancer cell cytotoxicity. A pharmacokinetic analysis of systemic liposomal-C₆ delivery showed that the pegylated liposomal formulation follows first-order kinetics in the blood and achieves a steady-state concentration in tumor tissue. Confirming the therapeutic utility of i.v. liposomal-C₆ administration, we also shown diminution of solid tumor growth in a human xenograft model of breast cancer. Together, these results indicate that bioactive ceramide analogues can be incorporated into pegylated liposomal vehicles for improved solubility, drug delivery, and antineoplastic efficacy.

Sphingolipids not only serve a structural role in membranes but also are substrates for the generation of bioactive second messengers that influence mitogenesis and apoptosis. Metabolism of sphingomyelin, the major sphingolipid in membranes, forms ceramide, a potent lipid-derived second messenger that modulates the induction of cell differentiation, cell cycle arrest, and/or apoptosis (1–4). In addition, chemotherapeutic agents (5–7) and ionizing radiation (8, 9) are two of the multiple cellular stressors (10–16) that lead to the accumulation of ceramide within membranes. We and others have shown that ceramide-mediated signaling cascades induce apoptosis in part via the inhibition of Akt prosurvival pathways, mitochondrial dysfunction, and the stimulation of caspase activity, which ultimately leads to DNA fragmentation and cell death (17–21).

Although short-chain, cell-permeable ceramides, such as C₆-ceramide (C₆), have been shown to be antiproliferative and proapoptotic in numerous cancer cell types in vitro (22, 23), there are obstacles to the delivery of ceramide for systemic applications, such as cancer chemotherapy. Despite being more efficacious than physiologic long-chain ceramides (C₁₈-C₂₄-ceramide), the effectiveness of cell-permeable ceramide analogues remains limited due to their inherent hydrophobicity and potential precipitation as fine lipid micelle suspensions when given in aqueous solutions (24). Thus, to realize the therapeutic benefits of ceramide bioactivity, there is a need for improved drug delivery systems to assist with solubility, cell permeability, protection from enzymatic degradation, and systemic administration.

We and the Mayer Laboratory have shown that ceramide-incorporated liposomes are more cytotoxic than nonliposomal ceramide delivered via organic solvents (21, 25, 26). Our studies showed that C₆-incorporated pegylated liposomes significantly inhibited proliferation and induced apoptosis in human MDA-MB-231 breast adenocarcinoma cells (21). In addition, stable ceramide-containing liposomes exhibited an increase in life span of >20% in the J774 ascites tumor model compared with control liposomes following i.v. bolus and i.p. administration (26). The present study extends these findings to a solid tumor model of mammary adenocarcinoma. Our studies not only show the efficacy of inhibiting solid tumor growth following i.v. administration of liposomal-C₆ but also documents the underlying biophysical mechanisms of delivery of ceramide leading to tumor cell cytotoxicity.

Materials and cell culture. 1,2-Dioleoyl-sn-glycero-3-phosphocholine, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine, N-hexanoyl-d-erythro-sphingosine (C₆), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy polyethylene glycol-2000], N-octanoyl-sphingosine-1-[succinyl(methoxy polyethylene glycol-750)] (PEG(750)-C₈), N-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-d-erythro-sphingosine (NBD-C₆), and N,N-dimethyl-d-erythro-sphingosine (DMSph) were purchased from Avanti Polar Lipids (Alabaster, AL). N-hexanoyl-d-erythro-sphingosine [hexanoyl 6-³H] ([³H]C₆) was obtained from ARC (St. Louis, MO). Primary antibodies to caveolin-1 and CD31 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and BD PharMingen (San Jose, CA), respectively. Horseradish peroxidase–conjugated goat anti-rabbit IgG and rhodamine red-X–conjugated goat anti-rat IgG secondary antibodies were obtained from Santa Cruz Biotechnology and Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA), respectively. For pharmacokinetic studies, Solvable aqueous-based tissue solubilizer was purchased from Perkin-Elmer (Wellesley, MA). Murine 410.4 mammary adenocarcinoma cells were a kind gift from Dr. Amy Fulton (University of Maryland, Baltimore, MD) and grown at 37°C in RPMI 1640 supplemented with 10% fetal bovine serum. This cell line is a highly aggressive murine metastatic tumor cell line cultured from a BALB/c mouse lineage (27). Normal murine HC11 mammary epithelial cells, also from a BALB/c mouse lineage, were a kind gift from Dr. Bernd Groner (University of Frankfurt am Main, Frankfurt am Main, Germany) and grown at 37°C in RPMI 1640 supplemented with 10% fetal bovine serum, 10 ng/mL epidermal growth factor, and 5 μg/mL insulin. Human MDA-MB-231 breast adenocarcinoma cells were obtained from American Type Culture Collection (Manassas, VA) and grown at 37°C in RPMI 1640 supplemented with 10% fetal bovine serum. MDA-MB-231 is a highly aggressive human metastatic, estrogen receptor–negative, breast cancer cell line.

Liposome formulation. Lipids were solubilized in chloroform, combined in a specific molar ratio [1,2-dioleoyl-sn-glycero-3-phosphocholine/1,2-dioleoyl-sn-glycero-3-phosphoethanolamine/1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy PEG(2000)]/PEG(750)-C₈/C₆ (3.75:1.75:0.75:0.75:3.0 molar ratio)], dried under a stream of nitrogen above the lipid transition temperatures for each lipid, and hydrated with sterile isotonic 0.9% NaCl solution. The resulting solution underwent sonication for 2 minutes followed by extrusion through 100 nm polycarbonate membranes using the Avanti Mini Extruder (Avanti Polar Lipids). Incorporation efficiency was determined as described previously (21) by incorporating trace amounts of [³H]C₆ into the formulation, extracting constituent lipids in chloroform/methanol (2:1), and comparing radioactivity levels of equal aliquots before and after extrusion using a scintillation counter. In preliminary studies, there was no significant loss of C₆ during the formulation of liposomal vesicles. Moreover, liposomal-C₆ vesicles were formulated with an average size of 84 nm as measured by dynamic light scattering. The composition of formulated liposomes was validated by extracting constituent lipids in chloroform/methanol (2:1) followed by resolution on preheated silica gel 60 TLC plates using a chloroform/methanol/double-distilled water (60:25:4) solvent system. Lipids were visualized in an iodine chamber.

Cell cytotoxicity studies. To characterize the antiproliferative effects of liposomal-C₆ on normal and cancerous murine mammary epithelial cells, murine HC11 mammary epithelial cells and 410.4 mammary adenocarcinoma cells were plated out at 6 × 10³ per well in 96-well tissue culture plates. The following day, the cells were treated in their normal culture medium with liposomal-C₆, ghost liposomes, nonliposomal “free” C₆, and DMSO vehicle at the indicated doses for 48 hours. As a positive control for cell cytotoxicity, HC11 cells were also treated with DMSph. Following the 48-hour treatment, cell cytotoxicity was assessed using the Cell Titer 96 aqueous nonradioactive cell proliferation assay (Promega, Madison, WI) following manufacturer's instructions. The assay employs a tetrazolium compound that is bioreduced by viable cells into a soluble formazan product, which can be measured by absorbance at 490 nm. The quantity of formazan product as measured by its absorbance is directly proportional to the number of viable cells in culture.

Cell apoptosis analysis. As described previously (21), ceramide-induced apoptosis was assessed and quantified by flow cytometric analysis of Annexin V–stained cells using the Vybrant Apoptosis Assay Kit 3 (Molecular Probes, Eugene, OR). Briefly, 410.4 mammary epithelial cells were plated at 4 × 10⁵ per plate in 60 mm tissue culture plates. Following a 24-hour treatment with 25 μmol/L liposomal-C₆, adherent cells were harvested by trypsinization and all cells (floating and adherent) were washed once with cold PBS and pelleted at 300 × g. Pelleted cells were then stained with FITC-labeled Annexin V and propidium iodide according to the manufacturer's instructions. Labeled cells were immediately analyzed using flow cytometry. Viable cells were double negative, early apoptotic cells were positive for Annexin V staining and negative for propidium iodide staining, and late apoptotic cells were double positive. Cells positive for propidium iodide but negative for Annexin V were considered to be necrotic or debris and were excluded from apoptotic cell analysis.

In addition to Annexin V staining, an In situ Cell Death Detection kit (Roche Diagnostics Corp., Indianapolis, IN) was used, following the manufacturer's instructions, to confirm the induction of apoptosis. Briefly, 410.4 mammary epithelial cells were plated at 4 × 10⁵ per plate in 60 mm tissue culture plates and treated with 25 μmol/L liposomal-C₆, nonliposomal-C₆, and vehicle controls for 24 hours. Following treatment, adherent cells were harvested by trypsinization and all cells (floating and adherent) were washed once with PBS, fixed with 2% paraformaldehyde, permeabilized with 0.1% Triton X-100, and treated with terminal deoxynucleotidyl transferase in the presence of fluorescein-labeled nucleotide polymers. Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL)-positive cells were immediately analyzed using flow cytometry to quantify the induction of apoptosis. A one-color analysis, excluding small cellular debris, was done.

Confocal studies. To verify cell accumulation of C₆ into 410.4 murine mammary adenocarcinoma cells, we formulated a liposomal-C₆ vesicle with 10 molar percent NBD-C₆ as a marker for C₆. Cells were plated at 2.0 × 10⁴ per well in eight-well chamber slides and allowed to grow overnight. Liposomal-NBD-C₆ was treated at 25 μmol/L for a 2-hour treatment period. Cellular nuclei were counterstained with 4′,6-diamidino-2-phenylindole and MitoTracker Deep Red 633 (Molecular Probes) was used as a marker for mitochondria following manufacturer's instructions. C₆ delivery and accumulation was evaluated by confocal microscopy at ×63 magnification (Leica Microsystems, Heidelberg, Germany).

Sucrose gradient and Western blot. To evaluate whether ceramide preferentially accumulates into caveolin-1-enriched lipid rafts (known as caveolae), we did a sucrose gradient of cellular lysate as described previously (28). This velocity gradient centrifugation procedure permits the isolation of caveolin-1-enriched lipid raft microdomains within a cellular lysate. The liposomal-C₆ formulation was incorporated with trace amounts of [³H]C₆ to monitor C₆ levels. Briefly, 410.4 cells were plated at 1.5 × 10⁶ per plate into 100 mm tissue culture plates and grown to ∼90% confluency. Cells were treated with 25 μmol/L tritiated liposomal-C₆ for the indicated time points. Cells were lysed in 2 mL of 500 mmol/L NaCO₃ buffer containing 1 mmol/L phenylmethylsulfonyl fluoride and 1 mmol/L NaVO₃ and subsequently homogenized using a Dounce glass homogenizer. Cell lysates were mixed 1:1 (v/v) with 90% sucrose in 25 mmol/L MES (pH 6.5) and 0.15 mol/L NaCl (final 45% sucrose) and loaded into Beckman (Fullerton, CA) ultracentrifuge tube. The 45% sucrose solution was then overlayed with 4 mL of 35% sucrose/NaCO₃ solution and then 4 mL of 5% sucrose/NaCO₃ solution to form a 5%-35%-45% sucrose gradient. Samples were centrifuged at 35,000 rpm for 18 hours using a SW-40ti rotor. The 12 mL gradient was then aliquoted into 1 mL fractions. Equal aliquots of 800 μL were removed from each fraction and counted using a scintillation counter to assess ceramide levels. In addition, 50 μL of each fraction were loaded onto a Nupage 4% to 12% precasted SDS-PAGE gradient gel and probed for caveolin-1, which was visualized using enhanced chemiluminescence.

In vivo syngeneic mammary adenocarcinoma tumor model. To assess the in vivo efficacy of systemic liposomal-C₆ delivery in an immunocompetent, syngeneic solid tumor model, 5 × 10⁶ 410.4 cells were injected s.c. into the right hind flank of female BALB/c mice, which were purchased from Charles River Laboratories (Wilmington, MA). Four days later, when tumor size averaged 3 × 3 mm, tumor-bearing mice were randomized into seven groups with five mice per group. The mice received i.v. tail vein injections of either liposomal-C₆ (12, 24, or 36 mg/kg C₆) or empty ghost liposomes (equivalent total lipid mass as ceramide treatment group); the seventh group remained untreated. Mice were treated every 2 days, and immediately before treatment, mice were weighed and tumors were measured. Tumor size was measured with calipers, and tumor volume was calculated using the following formula: V = (L × W²) / 2, where V is tumor volume, L is length, and W is width. The study was halted when control tumors reached ∼14 mm in diameter ∼21 days following tumor cell injection.

In vivo human breast cancer xenograft tumor model. Further animal experimentation was done in athymic female nude mice, which were purchased from Harlan Sprague-Dawley (Indianapolis, IN). Tumor kinetics was measured by s.c. injection of 1 × 10⁷ MDA-MB-231 cells in 0.2 mL DMEM containing 10% fetal bovine serum above the left and right rib cages of 4- to 6-week-old nude mice. Based on the data obtained from the syngeneic tumor model, xenograft tumor-bearing mice were initially treated every 2 days with 36 mg/kg liposomal-C₆ or empty ghost liposomes (equivalent total lipid mass as ceramide treatment group). We also chose to use an escalating dose regimen (36-72 mg/kg) to further evaluate efficacy versus possible systemic toxicology. Three dimensional tumor size was measured on alternate days with calipers, and tumor volume was calculated using the following formula: V = (L × W × H), where V is tumor volume, L is length, W is width, and H is height. All of the animal studies were approved by the Institutional Animal Care and Use Committee at the Pennsylvania State College of Medicine (Hershey, PA).

Syngeneic tumor histology. A thorough histologic analysis was done to validate tumor growth inhibition. Tumors were removed from tumor-bearing mice following 1-week treatment with 40 mg/kg liposomal-C₆ and frozen in an ethanol/dry ice bath, and cryosections (4 microns) were generated for histologic analysis. To assess the degree of C₆-induced cellular apoptosis, tumor sections were stained using the In situ Cell Death Fluorescein Detection kit (Roche Diagnostics) following the manufacturer's instructions. Briefly, cryosections were fixed in 4% paraformaldehyde and permeabilized in 0.1% Triton X-100. Fixed and permeabilized cells were then incubated with TUNEL reaction mixture for 1 hour at 37°C in a humidified chamber. Positive control slides were treated with 300 units/mL DNase I for 10 minutes at 37°C to induce DNA strand breaks before the TUNEL labeling procedure.

Immunohistochemical procedures were used to evaluate proliferation and endothelial microvessel development in tumor tissue. To analyze proliferation, frozen sections were fixed in 4% paraformaldehyde, permeabilized with 0.2% Triton X-100 for 30 minutes at room temperature, and blocked with 5% normal donkey serum for 2 hours at room temperature. A goat anti-Ki-67 primary antibody (1:50) in 1% bovine serum albumin was applied to the tissue and incubated overnight at 4°C. A donkey anti-goat secondary antibody conjugated to rhodamine red-X was then applied (1:800) for 2 hours at room temperature in the dark. For microvessel endothelial cell staining, frozen sections were blocked with 10% normal goat serum and 1% bovine serum albumin for 2 hours at room temperature. A CD31 primary antibody (1:50) in 1% bovine serum albumin was applied to the tissue and incubated overnight at 4°C. A goat anti-rat secondary antibody conjugated to rhodamine red-X (1:800) was then applied for 2 hours at room temperature in the dark. All tissue sections were then mounted in 4′,6-diamidino-2-phenylindole-containing Vectashield mounting medium and visualized on a Leica confocal microscope using ×63 magnification. All histologic staining was quantified using the Leica confocal microscope software image analysis tools, which quantifies the number of pixels that represents positive staining, denoted “pixelation”.

In vivo pharmacokinetic analysis. To show that pegylated liposomes can be effectively employed to systemically deliver short-chain ceramide to solid tumor tissue, we evaluated the pharmacokinetic distribution of i.v. given liposomal-C₆ in 410.4 syngeneic tumor-bearing mice. Using [³H]C₆ as a marker for C₆ delivery, we injected tumor-bearing mice with 10 and 40 mg/kg liposomal-C₆ and removed blood (via cardiac puncture), tumor, spleen, kidney, liver, and heart tissue at chosen time points. Excised tissues were weighed, solubilized with Solvable, and counted using a scintillation counter. Using the specific activity of [³H]C₆ (20 Ci/mmol), we calculated the mass of [³H]C₆ within each tissue from the measured dpm counts. The mass of total C₆ per milligram of tissue (or milliliter of blood) was based on the fixed ratio (0.0026%) of [³H]C₆ to total C₆ in the liposome formulation and the apparent ceramide concentration was calculated for each tissue removed. Control tissue from mice that were not injected with radiolabeled C₆ was removed to subtract background dpm counts from treated tissues, and a pharmacokinetic profile was generated. WinNonLin software (Pharsight Corp., Mountain View, CA) was employed to analyze the pharmacokinetic data using a two-compartment model of drug distribution and elimination. A two-compartment model has wider application than a one-compartment model, because the body is divided into a central compartment with rapid mixing and a peripheral compartment with slower distribution.

Statistical analysis. Differences among two treatment groups were statistically analyzed using a two-tailed Student's t test for statistical analyses. Differences among three or more groups were analyzed using a one-way ANOVA test. A statistically significant difference was reported with Ps indicated where applicable. Data are reported as the mean ± SE from at least three separate experiments. Specifically, for both syngeneic and nude experiments, a quadratic random coefficient model was fit to tumor growth curves for each treatment group (29). The quadratic random coefficient model accounts for the repeated measurements over time. From the quadratic random coefficient model, 2 df contrasts (for the linear and quadratic coefficients) were used to compare any given two tumor growth curves. A Bonferroni correction was made to the Ps to account for the multiple comparisons testing of the tumor growth curves within the syngeneic experiment. All analyses were done using the SAS statistical software, particularly the ROBUSTREG and MIXED procedures (SAS Institute, Inc., Cary, NC).

Liposomal-C₆-ceramide induces adenocarcinoma cellular cytotoxicity and apoptosis. The liposomal-C₆ formulation was first tested for its in vitro cytotoxic effects in murine 410.4 mammary adenocarcinoma cells. The delivery of C₆ in our pegylated liposomal formulation reduced the IC₅₀ >2-fold from 12 to 5 μmol/L for nonliposomal-C₆ and liposomal-C₆, respectively. Our pegylated formulation displayed an improved dose-response inhibition of growth in 410.4 cells compared with nonliposomal “free” administration of C₆ in DMSO vehicle (Fig. 1A), thus indicating improved potency and efficacy. A pegylated “ghost” formulation without C₆ did not significantly affect 410.4 cell cytotoxicity, confirming previous studies noting that pegylated components (PEG(750)-C₈ and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy PEG(2000)]) are not cytotoxic (21). Similar cytotoxicity profiles were observed in another highly aggressive and metastatic murine adenocarcinoma cell line, T41 (ref. 30; data not shown), alluding to the potential applicability of liposomal-C₆ delivery as an experimental therapeutic. This in vitro study indicates that a pegylated liposomal formulation designed for systemic drug delivery is an effective vehicle for C₆-mediated induction of murine 410.4 mammary adenocarcinoma cell cytotoxicity.

In contrast to the murine 410.4 mammary adenocarcinoma cells, normal murine HC11 mammary epithelial cells did not show increased cytotoxicity to either liposomal-C₆ or “free” C₆ in DMSO vehicle (Fig. 1B). However, as a positive control, DMSph was toxic to the HC11 cells. DMSph is a known proapoptotic agent that shares the characteristic sphingoid backbone of all sphingolipids, including ceramide. DMSph inhibits sphingosine kinase activity, ultimately resulting in accumulation of ceramide as well as reduction in the prosurvival metabolite sphingosine-1-phosphate (31). This study shows the potential utility of C₆ as a chemotherapeutic with limited side effects, because the liposomal-C₆ formulation preferentially targets tumorigenic versus normal mammary epithelial tissue, a finding consistent with other in vitro studies using ceramide analogues (32, 33).

We next investigated if C₆-dependent growth inhibition correlated with enhanced apoptosis in 410.4 cells. To confirm that C₆ delivery leads to cell apoptosis, we did Annexin V staining of treated cycling 410.4 cells. Annexin V stains phosphatidylserine on the outer leaflet of apoptotic cells, a hallmark of apoptosis. Following a 48-hour treatment at 25 μmol/L, liposomal-C₆ induced a significantly greater amount of Annexin V staining compared with nonliposomal-C₆, whereas the ghost formulation had no effect (Fig. 1C). A TUNEL assay was used to confirm the Annexin V results. In a similar fashion, liposomal-C₆ induced a significant increase in TUNEL-positive 410.4 cells when compared with ghost formulations (Fig. 1D). Together, these cytotoxicity and apoptosis studies indicate that liposomal delivery of C₆ limits clonal expansion by promoting apoptosis in the 410.4 mammary adenocarcinoma cell line.

Liposomal-C₆-ceramide treatment allows preferential accumulation of C₆-ceramide in caveolae. We have shown previously that liposomal-C₆ partitions out of the liposomal bilayer into the plasma membrane bilayer without associated liposome/cell membrane fusion (21). We now extend these studies to assess if liposomal-C₆ accumulates within membrane rafts, sites of ceramide-regulated signaling cascades. Caveolin-1-enriched lipid rafts, or caveolae, are known to serve as signaling docking platforms for a multitude of signaling cascades (34, 35). We used trace amounts of [³H]C₆ within the liposomal formulation to assess time-dependent cellular accumulation of 25 μmol/L C₆ in caveolin-1-enriched lipid rafts. Within 15 minutes, ceramide was observed to preferentially accumulate within fractions 4 to 5 of a sucrose gradient in which caveolin-1 was found to be localized (Fig. 2A). It was evident that as time progressed, a percentage of the delivered C₆ migrated to non-caveolin-1-enriched fractions. Consistent with previous studies (21), total accumulation of ceramide within cell lysates (fractions 1-12) was linear over a 6-hour time period. We confirmed these data with NBD-C₆, demonstrating that that both radiolabeled and fluorolabeled C₆ accumulate within caveolae. Specifically, following a 2-hour treatment, 69.1% of total cellular NBD-C₆ (compared with 63.5% of total cellular [³H]C₆) were observed to accumulate within caveolin-1-enriched fractions (4-5) of the sucrose gradient. Thus, the delivery of liposomal-C₆ allows for the early preferential accumulation of ceramide into caveolae, which have been implicated as potential initiation sites for intracellular signaling cascades, including those that ultimately lead to cellular apoptosis (34, 35).

Liposomal-C₆-ceramide treatment delivers C₆-ceramide to mitochondria. Studies have shown that ceramide acts on the mitochondria to initiate antiproliferative and proapoptotic effects (36). Thus, we hypothesized that exogenous C₆ treatment would result in an accumulation of ceramide in the mitochondria of 410.4 adenocarcinoma cells. Using MitoTracker Deep Red as a marker for mitochondria, we showed that 25 μmol/L liposomal-NBD-C₆ colocalized with this mitochondrial marker (Fig. 2B), signifying the accumulation of exogenous ceramide. This study indicates that liposomal-C₆ localizes within the mitochondria, possibly contributing to mitochondrial dysfunction and resultant cell death.

Liposomal-C₆-ceramide treatment displays dose-dependent antitumor activity in a syngeneic tumor model. To study the in vivo anticancer activity of systemically delivered liposomal-C₆ on solid tumor tissue, we inoculated 410.4 cells in the right hind flank of female BALB/c mice to induce solid tumors. Preliminary toxicology studies showed that female BALB/c mice tolerated a single i.v. liposomal-C₆ dose of up to 100 mg/kg with no evident side effects (data not shown). However, i.v. treatment with 10 mg/kg nonliposomal “free” C₆ in an 80% ethanol solution was lethal in 50% of the mice (LD₅₀, 10 mg/kg), whereas 80% ethanol alone displayed no lethality. The acute, dose-limiting toxicity observed with free C₆ may be attributed to the extreme hydrophobicity of the molecule and its propensity to precipitate immediately in aqueous solutions and possibly induce an occlusion. Thus, liposomal delivery of C₆ significantly reduces the LD₅₀ when given i.v. in female BALB/c mice. It seems that in addition to serving as a suitable drug delivery vehicle our liposomal formulation has the further benefit of delivering ceramide in a much less toxic form.

Tumor-bearing mice were treated on an alternate day schedule with liposomal-C₆ or ghost liposomes via tail vein injections. Liposomal-C₆ treatments showed a significant dose-response inhibition of tumor growth compared with ghost and untreated controls (Fig. 3A). The greatest degree of tumor growth suppression was observed with the 36 mg/kg liposomal-C₆ formulation. All doses of the ghost formulation had no appreciable effects on tumor volume. The most significant tumor growth suppression was evident near the cessation of treatment, suggesting that liposomal-C₆ maintained its anticancer effects throughout the treatment period. At the cessation of tumor treatment on day 21, the mean tumor volume of the 36 mg/kg ghost-treated mice was ∼6-fold higher than that of the 36 mg/kg liposomal-C₆-treated mice in the 410.4 solid tumor model (P < 0.01). Moreover, as the treatment period progressed, the average mouse weight for liposomal-C₆-treated mice did not vary significantly from ghost-treated or untreated mice (Fig. 3B), thus signifying that the liposomal-C₆ treatment regimen did not display systemic toxicity to the tumor-bearing mice.

Liposomal-C6-ceramide treatment induces tumor apoptosis and reduces cellular proliferation in vivo. To validate that liposomal-C₆ induced apoptosis, resulting in diminished tumor growth, we did an immunofluorescent analysis of frozen tumor sections. Tumor sections were stained using an in situ TUNEL assay, which stains the nicked ends of DNA, a marker for cellular apoptosis. Tumor tissue from mice treated with liposomal-C₆ (Fig. 4A-4) had significantly more TUNEL staining than those of ghost-treated (Fig. 4A-3) or untreated mice (Fig. 4A-2). As a positive control, DNase I–treated tissue sections displayed marked TUNEL staining (Fig. 4A-1). Liposomal-C₆ treatment resulted in nearly a 20-fold increase in cellular apoptosis when compared with ghost treatment (Fig. 4B). Consistent with these apoptosis data, we also show that liposomal-C₆ (Fig. 4C-3) is more efficacious than ghost (Fig. 4C-2) or untreated mice (Fig. 4C-1) as measured by Ki-67 staining for proliferation. In fact, liposomal-C₆ treatment resulted in an ∼40% decrease in cellular proliferation (Fig. 4D). These studies show that the liposomal delivery of C₆ to solid tumor tissue induces significant apoptotic death, leading to limited tumor growth.

Liposomal-C₆-ceramide treatment reduces tumor microvessel formation in vivo. Further histologic analysis of frozen tumor tissue showed that liposomal-C₆ treatment (Fig. 4E-3) resulted in significantly less staining for CD31, a marker for endothelial cells, when compared with ghost-treated (Fig. 4E-2) or untreated mice (Fig. 4E-1). Tumors treated with liposomal-C₆ displayed nearly a 50% decrease in CD31 staining when compared with ghost-treated and untreated mice (Fig. 4F). We propose that in addition to inducing cellular apoptosis liposomal-C₆ inhibits the development of a microvessel network within solid tumor tissue, thus restricting the angiogenic-dependent growth of the solid tumor mass.

Liposomal-C₆-ceramide treatment displays favorable pharmacokinetic variables, consistent with its antitumor efficacy. To quantify the concentration of C₆ in various tissues following systemic i.v. administration in the 410.4 syngeneic tumor model, we again incorporated trace amounts of [³H]C₆ into the liposomal-C₆ formulation to serve as a marker for total C₆. Blood concentrations of liposomal-C₆ seemed to follow linear first-order kinetics following i.v. administration (Fig. 5A). Using a 40 mg/kg dose, blood concentrations of liposomal-C₆ were maintained well above the in vitro IC₅₀ concentration (∼5 μmol/L) throughout a 48-hour period. This concentration of liposomal-C₆ in the blood allows for the accumulation and maintenance of bioactive concentrations in the solid tumor tissue (Fig. 5B). Table 1 presents the tissue distribution profile in the blood, tumor, liver, kidney, spleen, and heart over a 48-hour period following i.v. administration of 40 mg/kg liposomal-C₆. Although a relative steady-state bioactive concentration of liposomal-C₆ is achieved within the tumor tissue, it is apparent that ceramide is rapidly cleared from the most highly perfused, first-pass organs that we analyzed. Using a two-compartmental model of distribution, Table 2 presents the associated pharmacokinetic variables of systemic liposomal-C₆ delivery in the blood. It is noteworthy that based on this kinetic modeling analysis that the apparent t_1/2 is ∼11 hours for the liposomal-C₆ formulation, further supporting the utility of these drug delivery vehicles. It is therefore concluded that liposomal-C₆ delivery not only improves the solubility of C₆ and fosters accumulation within tumor tissue but also allows for the sustained/controlled release of the drug over time, reducing the acute toxicity of C₆ and limiting tumor growth in vivo.

Liposomal-C₆-ceramide treatment displays antitumor activity in a human xenograft tumor model. To confirm and extend the above results from the syngeneic solid tumor model, we next evaluated the efficacy of a 36 mg/kg dose of liposomal-C₆ in a human xenograft nude mouse tumor model of breast adenocarcinoma. We chose human MDA-MB-231 breast adenocarcinoma cells, as we have shown previously that this highly metastatic, estrogen receptor–negative cell line responds to liposomal-C₆ with marked apoptosis in vitro (21). In vitro ceramide-induced apoptosis was validated previously using multiple methodologies, including Annexin V staining, TUNEL analysis, caspase stimulation, and propidium iodide labeling (21). We now report that the i.v. administration of liposomal-C₆, at an initial 36 mg/kg dose, statistically limits xenograft tumor growth upward of 5 weeks (Fig. 6). Despite an early reduction in tumor growth at this dose, it seemed that tumors developed some degree of resistance to treatment at approximately day 30. Thus, we escalated the dose to 72 mg/kg on day 32 for the duration of the study to potentially improve efficacy without inducing toxicity. This higher dose seemed to result in a short-term effect on efficacy without inducing apparent systemic toxicology. Although we show for the first time that systemic liposomal-C₆ delivery can limit tumor growth in both syngeneic and human xenograft models, it may require a combinatorial therapeutic approach to fully limit or regress solid tumors. It can be envisioned that additional chemotherapeutic agents with distinct mechanisms of action can be coencapsulated within ceramide-incorporated liposomes.

Many studies, including our own, have shown that ceramide can effectively induce apoptosis in tumor cells in vitro (23, 37). However, the potential chemotherapeutic utility of ceramide is limited by its inherent hydrophobicity. Using syngeneic and human xenograft models of breast adenocarcinoma, we now show that C₆-incorporated stable liposomes are an optimal “solution” for selective, bioefficacious, systemic delivery of ceramide. We designed the formulation to carry a large ratio of C₆, a bilayer-destabilizing lipid, to maximize the drug payload without sacrificing vesicular stability. The inclusion of PEG(750)-C₈ stabilized the formulation, allowing for the incorporation of up to 30 molar percent C₆. PEG(750)-C₈ is composed of a 750 molecular weight PEG head group conjugated to the free hydroxyl of C₈-cermamide. Without the PEG(750)-C₈, stable vesicles containing 30 molar percent C₆ could not be prepared or extruded. Additionally, the inclusion of PEG(750)-C₈ has been observed to facilitate time-release properties of liposomal bilayers with the added benefit of enhanced bioavailability due to the PEG chains (38).

Using this liposomal formulation, we investigated the membranous localization of liposomal-C₆ in 410.4 murine adenocarcinoma cells. We observed two distinct pools of exogenously delivered ceramide that are consistent with the ability of ceramide to induce apoptosis in breast cancer cells. First, endogenous ceramide preferentially accumulates within caveolin-1-enriched lipid rafts, structured microdomains that regulate the dynamic interaction between signaling molecules (34, 39, 40). In addition, the spontaneous ability of endogenous ceramide to partition into and expand lipid raft microdomains may also regulate signaling cascades within these lipid rafts (37, 40, 41). Ceramide-mediated inhibition of the prosurvival kinase Akt seems to occur within these caveolin-1-enriched domains.¹

Ceramide has been shown to target and activate protein kinase Cζ (17), which phosphorylates Akt at Thr³⁴ (42), contributing to the observed apoptotic cell death (43). A second putative mechanism responsible for ceramide-induced apoptosis of mammary cancer cells is accumulation of ceramide within the mitochondria. It is now well established that ceramide localization within mitochondria is responsible for mitochondrial dysfunction, including loss of electron potential and cytochrome c release, prerequisites for caspase activation (36, 44, 45). The fact that liposomally delivered exogenous C₆ accumulates within caveolin-1-enriched lipid rafts and mitochondria suggests that these two observations may be linked. It has been proposed that ceramide may use caveolae-dependent trafficking as an intracellular delivery route for mitochondrial accumulation (34, 35, 46). In fact, Li et al. have identified an intracellular depot for caveolin-1 in isolated mitochondria, suggesting the existence of a novel pathway of intracellular molecular trafficking, which may be used for delivering specific lipids to multiple cellular compartments (46). Interestingly, caveolin-1 has also been shown to be up-regulated in multidrug-resistant cancer cells (47) and has been suggested to be involved in an intracellular drug trafficking mechanism to shuttle exogenous agent to the plasma membrane where they may be exported by drug efflux pumps. It is possible that caveolin-mediated endocytosis and intracellular trafficking may serve as a pharmacodelivery system for ceramide, allowing for its delivery and accumulation within mitochondria in murine 410.4 mammary adenocarcinoma cells.

Consistent with apoptosis, systemic delivery of liposomal-C₆ induces cytotoxicity of adenocarcinoma solid tumors through decreased cell proliferation (Ki-67 staining) and diminished microvascularization (CD31 staining). Ceramide has been shown to decrease endothelial cell growth and induce apoptosis in vitro (23, 48–50). In addition, elevating endogenous ceramide through radiotherapy induces endothelial cell apoptosis in vivo (49, 51, 52). In addition, ceramide signaling is required for fenretinide-induced endothelial cell apoptosis and represents a molecular mechanism for the antiangiogenic effects of fenretinide (53). It is also possible that reduced CD31 staining is due to the immature vascular network of smaller ceramide-treated tumors. Regardless of mechanism, our studies are the first demonstration that exogenous short-chain ceramide can be systemically delivered to solid tumor tissue, resulting in significantly reduced tumor microvessel formation as well as induced apoptosis and diminished cellular proliferation.

In agreement with in vitro data where liposomal-C₆ was found to be preferentially cytotoxic to murine 410.4 mammary adenocarcinoma cells and not to normal murine HC11 mammary epithelial cells, we observed minimal toxicities in tumor-bearing mice, suggesting that the bioactive concentrations of ceramide achieved in the tumor tissue are not bioactive in normal tissues. Other studies have also suggested that cancerous tissue may generally be more susceptible to the antiproliferative and proapoptotic effects of ceramide analogues than normal tissue (32, 33). Our liposomal formulation may also have fewer side effects due to their pharmacokinetic properties. Steady-state concentrations of ceramide within solid tumors, due to the dysfunctional lymphatic and capillary drainage, may explain preferential adenocarcinoma and associated microvascular cell death. It can be envisioned that the liposomal-C₆ vesicles are trapped in the underdeveloped vasculature of the tumor while being cleared from other tissue compartments. In fact, a two-compartment model calculates a biological half-life of nearly 11 hours, a value consistent with establishing a steady-state therapeutic/cytotoxic dose within the solid tumor tissue. It is critical to note that although the solid tumor tissue is in the peripheral compartment an effective drug concentration is achieved and maintained for a 48-hour treatment period without overt toxicity in the animal.

We have shown systemic efficacy and a favorable pharmacokinetic profile of C₆-incorporated liposomes in two solid tumor models of breast adenocarcinoma. The clinical potential for the delivery of C₆ with additional therapeutic agents in liposomal vesicles is significant. Studies have shown that ceramide may act synergistically with chemotherapeutic agents, such as paclitaxel (54) and fenretinide (55). Thus, delivery of chemotherapeutic agents in C₆-formulated liposomes may further enhance apoptotic actions. Moreover, targeted immunoliposomes are expected to further benefit from C₆ incorporation by enhancing the tumor targeting index. Pegylated liposomal-C₆ vesicles may be conjugated to antibody moieties that target tumor-specific surface antigens, thus permitting targeted therapy of solid tumor and/or metastatic lesions (56, 57). These efforts will be critical in realizing a role for efficient, targeted, and potentially nontoxic ceramide drug delivery in cancer chemotherapy.

We thank Nate Sheaffer for assistance with flow cytometry, Young Shin Kim and Dr. Tao Lowe for assistance with the dynamic light scattering particle size. Dr. Amy Fulton for kindly providing the murine 410.4 mammary adenocarcinoma cells, and Dr. Bernd Groner for kindly providing the murine HC11 mammary epithelial cells.