Article pubs.acs.org/molecularpharmaceutics
Solid Nanocrystalline Dispersions of Ziprasidone with Enhanced Bioavailability in the Fasted State Avinash G. Thombre,*,† W. Brett Caldwell,‡ Dwayne T. Friesen,‡ Scott B. McCray,‡ and Steven C. Sutton†,§ †
Pfizer Inc., Center for Discovery and Development Sciences, Groton, Connecticut 06340, United States Bend Research Inc., Bend, Oregon 97701, United States
‡
ABSTRACT: Reducing the absorption difference between fed and fasted states is an important goal in the development of pharmaceutical dosage forms. The goal of this work was to develop and characterize a solid nanocrystalline dispersion (SNCD) to improve the oral absorption of ziprasidone in the fasted state, thereby reducing the food effect observed for the commercial formulation. A solution of ziprasidone hydrochloride and the polymer hydroxypropyl methylcellulose acetate succinate (HPMCAS) was spray-dried to form a solid amorphous spray-dried dispersion (SDD), which was then exposed to a controlled temperature and relative humidity (RH) to yield the ziprasidone SNCD. The SNCD was characterized using powder X-ray diffraction, thermal analysis, microscopy, and in vitro dissolution testing. These tools indicate the SNCD consists of a high-energy crystalline form of ziprasidone in domains approximately 100 nm in diameter but with crystal grain sizes on the order of 20 nm. The SNCD was dosed orally in capsules to beagle dogs. Pharmacokinetic studies showed complete fasted-state absorption of ziprasidone, achieving the desired improvement in the fed/fasted ratio. KEYWORDS: ziprasidone, solubilization, oral bioavailability, food effect
■
INTRODUCTION Ziprasidone (5-{2-[4-(1,2-benzisothiazol-3-yl)-1-piperazinyl]ethyl}-6-chloro-1,3-dihydro-2H-indol-2-one, Scheme 1), a
fasted state cannot be compensated for simply by increasing the prescribed dose.9,10 The variable absorption of ziprasidone can be attributed to its physicochemical characteristics. Ziprasidone is poorly watersoluble (free-base solubility in pH 6.5 buffered media ∼0.3 μg/ mL), basic (pKa ∼6), and highly lipophilic (c log P = 3.6).11 Ziprasidone dissolution is greatly enhanced when the drug partitions into bile-salt micelles; we measured the neutral form of the drug to have a micelle-to-aqueous partition coefficient of approximately 4000. These physicochemical characteristics of ziprasidone result in poor and variable oral absorption in the fasted state where bile concentrations are relatively low.6 Numerous solubilization techniques have been developed to enhance the absorption of low-solubility compounds such as ziprasidone. These technologies include solid amorphous dispersions made by spray drying or melt extrusion,12−17 nanocrystals,18−20 prodrugs,21 lipid-based formulations22 including self-emulsifying drug delivery systems (SEDDS),23 and complexation using cyclodextrins.24 Solid amorphous dispersions represent an attractive formulation method for compounds with low solubility and high lipophilicity because they can be stable, solid-state, high-energy drug forms.12−17 Use of high-energy drug forms significantly
Scheme 1. Ziprasidone
dopamine (D2) receptor antagonist, is an orally active atypical antipsychotic drug used in the treatment of schizophrenia and bipolar disorder.1−3 It is commercially marketed as Geodon in the U.S. and as Zeldox in other parts of the world. When commercial Geodon capsules are dosed, a large fed/ fasted ratio is observed: ziprasidone absorption is increased up to 2-fold in fed versus fasted patients.4,5 As a result, administration of Geodon with food is considered crucial to ensure optimal, reliable, dose-dependent bioavailability and achieve predictable symptom control and patient toleration.6 Studies have shown two factors are important in ziprasidone absorption: (1) the number of calories consumed (it should exceed 500 kcal, irrespective of the fat content) and (2) the amount of time between food intake and dosing (should be less than 2 h).7,8 Reduced oral absorption of ziprasidone in the © 2012 American Chemical Society
Received: Revised: Accepted: Published: 3526
June 29, 2012 October 5, 2012 October 16, 2012 October 16, 2012 dx.doi.org/10.1021/mp3003607 | Mol. Pharmaceutics 2012, 9, 3526−3534
Molecular Pharmaceutics
Article
Figure 1. General scheme for formation of a solid nanocrystalline dispersion (SNCD) from a spray-dried dispersion (SDD) using controlled temperature and humidity.
process schematically shown in Figure 1. In this paper, an amorphous ziprasidone SDD is formed through a conventional spray-drying process, followed by crystallization of the drug using controlled temperature and relative humidity (RH). The dispersion loading and crystallization conditions are selected to strongly favor nucleation over crystal growth. When water is used as a plasticizer for dispersions of drugs with low aqueous solubility, the solubility of drug in the polymer matrix is often lower than the solubility of the drug in the dry polymer, further promoting rapid nucleation rates. As a result, ziprasidone SNCDs consist of extremely small crystalline domains, approximately 20 to 100 nm in diameter, embedded in a polymer-rich matrix. Significantly, these crystals are a higherenergy drug form than the lowest-energy crystalline polymorph as evidenced by melting point shifts and smaller crystal size. As a result, ziprasidone SNCDs provide not only rapid dissolution but also enhanced drug concentrations relative to bulk crystalline drug. Although a variety of matrix polymers may be used to form the SNCDs (e.g., hydroxypropyl methylcellulose [HPMC] or polyvinyl pyrrolidone [PVP]),29 the results presented in this paper for ziprasidone have been achieved using HPMCAS as the polymeric matrix for the initial SDD and resulting SNCD. Importantly for increasing fasted-state absorption, the colloidal nature of HPMCAS in media at intestinal pH results in concentration enhancement as well as sustainment that is unusual among dispersion polymers. In addition, the high Tg and relatively low water uptake of HPMCAS result in low drug mobility at the storage conditions employed for stability testing. These polymer properties prevent growth of the high-energy crystals of the SNCD at normal storage conditions. Although the data are not shown in this paper, ziprasidone:HPMCAS SNCDs have shown excellent physical stability (due to the high-Tg, low mobility polymer matrix) and excellent chemical stability (due to the crystalline drug form). This paper discusses the formation of ziprasidone SNCDs from amorphous SDDs and physical characterization using powder X-ray diffraction (PXRD), modulated differential scanning calorimetry (mDSC), and microscopy. The ziprasidone SNCDs enhanced dissolution rates and free-drug concentrations in in vitro performance tests, and these improvements correlated with enhanced fasted-state absorption and elimination of the fed/fasted effect when capsules containing the ziprasidone SNCDs were dosed orally to beagle dogs.
enhances drug dissolution rate and solubility, and the stable, solid-state nature of dispersions facilitates incorporation in a variety of traditional dosage forms. When the drug is in a highenergy form and a concentration-enhancing polymer such as hydroxypropyl methylcellulose acetate succinate (HPMCAS) is used as the dispersion matrix, the increases in drug dissolution rate and solubility can be maintained for times relevant to absorption (e.g., up to several hours).12,16,25,26 However, not all solid amorphous dispersions are stable at normal storage conditions. For example, when the drug loading in the dispersion is increased above the solubility of the amorphous drug in the dispersion polymer, the drug may phase separate from the polymer and crystallize (i.e., as the thermodynamic driving force for crystallization increases). This rate of conversion to the crystalline state generally increases as drug mobility increases. Drug mobility in an amorphous dispersion increases exponentially as the ratio of the storage temperature (in kelvins) to the glass-transition temperature (Tg) of the dispersion at storage conditions increases. Thus, in the common situation in which the Tg of the drug is less than that of the polymer, increasing the drug loading in a dispersion will decrease the Tg of the dispersion, thereby increasing the drug mobility and rate of crystallization at a given condition of temperature and humidity.12 Therefore, to avoid potential changes in physical state and performance as a result of controlled temperature and humidity storage, formulators limit the drug loading of amorphous dispersions. In the case of ziprasidone, the active dispersion loading must be maintained at 10 wt % to maintain good physical stability.12 Alternatively, storage conditions are chosen or controlled through packaging to maintain a high dispersion Tg and subsequent low drug mobility within the dispersion (i.e., storage at low relative humidity [RH] and/or low temperature). Another technique used to increase the dissolution rate and subsequent absorption of lipophilic compounds involves reducing crystal size by “top-down” attrition milling of a lowenergy crystal polymorph, usually in the presence of excipients.18−20 This technique yields crystals of the ingoing polymorph (usually the lowest-energy polymorph) with diameters of approximately 100 to 400 nm. The increased drug surface area achieved through attrition milling increases dissolution rates but usually does not significantly increase solubility unless the form of the drug changes. By enhancing the dissolution rate, attrition milling has been shown to increase oral absorption of some low-solubility drugs.27,28 Solid nanocrystalline dispersions (SNCDs)the focus of this workprovide an alternative technique for the improvement of oral absorption of drugs in the fasted state. In contrast to attrition milling, SNCDs are formed by the “bottom-up”
■
EXPERIMENTAL SECTION Materials. Ziprasidone hydrochloride (HCl) and other chromatography standards were obtained from Pfizer Inc. (Groton, CT). The HG grade of HPMCAS (which is also
3527
dx.doi.org/10.1021/mp3003607 | Mol. Pharmaceutics 2012, 9, 3526−3534
Molecular Pharmaceutics
Article
orientation effects. The X-ray source (Cu Kα, λ = 1.54 Å) was operated at 45 kV and 40 mA. Data for each sample were collected from 4° to 40° on the 2Θ scale over 30 min in continuous-detector scan mode at a scan speed of 2 s/step and a step size of 0.04°/step. In Vitro Membrane-Permeation Test. In the membranepermeation test, the drug formulation was evaluated in two different feed solutions (based on two different intestinal buffer solutions, which are described below). Each feed solution was placed on the feed side of a microporous membrane that had been treated with an argon plasma to render the feed-side surface hydrophilic. A water-immiscible permeate solution (a 60:40 mixture of decanol and decane) was placed on the permeate side of the membrane, filling the pores of the microporous membrane. Both the feed and permeate solutions were stirred during the test, and the apparatus was placed into a 37 °C, temperature-controlled box. Small aliquots (i.e., 50 μL) of the permeate solution were collected periodically. The concentration of drug in the permeate solution was analyzed by high-performance liquid chromatography (HPLC) with ultraviolet−visible (UV−vis) spectroscopy as described below. From these data, the mass and concentration of drug in the permeate were determined over time. Three feed solutions were used: (1) simulated gastric medium (pH 2.0, 0.1 N HCl), (2) model fasted duodenal (MFD) solution, and (3) 2% MFD solution. The “MFD” and “2% MFD” solutions were made at 2-fold higher concentrations and slightly more basic (pH 6.6) so that after 30 min in the simulated gastric solution the concentrates could be added 1:1 to the gastric solution to yield solutions with the following compositions. Both solutions were based on phosphate buffered saline (PBS), which consisted of 20 mM sodium phosphate (Na 2 HPO 4 ), 47 mM potassium phosphate (KH2PO4), 87 mM sodium chloride (NaCl), and 0.2 mM potassium chloride (KCl), adjusted to pH 6.5 with sodium hydroxide (NaOH), with an osmolarity of 290 mOsm/kg. MFD solution consisted of PBS containing 7.3 mM sodium taurocholic acid (NaTC) and 1.4 mM 1-palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC). With MFD solution, which was used to mimic fasted-state intestinal media, the taurocholate and phosphocholine compounds NaTC and POPC form mixed micelles and mimic those present in the gastrointestinal (GI) tract, solubilizing hydrophobic compounds. The 2% MFD solution, which was used to mimic fed-state intestinal media, consisted of PBS containing 29.2 mM NaTC and 5.6 mM POPC (4-fold that of MFD). The membrane-permeation test was performed by placing the SNCD formulation in simulated gastric medium at a theoretical concentration of 150 μgA/mL for 30 min, followed by addition of an equal amount of the concentrates to yield either the MFD solution or 2% MFD solution to yield a theoretical concentration of 75 μgA/mL. Ziprasidone was assayed in vitro using a normal-phase HPLC method (Agilent 1100, Agilent Technologies, Inc., Santa Clara, CA) using a Zorbax SB-CN column (4.6 mm × 150 mm, 5 μm) and a 1.2 mL/min flow rate, a 50 μm injection volume, a 25 °C column temperature, 230 nm UV detection, and a 2.6 min peakintegration retention time. The diluent was isopropyl alcohol (IPA), and the mobile phase was 16.3/83.7 (60/40 decanol/ decane)/IPA. This HPLC assay had a linear peak-area response from 0.1 to 17 μg/mL. In Vivo Dog Tests. The in vivo pharmacokinetic studies described here were reviewed and approved by the Pfizer
known as hypromellose acetate succinate, AQOAT) was purchased from Shin-Etsu Chemical Co. Ltd. (Tokyo, Japan). All other chemicals, reagents, and solvents were analytical grade and were purchased from commercial suppliers. Methods. SDD Preparation and SNCD Formation. A typical lot of the ziprasidone SDD and SNCD was prepared as follows. Ziprasidone HCl (same active polymorph as commercial Geodon capsules) and HPMCAS were dissolved at a 1:3 (w/w) ratio in methanol at ∼1 wt % solids. This solution was directed at ambient temperature to a pressure spray nozzle at 110 psi and a feed rate of 29 g/min into the stainless-steel chamber of a GEA-Niro PSD-1 spray dryer (GEA Process Engineering Inc., Columbia, MD). The inlet temperature of the nitrogen drying gas was maintained at 120 °C; the drying gas and evaporated solvent exited the dryer at 75 °C. The SDD was collected using a cyclone. Residual methanol was removed by spreading the SDD onto polyethylene-lined trays and drying it in a solvent-ready Gruenberg tray-dryer at 40 °C/ 15% RH for at least 8 h. The ziprasidone SNCD was formed by exposing the ziprasidone SDD to the desired temperature and RH inside humidified tray-drying chambers, as described in the Results and Discussion section. DSC Analysis. Modulated DSC analyses were performed using a Thermal Analysis Q1000 differential scanning calorimeter equipped with an autosampler (TA Instruments, New Castle, DE). Sample pans were equilibrated overnight at the desired RH, crimped and sealed, and then loaded into the differential scanning calorimeter. The samples were heated by modulating the temperature (e.g., at 1.5 °C/min) while increasing the temperature at a rate of 2.5 °C/min. Particle Size. Particle sizes were determined using a Mastersizer Scirocco 2000 powder-jet dry feeder (Malvern Instruments Ltd., Malvern, Worcestershire, U.K.). Specific Volume. The bulk and tapped density of the SDD and SNCD were determined by USP ⟨616⟩ method 1. Scanning Electron Microscopy (SEM). SEM was used to obtain particle size and morphology information for the SDD and SNCD using a Hitachi S-3400N microscope (Hitachi High-Tech Americas Inc., Pleasanton, CA). Samples were sputter-coated with gold/palladium to provide a conductive coating. Cryogenic Transmission Electron Microscopy (TEM). A two-part epoxy was prepared (Loctite 5 min epoxy, Henkel Corporation, Rocky Hill, CT). An uncured drop of viscous epoxy was transferred to the end of an aluminum mounting pin (3 mm diameter). Approximately 1 mg of sample powder was sprinkled onto the bead of uncured epoxy. Surface forces pulled particles into the epoxy. The sample was allowed to cure at ambient conditions for at least 24 h. The aluminum sample pin was mounted into an ultramicrotome from RMC Products (Boeckeler Instruments, Tucson, AZ), and
