Multilayered Nanoparticles for Controlled Release of Paclitaxel Formed by Near-Critical Micellization of Triblock Copolymers

Near-critical micellization (NCM), allowing for precise pressure-tuned control of sequential block collapse and micelle formation, can be synchronized with cancer-drug encapsulation with virtually no drug losses. NCM is demonstrated to produce benign, stable nanoparticles made of PEG-b-PLLA-b-PCL triblock copolymers that are not only solvent-free and paclitaxel-rich, which reduces the body exposure to the excipients, but also nearly burst-release-free, which reduces if not eliminates its toxic side effects while enhancing its therapeutic efficacy. ■ INTRODUCTION Drug carriers made of micelles of simple amphiphilic diblock copolymers, such as PEG-b-PLA micelles loaded with the hydrophobic cancer drug paclitaxel, are well-known and promising enough to undergo phase III clinical tests because, in general, they can be more benign than the powerful but nonspecific free drug itself. This way, a higher fraction of therapeutically productive drug ends up in the cancer tissue, as intended, instead of the healthy tissues, which alleviates the side effects relative to the free drug treatment. This therapeutically productive drug fraction could be even higher, and conversely, the counterproductive, toxic fraction could be even lower, if not for an excessive release of the initial drug fraction immediately following application. Such an undesirable “burst release”, a persistent fingerprint of most diblock micellar carriers, except for those that are plagued by low drug loading to begin with, significantly inhibits the probability of their clinical success. The root cause of burst release lies in the process of loading the drug into the micelles. Generally, both the drug and the block copolymer are dissolved in a water-miscible organic solvent, and then water, the selective antisolvent for the drug and hydrophobic block, is added to induce both drug nucleation and micelle formation. The popular but naively simplistic view is that both of these distinct phenomena occur simultaneously and the drug is encapsulated in the core of the micelles. However, this is not the case. A large fraction of the drug is actually adsorbed on the core surface, resulting in burst release. Subtle differences in the exact sequence of micellization and drug nucleation can prevent drug from reaching inside the core, which instead ends up trapped in the micelle corona on its way to the core, and hence have little resistance to be released prematurely. Regardless of its exact distribution within and around the core, and how it may affect the release rates, the pressing challenge is to find a robust and easily approvable approach to protecting the drug from burst release, recognized as one of the keys to increasing therapeutic efficacy of drug loaded micelles. There have been numerous known attempts to suppress burst release via new structures such as cross-linking of the micelle core or shell, conjugating drugs to the core, or even imparting an exotic protective layer designed to respond to external stimulus, such as pH or heat. However, such complicated modifications are hardly robust, not to mention a long and uncertain approval process they face. By contrast, the common blocks, such as PCL, PLA, and PEG, have all been approved for clinical use in other formulations, and in fact, PEG−PLA micelles loaded with paclitaxel are in phase III clinical trials (ClinicalTrials.gov), despite their serious burstrelease problems. Our goal, therefore, is to develop a translatable drug-loading process that can use these FDAapproved building blocks to fabricate drug-loaded micelles with minimized burst release and hence mitigate its side effects while enhancing its therapeutic efficacy. We previously developed a near-critical micellization (NCM) method and fabricated PEG−PCL micelles loaded with paclitaxel. The resulting micelles had much high drug loading but did little to reduce burst release. We thus further hypothesized that adding a protective layer on the drugcontaining core should reduce the burst-release probability. In principle, this can be accomplished with a suitable multiblock (at least triblock) copolymer via a precisely controlled sequential collapse of the blocks. Such a precise sequential Received: February 8, 2012 Revised: May 14, 2012 Published: May 24, 2012 Article pubs.acs.org/Macromolecules © 2012 American Chemical Society 4809 dx.doi.org/10.1021/ma300271k | Macromolecules 2012, 45, 4809−4817 block collapse, carefully synchronized with drug nucleation and encapsulation, is hard or impossible with the conventional liquid-solvent−antisolvent micellization method but it is attainable with the NCM method. This is because, instead of low-resolution, hard-to-control liquid solvents, the NCM method relies on compressed, near-critical gases that allow for precise, pressure-tuned control of each crucial structureforming stage separately, including micelle formation, drug nucleation, and encapsulation, then separately again, the protective layer formation upon the collapse of the middle block, as qualitatively illustrated in Figure 1, and, finally, corona collapse that triggers particle separation from the solvent. The particles formed in this manner are collected simply by rapid decompression of the solvent, which exists as a gas at ambient conditions and hence leaves no residual solvent traces. Our specific aim is to prove this hypothesis using ABC-type triblock copolymers with poly(ethylene glycol) (PEG) as the hydrophilic, corona-forming block on one end, poly(εcaprolactone) (PCL) as the hydrophobic, core-forming block on the other end, and a middle block that should form a protective “shell” around the core. For this auxiliary middle block, we select two models: poly(L-lactide) (PLLA), which is crystallizable (crystallinity around 37%, a glass transition temperature around 60−65 °C, a melting temperature between 173 and 178 °C), and poly(D,L-lactide) (PDLLA), which is amorphous. A structurally analogous reference diblock is PEG-b-(PDLLA-co-PCL), with a hydrophobic segment made of randomly distributed D,L-lactide and ε-caprolactone monomers. Our proof of concept will require characterizing all these micellar nanoparticles for drug loading content, drug encapsulation efficiency, overall drug loading efficiency, and drug release kinetics in water. ■ APPROACH Materials. Methoxypoly(ethylene glycol), ε-caprolactone, D,Llactide (3,6-dimethyl-1,4-dioxane-2,5-dione), and stannous octoate (Sn(Oct)2) were obtained from Sigma-Aldrich. Polycaprolactone and poly(D,L-lactide) were obtained from Polymer Source, Inc. Dimethyl ether and trifluoromethane were obtained from Airgas at 99.5% purity. Polymer Synthesis. PEG-b-(PDLLA-co-PCL) was synthesized in a manner consistent with the literature. Briefly, synthesis was accomplished by ring-opening polymerization of ε-caprolactone and D,L-lactide initiated by PEG-OH (5K) in the presence of stannous octoate (0.05 wt %) in a polymerization tube under nitrogen. The polymerization tube was placed in an oil bath at 160 °C for 3 h. The product was extracted using dichloromethane, precipitated with cold methanol, and then dried under vacuum for 48 h. PEG-b-PDLLA-b-PCL and PEG-b-PLLA-b-PCL were synthesized in two steps using a similar procedure. Polymer structures used in this work are summarized in Scheme 1. For PEG-b-PDLLA-b-PCL, first, PEG-b-PDLLA-OH was synthesized by ring-opening polymerization of D,L-lactide initiated by PEG-OH (5K) in the presence of Sn(Oct)2 catalyst in a polymerization tube under nitrogen and placed in an oil bath at 145 °C for 3 h. PEG-b-PDLLA was recovered by extracting with dichloromethane, precipitated with cold methanol, and dried under vacuum for 48 h to remove residual solvent. Second, PEG-bPDLLA-b-PCL was synthesized by ring-opening polymerization of εcaprolactone initiated by PEG-b-PDLLA-OH in the presence of Sn(Oct)2 catalyst in a polymerization tube under nitrogen and placed in an oil bath at 160 °C for 3 h. The product was extracted with dichloromethane, precipitated with cold methanol, and dried under vacuum for 48 h. Molecular weight and structure were determined by NMR, and molecular weight and polydispersity index (PDI) were confirmed by GPC. Cloud Point Measurements. The cloud point refers to the onset of a bulk transition of a binary solution from a homogeneous onephase region to a heterogeneous two-phase region. The cloud point transition for systems studied in this work can be induced either by decreasing temperature at constant pressure, which results in the cloud temperature, or by decreasing pressure at constant temperature, which results in the cloud pressure (CP). Upon increasing pressure or temperature beyond the cloud point boundary, the solution returns to its homogeneous one-phase state. The micellization pressure (MP) refers to the highest pressure at which micelles can be formed in a homogeneous solution upon decompression or, conversely, decomposed upon compression, at constant temperature. The nanosized micelle-containing phase is referred to as the micellar solution, in contrast to the molecular solution observed following micelle decomposition. The CP and MP transitions are measured in a small (about 1 cm in volume) high-pressure variable-volume cell coupled with transmittedand scattered-light intensity probes and with a borescope for visual observation of the phase transitions. This apparatus is equipped with a data acquisition and control systems that allow not only for constant temperature and pressure measurements but also for decreasing and increasing temperature and pressure measurements at a constant rate. The cloud points reported in this work are detected with a transmitted-light intensity probe. The micellization points are detected with a scattered-light intensity probe. A more detailed description of Figure 1. Cloud points in trifluoromethane of polycaprolactone (MW 2000), poly(L-lactide) (MW 1400), poly(ethylene glycol) (MW 5000), and poly(D,L-lactide) (MW 2500). The concentrations of the polymers are 1 wt %. Scheme 1. Structure of Multiblock Copolymers Used in This Work Note: 1) corresponds to the diblock PEG-b-(PDLLA-co-PCL) and 2) corresponds to the triblocks PEG-b-PDLLA-b-PCL and PEG-b-PLLAb-PCL, with the sole difference being that PDLLA contains both Dand L-isoforms of the monomer while PLLA contains only L-isoforms. Macromolecules Article dx.doi.org/10.1021/ma300271k | Macromolecules 2012, 45, 4809−4817 4810 the apparatus and of its transmittedand scattered-light intensity probes is given elsewhere. A known amount of the copolymer that will typically lead to a 1.0 wt % solution and solvent are loaded into the cell, which is then brought to and maintained at a desired pressure and temperature at which copolymer can be dissolved. Upon decompression, the bulk phase boundary (e.g., CP) is approached from the one-phase or micellar phase side, and the transmitted light intensity (TLI) starts decreasing. Conversely, upon compression, the phase boundary is approached from the two-phase side, and TLI starts increasing. A new data point is taken after equilibrating the mixture for 15 min in the one-phase region, well above the expected cloud temperature and pressure. In all cases, the TLI data are stored and analyzed as a function of time, temperature, and pressure. The cloud pressure in this work is taken as the inflection point on the TLI curve, which corresponds to a peak on its first derivative. Micelle formation is probed using high-pressure dynamic light scattering. The scattered light intensity and the hydrodynamic radius sharply increase on approaching the micellization pressure from the high-pressure side. For these measurements, we couple our highpressure equilibrium cell with an argon ion laser (National Laser) model 800BL operating at wavelength of 488 nm and a Brookhaven BI-9000AT correlator, as described previously. Nanoparticle Preparation. Aqueous drug-loaded micelle solutions are prepared by the following procedure. The polymer, drug, and selected solvent (trifluoromethane, dimethyl ether, or a mixture of the two) are loaded into the high-pressure cell. The solvent composition is chosen based on the relative phase behavior of the drug and the polymer. The polymer and drug are dissolved by setting the temperature and pressure well into the one-phase region, again determined from the phase diagrams of the polymer solution alone and the drug solution alone. Upon dissolution, the temperature is lowered to 35 °C at constant pressure, and the mixture is equilibrated for 15 min. The pressure is then lowered slowly (10 bar/min) to within 50 bar of the cloud point for the mixture to allow adequate time for the equilibration of the micellization process. From this point, the mixture is rapidly depressurized by releasing the pressurizing fluid (propane). The polymer is precipitated as the solvent rapidly evaporates. The solvent is then released slowly from the cell to prevent the loss of solids. The cell is washed with a known volume of distilled water into a flask and stirred. The volume of water is chosen to give a final concentration of 0.1 wt % polymer. Particle Size. Particle size measurements of the aqueous solutions are performed using dynamic light scattering after filtering the solution with a 0.2 μm PTFE filter. Critical Micelle Concentration. Critical micelle concentration (CMC) is measured by two methods. First, particle size of aqueous solutions is measured upon dilution at defined concentration intervals. The CMC is the point at which the particle size drops significantly, more than 30%. Once the CMC is approximated by this method, a second method is used for confirmation and higher resolution at low concentrations, as described previously in the literature. Briefly, aqueous micelle solutions are prepared with pyrene, which partitions between the micelle cores and solution. The partition coefficient can be determined by measuring the fluorescence spectra of the solution over a range of wavelengths and comparing the relative intensity of specific fluorescence intensity peaks. These peaks correspond to pyrene fluorescence in a hydrophobic microenvironment and a hydrophilic microenvironment. When the concentration falls below the CMC, the hydrophobic intensity peak, and hence the partition coefficient, changes abruptly. Drug Loading. Drug loading is characterized by the drug loading content (DLC), the drug encapsulation efficiency (DEE), and the drug loading efficiency (DLE) defined as follows: = ×