Method to Characterize Diffusion of Dye from Polymeric Partic

Aug 6, 2009 - Ketan Pancholi,* Eleanor Stride, and Mohan Edirisinghe. Department of Mechanical Engineering, University College London, Torrington Plac...
1 downloads 0 Views 3MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

In Vitro Method to Characterize Diffusion of Dye from Polymeric Particles: A Model for Drug Release Ketan Pancholi,* Eleanor Stride, and Mohan Edirisinghe Department of Mechanical Engineering, University College London, Torrington Place, WC1E 7JE United Kingdom Received February 26, 2009. Revised Manuscript Received May 19, 2009 The release profile of a drug delivery system is a key factor in determining its efficacy. In the case of a polymeric particle based system, the release profile is a function of several parameters including particle diameter and porosity. The effects of these parameters are usually investigated experimentally using UV-spectroscopy. Predicting the drug release profile from particles as a result of the interaction of many parameters is desirable in order to facilitate the design of more efficient drug delivery particles. In this work, a quantitative method of determining the diffusion profile is developed which removes the need for repetitive experimentation. Particles of polymethylsilsesquioxane were prepared using coaxial electrohydrodynamic atomization and collected in solutions containing different concentrations of Evans blue dye (6, 0.6, and 0.06 mg/mL) which was used to simulate a drug. The dye release profile was calculated by solving the unsteady state diffusion equation for parameters used in the experiments. It was demonstrated that the dye release profile from particles with diameters ranging from 400 nm to 9 μm can be calculated using a simple equation without addition of a dissolution term, if the volume ratio of surrounding liquid to particle in the unsteady second order solution is substituted by the surface area of particles to liquid volume ratio. The calculated data are found to be in good agreement with the experimental, indicating that this method can be used to determine the diffusion coefficient as a function of particle diameter and material. This study represents a crucial step toward developing a full drug release model.

Introduction In recent years, a range of novel drug delivery carriers such as microbubbles,1 liposomes,2 and solid and liquid nanoparticles3 have been the subject of intense research. Particle based drug delivery systems offer many advantages compared to liposomes and micelles4 including the capacity to incorporate both small and large molecules, hydrophobic and hydrophilic drugs, enhanced stability, and enhanced drug carrying capacity. Nanoparticle systems can contain therapeutic entities, such as small molecule drugs, proteins, and nucleic acids assembled with particles produced from lipids and polymers. These targeted nanoparticles can have enhanced anticancer effects compared to free drug because of specific targeting to tumor tissues and release of the drug locally, which improves the pharmacokinetics and pharmacodynamics, and intracellular delivery.5 The controlled release of the drug through nanoparticle design would minimize the side effects of anticancer drugs while enhancing efficacy in terms of sustained drug release from its matrix.6 Although nanoparticle based liposomal drug delivery systems, such as DaunoXome and Abraxane,7 have achieved encouraging results by solubilizing the drug for higher uptake and better biodistribution at the tumor site, they do not provide control over *Corresponding author. Present address: School of Engineering, Robert Gordon University, Schoolhill, Aberdeen, AB10 1FR Scotland. (1) Pancholi, K.; Stride, E.; Edirisinghe, M. Langmuir 2008, 24, 4388–4393. (2) Huang, S. Adv. Drug Delivery Rev. 2008, 60(10), 1167–1176. (3) Morimoto, N.; Winnik, F. M.; Akiyoshi, K. Langmuir 2007, 23(1), 217–223. (4) Hagan, S. A.; Coombes, A. G. A.; Garnett, M. C.; Dunn, S. E.; Davies, M. C.; Illum, L.; Davis, S. S.; Harding, S. E.; Purkiss, S.; Gellert, P. R. Langmuir 1996, 12(9), 2153–2161. (5) Davis, M. E.; Chen, Z.; Shin, D. M. Nat. Rev. Drug Discovery 2008, 7, 771– 782. (6) Pommier, Y. Curr. Med. Chem.: Anti-Cancer. Agens. 2004, 4, 429–434. (7) Zamboni, W. C. Clin. Cancer. Res. 2005, 11, 8230–8234.

Langmuir 2009, 25(17), 10007–10013

the release rate of the drug.7 The release of the drug from particles at a desired rate increases drug bioavailability and reduces the required dosing frequency. However, the rate at which drug is released from a particle depends on several parameters including the polymer diffusion coefficient, the particle diameter, porosity, the microenvironment surrounding the particle, the binding strength between the drug and polymer, the crystallinity of the polymer, the volume of liquid surrounding the particle, and the initial concentration of the drug inside the particles. Polymerdrug conjugation is one of the strategies used for slower release of hydrophilic drugs.8 Similarly, there are many possible combinations of the various parameters affecting the drug release profile, which will yield different results. It is therefore necessary to test a very wide range of parametric combinations in order to optimize the release profile. The most frequently used method for measuring the drug release profile is to disperse the drug loaded particles in aqueous solutions and then remove small samples of the liquid at constant time intervals in order to measure concentration of the drug using UV-spectroscopy. Plotting the concentration of the drug in the liquid at constant time intervals provides the drug release profile. However, in order to determine the drug release profile for a large number of parametric combinations, it is desirable to use a theoretical method. Mathematical models currently available in the literature include the Higuchi and the Peppas and Harland models.9,10 These models are based on the second order Fick’s equation and can accurately predict the dissolution and/or diffusion of a solid drug embedded in a solid matrix. To address the problem of modeling a rapid initial quick release (burst) of a drug, Harland (8) Sinha, V.; Trehan, A. Crit. Rev. Ther. Drug Carrier Syst. 2005, 22(6), 535– 602. (9) Higuchi, T. J. Pharm. Sci. 1963, 52, 1145–1149. (10) Peppas, N. A. Pharm. Acta Helv. 1985, 60, 110–111.

Published on Web 08/06/2009

DOI: 10.1021/la900694k

10007

Article

Pancholi et al.

Figure 1. Basic experimental setup.

added a dissolution term to the diffusion equation.11 However, this was only suitable for modeling the release from a single particle with a large diameter (>1 mm). Drug release from nanosized particles can be obtained using the heat conduction equation by assuming that the particle has negligible mass.12 In practice, however, there will be many particles present in the liquid contributing to the total particle mass comparable to the liquid volume.12 Moreover, the model does not work equally well with hydrophilic and hydrophobic drugs, and the form of the additional dissolution term in the equation is largely determined through trial and error rather than from the governing physical relationship. This prevents the solution from being applied to all drug types (hydrophilic and hydrophobic) and ranges of particle diameter.13 This work is aimed at removing the need for a trial and error method in selecting the model appropriate for the actual drug release and making the determination more predictive.

Materials and Methods Preparation of Solutions. To prepare the polymer solution, 72 g of ethanol (VWR international Ltd., Poole, U.K.) was stirred in a beaker for at least 120 s using a magnetic stirrer. Subsequently, 18 g of polymethylsilsesquioxane powder (Grade MK, Wacker Chemie, Burghausen, Germany) was gradually added to the ethanol until it was completely dissolved. The solution was then stored in an airtight container maintained at 5 C. Dye solutions of three different concentrations were prepared by adding 0.005, 0.01, and 0.1 g of Evans blue dye to 150 mL of millipure water to give concentrations of 0.033, 0.066, and 0.66 mg/mL, respectively. Particle Preparation. The experimental setup shown in Figure 1 consisted of two concentric needles. The inner needle had 150 μm inner diameter (ID) and 300 μm outer diameter (OD). The outer needle dimensions were 685 μm ID and 1100 μm OD. Both needles were connected to the same applied voltage source, and an earthed ring electrode (ID 15 mm and OD 20 mm) was placed 12 mm below the tip of the outer needle. The inner needle exit was kept ∼2 mm above the exit of the outer needle. Both of the needles were fixed inside a brass housing, and their outlets were connected to syringes supplying the fluids. The inner needle was connected to a Harvard syringe pump PHD-4400 (Harvard apparatus, Kent, U.K.) supplying air, whereas the outer needle was connected to a similar pump supplying the liquid. A vial (11) Harland, R. S.; Dubernet, C.; Benoit, J.-P.; Peppas, N. A. J. Controlled Release 1988, 7, 207–215. (12) Ng, E. Y. K.; Ng, W. K.; Chiam, S. S. J. Med. Syst. 2008, 32, 85–92. (13) Arifin, D. Y.; Lee, L. Y.; Wang, C. H. Adv. Drug. Delivery Rev. 2006, 58, 1274–1325.

10008 DOI: 10.1021/la900694k

containing a dye solution of known concentration as above was placed just beneath the ring electrode to collect the polymer droplets. The flow exiting the needles was monitored on a screen using a LEICA S6D JVC-color video camera attached to a zoom lens and a data DVD video recorder MP-600 using CDV Recorder/Editor DN-100. To obtain particles, the polymer solution and air were supplied simultaneously to the outer and inner needles, respectively, at a constant flow rate of 300 μL/min. As soon as coarse microbubbles started forming at the tips of the needles, the voltage across the needle tips and ground electrode was increased gradually up to 5.7 kV when disintegration of the composite liquid-air jet into droplets was observed. A total of 16 mL of polymer solution was electrosprayed into 18 mL of dye solution. The particles of different sizes (1, 3, 4, 7, 8 μm and 400 nm) were separated by a filtering process to retain similar sized particles only. The primary aim was to create hollow particles. However, the particles observed were not hollow, but coaxial jetting with air contributes to an increase in porosity of the particles.14 The droplets polymerized as they were collected in the dye solutions and were kept suspended in the dye solutions for 24 h to allow dye to diffuse into the particles. After 24 h, the particles were filtered and dried in a desiccator. The dried particles were then washed briefly using distilled water to remove any excess dye on the surface of the particles and again dried in the desiccator. Particle Size Measurement. Before measuring the particle size distribution, the zetasizer (Malvern Nano ZS, Malvern Instruments Ltd., Worcestershire, U.K.) was calibrated using latex particles of 0.5 and 1 μm diameter. To prepare samples for zetasizing, 0.5 mg of particles was dispersed in 25 mL of distilled water and stirred continuously before extracting approximately 1 mL of this suspension to load into the zetasizer. Ten samples from different experiments were used to measure the particle size, and all zetasizing experiments were performed at 15 C. Three reruns were performed on each sample to ensure data repeatability. To confirm the size distribution of particles obtained using the zetasizer, ∼0.1 mL of the particle suspension was extracted using a 1 mL syringe and placed on a glass slide to observe at least 100 particles under a Nikon Eclipse ME 600 microscope fitted with a CCD camera JVC KY-F55B. Micrographs of the particles were captured by the program AcQuis from Synoptics (version 4.0.1.3). These were used to measure the diameter of the particles in the micrometer range; the diameter of 400 nm particles was estimated using a JEOL JSM-6301F scanning electron microscope. Particle Density. To estimate the density of the particles, helium picnometry (Micromeretics AccuPyc 1330, Norcross, GA) was used. The 1 mL cuvette was filled with dry particles to obtain (14) Farook, U.; Stride, E.; Edirisinghe, M. J. J. R. Soc. Interface 2009, 6(32), 271–277.

Langmuir 2009, 25(17), 10007–10013

Pancholi et al.

Article

Figure 2. Typical graph obtained using the Malvern zetasizer showing the size distribution of particles. The diagram shows that diameters of most particles in this sample were around 1 μm. dense packing. Then the cuvette was infused with liquid helium to obtain the volume and weight of the particles every 60 s. Ten observations were made to find that the values of volume and weight were within (0.2% of the average values. Assuming a packing density of 0.7-0.5 in high shear limit for particles of R;

CL ¼ 0

ðBoundary condition 2Þ

DCd ¼0 Dr

ðBoundary condition 3Þ

 DC L DC d  ¼ K p As D e VL r ¼R Dt Dr 

ðBoundary condition 4Þ

t > 0; t > 0;

r ¼ 0; r ¼ R;

where VL, CL, Cd, Kp, R, and As are the suspending liquid volume, transiently varying dye concentration in the liquid, spatially DOI: 10.1021/la900694k

10009

Article

Pancholi et al.

varying dye concentration on the surface of the particle, partition coefficient of the Evans blue dye, radius of the particle, and area of the particle, respectively. Boundary condition 4 shows that the amount of dye entering the particles is equal to the amount of dye entering the liquid volume VL. With these boundary conditions, the solution of eq 1 can be obtained as16 ¥ X M tL 6σð1 þ σÞ De x n 2 t exp ¥ ¼ 12 2 ML R2 n ¼1 9 þ 9σ þ σ X n

! ð2Þ

where MtL/M¥ L is the ratio of the mass of the dye in the particle after time (t) and that in a finite amount of the liquid surrounding it. Effect of Size of the Particles on Diffusion. Generally, particles 50% of the total volume gave the same release profile. Below the value of such porosity, the dye release profile did not match with the experimental result (