Oxidation-Sensitive Polymeric Nanoparticles - Langmuir (ACS


Oxidation-Sensitive Polymeric Nanoparticles - Langmuir (ACS...

0 downloads 86 Views 404KB Size

Langmuir 2005, 21, 411-417

411

Oxidation-Sensitive Polymeric Nanoparticles Annemie Rehor,† Jeffrey A. Hubbell,† and Nicola Tirelli*,†,‡ Institute for Biomedical Engineering and Department of Materials, Swiss Federal Institute of Technology (ETH) and University of Zurich, Moussonstrasse 18, CH-8044 Zurich, Switzerland, and School of Pharmacy and Centre for Molecular Materials, University of Manchester, Oxford Road, M13 9PL Manchester, United Kingdom Received September 2, 2004 We have recently demonstrated the possible use of organic polysulfides for the design of oxidationsensitive colloidal carriers in the form of polymeric vesicles, which are particularly suitable for the encapsulation of hydrosoluble drugs. In the present research we extend our efforts to carriers specifically suitable for hydrophobic molecules. Exploiting the living emulsion polymerization of episulfides, we have produced new cross-linked polysulfide nanoparticles. Here we demonstrate how this process allows the production of stable nanoparticles with a good control over their size and functionality. The nanoparticles showed negligible cytotoxicity on a fibroblast model; furthermore, they exhibited sensitivity to oxidative conditions, which first produce swelling and then solubilize the material.

Introduction The “enhanced permeability and retention (EPR) effect”1 describes the leakage of colloidal carriers through the vasculature endothelium and their accumulation in the proximal tissues. The occurrence of EPR predominantly in inflamed tissues or in regions of angiogenesis, like in tumors, has suggested its exploitation for a targeted delivery of therapeutic drugs. Inflamed tissues and also, as lately being emphasized, certain tumors are populated with activated macrophages or tumor associated macrophages, respectively, that release oxygen-reactive species when activated.2-4 We have designed a new kind of colloidal carrier that combines the above-mentioned passive targeting via EPR effect and is at the same time responsive to oxidative conditions, using this sensitivity for a responsive release of encapsulated bioactive compounds. We have chosen polymeric nanoparticles as a colloidal carrier, because they offer a good combination of physicochemical stability (higher, e.g., than for liposomes) and reproducibility of the EPR effect (better than in smaller carriers, such as micelles). In recent literature reporting polymeric nanoparticle carriers,5,6 the most common systems are based on poly(lactic acid)7 and poly(lacticco-glycolic acid) or poly(alkyl cyanoacrylates)8 (both recently reviewed).9 Generally, nanoparticles show a hydrolysis-induced degradation and erosion of the carrier * To whom correspondence should be addressed. Tel.: +44 161 275 24 80. Fax.: +44 161 275 23 96. E-mail: [email protected]. † Swiss Federal Institute of Technology (ETH) and University of Zurich. ‡ University of Manchester. (1) Matsumura, Y.; Maeda, H. Cancer Res. 1986, 46, 6387-6392. (2) Coussens, L. M.; Werb, Z. Nature 2002, 420, 860-867. (3) Blackwell, J. E.; Dagia, N. M.; Dickerson, J. B.; Berg, E. L.; Goetz, D. J. Ann. Biomed. Eng. 2001, 29, 523-533. (4) Bingle, L.; Brown, N. J.; Lewis, C. E. J. Pathol. 2002, 196, 254265. (5) Couvreur, P.; Tulkens, P.; Roland, M.; Trouet, A.; Speiser, P. FEBS Lett. 1977, 84, 323-326. (6) Kreuter, J. Pharm. Acta Helv. 1978, 53, 33-39. (7) Gurny, R.; Peppas, N. A.; Harrington, D. D.; Banker, G. S. Drug Dev. Ind. Pharm. 1981, 7, 1-25. (8) Couvreur, P.; Kante, B.; Roland, M.; Speiser, P. J. Pharm. Sci. 1979, 68, 1521-1524. (9) Panyam, J.; Labhasetwar, V. Adv. Drug Delivery Rev. 2003, 55, 329-347.

structure,10 which results in a sustained release of drugs; however, this phenomenon is not necessarily responsive to the pathological situation that the system is supposed to address. On the other hand, if a responsive release is desired, the molecular design must provide (a) sensitivity to the pathological environment but also (b) the means to avoid the unspecific and phagocytic uptake of colloids, which often leads to fast blood clearance. By reducing protein adsorption with a poly(ethylene glycol) (PEG) coating, colloids derive a sort of “invisibility” from the cells of the mononuclear phagocyte system,11 if a sufficient coverage of the particle is provided.12 Furthermore, targeting of selected cells can be achieved by decorating the particle surface with cell-specific ligands.13 Having a response to the oxidative environment of inflammatory reactions as a target, we have selected poly(propylene sulfide) (PPS) as an oxidation-responsive material for the design of nanoparticles. PPS, a hydrophobic polymer, is readily converted to hydrophilic poly(sulfoxide) or poly(sulfone) by mild oxidizing agents. It could, therefore, serve as a matrix for hydrophobic drugs, which are released during the solubilization or swelling of the polysulfide upon oxidation. PPS-PEG block copolymers have been recently explored for the preparation of PEG-ylated carriers in form of selfassembled, oxidation-sensitive vesicles.14-17 We have also described the preparation of PPS through an emulsion polymerization technique; this process yields, in one step, nanoparticles directly from monomers.18 The nanoparticles (10) Uhrich, K. E.; Cannizzaro, S. M.; Langer, R. S.; Shakesheff, K. M. Chem. Rev. 1999, 99, 3181-3198. (11) Illum, L.; Davis, S. S. FEBS Lett. 1984, 167, 79-82. (12) Gbadamosi, J. K.; Hunter, A. C.; Moghimi, S. M. FEBS Lett. 2002, 532, 338-344. (13) Moghimi, S. M.; Hunter, A. C.; Murray, J. Pharmacol. Rev. 2001, 53, 283-318. (14) Napoli, A.; Tirelli, N.; Kilcher, G.; Hubbell, J. A. Macromolecules 2001, 34, 8913-8917. (15) Napoli, A.; Tirelli, N.; Wehrli, E.; Hubbell, J. A. Langmuir 2002, 18, 8324-8329. (16) Valentini, M.; Napoli, A.; Tirelli, N.; Hubbell, J. A. Langmuir 2003, 19, 4852-4855. (17) Napoli, A.; Valentini, M.; Tirelli, N.; Mu¨ller, M.; Hubbell, J. A. Nat. Mater. 2004, 3, 183-189. (18) Rehor, A.; Tirelli, N.; Hubbell, J. A. Macromolecules 2002, 35, 8688-8693.

10.1021/la0478043 CCC: $30.25 © 2005 American Chemical Society Published on Web 12/02/2004

412

Langmuir, Vol. 21, No. 1, 2005

still displayed on their surface the emulsifier used during the polymerization, Pluronic F-127 (a PEG block copolymer used in pharmaceutics19 and coating),11 which provides an optimal emulsion stability and, in principle, a PEGylated surface. The derivatization of the end groups of this polymer can further provide the possibility to decorate the nanoparticles with bio-active groups. In our previous report we polymerized propylene sulfide starting from a difunctional initiator, forming linear polymer chains. Unfortunately, the resulting nanoparticles showed an unsatisfactory long-term stability toward agglomeration; in the present paper, we describe the technique we have used to overcome this problem, based on the formation of a cross-linked polymer network. Furthermore, we demonstrate that our preparation allows for a precise control of the size of the stable PPS nanoparticles and, finally, that the nanoparticles are responsive to oxidizing conditions. Materials and Methods Solvents and reagents were purchased from Sigma-Aldrich (Buchs, Switzerland) and Molecular Probes (Leiden, The Netherlands) and used without further purification. 1H NMR spectra were recorded on a 300-MHz Bruker spectrometer. Fourier transform infrared (FT-IR) spectra were recorded in attenuated total reflection mode on a Spectrum One Perkin-Elmer spectrometer. Gel permeation chromatography (GPC) was performed on 0.1% polymer solutions in tetrahydrofuran (THF) on a GPCViscotek model 300 TDA equipped with triple detection, using a universal calibration with poly(styrene) standards. Absorbance was measured with a Perkin-Elmer Lambda20 UV/vis spectrometer and fluorescence with a Perkin-Elmer LS50B luminescence spectrometer. Synthesis of the Polymerization Initiator. Pentaerythritol triallyl ether (20 mL, 54.6 mmol) was dissolved in 200 mL of dry toluene under an inert atmosphere and cooled to 0 °C. A total of 3 equiv‚s of NaH (3.91 g, 163.8 mmol) were added under stirring. Gas evolved during 15 min; 5 equiv‚s of allyl bromide (23.5 mL, 273 mmol, diluted in 50 mL of toluene) were then added dropwise, and the mixture was stirred overnight at room temperature (r.t.). The solvent was evaporated, and the obtained material was dissolved in 100 mL of CH2Cl2 and extracted twice with water and then with a saturated salt solution. The organic solution was dried over Na2SO4, and then the solvent was evaporated. After purification over a silica gel column in CH2Cl2, 13.8 g of pure product were obtained (46.5 mmol, yield 85%, conversion 100% from 1H NMR). 1H NMR (CDCl3): δ ) 3.47 (s, 2H, sCs CH2sOsCH2sCHdCH2), 3.95 (d, 2H, sCsCH2sOsCH2sCHd CH2), 5.12 and 5.25 (dd, 2H, sCsCH2sOsCH2sCHdCH2), 5.85 (m, 1H, sCsCH2sOsCH2sCHdCH2). FT-IR (thin film): 3081 (ν vinyl CsH), 2909 (νas methylene CsH), 2868 (νs methylene CsH), 2851 (νs methine CsH), 1647 (ν CdC), 1478 (δin plane vinyl CsH), 1082 (νas CsOsC), 990 and 919 (δout of plane vinyl CsH) cm-1 (disappearance of νOH peak at 3400-3500 cm-1). Pentaerythritol tetraallyl ether (10 g, 33.73 mmol) was introduced into a Schlenk tube under inert conditions and dissolved into 90 mL of THF. A total of 0.06 equiv‚s of azobisisobutyronitrile (1.44 g, 8.192 mmol) and 2.4 equiv‚s of thioacetic acid (23.3 mL, 327.7 mmol) were added under stirring. The reaction mixture was degassed by three freeze-and-thaw cycles in liquid nitrogen. The reaction mixture was then stirred during 18-20 h at 60-65 °C. After cooling at r.t., 5.0 g of Dowex resin 1 × 8 (previously treated with NaOH and washed with water until the pH was around 8) were added for removing residual thioacetic acid. The mixture was stirred for 1 h and then filtered of the resin, the solvent was evaporated, and the product was redissolved in CH2Cl2. After extraction with water, 5% NaHCO3, and then water, the organic phase was dried over Na2SO4. After solvent evaporation, 17.8 g of product were obtained (29.6 mmol, yield 88%, conversion 99% from 1H NMR). 1H NMR (CDCl3): δ ) 1.82 (m, 2H, CsCH2sOsCH2sCH2sCH2sSs (19) Hazot, P.; Pichot, C.; Maazouz, A. Macromol. Chem. Phys. 2000, 201, 632-641.

Rehor et al. COCH3), 2.33 (s, 3H, CsCH2sOsCH2sCH2sCH2sSsCOCH3), 2.95 (t, 2H, CsCH2sOsCH2sCH2sCH2sSsCOCH3), 3.35 (s, 2H, CsCH2sOsCH2sCH2sCH2sSsCOCH3), 3.42 (t, 2H, Cs CH2sOsCH2sCH2sCH2sSsCOCH3). FT-IR (thin film): 2973 (νas methyl CsH), 2926 (νas methylene CsH), 1694 (ν CdO), 1099 (νas CsOsC), 942 cm-1 (disappearance of vinyl δs). Nanoparticles Preparation. Initiator Activation. In a typical experiment, pentaerythritol tetrathioester was mixed with a molar equivalent of 0.5 M sodium methanoate in methanol for 10 min [FT-IR (thin film): disappearance of νCdO thioester at 1687 cm-1 and appearance of a band at 1645 cm-1; see Supporting Information]. Emulsion Polymerization. The monomer emulsion was prepared by dissolving between 0.1 and 2% (w/v) of Pluronic F-127 (MW 12 6 00) in 25 mL of degassed, double-distilled, and filtered water. The system was continuously stirred at 1000 rpm and purged with nitrogen for 60 min. Propylene sulfide (1 mL, 12.8 mmol) was then added and 10 min later was followed by the deprotected initiator (0.06 g, 0.128 mmol, corresponding to 0.510 mequiv of thiol). Five minutes later 1.5 M equivalents (152 µL, 1.021 mmol) of diaza[5.4.0]bicycloundec-7-ene (DBU) were added. The reaction was stirred under inert conditions for 24 h. Cross-Linking. The nanoparticles were exposed to air to produce disulfide cross-linking or reacted by Michael-type addition with 1 equiv of tetra(ethylene glycol) diacrylate (0.077 g, 0.255 mmol) in the dark overnight [FT-IR (thin film) shift from 1720 (νs unsaturated CdO) to 1735 cm-1 (νs saturated Cd O)]. The end capping with diacrylates was proven by the production of soluble materials after the basic hydrolysis of the ester bonds. After water removal at the rotary evaporator, the dry material was dispersed in THF and 2 equiv (per ester group) of 40% tetrabutylammonium hydroxide (0.66 mL, 1.02 mmol) were added; after stirring for 15 min the material became soluble. A total of 3 equiv of acetic acid (0.092, 1.54) were then added together with Dowex 50 1 × 8 resin, to remove excess tetrabutylammonium hydroxide. The resin was filtered away, the solvent was evaporated, and the viscous polymer was dissolved in 40 mL of CH2Cl2 and extracted three times against water. The organic phase was dried over Na2SO4, the solvent was evaporated, and the material was washed with cold methanol and then dried under a vacuum at r.t. The molecular characterization data are analogous to those obtained for previously synthesized linear polymers.18 1H NMR (CDCl ): δ ) 1.35-1.45 (d, CH in PPS chain), 2.553 3 2.65 (m, 1 diastereotopic H of CH2 in PPS chain),16 2.85-3.0 (m, CH and 1 diastereotopic H of CH2 in PPS chain),16 1.75-1.85 (q, CsCH2sOsCH2sCH2sCH2sS), 2.9-3.0 (t, CsCH2sOsCH2s CH2sCH2sS), 3.36 (s, CsCH2sOsCH2sCH2sCH2sS), 3.43.5 (t, CsCH2sOsCH2sCH2sCH2sS). FT-IR (thin film): 33002500 (νs OsH), 2958 (νas CH3), 2865 (νs CH3), 2920 (νas CH2), 1720 (νs CdO, carboxylic acid), 1108 (νas CsOsC in Pluronic and tetrathiol), 800-600 (ν CsS in PPS) cm-1. Bulk Fluorescent Labeling. At the end of a typical polymerization experiment 5-iodoacetamido-fluorescein was added in a 1:100 ratio to the reactive thiols. After 2 h the emulsion was exposed to air for the formation of disulfides and then dialyzed as described below. Particles were imaged with a confocal microscope after embedding in a fibrin gel. Briefly, 1 µL of 2% (w/v) labeled particle emulsion were mixed with 49 µL of fibrinogen (5 mg/mL) in phosphate-buffered saline (PBS), pH 8, which was polymerized by the addition of 5 µL of thrombin (50 µU/mL) and 1 µL of factor XIII (1 U/mL). After a brief mixing the forming gel was immediately placed on microscope slides and covered. Surface Fluorescent Labeling. The OH groups of Pluronic F-127 were functionalized by dissolving 1 g of Pluronic F-127 in 30 mL of THF. 1-Hydroxybenzotriazole, diisopropyl carbodiimide, and dansylglycine were added in 2.5, 2, and 2 equiv per Pluronic OH group. The mixture was stirred overnight, then dialyzed against a water trough and 8-kDa molecular weight cutoff (MWCO) dialysis membrane, and finally freeze-dried. The polymer was redissolved in dichloromethane and precipitated in cold diethyl ether. Particles with various ratios of unfunctionalized and functionalized Pluronic were prepared, and their fluorescence (excitation 330 nm, emission 543 nm) was measured before and after dialysis through a 50-kDa membrane.

Oxidation-Sensitive Polymeric Nanoparticles

Langmuir, Vol. 21, No. 1, 2005 413

Scheme 1. Deprotection of Initiator and Formation of Cross-Linked Particles

Purification. Particles were purified from remaining monomers and base by 2 days of repeated dialysis against distilled water through a membrane with a MWCO of 6-8 kDa (Spectra/Por). Free Pluronic F-127 was removed in a second dialysis step by dialysis through 300-kDa membranes. Nanoparticles Characterization. Particle size distributions were determined by a hydrodynamic chromatography method based on poly(styrene) nanoparticles calibration (PL-PSDA, Polymer Laboratories). The dry weight was obtained by freezedrying until weight equilibrium. Assuming that no Pluronic was lost during dialysis, the weight of PPS was determined. Surface tensiometry by the Wilhelmy plate method (CAHN, dynamic contact angle analyzer) was used to monitor the aqueous phase surfactant concentration via a calibration curve obtained with Pluronic solutions in water [γ ) -1.99 ln CPl + 35.82, with γ being the surface tension (dyn/cm) and CPl being the Pluronic concentration in water % (w/v)]. Langmuir Balance. Langmuir balance isotherms were recorded with a Lauda FW2 instrument at 25 °C. A total of 30 µL of a solution of Pluronic F-127 in dichloromethane (0.1 µmol/mL) was carefully deposited on the water surface, and the solvent was evaporated during 20 min. Isotherms were recorded at a compression speed of 40 cm2 min-1. Negative Staining Transmission Electron Microscopy (TEM). Coated copper grids were rendered hydrophilic by glow discharge. The emulsions were diluted with water (1:2), and 5 µL was pipetted on the grid and adsorbed during 2-3 min. Then the grid was negatively stained by placing the grid on a drop of 2% acid phosphotungstic acid during 45 s. The samples were imaged with a Philips 208 transmission electron microscope. Cytotoxicity. Human foreskin fibroblasts were cultured to confluence in Dulbecco’s modified Eagle medium with 10% fetal calf serum and then incubated together with a 5 wt % suspension of dialyzed and sterile filtrated particle emulsion during up to 5 days at 37 °C, then incubated with 0.02% ethidium bromide and 0.1% calcein in PBS during 10 min at 37 °C, washed several times with PBS, and immediately checked for red (staining of DNA in the cell core of dead cells) and green (cytoplasma of living cells) fluorescence. Oxidation. Nanoparticle dispersions (2% w/w) were exposed to 10 and 5% (w/w) H2O2, respectively. Particle degradation was followed by turbidity (optical density at 600 nm) and particle size measurements.

Results and Discussion As a result of the low glass transition temperature of PPS (Tg ≈ -60 °C) and of the complete absence of crystallinity, nanoparticles of linear PPS are in fact viscous oil droplets in a water dispersion. This viscous character, combined to the mobility of the Pluronic, which is only physically adsorbed on the colloid surface, make highly probable the occurrence of anelastic interparticle collisions, which result in agglomeration. The possible interdiffusion of macromolecules renders then the agglomeration process irreversible. This tendency toward agglomeration in water suspension is such that nanoparticles formed of linear polysulfide chains18 tend to completely coalesce in the form of oily films within a few weeks of the preparation. Nanoparticles of glassy or semicrystalline polymers [e.g., poly(styrene), poly(lactic acid), and others] are inherently more stable, because of the more elastic nature of the polymer and the much slower diffusive motions of the polymer chains. The same applies to natural latex, where it is the chemical cross-linking that avoids interdiffusion and confers elasticity and stability to a low Tg polymer. Having the latter example in mind, we have overcome the problem of poor stability by preparing cross-linked polysulfide nanoparticles. We have used a tetra-armed hydrophobic initiator, which, taking advantage of the living character of the episulfide polymerization, provides tetra-armed, star, end-reactive macromolecules. The addition of a bifunctional end capper [tetra(ethylene glycol) diacrylate] or the simple exposure to air and consequent oxidation of the thiolates to disulfides readily cures the system (Scheme 1). Emulsion Polymerization. To avoid uncontrolled dimerization of free thiols to disulfides and, thus, an uncertainty in the number of initiator groups, we have employed a protected tetrathiol (in the form of a tetrathioacetate). This molecule was then transformed into a

414

Langmuir, Vol. 21, No. 1, 2005

reactive tetrathiolate in situ, according to a method already reported elsewhere.20 By exposing the initiator to 0.5 M sodium methanoate in methanol, the thioester groups were readily hydrolyzed with the formation of a four-arm thiolate and methyl acetate as a byproduct (Scheme 1). The thiolate was then added to the emulsified monomer together with an organic base (DBU), which at the same time enhances the initiator solubility in the oil phase and increases its reactivity (“naked” thiolate). After these additions, polymerization took place readily; a more detailed study of the polymerization has been reported elsewhere.18 Curing. The tetra-armed, thiol-terminated polymers were readily cross-linked by exposure to air and subsequent formation of disulfides. We assume the conversion of thiols into disulfides to be quantitative: after crosslinking free thiols seem to be absent, because a thiol scavenger such as iodoacetamide is incorporated in negligible quantities in the nanoparticles (as from IR analysis of the amide carbonyl stretching band). If kept under an inert atmosphere (no disulfide formation), thiols quantitatively reacted with Michael-type acceptors, such as acrylates. The use of bifunctional acrylates allowed for the introduction of cleavable linkers; in the simplest case, the cleaving sites are the acrylic esters, which can be hydrolyzed in basic conditions; this reaction converted the particles into polymers soluble in the most common organic solvents. GPC and 1H NMR provided M h n values for these polymers that corresponded to an almost quantitative monomer conversion. The fairly high polydispersity (1.54, as opposed to typical values of 1.1-1.3 for linear polymers) possibly indicates that some of the chains may nevertheless be bridged through disulfides, which are not readily cleaved in the basic hydrolysis step. Control Over Particle Size. Under the assumptions that (a) the emulsifier completely covers the surface of the monomer droplets and (b) this situation is “frozen” by the polymerization and the successive cross-linking, the size of the nanoparticles should depend on the ratio between the material in the bulk (polymerized monomer) and that at the water interface (surfactant). For example, if more surfactant is employed, a larger total surface can be covered, and, hence, smaller droplets and, later, smaller nanoparticles should be obtained. We have investigated in detail the dependence of the nanoparticle size on the Pluronic concentration, to validate the assumption of complete coverage by Pluronic (that ensures a perfect “PEG-ylation”) and also to accurately predict the particle size. In a series of experiments, we have prepared nanoparticles ranging in size from 25 to 225 nm by using different Pluronic concentrations and keeping fixed the monomer content. The average radius of a particle covered with surfactant, Rpart, is given by Rpart ) Rcore + RPl ) (3Vtot/Atot) + RPl, where Rcore is the radius of the “naked” particle, RPl is the thickness of the Pluronic layer, Vtot is the total volume of the polymer phase, and Atot is the total surface area covered by the surfactant at saturation. Knowing the polymer mass Mpol from dry weight measurements (equal to the mass of monomer times the polymerization conversion), Vtot can be expressed as a function of Mpol and the polymer density Vtot ) MPol/FPol. Always assuming a complete coverage of the PPS surface by Pluronic, Atot can be expressed as a function of the Pluronic surface area, APl, Atot ) APlNA(Pltot - Plaq)/MWPl, where Pltot is the total amount of Pluronic (in grams) and (20) Cellesi, F.; Tirelli, N.; Hubbell, J. A. Biomaterials 2003, in press.

Rehor et al. Table 1 Pluronic F-127 (wt %) 0.1

0.3

0.6

1

1.5

2

mean particle size (nm) 225 154 99 55 33 27 coefficient of variationa 35.4 33.2 47.3 52.9 25.4 35.8 b conversion (%) 49.2 75.4 64.24 72.5 61.6 64.8 free Pluronicc 2.59 1.61 0.95 0.19 0.11 0.08 (% of total content) a Standard deviation/mean particle size. b As determined by dry weight measurements. c As determined by surface tension measurement based on a calibration curve of Pluronic in water.

Plaq is the amount of Pluronic nonadsorbed on the surface; NA is the Avogadro number, and MWPl is the molecular weight of Pluronic. Plaq was obtained by surface tensiometry by the Wilhelmy plate method. Assuming that only nonadsorbed Pluronic contributes to the decrease in surface tension, the surfactant concentration in the aqueous phase was determined via a calibration curve of Pluronic solutions in water (see Supporting Information). As calculated with this method, the percentage of surface-adsorbed Pluronic was between 97.4 and 99.9% (Table 1) of its total amount. Rcore can, thus, be expressed as a function of the Pluronic/ monomer ratio, corrected by the polymerization conversion:

Rpart )

3MWPlMmon P1 + P2 F + RPl ) FPolAPlNAPltot x

(1)

where P1 ) 3MWPl/(NAAPlFPPS) and P2 ) RPl. We have assumed the density of cross-linked PPS to be roughly 1.1 g/cm3, which is intermediate between the values of crystalline and amorphous linear PPS. Variations up to 10% in this value will not seriously affect our calculations.21 Experimentally, the particle size was measured with hydrodynamic chromatography, a technique that provides differential and cumulative volume distributions (q3 respectively Q3). Similarly to our model, Rpart is calculated as the ratio between the total volume and the total surface of the particle population. By fitting the particle size data with eq 1, it is possible to calculate the area occupied by a single molecule and to compare it to the area of Pluronic on a water-air surface, which we separately determined by the extrapolation to zero pressure of the Langmuir isotherm of Pluronic F-127: 15.35 ( 0.23 nm2 (see Supporting Information). This is an upper limit for the area of Pluronic on the surface of nanoparticles: first, literature reports show that Pluronic adsorbed onto preformed hydrophobic nanoparticles may occupy a lower surface area than at the air/ water interface, probably because of partial interpenetration of the hydrophobic domain poly(propylene glycol) blocks [12 nm2 on poly(styrene)].22 Second, during polymerization and curing we indeed expect these molecules to feel increasing lateral pressure and “stretch”, because of the volume shrinkage of the colloids. We have shown that monomer droplets shrink already during the formation of linear polymers,18 and we expect this effect to be increased by cross-linking; we also expect that the formation of entanglements between Pluronic and PPS will not allow for a decrease in surface pressure by the release of emulsifiers in solution. Therefore, if the values extrapolated from eq 1 are similar to or below 15 nm2, we (21) Masamoto, J. In Polymer Data Handbook; Mark, J. E., Ed.; Oxford University Press: New York, 1999; p 792. (22) Stolnik, S.; Felumb, N. C.; Heald, C. R.; Garnett, M. C.; Illum, L.; Davis, S. S. Colloids Surf., A 1997, 122, 151-159.

Oxidation-Sensitive Polymeric Nanoparticles

Langmuir, Vol. 21, No. 1, 2005 415

Figure 1. Left: size distributions and negative stain TEM images for nanoparticles prepared with low Pluronic-PPS ratios. The bar in the images represents 500 nm. Right: dependence of the particle mean radius on Pluronic-PPS ratio (experimental values and modeling curve, R2 ) 0.91).

Figure 2. Fluorescently labeled nanoparticles embedded in fibrin gel (for avoiding Brownian motions) and visualized by confocal (left) and regular optical microscopy (right).

believe our assumption of complete surface coverage by Pluronic to be confirmed. Knowing the surface area occupied by Pluronic molecules, it will also be possible to estimate the number of functional groups (the termini of Pluronic chains) decorating an individual nanoparticle of a given size. The experimental data fitted with eq 1 provided the values of P2 ) 6.99 nm and P1 ) 6.7 nm (Figure 1); the latter converts into APl ) 8.16 nm2/molecule, a number that, if taken literally, should indicate a considerable reduction in the surface area of Pluronic. We think that, due to the approximation used in the Wilhelmy plate measurements (no interfacial activity of nanoparticles), the real value APl should be slightly superior to 8.16 nm2/ molecule, however, surely not above 15 nm2/molecule, ensuring the complete coverage of the nanoparticles. We have, therefore, demonstrated that the mean particle diameter can be controlled in a range of sizes (25-250 nm), which is appropriate for intravenously injected drug carriers. Additionally, the nanoparticles are characterized by a fairly narrow size distribution and their surface is likely completely covered by PEG chains (Table 1). Stability of the Nanoparticles. The stability of the Pluronic surface layer was the subject of a separate, recently published study conducted using pulsed gradient spin-echo (PGSE)-NMR.23 We have shown that unbound Pluronics (likely in a micellar state) can be easily removed (23) Valentini, M.; Vaccaro, A.; Rehor, A.; Napoli, A.; Tirelli, N.; Hubbell, J. A. J. Am. Chem. Soc. 2004, 6, 2142-2147.

Figure 3. Fluorescence of surface-labeled particles, before (filled squares) and after (open circles) dialysis.

by dialysis with membranes with a 300-kDa MWCO, without affecting the nanoparticles. Shelf life studies further showed that no free Pluronic was produced for periods up to several months, indicating a stable anchorage of the Pluronic to the nanoparticles. The study of the interactions of the nanoparticles with model proteins is the subject of an ongoing study, conducted through PGSENMR measurements. Preliminary results suggest that

416

Langmuir, Vol. 21, No. 1, 2005

Rehor et al.

Figure 4. Life/dead cell stain of human foreskin fibroblast after exposure to nanoparticles.

Figure 5. Particle emulsion turbidity during exposure to H2O2.

the Pluronic layer effectively hinders protein adsorption on the nanoparticles. Functionalization of the Nanoparticles. Bulk Functionalization. After emulsion polymerization, the thiolterminated polymers can be end-functionalized with “active” groups before curing. This could be explored for the introduction of fluorescent labels or bioactive molecules; the use of thiols for functionalization purposes is, however, limited to a degree that does not seriously modify the characteristics of the network As a proof of principle, the nanoparticles were labeled with fluorescein by reacting ∼1% of the terminal thiols

with an iodoacetamide-containing fluorescein derivative. The resulting particles were strongly fluorescent, even if a quantitative determination of the labeling yield has not been attempted. The labeled particles were successfully imaged with a confocal microscope (Figure 2) and will be used for future cell uptake and trafficking studies. Surface Functionalization. The PEG chains, anchored to the polysulfide core via hydrophobic interactions of the poly(propylene glycol) block, display terminal OH groups and are, therefore, amenable to functionalization. The presence of functional groups on a PEG surface could be employed for, for example, decorating the particles with targeting groups. Two routes of functionalization are possible: (1) functionalization of the adsorbed Pluronics after particles preparation or (2) use of a pre-functionalized Pluronic as a surfactant during the polymerization. The first approach allows for a greater flexibility, because the derivatization is performed as the last action, reducing the number of preparative steps. However, we have encountered purification problems when amphiphilic or lipophilic substances are used in the functionalization, because they can be incorporated in the bulk of nanoparticles in an unreacted form. Therefore, for a more general method, we have used pre-functionalized Pluronics; additionally, this method allows purification of the precursors and exactly determines the degree of functionalization by mixing functional and nonfunctional emulsifiers. Pluronics with different functional groups can be prepared and then employed in various combinations and concentrations as surfactants during the polymerization process.

Figure 6. Left: evolution of particle size distribution upon exposure to 10% H2O2. Right: nanoparticles at time zero and after 20 h of oxidation.

Oxidation-Sensitive Polymeric Nanoparticles

As a proof of concept we have used dansyl-labeled Pluronic; the resulting nanoparticles showed a good attachment of the fluorophores. The fluorescence intensity linearly increased with the dansyl content in the emulsifier mixture, demonstrating no preferential adsorption of one of the two compounds (Figure 3). Toxicity and Degradability of Material. The polymerization process is performed in water without any employment of solvents or toxic surfactants. Nonreacted monomer and the organic base are effectively removed by dialysis. The exposure of human foreskin fibroblast to the purified nanoparticles did not highlight any significant toxic effects for up to 5 days (Figure 4). The hydrophobic PPS core material is sensitive to oxidation, being transformed into poly(sulfoxides) and poly(sulfones), which are hydrophilic. We have used H2O2 for monitoring the oxidative degradation behavior of the nanoparticles, measuring both particle size and turbidity (Figures 5 and 6). Disulfide-bonded nanoparticles first swelled and finally completely dissolved in water, as a result of the oxidative cleavage of disulfides to sulfonates. The oxidized material permeated dialysis membranes with a 50-kDa cutoff, demonstrating the small size of the degradation products.

Langmuir, Vol. 21, No. 1, 2005 417

The polysulfide nanoparticles can be obtained at a size in the range 25-250 nm, depending on the monomer/ emulsifier ratio used in the polymerization. The nanoparticles are very stable (in nonoxidative conditions), display a PEG-ylated surface, and show no cytotoxic effect on cultured cells. An ongoing study, which will be the subject of a further publication, shows that they can be easily loaded with hydrophobic drugs. The nanoparticles are functional, and one can selectively label/derivatize their bulk, their surface, or both. Finally, the nanoparticles are responsive and can be solubilized by the action of water-based oxidizing substances, such as hydrogen peroxide. By using nanoparticles loaded with drugs, this phenomenon should trigger their release in the aqueous environment. The control on the molecular mass of the polymers, a consequence of the living character of the emulsion polymerization, allows the obtainment of soluble polymers with a mass low enough to allow for their excretion through kidneys. Acknowledgment. We thank Prof. Ba¨chi and his collaborators at the Electron Microscopy Central Lab, University Zurich, Switzerland, for help with confocal microscopy and TEM.

Conclusions We have demonstrated that the episulfide emulsion polymerization can be adapted to the preparation of colloidal structures with all the necessary features for an in vivo responsive behavior in an oxidative environment.

Supporting Information Available: Additional characterization data. This material is available free of charge via the Internet at http://pubs.acs.org. LA0478043