Pharmaceutical Research

, Volume 30, Issue 7, pp 1883–1895

Improving Intracellular Doxorubicin Delivery Through Nanoliposomes Equipped with Selective Tumor Cell Membrane Permeabilizing Short-Chain Sphingolipids


  • Lília R. Cordeiro Pedrosa
    • Laboratory Experimental Surgical Oncology Section Surgical Oncology, Department of SurgeryErasmus Medical Center
  • Albert van Hell
    • Division of Biological Stress ResponseThe Netherlands Cancer Institute - Antoni van Leeuwenhoek Hospital
  • Regine Süss
    • Department of Pharmaceutical Technology and BiopharmacyAlbert-Ludwigs University
  • Wim J. van Blitterswijk
    • Division of Cellular BiochemistryThe Netherlands Cancer Institute - Antoni van Leeuwenhoek Hospital
  • Ann L. B. Seynhaeve
    • Laboratory Experimental Surgical Oncology Section Surgical Oncology, Department of SurgeryErasmus Medical Center
  • Wiggert A. van Cappellen
    • Optical Imaging CentreErasmus Medical Center
  • Alexander M. M. Eggermont
    • Laboratory Experimental Surgical Oncology Section Surgical Oncology, Department of SurgeryErasmus Medical Center
    • Institut de Cancerologie Gustave Roussy
  • Timo L. M. ten Hagen
    • Laboratory Experimental Surgical Oncology Section Surgical Oncology, Department of SurgeryErasmus Medical Center
  • Marcel Verheij
    • Division of Biological Stress ResponseThe Netherlands Cancer Institute - Antoni van Leeuwenhoek Hospital
    • Department of RadiotherapyThe Netherlands Cancer Institute - Antoni van Leeuwenhoek Hospital
    • Laboratory Experimental Surgical Oncology Section Surgical Oncology, Department of SurgeryErasmus Medical Center
Research Paper

DOI: 10.1007/s11095-013-1031-6

Cite this article as:
Pedrosa, L.R.C., van Hell, A., Süss, R. et al. Pharm Res (2013) 30: 1883. doi:10.1007/s11095-013-1031-6



To improve nanoliposomal-doxorubicin (DoxNL) delivery in tumor cells using liposome membrane-incorporated short-chain sphingolipids (SCS) with selective membrane-permeabilizing properties. DoxNL bilayers contained synthetic short-chain derivatives of known membrane microdomain-forming sphingolipids; C8-glucosylceramide (C8-GluCer), C8-galactosylceramide (C8-GalCer) or C8-lactosylceramide (C8-LacCer).


DoxNL enriched with C8-GluCer or C8-GalCer were developed, optimized and characterized with regard to size, stability and drug retention. In vitro cytotoxic activity was studied in a panel of human tumor cell lines and normal cells. Intracellular Dox delivery was measured by flow cytometry and visualized by fluorescence microscopy. For a further understanding of the involved drug delivery mechanism confocal microscopy studies addressed the cellular fate of the nanoliposomes, the SCS and Dox in living cells.


C8-LacCer-DoxNL aggregated upon Dox loading. In tumor cell lines SCS-DoxNL with C8-GluCer or C8-GalCer demonstrated strongly increased Dox delivery and cytotoxicity compared to standard DoxNL. Surprisingly, this effect was much less pronounced in normal cells. Nanoliposomes were not internalized, SCS however transfered from the nanoliposomal bilayer to the cell membrane and preceded cellular uptake and subsequent nuclear localization of Dox.


C8-GluCer or C8-GalCer incorporated in DoxNL selectively improved intracellular drug delivery upon transfer to tumor cell membranes by local enhancement of cell membrane permeability.


doxorubicin-nanoliposomeshort chain sphingolipidTumor cell membrane permeabilization








D:PL ratio

Drug to phospholipid initial mass ratio




Fetal calf serum


Hydrogenated soy phosphatidylcholine




Mean of fluorescence intensity






Polydispersity index


Short chain sphingolipids


Standard error of the mean


Tumor cell membrane modulation


Insufficient uptake of chemotherapeutic agents by tumor cells remains an important limitation in clinical cancer treatment. Several factors play a role such as suboptimal dosing due to systemic toxicity, rapid drug clearance from circulation and limited drug traversal across the tumor cell membrane. Here we addressed these hurdles using advanced drug delivery technologies. Liposomes, small nanovesicles composed of phospholipids, cholesterol and poly(ethylene glycol) (PEG)-lipids, are used to entrap the chemotherapeutic agent prolonging systemic drug concentrations and reducing dose-limiting toxicities (1). In addition, nanoliposomes (NL), due to their small size (< 100 nm), can accumulate in tumor tissue by virtue of the enhanced permeability and retention effect (EPR) (25). In this study nanoliposomal delivery is combined with a novel strategy using short-chain sphingolipids (SCS) to enhance transmembrane drug transport (68).

Sphingolipids are key molecules for assembly of microdomains in the cell membrane, (911). Evidence exists that SCS self-association may lead to domain or channel formation in the membrane (12) and may explain the enhanced passage of amphiphilic drugs across cell membranes (6). In this study we exploited this property of SCS to develop pegylated-nanoliposomal doxorubicin (DoxNL) formulations.

Dox is a chemotherapeutic agent whose mode of action includes intercalation between adjacent base pairs of the DNA double helix, binding to DNA-associated enzymes such as topoisomerase, and effects on membranes (13,14). DoxNL (Caelyx®/Doxil®) increased drug accumulation in solid tumors and reduced dose-limiting toxicities such as myelosuppression and cardiotoxicity (1517). DoxNL is currently approved for use in AIDS-related Kaposi’s sarcoma (18), refractory ovarian cancer (19), myeloma (20) and metastatic breast cancer (16,18,21). Although DoxNL, due to its favorable pharmacokinetic profile and small size, accumulates in tumors, its ability to deliver the Dox content to its active site intracellularly (bioavailability) remains a main issue. DoxNL is characterized by high stability to prevent drug release in circulation, causing slow and suboptimal drug delivery to tumor cells upon accumulation in the tumor area (22,23). The inadequate Dox release together with its slow intracellular uptake likely are main reasons for the limited efficacy increase of DoxNL in cancer patients (16,18,24). In order to improve intracellular Dox delivery we combined nanoliposomal drug delivery with the concept of tumor cell membrane modulation (TCMM) using SCS.

Regarding the mechanism of enhanced drug delivery from SCS enriched nanoliposomes we hypothesize that they transfer spontaneously from the NL bilayer to cell membranes creating enhanced cellular accessibility for amphiphilic compounds by the formation of specific domains with increased drug permeability or transient pores which improve drug influx. The aim of this study is to further optimize and explore SCS-based nanoliposomal drug delivery using various synthetic SCS and investigate the involved drug delivery mechanism.

Materials and Methods


Hydrogenated soy phosphatidylcholine (HSPC) and distearylphosphatidylethanolamine (DSPE)-PEG2000 were from Lipoid (Ludwigshaven, Germany). Short chain sphingolipids, C8-glucosylceramide (C8-GluCer), C8-galactosylceramide (C8-GalCer), C8-lactosylceramide (C8-LacCer) and fluorescent lipid NBD PE (1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-1,3-benzoxadiazol-4-yl) (ammonium salt) and C6-NBD Galactosyl Ceramide N-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-D-galactosyl-ß1-1’-sphingosine were from Avanti Polar Lipids (Alabaster, AL, USA).

Polycarbonate filters were from Northern Lipids (Vancouver, BC, Canada) and PD-10 Sephadex columns were from GE Healthcare (Diegem, Belgium). Dox-HCl (Dox) was from Pharmachemie (Haarlem, The Netherlands). Cholesterol, HEPES (2-[4-(2-hydroxyethyl)piperazin-1-yl] ethanesulfonic acid), trichloroacetic acid (TCA), acetic acid, Triton-X, sulforhodamine B (SRB) were from Sigma Aldrich (Zwijndrecht, The Netherlands). Hoechst was from Molecular Probes (Leiden, The Netherlands). PBS was from Boom and FACS flow fluid from BD Biosciences.

Preparation of SCS-NL

Nanoliposomes were formulated of HSPC/Cholesterol/DSPE-PEG2000 in a molar ratio of 1.85: 1: 0.15. Liposomes of 85–100 nm in diameter were prepared by lipid film hydration and extrusion at 65°C. Drug loading was based on ammonium sulfate gradient method (25). To the mixture of lipids, 0.05, 0.1 or 0.15 mol of SCS was added per mole of total amount of lipid (including cholesterol) and liposomes were formulated as described previously (8). Dox was added to liposomes in different drug to phospholipid ratios (w/w) - 0.25:1, 0.2:1, 0.15:1 and 0.1:1 for 1 h at 65°C. Non encapsulated Dox was separated by ultracentrifugation at 29000 rpm in a Beckman ultracentrifuge (Ti50.2 rotor). The liposome pellet was resuspended in buffer containing 135 mM NaCl, 10 mM HEPES, pH 7.4.

Size and polydispersity index (pdi) were determined by light scattering using a Zetasizer Nano ZS (Malvern Instruments, Malvern, UK). Lipid concentration was measured by phosphate assay (26).

NBD PE was incorporated in the formulation at 0.25 mol% of total lipid as a stable fluorescent bilayer marker. Studies assessing the fate of liposomal SCS used a similar mol% of bilayer incorporated C6-NBD-Galactosyl Ceramide.

Loading Efficiency

After separation of free non-encapsulated Dox from nanoliposomal encapsulated Dox, the amount of entrapped Dox was measured by fluorimetry (λExitation 475 nm; λEmission 590 nm). Total amount of drug was measured after entire liposome solubilization with 1% (v/v) Triton in water to a sample from stock liposomal Dox.


Long-term storage conditions stability (4°C) was based on size, pdi and Dox content measurements for a period of at least 6 months. All measurements were performed in triplicate. Dox content was measured by fluorimetry (λExitation 475 nm; λEmission 590 nm), after separation of free/entrapped drug by gel filtration chromatography. Stability at 37°C, in the presence and absence of serum was studied for a period of 24 h by fluorimetry. Total drug release was measured after entire liposomal solubilization by adding 1% (v/v) of Triton-X. The percentage of drug content was calculated following the formula:
$$\frac{{100 - {\left( {{\text{Flourescence}}_{{{\text{sample}}}} - {\text{blank}}} \right)}}} {{{\text{Flourescence}}_{{{\text{total \, release}}}} - {\text{blank}}}} \times 100 $$

Cryo-transmission Electron Microscopy (TEM)

Cryo-TEM was used to characterize the detailed structure of the liposomal formulations (non-enriched, 10 mol% C8-GluCer and 10 mol% C8-GalCer-NL) of Dox and the physical state of the encapsulated drug, as previously described (27,28). The freezing was performed in a cooling chamber, which was permanently cooled with liquid nitrogen. A sample droplet was placed on a microperforated copper grid and blotted by a filter paper to result in a thin liquid film. The grid was plunged into liquid ethane for immediate freezing. A Leo 912 Omega TEM microscope (Carl Zeiss NTS GmbH, Oberkochen, Germany) was used.

Cell Culture

In vitro anti-tumor activity was studied towards a panel of human tumor cell lines: BLM and Mel57 melanomas, MCF7 and SKBR3 breast carcinoma and ASPC1 and Panc1 pancreatic carcinomas. All tumor cell lines were cultured in Dulbeccos’s modified Eagle medium, supplemented with 10% fetal calf serum and 4 mM L-glutamine. HUVEC were isolated by collagenase digestion using the method described by Jaffe et al. (29) and cultured in HUVEC medium containing human endothelial serum free medium (Invitrogen), 20% heat inactivated newborn calf serum (Cambrex), 10% heat inactivated human serum (Cambrex), 20 ng/ml human recombinant epidermal basic fibroblast growth factor (Peprotech EC Ltd), 100 ng/ml human recombinant epidermal growth factor (Peprotech EC Ltd) in fibronectin (Roche Diagnostics) coated flasks. Fibroblasts (3T3) were purchased from Biowhitakker and cultured in Dulbeccos’s modified Eagle medium containing nutrient mixture F12, supplemented with 10% fetal calf serum and 4 mM L-glutamine.

Cells were subcultured weekly by trypsinization when a confluency of 80–90% was achieved and maintained in a water saturated atmosphere of 5% CO2 at 37°C.

Cell Toxicity

All cell lines were plated in flat bottom 96 well-plates. Tumor cells were seeded at cell densities obtained from respective tumor growth curves to ensure use of cells in their log-phase of growth. BLM and Mel57 melanoma cells were seeded at a density of 3 × 103 and 6 × 103 cells/well, respectively. 1.25 × 104 cells/well were used for breast carcinoma cells MCF7 and SKBR3. For ASPC1 and Panc1 pancreatic carcinoma cell lines 6 × 103cells/well were seeded. HUVEC and 3T3 fibroblast cells were seeded at a density of 1 × 104 cells/well and 5 × 103cells/well, respectively, in order to achieve the same confluency.

After 24 h, cells were exposed to serial concentrations of Dox-liposomal formulations (SCS enriched and non-enriched) in culture medium for 24 h.

Cell survival was determined by measuring total cellular protein levels using the sulforhodamine-B (SRB) assay (30). Cells were washed twice with PBS, incubated with 10% trichloric acetic acid (1 h, 4°C) and washed again. Cells were stained with 0.4% SRB (Sigma) for 15 min, washed with 1% acetic acid. After drying, protein-bound SRB was dissolved in TRIS buffer (10 mM, pH 9.4) and absorbance was measured at a wavelength of 540 nm in a plate reader. Cell survival was calculated as a percentage relative to control (untreated cells), which was set at 100%.

IC50 was determinate for each Dox liposomal formulation, by plotting the cell survival observed for each concentration versus the log of the concentration and fitting a non linear regression curve using GraphPad Prism software v5.0. All experiments were performed in triplicates and were repeated at least 3 times, independently.

Cell survival quantification was complemented by evaluation of cellular morphology. Following the same cellular concentration from SRB assay, BLM, Mel57, MCF7, SKBR3, Panc1 and ASPC1 cells were plated in a 24 well plate and allow to growing for 24 h. The cells were treated with Dox (1 μM, 10 μM or 100 μM) in its non-enriched and enriched (C8-GluCer and C8-GalCer) liposomal form. After 24 h cell morphology was examined with a 10× N.A 0.30 Plan Neofluar objective lens using an inverted Zeiss Axiovert 100M microscope equipped with an Axiocam digital camera (Carl Zeiss).

Intracellular Drug Uptake by Flow Cytometry

Intracellular Dox levels in BLM, MCF7 and ASPC1 cells were measured by flow cytometry. Cells were seeded on flat bottom 24 well plates at a final concentration of 6 × 104 cells/well and incubated for 24 h. Non-enriched and SCS enriched-DoxNL, with C8-GluCer or C8-GalCer, diluted in growth medium were added in a drug concentration of 1 μM and 10 μM, after which cells were incubated for 1, 4 and 24 h. After incubation, cells were washed twice to discard non-incorporated drug and trypsinized for 2 min. Cell suspensions were washed twice in medium and resuspended in PBS. Cellular uptake was measured on a Becton Dickinson FACScan using Cell Quest software. Excitation was set at 488 nm and detection by FL2 detector channel.

Intracellular Drug Uptake by Fluorescence Microscopy

MCF7, BLM and ASPC1 cells in a concentration of 1 × 105 cells/well were seeded on a cover glass coated with 0.1% collagen and incubated for 24 h. Hoechst (1:100) was added for 20 min cells and washed twice before adding 10 μM of Dox in form of non enriched-DoxNL or SCS-enriched DoxNL. Cells were incubated at 37°C for 1 and 4 h and using fluorescence microscopy intracellular drug uptake was evaluated in living cells, using a filter set consisting of a 450–490 nm band pass excitation filter for Hoechst and a 510 nm beam splitter and a 520–543 nm long pass excitation filter for Dox. Cells were imaged using a 40× oil immersion objective.

Intracellular Drug Uptake by Confocal Microscopy

BLM melanoma, MCF7 breast carcinoma and ASPC1 pancreatic carcinoma cells were cultured on a cover glass of 25 mm diameter coated with 0.1% collagen and incubated for 24 h in cultured medium. Cell density was 1 × 105 cells for BLM melanoma cells, MCF7 breast carcinoma and ASPC1 pancreatic carcinoma cells. Non-enriched liposomal Dox and SCS enriched liposomal Dox was added in a concentration of 10 μM. Cells were incubated for 4 h. Dox uptake was studied using a Zeiss LSM 510 META confocal microscope using 488 argon laser for NBD-PE (500–550 nm band pass filter), a 405 nm Diode laser for Hoechst (420–480 nm band pass filter) and a 543 nm Helium-Neon laser for Dox (560–615 nm band pass filter). Images of the different incubations were taken using identical settings for laser power and photomultipliers.

Intracellular Fate of SCS and Doxorubicin

Labelled NBD-C6-GalCer-DoxNL was followed in time in BLM melanoma cells (1 × 105 cells) during treatment for 40 min. A Leica SP5 (Leica Mannheim) confocal microscope was used with 488 nm excitation and BP 500–550 nm emission of the SCS and 561 nm excitation and BP 570 nm–650 nm emission for Dox with a 63× plan apo (na 1.4) lens. Six optical planes were recorded with an interval of 1.22 μm and a pinhole at 1 airy unit.

Doxorubicin was giving by direct excitation with the 488 excitation line, a crosstalk in the SCS emission channel. The amount of crosstalk was measured (35%) and the SCS images were corrected for this crosstalk. To remove some photon noise from the images, a Gaussian filter was applied to the time lapse images with ImageJ (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA,, 1997–2011). To increase the weak SCS signal a contract stretch was performed.


Statistical analysis was performed using the ANOVA test. P-values less than 0.05 were considered statistically significant. All parametric values are expressed as mean ± standard error of the mean (SEM). Calculations were performed using GraphPad Prism v5.0.


Optimizing C8-GluCer-DoxNL

C8-GluCer-DoxNL with optimal density of the SCS at 10 mol% (8) were optimized with regard to drug loading. Formulations were loaded at Dox:PL ratios of 0.25:1, 0.2:1, 0.15:1 or 0.1:1 (w/w). The size of the final liposomes after loading was smaller than 100 nm and the pdi ≤ 0.1 and a loading efficiency > 90% was achieved at a Dox:PL ratio of 0.1:1 for both C8-GluCer-DoxNL and DoxNL (Table I). To study the effect of Dox:PL ratio on in vitro anti-tumor efficacy, 10 mol% C8-GluCer–DoxNL loaded with the different Dox:PL ratios were tested in human BLM melanoma (Fig. 1a) and human MCF7 breast carcinoma (Fig. 1b) cell lines. Treatment with 10 mol% C8-GluCer–DoxNL for 24 h, decreased cell viability at increasing DoxNL concentration, but was not affected by different Dox:PL ratios in either cell line (Fig. 1 and Supplementary Material Table SI).
Table I

Characterization of C8-GluCer Enriched Liposomes Loaded at Different Drug to Phospholipid Mass Ratios

Drug:PL (w/w)i


Size (nm) ± SEM

pdi ± SEM

% Load ± SEM



109 ± 6.7

0.08 ± 0.01

79 ± 4.1

10% C8-GluCer

99 ± 5.7

0.08 ± 0.01

74 ± 4.6



95 ± 4.9

0.1 ± 0.03

68 ± 10.2

10% C8-GluCer

98 ± 7.8

0.19 ± 0.04

69 ± 10.9



89 ± 3.7

0.09 ± 0.01

92 ± 2.0

10% C8-GluCer

86 ± 2.4

0.09 ± 0.02

78 ± 8.2



89 ± 2.0

0.07 ± 0.01

96 ± 5.7

10% C8-GluCer

84 ± 1.1

0.07 ± 0.01

91 ± 5.0
Fig. 1

In vitro efficacy study of C8-GluCer-DoxNL loaded with doxorubicin at different Dox:PL ratios (w/w), 0.25: 1 (●), 0.2:1 (■), 0.15:1 (▲) and 0.1:1 (▼). Different Dox:PL loading ratios did not affect the efficacy of enriched C8-GluCer-DoxNL in a BLM, melanoma and b MCF7 breast carcinoma cell line when incubated for 24 h at 37°C. Cell survival was analyzed by SRB assay and values represent the mean ± SEM, n ≥3.

To assess C8-GluCer-DoxNL stability, drug retention was studied during long-term storage at 4°C and under cell culture conditions at 37°C in the presence of 10% FCS. During long-term storage, C8-GluCer-DoxNL had similar stability as DoxNL, showing a minor release of ≤ 10% Dox over 6 months (Table II). Also under culture conditions (37°C with 10%FCS), C8-GluCer-DoxNL efficiently retained more than 90% of their contents (Table II and Supplementary Material Fig. S1), and was comparable to DoxNL.
Table II

Characterization and Stability of Different Glycoceramide-Enriched Liposomes

Drug:PL (w/w)i


Size (nm) ± SEM

pdi ± SEM

%Load ± SEM

% Encapsulated Dox (after 6 months at 4°C) ± SEM

% Encapsulated Dox (after 15 h at 37°C) ± SEM



89 ± 2.0

0.07 ± 0.01

96 ± 5.7

91 ± 4.4

91 ± 9.1

10% C8-GluCer

84 ± 1.1

0.07 ± 0.01

91 ± 5.0

88 ± 5.3

98 ± 1.9

10% C8-GalCer

92 ± 4.0

0.09 ± 0.02

96 ± 2.8

91 ± 3.3

100 ± 9.5

10% C8-LacCer

482 ± 36.2

0.47 ± 0.07

9 ± 2.9

nd *



10% C8-GalCer

93 ± 4.0

0.16 ± 0.05

98 ± 9.5



10% C8-LacCer

395 ± 29.0

0.58 ± 0.07

7 ± 2.9



* nd not determined

Novel DoxNL Enriched with C8-GalCer or C8-LacCer

In addition to C8-GluCer, two other short-chain C8-glycosphingolipids, with galactosyl (C8-GalCer) and lactosyl (C8-LacCer) head groups, (Fig. 2) were evaluated as possible alternative SCS to enhance drug uptake from DoxNL. Formulations were characterized with regard to size, pdi, loading efficiency and stability. At a SCS density of 10 mol% and loaded with Dox at a 0.1:1 Dox:PL (w/w) ratio, C8-GalCer-DoxNL presented ideal size (<100 nm) and pdi (<0.1), high loading efficiency and a long-lasting stability. At a higher Dox:PL ratio of 0.2:1, Dox loading of C8-GalCer-DoxNL was still > 90%, but pdi increased to 0.16 (Table II). In contrast, C8-LacCer-DoxNL could not be prepared successfully at either ratio as they strongly aggregated immediately upon Dox loading, resulting in large particle size, pdi and low drug loading efficiency.
Fig. 2

Short chain sphingolipid molecular structures. C8-GluCer and C8-GalCer differ in the position of one hydroxyl group in the sugar moiety position, which is equatorial and axial, respectively. C8-LacCer presents a more complex sugar moiety. Chem Draw v8.0 was used to draw molecular structures.

Upon long-term storage at 4°C, C8-GalCer-DoxNL had similar stability profile as C8-GluCer-DoxNL and DoxNL, showing a minor release of ≤ 10% Dox over 6 months (Table II) and no changes in particle size or pdi. (Supplementary Material Table SII). Under culture conditions (37°C with 10%FCS) C8-GalCer–DoxNL efficiently retained more than 90% of their contents (Table II and Supplementary Material Fig. S1), comparable to both C8-GluCer-DoxNL and DoxNL.

Optimal C8-GalCer Enriched DoxNL

C8-GalCer-DoxNL was prepared at different SCS densities and at a Dox:PL initial mass ratio of 0.1:1. Densities of C8-GalCer up to 15 mol% did not negatively affect size, pdi or drug loading of the DoxNL (Table III).
Table III

Optimal Density of C8-GalCer-enriched Pegylated DoxNL

Drug:PL 0.1:1 (w/w)i

Size (nm) ± SEM

pdi ± SEM

% Load ± SEM

0% C8-GalCer-DoxNL

89 ± 2.0

0.07 ± 0.01

96 ± 5.7

5% C8-GalCer-DoxNL

84 ± 1.2

0.06 ± 0.01

97 ± 4.3

10% C8-GalCer-DoxNL

92 ± 4.0

0.09 ± 0.02

96 ± 2.8

15% C8-GalCer-DoxNL

83 ± 3.2

0.06 ± 0.01

103 ± 1.1

In vitro efficacy studies on MCF7 breast, human BLM melanoma and ASPC1 pancreatic carcinoma cells (Fig. 3 and Supplementary Material Table SIII) demonstrated that a density of 10 mol% resulted in the strongest anti-tumor activity in all cell lines. A higher density of 15 mol% did not further increase efficacy relative to 10 mol%.
Fig. 3

In vitro efficacy study in MCF7, breast carcinoma cell line incubated with different density of C8-GalCer in DoxNL, 0 mol% (●), 5 mol% (■), 10 mol% (▲) and 15 mol%(▼) in a Dox:PL mass ratio of 0.1:1. 15 mol % content of C8-GalCer did not give additional cytotoxicity compared to 10 mol% content. Values represent the mean ± SEM (n ≥ 3).

Transmission Electron Microscopy of SCS-DoxNL

Non-enriched DoxNL and DoxNL enriched with C8-GluCer or C8-GalCer were analyzed by cryogenic Transmission Electron Microscopy (cryo-TEM) (Fig. 4). DoxNL were round shaped, uniform in size and most particles were characterized by an intraliposomal gel like precipitate of Dox. C8-GluCer-DoxNL and C8-GalCer-DoxNL, which had a normal round shape before Dox-loading (data not shown), were observed as round and rod-shaped vesicles after Dox-loading. Round vesicles were observed with and without Dox precipitate, rod-shaped particles all contained the typical Dox precipitate and had a width of approximately 40 nm and varied in length between 100 and 200 nm. In the C8-GalCer-DoxNL formulation occasionally more elongated rod-like particles with a length of up to 500 nm were observed.

SCS-enriched-DoxNL exert selective cytotoxicity towards tumor cells
Fig. 4

Cryo-TEM images of HSPC/Chol/PEG liposomes with or without SCS, loaded with doxorubicin in a Dox:PL ratio of 0.1:1. A clear drug precipitate is visible inside liposomes. SCS-enriched DoxNL have more elongated rod-like structures upon loading with Dox. The bar in the micrograph represents 200 nm. A 12500× magnification was used.

In vitro drug efficacy of the optimized SCS enriched DoxNL was tested in a panel of human tumor cell lines, including BLM and Mel57 melanoma, MCF7 and SKBR3, breast and Panc1 and ASPC1 pancreatic carcinoma, as well as in normal cells: endothelial cells (HUVEC) and 3T3 fibroblasts. In all tumor cell lines, C8-GluCer-DoxNL and C8-GalCer-DoxNL exerted increased cytotoxicity, compared to DoxNL during 24 h incubation. IC50 values of tumor cell treatments with both SCS-enriched formulations dropped significantly by 5 to 50-fold in comparison with DoxNL and commercially available Doxil®, P < 0.05 (Table IV and Supplementary Material Fig. S2a). Despite the fact that SCS-enrichment of DoxNL increased toxicity towards Mel57 and Panc1 cell lines, the differences were somewhat less pronounced as compared to the other tumor cell lines. No major differences were observed between C8-GluCer-DoxNL and C8-GalCer-DoxNL efficacy towards the various tumor cell lines. Strikingly, cytotoxicity of SCS-DoxNL towards normal endothelial cells and fibroblasts was much less pronounced, with IC50 values that are either indifferent between SCS-DoxNL or DoxNL or show a slight drop of less than 2-fold. In HUVEC, C8-GluCer-DoxNL showed a toxicity profile comparable to DoxNL and Doxil®. On the other hand, C8-GalCer-DoxNL had a somewhat more pronounced toxicity towards HUVEC than C8-GluCer-DoxNL at high drug concentrations. In turn, C8-GluCer-DoxNL showed to be of equally low toxicity to fibroblasts as C8-GalCer-DoxNL with comparable IC50 values. Empty SCS-NL were tested on MCF7 breast carcinoma cell line and no toxicity was caused by the presence of C8-GluCer or C8-GalCer in the liposomal bilayer (data not shown) at equimolar lipid concentrations to SCS-DoxNL.
Table IV

In Vitro Cytotoxicity (IC50, μM) of SCS-enriched-DoxNL and Standard Pegylated Liposomal Doxorubicin in Tumor and Non-tumor Cell Lines. At least three independent experiments were performed and values represent the mean ± SEM. * P < 0.05 versus DoxNL


IC50 (μM)



10% C8-GluCer DoxNL

10% C8-GalCer DoxNL

Tumor cells




10.0 ± 2.2*

9.1 ± 0.4*




31.0 ± 7.2*

34.1 ± 18.2*




10.8 ± 2.3*

6.9 ± 2.0*




3.1 ± 0.9*

5.3 ± 4.5*




29.2 ± 7.1*

25.4 ± 3.3*




10.1 ± 1.6*

23.1 ± 17.3*

Non tumor cells





87.7 ± 7.1




101.7 ± 17.4

103.3 ± 8.6
Fig. 5

Intracellular Dox uptake after treatment with DoxNL, 10 μM was quantified by flow cytometry in BLM melanoma, MCF7 breast carcinoma and ASPC1 pancreatic carcinoma (a). Doxorubicin was formulated in non enriched-DoxNL (open), 10% C8-GluCer-DoxNL (dark grey) and 10% C8-GalCer-DoxNL (black). Fluorescence of non treated cells was measured as a control (bright grey). At least three independent experiments were performed and values represent the mean ± SEM; *, P < 0.001; **, P < 0.01 and #, P < 0.05 versus non enriched-DoxNL at the same time point. (b) Overview of Dox uptake by 3 different tumor cell lines incubated with 10 μM C8-GluCer or C8-GalCer-DoxNL after 4 h of incubation. C8-GalCer-DoxNL gave similar results in all three cell lines, C8-GluCer seemed to increase Dox uptake in BLM (open) cells when compared to MCF7 (grey) and ASPC1 (black) cells.

In vitro drug efficacy was qualitatively confirmed by evaluating cellular morphology after treatment with free Dox, Doxil, non-enriched and SCS-enriched-DoxNL (Supplementary Material Fig. S2b). All other tested tumor cell lines showed similar results as the BLM melanoma cell line (data not shown).

Quantification of SCS-enriched-DoxNL Potential as Drug Uptake Enhancers

The effect of SCS enrichment of DoxNL on intracellular Dox uptake was quantified by flow cytometry. BLM melanoma, MCF7 breast carcinoma and ASPC1 pancreatic carcinoma cell lines were treated with DoxNL, C8-GluCer-DoxNL or C8-GalCer-DoxNL at drug concentrations of 10 or 1 μM for 1, 4 or 24 h (Fig. 5a and Supplementary Material Fig. S4a, respectively). Intracellular drug uptake increased in time for both SCS-DoxNL formulations in all cell lines, whereas only minor drug uptake was observed in cells incubated with DoxNL for 24 h. At all drug concentrations and incubation times tested, SCS-DoxNL induced higher intracellular drug levels than DoxNL. Upon incubation with 10 μM SCS-DoxNL a high intracellular drug uptake was achieved after 4 h, with little to no increase in uptake upon prolonging the incubation to 24 h. This is caused by cytotoxicity effects affecting drug uptake measurements at 24 h. At a 10-fold lower concentration, which causes much less cytotoxicity during 24 h, drug levels continued to increase up to 24 h (Supplementary Material Fig. S4). Whereas C8-GalCer after 4 h of incubation at 10 μM gives similar results in all three cell lines, C8-GluCer seemed to increase Dox uptake in BLM cells (45.2 ± 1.2) to a higher extent than MCF7 (22.3 ± 5.9) and ASPC1 cells (24.4 ± 3.9) (Fig. 5b). A similar trend can be seen at 1 μM, where C8-GluCer improves Dox delivery to BLM more strongly than C8-GalCer (Supplementary Material Fig. S4b).

Fluorescence Microscopic Imaging of Intracellular Drug Delivery

Evaluation of intracellular drug uptake by life cell fluorescence microscopy (Fig. 6) supported the outcome of flow cytometry measurements. After 4 h of incubation, non-enriched DoxNL still did not show considerable intracellular drug delivery, whereas both SCS-enriched DoxNL established high intracellular Dox levels. Similar to the observations obtained by flow cytometry, BLM cells showed increased Dox uptake from C8-GluCer-DoxNL compared to C8-GalCer-DoxNL. Dox from SCS enriched-DoxNL showed faint cytoplasmic staining, and mainly accumulated in the nucleus as confirmed by a Hoechst co-staining.
Fig. 6

Fluorescence microscopy images of BLM, melanoma cell line (a), MCF7 breast carcinoma cell line (b) and ASPC1, pancreatic carcinoma cell line (c), after 4 h of treatment with 10 μM of doxorubicin formulated in standard non-enriched, C8-GluCer or C8-GalCer enriched nanoliposomes. Nucleus was stained with Hoechst (blue). Doxorubicin is fluorescent by itself (red). At least three independent experiments were performed.

Cellular Fate of SCS-NL and Doxorubicin

Confocal microscopy experiments were performed to further confirm the intracellular localization of Dox and nanoparticles. After 4 h incubation, both enriched DoxNL achieved high intracellular drug levels. Dox accumulated in the nucleus to a high extent, whereas also cytoplasmic Dox was observed. Liposome internalization was not observed by MCF7, breast carcinoma cells (Fig. 7), BLM melanoma or ASPC1 pancreatic carcinoma cell lines (Supplementary Material Fig. S3). The absence of tumor cell DoxNL internalization was evidenced by the lack of cell associated liposomal bilayer marker NBD-PE, which was found extracellularly surrounding the cells whereas at the same time Dox showed some cytoplasmic staining and high levels in the nucleus.
Fig. 7

Intracellular-localization of NBD-PE-labeled liposomes (green) and Dox (red) in MCF7 breast carcinoma cell line by confocal microscopy. Treatment was added (10 μM Dox) in the form of C8-GluCer-DoxNL (a) or C8-GalCer-DoxNL (b) and after 4 h of incubation at 37°C, nuclear (blue) and cytoplasmatic drug uptake were analyzed.

Intracellular Localization of SCS and Doxorubicin

Time-lapse confocal microscopy experiments were performed directly after starting the incubation of cells with SCS-enriched liposomes to further study the intracellular localization of Dox and SCS. Three-dimensional reconstructions of cellular stacks in time were used to create a movie representing cellular uptake of the SCS, Dox and their co-localization in time (Supplementary Material Movie 1).

In time, fluorescently labeled SCS were observed to transfer from the DoxNL bilayer to the cell membrane (Fig. 8 and Supplementary Material Movie 1). This transfer was observed already at 5 min after start of the incubation and preceded cellular Dox uptake, which was observed only at 10 min after adding liposomes to the cells. Doxorubicin entered the cell and prominently accumulated in the nucleus within the subsequent 30 min. In addition, appreciable Dox-levels were detected in the cytoplasm at 25–40 min (Fig. 8).
Fig. 8

Localization of SCS and Dox in the plasma membrane by confocal microscopy. Directly after starting BLM melanoma cell treatment, with 5 μM NBD labeled-SCS liposomal Dox, the green fluorescent labelled-C8-GalCer and fluorescent Dox were followed in time. The SCS accumulates in the plasma membrane after which Dox accumulation in the nucleus takes place rapidly.

SCS transferred to the cell membrane were later also found intracellular in the cytoplasm, but were not associated with the nucleus or with its membrane.


The use of SCS-enriched nanoliposomes to improve drug-cell membrane traversal is a novel drug delivery concept and represents an advanced and versatile technology to enhance Dox delivery into tumor cells and thereby its efficacy as a chemotherapeutic drug. Previously, we were able to demonstrate that DoxNL enriched with C8-GluCer strongly enhanced Dox delivery and subsequent anti-tumor efficacy both in vitro and in vivo (7,8). In the present study a further optimization and characterization of C8-GluCer-DoxNL was performed, in terms of drug loading, liposome size, pdi, drug entrapment efficiency and morphology, all representing important criteria for translation of this formulation towards future clinical application. In addition, DoxNL with other synthetic short-chain glycosphingolipids were studied for drug uptake enhancement, of which C8-GalCer, but not C8-LacCer appeared promising and displayed similar drug uptake enhancing properties as C8-GluCer. Drug uptake enhancement by SCS-DoxNL significantly increased in vitro antitumor activity towards various human tumor cell lines. This effect was observed to a much lesser extent in normal endothelial cells and fibroblasts, demonstrating an important preference of this drug delivery process for tumor cells. Finally, we were able to demonstrate that enhanced Dox delivery by SCS-Dox-NL was not related to uptake of the nanoliposomes as such by the tumor cells, but was preceded by a rapid and apparently critical transfer of the SCS from the nanoliposomal bilayer to the cell membrane. A similar transfer process was proposed previously by Zolnik and co-workers for the apoptosis-inducing N-hexanoyl-d-erythrosphingosine (C6-ceramide) (31). In the present study we demonstrated that the C8-GluCer or C8-GalCer upon their transfer to the tumor cell membrane do not affect cell survival in the absence of Dox, but are able to potentiate Dox cytotoxicity.

Optimal loading of Dox (> 90% encapsulation efficiency) in DoxNL enriched with 10 mol% C8-GluCer or C8-GalCer was obtained at a Dox:PL initial mass ratio of 0.1:1 resulting in homogeneous particles (pdi < 0.1) with a size between 80 and 100 nm, which is considered ideal for efficient extravasation through leaky tumor vasculature (1,32). For SCS-DoxNL higher Dox:PL ratios caused somewhat less efficient drug loading, but did not affect drug efficacy towards BLM, melanoma and MCF7 breast carcinoma cell lines (Fig. 1a). This result suggests that the quantity of SCS used at the highest Dox:PL ratio of 0.25:1 was sufficient to induce maximal enhancement of drug uptake. This is in accordance with Veldman et al. who demonstrated that uptake of Dox, added to cells in free form, reached saturation upon treating these cells with increasing concentrations of C6-sphingomyelin (6). Moreover, the higher quantity of SCS, in enriched liposomal formulations with lower Dox:PL mass ratios did not induce additional toxicity by the SCS. This correlates well with the finding that cytotoxicity is related to doxorubicin treatment and not to the SCS.

Formulation of C8-LacCer in Dox-NL was unsuccessful. Despite the fact that C8-LacCer enriched-NL could be prepared, Dox loading of these NL resulted in strong aggregation and low loading efficiency. Likely, the more complex sugar moiety of C8-LacCer in comparison to C8-GluCer or C8-GalCer (Fig. 2) interfered with the liposomal membrane preventing a stable transmembrane ammonium sulfate gradient.

Whereas density of C8-GalCer did not affect DoxNL size, pdi and Dox loading, in vitro drug efficacy was improved only for the 10 and 15 mol% enriched formulations in comparison to standard DoxNL. Since 10 mol% C8-GalCer-DoxNL were equally effective as those with 15 mol%, we concluded that the former loaded at an initial Dox:PL mass ratio of 0.1:1 represented the optimal C8-GalCer-DoxNL formulation. With a 10 mol% SCS-density and a 0.1:1 D:PL (w/w) loading ratio this optimal C8-GalCer-DoxNL has similar characteristics as the optimal C8-GluCer-DoxNL (8).

Stability studies under storage or cell culture conditions demonstrated both SCS-DoxNL formulations provided high stability and did not reveal significant differences between C8-GluCer or C8-GalCer enriched and standard DoxNL. Cryo-TEM elucidated a typical Dox precipitate inside all DoxNL (28) with a more prominent rod-shaped morphology for SCS-Dox-NL than DoxNL. Apparently the presence of SCS in the bilayer of DoxNL is responsible for the change in shape of the nanoparticles during Dox-loading, as SCS-liposomes before loading were round-shaped and after loading all rod-shaped particles did contain a clear Dox precipitate. Liposomal bilayers, when enriched with SCS, seemingly are more flexible or deformable. It remains speculative whether this enhanced flexibility of enriched liposomal bilayers contributes to the improved intracellular Dox delivery.

Both C8-GluCer-DoxNL and C8-GalCer-DoxNL significantly increased cytotoxicity in comparison to standard DoxNL and clinically applied DoxNL (Doxil/Caelyx) in all tumor cell lines tested. Stable Dox entrapment in Doxil is known to reduce toxic side effects related to free Dox. However, this high stability in combination with the low tumor cell membrane permeability for DoxNL strongly limits drug delivery and efficacy (22,23,33). It is at this point that drug delivery is improved when using SCS-DoxNL. Enrichment of DoxNL with C8-GluCer or C8-GalCer did not affect Dox release for up to 24 h in cell culture medium containing 10% of serum (Table II and Supplementary Material Figure S1), but was in a similar period able to improve in vitro drug efficacy, due to SCS potential as drug uptake enhancer. Remarkably, the in vitro cytotoxic effect of SCS-DoxNL was much less pronounced on endothelial cells and fibroblasts demonstrating a cell-type dependency and specificity of the SCS-drug delivery process in favor of tumor cells. In addition, this finding holds promise that SCS-DoxNL during circulation will not readily affect the normal endothelial lining of healthy blood vessels. Upon accumulation at the tumor site, SCS can act as a tumor cell membrane selective drug permeabilizer. In this context, similar results using SCS and Dox administered in free form, were reported by Veldman et al. with cultured rat cardiac myoblasts (6). For yet unknown reasons, non-tumor cells are not highly susceptible towards the sphingolipid analogue mediated drug uptake enhancement, when compared to tumor cells.

The increased in vitro efficacy of SCS-DoxNL coincided with an increased and rapid intracellular drug delivery in different tumor cell lines as demonstrated by fluorescence microscopy and flow cytometry. Already within 1 h notable differences in cellular Dox fluorescence were measured between SCS-enriched and non-enriched DoxNL, in favor of the former. These differences appeared maximal after 4 h of incubation reaching up to 5-8-fold more Dox delivered using SCS-DoxNL and remained high until 24 h. In these drug uptake studies some variation between the sensitivity of various tumor cell lines towards SCS- DoxNL was observed. Dox uptake after incubation for 4 h with C8-GluCer-DoxNL was higher in BLM melanoma than after C8-GalCer-DoxNL. This difference was not observed with MCF7 breast and ASPC1 pancreatic carcinoma. These observations were confirmed by living cell fluorescence microscopy and suggest a difference in the potential of C8-GluCer versus C8-GalCer as drug uptake enhancers and that even among tumor cells certain cell specificity for each distinct SCS exists. Future studies will investigate whether differences in cell membrane lipid composition can explain the differences in SCS-mediated drug uptake in normal and tumor cells. An important question is how Dox is transported from the liposomes into the cell. Seynhaeve et al. concluded that, in addition to passive release, DoxNL may be taken up completely by tumor cells upon long-term incubations, followed by intracellular degradation, after which released Dox enters the nucleus (22). Others suggested that liposomes can be degraded in the interstitial space followed by uptake of the released drug by tumor cells (4,5,3436). Our living cell confocal microscopy studies using SCS-DoxNL with a fluorescent bilayer marker, showed virtually no internalization of the nanoliposomal carrier by MCF7, ASPC1 or BLM cell lines during 4 h of incubation, whereas at the same time high levels of Dox where detected intracellularly in the nucleus and to a lesser extent in the cytoplasm. Therefore intracellular Dox delivery using SCS-NL is not dependent on internalization of the nanocarrier as such. By contrast, transfer of the SCS from the nanoliposomal bilayer to the tumor cell membrane was demonstrated to occur rapidly and to precede Dox influx. Together these data demonstrate that the selective transfer of the SCS to the tumor cell membrane promotes Dox internalization. Transfer of liposomal bilayer constituents to tumor cell membranes has been described previously for a lipophilic prodrug of 5-fluorodeoxyuridine and the anti-angiogenic drug fumagillin (3739).

It remains to be seen whether the SCS transfer also enhances drug release from the liposomes. The transfer of SCS from the rod-shaped liposomal bilayer to the cell membrane when in close proximity may well be responsible for a locally enhanced release of Dox from the particles in which the SCS stabilized rod-like shape is deformed causing local content release at the sites of SCS transfer. The released content is then taken up more rapidly due to SCS induced permeability of the cell membrane. Pinnaduwage and Huang referred to a similar delivery process by liposomes that upon direct contact with a membrane-bound antigen, rapidly and spontaneously destabilize, releasing entrapped contents (40).


The findings presented here identified, C8-GluCer and C8-GalCer, as specific SCS to improve the therapeutic potential of nanoliposomal Dox by selectively enhancing cellular membrane permeability of tumor cells to Dox upon their transfer to the cell membrane. Optimized formulations of SCS-DoxNL were developed, which will now undergo in vivo evaluation to investigate if they are able to overcome the current issue of limited Dox bioavailability related to DoxNL.

Acknowledgments and Disclosures

This work was financed by the Dutch Cancer Society.

Supplementary material

11095_2013_1031_MOESM1_ESM.avi (44.1 mb)
Movie 1Localization of NBD-PE labeled C8-GalCer-DoxNL (green) and Dox (red) by confocal microscopy. BLM melanoma cell line was treated with 5 μM Dox formulated in NBD labelled- C8-GalCer-DoxNL and fluorescence of SCS and Dox was followed in time for a total of 40 min (AVI 45116 kb)
11095_2013_1031_Fig9_ESM.jpg (16 kb)
Supplemental Figure 1a

Stability at 37°C in the presence of 10%FCS of enriched C8-GluCer-DoxNL (■) was measured for 24 h (A) and enriched C8-GalCer-DoxNL (▲) for 15 h (B). Measurements were based on doxorubicin fluorescence following time and were performed continuously. Both enriched-DoxNL presented a better stability profile than non enriched-DoxNL (●) (JPEG 16 kb)

11095_2013_1031_Fig10_ESM.jpg (17 kb)
Supplemental Figure 1a

Stability at 37°C in the presence of 10%FCS of enriched C8-GluCer-DoxNL (■) was measured for 24 h (A) and enriched C8-GalCer-DoxNL (▲) for 15 h (B). Measurements were based on doxorubicin fluorescence following time and were performed continuously. Both enriched-DoxNL presented a better stability profile than non enriched-DoxNL (●) (JPEG 16 kb)

11095_2013_1031_MOESM2_ESM.tif (2.2 mb)
High Resolution Image(TIFF 2211 kb)
11095_2013_1031_MOESM3_ESM.tif (2.2 mb)
High Resolution Image(TIFF 2211 kb)
11095_2013_1031_Fig11_ESM.jpg (133 kb)
Supplemental Figure 1a

Stability at 37°C in the presence of 10%FCS of enriched C8-GluCer-DoxNL (■) was measured for 24 h (A) and enriched C8-GalCer-DoxNL (▲) for 15 h (B). Measurements were based on doxorubicin fluorescence following time and were performed continuously. Both enriched-DoxNL presented a better stability profile than non enriched-DoxNL (●) (JPEG 16 kb)

11095_2013_1031_MOESM4_ESM.tif (138 kb)
High Resolution Image(TIFF 138 kb)
11095_2013_1031_MOESM5_ESM.doc (273 kb)
Supplemental Figure 2b(DOC 273 kb)
11095_2013_1031_MOESM6_ESM.ppt (1.3 mb)
Supplemental Figure 3Co-localization of NBD-PE labeled liposomes (green) and Dox (red) by Confocal Microscopy. After nucleus staining with Hoechst (blue), BLM melanoma cell line was treated with 10 μM Dox in form of C8-GluCer-DoxNL (A) and C8-GalCer-DoxNL (B). ASPC1 pancreatic carcinoma cells were equally treated with 10 μM Dox in form of C8-GluCer-DoxNL (C) and C8-GalCer-DoxNL (D). After treatment cells were incubated for 4 h at 37°C. Nuclear and cytoplasmatic drug uptake was analysed (PPT 1299 kb)
11095_2013_1031_Fig12_ESM.jpg (53 kb)
Supplemental Figure 1a

Stability at 37°C in the presence of 10%FCS of enriched C8-GluCer-DoxNL (■) was measured for 24 h (A) and enriched C8-GalCer-DoxNL (▲) for 15 h (B). Measurements were based on doxorubicin fluorescence following time and were performed continuously. Both enriched-DoxNL presented a better stability profile than non enriched-DoxNL (●) (JPEG 16 kb)

11095_2013_1031_MOESM7_ESM.tif (12.6 mb)
High Resolution Image(TIFF 12946 kb)
11095_2013_1031_MOESM8_ESM.doc (30 kb)
Supplemental Table 1In vitro IC50 values toward tumor cells BLM, melanoma and SKBR3, breast carcinoma, after treatment with C8-GluCer-DoxNL following different drug to phospholipid ratios: 0.25:1; 0.2:1; 0.15:1; 0.1:1 (w/w) (DOC 30.5 kb)
11095_2013_1031_MOESM9_ESM.doc (30 kb)
Supplemental Table 2Stability at 4°C after 6 months of different glycoceramide-enriched liposomes in a drug to phospholipid ratios of 0.1:1 (w/w). Size and pdi were analyzed. (DOC 30 kb)
11095_2013_1031_MOESM10_ESM.doc (34 kb)
Supplemental Table 3In vitro IC50 values of doxorubicin in different tumor cell lines, BLM, melanoma, MCF7 breast carcinoma and ASPC, pancreatic carcinoma, following different densities for C8-GalCer-DoxNL (DOC 34 kb)

Copyright information

© Springer Science+Business Media New York 2013