Langmuir 2008, 24, 9295-9301
9295
Novel pH-Responsive Nanoparticles Sergey Filippov,* Martin Hruby´, Cˇestmı´r Konˇa´k, Hana Mackova´, Milena Sˇpı´rkova´, and Petr Sˇteˇpa´nek Institute of Macromolecular Chemistry, Academy of Sciences of the Czech Republic, HeyroVsky Sq. 2, 162 06 Prague 6, Czech Republic ReceiVed March 21, 2008. ReVised Manuscript ReceiVed June 24, 2008 In this work we report a new type of pH-responsive micelle-like nanoparticle. Reversible nanoscale structures are formed in solutions of a pH-sensitive hydrophobic polyelectrolyte, poly(N-methacryloyl-L-valine) or poly(N-methacryloylL-phenylalanine), and nonionic surfactant (Brij 98) in the presence of hydrochloric acid. The influence of composition and pH on particles size and shape was investigated by a variety of methods. An entity’s size and polydispersity could be varied in a broad range making them a perspective candidate as a drug carrier. Unlike the case of typical micelles, our results indicate the presence of cavities in the formed particles. A hypothetical model of a nanoparticle and mechanism of formation are proposed.
Introduction Polymeric micelles recently attracted much attention in the development of drug delivery systems because of their nanometerscale size, ability to solubilize hydrophobic drugs in large amounts, and the ability to achieve site-specific delivery. Polymeric micelles built from amphiphilic block copolymers usually have a hydrophobic core surrounded by a hydrophilic corona in aqueous environment. The thermodynamic and kinetic stability of such micelles is controlled by the mass ratio of hydrophobic and hydrophilic blocks. The hydrophobic domain has a dual assignment. First, hydrophobicity forces the polymer molecules to assemble in aqueous environment and second, the hydrophobic nature can be used for the entrapment of hydrophobic drugs. There are considerable numbers of reports on the controlled release of hydrophobic drugs from the block copolymer micelles in aqueous solutions.1 In some cases it is advantageous to have nanoscale structures, which can disassemble and release their cargo as a function of pH, because the pH in the target compartment is different from that in the environment [e.g., the pH in stomach is 1-2, while in the small intestine it is ca. 8-9; * To whom correspondence should be addressed. E-mail: filippov@ imc.cas.cz. (1) Bae, Y.; Cabral, H.; Kataoka, K. Block Copolymers in Nanoscience, 1st ed.; Wiley-VCH: Weinheim, 2006; Chapter 4. (2) Lee, E. S.; Oh, K. T.; Kim, D; Youn, Y. S.; Bae, Y. H. J. Controlled Release 2007, 123, 19–26. (3) Gillies, E. R.; Frechet, J. M. Bioconjugate Chem. 2005, 16, 361–368. (4) Sawant, R. M.; Hurley, J. P.; Salmaso, S.; Kale, A.; Tolcheva, E.; Levchenko, T. S.; Torchilin, V. P. Bioconjugate Chem. 2006, 17, 943–949. (5) Bae, Y. H.; Nishiyama, N. S.; Fukushima Koyama, H.; Yasuhiro, M.; Kataoka, K. Bioconjugate Chem. 2005, 16, 122–130. (6) Na, K.; Lee, E. S.; Bae, Y. H. J. Controlled Release 2003, 87, 3–13. (7) Lee, E. S.; Shin, H. J.; Na, K.; Bae, Y. H. J. Controlled Release 2003, 90, 363–374. (8) Lee, E. S.; Na, K.; Bae, Y. H. J. Controlled Release 2003, 91, 103–113. (9) Na, K.; Lee, K. H.; Bae, Y. H. J. Controlled Release 2004, 97, 513–525. (10) Lee, E. S.; Na, K.; Bae, Y. H. Nano Lett. 2005, 5, 325–329. (11) Kim, G. M.; Bae, Y. H.; Jo, W. H. Macromol. Biosci. 2005, 5, 1118– 1124. (12) Gao, Z. G.; Lee, D. H.; Kim, D. I.; Bae, Y. H. J. Drug Target. 2005, 13, 391–397. (13) Lee, E. S.; Na, K.; Bae, Y. H. J. Controlled Release 2005, 103, 405–418. (14) Licciardi, M.; Giammona, G.; Du, J.; Armes, S. P.; Tang, Y.; Lewis, A. L. Polymer 2006, 47, 2946–2955. (15) Wei, H.; Zhang, X. Z.; Cheng, H.; Chen, W. Q.; Cheng, S. X.; Zhuo, R. X. J. Controlled Release 2006, 116, 266–274. (16) Schmaljohann, D. AdV. Drug DeliV. ReV. 2006, 58, 1655–1670. (17) Sakai, K; Smith, E. G.; Webber, G. B.; Schatz, C.; Wanless, E. J.; Butun, V.; Armes, S. P.; Biggs, S. J. Phys. Chem. B 2006, 110, 14744–14753.
tumor and inflammated places are generally more acidic (pH ca. 6.5) than normal tissues (pH 7.4)]. Several groups have reported micelles composed of pHresponsive polymers.2-17 A biologically active unit (drug) is released by disassembling the micelle in response to the change of the external pH. The pH-dependent disassembly of the micelles is in the most cases given by the presence of an ionizable group (a weak acid or a weak base) on the relatively hydrophobic polymer forming the micelle core. When the external pH increases, the ionization degree of the polymer may increase (weak polyacid) or decrease (weak polybase). In the case of relatively hydrophobic polymer backbone, changing the ionization degree strongly influences the hydrophilic-hydrophobic balance, since ionized forms are always much more hydrophilic than the corresponding nonionized species and also Coulombic interactions among the species of the same charge are increased when the net polymer chain charge is increased. Such an increase of polarity due to the increase in ionization degree finally leads to the suppression of the hydrophobic forces that drive the polymer assembly and a reversible pH-dependent disassembly occurs. A weak hydrophobic polyacid as a micelle core-forming block thus disassembles when the pH is increased and vice versa. Recently, a new way to construct polymeric nanoparticles has been proposed. The temperature or pH of the solution is gradually changed, so the polymer, which is in the beginning in an environment where it is molecularly soluble (i.e., in the mostly ionized form), starts to precipitate. Adding a surfactant will terminate the phase separation, leading to well-defined micellelike nanoparticles with low polydispersity.18,19 The aim of this work was to prepare and study the properties of such micelle-like nanoparticles built from the newly synthesized pH-responsive biomimetic hydrophobic weak polyacids poly(Nmethacryloyl-L-valine) and poly(N-methacryloyl-L-phenylalanine) (see Figure 1 for monomer structures). We investigated the aggregation behavior of these polymers as a function of pH and temperature for variable polymer/surfactant composition with respect to size, polydispersity, and shape of the nanoparticles in aqueous solution. The resulting nanoparticles were characterized by a combination of dynamic and static light scattering, smallangle neutron scattering (SANS), and atomic force microscopy (18) Konak, C; Hruby, M. Macromol. Rapid Commun. 2006, 27, 877. (19) Konak, C; Panek, J.; Hruby, M. Colloid Polym. Sci. 2007, 285, 1433.
10.1021/la801472x CCC: $40.75 2008 American Chemical Society Published on Web 08/08/2008
9296 Langmuir, Vol. 24, No. 17, 2008
FilippoV et al. Table 1. Samples’ Molecular Mass (Mw), Polymerization Degree (z), and Critical pH Transition (pHtr)
Figure 1. Chemical structure of poly(N-methacryloyl-L-valine) (P1) and poly(N-methacryloyl-L-phenylalanine) (P2) (sodium salt forms).
(AFM). Finally we assessed the observed structures in the context of their possible use as drug carriers.
Experimental Section Materials. All chemicals were purchased from Sigma-Aldrich Ltd. N-Isopropylacrylamide and N-isopropylmethacrylamide were recrystallized from hexane, and other chemicals were used without additional purification. Spectra/Por 3 dialysis tubing (molecular weight cutoff 3500 g/mol) was purchased from Serva Electrophoresis GmbH. Preparation of Hydrophobic Polyanions. Synthesis of Monomers. The amino acids L-valine and L-phenylalanine were methacroylated by a modified procedure:20 The particular amino acid (21.3 mmol, i.e. 2.49 g of L-valine or 3.51 g of L-phenylalanine, respectively) was dissolved in a solution of sodium hydroxide (42.5 mmol, 1.70 g) in water (20 mL) containing a few crystals of sodium nitrite as polymerization inhibitor. The clear solution was cooled to 0-5 °C and crushed ice (ca. 20 g) was added into the mixture. Methacryloyl chloride (21.3 mmol, 2.04 mL) was then added in one portion while the mixture was vigorously stirred. The mixture was allowed to warm to room temperature and allowed to stay at room temperature for 1 h. After that 35% aqueous hydrochloric acid (85 mmol, 7.5 mL) was added and the cloudy aqueous phase was extracted five times with chloroform (ca. 30 mL). The mixed chloroform extracts were washed two times with water (ca. 50 mL) and the chloroform layer was dried with anhydrous magnesium sulfate and evaporated in vacuo. The resulting crude oily product was dissolved in a minimum amount of methanol, toluene (100 mL) was added, and the solvents were evaporated in vacuo. This procedure was repeated three or four times until there was no methacrylic acid in the product according to 1H NMR [CDCl3; δ ) 5.60 (s)]. Purification was then finalized by drying the product in vacuo overnight. N-Methacryloyl-L-Valine (P1). Yield: 2.40 g (61%) of colorless oil. 1H NMR (CDCl3): δ ) 0.97 (m, 6H), 1.98 (s, 3H), 2.27 (m, 1H), 4.63 (m, 1H), 5.39 (s, 1H), 5.76 (s, 1H), 6.43 (d, 1H), 9.56 (s, 1H). Anal. calcd/found: 58.36/58.10 C, 8.16/8.25 H, 7.56/7.60 N. N-Methacryloyl-L-phenylalanine (P2). Yield: 3.57 g (72%) of slightly brownish oil. 1H NMR (CD3OD): δ ) 1.84 (s, 3H), 3.13 (m, 2H), 4.69 (m, 1H), 5.31 (s, 1H), 5.58 (s, 1H), 7.23 (m, 5H). Anal. calcd/found: 66.94/67.10 C, 6.48/6.70 H, 6.00/5.85 N. Synthesis of Polymers. All the monomers were polymerized in bulk at 60 °C for 48 h using azobis(isobutyronitrile) (AIBN) as an initiator (5 wt %). The resulting solid was crushed and dissolved in aqueous sodium hydroxide (2 wt % solution, 2 mol of NaOH/mol of monomer in polymerization mixture). The resulting solution was filtered through a ceramic filter (density S4) and dialyzed against a large excess of water for 48 h; water was replaced after 8 and 24 h. The solution of the particular polymer poly(N-methacryloyl-L-valine) (P1) or poly(N-methacryloyl-L-phenylalanine) (P2), respectively, in the sodium salt form (see Figure 1 for structures) was then filtered through a 0.22 µm filter and freeze-dried. The typical yield is ca. 30%. Molecular weights of the polymers were determined by static light scattering in formic acid and are presented in Table 1. Preparation of Nanoparticles. The particular polymer (P1 or P2, respectively) and the surfactant Brij 98 (B98) were dissolved in (20) Lynn, J. W. J. Org. Chem. 1959, 24, 1030.
sample
Mw, g/mol
z
pHtr
P1 P2
78 000 29 000
376 113
4.0 4.9
water and the pH was set to 8.0 with sodium hydroxide solution (1 mol/L). Nanoparticles were prepared by decreasing the pH of the Pi/B98 mixed solutions below the phase separation pHtr (see further). The pH drop was realized by an addition of HCl solution (0.1 M) into the stirred Pi/B98 solutions. A series of solutions with the same polymer concentration but with different surfactant concentrations was prepared. All concentrations are given in weight percent. Cmc and Cac Measurements. The cac (critical aggregation concentration) values were determined from concentration variations of the fluorescence intensity of the I1 band (λem ) 367 nm; the first peak on the emission spectra with λex ) 339 nm) of pyrene fluoroprobes in solutions with different concentrations of the polymer and a constant concentration of pyrene according to a previously reported method.21 Dynamic and Static Light Scattering. Measurements were carried out on an ALV instrument equipped with a 30 mW He-Ne laser in the angular range 30°-140°. The SLS data were analyzed using the Zimm plot procedure. The obtained correlation functions have been analyzed by REPES22 analytical software, providing a hydrodynamic radius distribution function, G(Rh). To account for the logarithmic scale on the Rh axis, all DLS distribution diagrams are shown in the equal area representation,23 RhG(Rh). pH and temperature dependences of the particle hydrodynamic radius, Rh, and scattering intensity, Is, were automatically measured at the scattering angle θ ) 173° on a Zetasizer Nano-ZS, Model ZEN3600 (Malvern Instruments, UK). For evaluation of data, the DTS (Nano) program was used. It provides a Rh distribution function together with polydispersity index derived from a cumulants analysis (PDI ) Γ2/Γ12). Here Γ2 and Γ1 are the second and first cumulants.24 Small-Angle Neutron Scattering. SANS experiments were performed at CEA-Saclay on the spectrometer PAXY of the Laboratoire Leon-Brillouin. Measurements were performed with a 128 × 128 multidetector (pixel size 0.5 × 0.5 cm) using a nonpolarized, monochromatic (wavelength λ set by a velocity selector) incident neutron beam collimated with circular apertures for two sample-to-detector distances, namely, 1 m (with λ ) 6 Å) and 7 m (with λ ) 8 Å). With such a setup, the investigated range of scattering wave vector modulus is from 5 × 10-3 to 4 × 10-2 Å-1. In all the cases reported in this paper, the twodimensional scattering patterns were isotropic so that they were azimuthally averaged to yield the dependence of the scattered intensity I(q) on the scattering vector q. Atomic Force Microscopy (AFM). The nanoparticles (concentration 2 mg/mL) were deposited on fresh mica substrate and characterized by atomic force microscopy. The surface morphology (height image) and the sum of tip-sample interactions (phase image) were characterized by AFM. All measurements were performed under ambient conditions using a commercial atomic force microscope (NanoScope Dimension IIIa, MultiMode Digital Instruments, Santa Barbara, CA) equipped with an NCLR POINTPROBE-Silicon SPM-sensor [probe covered by reflex aluminum coating (NanoWorld AG; spring constant 45 N m-1, resonant frequency 154 kHz]. The tapping mode AFM technique was used for collecting all images. This technique allows one to obtain two- and/or three-dimensional information of both height and material heterogeneity contrast with high resolution when recording height and phase shifts simultaneously. (21) Lee, E. S.; Shin, H. J.; Na, K.; Bae, Y. H. J. Controlled Release 2003, 90, 363. (22) Jakes, J. Czech. J. Phys. 1988, B 38, 1305. (23) Stepanek, P. Dynamic Light Scattering: The Method and Some Applications; Clarendon Press: Oxford, 1993; pp 177-240. (24) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814.
NoVel pH-ResponsiVe Nanoparticles
Figure 2. The pH dependence of the scattered light intensity (θ ) 173°) for (a) the polymer P1 solution (cP1 ) 1%) (0) and (b) the polymer P2 solution (cP2 ) 1%) (b).
Results and Discussion pH-Dependent Behavior of Hydrophobic Weak Polyacids. In order to monitor the aggregation behavior of both polymers, we conducted a stepwise titration of 1% solution of each of the polymers with 0.1 M hydrochloric acid. pH, intensity of scattered light, and particle size (DLS) measurements followed in 10 min. after of each HCl addition. The HCl addition decreases the ionization of polymers, which results in decreasing of solubility and macroscopic phase separation of both polymers. The phase transition manifests itself by an increase of solution turbidity followed by flocculation. Consequently, the intensity Is increases at the transition as demonstrated in Figure 2a,b. Arrows in the Figure 2a,b mark the location of the transition pH, pHtr (pHtr ) 4.0 and 4.9, respectively). pHtr was determined by finding a point of the sharpest change of the first-order derivative. Formation of Nanoparticles. Following a strategy proposed previously,18,19 we have tried to design nanoparticles by adding a surfactant into the polymer solution to prevent a macroscopic polymer phase separation. A suitable volume of 0.1 M hydrochloric acid was added to the polymer-surfactant mixed solution to decrease the solution pH below pHtr. We used the surfactant nonionic Brij 98. B98 is the commercial name for C18H35(OCH2CH2)20OH (M ) 1150 g/mol), a polyoxyethylene(20) oleyl ether surfactant. The number 98 refers to the sum of the length of the alkyl and the ethylene oxide chains. It is a nonionic surfactant containing poly(ethylene oxide) chains as the hydrophilic part and n-alkyl chains as the hydrophobic part and it forms micelles in water. The micellar structure and properties of similar surfactants have been extensively studied earlier.25-33 For all cases, a concentration of surfactant used for nanoparticles preparation was substantially above the cmc value (cmc ) 2.9 (25) Schefer, J.; McDaniel, R. P.; Schoenborn, B. J. Phys. Chem. 1988, 92, 729–732. (26) Borbely, S. Langmuir 2000, 16, 5540–5545. (27) Dutt, G. B. J. Phys. Chem. B. 2003, 107, 10546–10551. (28) Sharma, K. S.; Patil, S. R.; Rakshit, A. K.; Glenn, K.; Doiron, M.; Palepu, R. M.; Hassan, P. A J. Phys. Chem. B. 2004, 108, 12804–12812. (29) Sommer, C.; Pedersen, J. S.; Garamus, V. M. Langmuir 2005, 21, 2137– 2149. (30) Toth, G.; Madara, A. A. Langmuir 2006, 22, 590–597. (31) Heins, A.; Garamus, V. M.; Steffen, B.; Sto¨ckmann, H.; Schwarz, K. Food Biophysics 2006, 1, 189–201. (32) Suradkar, Y. R.; Bhagwat, S. S. J. Chem. Eng. Data 2006, 51, 2026– 2031. (33) Sommer, C.; Deen, G. R.; Pedersen, J. S.; Strunz, P.; Garamus, V. M. Langmuir 2007, 23, 6544–6553.
Langmuir, Vol. 24, No. 17, 2008 9297
Figure 3. Intensity weighted distribution function of nanoparticles obtained from DLS data: (a) (cP1 ) 0.8 wt % and cB98 ) 1.0%, pH 1.4; (b) cP2 ) 1.0% and cB98 ) 0.9%, pH 1.7.
× 10-3%) reported for Brij 98.34 Our surface tension measurements by Wilhelmy plate method give a lower cmc value (1.0 × 10-3%). Micelles formed above the cmc have an aggregation number of 3735 or 4936 (depending on the surfactant concentration) and a radius of about 5 nm (our data). The distributions of hydrodynamic radii obtained from dynamic light scattering measurements for nanoparticles prepared under similar conditions from both the polymers are shown in Figure 3. The hydrodynamic radius, Rh ) 25 and 15 nm was found for P1 and P2, respectively. A detailed inspection of a distribution function for nanoparticles (in volume-weighted and numberweighted representations) does not reveal any other peaks at lower Rh value, indicating that surfactant micelles existing prior to HCl injection have been partially disassembled. The polydispersity (PDI value) of nanoparticles is about 0.1. pH Properties of Pi/B98 Solutions. We monitored by DLS and SLS measurements the particles formation during systematic variation of pH (from pH 8 to 2). The distribution of Rh (volume weighted) as a function of pH is shown for polymer P1 in Figure 4. It is evident from Figure 4 that the formation of nanoparticles starts in the vicinity of the pHtr, which is manifested by a sharp increase in Rh from 5 nm (B98 micelles) up to 8 nm (nanoparticles have been assembled). The Rh of nanoparticles continuously grows below pHtr with deacreasing pH. The observed formation of well-defined pH-sensitive nanoparticles can be qualitatively explained by the model proposed in our previous papers.18 The nuclei of a condensed phase formed after the polymer solution passes through the phase transition temperature are monodisperse at the early stage of phase separation.37,38 They are hydrophobic and, therefore, they attract surfactants molecules, which leads to partial disassembly of existing surfactant micelles. If the adsorbed surfactant reaches a critical surface concentration, the nanoparticles are sufficiently solubilized and their growth stops. Growth of formed nanoparticles with decreasing pH below pHtr in our opinion is due to decrease of polymer ionization degree, which (34) Klammt, C.; Schwarz, D.; Fendler, K.; Haase, W.; Dotsch, V.; Bernhard, F. FEBS J. 2005, 272, 6024–6038. (35) Moore, Stephanie, A.; Glenn, Karen, M.; Palepu, Rama, M. J. Solution Chem. 2007, 36, 563–571. (36) Moore, Stephanie, A.; Harris, Adam, A.; Palepu Rama, M. Fluid Phase Equilib. 2007, 251, 110–113. (37) Cahn, J. W. Acta Metall. 1956, 4, 449. (38) Christian, L. W. The Theory of Transformations in Metals and Alloys, Pergamon: New York, 1981; Part I..
9298 Langmuir, Vol. 24, No. 17, 2008
Figure 4. Scattered light intensity Is (O) (logarithmic scale) and volume weighted Rh (2) (linear scale) plotted as a function of solution pH; cP1 ) 0.2 wt % and cB98 ) 0.2 wt %. For comparison, the intensity pH dependence for the pure polymer is partially depicted (9) (one can see the full curve in the inset).
Figure 5. Dependence of particles’ hydrodynamic radius (Rh) as a function of composition ratio f ) cP1/cB98: cP1 ) 0.1% (9), cP1 ) 0.2% (0), cP1 ) 0.4% (b), cP1 ) 0.8% (O).
further increases the hydrophobicity of polymer. Thus, polymers attract more and more unimeric surfactant molecules. To our best knowledge, there are quite a few reports on nanoparticle formation of polymers by interaction with nonionic surfactants.19,39 These findings suggest that the particles formed might exhibit the ability to carry and release a drug in a pHdependent fashion, e.g., nanoparticles would be stable in the stomach and they would dissolve in the intestines and release their cargo. Effect of Solution Composition f and Polyacid Concentration. We further investigated systematically the size and polydispersity of the particles as a function of solution composition ratio f ) cPi/cB98 and polymer concentration cPi. The hydrodynamic radii, Rh, of P1 and P2 nanoparticles are plotted for several polymer concentrations as a function of f, in Figure 5 and 6, respectively. Two trends were noticed: (i) The particle size increases with increasing f. While there is a nearly linear increase of the radius or diameter, Rh, for low f values, a sharp increase in particle size followed by phase separation is observed at high f values. The particle size can be changed 3-4 times by varying f. This is (39) Deo, P.; Somasundaran, D. Langmuir 2005, 21, 3950–3956.
FilippoV et al.
Figure 6. Dependence of particles’ hydrodynamic radius (Rh) as a function of composition ratio f ) cP2/cB98: cP2 ) 0.1% (b), cP2 ) 0.2% (O), cP2 ) 0.4% (9).
Figure 7. Dependence of nanoparticles’ polydispersity (PDI) as a function of composition ratio (f ) cPi/cB98): (a) cP1 ) 0.1% (0), (b) cP1 ) 0.2% (b), (c) cP1 ) 0.4% (O), (d) cP1 ) 0.8% (∆), and (e) cP2 ) 0.2% (9). Solid lines are the best fits of a polynomial of the second-order to the data. For simplicity reason, the PDI data are plotted with different offsets ∆Y ) 0.18 (b), 0.3 (c), 0.4 (d) in the Y axis. The Y axis scale is correct for samples a and e.
probably caused by the fact that the more surfactant present in solution relative to polymer, the more effectively the surfactant prevents phase-separated polymer from intermolecular aggregation by increased surface coverage. (ii) The particle size is also sensitive to polymer concentration for a fixed composition ratio f. The higher the polymer content is, the bigger the particles that are formed. It is probably due to the faster polymer aggregation before the surfactant stabilization at higher polymer concentrations. In the case of P1, phase separation occurs at lower values of f than for P2 at the same polymer concentration. We attribute this behavior to the different hydrophobicity of the amino acids that has been used for synthesis. The nanoparticles have a polydispersity index (PDI) in a range from 0.06 to 0.3 (see Figure 7). The PDI dependence on solution composition f can be well-fitted to a polynomial of the secondorder with a local minimum. The minimum shifts to lower f values with increasing polymer concentration (Figure 7). This fact could be explained by keeping in mind that the higher the polymer concentration, the bigger the nanoparticles that are
NoVel pH-ResponsiVe Nanoparticles
Langmuir, Vol. 24, No. 17, 2008 9299
Figure 9. The fluorescence intensity of the I1 band of pyrene fluoroprobes (the first peak of the emission spectra λem ) 367 nm; λex ) 339 nm) plotted versus log c (mg/cm3) of P1/B98; cP1 ) 0.1%, f ) 1.0.
Figure 8. (a) Volume-weighted Rh and reduced intensity Is/c for P1/ B98 (cP1) 0.5%, f ) 1.0) measured at four consecutive assemblingdisassembling cycles (adding hydrochloric acid to pH 1.4 with following neutralization of acid by NaOH). (b) Intensity-weighted distribution function of nanoparticles obtained from DLS for each cycle at pH 1.4.
formed (Figures 5 and 6) with higher surface area. Naturally, more surfactant is needed in solution to solubilize the created nanoparticles. In this way, the particles’ polydispersity can be tuned to a desired value. The polydispersity of nanoparticles reported so far was an invariable quantity or erratic in a limited range. Reversibility of Particle Formation and Critical Aggregation Concentration. The size and the polydispersity of the nanoparticles were the central issues up to now. Another important question is, are the pH changes of particle parameters reversible? To answer this question, the solution with already created nanoparticles was neutralized by the same molar amount of base (sodium hydroxide 0.1 M). Each assembling and disassembling cycle was repeated several times with continuous monitoring of Rh and reduced scattering intensity Is/c for both polymers (Figure 8a). The intensity-weighted distribution functions obtained after each addition of acid are shown in Figure 8b. One can see that neutralization disassembled the nanoparticles. The repeated addition of hydrochloric acid creates nanoparticles of a similar size. The distributions of original and new particles are not perfectly the same, although their average Rh values are almost the same. The origin of this difference might come from different preparation history, because the distribution function of new particles was measured immediately after addition of HCl and also from the continuous dilution of the solution by adding HCl
and NaOH. It reduces the scattering intensity, resulting to a more noisy correlation function. The nanoparticles thus can be practically reversibly assembled and disassembled. Another important parameter for nanoparticles is the so-called critical aggregation concentration. The cac was determined by fluorescent spectroscopy. The fluorescence intensity data of the I1 band of pyrene fluoroprobes shows the presence of cac for both the investigated systems (see an example in Figure 9). The cac values found at cPi ) 0.1 and f )1.0 are 2.6 × 10-3% and 3.4 × 10-3% for P1 and P2 polymers, respectively. These values are close to the cmc value of Brij 98 surfactant (1.0 × 10-3%). This is indirect evidence for a connection of Brij 98 micellization with nanoparticle formation. To get detailed information on the structure of the nanoparticles, SANS experiments have been realized. Small Angle Neutron Scattering. Three different solutions of P1 particles have been chosen for SANS measurements. Figure 9 shows the q dependence of the intensity of neutron scattering Is for these particles measured at 36 °C in D2O/H2O mixture. The H2O comes from 0.1 M HCl solution used for the preparing the nanoparticles. The contrast, nevertheless, was big enough to accumulate a good scattering profile. Concentrations of polymer cP1 and compositions f of P1/B98 are given in the Figure 10 caption. In addition to SANS measurements, we also ran SLS experiments for all three samples. Measured Raleigh ratios have been scaled by certain factors and juxtaposed with SANS data. Scaled data are shown together with SANS results in Figure 10. The values of Rg and Mw were obtained by analyzing the Guinier plot, and calculated results are given in Table 2. The Rg values are proportional to the composition ratio f and are in a reasonable relation to Rh. Rg/Rh values are approximately 0.8-1 (see Table 2) for all samples to be higher than the theoretical prediction for hard spheres (Rg/Rh ) 0.775).40 The Rg/Rh values of about 1 are predicted for random polycondensates (so-called soft balls) and star molecules.41 Experimentally, Rg/Rh values about 1 were found also for polyelectrolyte complexes.42 Thus, it was concluded that the particles in this study may be considered as “soft balls”. Calculated molecular weight also follows to Rg, Rh dependence. Using obtained Rg and Mw, one can calculate a particle’s density, (40) Burkhard, W. AdV. Polym. Sci. 1999, 143, 113–194. (41) Burchard, W. AdV. Polym. Sci. 1983, 48, 1. (42) Oupicky´, D.; Konˇa´k, Cˇ.; Ulbrich, K. J. Biomater. Sci. Polym. Ed. 1999, 10, 573.
9300 Langmuir, Vol. 24, No. 17, 2008
FilippoV et al.
Figure 10. Scattered intensity Is as a function of the scattering vector q at a temperature of 36 °C for P1/B98 sample with cP1 ) 1.0% and with a variety of composition ratios f: (∆) f ) 1.0, (9) f ) 1.8, (O) f ) 0.5. The scaled SLS data are in the left part of the graph. Table 2. Samples Composition f, Radius of Gyration (Rg), Hydrodynamic Radius (Rh), Structure Ratio (Rg/Rh), Molecular Weight (Mw), and Particle Density (G) of Nanoparticles f ) cP1/ cB98 0.5 1.0 1.8
Rg (heavy water, 36 °C), nm 23.5 31.1 43.2
Rh (H2O, 25 °C), nm 24.1 34.5 53.5
Rg/Rh
Mw, g/mol
0.98 0.90 0.81
1.4 × 10 3.3 × 106 7.1 × 106
F, % 6
4.3 4.4 3.5
F (Table 2). The results for the nanoparticles used for SANS measurements are 4-5%. The F values are 4-5 times smaller than those found for copolymer micelles.43 That indicates that entities created in solution probably have some cavities. Therefore, we have utilized the DAMMIN software,44,45 which is an ab initio method for reconstructing a particle’s shape and possible internal structure from small-angle scattering data. This model starts with a particle built from a sphere of densely packed dummy scattering centers with subsequent annealing46 of those scattering centers to find a configuration that fits the scattering data. An advantage of this approach is that such calculations require no a priori knowledge about particles architecture. A single parameter required for simulation is a particle’s maximum size. It has been shown that this method correctly reconstructs the general shape of a number of biological objects, e.g., lysosomes44,45 The results of simulations on our nanoparticles are shown in Figure 11a-c. The stability of the calculations was inspected by performing 10 runs. Although structures simulated in each run for the same sample were not perfectly identical, general features were the same. In principle, this approach is valid for systems of identical particles. In our case, the use of the ab initio method could be justified by DLS data, showing very narrow and homogeneous distribution (Figure 7) of nanoparticles in all samples. A variation in shape was most prominent for the sample with f ) 0.5 (Figure 11a). For the other two samples, variation was manifested in minor details. These structures deserve more attention. The standard micelles are spheres with a dense hydrophobic core and loose hydrophilic corona. The observed nanoparticles are different: they resemble (43) Tuzar, Z.; Plestil, J.; Konak, C.; Hlavata, D.; Sikora, A. Makromol.Chem. 1983, 184, 2111–2121. (44) Svergun, D. I. Biophys. J. 1999, 76, 2879–2886. (45) Svergun, D. I. J. Appl. Crystallogr. 2007, 40, S10–S17. (46) Kirkpatrik, S.; Gelatt, C. D.; Vecci, M. P. Science 1983, 220, 671–680.
Figure 11. A hypothetical structures modeled by DAMMIN. Left and right images are front view and side view projections of the structure calculated in a single run.
oblate particles with surfactant shells wrapped around a porous polymer hydrophobic core. Such entities have much more cavities (holes) comparing to typical micelles. More developed and compact particles are observed at high f values. Models simulated by DAMMIN could be juxtaposed with the one proposed by Cabane and Duplessix.47 They investigated nanoparticles created in a mixture of PEO and SDS by the SANS method. Using a contrast variation method, they conclude that stoichiometric particles consist of SDS aggregates that are clustered in subunits randomly distributed around polymer chain. The authors came to the conclusion about the presence of SDS aggregates due to a peak in SANS curves when only SDS was visible. We did not find any peak or shoulders on our SANS curves. This fact implies that surfactant is mostly located on nanoparticles surface. Nevertheless, we cannot eliminate the Cabane and Duplessix model, since our SANS experiments were realized with protonated polymers and surfactant and, moreover, there is evidence for a connection between Brij 98 micellization and nanoparticle formation (cmc and cac values are close each other). The above results are in agreement with calculation of the mean density nanoparticle (see above). We believe that such features as the possibility of tuning of the shape, polydispersity, and internal structure (space for carrying of a drug) make these nanoparticles a prospective drug-delivery system. Atomic Force Microscopy. The AFM method in phase and height modes enables one to characterize local homogeneities (47) Cabane, B.; Dupleessx, R. J. Physique 1982, 43, 1529–1542.
NoVel pH-ResponsiVe Nanoparticles
Langmuir, Vol. 24, No. 17, 2008 9301
Figure 12. AFM images of sample P1 cP1 ) 0.2%, f ) 1.0: (a, b) 3-D height images (200 nm zoom) and (c) 2-D phase images (200 nm zoom).
of particles. According to DAMMIN simulation, we must observe a microstructure of nanoparticles. AFM proves that all particles have surface heterogeneities, as evident from the height image (Figure 12a,b) and more clearly from the phase image (Figure 12c). Generally, particles resemble an egg-shell structure. Such findings support the DAMMIN hypothetical structures (see above), indicating that surfactant molecules are wrapping around a porous polymer hydrophobic core.
Conclusion We have successfully tested a simple protocol for the preparation of nearly monodisperse micelle-like nanoparticles. The particle size and polydispersity are controlled by varying pH and the concentrations of polymer and surfactant. Particles with very narrow distributions (PDI ∼ 0.06) can be prepared. Analyses by DLS, SANS, and AFM support the formation of nanoparticles and show a reversible transition (assemblingdisassembling) in an aqueous solution by increasing or decreasing the acidity. The average diameter of the nanoparticles could be
varied from 25 to 200 nm. In contrast with standard micelles, these particles have a more porous structure, the density of which can be also adjusted by variation of the concentration of polymer and surfactant. Our experiments indicate that such particles could be attractive candidates for nanoscale drug-delivery applications by releasing their load on a pH change with a high sensitivity on the exact pH value. Acknowledgment. The authors acknowledge support by the Grant Agency of the Academy of Sciences of the Czech Republic (grants No. A100500501, A4050403 and A400480616) and by the Grant Agency of the Czech Republic (SON/06/E005) within the EUROCORES Programme SONS of the European Science Foundation, which is also supported by the European Commission, Sixth Framework Programme. The authors gratefully acknowledge support by the NMI3 Programme of the European Commission (Contract No. HII3-CT-2003-505925). Also we wish to thank Dr Frederic Nallet for his help with SANS experiments. LA801472X
