Barbara
Pili
a,
L. Harivardhan
Reddy†
a,
Claudie
Bourgaux
a,
Sinda
Lepêtre-Mouelhi
ab,
Didier
Desmaële
b and
Patrick
Couvreur
*a
aUniversité Paris-Sud XI, UMR CNRS 8612, 5 rue J.B. Clément, 92290, Châtenay-Malabry, France. E-mail: patrick.couvreur@u-psud.fr
bUniversité Paris-Sud XI, UMR CNRS 8076 Biocis, 5 rue J.B. Clément, 92290, Châtenay-Malabry, France
First published on 5th July 2010
A new prodrug of gemcitabine, based on the covalent coupling of squalene to gemcitabine (GemSQ), has been designed to enhance the anticancer activity of gemcitabine, a nucleoside analogue active against a wide variety of tumors. In the present study, the feasibility of encapsulating GemSQ into liposomes either PEGylated or non-PEGylated has been investigated. The in vivo anticancer activity of these formulations has been tested on subcutaneous grafted L1210wt leukemia model and compared to that of free gemcitabine. The liposomal GemSQ appears to be a potential delivery system for the effective treatment of tumors.
Liposomes have been employed for the encapsulation of gemcitabine with the aim to modify the drug pharmacokinetics and biodistribution and to deliver gemcitabine more efficiently to tumors. It has been found that liposomal gemcitabine exhibited greater anticancer activity than did free gemcitabine.6–9 Composition used for liposomal formulation and the preparation techniques were, however, observed to greatly influence the loading capacity of gemcitabine into liposomes and the cell penetration properties of the entrapped drug.10 But more importantly, gemcitabine which is a low-molecular weight hydrophilic molecule as well as Ara-C or 5-fluorouridine, the drugs closely related to gemcitabine, may rapidly leak out from liposomes. In order to limit this problem, various alkyl derivatives of Ara-C have been synthesized. These lipophilic pro-drugs have been encapsulated into liposomes with high efficiency and these liposomal formulations displayed in vivo antitumor activity superior to that of the pure drug.11 Similarly, a series of acyl derivatives of gemcitabine have been synthesized and encapsulated into liposomal formulations leading to modified pharmacokinetics.12 For instance, liposomes loaded with a 4-(N)-stearoyl derivative of gemcitabine showed an improved in vivo anticancer activity against HT-29 colon adenocarcinoma and KB396p nasopharyngeal carcinoma.13 Thus, liposomal prodrugs may be beneficial in terms of drug incorporation efficiency, stability, and biopharmaceutical properties.
In this context, we have synthesized a new derivative of gemcitabine by coupling the acyclic isoprenoid chain of squalene on the 4-amino function of gemcitabine. The resulting bioconjugate, i.e. 4-(N)-1,1′,2-trisnor-squalenoylgemcitabine (GemSQ), displayed more potent cytotoxicity in vitro, and exhibited considerably higher anticancer activity in vivo compared to gemcitabine against P388 and L1210wt leukemia following intravenous dosing.14–17 However, nanoassemblies of this prodrug showed considerable accumulation in organs of the reticuloendothelial system such as the spleen and liver after intravenous administration, probably due to important capture by macrophages.18 This is the reason why we performed the PEGylation of GemSQ nanoassemblies by the addition of polyethyleneglycol coupled to squalene (PEGSQ). This nanoconstruction was, however, found to be very unstable in biological media, due to the formation of unstable micelles, resulting from the solubilization of GemSQ nanoassemblies by PEGSQ in the presence of proteins.
Thus, in the present study, we have investigated the feasibility of encapsulating GemSQ into PEGylated liposomes to confer long circulating properties to this anticancer prodrug. The in vivo anticancer activity of this formulation has been tested on a subcutaneous grafted L1210wt leukemia model and compared with the activity of GemSQ encapsulated in non-PEGylated liposomes of the same composition or with free gemcitabine.
Gemcitabine hydrochloride (2′-deoxy-2′,2′-difluorocytidine monohydrochloride; C9H11F2N3O4 HCl, β-isomer) was purchased from Sequoia Research Products (Pangbourne, UK). Squalene was purchased from Sigma-Aldrich (Saint-Quentin, Fallavier, France).
Fig. 1 Chemical structure of gemcitabine squalene (GemSQ). |
The hydrodynamic diameters of the resulting liposome suspensions were determined at 25 °C by quasi-elastic light scattering (QELS). The selected scattering angle was 90° (Zetasizer Nano ZS, Malvern Instruments Ltd, UK). The measurements were done 15 min after preparation of the liposomal formulations, in order to have the initial size. The size stability of the suspensions was then investigated over two weeks. The zeta potential of these liposomal formulations was also determined. Each sample was diluted (1/10) in 1 mM NaCl. Analysis of the samples was performed at 25 °C in triplicate.
The final GemSQ concentration in both formulations, GemSQ/DPPC and GemSQ/DPPC/DSPE-PEG, was 2 mg ml−1 for all samples.
Data analysis has been performed using TA Universal Analysis program (New Castle, Delaware, USA). The transition temperatures were taken at the onset of the transitions (Tonset) i.e., the intersection of the tangent to the left side of the endothermic peak with the baseline.
The L1210wt leukemia subcutaneous tumor model was developed by injecting the exponentially growing L1210wt leukemia cells in suspension, containing 30% growth factor reduced Matrigel, subcutaneously (1 × 106) into the hind flank region of mice. A palpable tumor was allowed to grow at the injection site. The mice were randomly divided into four groups of eight mice each: untreated, treated with 100 mg kg−1 gemcitabine (MTD), treated with GemSQ/DPPC formulation at 20 mg kg−1 (equivalent in gemcitabine) and treated with GemSQ/DPPC/DSPE-PEG liposomal formulation at 20 mg kg−1 (equivalent in gemcitabine). Six days after tumor implantation, when the mice developed palpable tumors, all groups of mice received the treatment by retro-orbital sinus injection on days 0, 4, 8 and 13, with the exception of the untreated group. The mice were monitored regularly for tumor volume, and survival to assess the anticancer efficacy. All groups were considered statistically valid up to (n − 3) surviving animals. Tumor size was measured across its two perpendicular diameters, and its volume was calculated using the following formula:
V = 0.5 × (W2 × L) |
Fig. 2 DSC scans of MLV mixtures. Heating cycles were performed at 5 °C min−1. |
It is noteworthy that the extrusion of these two highly diluted GemSQ formulations led to the formation of liposomes stable in the 20–50 °C temperature range. The inverse bicontinuous cubic structure of the GemSQ/DPPC mixture was no longer maintained in the nanoassemblies (see below).
Both formulations of liposomes (PEGylated or not PEGylated) obtained by extrusion displayed monodisperse population with diameters close to the average pore size of the last used filter (0.1 μm). Nevertheless, non-PEGylated liposomal formulations exhibited a mean hydrodynamic diameter (133 ± 25 nm, polydispersity index 0.058) tending to be slightly larger than PEGylated liposomes (113 ± 24 nm, polydispersity index 0.035) due to a possible difference in the bilayer curvature.
Both GemSQ/DPPC and GemSQ/DPPC/DSPE-PEG liposomes did not show any visual sedimentation over time. In order to check if aggregation or fusion/coalescence occurred, we have monitored the particle size of the samples during a 15 day period (Table 2). This stability study indicated that the size of the liposomal formulations remained unchanged during this period of evaluation.
The size evolution of GemSQ/DPPC liposomes has also been monitored as a function of temperature. The size of liposomes increased in a reversible manner up to about 145 nm at 50 °C. Interestingly, the ratio between the areas of liposomes at 50 °C and 25 °C (respectively ∼66020 nm2 and ∼53066 nm2) was close to the ratio of the areas per DPPC molecule at these temperatures (respectively 64 Å2 and 47 Å2).22,23 This confirmed that the liposomes formed by extrusion at 50 °C were preserved upon cooling to room temperature.
The zeta potential of the two liposomal formulations has been measured in the presence of 10−3 M NaCl. As expected, GemSQ/DPPC liposomes displayed zeta potential values close to zero. Indeed, DPPC is a zwitterionic phospholipid, globally neutral, and the head-group of GemSQ does not carry charges at neutral pH. On the contrary, the zeta potential of the PEGylated liposomes has been found to be negative (i.e., −35 mV) which may be explained by the net negative charge of the PEG-lipid head-group. This result confirmed that DSPE-PEG was inserted into the bilayer.
The vesicular morphology of the two extruded formulations has been demonstrated by small-angle X-ray scattering (SAXS) measurements (Fig. 3). Regarding the GemSQ/DPPC mixture, the SAXS curve did not show the Bragg reflections indicative of the inverse bicontinuous cubic phase previously evidenced in not extruded GemSQ/DPPC samples at room temperature. A broad bump followed by periodic oscillations was seen in the q-range 0.05–0.5 Å−1. This pattern was characteristic of the bilayer form factor of unilamellar vesicles. From the positions of the curve minima, the thickness of the bilayer could be estimated to be ∼50 Å, consistent with already reported data for pure DPPC. It is noteworthy that the pattern obtained for the PEGylated formulation was the same. The bilayer form factor did not appear to be affected by the hydrophilic PEG layer at the surface of the membrane. This could be explained by the low excess scattering induced by the presence of a small molar fraction of PEG in water.24 Only the electron density profile of the bare bilayer was seen with X-rays. It should be emphasized that the complementary shapes of “inverted wedge-shaped” GemSQ and “wedge-shaped” DSPE-PEG molecules helped to stabilize the vesicles in the PEGylated formulation.
Fig. 3 SAXS patterns of GemSQ/DPPC and GemSQ/DPPC/DSPE-PEG extruded liposomes at room temperature (log. scale). |
Fig. 4 Tumor volume analysis of L1210wt subcutaneous tumor in mice after i.v. injection of gemcitabine (GEM) 100 mg kg−1 and liposome formulations of gemcitabine squalene (GemSQ) 20 mg kg−1 eq. of gemcitabine. A palpable tumor was developed on the hind flank region of mice by subcutaneously injecting 1 × 106 L1210wt cells in suspension containing 30% growth factor reduced Matrigel. The above formulations were injected intravenously into the tumor-bearing mice on days 0, 4, 8, 13, indicated by ↑ in the figure. All groups were considered statistically valid up to (n − 3) surviving animals. The results are expressed as mean ± S.D. Statistical analysis was performed using Student's t-test considering 95% confidence interval at significance level p < 0.05 (* indicates p < 0.05). |
In general, the lipophilic prodrug approach facilitates the incorporation of both hydrophilic and hydrophobic small therapeutic molecules into liposomal vesicles.28 For this reason, the design of an altered lipophilic prodrug of gemcitabine that would be more efficiently retained in liposomes is an interesting alternative approach. Thus, we have employed this approach in the present study by coupling gemcitabine with squalene, a natural lipid and a precursor in the biosynthesis of cholesterol. It was expected that the stable anchoring of this lipophilic bioconjugate in the liposomal bilayer could lead to significant anticancer activity in experimental tumors.
The incorporation of GemSQ into DPPC and DPPC/DSPE-PEG bilayers has been evidenced by DSC analysis, whereas the vesicular morphology of both formulations has been demonstrated by SAXS measurements. The prepared GemSQ/DPPC and GemSQ/DPPC/DSPE-PEG formulations were found to be stable for at least 2 weeks allowing us to perform in vivo experiments in good conditions; the size of the liposomes (i.e. 113 nm and 130 nm) were compatible with intravenous injection.
We show in this paper that the liposomal formulations of GemSQ displayed at doses as low as 20 mg kg−1 (equiv. of gemcitabine) (the maximum tolerable dose of GemSQ) a similar anticancer activity to free gemcitabine injected at a dose of 100 mg kg−1. However, contrary to what was expected, the PEG-coated liposomes did not improve considerably the antitumoral effect as compared with plain, non-PEGylated liposomes, although the antitumor activity difference of the PEGylated liposomal formulation over the non-PEGylated liposomal formulation was statistically significant at a single time point on day 14. Vascularization and microvessels density have, however, been previously evidenced by immunohistochemistry in the leukemia solid tumor model chosen for this study. It is noteworthy that the subcutaneous implanted version of the L1210wt leukemia has already been employed to study the tumor penetration of fluorescent latex particles mimicking liposomes29 and to determine the antitumor activity of antisense oligonucleotide loaded into lipid nanoparticles against the anti-apoptotic Bcl2.30 In our case, it remains, however, possible that the subcutaneously growing L1210wt leukemia did not display the increased permeability of tumor vessels, needed for the “enhanced permeability and retention” (EPR) effect31 to occur, possibly attributed to the smaller initial tumor size as compared to that reported in Pan et al.29 This may perhaps be one of the reasons why the PEGylated GemSQ liposomes did not display superior anticancer activity compared to their non-PEGylated counterpart. It might also be speculated that a fraction of GemSQ was transferred from liposomes to lipoproteins in the blood stream. Indeed, this effect has already been reported for annamycin, a lipophilic derivative of doxorubicin.32 When liposomes containing annamycin were incubated in plasma, the major part of annamycin was recovered in the HDL fraction of plasma lipoproteins. “Stealth” liposomes did not display any advantage concerning the tumor targeting compared to bare liposomes. It is likely that they did not significantly prevent the transfer of annamycin to blood components.33 The same effect was observed for a lipophilic derivative of Ara-C, NHAC. The drug was rapidly transferred from liposomes to plasma proteins and erythrocytes, regardless of the liposome’s surface coating with PEG.11 Dosages of GemSQ in LDL, HDL and/or VLDL should allow one to clarify if an exchange of the prodrug may exist between liposomes and lipoproteins. If so, a modification of the lipid composition of the liposomes should be considered in order to reduce GemSQ exchange with lipoproteins. The body weight differences in mice following the treatments were assessed to understand if the PEGylated liposomal formulation could make a difference over non-PEGylated formulation in terms of toxicity; however no meaningful conclusion could be made from these observations. This is because the L1210wt subcutaneous tumor chosen in this study is a rapidly growing tumor and hence the body weights of the tumor-bearing mice increased faster than the decrease in body weights, if any, caused by the treatment.
Thus, in the tumor model tested in our study, it may be hypothesized that the anticancer efficacy of the liposomal formulations at considerably low doses compared to the free drug was principally due to the protection of gemcitabine from deamination in the blood stream since the PEGylation did not lead to a considerable advantage compared to non-PEGylated liposomes.
Footnote |
† Current address: Sanofi-aventis, 13 Quai Jules-Guesdes, 94403, Vitry-sur-Seine, France. |
This journal is © The Royal Society of Chemistry 2010 |