Controlled Release, Intestinal Transport, and Oral Bioavailablity of Paclitaxel Can be Considerably Increased Using Suitably Tailored Pegylated Poly(Anhydride) Nanoparticles

Materials

Poly(methyl vinyl ether-co-maleic anhydride) or poly(anhydride) (PMV/MA) [Gantrez® AN 119; molecular weight 200 kDa] was purchased from ISP/Ashland Inc. (Barcelona, Spain). PTX (USP 26, grade > 99.5%) and docetaxel (DCX) (99.0%) were supplied by 21CECpharm (London, UK). Taxol® was provided by Bristol-Myers-Squibb (Princeton, NJ, USA). Verapamil, glutamine, glycine, polysorbate 80, β-cyclodextrin (CD), and 2-hydroxylpropyl-β-cyclodextrin (HPCD) were obtained from Sigma–Aldrich (St. Louis, MO, USA) and ethylenediaminetetraacetic acid disodium salt (EDTA) and PEG 2000 (PEG2000) were provided by Fluka (Buchs, Switzerland). Phosphate-buffered saline (PBS) was purchased from Medicago AB (Uppsala, Sweden). Acetone, ethanol, and acetonitrile were obtained from Merck (Darmstadt, Germany). Deionized water (18.2 MΩ resistivity) was prepared by a water purification system (Wasserlab, Pamplona, Spain). All reagents and chemicals used were of analytical grade.

Preparation of PTX–Cyclodextrin Complexes

The preparation of inclusion complexes of PTX and cyclodextrins (HPCD or CD) was performed, by an evaporation method, in the molar ratio 1:1 as described previously.23, 27 Ten milligrams PTX was dissolved in 2 mL ethanol and then added to 8 mL water containing the cyclodextrin. After magnetic agitation of the mixture for 72 h, ethanol was evaporated under reduced pressure (Büchi R-144, Postfach, Switzerland) and the resulting suspensions filtered through a 0.45 μm membrane filter. Finally, the obtained filtrate was evaporated under reduced pressure at a temperature of 50°C to obtain a solid dry residue.

Preparation of Nanoparticles

Paclitaxel–cyclodextrin inclusion complexes (PTX–HPCD or PTX–CD) were encapsulated in poly(anhydride) nanoparticles by a solvent displacement method as described previously.27 Pegylated nanoparticles with PEG2000 were prepared following the method published by Zabaleta et al.31 with minor modifications.

Characterization of the Different Nanoparticles

In Vitro Release Study

Release experiments were conducted under sink conditions at 37°C using simulated gastric fluid (SGF; pH 1.2; pepsin 0.32%, w/v) and simulated intestinal fluid (SIF; pH 6.8; pancreatin 1%, w/v) containing 1% of polysorbate 80 (Tween 80) as solubilizing agent for PTX. The studies were performed under agitation in a Vortemp 56^TM Shaking Incubator (Labnet International Inc., Edison, NJ, USA) after the dispersion of the nanoparticles in the appropriate medium.

For each time point, 18 μg of PTX formulated in nanoparticles was resuspended in 1 mL of the corresponding simulated fluid. The different formulations were kept in the SGF for 2 h and then, for 20 h in SIF. At different time points, sample tubes were collected and centrifuged at 27,000g for 20 min. The amount of PTX released from the formulations was quantified by HPLC from the supernatants (calibration curves of free PTX in supernatants obtained from SGF and SIF, r² > 0.999).

In Vitro Studies in Ussing Chambers

Studies were carried out after approval by the responsible Ethical Committee of the University of Paris-Sud (agreement number A092–019–01) in strict accordance with the European legislation on animal experiments.

Permeation Assays

Ussing Chambers (Warner Instruments, Harvard Apparatus, Hamden, CT, USA) were used to determine the permeability of fresh rat intestinal tissue to PTX formulated in nanoparticles or commercial Taxol®. The methodology and terminology used was previously described.33 Jejunum portions from fresh small intestine of sacrificed male Wistar rats (200–250 g; Charles River, Chatillon-sur-Chalaronne, France) were excised, rinsed with chilled NaCl 0.9%, and cut into segments of 2–3 cm length, discarding sections containing Peyer's patches. The selected portions were cut along the mesenteric border and mounted in Ussing chambers (apparent intestinal surface of 1 cm²) while bathed with PBS at pH 7.4 containing glutamine 0.2 M.

Each compartment (donor and receptor) was filled with 4 mL PBS, maintained at 37°C and continuously oxygenated with O₂/CO₂ (95/5%). In the receptor compartment, 0.1% (w/v) HPCD was also included as solubilizing agent for PTX. A 30-min equilibration time was allowed to achieve steady-state electrophysiological conditions. After equilibration, 2 mg PTX loaded in the different nanoparticle formulations or as Taxol® were diluted in 300 μL PBS and added to the donor compartment. Four tissue portions of four different animals were assessed for each formulation and experiments were repeated on different days.

Aliquots of 200 μL were withdrawn from the receptor chambers every 30 min up to 2 h and the volume was immediately replaced by fresh medium pre-equilibrated at the experimental temperature (37°C). In addition, at the beginning and end of the experiments, samples (200 μL) were also obtained from the donor compartments, in order to monitor any changes in this compartment during the experiment and to safeguard mass balance. All samples were immediately frozen and stored at −20°C until analysis. PTX amount was quantified by HPLC–UV as described above.

The cumulative amounts of drug permeated in the mucosal-to-serosal (M–S; absorptive transport, or apical to basolateral) and serosal-to-mucosal (S–M, excretory transport or basolateral to apical) directions were calculated. In addition, the influence of the presence of verapamil (selective Pgp inhibitor) was evaluated by adding 0.2 mM of the inhibitor to the PBS solution in the donor chamber.

For the comparison between the different formulations, the following parameters were calculated: apparent permeability coefficient (P_app) and the absorption enhancement ratio (R). The apparent permeability coefficient was calculated by the following equation:

(1)

where (dQ/dt) is the flux of PTX from the donor to receiver compartment, C₀ is the initial drug concentration in the donor chamber, and A is the surface of the tissue membrane (1 cm²).

In order to standardize calculations, the P_app values were estimated between 30 and 90 min after the addition of the formulation in study.

The absorption enhancement ratio (R) was calculated from the P_app values:

(2)

where P_app (sample) is the apparent permeability of jejunum to PTX when included in the tested nanoparticle formulation and P_app (control) is the apparent permeability to commercial Taxol®.

In Vivo Pharmacokinetic Studies in C57BL/6J Mice

Pharmacokinetic Data Analysis

The pharmacokinetic analysis of plasma concentration plotted against time data, obtained after administration of the different PTX formulations, was performed according to a non-compartmental model with the WinNonlin 5.2 software (Pharsight Corporation, Princeton, NJ, USA). The following parameters were estimated: maximal plasma concentration (C_max), time in which C_max is reached (T_max), area under the concentration–time curve from time 0 to ∞ (AUC), mean residence time (MRT), clearance (Cl), volume of distribution (V), and half-life in the terminal phase (t_1/2).

Furthermore, the relative bioavailability (Fr%) of PTX was estimated by the following equation:

(3)

where AUC_i.v. and AUC_oral are the areas under the curve for the i.v. Taxol® or oral administrations, respectively.

Statistical Analysis

Data are expressed as the mean ± SD of at least three experiments. The non-parametric Kruskall–Wallis followed by Mann–Whitney U-test was used to investigate statistical differences. In all cases, p < 0.05 was considered to be statistically significant. All data processing was performed using GraphPad Prism 4.0 statistical software (GraphPad Software, La Jolla, CA, USA).

Preparation and Characterization of Poly(Anhydride) Nanoparticles

Table 1 summarizes the main physicochemical properties of the different poly(anhydride) nanoparticles containing PTX. When nanoparticles were loaded with the PTX–cyclodextrin complexes, the mean size of the resulting carriers was significantly higher than when PTX was formulated with PEG2000 (p < 0.05). Herein, pegylation of nanoparticles significantly decreased the mean size of the resulting nanoparticles. Thus, PTX–CD–NP–PEG displayed a mean diameter about 25% lower than PTX–CD–NP (Table 1). Concerning the zeta potential of nanoparticles, all formulations were formed by nanoparticles with negative surface charge. However, again, when nanoparticles were pegylated with PEG2000, the zeta potential appeared to be slightly more negative, around −55 mV. For the formulation with just the PTX–CD complexes, the surface charge was −48 mV. On the other hand, the yield of the production of nanoparticles was calculated to be between 60% and 75% for the PEG and cyclodextrin containing formulations, respectively.

Regarding the drug loading, the amount of PTX encapsulated in the nanoparticles varied according to the excipient used. On one hand, looking at the amount of drug encapsulated, when PTX was complexed with HPCD, the drug loading (150 μg/mg nanoparticles) was approximately three times higher than when PTX complexed with CD was loaded in nanoparticles (46 μg/mg nanoparticles). On the other hand, when PEG was incorporated to the formulation, the amount of PTX encapsulated was around 110 μg/mg NP. In contrast, when PTX was complexed with CD and loaded in pegylated nanoparticles, there was around a 40% increase in the drug loading, from 46 μg/mg nanoparticles to 65 μg/mg nanoparticles, compared with PTX–CD–NP.

Figure 1 shows the microphotographs obtained by SEM and cryo-EM of the nanoparticles. Figure 1a corresponds to SEM image of the PTX–HPCD–NP, Figure 1b corresponds to cryo-EM image of PTX–CD–NP and Figures 1c and 1d belong to SEM images of PTX–CD–NP–PEG and of PTX–NP–PEG, respectively. Nanoparticles displayed spherical shapes and, in all cases, their apparent sizes were similar to the obtained values by PCS. It is noteworthy that formulations containing cyclodextrins (Figs. 1a–1c) presented more irregular and rough surfaces, whereas the nanoparticles formulated with PEG (Figs. 1c and 1d) showed more diffuse surfaces in the photographs.

In Vitro Release Study

Paclitaxel release kinetics from nanoparticles was evaluated in two different media: SGF and SIF fluids with polysorbate 80 as solubilizing agent for PTX. Figure 2 represents the release profiles of PTX from the different assayed formulations as cumulative percentage of drug released as a function of time.

For all the evaluated formulations, there was no release of the drug when nanoparticles were dispersed in SGF. In contrast, when nanoparticles were dispersed in the SIF, the release behavior of PTX from the poly(anhydride) nanoparticles exhibited a biphasic pattern. There was an initial rapid drug discharge for PTX–NP–PEG (60% released) followed by a more sustained release up to 18 h. For the formulations containing cyclodextrins (PTX–CD–NP and PTX–CD–NP–PEG), there was a first slow release of the drug maintained for 6 h. After 6 h, there was a higher release rate prolonged for the next 10–12 h.

Focusing on the time to achieve 50% release of PTX from the nanoparticles, differences were observed. For pegylated nanoparticles (PTX–NP–PEG), 30 min of incubation in SIF appeared to be enough to promote the release of 50% of the loaded drug. However, when the nanoparticles contained CD, this time lapse appeared to be longer. Thus, for the PTX–cyclodextrin complexes loaded in the nanoparticles, the time to achieve the 50% release was around 5 h of incubation in the intestinal simulated fluid. For the combined strategy incorporating CD and PEG in the nanoparticles (PTX–CD–NP–PEG), the release rate appeared to increase because just 2 h of incubation in the SIF was necessary to release 50% of the loaded drug. Interestingly, at 18–20 h, PTX was completely released to the medium in all the assayed formulations.

In Vitro Ussing Chambers

The transport and permeability of PTX when loaded in nanoparticles, through the intestinal mucosa of rats was investigated in vitro by the Ussing chamber technique.

Figure 3 represents the percentage of the dose of PTX absorbed through the jejunum portions in the Ussing chambers plotted against time for the absorptive direction (M–S). When the experiments were carried out with the nanoparticle formulations, there was a prolonged increase of the amount of PTX absorbed through the intestine for the first 30 min. In contrast, after 90 min a slower absorption profile was observed, becoming rather stable. At the end of the study, the amount of PTX in the receptor chamber for the nanoparticle formulations was five to eight times higher than for the commercial Taxol®. Thus, significantly higher levels of PTX present in the receptor chamber were observed for pegylated containing formulations than for PTX–CD–NP (p < 0.05).

In M–S direction, for Taxol®, the apparent permeability coefficient (P_app) was very low (1.77 × 10⁻⁶ cm/s). However, when PTX was encapsulated in nanoparticles, the intestinal permeability increased almost sevenfold, ninefold, and 12-fold for PTX–CD–NP, PTX–CD–NP–PEG, and PTX–NP–PEG, respectively. Comparing between nanoparticles, the highest permeability values were obtained for the PEG2000 containing formulation (PTX–NP–PEG) (P_app = 20.6 × 10⁻⁶ cm/s). For the two formulations containing the PTX–CD complexes, there were different values of permeability (P_app PTX–CD–NP = 11.4 × 10⁻⁶ cm/s versus P_app PTX–CD–NP–PEG = 15.1 × 10⁻⁶ cm/s) (p < 0.05).

The absorption enhancement ratios (R) are summarized in Figure 4 where the variation in absorption for the different formulations is plotted after a 2-h experiment in the Ussing chambers for all the experimental conditions evaluated.

When the apparent intestinal permeability was evaluated on the absorptive direction (M–S), differences were observed between the poly(anhydride) nanoparticle formulations. PTX–HPCD–NP presented a higher absorption ratio than PTX–CD–NP, around 1.5 times higher. Interestingly, when the PTX–CD complex was encapsulated in pegylated nanoparticles (PTX–CD–NP–PEG), a rise in the absorption ratio was evidenced, from an R of 6.7–9. On the other hand, PTX–NP–PEG presented the highest value of absorption ratio (R = 12.1). In addition, statistical differences were found between PTX–CD–NP formulation and the rest of poly(anhydride) nanoparticles (PTX–HPCD–NP, PTX–CD–NP–PEG, and PTX–PEG–NP) (p < 0.01).

In contrast, when the permeability was evaluated in the excretory direction (S–M), no statistical differences were observed between the commercial Taxol® and the poly(anhydride) nanoparticles with similar values of absorption ratio (R ≈ 18–20). On the other hand, when the intestinal permeability of PTX was assessed in the presence of verapamil in the medium, the values of P_app increased compared with the mucosal-to-serosal transport. In these circumstances, the absorption ratio for commercial Taxol® and for the different formulations rose to 20 approximately, in all cases (Fig. 4) with no statistical differences.

Pharmacokinetic Studies in C57BL/6J Mice

When Taxol® was administered as a single i.v. injection (dose 10 mg/kg) to laboratory animals, the PTX plasma concentration decreased rapidly in a biphasic way and the data were adjusted to a non-compartmental model (data not shown). The peak plasma concentration (C_max) of PTX was about 65 μg/mL. The values obtained for AUC and half-life (t_1/2) were 71 μg h/mL and 3.1 h, respectively. The MRT was 2.22 h (Table 2).

Figure 5 shows the plasma concentration versus time profile after oral administration of PTX (single dose of 10 mg/kg) included in the different nanoparticle formulations. Although orally administered Taxol® showed no detectable plasma levels in mice, the poly(anhydride) formulations offered sustained plasma curves characterized by increasing amounts of PTX for the first 1–2 h (T_max) reaching the maximum concentration (C_max) followed by another phase of slow and prolonged decline of the plasma levels until at least 24 h.

Comparing the different nanoparticle formulations, the PTX plasma levels reached with either PTX–NP–PEG or PTX–HPCD–NP or PTX–CD–NP–PEG were found to be between 1.5-fold and twofold higher than for PTX–CD–NP. In fact, PTX–NP–PEG displayed the highest plasma levels and, more importantly, remained within the detection range for at least 72 h postadministration. On the other hand, both PTX–HPCD–NP and PTX–CD–NP displayed PTX plasma levels for at least 24 h, whereas the pegylation of nanoparticles loaded with PTX–CD complexes offered plasma levels for at least 48 h. In addition, for this formulation (PTX–CD–NP–PEG), the second phase of the plasma curve followed a similar trend than the one obtained for the PTX–NP–PEG (see Fig. 5).

Table 2 summarizes the pharmacokinetic parameters estimated for a non-compartmental analysis of the experimental data obtained after the administration of the different PTX formulations to mice. The AUC of PTX in nanoparticles varied depending on the formulation. Thus, for PTX–HPCD–NP and PTX–NP–PEG, AUC was around 52–56 μg h/mL. For the nanoparticles containing just the PTX–CD complexes (PTX–CD–NP), AUC was around 44 μg h/mL. In contrast, when PEG was added to this formulation (PTX–CD–NP–PEG), the AUC increased (around 49 μg h/mL). Within the different nanoparticle formulations, the rank order of the mean AUC of plasma drug concentration versus time was as follows: PTX–NP–PEG > PTX–HPCD–NP = PTX–CD–NP–PEG > PTX–CD–NP (p < 0.05). In addition, the AUC values of poly(anhydride) nanoparticles were always significantly higher than those of orally administered Taxol® as no detectable levels were obtained. The peak plasma concentration (C_max) of PTX in the poly(anhydride) nanoparticles was between 2.4 and 4 μg/mL. These maximum plasma concentrations for the poly(anhydride) nanoparticles were around 16–35 times lower than that of the i.v. Taxol®.

In addition, MRT of the drug in plasma and the t_1/2 were found around 10 times higher when PTX was administered in the nanoparticle formulations by the oral route than when i.v. administered as Taxol®.

When Taxol® was i.v. administered, the Cl of the drug was 2.7 L/h and the V was 7 L, approximately. Interestingly, the Cl values of the anticancer drug when loaded in nanoparticles were always relatively lower than for the commercial Taxol®.

Finally, the relative oral bioavailability of PTX when incorporated to poly(anhydride) nanoparticles varied from 60% to 80% depending on the formulation. The highest relative oral bioavailability of the drug was obtained by the pegylated nanoparticles PTX–NP–PEG (79%), followed by the PTX–HPCD–NP (72%), PTX–CD–NP–PEG (68%), and finally, the lowest achieved by PTX–CD–NP (61%).