Intracellular drug delivery in Leishmania-infected macrophages ...

1 downloads 0 Views 941KB Size Report
Mar 7, 2011 - lead structures. RESEARCH ARTICLE. Intracellular drug delivery in Leishmania-infected macrophages: Evaluation of saponin-loaded PLGA.
Journal of Drug Targeting, 2012; 20(2): 142–154 © 2012 Informa UK, Ltd. ISSN 1061-186X print/ISSN 1029-2330 online DOI: 10.3109/1061186X.2011.595491

RESEARCH ARTICLE

Intracellular drug delivery in Leishmania-infected macrophages: Evaluation of saponin-loaded PLGA nanoparticles H. Van de Ven1, M. Vermeersch2, R.E. Vandenbroucke3,4, A. Matheeussen2, S. Apers5, W. Weyenberg1, S.C. De Smedt3, P. Cos2, L. Maes2, and A. Ludwig1 University of Antwerp, Laboratory of Pharmaceutical Technology and Biopharmacy, Faculty of Pharmaceutical, Biomedical and Veterinary Sciences, Universiteitsplein 1, CDE, Antwerpen (Wilrijk), 2610 Belgium, 2University of Antwerp, Laboratory of Microbiology, Parasitology and Hygiene (LMPH), Faculty of Pharmaceutical, Biomedical and Veterinary Sciences, Groenenborgerlaan 171, Antwerp, 2020 Belgium, 3Ghent University, Laboratory of General Biochemistry and Physical Pharmacy, Department of Pharmaceutics, Harelbekestraat 72, Ghent, 9000 Belgium, 4 Department for Molecular Biomedical Research (DMBR), Molecular Mouse Genetics, Ghent University-VIB, Belgium, and 5University of Antwerp, Laboratory of Pharmacognosy and Pharmaceutical Analysis, Faculty of Pharmaceutical, Biomedical and Veterinary Sciences, Universiteitsplein 1, Antwerp (Wilrijk), 2610 Belgium 1

Abstract Drug delivery systems present an opportunity to potentiate the therapeutic effect of antileishmanial drugs. Colloidal carriers are rapidly cleared by the phagocytic cells of the reticuloendothelial system (RES), rendering them ideal vehicles for passive targeting of antileishmanials. This paper describes the development of poly(D,L-lactide-co-glycolide) (PLGA) nanoparticles (NPs) for the antileishmanial saponin β-aescin. NPs were prepared using the combined emulsification solvent evaporation/salting-out technique. Confocal microscopy was used to visualise the internalisation and intracellular trafficking of fluorescein- and nile red-labelled PLGA NPs in J774A.1 macrophages infected with GFP-transfected Leishmania donovani. The in vitro activity of aescin and aescin-loaded NPs on L. infantum was determined in the axenic model as well as in the ex vivo model. The developed PLGA NPs were monodispersed with Zave 32 µg/mL) against bacteria, yeasts, fungi, viruses and protozoa other than Leishmania, indicating high selectivity (Maes et  al., 2004a; Maes et  al., 2004b). Bioassay-guided fractionation of the extract led to the isolation and structure elucidation of six closely related oleanane triterpene saponins, each with a different esterification pattern on the aglycone (Germonprez et al., 2005). The most active constituent of PX-6518, i.e. maesabalide III (MB-III), was administered in a single dose (0.8 mg/kg) to golden hamsters infected with L. donovani and was shown to have in vivo potency comparable to that of a single dose of AmBisome® (5 mg/kg) (Maes et al., 2004a). MB-III was, therefore, considered a promising new antileishmanial lead compound; however, systemic intolerance delayed its further development (Inocêncio da Luz et al., 2011). To enhance therapeutic efficacy and selectivity, a rewarding strategy consists of entrapping drugs in drug delivery systems (DDS) such as liposomes, polymeric nanoparticles and solid lipid nanoparticles (SLNs), hereby altering their biodistribution. The latter has nicely been illustrated by Ambisome®, the liposomal form of amphotericin B (Adler-Moore and Proffitt, 2008). Once in the bloodstream, DDS are rapidly recognized and taken up by the phagocytic cells of the reticuloendothelial system (RES). In fact, nanoparticles (NPs) with a size >150 nm and hydrophobic surface are cleared from the bloodstream within minutes upon intravenous injection (Chellat et al., 2005; Wong et al., 2008; Sheng et al., 2009). Drug selectivity is obtained because of this passive targeting towards the RES cells, which are the normal host cells for Leishmania amastigotes (Date et  al., 2007; Briones et al., 2008). Further, in order to improve uptake and tissue distribution to all macrophage-rich tissues the NPs’ size should be within the range 150–300 nm (Nahar and Jain, 2009). According to previous reports (Sinha et  al., © 2012 Informa UK, Ltd.

2002; Tyagi et  al., 2005), biodegradable polymeric NPs of poly(D,L-lactide-co-glycolide) (PLGA) may be a good choice as DDS for saponins, although these publications did not report on the targeting potential of PLGA NPs at the macrophage level. In this work, we emphasize the development of PLGA NPs to treat VL. Saponin-loaded PLGA NPs were prepared by means of an oil-in-water (O/W) combined emulsification solvent evaporation/salting-out technique. We preferred using the commercially available aescin as model triterpene saponin to the described maesabalides (Maes et al., 2004a) since the analytical method for aescin was available (Apers et  al., 2006). In order to investigate the influence of organic phase composition on the NPs’ physicochemical properties, a number of experiments were set up using simplex mixture designs. This allowed us to identify the optimal aescin-loaded PLGA NP formulation. Further, we provide in vitro evidence for the passive targeting potential of the developed NPs. This was accomplished by visualising the uptake of these particles by J774A.1 macrophages infected with L. donovani and subsequent determination of their in vitro antileishmanial activity. We performed a co-localisation study of fluorescently labelled PLGA NPs with green fluorescent protein (GFP)-transfected L. donovani amastigotes. Expression of GFP has been reported in both L. donovani and L. infantum (Dube et al., 2009; Bolhassani et  al., 2011); however, for the purpose of our study, the former was recommended. The infection of J774A.1 macrophages with GFP-expressing L. donovani promastigotes and subsequent intra-macrophage transformation to GFP-tagged amastigotes is already well established (Dube et  al., 2005; Dube et  al., 2009), whereas for L. infantum only GFP-tagged promastigotes are described (Kamau et al., 2001). Only very recently, Bolhassani et al. (2011) reported on the use of enhanced green fluorescent protein (EGFP)-transfected L. infantum promastigotes to infect bone marrow-derived macrophages and J774A.1 cells with the purpose of in vitro studies on the intracellular stages. For the antileishmania assay, we followed the primary drug screening protocol established in our laboratory (Maes et  al., 2004b; Vermeersch et  al., 2009) and therefore used L. infantum ex vivo amastigotes.

Methods Materials The PLGA polymer was Resomer® RG 503 (Boehringer Ingelheim, Germany) with a molecular weight of 40 kDa and a D,L-lactide:glycolide 52:48 molar ratio. Aescin was obtained from Fluka (Sigma, Belgium). Poly(vinylalcohol) (PVA), with an average molecular weight between 30 and 70 kDa, MgCl2.6H2O, p-anisaldehyde, NaHCO3, potato starch, Giemsa stain and resazurin were purchased from Sigma (Belgium), whereas RPMI-1640 medium, L-glutamine and foetal calf serum (FCS) were supplied by Invitrogen (Merelbeke, Belgium). The solid phase extraction (SPE) C18ec cartridges were obtained from Macherey-

144  H. Van de Ven et al. Nagel Filter Service (Eupen, Belgium), the silica gel 60 F254 HPTLC-plates were from Merck (Belgium). Milli-Q water was prepared with a Millipore water purification system (Millipore Co., Bedford, USA). The organic solvents acetone, dichloromethane (DCM), dimethyl sulfoxide (DMSO) and methanol (MeOH) were of analytical grade and purchased from Sigma (Belgium). The experimental drug PX-6518 was available in the laboratory from previous investigations. In brief, the leaves of the shrub Maesa balansae Mez. (Myrsinaceae) were collected in the Thai Nguen province in Vietnam (voucher specimens deposited at the Institute of Ecology and Biological Resources at National Center for Natural Sciences and Technology, Hanoi, Vietnam) and were dried, ground and extracted in a percolator in MeOH-water (9:1). The obtained PX-6518 extract was filtered, evaporated to dryness and stored at −20°C until further use (Maes et al., 2004b).

Parasites, animals and cell cultures L. infantum MHOM/MA(BE)/67 was kindly provided by the Institute of Tropical Medicine in Antwerp (Belgium) and was maintained in the laboratory by serial passage in golden hamsters (Mesocricetus auratus). Fresh ex vivo amastigotes were obtained from the spleen of an infected donor hamster. L. infantum ex vivo amastigotes were allowed to transform to extracellular promastigotes as described previously (Vermeersch et al., 2009). Primary mouse macrophages (PMM) were collected from Swiss CD-1 mice (Elevage Janvier, France) 2 days after intraperitoneal stimulation with 1 mL of a 2% (w/v) aqueous potato starch suspension. Cells were collected and grown in RPMI-1640 medium supplemented with 200 mM L-glutamine, 16.5 mM NaHCO3 and 5% (v/v) inactivated FCS at 37°C under 5% CO2. All animal experiments were approved by the Ethical Committee of the University of Antwerp (Belgium).

Preparation of PLGA NPs The PLGA NPs were prepared by means of the O/W combined emulsification solvent evaporation/salting-out technique. The organic phase consisted of a solution of 500 mg of PLGA and 100 mg of aescin in 10 mL of solvent mixture that was composed of various percentages of acetone, DCM and DMSO (or methanol). An O/W emulsion was obtained by dispersing this organic phase in 25 mL 1% (w/v) PVA and 60% (w/v) MgCl2.6H2O solution and sonicating for 1 min at 29 ± 1 W (amplitude set at 40%) (Branson Sonifier® Model S-450D, Branson, UK) on ice. The emulsion was then diluted in 120 mL 0.3% (w/v) PVA stabilizer solution. This dilution step induces the diffusion of water miscible organic solvents to the water phase and consequently the partial precipitation of PLGA into NPs within less than 30 min. Subsequently, the preparation was agitated with a magnetic stirrer (700 rpm) for 4 h at room temperature to fully evaporate DCM. Fluorescent PLGA NPs were prepared (i) by using PLGA covalently coupled to carboxyfluorescein and (ii) by dissolving nile red in the organic phase (ratio of 1:10 w/w to PLGA). The 

fluorescein-PLGA conjugate was synthesised as reported previously (Tosi et al., 2005; Dillen et al., 2008). The obtained PLGA NPs were purified by cross-flow filtration to remove non-incorporated aescin, excess stabiliser and salting-out agent, and the filtrate was collected. The NPs were filtered three times over regenerated cellulose membranes with a MWCO of 100 kDa (Vivaflow® 50, Sartorius) using a Masterflex® L/S pump (model 7518-00) and tubing (Sartorius, Germany). After the addition of mannitol (5% w/v) as cryoprotectant, the purified nanosuspensions were cooled down to −18°C and subsequently freeze-dried (Leybold-Heraeus D8B, GT-2A, Germany). The freeze-dried NPs were stored at +4°C.

Physical characterisation of PLGA NPs The mean particle size (Zave) of the NPs was determined by Photon Correlation Spectroscopy (PCS) with a Zetasizer 3000 (Malvern Instruments, UK). Freshly prepared PLGA NPs were diluted with deionised water. The measurements were performed in triplicate. The zeta potential of the NPs was determined with the Zetasizer 3000 (Malvern Instruments, UK) using Electrophoretic Light Scattering (ELS). Averages and standard deviations were obtained from ten consecutive measurements on the same sample.

Determination of drug loading The saponin drug loading of the PLGA NPs was determined directly by measuring the amount of aescin entrapped in the NPs. With this purpose, the NPs were redissolved in DCM followed by the extraction of aescin in water. However, the saponin extraction is incomplete because of the surface-active properties of these compounds. In order to circumvent this problem, DCM was allowed to evaporate during sample preparation. Briefly, 200 mg of freeze-dried PLGA NPs was accurately weighed and dispersed in 10 mL DCM to which subsequently 10 mL of water was added. The mixture was stirred on a magnetic stirring plate (400 rpm) during 30 min. During this period, the beaker was covered in order to prevent DCM from evaporating, so that the PLGA polymer could dissolve quantitatively. After 30 min, the DCM was allowed to evaporate, inducing precipitation of PLGA. The dispersion was centrifuged at 4000 rpm for 1 h at 4°C. The supernatant was subsequently transferred quantitatively on a 500 mg SPE C18ec cartridge previously conditioned with MeOH and water. From this point onwards, a SPE purification step was introduced to eliminate the mannitol that interferes in the HPTLC determination of aescin. After washing with water, aescin was eluted from the SPE cartridge with 18 mL MeOH. The eluate was diluted to 20.0 mL with MeOH and analysed for aescin content. The sample solutions were spotted in duplicate using a Camag automatic TLC sampler 4 (Camag, Switzerland) onto silica gel HPTLC-plates. The plate was developed using the upper layer of a mixture acetic acid-waterbutanol (10:40:50) as mobile phase, subsequently dried, dipped in p-anisaldehyde detection reagent EP 1007301 Journal of Drug Targeting

Drug delivery in Leishmania-infected macrophages  145 and heated for 5 min at 100–105°C. The remission absorption of the coloured product was measured at 535 nm by a Camag TLC scanner 3 (Apers et  al., 2006). Validation of the method according to ICH guidelines showed that it was precise (RSDbetween days 2.9%) and accurate (mean recovery 98.2%).

Confocal microscopy J774A.1 cells were seeded on sterile glass bottom culture dishes (MatTek Corporation, Ashland, MA) at a density of 3 × 104 cells per dish in RPMI-1640 culture medium, supplemented with Glutamax®, and allowed to attach overnight. The J774A.1 cells were infected with metacyclic promastigotes of a GFP-transfected L. donovani strain (Barak et al., 2005) at a ratio of 15 promastigotes per cell. The promastigotes were grown at 25°C and pre-conditioned in acidified promastigote medium (pH 5.4) at 37°C for 24 h prior to infection of the cells to enhance infectivity (Inocêncio da Luz et  al., 2009). Infected cells were cultured for 5 days during which period the internalised promastigotes transformed to dividing amastigotes. Freshly prepared fluorescent PLGA NPs were diluted 1/2 in culture medium and 20 µL (concentration, 1.8 mg/mL PLGA NPs) was transferred onto the (infected) J774A.1 cells in 300 µL culture medium. After incubation for 2 h, the cells were washed once with 1 mL culture medium. The distribution of fluorescence in the cells was visualised using a Nikon C1si confocal laser scanning module attached to a motorised Nikon TE2000-E inverted microscope (Nikon Benelux, Brussels, Belgium). For lysosome staining, cells were incubated for 60 min with 1:15,000 diluted LysoTracker Red™ (Molecular Probes) prior to image recording. Images were captured with a Nikon Plan Apochromat 60× oil immersion objective (numerical aperture of 1.4) using the 488, 561 and 639 nm line from an Ar-ion and a diode laser for the excitation of the fluorophores. Slides of untreated J774A.1 cells incubated under the same conditions were used as a negative control to determine microscope settings, which were maintained for all other experiments.

In vitro antileishmania assay Compounds Stock solutions of the reference saponins PX-6518 and aescin were prepared by dissolving the compounds in DMSO, and further dilution in Milli-Q water was performed so that the final in-test concentration of DMSO was below 0.6% (v/v). The aescin-loaded PLGA NPs were reconstituted in sterile water in aseptic conditions by means of the Branson Sonifier®. Serial two-fold dilutions of the compounds and NPs were performed in sterile Milli-Q water in sterile 96-well microtiter plates and were stored, if necessary, at 4°C. The in-test concentration range was 16 to 0.03 µg/mL saponin equivalents, taking the drug loading of the different aescin-loaded PLGA formulations into account. © 2012 Informa UK, Ltd.

Intracellular drug sensitivity assay The collected PMM were seeded in sterile 96-well microtiter plates at 3 × 105 cells/well and were left for adhesion and differentiation. After 48 h of incubation at 37°C and 5% CO2, the cells were infected with L. infantum ex vivo amastigotes at a multiplicity of 15 amastigotes per cell. Two hours post infection, cells were washed and the compound dilutions were added. After incubation for another 120 h, the cells were fixed with MeOH and stained with 10% (v/v) Giemsa solution for microscopic reading. The total intracellular amastigote burden (=average number of amastigotes per cell) in treated wells was compared to that of control wells (Vermeersch et al., 2009). All tests were performed in duplicate in two independent repeats. The 50% inhibitory concentration (IC50) was determined from the concentration-response curves using linear regression analysis. A compound is classified as active when the IC50 is lower than 1 µg/mL, a compound with IC50 between 1 and 16 µg/mL is classified as moderately active and a compound is classified as inactive when the IC50 is higher than 16 µg/mL (Cos et al., 2006). The antileishmania reference drug miltefosin (IC50, 5.2 ± 0.8 µM) was included in each test run. Axenic drug sensitivity assay Assays were performed in sterile 96-well microtiter plates, with each well containing 10 µL of the compound dilution and 200 µL of parasite inoculum (1 × 105 log-phase promastigotes/well), at 25°C under normal atmospheric conditions. Parasite multiplication was compared to that for untreated controls (100% growth) and uninfected controls (0% growth). After 72 h of incubation, resazurin was added and cell viability was measured fluorometrically (excitation λ, 550 nm; emission λ, 590 nm). The results were expressed as the percentage of reduction in the parasite burden compared to that in untreated control wells (Vermeersch et al., 2009).

Statistical analysis Statistical analysis of results was performed with Statistica® software (Statsoft, Tulsa, USA). One-way analysis of variance and post-hoc analysis (Tukey HSD test) were used to determine whether data groups differed significantly from each other. A p-value lower than 0.05 was considered statistically significant.

Results and discussion Preparation and characterisation of PLGA NPs Emulsification/solvent evaporation, emulsification/ solvent diffusion, emulsification/salting-out and solvent displacement (nanoprecipitation) are the four most commonly used methods for the production of PLGA NPs (Reis et  al., 2006; Vauthier and Bouchemal, 2009). The former three are based on the formulation of NPs starting from a nano-emulsion template, which generally consists of an organic PLGA solution and an

146  H. Van de Ven et al. aqueous stabiliser solution. The latter is a simple, fast and one-step method based on the Marangoni and/or ouzo effect, avoiding the use of high surfactant concentrations or high energy input (Quintanar-Guerrero et al., 1998; Beck-Broichsitter et  al., 2010). Nanoprecipitation is solely recommended for the encapsulation of hydrophobic drugs, whereas emulsification techniques, and more precisely the double or W/O/W emulsion technique, are also suited for hydrophilic drugs (Cohen-Sela et al., 2009). That is exactly why the preparation of NPs loaded with amphiphilic and water-soluble saponins represents a challenge. An extremely low drug encapsulation can be expected due to the surface-active properties of saponins, which drive these molecules to settle at the O/W interface. We hypothesized that a rapid Table 1.  Influence of organic phase composition on Zave and polydispersity index (PI) of aescin-loaded PLGA NPs prepared by the emulsification/salting-out technique. Formulation code

Zave ± SD* (nm)

PI* (−)

I II III IV V SD, standard deviation. *n = 3.

526.4 ± 238.8 479.5 ± 41.6 215.7 ± 15.5 247.9 ± 9.8 272.9 ± 1.5

0.77 0.93 0.11 0.09 0.05

particle formation would be preferable in this case, as drug leakage to the continuous aqueous phase can be reduced by decreasing the emulsion droplets’ solidification time (Park et  al., 1998; Cohen-Sela et  al., 2009). Emulsification/salting-out was thus considered method of choice and DMSO, an organic solvent that is completely miscible with water, was used as co-solvent. A minimum amount of 25% (v/v) DMSO was sufficient for complete solubilisation of the antileishmanial saponin in the organic PLGA solution. Methanol and ethanol are valuable alternatives to DMSO; however, these solvents have a negative influence on the solubility of the Resomer® RG 503 in the resulting organic mixtures. Besides, methanol belongs to the class 2 residual solvents (solvents to be limited), whereas DMSO is a class 3 solvent (solvents with low toxic potential) according to the International Conference on Harmonization (ICH) classification of residual solvents in pharmaceutical products (Reis et al., 2006). Application of the proposed classic approach with solely water-miscible solvents, such as acetone and DMSO in our case, led to the formation of very polydispersed particles, as can be derived from the measured polydispersity index (PI) values (Table 1 and Figure 1). Moreover, this occurred irrespective of the concentration of DMSO or acetone (Table 1 and Figure 1). This observation was corroborated by the work of McCarron et al. (2006). For further investigation,

Figure 1.  Composition of the organic phase of the 5 PLGA NP formulations prepared by the emulsification/salting-out technique (formulation I and II) or the combined emulsification solvent evaporation/salting-out technique (formulation III, IV and V). 

Journal of Drug Targeting

Drug delivery in Leishmania-infected macrophages  147 the volatile DCM was added to the organic PLGA/aescin solutions. This allowed us to produce PLGA NPs with acceptable PI values (PI ≤ 0.25) (Table 1 and Figure 1). However, as a consequence, the formation of the PLGA NPs will take longer, affecting the encapsulation of the amphiphilic compound. The extraction of acetone from the emulsion droplets upon dilution of the emulsion with a large excess of water occurs in a time scale of milli-seconds (Vauthier and Bouchemal, 2009), whereas the

diffusion and subsequent evaporation of DCM is a slow process because of its very low water-solubility (Park et  al., 1998). Consequently, a systematic study investigating the effect of organic phase composition on size, zeta potential and drug loading of the PLGA NPs was performed. The experiments were set up using a mixture design template (Statistica®) (Figure 2). The composition of the organic polymer solution had a statistically significant effect (p ≤ 0.001) on the average particle size (Zave).

Figure 2.  Composition of the organic phase of the 4 PLGA NP formulations prepared by the combined emulsification solvent evaporation/ salting-out technique.

Figure 3.  Effect of organic phase composition on (A) particle size, **p ≤ 0.001, between all formulations, (B) zeta potential, *p 16** PLGA NP B β-aescin-loaded 0.59 ± 0.24** ‡ >16** PLGA NP C β-aescin-loaded 0.76 ± 0.47** ‡ >16** PLGA NP D PLGA NP blank >246 >246 *n = 4 **IC50 expressed as equivalent concentration of aescin entrapped in PLGA NP † significant difference with free drug aescin, p ≤ 0.001 ‡ significant difference with free drug aescin, p