European Journal of Pharmaceutical Sciences 41 (2010) 71–77 Contents lists available at ScienceDirect European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps Predicting the in vivo release from a liposomal formulation by IVIVC and non-invasive positron emission tomography imaging Eva Hühn a, Hans-Georg Buchholz b, Gamal Shazly a, Stephan Maus b, Oliver Thews c, Nicole Bausbacher b, Frank Rösch d, Mathias Schreckenberger b,1, Peter Langguth a,∗,1 a Institute of Pharmacy and Biochemistry, Johannes Gutenberg University, Staudinger Weg 5, 55099 Mainz, Germany b Department of Nuclear Medicine, University Medical Center, Mainz, Germany c Institute of Physiology and Pathophysiology, University Medical Center, Mainz, Germany d Institute of Nuclear Chemistry, Johannes Gutenberg University, Mainz, Germany article info abstract Article history: This study aimed to predict the in vivo performance from the in vitro release of a low-molecular weight Received 19 March 2010 model compound, [18F]-2-fluoro-2-deoxy-d-glucose ([18F]FDG), from liposomes and by means of positron Received in revised form 12 May 2010 emission tomography (PET). Liposomes composed of hydrogenated phosphatidylcholine (HPC) were pre- Accepted 30 May 2010 pared by a freeze–thaw method. Particle size distribution was measured by dynamic light scattering (DLS). Available online 8 June 2010 In vitro release was examined with a dispersion method detecting the radioactivity of [18F]FDG. In vivo release of [18F]FDG, following i.p. injection of the liposomes in rats, was determined by using a Micro-PET Keywords: scanner. Convolution was performed to predict the in vivo profiles from the in vitro data and to establish Convolution In vitro–in vivo correlation (IVIVC) an in vitro–in vivo correlation (IVIVC). The in vivo predictions slightly underestimated the experimentally Liposome determined values. The magnitude of the prediction errors (13% and 19%) displayed a satisfactory IVIV Positron emission tomography (PET) relationship leaving yet room for further improvement. This study demonstrated for the first time the use Release of PET in attaining an IVIVC for a parenterally administered modified release dosage form. It is therefore Dissolution possible to predict target tissue concentrations, e.g., in the brain, from in vitro release experiments. IVIVC using non-invasive PET imaging could thus be a valuable tool in drug formulation development, resulting in reduced animal testing. © 2010 Elsevier B.V. All rights reserved. 1. Introduction The rate of release of an agent from liposomes is governed by several factors (Drummond et al., 2008): the lipid composition, In the last 15 years liposomes have frequently been used as drug the active’s physico-chemical properties and its drug–lipid interac- carriers. They may reduce toxicity and deliver the drug to its sites of tions and the liposome size. The lipid type determines the stability action. Other advantages include reduced dosage, sustained drug and fluidity and, consequently, the permeability of the liposomal release and, consequently, reduced frequency of administration. membrane. An important factor with significant impact is the rate of release of The principal mechanisms of drug release are diffusion and the drug from the carrier, since only the free fraction of the drug in vivo disintegration of the liposomal membrane. The latter extends the therapeutic effect. could lead to the immediate release of most of the entrapped In the pharmaceutical industry, in vitro release tests are an drug from a disrupted liposome. Both mechanisms can appear essential tool in (i) taking indirect measurements of drug avail- simultaneously. ability, especially in preliminary stages of product development, In contrast to oral modified release formulations today exist (ii) assuring stable release characteristics of the product over time, no regulatory standards for dissolution tests for colloidal systems. (iii) assessing the effect of manufacturing process changes on prod- Due to the considerable differences in formulation and hence in uct performance and (iv) quality control and assurance to support physico-chemical and release characteristics, it is hard to estab- batch release. lish a generally applicable method. Thus, different techniques are applied. The most widely used methods for determination of release profiles from disperse systems can be divided into four ∗ groups: membrane diffusion, sample and separate techniques, in Corresponding author. Tel.: +49 6131 39 25746; fax: +49 6131 39 25021. situ and continuous flow methods. All these methods have been E-mail addresses: [email protected], [email protected] (P. Langguth). comprehensively reviewed by Washington (1990) and represent 1 Both authors contributed equally to this work. their advantages and disadvantages. 0928-0987/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2010.05.020 72 E. Hühn et al. / European Journal of Pharmaceutical Sciences 41 (2010) 71–77 In vivo release is mainly determined by measuring the concen- and the gel was compressed via centrifugation. After washing five tration of free drug in blood, plasma and/or serum. It is also possible times with isotonic saline solution the MLV (300 ␮l) were loaded to ascertain in vivo release with imaging methods such as mag- on the top of the dehydrated gel and centrifuged for exactly 1 min netic resonance imaging (MRI) or positron emission tomography at 3200 rpm. The elute containing the final liposomes was used for (PET). PET is an imaging technique with a high temporal and spatial the following experiments. resolution that produces three-dimensional images of the human or animal body using short-lived radioactive substances. One of 2.2.2. Preparation of [18F]FDG-containing extruded MLV (eMLV) the ultimate goals of biopharmaceutical research is to establish a MLV were prepared as described above by the freeze–thaw suitable approach to describe the relationship between an in vitro method except for taking 70 mg instead of 100 mg Ph90H. There- property of a dosage form (e.g., the rate or extent of drug release) after, the MLV (still containing free [18F]FDG) were extruded and a relevant in vivo response (e.g., plasma drug concentration or through a polycarbonate membrane (pore diameter 100 nm) 11 amount of drug absorbed). Guidances both by the Food and Drug times using a LiposoFast-Basic extruder at 68 ◦C. The resulting Administration (FDA) and the European Medicines Agency (EMA) eMLV were cleared from the free [18F]FDG by cGPC. deal with the development and role of in vitro–in vivo correlations (IVIVC) in oral solid dosage forms (EMEA, 1999; FDA, 1997). As 2.2.3. Particle size modified release parenteral dosage forms become more and more Size distribution was determined by dynamic light scattering important, recent developments investigate the role of IVIVC for (DLS). these dosage forms as well (Burgess et al., 2004; Martinez et al., 2008). 18 The aim of this study is to establish a quantitative relationship 2.3. In vitro release of [ F]FDG between the in vitro and in vivo release-rate of an active from such a liposomal dosage form with the aid of PET. 2.3.1. Dispersion method 1 ml liposome-dispersion was diluted to 2 ml with isotonic saline solution in a 2.2 ml MicroCentrifuge Tube. The tubes were 2. Materials and methods rotated using a rotator at 25 rpm at 37 ± 0.5 ◦C. At 15, 30, 45, 60, 75 and 90 min, 0.2 ml of the dispersion were withdrawn and 2.1. Materials replaced with 0.2 ml of fresh isotonic saline solution. The sample was separated from the free [18F]FDG by cGPC. The radioactivity ® Phospholipon 90 H (Ph90H) was a generous gift by Phopholipid was measured before (act1) and after (act2) the cGPC by a VDC-405 GmbH (Cologne, Germany). Isotonic saline solution was pur- dose calibrator. chased from Fresenius Kabi Deutschland GmbH (Bad Homburg, Germany). [18F]-2-Fluoro-2-deoxy-d-glucose ([18F]FDG) was pro- 2.4. In vivo release of [18F]FDG cured from PET Net GmbH (Erlangen, Germany). Rats were anaesthetized with Narcoren (Merial, Hallbergmoos, Germany). In order to assess the in vivo release of encapsulated [18F]FDG, SephadexTM G-25 (medium) was obtained from GE Healthcare 3 rats per different formulations were investigated in a small Bio-Sciences AB (Uppsala, Sweden) and Penefsky columns from animal PET scanner. The model used was as follows: The [18F]FDG- HWS Labortechnik (Mainz, Germany). LiposoFast-Basic extruder containing liposomes were injected i.p. It is assumed that these and polycarbonate membranes (pore diameter 100 nm) were liposomes stay in the peritoneum for several hours. The [18F]FDG from Avestin Europe GmbH (Mannheim, Germany). A rota- released from the liposomes is circulating in the blood pool. tor was purchased from neoLab Migge GmbH (Heidelberg, The dominant part of this “free” [18F]FDG will be transported Germany) and a dose calibrator from Veenstra Instruments to and trapped in brain cells because of the high-glucose con- (Ahlerstedt, Germany). DLS measurements were performed on sumption rate of brain cells. The trapping mechanism of [18F]FDG a ProSpecD 501 photon correlation spectrometer with ALV- reflects phosphorylation of [18F]FDG by intracellular hexokinase. correlator and ALV DLS-software 3.0 (Nanovel, Langenlonsheim, As [18F]FDG-6-phosphate is not a substrate of glucose-6-phosphate Germany). isomerase, this [18F]-labelled compound is retained inside the brain cells. For comparison and to test our model, free [18F]FDG solution 2.2. Preparation of liposomes and characterization was directly injected i.p. and the brain uptake was measured. The rats were anaesthetized with pentobarbital (40 mg/kg, i.p.) 2.2.1. Preparation of [18F]FDG-containing multilamellar vesicles before PET imaging. The animals breathed room air spontaneously (MLV) through a tracheal tube. For brain imaging they were fixed on The MLV were prepared by the freeze–thaw method (Mayer et the scanner’s bed in supine position.