Journal of Controlled Release 108 (2005) 97 – 111 www.elsevier.com/locate/jconrel
Physico-chemical characterization of polysaccharide-coated nanoparticles Caroline Lemarchand a,b, Ruxandra Gref a,*, Sylviane Lesieur a, Hubert Hommel c, Be´atrice Vacher d, Ahmed Besheer e, Karsten Maeder e, Patrick Couvreur a a
UMR CNRS 8612, School of Pharmacy, Chaˆtenay Malabry, France b BioAlliance Pharma, Paris, France c ESPCI, Paris, France d UMR 5513, Ecole Centrale Lyon, Ecully, France e Institute of Pharmaceutics and Biopharmaceutics, Halle, Germany Received 5 January 2005; accepted 19 July 2005 Available online 19 September 2005
Abstract A series of amphiphilic copolymers (PCL–DEX) made of poly(q-caprolactone) (PCL) side chains grafted onto a dextran (DEX) backbone, was used to modify the surface of PCL nanoparticles. PCL–DEX nanoparticles were prepared by a technique derived from emulsion-solvent evaporation. The purpose of the present study was to investigate the DEX coating (quantification, conformation, mobility) in order to better understand particle surface–protein interactions. The DEX coating was deeply examined using different complementary methods: zeta potential measurement, specific degradation of the DEX shell by dextranase, energy-filtering transmission electron microscopy coupled to image-spectrum electron energy-loss spectroscopy, electronic paramagnetic resonance, high performance size exclusion chromatography as well as nonspecific bovine serum albumin adsorption. All our data together supported a core–shell structure of the nanoparticles, DEX moieties constituting the external coating. The amount of DEX located on the nanoparticle surface was estimated to 70%. The organisation of the shell including chains density and mobility was found to be dramatically influenced by DEX molar mass. The steric repulsion conferred by the presence of DEX at the surface of the nanoparticles decreased the adsorption of albumin. The nanoparticle– protein interaction was, however, greatly influenced by the polysaccharide conformation onto the surface. D 2005 Elsevier B.V. All rights reserved. Keywords: Nanoparticle; Dextran; Poly(q-caprolactone); Surface; Core–shell
1. Introduction * Corresponding author. Universite´ Paris Sud, Faculte´ de Pharmacie, UMR CNRS 8612, tour D5, 2e`me e´tage, 5 rue JB Cle´ment, 92926 Chaˆtenay Malabry, France. Tel.: +33 1 46 83 59 09; fax: +33 1 46 61 93 34. E-mail address:
[email protected] (R. Gref). 0168-3659/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2005.07.014
Nanoparticles based on biodegradable polymers such as polyester or poly(alkyl cyanoacrylate) (PACA), are being extensively investigated as delivery systems for small drug molecules, proteins, peptides,
98
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
nucleic acids. These colloidal carriers have shown many advantages in terms of drug protection, transport and delivery [1]. However, since nanoparticles are recognized as foreign particles and quickly eliminated from the bloodstream after intravenous administration, many studies have been devoted to modification of their surface. Poly Ethylene Glycol (PEG) has often been used as a hydrophilic polymer to coat nanoparticles, causing their blood circulation time to increase [2–4]. Nevertheless, PEG only delays but does not abolish the capture of these nanoparticles by the monocyte phagocytosis system (MPS) [5]. Moreover, the covalent linkage of a specific ligand for active targeting onto the surface of PEG-coated nanoparticles is not easy from a chemical point of view [6]. For these reasons, the development of carriers coated with polysaccharides has progressed in parallel to that of PEG-coated particles. For example, in the liposome field, dextran, pullulan or glycolipids have been used to decrease the uptake of liposomes by the MPS [7,8], hyaluronic acid to confer bioadhesive properties for the local depot of drugs [9], and functionalized dextran to target specifically vascular smooth muscle cells [10]. In the case of nanoparticles, dextran or heparin have been used for the preparation of coated poly (methyl methacrylate) (PMMA) nanoparticles by radical polymerisation initiated by cerium IV ions [11]. The surface modification conferred by the presence of dextran or heparin was shown to reduce or to avoid complement activation [12], to decrease interactions with macrophage-like cell lines [13], and to increase the circulation time of the nanoparticles in a mouse model [14]. The same polysaccharides have been used to modify the surface of biodegradable nanoparticles such as those formulated using PACA [15]. Hydrophobically modified dextran with phenoxy groups or alkyl chains, used as emulsion stabilizer, also contributed to reducing the protein adsorption onto the surface of poly(lactic acid) (PLA) nanoparticles [16,17]. In addition to their ability to reduce nanoparticle recognition by the MPS, polysaccharides allow organs and tissues to be targeted in a specific way, due to their mucoadhesive properties. For example, a cationic chitosan coating was shown to enhance the interactions of nanocapsules made of poly(q-caprolactone) (PCL) with the corneal epithelium [18] and bioadhe-
sive hyaluronic acid was used to obtain a prolonged action of drugs in the eye [19]. Recently, we have synthesized a new family of comb-like copolymers in which hydrophobic polyesters (i.e. PLA, PLGA (poly(lactic co-glycolic acid) or PCL) were covalently grafted onto a polysaccharide backbone (dextran, chitosan, hyaluronic acid, amylose) [20]. Among the large family of amphiphilic copolymers, PCL–DEX copolymers have been particularly studied for their abilities to stabilize emulsions [21] and to prepare nanoparticles by an original technique of binterfacial migration and solvent evaporationQ [22]. However, none of these studies have investigated whether the polysaccharide truly coats the nanoparticles and, if so, whether a core–shell structure is obtained. Thus, the aim of this study was the direct demonstration of the surface modification of PCL– DEX nanoparticles by the DEX moieties. It also focuses on the organisation of the graft copolymer within the nanoparticles.
2. Materials and methods 2.1. Polymer synthesis and materials The detailed synthesis and characterization of the family of PCL–DEX comb-like copolymers is described elsewhere [20]. It was carried out by grafting preformed PCL chains onto the DEX backbone. Briefly, low weight average molar mass (M W) (2000– 3000 g/mol) PCL polymers with low polydispersity (b 1.2), bearing one carboxyl end group, were obtained by polymerising freshly distilled caprolactone at 230 8C in the presence of capric acid. The carboxyl function was activated with carbonyldiimidazole, allowing the linkage to DEX (Fluka, M W 5000 or 40,000 g/mol). DEX–PCLn copolymers were analysed by gel permeation chromatography (GPC) using a triple detection system (Viscotek, Houston, Texas, US) using two GMH-HR H columns mounted in series and heated at 60 8C. The mobile phase was N, N dimethylacetamide (DMAC) containing 0.4% LiBr at a flow rate of 0.5 ml/min. The injected volumes were 100 Al and sample concentrations ranged from 5 to 10 mg/ml. 1 H NMR spectra of the copolymers dissolved in DMSO-d 6 (99.9 atom % deuterium (D), Sigma-
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
Aldrich, Steinheim, Germany) were recorded with a 200 MHz Bruker B-ACS 60 spectrometer. Infrared spectra of dried polymer powders were recorded using a Bruker Vector 22 spectrometer and 16 scans were averaged for each sample. The copolymers used in this study are designated PCL–DEXx where x is the DEX M W value (5 or 40.103 g/mol). The DEX weight content in the copolymers was 33 wt.%. The solvents used for nanoparticle preparation were of analytical grade. All other chemicals were commercially available reagent grade. Distilled water (GibcoR) was used. 4-Amino TEMPO and 1,1V-carbonyl diimidazol (CDI) were bought from Sigma-Aldrich, Germany. 2.2. Determination of the water soluble fraction in the copolymer 1.5 mg copolymers were dispersed in 0.5 ml of distilled water and mixed by vortex and/or sonication. The copolymer remained insoluble in water. After ultracentrifugation of the dispersion (30 min, 4 8C, 540,000 g, Beckmann TL100 ultracentrifuge, USA), DEX content (unreacted DEX or DEX with lower degree of PCL substitution) was determined in the supernatant by anthrone assay [25] as described in the section bDex content of the nanoparticlesQ. 2.3. Nanoparticle preparation Nanoparticles were formed by a technique derived from emulsion-solvent evaporation described elsewhere [22]. Briefly, 5 mg of PCL–DEX copolymer were added into a vial with 1 ml of organic solvent (ethyl acetate) and 5 ml of pure water. The graft copolymer was insoluble in this mixture. However, it was progressively dispersed leading the o/w emulsion formation under magnetic stirring (3 min, 500 rpm, room temperature). The size of the emulsion droplets was then considerably reduced by sonication (Microprobe, 60 s, pulses of 1 s each, Vibra Cellk VC750, Sonics and Materials, Newtown, CT, USA). Finally, the organic solvent was evaporated (RotavaporR, 20 to 30 min at a constant temperature (25 8C), 10–30 mbar), leading to PCL–DEX precipitation in the form of nanoparticles. PCL nanoparticles were prepared by simple emulsion-solvent evaporation. Sodium cholate solution
99
was used as aqueous dispersion media, instead of pure water. To eliminate the free surfactant, the PCL nanoparticle suspension was dialysed 3 times against MilliQR water (dialysis membrane Spectra Por 4, cutoff 12–14 000 g/mol, Spectrum). For comparison purposes, PEG5000–PCL45000 nanoparticles were prepared by simple emulsion-solvent evaporation as previously described [2]. 2.4. Physico-chemical characterization of the nanoparticles 2.4.1. Nanoparticle sizing and zeta potential The hydrodynamic diameter of the nanoparticles was measured at 20 8C at pH 6.5 by Quasi-Elastic Light Scattering (QELS) after a 1/100 dilution in water using a Nanosizer (CoulterR N4MD, Coulter Electronics, Inc., Hialeath, FL, USA). Their surface charge was investigated through zeta potential measurements after a 1/100 dilution in KCl 1 mM (Zetasizer 4, with a multi-8 correlator 7032, Malvern Instruments). The zeta potential values were calculated using the Smoluchowski equation. 2.4.2. Nanoparticle morphology The morphology of the nanoparticles was analysed by scanning electronic microscopy (SEM) (LEO 9530, France) with a Gemini Column. The nanoparticles were dried on supports at room temperature and coated with a Pt/Pd layer (Cressington, 208 HR) under an argon atmosphere. 2.4.3. Chemical bond mapping of PCL–DEX nanoparticles Chemical bond mapping was performed using an Energy-Filtering Transmission Electronic Microscopy (EFTEM) and image-spectrum acquisition technique, called Image-Spectrum Electron Energy-Loss Spectroscopy (IMEELS) analysis, as described elsewhere [23,24]. Practically, the samples were immobilized on copper grids covered by a holed carbon film and thin film (3 nm) and analysed using a microscope LEO 912 fitted with an image analyser. During all IMEELS analysis the sample was maintained at low temperature (97 K) using a liquid nitrogen cooler. The IMEELS method consists of acquiring sequences of energy-filtered images with a small energy window (10 eV) in the back focal plane of the spectrometer.
100
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
This permits a part of the EELS spectrum to be obtained at any location of an energy-filtered image, interactively. Using the IMEELS technique, the postexperiment choice of the images is made easier and mapping is greatly optimised. Experimentally, the energy loss was scanned between 239 and 343 eV with an energy step of 1 eV in order to record the carbon K-edge. 2.5. BSA adsorption studies The bovine serum albumin (BSA) adsorption experiments were carried out in a 10 mM phosphate buffer (pH 7.4). A precise amount of nanoparticles was mixed with a BSA solution composed of 14CBSA (specific activity 14 ACi/mg) and BSA solution (100 Ag/ml). After 6 h of incubation at 37 8C, the samples were ultracentrifuged (30 min, 4 8C, 145,000 g, Beckmann L7 ultracentrifuge, USA) and the BSA concentration in the supernatant was determined by h-scintillation. We calculated C, the amount of BSA (mg) adsorbed by unit of nanoparticle surface area (m2) as follows: C ¼ A=S
ð1Þ
where A is the weight of the BSA adsorbed onto the nanoparticles from the BSA determination and S the surface area of the nanoparticles. S was calculated as follows: S ¼ 3=rd d
ð2Þ
where r is the mean hydrodynamic diameter of the nanoparticles measured by QELS, and d the density of the PCL–DEX or PCL nanoparticles measured using a high precision digital densimeter (Analyser Beer 2, Anton-Paar, Austria). Then, C was calculated by the equation: C ¼ ddrdA=3:
ð3Þ
2.6. DEX content of nanoparticles The content of dextran was measured by a modified anthrone method [25]. To determine the amount of DEX at the surface of the nanoparticles, enzymatic degradation of nanoparticles with endo-dextranase was carried out. Practically, nanoparticles were incubated in 50 mM acetate buffer (pH 5.5) with dextra-
nase (0.1 mg/ml) for 5, 15, and 30 min, 1, 5 and 24 h at 37 8C under shaking (900 rpm). Samples were then transferred into a water bath at 90 8C for 15 min in order to inactivate the enzyme. DEX content was measured in the following samples: Sample 1: the whole nanoparticle suspension which included the DEX available at the surface of the particles, the DEX anchored within their core and the DEX free in solution, Sample 2: the supernatant obtained after ultracentrifugation (30 min, 4 8C, 540,000 g, Beckmann TL100 ultracentrifuge, USA) of the nanoparticle suspension which contained the DEX free in solution and released from the nanoparticle surface after enzymatic treatment, if any, Sample 3: the pellet of the nanoparticles which contained the DEX remaining firmly associated with the particles (DEX in the core of the particles after enzymatic treatment). Briefly, 200 Al of samples were pipetted into vials placed in ice bath and covered with 5 ml of 0.5% solution of anthrone in sulphuric acid (84% w/w). After shaking the samples, they were heated to 90 8C for exactly 16 min. After cooling, the absorbance was read at 620 nm against a control containing water and anthrone. The concentration of DEX in each sample was calculated using a series of DEX solutions between 0.1 and 0.5 mg/ml as a calibration curve. For the samples corresponding to the supernatant and the pellet of the nanoparticles, the amount of DEX was calculated as a percentage of the total amount of DEX in the whole nanoparticle suspension (Sample 1). Furthermore, samples were assayed for the presence of reducing oligosaccharides, such as isomaltose using the Sunner reagent, as described by Miller [26]. Practically, 50 Al of the samples were added to 1.5 ml of Sunner reagent or DNS reagent (10 mg/ml 3–5 dinitro-salicylic acid, 2 mg/ml phenol, 10 mg/ml sodium hydroxide) in the presence of 100 Al of freshly prepared sodium sulfite. This mixture was incubated for 15 min at 95 8C. One ml of 40% Rochelle salt was added immediately to the mixture after the appearance of the coloration and before the mixture was cooled. After cooling, absorbance was read at 570 nm against a control containing water and DNS reagent. A range of solutions of glucose was used to calibrate the assay.
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
The determination of the weight average molar masses of reducing oligosaccharides was performed by HPLC–SEC using previously described experimental set-up [27–29]. A 30 0.75 cm TSK-G4000 PW column (Toyo Soda, Tokyo, Japan) preceded by a 2Am filter (Rheodyne, CA) was equipped with a Hitachi pump (model L-6000), a precision injection valve (Rheodyne, 50 AL sample loading) and, for sample detection, with a differential refractometer (R401 Waters Associates, France) connected to a computer. The eluent was an aqueous solution of 100 mM sodium acetate at pH 5.5 (1 ml/min flow rate). Polysaccharide standards (M W / M n b 1.2) and glucose, both supplied by Polymer Laboratories (Amherst, MA), with weight average molar masses M W of 1,660,000; 47,300; 22,800; 11,800; 5900; 738 and 180 g/mole were injected at a concentration of 1 mg/ ml. The elution parameter K d was calculated according to the following equation: Kd ¼ ðVe V0 Þ=ðVt V0 Þ
ð4Þ
where Ve, V 0 and V t are the sample elution volume, void and total volume. The void volume was determined experimentally from the intercept of the baseline with the half-height tangent to the left side of the elution peak given by the 1,660,000-standard. The total volume was given by the maximum of the elution peak of an aqueous solution of 105 mM sodium acetate. The particle size of the degraded PCL–DEX nanoparticles was followed by taking samples of the homogenized nanoparticles suspension during the degradation process. The surface charge of the degraded PCL– DEX nanoparticles was followed by zeta potential measurements. The morphology of the degraded PCL–DEX nanoparticles was analysed by SEM. Calculations of the average DEX surface density were made as follows. The surface density of DEX chains (1 / S) could be calculated from the ratio between the total number of DEX molecules at the surface of the nanoparticles (N DEX), as experimentally determined by enzymatic degradation followed by the anthrone assay and the nanoparticle surface area: 1=S ¼ NDEX =Snp :
ð5Þ
101
Nanoparticle surface area (S np) was calculated by the following equation: Snp ¼ nd4kr2 :
ð6Þ
Since, Vnp ¼ n4=3kr3
ð7Þ
Snp ¼ 3Vnp =r ¼ 3Mnp =rd
ð8Þ
where V np is the volume of n spherical particles with a mean hydrodynamic radius r, M np the weight of nanoparticles per milliliter of suspension and d the density of the nanoparticles. Afterwards, the total number of DEX molecules at the surface was calculated as follows: NDEX ¼ ðMDEX =MW ÞdN
ð9Þ
where M DEX is the mass of DEX at the nanoparticle surface (per ml of suspension), M W the DEX weight average molar mass and N Avogadro’s number. Introducing fraction a = (M DEX / M np) into Eq. (9), Eq. (5) becomes: 1=S ¼ N dddrda=3MW :
ð10Þ
2.7. Determination of dextran chain mobility Aqueous suspensions of PCL–DEX5000 or PCL– DEX40000 nanoparticles were prepared according to the method previously described and dialysed against distilled water (dialysis membrane 12–14 000 Da MWCO, spectra Por 4). The nanoparticles were then labelled with a nitroxide free radical containing probe, 4-amino TEMPO, using a specific method for the labelling of dextran. For this, 1 mg 4-Amino TEMPO and 3 mg of CDI were added to 1 ml of nanoparticle suspension with concentration 1 mg/ml. The mixture was stirred gently for 20 h after which it was dialyzed 3 times, each for 2 h against 2 l distilled water using ServaporR dialysis membrane (molecular weight cut-off 12,000–14, 000, pore dia˚ ). ESR measurements were carried out meter ca. 25 A using X-band spectrometer MiniScope MS200 9.3– 9.55 GHz (Magnettech, Berlin). Measurement parameters were: modulation 0.1 mT, microwave power 10 mW, sweep time 60 s., number of sweeps 90 and
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
gain 800. The oscillating magnetic field was maximal at the centre of the cavity whereas the electric field was minimal. The shape of the EPR spectra of nitroxide free radicals such as those of the 4-amino TEMPO is very sensitive to the Brownian motion of the probe. The absorption results from transitions between the energy levels of the Hamiltonian spin, as described elsewhere [30]. When the motion is fast, the spectrum consists of three narrow well resolved Lorentzian lines. In the slow tumbling region, the anisotropic part of the Hamiltonian spin is not completely averaged and broader lines appear. The fast motion spectrum is attributed to loops and tails protruding into solution with a high mobility. The slow motion spectrum is attributed to trains adsorbed on the solid surface with a restricted mobility [31].
3. Results 3.1. Characterization of nanoparticle suspensions The PCL–DEX5000 nanoparticles exhibited a mean hydrodynamic diameter of 200 nm slightly lower than that of the PCL–DEX40000 nanoparticles (270 nm). Zeta potential measurement indicated that part of DEX was indeed located at the surface of PCL–DEX nanoparticles since the strong negative surface charge of PCL (n = 51 mV) was partially (PCL–DEX5000: n = 23 mV) or almost completely (PCL–DEX40000 n = 14 mV) shielded when PCL was replaced by its PCL–DEX counterpart. 3.2. BSA adsorption studies The adsorption of BSA onto PCL, PCL–DEX and PEG–PCL nanoparticles was measured for comparison purposes. As can be seen in Fig. 1, when the weight DEX content in the copolymer was maximal (33 wt.%), the amount of BSA adsorbed onto the surface of the nanoparticles decreased, compared with the uncoated PCL nanoparticles. The BSA adsorption appeared to be lower onto the PCL– DEX5000 nanoparticles than onto the PCL–DEX40000. Furthermore, the amount of BSA adsorbed onto the surface of PCL–DEX5000 nanoparticles was similar to the adsorption onto PEG–PCL nanoparticles.
1.4
BSA adsorption (mg/m2)
102
1.2 1.0 0.8 0.6 0.4 0.2 0.0 PCL
PCL-DEX5000
PCL-DEX40000
PEG-PCL
Fig. 1. BSA adsorption onto the surface of PCL, PCL–DEX5000, PCL–DEX40000, and PEG–PCL nanoparticles, the last used as a control of sterically stabilized nanoparticles.
3.3. Enzymatic treatment of nanoparticles and their DEX content The DEX content of the whole PCL–DEX40000 and PCL–DEX5000 nanoparticle suspensions were similar to the DEX content in the corresponding graft copolymer (Table 1). The content of DEX detected in the supernatant of the PCL–DEX40000 nanoparticles without enzyme treatment was negligible (less than 1 wt.%). In contrast, 10 wt.% of DEX was measured in the supernatant of the PCL–DEX5000 33 wt.% nanoparticles. During the enzymatic degradation of the nanoparticles by dextranase, either modified DEX or free DEX or reducing oligosaccharides could be released from the copolymer matrix. The anthrone assay allows carbohydrates from DEX to glucose to be analysed quantitatively [25]. Indeed, DEX was hydrolysed into glucose by concentrated sulphuric acid and high temperature (90 8C) conditions into glucose, allowing detection by the anthrone reagent [32]. Fig. 2 shows the cumulative release of DEX from nanoparticles in the supernatant as a function of time. Initially, in the first 5 min of the degradation process, a fast release of DEX from the nanoparticles of PCL–DEX5000 (Fig 2a) or PCL–DEX40000 (Fig 2b) was observed. Thereafter, the amount of DEX released gradually reached a plateau. The amount of DEX in the supernatant correlated perfectly with the DEX content remaining in the nanoparticle pellet: DEX content increased in the supernatant, whereas it decreased in the same proportion in the pellet. The sum of the amount of DEX found in the supernatant and in the pellet was indeed close to 100%, corre-
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
103
Table 1 Determination of DEX content in graft copolymers and nanoparticles DEX content in copolymer (wt.%)
PCL–DEX5000 33 wt.% PCL–DEX40000 33 wt.%
DEX content in NP (mg DEX/100 mg NP)
Reaction mixture
NMR
IR
Soluble fraction
33 33
31 29
26 31
8 b1
a
Whole
Supernatant
29 28
10 b1
a The soluble fraction of DEX in the copolymer was determined using anthrone assay in the supernatant, after dispersion of copolymers in distilled water.
sponding to the total quantity of DEX in the whole nanoparticle suspension. Even when the concentration of dextranase was increased (N 400 Ag/ml) or the incubation time was prolonged over 24 h in the presence of dextranase, the amount of DEX finally recovered in the nanoparticle pellets was between 25% and 35%. Reduced oligosaccharides were assayed using the Sunner reagent, the concentration of isomaltose released in the supernatant was found to be less than 0.05 mg/ml, whereas the content of DEX in the nanoparticles was about 0.35 mg/ml. Increas-
a
100
ing the concentration of dextranase (500 Ag/ml) or the incubation period (48 or 72 h) did not increase the concentration of isomaltose found in the supernatant. Fig. 3 shows the result of the HPLC–SEC analysis of the dextran fractions contained in the supernatant. The action of dextranase on the nanoparticles led to the release of dextran oligomers whose weight average molecular mass corresponds to an average length close to five glucosidic units whatever the length of DEX in the nanoparticle-forming copolymers (Fig. 3). This result confirmed that dextranase degradation produced oligosaccharide chains significantly shorter than the initial DEX chains but did not lead to a complete hydrolysis into isomaltose or glucose.
DEX %
80 60
100000
40 20
10000
0
b
6
12 Time (h)
18
24
Mw
0
1000
100 100
DEX %
80 60
10 0.40
40
0.60
0.80
1.00
Kd
20 0 0
12
24
36
48
Time (h)
Fig. 2. Time-course of nanoparticle degradation by dextranase: weight percent of DEX released in the supernatant (5) and remaining in the pellet (n): (a) PCL–DEX5000 and (b) PCL–DEX40000.
Fig. 3. Variation of weight average molar mass of dextrans M W as a function of column parameter K d from standard polymers and glucose ( S ) and oligosaccharides produced by the action of dextranase on PCL–DEX5000 (+) and PCL–DEX40000 (D) coated nanoparticles. The dashed curve corresponds to the selectivity curve of the column fitted from the standard polymers and glucose by a polynomial of the third degree (log M W = 5.875 4.123 K d + 6.380 K 2d 6.740 K 3d).
104
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111 Degradation time (h)
3.4. Size and surface charge of dextranase-treated nanoparticles
1
6
24
-10 Zeta Potential (mV)
3.4.1. Nanoparticle diameter The mean hydrodynamic diameter of PCL nanoparticles increased to more than 1 Am in the acetate buffer in the presence of dextranase, whereas it still remained in acetate buffer alone (150 nm). The mean diameter of the dextranase-treated PCL–DEX40000 nanoparticles decreased by 50 nm during the 24-h degradation process (Fig. 4a). In contrast, PCL–DEX5000 nanoparticles increased in size within the first hours of the dextranase degradation process (Fig. 4b). After 24 h, the mean diameter increased and the polydispersity was higher than 0.6, due to nanoparticle aggregation. When the concentration of dextranase was increased to 500 Ag/ml, the size of nanoparticles from PCL–DEX40000 or PCL– DEX5000 decreased slightly during the first hour of reaction, and then extensive aggregation occurred (data not shown).
0
-15 -20 -25 -30 -35
Fig. 5. Zeta potential of PCL–DEX5000 (5) and PCL–DEX40000 (n) nanoparticles after treatment at various time with dextranase.
3.4.2. Nanoparticle surface potential Fig. 5 shows the evolution of the zeta potential of the PCL–DEX nanoparticles during treatment with dextranase. The zeta potential values decreased towards 30 mV for both copolymers. When PCL, used as a control, was incubated with dextranase during 6 h, its zeta potential increased from 51 to 35 mV. 3.5. Morphology of nanoparticles before and after enzymatic degradation
0.7
300
0.6 0.5
250
0.4 0.3 200
P.I.
Mean diameter (nm)
a
0.2 0.1
150
0.0 0
1
6
24
Degradation time (h)
3.6. Chemical bond mapping of nanoparticles 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
700 600 500 400 300 200 100 0 0
1
6
P.I.
Mean diameter (nm)
b
Scanning electron microscopy was performed on PCL–DEX nanoparticles before degradation (Fig. 6a) and after treatment with dextranase during 24 h (Fig. 6b). In Fig. 6a, the nanoparticle surface shows ripples which are probably due to the presence of the polysaccharide coating. After degradation with dextranase, the structure of nanoparticle was preserved but the surface appeared smoother.
24
Degradation time (h)
Fig. 4. Mean hydrodynamic diameter (5) and polydispersity index ( S ) of PCL–DEX40000 (a) and PCL–DEX5000 (b) nanoparticles after treatment with dextranase.
Copolymers made of PCL and DEX contain both oxygen and carbon. Using EFTEM coupled with IMEELS, a series of 105 images was acquired, which permitted the location of C–O bonds within the nanoparticle to be studied. After alignment of these images, mapping was obtained by using three image recording at a specific threshold and a background subtraction model. Compared to the image obtained at zero loss (Fig. 7a), Fig. 7b shows IMEELS data for mapping C–O bonds within the nanoparticle. The C–O bond was found in all the nanoparticles, in accordance with the molecular structure of the copo-
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
105
lymer PCL–DEX. However, the mapping of this bond shows a heterogeneous distribution between the core and the shell of the nanoparticles. A large number of C–O bonds were located inside the nanoparticles with higher density than that around the particles. 3.7. DEX chain mobility Fig. 8 shows the EPR spectra of the PCL–DEX5000 (a) and PCL–DEX40000 (b) amino TEMPO-labelled nanoparticles. The spectra in Fig. 8a and b are a superposition of a fast and a slower motion component for each of DEX5000 and DEX40000, respectively. We simulated the fast motion and slower motion components for each spectrum to determine the rotational correlation time. It was found to be of the order of 10 11 s for the fast motion part and 10 9 s for the slow motion part. We also integrated the experimental spectra as well as the simulated fast motion parts for DEX5000 (Fig. 8c) and DEX40000 (Fig. 8d). The inte-
Fig. 7. IMEELS in nanoparticle: image at zero-loss (a) and image of the C–O bond at 301 eV (b).
grated spectra clearly show that the slower motion component represents a considerable part of both spectra. Indeed, the ratio of the slower motion component to the fast motion component in DEX5000 nanoparticles is 4.37, while it is 1.9 for the DEX40000 nanoparticles. In order to make the comparison between the two types of nanoparticles clear, we superposed the slower motion components of both types of nanoparticles as seen in Fig. 8e. Using this approach, we found out that the fast motion component of DEX40000 is 2.03 times that of DEX5000.
4. Discussion
Fig. 6. Morphology of PCL–DEX40000 nanoparticles by SEM before (a) and after (b) 24 h treatment with dextranase.
The aim of the present study was to characterize the core–shell structure of nanoparticles prepared from two copolymers of the PCL–DEX family with a high content of DEX (33 wt.%). Whatever the
106
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
b
a
330
332
336 334 Field / mT
338
340
c
330
332
336 334 Field / mT
332
336 334 Field / mT
338
340
d
332
336 334 Field / mT
338
340
e
330
330
330
340
338
f
332
336 334 Field / mT
338
340
330
332
336 334 Field / mT
338
340
Fig. 8. ESR-spectra of spin labeled (a) PCL–DEX5000 and (b) PCL–DEX40000 nanoparticles labelled with 4-amino TEMPO. Bold lines represent experimental spectra, while dotted lines are simulated fast motion components. Integration of experimental (bold) and simulated (dotted) ESRspectra for (c) PCL–DEX5000 and (d) PCL–DEX40000. (e) Superposition of integrated experimental ESR spectra for PCL–DEX5000 (black) and PCL–DEX40000 (gray). (f) Comparison of simulated fast motion components of PCL–DEX5000 (black) and PCL–DEX40000 (gray). The curves were fitted from the superposed integrated experimental spectra (see box e of this figure).
molecular weight of DEX (5000 or 40,000 g/mol), these PCL–DEX copolymers allowed the formation of nanoparticles spontaneously without additional surfactant. The hydrodynamic diameter of the PCL– DEX5000 nanoparticles and the corresponding polydispersity index were somewhat lower than those of PCL–DEX40000 ones: 196 nm (IP 0.2) and 260 nm (IP 0.4), respectively, probably related to the thickness of
the DEX shell and its hydration behaviour (see discussion below). Of course, the DEX content in the graft copolymer depended on the amount of DEX added to the reaction mixture during copolymer synthesis. However, NMR and IR analyses have shown that the percentage of DEX was often slightly lower than the theoretical expected value because of the loss of less substituted copolymers during the purification
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
steps [20]. Thus, in the nanoparticles, 29 wt.% of DEX were recovered in the PCL–DEX5000 nanoparticles and 28 wt.% in the PCL–DEX40000 nanoparticles instead of the theoretical 33 wt.% (Table 1). Moreover, the total DEX content of the nanoparticle suspensions, as carried out by anthrone assay, was in accordance with the analysis of the graft copolymers. In the case of PCL–DEX5000 nanoparticles, the determination of the amount of DEX in the different nanoparticle fractions (supernatant and pellet) revealed the presence of 10 wt.% of DEX free in the supernatant whereas the soluble fraction of DEX in the graft copolymer represented 8 wt.% (Table 1). Thus, the final nanoparticle content mostly reflected the composition of the copolymer used for their preparation. The presence of DEX shell was identified using EFTEM coupled with IMEELS. Although the C–O bond is present in PCL as well as in DEX, the mapping of this bond in the nanoparticle clearly showed a dense inner core surrounded by a coating of different density (Fig. 7). In order to quantify more precisely this DEX coating, we performed a specific degradation of the DEX shell using an endo-dextranase which hydrolyses the (1Y6) linkages in DEX. It was clearly observed that this enzymatic treatment released DEX from the matrix of PCL–DEX nanoparticles (Fig. 2) without altering the nanoparticulate structure (Fig. 6). DEX release, which increased during the reaction period, proved that the polysaccharide was mostly located at the surface of the PCL–DEX nanoparticles. Since at the end of the degradation all DEX accessible to the enzyme was already released, it was possible to determine the quantitative distribution of the DEX in the nanoparticles by analysis: 60–80 wt.% of DEX were found to be at the surface of PCL–DEX nanoparticles, whereas 20–40% remained in the solid core structure. The persistence of negative charge around PCL–DEX nanoparticles ( 14 or 23 mV) as observed by zeta potential measurements could be explained by the presence of water dipole created by free hydroxyls of DEX. Moreover, the pellet of the dextranase-treated PCL–DEX nanoparticles could be easily solubilized in THF, a solvent in which PCL is very soluble. In contrast, pellets of PCL–DEX nanoparticles which were not in contact with the enzyme were not soluble in THF (data not shown). Nevertheless, the dinitro-salicylic acid assay did not detect isomaltose, which suggested that the fragments of
107
DEX released from the nanoparticles were not further reduced into isomaltose upon the action of dextranase. These fragments probably corresponded to very short DEX molecules. In order to evaluate the weight average molar mass of the DEX released from the nanoparticles, size exclusion chromatography (SEC) was performed. SEC results (Fig. 3) confirmed that dextranase degradation produced oligosaccharide chains significantly shorter than the initial DEX chains but did not lead to complete hydrolysis into isomaltose or glucose. It is also suggested that DEX backbone formed regular loops at the nanoparticle surface (for PCL–DEX5000 at least), which were cut into regular fragments of nearly 5 glucose units. To investigate the thickness of the DEX shell, we measured the evolution of the nanoparticle mean diameter as a function of the degradation time (Fig. 3). The thickness of the DEX shell of PCL–DEX40000 could be estimated to 22–25 nm but, in the case of PCL–DEX5000, the difference was too small for reliable measurement (Fig. 4), although calculation predicted that 0.40 DEX chains/nm2 covered the nanoparticles (Table 2). The compactness of this type of nanoparticles may explain this result, in accord with the zeta potential measurement. Indeed, as discussed above, the zeta potential of the PCL– DEX5000 ( 23 mV) was less than the PCL–DEX40000 one ( 14 mV) meaning that the potential at the shear plan between the particle and the fixed layer of ions is more negative relative with PCL–DEX5000 than with PCL–DEX40000. This suggests a difference of thickness or compactness of the DEX coating modifying the orientation of the water dipole around nanoparticles. Furthermore, the lower thickness of the DEX shell in PCL–DEX5000 could explain the size of the nanoparticle (196 nm), lower than those made from PCL–DEX40000 (260 nm). During the degradation, DEX was eliminated from the nanoparticle surface leading to the decrease of zeta potential (Fig. 5). The surface of the nanoparticles then became more
Table 2 DEX surface density (1 / S) for PCL–DEX5000 and PCL–DEX40000
PCL–DEX5000 PCL–DEX40000
DEX content at the surface (wt.%)
1 / S (DEX chains /nm2)
62 80
0.40 0.13
108
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
and more negative and hydrophobic, because of the progressive disappearance of the DEX coating layer. The fact that dextranase, as a protein, could be adsorbed onto the surface of the degraded nanoparticles also explained that the zeta potential of the degraded nanoparticles reached a plateau value ( 30 mV), comparable to that obtained with PCL nanoparticles incubated with dextranase ( 35 mV) for 6 h. The adsorption of dextranase onto the surface of the nanoparticles was probably responsible for the observed aggregation of PCL–DEX5000 nanoparticles, as for PCL nanoparticles, in accordance with previous studies of Mu¨ller et al. [33]. In order to evaluate the adsorption of protein onto the nanoparticle surface, we used BSA as a model protein since this protein is the most abundant in the blood (40 g/l). As shown in Fig. 1, BSA adsorption depends on the (co)polymer composition: it was between two and four times higher on uncoated
a
PCL nanoparticles than on PCL–DEX or on the PEG–PCL nanoparticles used as controls due to their well-known protein steric repulsion properties. Nevertheless, the amount of BSA adsorbed onto the uncoated PCL nanoparticles (1.2 mg/m2) was found to be slightly less than that on the PLA nanospheres (1.8–2 mg/m2) [16]. Although the DEX content is the same in PCL–DEX5000 and PCL–DEX40000 copolymers, their protein-rejecting abilities were different. Indeed, twice as much BSA adsorbed onto the PCL– DEX40000 as onto the PCL–DEX5000 nanoparticles. Taking into account the structure of the PCL–DEX copolymers, bside-onQ is the only possible conformation of DEX. Whether DEX or PEG, the bside-onQ configuration was shown to be more effective than the bend-onQ one with respect to protein-rejecting abilities [34,35]. Nevertheless, in our case, although DEX was in bside-onQ conformation in both PCL– DEX5000 and PCL–DEX40000, their protein-rejecting
b
Protein
Protein rejecting
Hydrodynamic diameter
Mobility of DEX chains
Fig. 9. Hypothetical model of PCL–DEX5000 (a) and PCL–DEX40000 (b) nanoparticles.
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
abilities were different and could be related to the molecular weight of DEX. We calculated that DEX was three times more dense at the surface of PCL– DEX5000 nanoparticles than at the surface of PCL– DEX40000 one (Table 2). Furthermore, in the case of PCL–DEX40000, some hydrophobic regions could remain at the surface (see the hypothetical structure of the nanoparticles in Fig. 9). This could be due to the large loop conformation of the DEX units and possibly to the presence of chains more mobile, thus allowing a better penetrability for protein molecules. On the other hand, PCL–DEX5000 would be more anchored to the nanoparticle core with a more uniform coating, ensuring a better protection against BSA adsorption. Interestingly, this hypothesis is supported by the EPR results (Fig. 8) showing that the DEX chains of PCL–DEX40000 nanoparticles were more flexible and mobile than those of PCL–DEX5000. Indeed, the fast motion component of DEX40000 is about two times that of DEX5000. The slower motion component represents a considerable part of both spectra, probably due to the fact that the dextran coat is in the form of short loops because of the PCL anchorages with the core. DEX5000 chains appear to be more compacted onto the surface of the nanoparticles, whereas DEX40000 ones are looser. This might be the consequence of an inhomogeneity in the distance between the PCL grafts in PCL–DEX5000 and PCL–DEX40000. These two copolymers have similar mass composition of both hydrophilic backbone (DEX) and hydrophobic grafts (PCL). Thus, if grafting reaction was homogeneous, the distance between the grafts (or the number of glucose units) would have been the same in both copolymers. It was noticed that, during PCL grafting, the solubility of the long chain PCL–DEX40000 copolymer decreased. Therefore, grafting reaction might have occurred preferentially in certain domains of the DEX40000 backbone. This could explain why PCL– DEX40000 copolymer forms larger, looser loops at the nanoparticles’ surface. It is worth noting that the inhibition of BSA adsorption on the PCL–DEX5000 nanoparticles was comparable to that of the PEG–PCL nanoparticles. The surface modification of PEG–PCL was conferred by the presence of a PEG brush, meaning that the PEG was anchored to the nanoparticle core through ¨ sterone terminal site in a bend-onQ conformation. O
109
berg et al. [34] demonstrated that bside-onQ DEX immobilized onto a polystyrene substrate was effective in preventing the adsorption of protein, such as fibrinogen, and was comparable to the bend-onQ PEG, which is confirmed by our own data. Thus, the BSA adsorption demonstrates the steric protection of the dextran coating against protein adsorption and reveals the importance of the molecular weight of DEX and its consequences on the conformation of the polymer at the surface, for the protein-rejecting properties.
5. Conclusion Amphiphilic PCL–DEX copolymers were found to be able to form polysaccharide-coated nanoparticles, with evidence of a core–shell structure. The majority of the DEX (60–80 wt.%) was indeed exposed at the surface of the nanoparticles, modifying their interactions with proteins. The weight average molar mass of dextran was shown to influence the density of the DEX coating and the chain conformation which clearly modulated the protein-rejecting abilities. This study also shows that the synthesis of PCL–DEX copolymers with different molecular weight characteristics allows a good control over the polysaccharide shell of the particles obtained.
Acknowledgments The authors acknowledge CNRS and Bioalliance Pharma for their financial support, Dominique Costantini from Bioalliance Pharma for her helpful discussion, Jean Louis Pastol and Audrey Valette from CECM for the SEM experiments, Madeleine Besnard for help with the experimental part, He´le`ne Chacun for radioactivity assays, Vincent Dupuis for his helpful discussion in IMEELS experiments and Gillian Barratt for help in revising the style of the manuscript.
References [1] K.S. Soppimath, T.M. Aminabhavi, A.R. Kulkarni, W.E. Rudzinski, Biodegradable polymeric nanoparticles as drug delivery devices, J. Control. Release 70 (2001) 1 – 20.
110
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111
[2] R. Gref, Y. Minamitake, M.T. Peracchia, V. Trubetskoy, V. Torchilin, R. Langer, Biodegradable long-circulating polymeric nanospheres, Science 263 (1994) 1600 – 1603. [3] M.T. Peracchia, E. Fattal, D. Desmaele, M. Besnard, J.P. Noel, J.M. Gomis, M. Appel, J. d’Angelo, P. Couvreur, Stealth PEGylated polycyanoacrylate nanoparticles for intravenous administration and splenic targeting, J. Control. Release 60 (1999) 121 – 128. [4] R. Gref, Surface-engineered nanoparticles as drug carriers, in: M.-I. Baraton (Ed.), Synthesis, Functionalization and Surface Treatment of Nanoparticles, American Scientific Publishers, 2003, pp. 233 – 256. [5] D. Bazile, C. Prud’homme, M.T. Bassoullet, M. Marlard, G. Spenlehauer, M. Veillard, Stealth MePEG–PLA nanoparticles avoid uptake by the mononuclear phagocytes system, J. Pharm. Sci. 84 (1995) 493 – 498. [6] B. Stella, S. Arpicco, M.T. Peracchia, D. Desmaele, J. Hoebeke, M. Renoir, J. D’Angelo, L. Cattel, P. Couvreur, Design of folic acid-conjugated nanoparticles for drug targeting, J. Pharm. Sci. 89 (2000) 1452 – 1464. [7] M. Woodle, D. Lasic, Sterically stabilized liposomes, Biochim. Biophys. Acta 1113 (1992) 171 – 199. [8] T.M. Allen, The use of glycolipids and hydrophilic polymers in avoiding rapid uptake of liposomes by the mononuclear phagocyte system, Adv. Drug Del. Rev. 13 (1994) 285 – 309. [9] N. Yerushalmi, R. Margalit, Hyaluronic acid-modified bioadhesive liposomes as local drug depots: effects of cellular and fluid dynamics on liposome retention at target sites, Arch. Biochem. Biophys. 349 (1998) 21 – 26. [10] D. Letourneur, C. Parisel, S. Prigent-Richard, M. Cansell, Interactions of functionalized dextran-coated liposomes with vascular smooth muscle cells, J. Control. Release 65 (2000) 83 – 91. [11] C. Passirani, L. Ferrarini, G. Barratt, J.P. Devissaguet, D. Labarre, Preparation and characterization of nanoparticles bearing heparin or dextran covalently-linked to poly(methyl methacrylate), J. Biomater. Sci., Polym. Ed. 10 (1999) 47 – 62. [12] C. Passirani, G. Barratt, J.P. Devissaguet, D. Labarre, Interactions of nanoparticles bearing heparin or dextran covalently bound to poly(methyl methacrylate) with the complement system, Life Sci. 62 (1998) 775 – 785. [13] N. Jaulin, M. Appel, C. Passirani, G. Barratt, D. Labarre, Reduction of the uptake by a macrophagic cell line of nanoparticles bearing heparin or dextran covalently bound to poly(methyl methacrylate), J. Drug Target. 8 (2000) 165 – 172. [14] C. Passirani, G. Barratt, J.P. Devissaguet, D. Labarre, Longcirculating nanoparticles bearing heparin or dextran covalently bound to poly(methyl methacrylate), Pharm. Res. 15 (1998) 1046 – 1050. [15] C. Chauvierre, D. Labarre, P. Couvreur, C. Vauthier, Radical emulsion polymerization of alkylcyanoacrylates initiated by the redox system dextran–cerium (IV) under acidic aqueous conditions, Macromolecules 36 (2003) 6018 – 6027. [16] C. Rouzes, R. Gref, M. Leonard, A. De Sousa Delgado, E. Dellacherie, Surface modification of poly(lactic acid) nanospheres using hydrophobically modified dextrans as stabilizers
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25] [26] [27]
[28]
[29]
[30] [31] [32]
[33]
in an o/w emulsion/evaporation technique, J. Biomed. Mater. Res. 50 (2000) 557 – 565. C. Rouzes, M. Leonard, A. Durand, E. Dellacherie, Influence of polymeric surfactants on the properties of drug loaded PLA nanospheres, Colloids Surf., B Biointerfaces 32 (2003) 125 – 135. P. Calvo, J.L. Vila-Jato, M.J. Alonso, Evaluation of cationic polymer-coated nanocapsules as ocular drug delivery, Int. J. Pharm. 153 (1997) 41 – 50. S. Barbault-Foucher, R. Gref, P. Russo, J. Guechot, A. Bochot, Design of poly-epsilon caprolactone nanospheres coated with bioadhesive hyaluronic acid for ocular delivery, J. Control. Release 83 (2002) 365 – 375. R. Gref, J. Rodrigues, P. Couvreur, Polysaccharides grafted with polyesters: novel amphiphilic copolymers for biomedical applications, Macromolecules 35 (2003) 9861 – 9867. C. Lemarchand, P. Couvreur, C. Vauthier, D. Costantini, R. Gref, Study of emulsion stabilization by graft copolymers using the optical analyser Turbiscan, Int. J. Pharm. 254 (2003) 77 – 82. C. Lemarchand, P. Couvreur, M. Besnard, D. Costantini, R. Gref, Novel polyester-polysaccharide nanoparticles, Pharm. Res. 20 (2003) 1284 – 1292. J.M. Martin, B. Vacher, L. Ponsonnet, V. Dupuis, Chemical bond mapping of carbon by image-spectrum EELS in the second derivative mode, Ultramicroscopy 65 (1996) 229 – 238. L. Ponsonnet, B. Vacher, J.M. Martin, High resolution chemical mapping in the energy-filtering TEM: application to interface layers in ceramics, Thin Solid Films 324 (1998) 170 – 175. T.A. Scott, E.H. Melvin, Determination of dextran with anthrone, Anal. Chem. 25 (1953) 1656 – 1661. G.L. Miller, Use of dinitrosalicylic acid reagent for determination of reducing sugar, Anal. Chem. 31 (1959) 426 – 428. A. Walter, S. Lesieur, R. Blumenthal, M. Ollivon, Size characterization of liposomes by HPLC, in: G. Gregoriadis (Ed.), Liposome Technology, 2nd ed., CRC Press, Boca Raton, 1993, pp. 271 – 290. S. Lesieur, C. Grabielle-Madelmont, M. Paternostre, M. Ollivon, Study of size distribution and stability of liposomes by high performance gel exclusion chromatography, Chem. Phys. Lipids 64 (1993) 57 – 82. K. Andrieux, P. Lesieur, S. Lesieur, M. Ollivon, C. GrabielleMadelmont, Characterization of fluorescein isothiocyanate– dextrans used in vesicle permeability studies, Anal. Chem. 74 (2002) 5217 – 5226. L.J. Berliner, Spin labeling, in: L.J. Berliner (Ed.), Theory and Applications, A. press, New York, 1976, p. 592. H. Hommel, Polymers at solid interfaces: a spin labeling approach, Adv. Colloid Interface Sci. 54 (1995) 209 – 277. K. Langer, E. Seegmu¨ller, A. Zimmer, J. Kreuter, Characterization of polybutylcyanoacrylate nanoparticles: I. Quantification of PBCA polymer and dextrans, Int. J. Pharm. 110 (1994) 21 – 27. R.H. Mu¨ller, C. Lherm, J. Herbert, P. Couvreur, In vitro model for the degradation of alkylcyanoacrylate, Biomaterials 11 (1990) 590 – 595.
C. Lemarchand et al. / Journal of Controlled Release 108 (2005) 97–111 [34] E. Osterberg, K. Bergstrom, K. Holmberg, T.P. Schuman, J.A. Riggs, N.L. Burns, J.M. Van Alstine, J.M. Harris, Proteinrejecting ability of surface-bound dextran in end-on and sideon configurations: comparison to PEG, J. Biomed. Mater. Res. 29 (1995) 741 – 747.
View publication stats
111
[35] M.T. Peracchia, C. Vauthier, D. Desmaele, A. Gulik, J.C. Dedieu, M. Demoy, J. D’Angelo, P. Couvreur, Pegylated nanoparticles from a novel methoxypolyethylene glycol cyanoacrylate–hexadecyl cyanoacrylate amphiphilic copolymer, Pharm. Res. 15 (1998) 550 – 556.
