Article pubs.acs.org/Biomac
Stability of Chitosan Nanoparticles Cross-Linked with Tripolyphosphate Helene Jonassen,* Anna-Lena Kjøniksen, and Marianne Hiorth Department of Pharmacy, School of Pharmacy, University of Oslo, P.O. Box 1068 Blindern, 0316 Oslo, Norway S Supporting Information *
ABSTRACT: The physical stability of chitosan nanoparticles cross-linked with sodium tripolyphosphate (TPP) was investigated over a period of 1 month. Special emphasis was placed on changes in the particle size and the particle compactness, which are two important physicochemical parameters of nanoparticulate drug delivery systems. The chitosan−TPP particles were prepared at different ionic strengths, chitosan chloride concentrations, and TPP-to-chitosan ratios. In the presence of monovalent salt, the positive ζ potential of the particles was reduced. In spite of this, the particles were more stable when prepared and stored under saline conditions compared to water. This could be attributed to the smaller particle sizes found in the presence of sodium chloride. Most of the particles prepared in saline solvents were stable with respect to changes in the size and the compactness of the particles. However, instability was observed at the highest cross-linker-to-polymer ratios. Generally, a reduction in the ζ potential and an increase in the particle compactness were observed at increasing TPP-to-chitosan ratios. This combined with the size increase induced by a high concentration of chitosan, increased the aggregation and sedimentation tendency of the particles and reduced the colloidal stability of the particles.
■
INTRODUCTION Nanoparticulate systems offer many advantages in drug delivery, by, e.g., improving bioavailability, extending the therapeutic effect of the drug at the target site, and by improving the stability of the drug against chemical and enzymatic degradation.1 Essential physicochemical properties of nanoparticles, i.e., particles in the size range 1−1000 nm, include size, shape, compactness, and surface charge. These properties have been linked to many important drug delivery aspects, such as drug release profiles, bioavailability, biodistribution, and mucoadhesion.2−6 Preparation of polymer-based nanoparticles can be performed by cross-linking polymer chains together with covalent bonds, or by exploiting physical interactions between the polymer chains, such as hydrogen bonds, electrostatic forces, or hydrophobic associations.7 The attractive forces that can be exploited in the preparation of the particles, may lead to both contraction and aggregation of the particles over time. Estimation of the compactness of the nanoparticles, along with their size, can help to discriminate between intra- and interparticle associations. The particle compactness is also in itself an important parameter of polymer-based nanoparticles. For instance, a higher degree of swelling (i.e., lower compactness) of the nanoparticles has been linked to increasing drug release rates.2,4,8 In spite of these considerations, the compactness of nanoparticulate systems is seldom reported, probably due to experimental difficulties.7 Carbohydrates such as alginate and chitosan are often employed in the fabrication of nanoparticles due to their biocompatibility and biodegradability. Chitosan is a derivative of the naturally occurring carbohydrate chitin, and consists of beta 1−4 linked glucosamine units with varying amounts of N© 2012 American Chemical Society
acetylated units. The pKa value of chitosan has been reported to be approximately 6.3−6.6,9−11 and is affected by the degree of Nacetylation. Chitosan is readily soluble in an acidic environment due to protonation of the amine groups. The resultant positive charge makes it possible to prepare nanoparticles by ionotropic gelation with multivalent anions, such as tripolyphosphate (TPP). Chitosan−TPP nanoparticles have been studied for improved drug delivery of both small and large molecular weight drugs,3,10 and may therefore be considered a promising drug delivery system. Important factors that have been shown to affect the characteristics of these nanoparticles are the chitosan concentration, the TPP-to-chitosan ratio, the chitosan’s molecular weight, the pH, and the ionic strength of the dissolution medium used in the particle preparation.12−19 Recently, we investigated how the ionic strength of the solvent employed in the preparation of chitosan−TPP nanoparticles affected the size and the compactness of the particles 1 day after preparation.20 It was concluded that much smaller particles could be formed in saline solvents compared to pure water. In addition, when the particles were prepared under saline conditions, only moderate changes in the average particle size were found with increasing TPP-to-chitosan ratios. Furthermore, the compactness of the chitosan−TPP nanoparticles was found to increase with increasing solvent salinity and TPP-to-chitosan ratios. The size of the nanoparticles could be controlled by adjusting the chitosan concentration employed in the particle preparation, without affecting the compactness of the particles at low to Received: July 31, 2012 Revised: September 14, 2012 Published: October 9, 2012 3747
dx.doi.org/10.1021/bm301207a | Biomacromolecules 2012, 13, 3747−3756
Biomacromolecules
Article
Preparation of Particles and Storage Conditions. The chitosan−TPP nanoparticles were prepared by ionotropic gelation, as previously reported.20 Briefly, a TPP solution was added dropwise to a chitosan solution under magnetic stirring at room temperature. The solutions were filtered into prerinsed vials prior to mixing. In order to prepare particles at different TPP:chitosan ratios, the concentration of the TPP solution was varied. Preparation parameters such as stirring speed and batch size were kept constant. The particles were stored at 25 °C under minimal light exposure. Characterization of the Particles. The particle dispersions were characterized without any further purification or alteration, by dynamic light scattering (DLS), turbidity, and ζ potential measurements 1 day, 1 week, and 1 month after preparation. The pH of the nanoparticle suspensions was measured after 2 days, 1 week, and 1 month. Data obtained at day 1/2 are reported from Jonassen et al.20 ζ Potential. The ζ potential measurements were performed on a Malvern Zetasizer 3000HSA at 25 °C. The ζ potential, ζ, was calculated from the electrophoretic mobility, UE, through the Henry equation as follows:
moderate TPP-to-chitosan ratios. The results can be regarded as promising with respect to controlling both the size and the compactness of the particles. However, the particles’ ζ potential was reduced when prepared under saline conditions, imposing important questions on the particles’ long-term stability in suspension. In order to increase the storage stability, freeze-drying of the nanoparticles is often performed. However, polymer-based nanoparticles may be labile to the freeze-drying process, even in the presence of a cryoprotectant.21 For simple screening tests, the addition of a cryoprotectant may also undesirably complicate the system under investigation. These considerations, along with the general demand for less time-consuming preparation processes, highlight the need to evaluate the stability of nanoparticulate drug formulations in suspension. Particles in a colloidal dispersion may adhere to one another and form aggregates of successively increasing size, which may settle out under the influence of gravity.22,23 The colloidal stability of the particle suspension is thus affected by the ability to withstand flocculation or aggregation of the particles. Consequently, a nanoparticle dispersion may remain colloidally stable if the particles have repulsive forces strong enough to resist flocculation/aggregation. If the particles have little or no repulsive forces, then aggregation/flocculation or coagulation will eventually take place. The stability of polymer-based nanoparticles should also be evaluated with regards to structural changes of the particles, such as contraction and swelling. The degree of aggregation and contraction of the nanoparticles is expected to be influenced by any incorporated drug substances and their physicochemical properties. Furthermore, chemical processes such as oxidation and hydrolysis may affect the shelf life of polymer-based nanoparticle suspensions.24 The aim of this study was to investigate the physical stability of chitosan−TPP nanoparticles prepared and stored in solvents with different ionic strengths. The stability over a period of 1 month was investigated with special emphasis on changes in the average hydrodynamic radius and the particle compactness. The findings herein will be discussed by applying both classical and more recent theories on the behavior of nanoparticle suspensions. To our knowledge, only a few studies have been performed on the long-term stability of chitosan−TPP nanoparticles in suspension,12−15,24,25 and no studies quantitatively report or discuss any changes in the particle compactness over time.
■
UE =
2εζ f (Ka) 3η
(1)
where η and ε are the viscosity and the dielectric constant, respectively, of the solvent at the given temperature. The Smoluchowski approximation to Henry’s function ( f(Ka) = 1.5) was applied, and the viscosity and the dielectric constant of water were used (irrespective of the solvent salinity applied). The ζ potential was not measured on samples in which sedimentation was observed. pH. The pH of the particle suspensions was measured at room temperature using a Mettler Toledo 220 pH Meter. Dynamic Light Scattering. DLS measurements were performed at 25 °C with an ALV/CGS-8F Compact Goniometer System, with eight fiber-optical detection units (from ALV-GmbH, Langen, Germany). Further details of the measurements, instrumentation, and theory can be found elsewhere.20,26 The experimentally measured homodyne intensity autocorrelation function g2(q,t) in DLS measurements is directly linked to the theoretically amenable first-order electric field autocorrelation function g1(t) through the Siegert relationship, assuming that the scattering of the incoming light exhibits Gaussian statistics:26−28
g 2(q , t ) = 1 + B |g 1(t )|2
B≤1
(2)
where B is an instrumental parameter, t is the time, q is the wave vector defined as q = 4πn sin(θ/2)/λL, where λL is the wavelength of the incident light in vacuum, θ is the scattering angle, and n is the refractive index of the sample. In this study, n was measured with a PTR 46 refractometer from Index Instruments (U.K.). In the data analysis, g1(t) was fitted to a stretched exponential function:
EXPERIMENTAL SECTION
g 1(t ) = exp[− (t /τfe)β ]
Materials. Water-soluble chitosan chloride (PROTASAN UP CL 213) with a degree of deacetylation of 83% was supplied from Novamatrix, FMC Biopolymer, Norway. The weight average molecular weight (Mw) and number average molecular weight (Mn) were 3.07 × 105 and 1.15 × 105, respectively, and the polydispersity index (Mw/Mn) was 2.7.20 TPP (sodium triphosphate pentabasic, purum p.a. grade, ≥ 98.0%) was supplied from SigmaAldrich, and sodium chloride (for parenteral use) was supplied from Apotekproduksjon AS, Norway. The water used throughout the study was purified with a Millipore Milli-Q system. Experimental Design. Three factors affecting the characteristics of the nanoparticles were investigated: the chitosan concentration (0.05% and 0.10%), the TPP-to-chitosan ratio (5:95, 10:90, 15:85 and 20:80), and the ionic strength of the solvent (pure water, 0.05 M NaCl and 0.15 M NaCl), resulting in a total of 24 combinations. The chitosan concentrations given are the concentrations (w/w) in the final suspensions, and the TPP:chitosan ratios are given as weight-to-weight (w:w) ratios. Samples where sedimentation was observed immediately after particle preparation20 were excluded from further analysis.
0 0.997. The mean relaxation time (τf) is given by
τf =
τfe 1 Γ β β
(4)
where Γ(1/β) is the gamma function. The relaxation times were found to be diffusive (τf ∼ q−2), and the mutual diffusion coefficient is given by D = 1/τfq2. The apparent hydrodynamic radius, Rh, of the particles can thus be calculated using the Stokes−Einstein relationship: Rh =
kBT 6πηD
(5)
where kB is the Boltzmann constant, T is the absolute temperature, and η is the viscosity of the solvent at the given temperature. The viscosity of 3748
dx.doi.org/10.1021/bm301207a | Biomacromolecules 2012, 13, 3747−3756
Biomacromolecules
Article
water was applied irrespective of the ionic strength of the solvent employed in the preparation of the particles. Turbidity. The transmittance of the nanoparticle suspensions was measured at room temperature (ca. 25 °C) with a Helios Gamma (Thermo Spectronic, Cambridge, UK) spectrophotometer at a wavelength (λT) of 500 nm, as previously reported.20 The turbidity, τ, of the suspensions was calculated from the transmittance using the Lambert−Beer law: τ=−
1 ⎛ It ⎞ ln⎜ ⎟ L ⎝ I0 ⎠
(6)
where L is the light path length in the sample cell (1 cm), It is the intensity of the light transmitted through the sample, and I0 is the intensity of the light transmitted through the solvent (water). Particle Compactness. The local polymer concentration inside the nanoparticles, cNP, (i.e., the particle compactness) was calculated for particles prepared in the presence of NaCl by the following equation:7
τ=
⎞⎤ 3ct ⎡ 2 ⎛ 1 ⎢1 − (1 − cos(wc NP))⎟⎥ ⎜sin(wc NP) − ⎢ 2c NPR h ⎣ wc NP ⎝ wc NP ⎠⎥⎦
(7) where ct is the total polymer concentration in suspension, and w = [4πRh(dn/dc)]/λTn0, where n0 is the refractive index of the solvent, and dn/dc is the refractive index increment of the polymer. For the chitosan used in this study, dn/dc was previously determined to be 0.157 g/mL.7 The method is developed for spherical, monodisperse particles, and the total polymer concentration that exists in the form of nanoparticles must be known. In this study, the particles are assumed to be close to spherical, which is reasonable considering the q2-dependency of the relaxation times. The relatively high cross-linker concentrations applied, along with the presence of NaCl, promoting aggregation of the polymer chains, and the fact that the measured correlation functions in the DLS experiments exhibited only one relaxation mode, support the additional assumption that practically all chitosan is in the form of nanoparticles.7
Figure 1. The average ζ potential values as a function of NaCl concentration for combinations with (a) 0.05% chitosan and (b) 0.10% chitosan after 1 month of storage. The average ζ potential values after 1 day are given in the inset-plots.20 The samples in which sedimentation was observed are omitted from the plots. The points without error-bars have standard deviations equal to or smaller than the size of the symbols. Connection lines between the symbols are added in order to guide the eyes.
■
RESULTS As previously reported,20 particles were successfully prepared in almost all combinations. However, sedimentation was observed immediately for the particles prepared in pure water at 0.10% chitosan and high TPP:chitosan ratios (15:85 and 20:80). These combinations were therefore excluded from the stability studies reported in the present work. The reproducibility was checked for selected particle combinations (see Supporting Information for details). The relative standard deviations (RSDs) calculated were