Miniaturization of Drug Solubility and Dissolution Testings - Doria
Miniaturization of Drug Solubility and Dissolution Testings - Doria
Miniaturization of Drug Solubility and Dissolution Testings - Doria
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Division <strong>of</strong> Pharmaceutical Technology<br />
Faculty <strong>of</strong> Pharmacy<br />
University <strong>of</strong> Helsinki<br />
<strong>Miniaturization</strong> <strong>of</strong> <strong>Drug</strong> <strong>Solubility</strong> <strong>and</strong> <strong>Dissolution</strong><br />
<strong>Testings</strong><br />
Tiina Heikkilä<br />
ACADEMIC DISSERTATION<br />
To be presented, with the permission <strong>of</strong> the Faculty <strong>of</strong> Pharmacy <strong>of</strong> the University <strong>of</strong><br />
Helsinki, for public examination in auditorium 1041, Viikki Biocenter 2 (Viikinkaari 5 E),<br />
on 11 th June 2010, at 12 noon.<br />
Helsinki 2010
Supervisors: Docent Leena Peltonen<br />
Division <strong>of</strong> Pharmaceutical Technology<br />
Faculty <strong>of</strong> Pharmacy<br />
University <strong>of</strong> Helsinki<br />
Helsinki, Finl<strong>and</strong><br />
Pr<strong>of</strong>essor Marjo Yliperttula<br />
Division <strong>of</strong> Biopharmacy <strong>and</strong> Pharmacokinetics<br />
Faculty <strong>of</strong> Pharmacy<br />
University <strong>of</strong> Helsinki<br />
Helsinki, Finl<strong>and</strong><br />
Pr<strong>of</strong>essor Jouni Hirvonen<br />
Division <strong>of</strong> Pharmaceutical Technology<br />
Faculty <strong>of</strong> Pharmacy<br />
University <strong>of</strong> Helsinki<br />
Helsinki, Finl<strong>and</strong><br />
Reviewers: Pr<strong>of</strong>essor Jukka Rantanen<br />
Department <strong>of</strong> Pharmaceutics <strong>and</strong> Analytical Chemistry<br />
Faculty <strong>of</strong> Pharmaceutical Sciences<br />
University <strong>of</strong> Copenhagen<br />
Copenhagen, Denmark<br />
Associate Pr<strong>of</strong>essor Christos Reppas<br />
Department <strong>of</strong> Pharmaceutical Technology<br />
Faculty <strong>of</strong> Pharmacy<br />
University <strong>of</strong> Athens<br />
Athens, Greece<br />
Opponent: Doctor Juha Kiesvaara<br />
Head <strong>of</strong> Formulation Research<br />
Orion Pharma R&D<br />
Turku, Finl<strong>and</strong><br />
© Tiina Heikkilä<br />
ISBN 978-952-10-6190-5 (paperback)<br />
ISBN 978-952-10-6191-2 (PDF, http://e-thesis.helsinki.fi)<br />
ISSN 1795-7079<br />
Helsinki University Print<br />
Helsinki, Finl<strong>and</strong>, 2010
Abstract<br />
<strong>Solubility</strong> <strong>and</strong> drug dissolution are <strong>of</strong> crucial importance for drug formulations. Poor<br />
water-solubility <strong>of</strong> drug c<strong>and</strong>idates is a major obstacle in drug development, since the oral<br />
route is the most patient convenient <strong>and</strong> cost effective way to deliver drugs. In some cases<br />
the low aqueous solubility may limit the bioavailability when the absorption <strong>of</strong> the drug is<br />
dissolution limited. About 40% <strong>of</strong> the current lead optimization compounds suffer from<br />
poor solubility.<br />
Improvements in drug solubility/dissolution testing technologies (e.g. high throughput<br />
screening, HTS) can enhance the possibilities <strong>of</strong> the lead compounds to success in the later<br />
stages <strong>of</strong> drug development process. Typically HTS protocols measure the kinetic<br />
solubility involving co-solvents (e.g. dimethyl sulfoxide), which might enhance the in<br />
vitro solubility or give erroneous results if the potential drug c<strong>and</strong>idates are eliminated.<br />
Also, measurement <strong>of</strong> dissolution pr<strong>of</strong>iles is not available by the present HTS methods.<br />
For these reasons, there is a need for improvements in HTS solubility formats.<br />
Traditionally the in vitro dissolution tests are studied by pharmacopoeial methods,<br />
which are not utilizable in the drug discovery stage because <strong>of</strong> the large amount <strong>of</strong><br />
compounds needed. However, dissolution studies at the drug discovery stage could be<br />
useful e.g. to classify compounds based on their dissolution rates. In addition, the initial<br />
steps <strong>of</strong> dissolution process might be lost by the regulatory dissolution methods.<br />
The aim <strong>of</strong> this thesis was to miniaturize traditional drug dissolution <strong>and</strong> solubility<br />
testing methods. Systematic down scaling <strong>of</strong> methods was done towards the development<br />
<strong>of</strong> both equilibrium <strong>and</strong> kinetic 96-well plate solubility/dissolution methods.<br />
<strong>Miniaturization</strong> <strong>of</strong> the regulatory dissolution methods <strong>and</strong> shake-flask solubility<br />
measurements on the 96-well plates was successful. 96-well plate methods for equilibrium<br />
drug solubilities, as well as for drug dissolution pr<strong>of</strong>iles as a function <strong>of</strong> time, were<br />
developed. The former method is the first true equilibrium solubility method, which can<br />
be used in the screening <strong>of</strong> drug solubility <strong>and</strong> dissolution phenomena at the early stages<br />
<strong>of</strong> drug development process. This method was also tested using fasted state human<br />
intestinal fluid as a medium for the first time. Human intestinal fluid <strong>and</strong> data obtained<br />
might turn out to be important for very low water-solubility compounds. Surface tension<br />
based microtensiometry was also presented as an alternative method for kinetic HTS <strong>of</strong><br />
drug solubility properties, e.g. for classifying solubility <strong>of</strong> compounds not suitable for UVanalysis.<br />
Channel flow methodology was introduced enabling especially the kinetic<br />
follow-up at the very beginning <strong>of</strong> the dissolution process.<br />
This thesis provides directly applicable new miniaturized methods for the drug<br />
solubility <strong>and</strong> dissolution experiments. These methods enhance the throughput <strong>and</strong><br />
underst<strong>and</strong>ing <strong>of</strong> drug solubility/dissolution phenomena <strong>and</strong> pr<strong>of</strong>iles in drug discovery <strong>and</strong><br />
improve success in the later stages <strong>of</strong> the drug development process.
Acknowledgements<br />
This work was carried out at the Division <strong>of</strong> Pharmaceutical Technology, Faculty <strong>of</strong><br />
Pharmacy, University <strong>of</strong> Helsinki, Finl<strong>and</strong>.<br />
I wish to thank all the people who have been involved with this study.<br />
First, I would like to express my warmest gratitude to my supervisors Pr<strong>of</strong>essor Jouni<br />
Hirvonen, Docent Leena Peltonen <strong>and</strong> Pr<strong>of</strong>essor Marjo Yliperttula. In particular, I want to<br />
thank Leena for her constant patience <strong>and</strong> practical support. You were precent when<br />
needed <strong>and</strong> I felt like no stupid questions were. I wish everybody had a supervisor like<br />
you! Jouni <strong>and</strong> Marjo, first <strong>of</strong> all, thank you for this opportunity. Thank you also for your<br />
comments to all the versions <strong>of</strong> manuscripts <strong>and</strong> reports I have written, your scientifical<br />
guidance <strong>and</strong> providing excellent working facilities.<br />
The reviewers <strong>of</strong> the thesis, Pr<strong>of</strong>essor Jukka Rantanen <strong>and</strong> Associate Pr<strong>of</strong>essor<br />
Christos Reppas, are acknowledged for their careful work <strong>and</strong> valuable comments on<br />
impoving the manuscrpit.<br />
I am also grateful to all my co-authors, Pr<strong>of</strong>essor Patrick Augustijns, Doctor Milja<br />
Karjalainen, Pr<strong>of</strong>essor Kyösti Kontturi, Doctor Timo Laaksonen, Pr<strong>of</strong>essor Frank<br />
Lammert, Doctor Peter Liljeroth, M.Sc. Krista Ojala, M.Sc. Kirsi Partola, M.Sc. Saila<br />
Taskinen <strong>and</strong> Pr<strong>of</strong>essor Arto Urtti for their co-operation <strong>and</strong> scientific contributions.<br />
It has been very pleasent to work at Viikki because <strong>of</strong> all you co-workers. I felt warmly<br />
wellcome from the very beginning <strong>and</strong> I have got all the practical help I ever needed to<br />
perform this work.<br />
I would also like to thank my co-workers at the Lauttasaari Centralpharmacy for<br />
keeping me updated in practical pharmacy work. Especially, I want to thank pharmacy<br />
owner Liisa Heikkinen.<br />
In vitro in vivo relevance (IVIVRe) project in co-operation with Orion Pharma is<br />
acknowledged for the financial support.<br />
Finally, I would like to thank my most lowed ones. Especially, I would like to thank<br />
my Mum for introducing me the pharmacy <strong>and</strong> encouraging me to study this far, <strong>and</strong><br />
Sanna, thank you for endless support. Emmi, Elsa, Petter <strong>and</strong> Alpo, thank you for just<br />
being there, <strong>and</strong> Timo, thank you for your love all the way. I made it!<br />
Espoo, May 2010<br />
Tiina Heikkilä
Contents<br />
Abstract 3<br />
Acknowledgements 4<br />
List <strong>of</strong> original publications 7<br />
Abbreviations 8<br />
1 Introduction 9<br />
2 Review <strong>of</strong> the literature 11<br />
2.1 <strong>Solubility</strong> in drug discovery <strong>and</strong> development 11<br />
2.1.1 Kinetic solubility 13<br />
2.1.1.1 Dimethyl sulfoxide (DMSO) as a co-solvent 13<br />
2.1.1.2 Kinetic solubility assays 14<br />
2.1.2 Equilibrium solubility 15<br />
2.2 In vitro dissolution methods for solid dosage forms 16<br />
2.2.1 Regulatory dissolution methods 18<br />
2.2.2 Initial steps <strong>of</strong> the dissolution process 19<br />
2.2.3 In vivo mimicking <strong>of</strong> dissolution methods 22<br />
2.3 In vivo dissolution 23<br />
2.3.1 Effect <strong>of</strong> pH 23<br />
2.3.2 Role <strong>of</strong> bile acids <strong>and</strong> phospholipids 24<br />
2.3.3 Hydrodynamics in gastrointestinal (GI) tract 25<br />
2.4 In vitro in vivo (IVIV) correlation 26<br />
2.4.1 Biopharmaceutics Classification System (BCS) 26<br />
2.4.2 Biopharmaceutics <strong>Drug</strong> Disposition Classification System (BDDCS) 27<br />
2.4.3 BCS Biowaivers 28<br />
3 Aims <strong>of</strong> the study 30<br />
4 Experimental 31<br />
4.1 Materials 31<br />
4.1.1 Chemicals (I-IV) 31<br />
4.1.2 Human intestinal fluid (HIF) (IV) 31<br />
4.2 Methods 31<br />
4.2.1 Preparation <strong>of</strong> tablets (I, II) 31<br />
4.2.2 <strong>Dissolution</strong> methods 32<br />
4.2.2.1 Basket method (I, II) 32
4.2.2.2 Intrinsic dissolution rate method (I, II) 33<br />
4.2.2.3 Channel flow method (I, II) 33<br />
4.2.2.4 Micro dissolution method (previously unpublished data) 33<br />
4.2.2.5 96-well plate method (IV) 34<br />
4.2.3 <strong>Solubility</strong> methods 34<br />
4.2.3.1. Shake-flask method (III, IV) 34<br />
4.2.3.2 96-well plate method (IV) 35<br />
4.2.3.3 Surface tension based kinetic HTS solubility method (III) 35<br />
4.2.4 Analyses 35<br />
4.2.4.1 Analyses <strong>of</strong> human intestinal fluid (HIF) (IV) 35<br />
4.2.4.2 Solid-state characterization (IV) 35<br />
5 Results <strong>and</strong> discussion 37<br />
5.1 <strong>Dissolution</strong> testing 37<br />
5.1.1 Channel flow method vs. the regulatory dissolution methods (I, II) 37<br />
5.1.2 Micro dissolution studies in different media (previously unpublished data) 40<br />
5.1.3 <strong>Dissolution</strong> pr<strong>of</strong>iles in 96-well plate (IV) 41<br />
5.1.4 Scaling down the dissolution testings (I, II, III, IV) 41<br />
5.2 <strong>Solubility</strong> screening 42<br />
5.2.1 HTS 96-well plate methods (III, IV) 42<br />
5.2.2 Polymorphic changes (IV) 47<br />
5.3 Human intestinal fluid (HIF) 49<br />
5.3.1 Characterization <strong>of</strong> fasted state HIF (IV) 49<br />
5.3.2 HIF in solubility testing (IV) 50<br />
6 Conclusions 52<br />
7 References 53
List <strong>of</strong> original publications<br />
This thesis is based on the following publications, which are referred to in the text by their<br />
respective roman numerals (I-IV):<br />
I Peltonen, L., Liljeroth, P., Heikkilä, T., Kontturi, K., Hirvonen, J., 2003.<br />
<strong>Dissolution</strong> testing <strong>of</strong> acetylsalicylic acid by a channel flow method –<br />
correlation to USP basket <strong>and</strong> intrinsic dissolution methods. European<br />
Journal <strong>of</strong> Pharmaceutical Sciences, 19, 395-401.<br />
II Peltonen, L., Liljeroth, P., Heikkilä, T., Kontturi, K., Hirvonen, J., 2004. A<br />
novel channel flow method in determination <strong>of</strong> solubility properties <strong>and</strong><br />
dissolution pr<strong>of</strong>iles <strong>of</strong> theophylline tablets. STP Pharma Sciences, 14, 389-<br />
394.<br />
III Heikkilä, T., Peltonen, L., Taskinen, S., Laaksonen, T., Hirvonen, J., 2008.<br />
96-well plate surface tension measurements for fast determination <strong>of</strong> drug<br />
solubility. Letters in <strong>Drug</strong> Design <strong>and</strong> Discovery, 5, 471-475.<br />
IV Heikkilä, T., Karjalainen, M., Ojala, K., Partola, K., Lammert, F.,<br />
Augustijns, P., Urtti, A., Yliperttula, M., Peltonen, L., Hirvonen, J., 2010.<br />
Equilibrium drug solubility measurements in 96-well plates. International<br />
Journal <strong>of</strong> Pharmaceutics, submitted.<br />
Reprinted with the permission from the Elsevier Ltd (I, IV), Edition de Santé (II), <strong>and</strong><br />
Bentham Science Publishers Ltd (III).<br />
7
Abbreviations<br />
BCS Biopharmaceutics Classification System<br />
BDDCS Biopharmaceutics <strong>Drug</strong> Disposition Classification System<br />
CFC channel flow cell<br />
CMC critical micelle concentration<br />
CSD Cambridge Structural Database<br />
DMSO dimethyl sulfoxide<br />
DSC differential scanning calorimetry<br />
EMEA European Medicines Agency<br />
FaSSIF Fasted state simulated intestinal fluid<br />
FDA US Food <strong>and</strong> <strong>Drug</strong> Administration<br />
GI gastrointestinal<br />
HIF human intestinal fluid<br />
HPCD hydroxypropyl-β-cyclodextrin<br />
HPLC high pressure liquid chromatography<br />
HTS high throughput screening<br />
IDR intrinsic dissolution rate<br />
Int.Ph. The International Pharmacopoeia<br />
IR immediate release<br />
IUPAC The International Union <strong>of</strong> Pure <strong>and</strong> Applied Chemistry<br />
IVIV in vitro in vivo<br />
JP Japanese Pharmacopoeia<br />
MAD maximum absorbable dose<br />
MCC microcrystalline cellulose<br />
Ph. Eur. The European Pharmacopoeia<br />
rpm rotations per minute<br />
SDS sodium dodecyl sulphate<br />
USP The United States Pharmacopeia<br />
WHO World Health Organization<br />
XRPD X-ray powder diffraction<br />
8
1 Introduction<br />
Oral route is the most patient convenient <strong>and</strong> cost effective way to deliver drugs.<br />
However, inadequate aqueous solubility <strong>and</strong> dissolution in intestines <strong>and</strong> colon limits the<br />
bioavailability <strong>of</strong> drugs. Nowadays even 40% <strong>of</strong> the lead optimization compounds are<br />
poorly water-soluble (Lipinski et al., 1997). Although the poor solubility can be<br />
occasionally overcome by novel formulations, e.g. oral lipid-based (Porter <strong>and</strong> Charman,<br />
2001; Hauss, 2007) or amorphous formulations (Kawakami 2009), or delivery systems,<br />
e.g. cyclodextrins (Szejtli, 1988; Brewster <strong>and</strong> L<strong>of</strong>tsson, 2007), it still increases the cost<br />
<strong>and</strong> time <strong>of</strong> the new drug product to enter the market. Instead, improvement <strong>of</strong> drug<br />
solubility/dissolution testing methods (e.g. by high throughput screening, HTS) in the drug<br />
discovery stage can enhance the possibilities <strong>of</strong> the lead compound to success in the later<br />
stages <strong>of</strong> the drug development process.<br />
HTS solubility methods measure the solubility <strong>of</strong> numerous compounds at small<br />
quantities (Pan et al., 2001; Chen et al., 2002b; Takano et al., 2006; Bard et al., 2008; Dai<br />
et al., 2008; Alelyunas et al., 2009). Discovery protocols typically measure the kinetic<br />
solubility involving small amounts <strong>of</strong> co-solvents like dimethyl sulfoxide (DMSO). The<br />
co-solvents might enhance the in vitro solubility due to the co-solvent effect (Chen et al.,<br />
2002b) <strong>and</strong> the solubility in vivo might be totally different. Especially, for the poorly<br />
soluble drugs the solubility can increase manifold with co-solvents. Also, erroneous<br />
decisions can be made if wrong drug c<strong>and</strong>idates are eliminated because <strong>of</strong> rapid uptake <strong>of</strong><br />
water by DMSO. When the water concentration in the DMSO solution gets too high, some<br />
lead compounds may fall out <strong>of</strong> solution <strong>and</strong> give a false negative result (Homon <strong>and</strong><br />
Nelson, 2006). For these reasons, there is a need for improvements in HTS solubility<br />
formats.<br />
<strong>Drug</strong> dissolution testing <strong>of</strong> solid dosage forms is a routine process in pharmaceutical<br />
industry to evaluate drug release characteristics <strong>and</strong> to predict in vivo behavior.<br />
Traditionally the in vitro dissolution tests are studied by pharmacopoeial methods (e.g.<br />
paddle <strong>and</strong> basket methods) in the product development process for quality control, e.g. to<br />
assess the lot-to-lot quality <strong>and</strong> to detect changes in the drug formulation or manufacturing<br />
process. In the development <strong>of</strong> the drug product dissolution tests have been used to study<br />
the excipients (Omelczuk <strong>and</strong> McGinity, 1993; Uesu et al., 2000), <strong>and</strong> also to determine<br />
bioequivalence. In the drug discovery process pharmacopoeial dissolution methods are not<br />
utilizable because <strong>of</strong> the large amount <strong>of</strong> lead compounds <strong>and</strong>/or drug c<strong>and</strong>idates needed,<br />
however, dissolution studies at the drug discovery stage could be used to e.g. to classify<br />
compounds based on their dissolution rates or to show whether the dissolution has taken<br />
place.<br />
Biopharmaceutics Classification System (BCS), which divides drugs into four classes<br />
according to their solubility <strong>and</strong> permeability properties, correlates in vitro drug product<br />
dissolution <strong>and</strong> in vivo bioavailability (Amidon et al., 1995). After BCS the meaning <strong>and</strong><br />
application <strong>of</strong> relevant solubility/dissolution media has become more important <strong>and</strong> a<br />
broad range <strong>of</strong> dissolution media have been developed, e.g. FaSSIF (fasted state simulated<br />
intestinal fluid) <strong>and</strong> FeSSIF (fed state simulated intestinal fluid) (Galia et al., 1998). Still,<br />
the most physiological medium for solubility <strong>and</strong> dissolution studies would be human<br />
9
intestinal fluid (HIF), because drugs are usually dissolved <strong>and</strong> absorbed in the small<br />
intestine. The miniaturized drug solubility <strong>and</strong> dissolution methods requested above<br />
should enable the use <strong>of</strong> HIF when appropiate.<br />
In this thesis, different approaches were studied to miniaturize solubility <strong>and</strong><br />
dissolution testing methods. The in vitro in vivo (IVIV) correlation <strong>of</strong> the solubility testing<br />
was studied by using different buffer solutions <strong>and</strong> fasted state human intestinal fluid as<br />
media. The literature review focuses on the different available solubility <strong>and</strong> dissolution<br />
methods <strong>and</strong> on the IVIV correlation.<br />
10
2 Review <strong>of</strong> the literature<br />
2.1 <strong>Solubility</strong> in drug discovery <strong>and</strong> development<br />
IUPAC (The International Union <strong>of</strong> Pure <strong>and</strong> Applied Chemistry) defines solubility as<br />
‘the analytical composition <strong>of</strong> a mixture or solution which is saturated with one <strong>of</strong> the<br />
components <strong>of</strong> the mixture or solution, expressed in terms <strong>of</strong> the proportion <strong>of</strong> the<br />
designated component in the designated mixture or solution’ (Lorimer <strong>and</strong> Cohen-Adad,<br />
2003). <strong>Solubility</strong> can further be defined in a more specific manner, like unbuffered,<br />
buffered <strong>and</strong> intrinsic solubility. Unbuffered solubility means the solubility <strong>of</strong> a saturated<br />
solution <strong>of</strong> the compound at the final pH <strong>of</strong> the solution. Buffered (apparent) solubility<br />
refers to solubility at a given pH, measured in a defined pH-buffered system, <strong>and</strong> usually<br />
does not perceive the influence <strong>of</strong> salt formation with counterions <strong>of</strong> the buffering system<br />
on the measured solubility value. Intrinsic solubility (So) is the solubility <strong>of</strong> the neutral<br />
form <strong>of</strong> an ionizable compound. For neutral compounds all three definitions coincidence.<br />
<strong>Solubility</strong> measurements determine either kinetic or equilibrium (thermodynamic)<br />
solubility depending on the experimental set-up (Alsenz <strong>and</strong> Kansy, 2007).<br />
Compounds can be divided into highly or poorly soluble ones. Traditionally<br />
compounds were considered to be poorly soluble if the solubility was less than 100 μg/ml.<br />
In 2006 Fligge <strong>and</strong> Schuler redefined the concept <strong>of</strong> poor solubility by HTS method; a<br />
compound is poorly soluble when the solubility is less than 20 μg/ml, partly soluble when<br />
the solubility is 20-80 μg/ml <strong>and</strong> soluble when the solubility is over 80 μg/ml. According<br />
to Lipinski (2004), 15% <strong>of</strong> the drug-like compounds <strong>and</strong> 40% <strong>of</strong> lead optimization<br />
compounds were insoluble at a concentration <strong>of</strong> less than or equal to 20 μg/ml. The<br />
definitions <strong>of</strong> drug-likeness <strong>and</strong> lead-likeness have been discussed e.g. by Oprea et al.<br />
(2001), but in general it can be said that in the lead optimization process a lead compound<br />
is moved towards a drug c<strong>and</strong>idate having drug-like properties. In the pharmacopoeias<br />
solubility is indicated by the descriptive terms (Table 1).<br />
Table 1. The descriptive terms for the solubility used in the pharmacopoeias (USP,<br />
2009;Ph.Eur., 2009).<br />
Term Parts <strong>of</strong> solvent required for 1 part <strong>of</strong> solute<br />
Very soluble<br />
Less than 1<br />
Freely soluble<br />
From 1 to 10<br />
Soluble<br />
From 10 to 30<br />
Sparingly soluble<br />
From 30 to 100<br />
Slightly soluble<br />
From 100 to 1000<br />
Very slightly soluble<br />
From 1000 to 10 000<br />
Practically insoluble, or insoluble<br />
Greater than or equal to 10 000<br />
11
The requirement to the drug c<strong>and</strong>idate’s solubility has been studied by the concept <strong>of</strong><br />
maximum absorbable dose (MAD) (Johnson <strong>and</strong> Swindell, 1996):<br />
MAD = S 6.5 VL k a t (1)<br />
The unit <strong>of</strong> MAD is milligram, the transit time, t, is 270 minutes <strong>and</strong> the estimated luminal<br />
volume, VL, is 250 ml. S 6.5 marks the solubility at pH 6.5 (mg/ml) <strong>and</strong> ka the<br />
transintestinal absorption rate constant (min -1 ). In the MAD equation the small intestine<br />
can be considered as a tube, which has the volume <strong>of</strong> 250 ml. The saturated drug<br />
suspension is set into that tube, <strong>and</strong> the timer is started. The absorption rate constant<br />
describes how fast the dissolved quantity <strong>of</strong> the drug ‘leaks’ out <strong>of</strong> the tube each minute.<br />
After 270 minutes the total amount <strong>of</strong> ‘leakage’ is the estimate <strong>of</strong> the amount <strong>of</strong> the drug<br />
that should be intestinally absorbed. If that estimated quantity is close to the entire drug<br />
dosage, the intestinal absorption should be 100%.<br />
Lipinski (2000) has developed further the MAD concept based on the most common<br />
clinical potency <strong>of</strong> drug dose, 1 mg/kg. According to Lipinski, in the case <strong>of</strong> permeability<br />
being the average, the equilibrium (thermodynamic) solubility should be 52 μg/ml at pH<br />
6.5-7. For the low permeability drugs the equilibrium solubility should be 207 μg/ml <strong>and</strong><br />
for the high permeability drugs 10 μg/ml, in order to be therapeutically effective. Low<br />
dose drugs (0.1 mg/kg) with high permeability can be effective even at the equilibrium<br />
solubility <strong>of</strong> 1 μg/ml <strong>and</strong>, on the opposite cases, 10 mg/kg dose drugs having low<br />
permeability properties might need equilibrium solubilities <strong>of</strong> as high as 2100 μg/ml to be<br />
effective. The advantage <strong>of</strong> the MAD model is that it easily shows the impact that the<br />
required dose might have to the permeability, solubility <strong>and</strong> fraction absorbed <strong>of</strong> the dosed<br />
compound.<br />
The solubility determinations are performed several times along the drug discovery<br />
<strong>and</strong> development processes, only varying the focus, the amount <strong>of</strong> the drug used <strong>and</strong> also<br />
the level <strong>of</strong> the purity <strong>and</strong> crystallinity <strong>of</strong> the drug c<strong>and</strong>idate. In the early drug discovery<br />
stage the goal is to rank compounds according to the solubility <strong>and</strong> to aid in the structure<br />
activity chemistry relationships; in the drug development stage the goal is a more accurate<br />
numerical solubility number (Lipinski, 2004). More precisely, during the lead<br />
identification the solubility assays are, e.g., used to recognize compounds with potential<br />
liabilities, while in the lead optimization solubility information is used to make informed<br />
decisions on potential development c<strong>and</strong>idates. Also the equilibrium solubility is used to<br />
confirm earlier kinetic solubility results in the lead optimization. In the drug development<br />
stage the dissolution testing is an important area for solubility assays (Dokoumetzidis <strong>and</strong><br />
Macheras, 2006). In the earlier drug development stages the goal is, e.g., to identify ratelimiting<br />
steps, while in the later stages the dissolution tests are used, e.g., in the quality<br />
control <strong>and</strong> bioequivalence studies. Figure 1 depicts the solubility <strong>and</strong> dissolution assays<br />
in the drug discovery <strong>and</strong> development processes.<br />
12
New<br />
medicines<br />
proposal<br />
DISCOVERY DEVELOPMENT<br />
Target<br />
assessment<br />
Lead<br />
identi-<br />
fication<br />
Figure 1 <strong>Solubility</strong> <strong>and</strong> dissolution determinations in the drug discovery <strong>and</strong> development<br />
stages.<br />
2.1.1 Kinetic solubility<br />
Lead<br />
optimi-<br />
zation<br />
Kinetic Equilibrium<br />
solubility solubility<br />
(HTS) (shake-flask)<br />
Entry into<br />
human<br />
enable<br />
The measurement <strong>of</strong> the kinetic (non-equilibrium) solubility usually starts from the<br />
dissolved compound <strong>and</strong> represents the maximum solubility <strong>of</strong> the fastest precipitating<br />
material, which is typically not determined. The precipitating material can be any <strong>of</strong> the<br />
existing different solid forms <strong>of</strong> the compound, for example, polymorphic changes can<br />
happen even within minutes (Aaltonen et al., 2006). Also the formation <strong>of</strong> different salts<br />
may be possible depending on the presence <strong>of</strong> ionic species in the medium. In practice, the<br />
small volume (μl) <strong>of</strong> the stock solution (most <strong>of</strong>ten DMSO), in which the compound is<br />
dissolved, is added to the aqueous solution until the solubility limit is reached.<br />
Experiments can be performed by adding equal volumes <strong>of</strong> DMSO stock solution to the<br />
aqueous medium spaced a minute apart (Lipinski et al., 1997). In an alternative<br />
experiment, a dilution series <strong>of</strong> DMSO stock solution is prepared into separate vials<br />
(Kibbey et al., 2001). The kinetic solubility is defined to be the concentration at which the<br />
precipitation starts.<br />
2.1.1.1 Dimethyl sulfoxide (DMSO) as a co-solvent<br />
The amount <strong>of</strong> DMSO in the aqueous solution is increased in the kinetic solubility<br />
experiments by every addition <strong>of</strong> the DMSO stock solution. The final DMSO<br />
concentration varies <strong>and</strong> concentrations 0.5% (Bard et al., 2008), 1% (Chen et al., 2002b;<br />
Taub et al., 2002; Bard et al., 2008), 2% (Kibbey et al., 2001; Dehring et al., 2004), 5%<br />
(Pan et al., 2001; Bard et al., 2008) <strong>and</strong> even up to 20% (Kramer et al., 2009) have been<br />
reported. DMSO might induce a higher water-solubility than the compound would have<br />
without the DMSO due to the supersaturation or co-solvent effects (Chen et al., 2002b;<br />
L<strong>of</strong>tsson <strong>and</strong> Hreinsdottir, 2006; Bard et al., 2008). The co-solvents increase the solubility<br />
Phase<br />
I<br />
13<br />
Phase<br />
II<br />
Phase<br />
III<br />
<strong>Dissolution</strong><br />
(regulatory methods, intrinsic dissolution rate)<br />
Registration<br />
On<br />
market
y being miscible with water <strong>and</strong> acting as better solvents for the solute in question.<br />
DMSO is an efficient solvent for poorly water-soluble compounds because <strong>of</strong> its<br />
amphipathic structure with highly polar domain <strong>and</strong> two apolar groups enabling it to mix<br />
both hydrophilic (with water) <strong>and</strong> hydrophobic (organic compounds) environments<br />
(Santos et al., 2003). Supersaturated solutions are solutions, which have higher solute<br />
concentration than in the equilibrium state. The equilibrium is established between the<br />
solute in the solution <strong>and</strong> the excess <strong>of</strong> (undissolved) compound. DMSO enhances the<br />
solubility in a compound specific way (Chen et al., 2002b; Taub et al., 2002). Taub et al.<br />
(2002) investigated the possibilities <strong>of</strong> replacing DMSO by another co-solvent, for<br />
example ethanol or HPCD (hydroxypropyl-β-cyclodextrin), but the effect <strong>of</strong> enhanced<br />
solubility remained. According to Alsenz <strong>and</strong> Kansy (2007), the DMSO concentration<br />
should be kept as low as possible (1%) to reduce the potential co-solvent effect <strong>and</strong> to<br />
achieve better correlation with the equilibrium solubility methods. The final co-solvent<br />
<strong>and</strong> the amount <strong>of</strong> it should always be reported, although the solubility results obtained<br />
from different kinetic methods are typically not comparable to each other <strong>and</strong> the kinetic<br />
solubility values are not even expected to be reproducible between different kinetic<br />
methods (Stuart <strong>and</strong> Box, 2005). Comparison <strong>of</strong> the results is complicated because <strong>of</strong> the<br />
different DMSO concentrations at the end-point (Chen et al., 2002b). Di <strong>and</strong> Kerns (2006)<br />
have reviewed the strategies for optimizing bioassays with DMSO, including<br />
improvements in storage <strong>and</strong> h<strong>and</strong>ling. Serial dilutions in DMSO are recommended, <strong>and</strong><br />
these dilutions should be added directly to the assay media. Assays should also run at the<br />
lowest useful concentrations.<br />
DMSO, the current industry st<strong>and</strong>ard solvent for screening purposes, has some very<br />
problematic features, like it suffers from the rapid uptake <strong>of</strong> water; if the water content<br />
gets too high in the DMSO solutions, some compounds may fall out <strong>of</strong> solution <strong>and</strong> give a<br />
false negative result (Homon <strong>and</strong> Nelson, 2006). DMSO is hygroscopic <strong>and</strong> can absorb<br />
water up to 10% <strong>of</strong> its original volume (Di <strong>and</strong> Kerns, 2006), which also causes reduced<br />
solubility with the crystallization in the freeze-thaw cycles. DMSO has also several<br />
systemic adverse effects, e.g. nausea, vomiting, diarrhea, anaphylactic reactions,<br />
bradycardia <strong>and</strong> heart block are reported (O’Donnell et al., 1981; Davis et al., 1990;<br />
Rapoport et al., 1991; Stroncek et al., 1991; Styler et al., 1992). Besides, even 20% <strong>of</strong> the<br />
compounds in the commercial libraries are poorly soluble in DMSO according to Balakin<br />
et al., 2004.<br />
2.1.1.2 Kinetic solubility assays<br />
In general, kinetic solubility values are measured in the drug discovery process to screen<br />
the lead compounds. A high-throughput turbidimetric method for kinetic solubility assay,<br />
in which the detection was based on nephelometry (Davis <strong>and</strong> Parke, 1942), was<br />
introduced by Lipinski et al. (1997). Detection is based on the intensity <strong>of</strong> the scattered<br />
light as a function <strong>of</strong> concentration. In general, kinetic solubility measurements are fast<br />
<strong>and</strong> suitable for the HTS processes h<strong>and</strong>ling more than 100.000 samples per day <strong>and</strong> cover<br />
14
the required concentration range in assays used in early discovery (Homon <strong>and</strong> Nelson,<br />
2006). More details about the methods are presented in the Table 2.<br />
Table 2. Kinetic solubility in the drug discovery stage. Precipitation quantified by<br />
nephelometry or UV-absorbance.<br />
Sensitiveness<br />
Sample volume<br />
DMSO conc.<br />
Advantages<br />
Nephelometry a-d Diode array UV a UV plate reader d-g<br />
20-54 μM<br />
200-300 μl<br />
0.67-5%<br />
Colored compounds do<br />
not interfere,<br />
Shaking or<br />
ultrasonication is not<br />
required.<br />
5-65 μg/ml<br />
2.5 ml<br />
0.67%<br />
Accurate determination based<br />
on a calibration curve.<br />
0.5-1 μM<br />
200 μl<br />
0.5–5%<br />
Able to detect impurities<br />
or low UV absorption.<br />
Disadvantages Small undissolved Colored compounds miscalled Compounds should<br />
particles cause as insoluble.<br />
have sufficient<br />
misinterpretations;<br />
absorbance in the<br />
High/low<br />
region > 230 nm,<br />
concentrations<br />
Absorption <strong>of</strong> residues<br />
can be problematic.<br />
or other materials at the<br />
same wavelength as<br />
drug compounds.<br />
a b c d<br />
Lipinski et al., 1997; Bevan <strong>and</strong> Lloyd, 2000; Dehring et al., 2004; Hoelke et al., 2009;<br />
e f g<br />
Pan et al., 2001; Chen et al., 2002b; Bard et al., 2008<br />
2.1.2 Equilibrium solubility<br />
The equilibrium (thermodynamic) solubility test is performed by dispensing a surplus <strong>of</strong><br />
solid compound into a liquid medium. After reaching the equilibrium state between the<br />
excess <strong>of</strong> undissolved compound <strong>and</strong> saturated solution, the saturation solubility can be<br />
defined. To confirm that equilibrium state has been reached, the solubility has to be<br />
constant for several consecutive time-points, which takes several hours or days, even<br />
months to reach. Usually equilibrium solubility is studied in the later stages <strong>of</strong> drug<br />
15
discovery process <strong>and</strong> it is regarded as the golden st<strong>and</strong>ard for development needs (Alsenz<br />
<strong>and</strong> Kansy, 2007).<br />
A classical equilibrium solubility assay is the shake-flask method. In this method an<br />
excess <strong>of</strong> solid drug is dispensed into aqueous solution, after which the container is shaken<br />
at certain speed <strong>and</strong> temperature. Samples are collected regularly, usually during 24-48<br />
hours, to make certain that the equilibrium plateau has been reached. The shake-flask<br />
method has not been <strong>of</strong>ficially defined in pharmacopoeias, so several protocols exist<br />
(Avdeef et al., 2000; Blasko et al., 2001; Kerns, 2001; Bergström et al., 2002; Lipinski et<br />
al., 2003; Chen <strong>and</strong> Venkatesh, 2004; Glomme et al., 2005; Avdeef, 2007; Zhou et al.,<br />
2007) (Table 3).<br />
Table 3. Different protocols to perform shake-flask solubility measurements (N.A. not<br />
available in the article).<br />
Blasko et al., 2001<br />
Kerns, 2001<br />
Bergström et al., 2002<br />
Chen <strong>and</strong> Venkatesh, 2004<br />
Glomme et al., 2005<br />
Avdeef, 2007<br />
Zhou et al., 2007<br />
Medium<br />
volume (ml)<br />
5.0<br />
0.45<br />
0.05-1.0<br />
1.0<br />
2.0<br />
2.0-20.0<br />
1.0-5.0<br />
Shaking<br />
(rpm)<br />
N.A.<br />
N.A.<br />
300<br />
N.A.<br />
450<br />
N.A.<br />
600<br />
16<br />
Residue removal Quantification<br />
filtration<br />
filtration<br />
centrifugation<br />
filtration<br />
filtration<br />
filtration or<br />
centrifugation<br />
filtration or<br />
centrifugation<br />
2.2 In vitro dissolution methods for solid dosage forms<br />
<strong>Dissolution</strong> (dissolution rate) can be expressed by the Noyes-Whitney equation:<br />
HPLC<br />
HPLC<br />
HPLC-UV/fluorescence<br />
HPLC-UV<br />
HPLC-UV<br />
HPLC-UV<br />
HPLC-UV<br />
(LC-MS)<br />
dm DA<br />
= (Cs – Cb). (2)<br />
dt h<br />
The rate <strong>of</strong> mass transfer <strong>of</strong> solute molecules or ions through a static diffusion layer,<br />
dm/dt, is directly proportional to the area available for molecular or ionic migration, A, the<br />
concentration difference, Cs-Cb, <strong>and</strong> D, the diffusion coefficient <strong>of</strong> the solute (usually<br />
m 2 /s) <strong>and</strong> is inversely proportional to the thickness <strong>of</strong> the boundary layer, h. Cs is the
saturation solubility <strong>and</strong> Cb is the concentration <strong>of</strong> drug at a particular time in the bulk<br />
dissolution medium (Noyes <strong>and</strong> Whitney, 1897). If the Cb does not exceed 10% <strong>of</strong> the Cs,<br />
the dissolution is said to occur under sink conditions. In the table 4 are the factors<br />
affecting the dissolution rate.<br />
Table 4. Factors affecting the dissolution rate.<br />
Term Factor Examples<br />
A Particle size <strong>and</strong> shape<br />
- Aggregation<br />
Dokoumetzidis <strong>and</strong> Macheras,<br />
2006<br />
- Crystal shape affected by crystallization process<br />
Fini et al., 1995<br />
Cs<br />
Cb<br />
h<br />
D<br />
Solid-state form<br />
- Change in the solid form properties<br />
Complex formation<br />
Salts<br />
- common ion-effect<br />
Solubilizing agents<br />
Temperature<br />
<strong>Dissolution</strong> medium<br />
- e.g. pH (ionizable compounds), co-solvents<br />
Volume <strong>of</strong> medium<br />
Processes that remove dissolved solute from the medium<br />
- e.g. adsorption<br />
Thickness <strong>of</strong> boundary layer<br />
- affected by agitation<br />
Diffusion coefficient <strong>of</strong> solute in the dissolution medium<br />
- affected by the viscosity <strong>of</strong> medium<br />
17<br />
Pudipeddi <strong>and</strong> Serajuddin, 2005<br />
Brewster <strong>and</strong> L<strong>of</strong>tsson, 2007<br />
Serajuddin, 2007<br />
Bakatselou et al., 1991<br />
Hill <strong>and</strong> Petrucci, 1999<br />
Taub et al., 2002<br />
Dressman et al., 1998<br />
Lindenberg et al., 2005<br />
Alsenz et al., 2007<br />
Dressman et al., 1998
The importance <strong>of</strong> dissolution to the drug absorption <strong>and</strong> bioavailability is emphasized in<br />
the Biopharmaceutics Classification System (see also later chapters) by Dn, dissolution<br />
number (Amidon et al., 1995):<br />
Dn =<br />
t<br />
t<br />
res<br />
Diss<br />
where tres is the mean intestinal residence time (ca. 180 min, which differs from MAD<br />
model time) <strong>and</strong> tDiss is the time required for a drug particle to dissolve. From the above<br />
equation it is obvious that the higher the Dn, the better the drug absorption, <strong>and</strong> the<br />
maximal drug absorption is expected when Dn>1. In the case <strong>of</strong> Dn
Table 5. The <strong>of</strong>ficial dissolution testing conditions.<br />
Medium<br />
Volume (ml)<br />
pH <strong>of</strong> buffers<br />
Temperature ( o C)<br />
Enzymes<br />
Agitation (rpm)<br />
Basket method<br />
Paddle method<br />
a<br />
FDA, 1997a<br />
Ph. Eur. USP a JP<br />
900<br />
1.0-7.5<br />
37±0.5<br />
case-by-case<br />
100<br />
50<br />
500, 900, 1000<br />
1.2-6.8<br />
37±0.5<br />
case-by-case<br />
50-100<br />
50-75<br />
2.2.2 Initial steps <strong>of</strong> the dissolution process<br />
19<br />
900<br />
1.2-6.8<br />
37±0.5<br />
case-by-case<br />
The main difference between the stirring methods described above <strong>and</strong> intrinsic<br />
dissolution method (IDR) (Ph. Eur., 2009; USP, 2009) or channel flow cell technique<br />
(Compton et al., 1988; Brown et al., 1993; Compton et al., 1993) is that in the latter<br />
methods only one tablet side is exposed to the dissolution medium (Figure 2).<br />
100<br />
50
Figure 2 Schematic illustrations <strong>and</strong> liquid flows <strong>of</strong> a) Intrinsic dissolution apparatus <strong>and</strong> b)<br />
Channel flow cell method.<br />
Intrinsic dissolution rate (IDR)<br />
The IDR method is based on the idea that the dissolution environment (surface area,<br />
temperature, pH, stirring speed <strong>and</strong> ionic strength) <strong>of</strong> pure drug substances is kept constant<br />
during the whole measurement (Nicklasson et al., 1981; Nicklasson et al., 1985). The fluid<br />
flow has been studied (Mauger, 1996; Mauger et al., 2003) <strong>and</strong> the IDR method has also<br />
been successfully miniaturized requiring only 5 mg <strong>of</strong> pure drug (Berger et al., 2007;<br />
Persson et al., 2008). As an advantage the compression pressure, dissolution medium<br />
volume or die position have no significant effect on IDR (Yu et al., 2004). The IDR <strong>of</strong> the<br />
pure drug is an important physicochemical property, which is useful, e.g., when predicting<br />
20
problems related to the absorption <strong>of</strong> the drugs. The evaluation <strong>of</strong> the IDR <strong>of</strong> pure drug<br />
substances demonstrates chemical purity <strong>and</strong> equivalence.<br />
The intrinsic rate (j) is expressed in terms <strong>of</strong> dissolved mass <strong>of</strong> compound per time per<br />
exposed area (mg min -1 cm -2 ):<br />
V dc<br />
j = , (4)<br />
A dt<br />
where V is volume, A is area <strong>of</strong> the drug disk, c is concentration <strong>and</strong> t is time. IDR can<br />
also be calculated by Levich equation (Yu et al., 2004):<br />
2 / 3<br />
D <br />
j = 0.62 ( 1/<br />
6<br />
v<br />
1/<br />
2<br />
) Cs, (5)<br />
where Cs is saturated solubility or concentration, D is the diffusion coefficient, ω is the<br />
angular velocity <strong>of</strong> the rotating disk <strong>and</strong> v is the kinematic viscosity <strong>of</strong> the dissolution<br />
medium. It is assumed in this model that equilibrium is predominant at the interface <strong>of</strong><br />
solid <strong>and</strong> liquid, <strong>and</strong> the liquid at the interface is the same as the liquid in bulk dissolution.<br />
Since the intrinsic dissolution is primarily a rate phenomenon, it is expected to correlate<br />
more closely with in vivo drug dissolution dynamics than solubility. Still, the correlation<br />
between the solubility <strong>and</strong> intrinsic dissolution rate <strong>of</strong> a drug compound has been found<br />
both in water or aqueous buffer solutions (Nicklasson et al., 1981; Gu et al., 1987;<br />
Nicklasson et al., 1988) <strong>and</strong> water-ethanol binary solvent systems (Nicklasson <strong>and</strong> Brodin,<br />
1984). Also the correlation between the IDR <strong>and</strong> BCS solubility classification has been<br />
found by Yu et al. (2002).<br />
Channel flow cell (CFC) method<br />
The hydrodynamics (laminar flow) in the channel flow cell (CFC) is well defined<br />
(Compton et al., 1988; Brown et al., 1993; Compton et al., 1993), which enables<br />
acquisition <strong>of</strong> quantitative information <strong>of</strong> the dissolution kinetics <strong>and</strong> mechanism, even at<br />
early stages <strong>of</strong> the process, <strong>and</strong> modelling. This information is <strong>of</strong>ten lost in the traditional<br />
dissolution methods. In addition, accurate modelling <strong>of</strong> the mass transfer in the CFC<br />
enables comparisons <strong>of</strong> different dissolution mechanisms <strong>and</strong> experimental results. The<br />
flow rate in the channel presents an additional experimental parameter that can be used to<br />
tune the rate <strong>of</strong> mass transfer in the flowing aqueous phase.<br />
When the dissolution experiments involve pure compounds, the model consists <strong>of</strong> a<br />
convective mass transfer in the medium coupled to a dissolution reaction at the tablet<br />
surface with an equilibrium constant, K (Singh <strong>and</strong> Dutt, 1985). Theoretical expression to<br />
the concentration, c b , as a function <strong>of</strong> time:<br />
21
c b mA<br />
(t) = K (1 – exp (- t)) (6)<br />
V<br />
where m is the mass transfer coefficient, A is the exposed area <strong>of</strong> the tablet <strong>and</strong> V is the<br />
volume <strong>of</strong> the dissolution medium <strong>and</strong> t is time.<br />
In the CFC the tablet is located in the cavity so that the disintegration <strong>of</strong> the tablet is<br />
hindered when, e.g., the properties <strong>of</strong> disintegration excipients do not affect the<br />
dissolution as in the traditional methods. In the case <strong>of</strong> tablet formulations, it is not<br />
expected that the mass transfer coefficient would be equal for all tablets <strong>and</strong> only a<br />
function <strong>of</strong> the flow rate. With tablet formulations including excipients, diffusion <strong>and</strong><br />
dissolution inside the tablet <strong>and</strong> the sink conditions were included in the model (Higuchi,<br />
1961):<br />
2 b<br />
b<br />
V c ( t)<br />
V V 2K<br />
K − c ( t)<br />
t + + ( + ) ln ( ) = 0 (7)<br />
2 o<br />
2 o<br />
D A c mA D A c K<br />
T<br />
T<br />
where DT is the diffusion coefficient <strong>of</strong> the drug inside the tablet <strong>and</strong> c 0 the concentration<br />
<strong>of</strong> the drug in the dry tablet.<br />
The combination <strong>of</strong> dissolution testing <strong>and</strong> solid-state analysis has been proposed to be<br />
called physico-relevant dissolution testing (Aaltonen <strong>and</strong> Rades, 2009). Recently, the<br />
channel flow method has been utilized, e.g., to study the solvent-mediated solid-state<br />
transformation during dissolution (Aaltonen et al., 2006; Lehto et al., 2008), the solid-state<br />
behavior during dissolution <strong>of</strong> amorphous drugs (Savolainen et al., 2009) <strong>and</strong> the effect <strong>of</strong><br />
texture on the intrinsic dissolution behavior (Tenho et al., 2007).<br />
2.2.3 In vivo mimicking <strong>of</strong> dissolution methods<br />
Alternative methods have been developed to complete the lacks <strong>of</strong> the regulatory methods.<br />
In general, the <strong>of</strong>ficial basket <strong>and</strong> paddle dissolution techniques suffer from high<br />
variability in results (Achanta et al., 1995; Qureshi <strong>and</strong> McGilvery, 1999). Further, the<br />
basket method suffers, e.g., from the hydrodynamic ‘dead zone’ under the basket,<br />
blockings in the mesh by drug or excipient <strong>and</strong> hindered visual examination <strong>of</strong> the<br />
disintegration process (Banaker, 1992). The paddle method, for its part, suffers above all<br />
from the coning effect (Beckett et al., 1996).<br />
Several un<strong>of</strong>ficial in vitro dissolution methods exist (Pernarowski et al., 1968;<br />
Shepherd et al., 1972; Shah et al., 1973; Cakiryildis et al., 1975), some <strong>of</strong> them modified<br />
from the regulatory methods (Qureshi <strong>and</strong> Shabnam, 2003; Röst <strong>and</strong> Quist, 2003). The<br />
new methods, e.g. the dissolution stress-test device, are more complicated (Garbacz et al.,<br />
2008; Garbacz et al., 2009). This method mimics the discontinuous movements <strong>of</strong> the<br />
dosage form <strong>and</strong> breaks <strong>of</strong> it from the intestinal fluid in the GI tract. Even more<br />
complicated release <strong>and</strong> solubility testing methods are the physiologically based TIM-1<br />
22
(stomach <strong>and</strong> small intestine) <strong>and</strong> TIM-2 systems (large-intestine) (Havenaar <strong>and</strong><br />
Menikus, 1994). For example the TIM-1 has four compartments simulating stomach,<br />
duodenum, jejunum <strong>and</strong> ileum. The TIM devices mimic, e.g., the peristaltic movements,<br />
GI transition times, pH <strong>of</strong> GI tract <strong>and</strong> secretion <strong>of</strong> bile, gastric acids <strong>and</strong> enzymes. Results<br />
with the TIM-1 system compared to the in vivo data have been similar (Blanquet et al.,<br />
2004; Souliman et al., 2006). Still, the new dissolution devices need a large volume <strong>of</strong><br />
medium, which enables their use only in the later stages <strong>of</strong> drug development process.<br />
2.3 In vivo dissolution<br />
The most common <strong>and</strong> patient convenient way to deliver drugs is the oral route. After oral<br />
administration, the physiology <strong>of</strong> the GI tract affects the dissolution <strong>and</strong> absorption <strong>of</strong> the<br />
drug. The GI tract is divided into three major anatomical regions (the stomach, the small<br />
intestine <strong>and</strong> the colon), from which the small intestine is the most important site for drug<br />
absorption. The length <strong>of</strong> the small intestine is about 3.3 meters (Ho et al., 1983) <strong>and</strong> the<br />
area available for absorption is about 200 m 2 (Ashford, 2007). The large area <strong>of</strong> small<br />
intestine is possible because <strong>of</strong> the folds <strong>of</strong> Keckring, the villi <strong>and</strong> the microvilli. The<br />
absorption sites <strong>of</strong> drugs in the small intestine differ from each other. Highly permeable<br />
drugs absorb mainly at the tips <strong>of</strong> the villi <strong>and</strong> the low permeability drugs may diffuse<br />
down the length <strong>of</strong> the villi to be absorbed over a larger surface area (Artursson et al.,<br />
2001). The main anatomical barrier for the drug absorption is the enterocytes, which<br />
apical side constitutes <strong>of</strong> densely packed microvilli <strong>and</strong> the lateral membranes connected<br />
through tight junctions. The absorption routes from the intestinal epithelium are based on<br />
paracellular passive diffusion, transcellular pathway (passive diffusion, carrier-mediated,<br />
receptor-mediated) or carrier-mediated efflux pathways (Back <strong>and</strong> Rogers, 1987; Tsuji<br />
<strong>and</strong> Tamai, 1996; Stenberg et al., 2000; Artursson et al., 2001; Martinez <strong>and</strong> Amidon,<br />
2002).<br />
Before possible absorption the drug needs to be released <strong>and</strong> dissolved from the solid<br />
dosage form to the GI fluids. The pH, composition, volume <strong>and</strong> hydrodynamics <strong>of</strong> the<br />
contents in the lumen are the main factors in the dissolution <strong>of</strong> drugs in the GI tract<br />
(Dressman et al., 1998).<br />
2.3.1 Effect <strong>of</strong> pH<br />
The pH in the GI tract varies depending on the anatomical region <strong>and</strong> the nutrition state.<br />
Stomach has an acidic pH in the fasted state, median values are found between 1.7 <strong>and</strong> 2.4<br />
with high interindividual variations. After meal the gastric median pH increases up to 6.4-<br />
6.7 depending on the type <strong>and</strong> volume <strong>of</strong> the meal. The pH <strong>of</strong> stomach is returned acidic<br />
2-4 hours after the meal (Dressman et al., 1990; Kalantzi et al., 2006a). Intestinal pH is<br />
higher than the gastric pH. In the fasting state the median pH is about 6.1-7.2 (Lindahl et<br />
al., 1997; Kalantzi et al., 2006a; Moreno et al., 2006; Brouwers et al., 2006; Clarysse et<br />
al., 2009b) <strong>and</strong> in the fed state 6.1-6.6 (Dressman et al., 1990; Kalantzi et al., 2006a;<br />
23
Clarysse et al., 2009b), with less interindividual variation compared to the gastric pH<br />
(Table 6). The pH has been reported to be similar in the jejunum (median 7.0) <strong>and</strong><br />
duodenum (median 6.5) in the fasted state (Moreno et al., 2006). The pH in the colon<br />
depends on the region. In the caecum the pH is around 5.6-6.6, increasing up to 7-7.5 in<br />
the distal parts <strong>of</strong> the colon (Fallingborg et al., 1989). Especially, for ionizable compounds<br />
the pH <strong>of</strong> the GI tract is important; Ratio <strong>of</strong> unionized <strong>and</strong> ionized forms depends on pH.<br />
Usually the ionized form has higher solubility than unionized form. For example weak<br />
acids has slow solubility <strong>and</strong> dissolution in stomach, whereas they can achieve high<br />
solubility <strong>and</strong> dissolution in intestine. On the contrary, for poorly soluble compounds the<br />
pH <strong>of</strong> GI is not the rate-limiting step for dissolution, but in the GI tract is not enough fluid<br />
available or the entire dose <strong>of</strong> the drug dissolves too slowly (Dressman et al., 1998).<br />
2.3.2 Role <strong>of</strong> bile acids <strong>and</strong> phospholipids<br />
Bile is secreted to the small intestine. The secretion <strong>of</strong> bile is at maximum about 30<br />
minutes after ingestion <strong>of</strong> a meal <strong>and</strong> the gallbladder is emptied up to 73% (Mathus-<br />
Vliegen et al., 2004). The major components <strong>of</strong> bile are the bile acids, phospholipids,<br />
cholesterol <strong>and</strong> free fatty acids. The bile salt concentration compared to the phospholipid<br />
concentration has been reported to be 2-3:1 in the fed state <strong>and</strong> 9:1 in the fasted state<br />
(Persson et al., 2005; Kalantzi et al., 2006b). Bile has an important role in improving<br />
solubility <strong>and</strong> dissolution <strong>of</strong> poorly water-soluble drugs. Especially, bioavailability <strong>of</strong><br />
poorly soluble drugs increases manifold when administered in fed state or with lipid-based<br />
formulations (Persson et al., 2005; Sunesen et al., 2005; Kalantzi et al., 2006b). The effect<br />
could result from decreased first pass metabolism, lymphatic uptake (Porter et al., 1996),<br />
inhibition <strong>of</strong> efflux transporters (Wu et al., 2005) or increased solubility (Persson et al.,<br />
2005; Kalantzi et al., 2006b) <strong>and</strong> dissolution. Bile acids enhance solubility due to the<br />
wetting (Bakatselou et al., 1991) <strong>and</strong>/or solubilization effect (Bakatselou et al., 1991;<br />
Charman et al., 1993) which has been demonstrated e.g. with steroidal compounds<br />
(Mithani et al., 1995).<br />
For the prediction <strong>of</strong> in vivo dissolution <strong>and</strong> solubility, simulated intestinal fluids, like<br />
FaSSIF <strong>and</strong> FeSSIF, have been developed (Galia et al., 1998). Jantratid et al. (2008)<br />
proposed further a set <strong>of</strong> media to represent various regions <strong>of</strong> the gastrointestinal tract:<br />
fasted state simulated gastric fluid (FaSSGF) (Vertzoni et al., 2005), fed state simulated<br />
gastric fluid (FeSSGF), fasted <strong>and</strong> fed state simulated upper small intestine fluids<br />
(FeSSIF-V2, FaSSIF-V2) <strong>and</strong> various media for conditions that represent digestion.<br />
Simulated media contain e.g. lecithin <strong>and</strong> sodium taurocholate to represent the<br />
phospholipids <strong>and</strong> the bile salts.<br />
24
Table 6. The compositions <strong>of</strong> fasted <strong>and</strong> fed state HIF.<br />
Bile Acids/salts (mM)<br />
Taurocholic acid (%)<br />
Glycocholate (%)<br />
Glycochenodeoxycholate (%)<br />
Glycodeoxycholate (%)<br />
Taurochenodeoxycholate (%)<br />
Taurodeoxycholate (%)<br />
Phospholipids (mM)<br />
Lyso-phosphatidylcholine (%)<br />
Phosphatidylcholine (%)<br />
Parameters<br />
Osmolality (mOsm/kg)<br />
pH<br />
Buffer capacity (mmol/l pH -1 )<br />
Fasted HIF Fed HIF<br />
2.0-3.5 a-e<br />
22.0 a , 44.95±18.17 d<br />
25.0 a , 17.21±6.37 d<br />
18.0 a , 13.77±9.59 d<br />
1.0 a , 10.40±7.32 d<br />
20.0 a , 7.89±2.16 d<br />
1.0 a , 4.71±2.47 d<br />
0.2–0.6 a,i<br />
99.0 a<br />
1.0 a<br />
200–271 b-d<br />
6.6-7.5 a-f<br />
2.4-2.8 a-e<br />
25<br />
8.0-15.0 a,g,i<br />
18.0 a<br />
25.0 a<br />
20.0 a<br />
13.0 a<br />
15.0 a<br />
9.0 a<br />
0.07-4.8 a,g,i<br />
87.0 a<br />
13.0 a<br />
287 k<br />
5.1-6.1 f,j,k<br />
13.2-14.6 a<br />
a Persson et al., 2005; b Clarysse et al., 2009b; c Lindahl et al., 1997; d Moreno et al., 2006;<br />
e Brouwers et al., 2006; f Kalantzi et al., 2009a; g Mansbach et al., 1975; h Sheng et al., 2009;<br />
i Arm<strong>and</strong> et al., 1996; j Clarysse et al., 2009a; k Dressman et al., 1998<br />
2.3.3 Hydrodynamics in gastrointestinal (GI) tract<br />
In the fasted state GI tract has only little or no motoric activity. Still, the ‘’house-keeping’’<br />
waves empty the stomach regularly about every two hours from the stomach to the small<br />
intestine. In the fed state mixing is more active <strong>and</strong> dissolution is expected to be more<br />
efficient. The gastric emptying time is higher after meal ingestion or after large liquid<br />
volume compared to the fasted state <strong>and</strong> the lipid contents <strong>of</strong> the meal increases the transit<br />
time further (Sunesen et al., 2005). For small intestine the transit time varies from three to<br />
four hours <strong>and</strong> stays the same whether the drug is taken with meal or not (Sunesen et al.,<br />
2005). Instead, the flow velocity varies in different parts <strong>of</strong> the small intestine e.g. being<br />
higher in the ileum as in the jejunum (Ho et al., 1983). The volume <strong>of</strong> stomach <strong>and</strong><br />
intestine in fasted state is 20-30 ml <strong>and</strong> 120-350 ml (Dressman et al., 1998). Under fed<br />
conditions the volume might be as high as 1500 ml in the upper intestine.
2.4 In vitro in vivo (IVIV) correlation<br />
IVIV correlation has been defined as ‘a predictive mathematical model describing the<br />
relationship between in vitro property <strong>of</strong> a dosage form <strong>and</strong> relevant in vivo response, for<br />
example plasma drug concentration or amount <strong>of</strong> drug absorbed’ (FDA, 1997b). IVIV<br />
correlation serves as a surrogate for in vivo data <strong>and</strong> is also used to validate the dissolution<br />
methods <strong>and</strong> in quality control. Three main levels <strong>of</strong> the IVIV correlation have been<br />
determined: level A, level B <strong>and</strong> level C (FDA, 1997b). The level A correlation represents<br />
the linear relationship between in vitro dissolution rate <strong>and</strong> in vivo input curves <strong>of</strong> the drug<br />
from the dosage form. Level A correlation is the most informative <strong>and</strong> recommended, if<br />
possible. Table 7 compares the release testing in vitro <strong>and</strong> in vivo. Medium, volume,<br />
hydrodynamics, temperature, location <strong>of</strong> the formulation <strong>and</strong> amount <strong>of</strong> the drug are<br />
considered.<br />
Table 7. Differences between the in vitro <strong>and</strong> in vivo release testing <strong>of</strong> immediate release<br />
drug products.<br />
Parameter IN VIVO IN VITRO<br />
Medium GI fluids<br />
Water,<br />
Buffer solutions,<br />
Simulated media<br />
Volume 250 ml (fasted state)<br />
Depending on the device<br />
(interindividual variation) (from μl:s to1000 ml)<br />
Hydrodynamics GI motility Mechanical agitation,<br />
Depending on the device<br />
Temperature Body temperature Depending on the test<br />
circumstances<br />
Location <strong>of</strong> the formulation Variation with time Depending on the device<br />
Amount <strong>of</strong> the drug Decreases as it is absorbed Constant in closed systems,<br />
(e.g. open system)<br />
Decreases in open systems<br />
2.4.1 Biopharmaceutics Classification System (BCS)<br />
The meaning <strong>of</strong> BCS is to correlate in vitro drug product dissolution <strong>and</strong> in vivo<br />
bioavailability (Amidon et al., 1995). In practice, BCS divides drugs into four classes<br />
according to their solubility <strong>and</strong> permeability properties. The drugs are classified as highly<br />
soluble if the highest orally administered single dose can be completely dissolved in 250<br />
ml <strong>of</strong> water, independently <strong>of</strong> the pH value at the range <strong>of</strong> 1.0–7.5 at 37 o C. Further, the<br />
drug is considered to be highly permeable when 90% or more <strong>of</strong> the orally administered<br />
dose is absorbed in the small intestine. The ideal drug would be class I drug having high<br />
26
solubility <strong>and</strong> permeability. The dissolution test for these drugs is only to verify that the<br />
drug is rapidly released from the dosage form. For the BCS class I drugs the IVIV<br />
correlation exist if dissolution rate is slower than gastric emptying rate. For BCS class II<br />
drugs the drug absorption is expected to be dissolution limited <strong>and</strong> IVIV correlation<br />
should exist if the dose is not very high. For BCS class III drugs limited or no IVIV<br />
correlation is expected because permeability is the rate limiting step. These drugs dissolve<br />
rapidly, while the slow absorption restricts bioavailability during the passage in the<br />
absorption site in the GI tract. The BCS class IV drugs are <strong>of</strong>ten not very suitable for oral<br />
administration at all, so limited or no IVIV correlation is expected.<br />
Fleischer et al. (1999) noticed that food (high-fat meals) effects to the extent <strong>of</strong><br />
bioavailability could generally be predicted based on the BCS. For example, high-fat<br />
meals have no significant effect on the extent <strong>of</strong> availability for class I drugs, although<br />
food may delay the stomach emptying time. Class II drugs would benefit from high-fat<br />
meals, e.g., because <strong>of</strong> additional solubilization <strong>and</strong> inhibition <strong>of</strong> efflux transporters in the<br />
lumen. High-fat meal effects would be decreased if the class II drugs were formulated<br />
towards having better solubility.<br />
2.4.2 Biopharmaceutics <strong>Drug</strong> Disposition Classification System (BDDCS)<br />
In 2005 Wu <strong>and</strong> Benet represented a modified version <strong>of</strong> BCS called Biopharmaceutics<br />
<strong>Drug</strong> Disposition Classification System (BDDCS) (Table 8), which is used in the<br />
prediction <strong>of</strong> overall drug disposition, including the rates <strong>of</strong> elimination routes <strong>and</strong> the<br />
effects <strong>of</strong> efflux <strong>and</strong> absorptive transporters on oral drug absorption. Accordingly, the<br />
extent <strong>of</strong> drug metabolism would be a better predictor than the 90% absorption<br />
characteristic. It was noticed that BCS class I <strong>and</strong> II compounds are eliminated primarily<br />
via metabolism, while class III <strong>and</strong> IV compounds end up unchanged into the urine <strong>and</strong><br />
bile; if the major route <strong>of</strong> elimination for a drug is metabolism, then the drug exhibits the<br />
properties <strong>of</strong> high permeability compounds, <strong>and</strong> if the major route <strong>of</strong> elimination is renal<br />
<strong>and</strong> biliary excretion <strong>of</strong> unchanged drug, the drug is classified as low permeability<br />
compound. Using the extent <strong>of</strong> metabolism instead <strong>of</strong> permeability (BCS) changed the<br />
classification <strong>of</strong> 10 drugs, changed 11 multiple classified drugs to one class, <strong>and</strong> added 38<br />
drugs to the totally 168 drugs’ BDDCS list. The BDDCS predicted 93% <strong>of</strong> the drugs’<br />
permeabilities correctly related to major route <strong>of</strong> elimination (Benet et al., 2008).<br />
27
Table 8. The Biopharmaceutics <strong>Drug</strong> Disposition Classification System by Wu <strong>and</strong> Benet<br />
(2005). The major route <strong>of</strong> elimination (metabolized vs. unchanged) serves as the<br />
permeability criteria.<br />
Class I<br />
High <strong>Solubility</strong><br />
Extensive Metabolism<br />
Class III<br />
High <strong>Solubility</strong><br />
Poor Metabolism<br />
28<br />
Class II<br />
Poor <strong>Solubility</strong><br />
Extensive Metabolism<br />
Class IV<br />
Poor <strong>Solubility</strong><br />
Poor Metabolism<br />
The BDDCS can also predict the effects <strong>of</strong> transporters. According to the Wu <strong>and</strong> Benet<br />
(2005), the transporter effects are minimal for class I drugs; efflux transporter effects<br />
predominate for the class II drugs <strong>and</strong> transporter-enzyme interplay in the intestines is<br />
important for the CYP3A substrates <strong>and</strong> phase 2 conjugation enzymes; absorptive<br />
transporter effects predominate for the class III drugs <strong>and</strong>; absorptive or efflux transporter<br />
effects could be important for the class IV drugs.<br />
2.4.3 BCS Biowaivers<br />
BCS is utilized for the so-called biowaivers, which means that in vivo bioavailability<br />
<strong>and</strong>/or bioequivalence studies may be waived. In that case a dissolution test could be<br />
adopted as the surrogate basis for the decision as to whether two pharmaceutical products<br />
are equivalent. According to the regulatory authorities (e.g. FDA, 2000; EMEA, 2002),<br />
only highly soluble <strong>and</strong> highly permeable BCS class I drugs can have a biowaiver status in<br />
IR solid oral-dosage forms. Furthermore, biowaiver products may not contain excipients<br />
that could influence the absorption <strong>of</strong> a drug compound, drug compound can not have a<br />
narrow therapeutic index, <strong>and</strong> the product can not be designed to be absorbed from the<br />
oral cavity.<br />
Yu et al. (2002) have proposed to extend the biowaiver status to BCS class II <strong>and</strong> III<br />
drugs. They criticized the regulatory limits <strong>of</strong> solubility <strong>and</strong> permeability being too strict.<br />
The changes they proposed were to narrow the required solubility pH range to the 1.0-6.8<br />
<strong>and</strong> to reduce the high permeability requirement from 90% to 85%. Also the boundaries<br />
for solubility <strong>of</strong> acidic BCS class II drugs have been shown to be too restrictive by<br />
Yazdanian et al. (2004). Other studies (Cheng et al., 2004; Kortejärvi et al., 2005;<br />
Kortejärvi et al., 2007; Kortejärvi et al., 2010) <strong>and</strong> reports (WHO, 2006) have been<br />
published recommending a biowaiver status for BCS class III drugs. Benet et al. (2008)<br />
recommended the regulatory authorities to add the extent <strong>of</strong> drug metabolism (i.e. 90%
metabolized) as an alternative method for the extent <strong>of</strong> drug absorption (i.e. 90%<br />
absorbed) in defining class I drugs suitable for biowaivers. This would increase the<br />
number <strong>of</strong> class I drugs eligible for biowaivers.<br />
29
3 Aims <strong>of</strong> the study<br />
The overall aim <strong>of</strong> the study was to develop predictive, reliable <strong>and</strong> rapid kinetic <strong>and</strong><br />
equilibrium solubility <strong>and</strong> dissolution screening methods, which do not require large<br />
amounts <strong>of</strong> compounds <strong>and</strong> would (ultimately) correlate well with in vivo results. Several<br />
solubility <strong>and</strong> dissolution testing methods were developed <strong>and</strong> characterized by using<br />
different volumes <strong>of</strong> aqueous buffer solutions <strong>and</strong> fasted state HIF as media. The specific<br />
aims <strong>of</strong> the study were:<br />
1. To construct <strong>and</strong> characterize the channel flow method <strong>and</strong> to compare it with<br />
the regulatory USP basket <strong>and</strong> intrinsic dissolution methods (I, II).<br />
2. To miniaturize dissolution testing stepwise from the regulatory dissolution<br />
methods <strong>and</strong> shake-flask solubility measurements via micro dissolution method<br />
to the 96-well plates (micro dissolution testing is previously unpublished data).<br />
3. To develop 96-well plate HTS methodology to determine both the equilibrium<br />
solubility <strong>and</strong> dissolution rate while avoiding the co-solvent(s) related<br />
problems. The model compounds were initially in solid form as in the<br />
traditional equilibrium solubility method in shake-flasks (IV).<br />
4. To develop kinetic HTS drug solubility testing method suitable for compounds<br />
having problems with UV absorption. The method is based on compounds’<br />
surface tension measurements by a novel microtensiometer (III).<br />
5. To compare the aqueous equilibrium solubility/dissolution results in the 96well<br />
plate method with corresponding testings in fasted state HIF medium (IV).<br />
30
4 Experimental<br />
4.1 Materials<br />
4.1.1 Chemicals (I-IV)<br />
The model compounds <strong>and</strong> references to the corresponding publications are summarized<br />
in Table 9. The model compounds were from Orion Pharma (Finl<strong>and</strong>) except for<br />
antipyrine (Aldrich Chemical Company Inc., USA), dicl<strong>of</strong>enac sodium (Sigma-Aldrich<br />
Chemie GmbH, Germany), indomethacin (Fluka Biochemika, Italy), piroxicam (Hawkins<br />
Inc., USA) <strong>and</strong> theophylline anhydrous (Helm AG, Germany). As tablet excipients were<br />
used microcrystalline cellulose (MCC) (Avicel PH 102, FMC, Irel<strong>and</strong>) (I, II), lactose<br />
(monohydr., 80 M, DMV International, The Netherl<strong>and</strong>s) (I, II), magnesium stearate<br />
(University Pharmacy, Finl<strong>and</strong>) (II) <strong>and</strong> talc (Pharmia Oy, Finl<strong>and</strong>) (II). All the other<br />
laboratory chemicals were <strong>of</strong> analytical grade.<br />
4.1.2 Human intestinal fluid (HIF) (IV)<br />
The HIF was collected from five volunteers after an overnight fast by introducing one<br />
double-lumen catheter (Salem Sump Tube 14 Ch, external diameter 4.7 mm, Sherwood<br />
Medical, Petit Rechain, Belgium) in the duodenum (D2/D3) via the mouth. Sampling <strong>of</strong><br />
HIF was continued for 3 hours every 15 minutes. HIF was stored at -30 o C or colder <strong>and</strong><br />
before solubility studies the individual HIF samples were pooled together.<br />
4.2 Methods<br />
4.2.1 Preparation <strong>of</strong> tablets (I, II)<br />
Tabletting masses <strong>of</strong> 200 g (I) <strong>and</strong> 150 g (II) were prepared. The powders were mixed for<br />
15 min in turbula mixer (Turbula T2C, Willy A. Bach<strong>of</strong>en AG Machinenfabrik,<br />
Switzerl<strong>and</strong>). Tablets were directly compressed with a single punch tabletting machine<br />
(Korsch EK-0, Erweka Apparatebau, Germany) using a flat punch with a diameter <strong>of</strong> 9<br />
mm. The target weights <strong>of</strong> tablets were 250 mg.<br />
31
Table 9. The model compounds.<br />
Model compound<br />
Molecular<br />
structure<br />
BCS<br />
class<br />
32<br />
pKa<br />
Acidic<br />
/basic<br />
Paper<br />
Acetylsalicylic acid C16H12O6 I 3.48 A I<br />
Antipyrine C11H12N2O I 0.7 B IV<br />
Atenolol C14H22N2O3 III 9.16 B IV<br />
Caffeine C8H10N4O2 I 0.73 B IV<br />
Carbamazepine C15H12N2O II 13.94 A IV<br />
Dicl<strong>of</strong>enac sodium C14H10Cl2NNaO2 II 4.18 A IV<br />
Furosemide C12H11ClN2O5S IV 3.04 A IV<br />
Hydrochlorothiazide C7H8ClN3O4S2 III 8.95 A IV<br />
Ibupr<strong>of</strong>en C13H18O2 II 4.41 A III, IV<br />
Indomethacin C19H16ClNO4 II 3.93 A III, IV<br />
Ketopr<strong>of</strong>en C16H14O3 II 4.23 A IV<br />
Metformin HCl C4H12ClN5 III 13.86 B IV<br />
Metoprolol tartrate C15H25NO3 I 9.17 B IV<br />
Naproxen C14H14O3 II 4.84 A IV<br />
Phenytoin sodium C15H11N2NaO2 II 8.33 A IV<br />
Piroxicam C15H13N3O4S II 3.80,4.50,<br />
13.33<br />
A/B IV<br />
Propranolol HCl C16H22ClNO2 I 9.14 B IV<br />
Ranitidine HCl C13H23ClN4O3S III 8.40 B IV<br />
Rifampicin C43H58N4O12 II 9.92 A IV<br />
Spironolactone C24H32O4S IV N.A. N.A. IV<br />
Theophylline<br />
anhydrous<br />
C7H8N4O2 II 8.60,1.70 A/B II<br />
Trimethoprim C14H18N4O3 II 7.2 B IV<br />
4.2.2 <strong>Dissolution</strong> methods<br />
4.2.2.1 Basket method (I, II)<br />
USP basket method (Sotax AT 7, Sotax, Switzerl<strong>and</strong>) was used as a reference dissolution<br />
method (USP XXVI, 2003). The rotational speed was 50 rpm, the volume <strong>of</strong> medium was
900 ml <strong>and</strong> the temperature was 23 o C. The samples were analyzed by a spectrophotometer<br />
(Perkin-Elmer Lambda 2, Germany).<br />
4.2.2.2 Intrinsic dissolution rate method (I, II)<br />
The intrinsic dissolution test was done according to the USP XXVI (2003), which is based<br />
on Wood’s rotating disk method (Wood, 1965). The model compounds (100 mg) were<br />
pressed with a hydraulic pressure <strong>of</strong> 1500 kg (Pye Unicam, UK) <strong>and</strong> rotated at 50 rpm in<br />
900 ml <strong>of</strong> the medium at the temperature <strong>of</strong> 23 o C. (Sotax AT 7, Sotax, Germany). The<br />
samples were conducted into a spectrophotometer (Perkin-Elmer Lambda 2, Germany) for<br />
analyses.<br />
4.2.2.3 Channel flow method (I, II)<br />
In the channel flow system only one tablet surface was in contact with the media (Figure<br />
3). The tablet was located at the centre <strong>of</strong> the channel flow cell in order to avoid the edgeeffects<br />
(Compton et al., 1988). The flow rate <strong>of</strong> the aqueous phase was controlled by a<br />
peristaltic pump (Ismatec IPN-12, Switzerl<strong>and</strong>) <strong>and</strong> temperature (23 o C) by a thermostatted<br />
bath. The channel flow cell was coupled to a UV-VIS spectrophotometer (HP8541A,<br />
Hewlett-Packard, USA).<br />
Figure 3 a) The channel flow cell <strong>and</strong> b) the channel flow test apparatus. Modified from<br />
publications I <strong>and</strong> II<br />
.<br />
4.2.2.4 Micro dissolution method (previously unpublished data)<br />
By this method the solubility <strong>of</strong> selected model compounds (ibupr<strong>of</strong>en, ketopr<strong>of</strong>en,<br />
indomethacin <strong>and</strong> piroxicam) was determined in 0.1 M hydrochloric acid solution in pH<br />
1.2 <strong>and</strong> 6.8 buffer solutions <strong>and</strong> in biorelevant medium, FaSSIF. Excess <strong>of</strong> drug<br />
33
compound was added into 2 ml flat-bottomed test tubes <strong>of</strong> the micro dissolution apparatus<br />
μDISS ® (pION, USA). <strong>Dissolution</strong> medium was added into the test tubes <strong>and</strong> the solution<br />
was stirred at 180 rpm using magnet stirrers. The temperature <strong>of</strong> the test solutions was<br />
maintained at 37 o C with the aid <strong>of</strong> a water bath surrounding the test tube block. The test<br />
solutions were kept under these circumstances for 3 hours after which the solutions were<br />
filtered using preheated Millipore PVDF 0.45 μm filters. The solutions were then left at<br />
37 o C for 24 hours <strong>and</strong> the concentration <strong>of</strong> the solutions were measured by detecting the<br />
UV absorbance <strong>of</strong> the solutions in situ.<br />
Quantification <strong>of</strong> the sample solutions was determined against a calibration curve with<br />
four to five dilutions <strong>of</strong> st<strong>and</strong>ard stock solution. The concentrations <strong>of</strong> the solutions were<br />
determined at the absorption maximum wavelength. Further, to assure the buffer capacity<br />
<strong>of</strong> the solutions, pH <strong>of</strong> the sample solutions was measured after the solubility runs.<br />
4.2.2.5 96-well plate method (IV)<br />
Model drugs were first dissolved into a presolvent (methanol or water), <strong>and</strong> the presolvent<br />
solutions were then pipetted into 96-well plates (96F untreated 260895, Nunc, Denmark),<br />
after which the presolvent was evaporated in order to avoid co-solvent effects. After<br />
evaporation, 250 μl <strong>of</strong> the medium was added to the wells <strong>and</strong> the 96-well plates were<br />
shaken orbitally at 600 rpm (Heidolph titramax 100, Germany). Samples were withdrawn<br />
after 30 min, 1, 2, 5, 24 <strong>and</strong> 30 hours <strong>and</strong> pipetted into UV 96-well plates (UV FB<br />
microtiter plates 8404, Thermo Fisher Oy, USA) for analyses with UV plate reader<br />
λ=230-400 nm (Varioskan, Thermo Electron corporation, Finl<strong>and</strong>). The tests were<br />
performed at 23 o C.<br />
4.2.3 <strong>Solubility</strong> methods<br />
4.2.3.1. Shake-flask method (III, IV)<br />
Shake-flask method was used as a reference method in the solubility tests (Lipinski,<br />
2003). The model compounds were weighted into 25 ml (III) or 40 ml (IV) glass vials,<br />
which were mixed by magnetic stirrer at 100 rpm at 23 o C during the tests. Samples were<br />
taken after 1, 2, 3, 4, 5 <strong>and</strong> 24, even after 30 hours for some compounds, until the<br />
equilibrium plateau was reached. Samples were filtered (0.45 μm, Minisart, Sartorius,<br />
UK) before analyses by UV-spectrophotometry (Pharmacia LKB, Ultrospec, Sweden).<br />
34
4.2.3.2 96-well plate method (IV)<br />
The method was described in chapter 4.2.2.5.<br />
4.2.3.3 Surface tension based kinetic HTS solubility method (III)<br />
The model drugs were first dissolved into a presolvent (methanol, ethanol or DMSO). 12-<br />
22 samples were prepared so that the first one had a drug concentration <strong>of</strong> ca. 100 mg/ml,<br />
<strong>and</strong> each subsequent sample was diluted to half the concentration <strong>of</strong> the previous one.<br />
Final sample was always a blank sample with no drug. Then, the presolvent solutions were<br />
added to the aqueous medium in the 96-well plate by a volume <strong>of</strong> 50 μl. Surface tension<br />
based measurements were performed with recently introduced Delta-8 multichannel<br />
microtensiometer (Kibron Inc., Finl<strong>and</strong>).<br />
4.2.4 Analyses<br />
4.2.4.1 Analyses <strong>of</strong> human intestinal fluid (HIF) (IV)<br />
Bile salts were measured by GC-MS-selected ion monitoring analysis developed by<br />
Lütjohann et al. (2004). Free bile acids were extracted with diethylether <strong>and</strong><br />
trimethylsilylated after alkaline hydrolysis <strong>and</strong> acidification. The chromatographic<br />
separation was performed with a DB-XLB capillary column in a constant flow mode <strong>of</strong><br />
0.8 ml/min with a final temperature <strong>of</strong> 290 o C having helium as a carrier gas.<br />
Total phospholipid levels were determinated enzymatically according to Gurantz et al.<br />
(1981), using phospholipase D <strong>and</strong> choline oxidase. The reagent kit was from Wako<br />
Chemicals (Neuss, Germany).<br />
4.2.4.2 Solid-state characterization (IV)<br />
X-ray powder diffraction (XRPD) was used to study the solid state forms <strong>of</strong> the model<br />
compounds <strong>and</strong> to confirm possible (pseudo)polymorphic changes. Measurements were<br />
done by an X-ray powder diffraction theta-theta diffractometer (Bruker AXS D8 advance,<br />
Bruker AXS GmbH, Germany). The XRPD experiments were performed in symmetrical<br />
reflection mode with CuKα radiation (1.54 Å) using Göbel Mirror bent gradient multilayer<br />
optics. The scattered intensities were measured with a scintillation counter. The angular<br />
range was from 5 o to 30 o with the steps <strong>of</strong> 0.1 o <strong>and</strong> the measuring time was 20 s/step.<br />
The reference codes to identify experimental polymorphs were from CSD database<br />
(Cambridge Structural Database, The Cambridge Crystallographic Data Centre, UK)<br />
35
(Allen, 2002). The reference crystal structure codes were for dicl<strong>of</strong>enac form I (P21/c)<br />
SIKLIH02 (Castellari <strong>and</strong> Ottani, 1997), dicl<strong>of</strong>enac form II (C2/c) SIKLIH05 (Muangsin<br />
et al., 2004), dicl<strong>of</strong>enac form III (Pcan) SIKLIH04 (Jaiboon et al., 2001), indomethacin αform<br />
INDMET2 (Chen et al., 2002a), indomethacin γ-form INDMET (Kistenmacher <strong>and</strong><br />
Marsh, 1972), piroxicam monohydrate CIDYAP0 (Bordner, 1984), piroxicam α-form<br />
BIYSEH06 (Vrecer et al., 2003), piroxicam β-form BIYSEH (Kojic-Prodic et al., 1982)<br />
<strong>and</strong> spironolactone polymorph 1 from acetone ATPRCL10 (Dideborg <strong>and</strong> Dupont, 1972).<br />
Differential scanning calorimetry (DSC) was used to further confirm the XRPD<br />
results. The thermal analyses were performed with a Mettler DSC 823e (Mettler-Toledo<br />
AG, Switzerl<strong>and</strong>) <strong>and</strong> analyzed by STAR e s<strong>of</strong>tware (Mettler-Toledo AG, Switzerl<strong>and</strong>).<br />
Samples were weighted (4-5 mg) to the aluminium pans <strong>and</strong> crimped by aluminium caps<br />
with a pinhole. The samples were heated at the rate <strong>of</strong> 10 o C/min from 25 o C to the 10 o C<br />
above the melting point <strong>of</strong> each compound. The measurements were performed under N2<br />
atmosphere <strong>and</strong> the calibration was performed with indium.<br />
36
5 Results <strong>and</strong> discussion<br />
5.1 <strong>Dissolution</strong> testing<br />
5.1.1 Channel flow method vs. the regulatory dissolution methods (I, II)<br />
<strong>Dissolution</strong> pr<strong>of</strong>iles <strong>of</strong> pure acetylsalicylic acid <strong>and</strong> theophylline from the channel flow<br />
method correlated well with the intrinsic dissolution pr<strong>of</strong>iles. In both methods the<br />
dissolution rate <strong>of</strong> acetylsalicylic acid at the time-point 25-min was around 5-10 mg/cm 2 at<br />
pH 1.2 <strong>and</strong> around 20-25 mg/cm 2 at pH 6.8 (I). The dissolution rates <strong>of</strong> theophylline at the<br />
time-point 10-min were around 11-12 mg/cm 2 at pH 1.2 <strong>and</strong> 10 mg/cm 2 at pH 6.8 by both<br />
the methods (II).<br />
The solubility <strong>of</strong> acetylsalicylic acid tablets (I) containing microcrystalline cellulose<br />
(MCC) (compositions A <strong>and</strong> C) was slightly increased by the lowering pH in the channel<br />
flow method opposite to the basket method. This could be due to the MCC swelling<br />
properties being in more important role in the channel flow method compared to the<br />
solubility <strong>of</strong> acetylsalicylic acid. Swelling <strong>of</strong> MCC affects the flow <strong>and</strong> the mass transfer<br />
characteristics hindering MCC’s disintegrating effect <strong>of</strong> the tablet in the CFC. The<br />
dissolution results with tablets containing only lactose as an excipient correlated well<br />
between the methods <strong>and</strong> the rank order <strong>of</strong> the dissolution pr<strong>of</strong>iles with all tablet batches<br />
(A-C) was similar by both methods (Figure 4).<br />
37
(a)<br />
(c)<br />
pH 1.2 pH 1.2<br />
pH 6.8<br />
Figure 4 Channel flow dissolution <strong>of</strong> acetylsalicylic acid tablet formulations A-C at pH 1.2 (a)<br />
<strong>and</strong> at pH 6.8 (c) (Flow rate was ca. 190 ml/h). <strong>Dissolution</strong> <strong>of</strong> acetylsalicylic acid<br />
with the USP basket method at pH 1.2 (b) <strong>and</strong> at pH 6.8 (d). Modified from paper I.<br />
The tablet compositions were more complicated with theophylline (II) containing four<br />
excipients (MCC, lactose, talc, magnesium stearate). The amounts <strong>of</strong> MCC <strong>and</strong> lactose<br />
varied between the tablet batches (A-D). The drug release rate measured by the channel<br />
flow method was decreased as the amount <strong>of</strong> MCC was increased. The USP basket<br />
method gave different result; the slowest drug release was from the tablets having the<br />
highest amount <strong>of</strong> lactose (A) or MCC (D). The difference can be explained by the<br />
different dissolution set-ups. In the channel flow method the position <strong>of</strong> the tablet hinders<br />
the disintegrating effect <strong>of</strong> MCC <strong>and</strong> even the formulation C (MCC 56%) remained as a<br />
coherent matrix. From the basket method the tablet formulation C disintegrated<br />
immediately. Also with theophylline the two methods lead to similar conclusions (Figure<br />
5).<br />
(b)<br />
(d)<br />
38<br />
B<br />
B A<br />
A<br />
C<br />
A C<br />
pH 6.8
(a)<br />
(c)<br />
pH 1.2<br />
pH 6.8<br />
(d)<br />
(b)<br />
Figure 5 Channel flow dissolution <strong>of</strong> theophylline from tablet formulations A-D at pH 1.2 (a)<br />
<strong>and</strong> at pH 6.8 (c) (Flow rate was ca. 190 ml/h). <strong>Dissolution</strong> <strong>of</strong> theophylline with the<br />
USP basket method at pH 1.2 (b) <strong>and</strong> at pH 6.8 (d) (A , B , C, D). Modified<br />
from paper II.<br />
The main difference between the channel flow <strong>and</strong> USP basket methods is that in the<br />
channel flow method only one tablet surface is exposed to the dissolution medium <strong>and</strong>,<br />
thus, the disintegration <strong>of</strong> the tablet is hindered. This enables collection <strong>of</strong> accurate<br />
information from the initial steps <strong>of</strong> the dissolution process. The intrinsic dissolution<br />
method is based on the idea <strong>of</strong> keeping the dissolution area <strong>of</strong> pure compound constant<br />
(Nicklasson et al., 1981; Nicklasson <strong>and</strong> Brodin, 1984) <strong>and</strong> achieving the dissolution by<br />
rotating or shear-like motion in the medium, which differs from the laminar flow in the<br />
channel flow method (Compton et al., 1988). Also the mass transfer can be followed with<br />
the channel flow method, which allows comparison between the different dissolution<br />
mechanisms <strong>and</strong> the experimental results. Channel flow cell method is also used in the<br />
modelling because <strong>of</strong> the well-defined laminar flow in the CFC (Aaltonen et al., 2006;<br />
Lehto et al., 2008).<br />
39<br />
pH 1.2<br />
pH 6.8
5.1.2 Micro dissolution studies in different media (previously unpublished<br />
data)<br />
Micro dissolution studies were performed in Orion Pharma with four model drugs<br />
(ibupr<strong>of</strong>en, indomethacin, ketopr<strong>of</strong>en <strong>and</strong> piroxicam) in three different media (aqueous<br />
solution pH 1.2, buffer solution pH 6.8 <strong>and</strong> FaSSIF) at +37 o C. Table 10 presents results<br />
comparing equilibrium solubilities with micro dissolution, shake-flask <strong>and</strong> equilibrium 96well<br />
plate methods. The latter were both performed at room temperature (23±1 o C).<br />
Table 10. Equilibrium solubilities (mg/ml) <strong>of</strong> four model compounds in different media<br />
(SF=Shake-flask, MD=Micro dissolution <strong>and</strong> HTS=equilibrium 96-well plate<br />
method, n=2).<br />
Compound<br />
Ibupr<strong>of</strong>en<br />
Indomethacin<br />
Ketopr<strong>of</strong>en<br />
Piroxicam<br />
SF<br />
pH 1.2<br />
0.07<br />
0.004<br />
0.08<br />
0.06<br />
MD<br />
pH 1.2<br />
0.05<br />
0.001<br />
0.13<br />
0.19<br />
SF<br />
pH 6.8<br />
2.9<br />
0.4<br />
5.5<br />
0.2<br />
40<br />
MD<br />
pH 6.8<br />
3.5<br />
0.6<br />
5.4<br />
0.6<br />
MD<br />
FaSSIF<br />
pH 6.5<br />
3.3<br />
0.6<br />
5.6<br />
0.4<br />
HTS<br />
HIF<br />
pH 6.2<br />
2.0<br />
2.4<br />
3.2<br />
0.8<br />
As expected, the equilibrium drug solubilities by the micro dissolution <strong>and</strong> shake-flask<br />
methods were higher at pH 6.8 than at pH 1.2 (pKa values <strong>of</strong> the model compounds are<br />
mostly around 4). The micro dissolution solubilities in the pH 6.8 buffer solution <strong>and</strong><br />
FaSSIF (pH 6.5) correlated well suggesting that the buffer might be an adequate medium<br />
at this solubility scale for the model compounds, although FaSSIF contains taurocholic<br />
acid <strong>and</strong> lecithin. Instead, the solubility results in HIF (pH 6.2) were somewhat lower for<br />
ibupr<strong>of</strong>en <strong>and</strong> ketopr<strong>of</strong>en <strong>and</strong> somewhat higher for piroxicam <strong>and</strong> indomethacin compared<br />
to the aqueous buffer pH 6.8 or FaSSIF pH 6.5. The compositions <strong>of</strong> fasted state HIF <strong>and</strong><br />
FaSSIF are more complicated than the composition <strong>of</strong> aqueous buffer at pH 6.8 containing<br />
e.g. bile acids, which enhance solubility due to the solubilization effect (Bakatselou et al.,<br />
1991; Charman et al., 1993) <strong>and</strong>/or wetting (Bakatselou et al., 1991). Here, the dissolution<br />
method (micro dissolution vs. 96-well plate) has important role to give similar results.<br />
According to our study, the medium (aqueous buffer, FaSSIF, HIF) has minor role. The<br />
use <strong>of</strong> HIF (<strong>and</strong> FaSSIF) in the solubility studies may have more pronounced importance<br />
for very low solubility compounds at the discovery stage, although the more simple, fast<br />
<strong>and</strong> economic methods/techniques are preferred also there.
5.1.3 <strong>Dissolution</strong> pr<strong>of</strong>iles in 96-well plate (IV)<br />
Besides the equilibrium solubility determinations, the 96-well plate method can be used to<br />
study the dissolution rates <strong>of</strong> drugs as a function <strong>of</strong> time (IV). Figure 6 distinguishes<br />
nicely the dissolution pr<strong>of</strong>iles <strong>of</strong> ibupr<strong>of</strong>en <strong>and</strong> indomethacin regardless the shaking<br />
speeds in question. At the moment the method is useful for the classification <strong>of</strong> dissolution<br />
rates because <strong>of</strong> rather long centrifugation times before the sampling. By shortening the<br />
centrifugation time, more detailed dissolution studies are possible for more rapidly<br />
dissolving materials.<br />
Concentration (mg/ml)<br />
Figure 6 Aqueous dissolution pr<strong>of</strong>iles <strong>of</strong> ibupr<strong>of</strong>en <strong>and</strong> indomethacin at different shaking<br />
speeds (300, 600 <strong>and</strong> 1200 rpm) (n=3). Modified from paper IV.<br />
5.1.4 Scaling down the dissolution testings (I, II, III, IV)<br />
The scaling down <strong>of</strong> the dissolution testing from the traditional basket method (e.g. USP,<br />
2009; Ph. Eur., 2009) via micro dissolution method to the 96-well plates was successful.<br />
Main focus in the dissolution testing has been in the quality control <strong>and</strong> the several stages<br />
<strong>of</strong> drug development (FDA, 1997a). The large amount <strong>of</strong> drug compound needed has <strong>of</strong>ten<br />
limited the use <strong>of</strong> dissolution testing at the drug discovery stage, which the new methods<br />
now enable.<br />
Alongside with the regulatory dissolution methods, the small scale dissolution methods<br />
can be used to study the results, e.g., at the initial states (0-5 min) <strong>of</strong> the dissolution<br />
process. <strong>Dissolution</strong> pr<strong>of</strong>iles from the early stage <strong>of</strong> drug discovery process can be used to<br />
improve the knowledge on lead c<strong>and</strong>idates. Further, while more complicated, hard to get<br />
<strong>and</strong> expensive media, e.g. HIF, can be utilized with these small scale methods, if needed.<br />
41
5.2 <strong>Solubility</strong> screening<br />
5.2.1 HTS 96-well plate methods (III, IV)<br />
Both kinetic (III) <strong>and</strong> equilibrium (IV) 96-well plate methods were developed in this thesis<br />
work. The fact that amphiphilic compounds decrease the surface tension (γ) <strong>of</strong> water<br />
linearly within a concentration range limited by air/water partitioning at the lower<br />
concentration end, <strong>and</strong> solubility/critical micelle concentration (CMC) at the higher end,<br />
was utilized in the development <strong>of</strong> the kinetic method. The solubility value can be<br />
determined from these isotherms, if the compound is more likely to precipitate than to<br />
form micelles. The phenomenon is described by the classical Gibbs adsorption isotherm:<br />
Γ = −<br />
1 ∂γ<br />
RT ∂ ln c<br />
where Γ is surface excess, R is gas constant, T is temperature <strong>and</strong> c is concentration.<br />
The measured Gibbs isotherm for ibupr<strong>of</strong>en is shown in Figure 7. The solubility <strong>of</strong><br />
ibupr<strong>of</strong>en can be obtained from the focal point between the sloping part <strong>and</strong> the level part<br />
<strong>of</strong> the isotherm. After this point, solid <strong>and</strong> solvated phases are in equilibrium <strong>and</strong> the<br />
amount <strong>of</strong> the free compound is constant in the solution. At higher concentrations, all<br />
material exceeding the solubility limit will be in the solid phase, which has no effect on<br />
the surface tension. Figure 7 also shows the effect <strong>of</strong> pH on the isotherm <strong>of</strong> ibupr<strong>of</strong>en. The<br />
solubility <strong>of</strong> ibupr<strong>of</strong>en is lower at pH 1.2 (0.19 mg/ml) than in pure water (0.23 mg/ml).<br />
This is seen from the shift <strong>of</strong> the adsorption curve.<br />
42<br />
(8)
Figure 7 Gibbs isotherm for ibupr<strong>of</strong>en in aqueous solution. Surface tension is reduced with<br />
increasing ibupr<strong>of</strong>en concentration until a solubility limit is reached. By lowering the<br />
pH decreases the solubility <strong>of</strong> ibupr<strong>of</strong>en, which is seen from the transition point at<br />
lower concentration (n=12, methanol was used as a presolvent). Modified from paper<br />
III.<br />
The slope <strong>of</strong> the linear part just before the transition point is inversely proportional to the<br />
cross-sectional area <strong>of</strong> the compound at the interface (Suomalainen et al., 2004) (Figure<br />
7). With ibupr<strong>of</strong>en the molecular area differed between the two pHs; the absolute value <strong>of</strong><br />
the slope for ibupr<strong>of</strong>en at pH 6.8 was 43.0 corresponding the cross-sectional area per<br />
molecule 0.0096 Å 2 , compared to the absolute value <strong>of</strong> the slope (17.9) <strong>and</strong> cross-sectional<br />
area (0.023 Å 2 ) at pH 1.2. With indomethacin the slopes <strong>and</strong> cross-sectional areas were<br />
similar in both media. These differencies may be due to the different interactions between<br />
solvent <strong>and</strong> solute molecule.<br />
A linear dependency was found between the solubilities in the microtensiometer <strong>and</strong><br />
shake-flask measurements for ibupr<strong>of</strong>en (Figure 8). Best fit <strong>of</strong> the linear curve gave a<br />
slope <strong>of</strong> 1.16 with an intercept at around 180 μg/ml. Differences in the solubility between<br />
the samples were clearly detected, <strong>and</strong> the correlation was at a reasonably good level for<br />
fast screening purposes <strong>and</strong>, thus, clear distinction between low <strong>and</strong> high solubility<br />
environments is possible. The indomethacin results did not fall on the same fitted line as<br />
the ibupr<strong>of</strong>en results, which is likely due to the differences in the precipitation rates <strong>of</strong> the<br />
two compounds. Still, also with indomethacin it was possible to distinguish between the<br />
high <strong>and</strong> low solubility conditions. Not at all unexpectedly, for both the model compounds<br />
the solubilities measured by the surface tension method were consistently higher than<br />
those from the (equilibrium) shake-flask method − due to the co-solvent effect (Fligge <strong>and</strong><br />
Schuler, 2006). This is common feature to all HTS kinetic methods (Pan et al., 2001; Chen<br />
et al., 2002b; Bard et al., 2008) (Figures 8 <strong>and</strong> 9).<br />
43
Figure 8 Measured ibupr<strong>of</strong>en solubilities with the novel surface tension method (y-axis) as a<br />
function <strong>of</strong> the solubilities from the shake-flask measurements (x-axis). Modified from<br />
paper III.<br />
Several variables were tested in order to study the effects <strong>of</strong> different solvent systems on<br />
the reliability <strong>and</strong> applicability <strong>of</strong> the microtensiometry measurements. Studied media<br />
environments included, e.g., electrolytes, <strong>and</strong> surfactants. NaCl <strong>and</strong> CaCl2 (0.9% <strong>and</strong> 2%)<br />
had only limited effects on the solubility <strong>of</strong> ibupr<strong>of</strong>en. Surfactants, sodium dodecyl<br />
sulphate (SDS) <strong>and</strong> Tween 80, increased the solubility slightly below their CMC values,<br />
but the increase was much more evident above the CMC values. The pH effect on the<br />
acidic model compounds was as expected; the higher the pH, the higher the solubility.<br />
Most importantly, the presence <strong>of</strong> other materials in the medium did not disturb the<br />
determinations.<br />
44
Figure 9 <strong>Drug</strong> solubilities by the surface tension (black, n=8-12) <strong>and</strong> shake-flask (grey, n=2)<br />
measurements; effect <strong>of</strong> the dissolution media. Modified from paper III.<br />
Most <strong>of</strong> the HTS methods measure the kinetic solubility when the lead compound is at<br />
dissolved state in the beginning <strong>of</strong> the measurement <strong>and</strong> the co-solvent(s) may<br />
overestimate the solubility (Chen et al., 2002b). The HTS equilibrium solubility method<br />
was developed to avoid that, so in the starting point the compound was in the solid form.<br />
The results from the equilibrium 96-well plate correlated well (R 2 =0.931) with the values<br />
from the shake-flask method over five orders <strong>of</strong> magnitude in aqueous solubility with<br />
twenty model drugs (Figure 10) (IV). The relative deviation <strong>of</strong> the 96-well plate results<br />
was always less than 7-fold. Relative overestimation by the 96-well plate method was the<br />
most significant for compounds with solubility values lower than 10 μg/ml (circled area in<br />
Figure 10). Still, the 96-well plate method gave reliable solubility pr<strong>of</strong>iling: 98% <strong>of</strong> the<br />
tests were within the predictive confidence intervals <strong>of</strong> 95%.<br />
45
Figure 10 The solubilities from shake-flask (y-axis) <strong>and</strong> equilibrium HTS methods (x-axis).<br />
Modified from paper IV.<br />
Despite the medium, the methods gave similar values. Even 10.000-fold changes in<br />
solubility were detected depending on the medium, but the correlation between the<br />
equilibrium HTS <strong>and</strong> shake-flask methods remained good. For example, the solubility <strong>of</strong><br />
dicl<strong>of</strong>enac sodium at pH 1.2 was 0.002 mg/ml <strong>and</strong> at pH 6.8 15.3 mg/ml using the HTS<br />
method; the corresponding values obtained by the shake-flask method were 0.003 mg/ml<br />
<strong>and</strong> 24.0 mg/ml, respectively. The solubilities were determined accurately over a wide<br />
solubility scale (0.002–169.2 mg/ml), which is a true advantage <strong>of</strong> the 96-well plate<br />
method. Compounds with vastly different solubilities can be determined with the method<br />
at small drug quantities <strong>and</strong> utilize method in the early drug discovery stage when only<br />
small quantities <strong>of</strong> compounds are available. Equilibrium 96-well plate method from<br />
Alelyunas et al. (2009) included HPLC, which might slow down the process <strong>and</strong>,<br />
traditionally, the equilibrium solubility has not been measured until the early phase <strong>of</strong><br />
drug development stage (Lipinski, 2003).<br />
Also the kinetic HTS method (microtensiometry) is suitable for drug solubility<br />
screening purposes at the drug discovery stage. Although the method has few limitations,<br />
e.g. it might be difficult to distinguish between the precipitation <strong>and</strong> micelle formation<br />
behavior or some very soluble compounds might not sufficiently partition to the water/air<br />
interface, it has several benefits compared to other kinetic HTS methods. The method does<br />
not require st<strong>and</strong>ard curves or filtrations, solid residues do not disturb the determinations,<br />
large concentration scales are measurable <strong>and</strong> low UV absorption <strong>of</strong> poorly soluble drugs<br />
does not matter (Bevan <strong>and</strong> Lloyd, 2000; Pan et al., 2001). It can be stated that the method<br />
is very well utilizable for poorly soluble drugs, which partition sufficiently to the water/air<br />
interface. Kinetic HTS method was also used to measure the solubility <strong>of</strong> ibupr<strong>of</strong>en as a<br />
46
function <strong>of</strong> time. Samples were remeasured after 5 hours, in which time the solubility<br />
values did not change too much indicating that overestimation was more due to the cosolvent<br />
effect than supersaturation. The dissolution <strong>of</strong> ibupr<strong>of</strong>en was fast <strong>and</strong> the<br />
dissolution pr<strong>of</strong>ile as a function <strong>of</strong> time just ascertained that dissolution had taken place. It<br />
is compound specific, how fast the dissolution takes place <strong>and</strong> the equilibrium solubility is<br />
reached.<br />
Further, ibupr<strong>of</strong>en <strong>and</strong> indomethacin results from both the 96-well plate methods were<br />
compared to each other at aqueous media pH 1.2 <strong>and</strong> 6.8. In both cases the 96-well plate<br />
methods gave maximum 7-fold higher solubility values compared to the shake-flask<br />
method. Only when using ethanol or methanol as presolvents the kinetic solubility values<br />
were more than 7-fold higher than the shake-flask solubility values (Figure 8). The<br />
overestimation by the HTS 96-well plate methods compared to the shake-flask method<br />
could be due to the different scales more than due to the partitioning during the kinetic <strong>and</strong><br />
equilibrium solubility testings.<br />
5.2.2 Polymorphic changes (IV)<br />
During the shake-flask solubility tests, changes in the solid-state form <strong>of</strong> piroxicam were<br />
recognized; the solubility <strong>of</strong> piroxicam was first increased in the aqueous solutions, after<br />
which it was decreased to the equilibrium level. Also the color <strong>of</strong> the solid residue was<br />
changed during the test from white (anhydrate) to yellow (monohydrate) indicating the<br />
change in the physical state <strong>of</strong> the material. The changes were confirmed by XRPD <strong>and</strong><br />
DSC techniques. Piroxicam raw material contains α- <strong>and</strong> β-forms (both anhydrates), but<br />
crystalline solid residues <strong>of</strong> piroxicam monohydrate <strong>and</strong> β-form (anhydrate) were formed<br />
in the solubility test at 5h time-point (Kojic-Prodic et al., 1982; Bordner, 1984; Vrecer et<br />
al., 2003) (Figure 11). For the other model compounds, no polymorphic changes were<br />
noticed during the shake-flask determinations.<br />
47
Figure 11 XRPD diffraction patterns <strong>of</strong> piroxicam in the shake-flask measurements. The<br />
diffractograms represent the solid residue recrystallized from methanol (starting<br />
material in HTS), raw material (starting material in shake-flask), pure piroxicam<br />
monohydrate <strong>and</strong> anhydrated α- <strong>and</strong> β-forms.<br />
In the equilibrium HTS method (pseudo)polymorphic changes occurred with four model<br />
drugs; piroxicam, indomethacin, dicl<strong>of</strong>enac sodium <strong>and</strong> spironolactone (Table 11) (Figure<br />
11). From XRPD <strong>and</strong> DSC results it was obvious that the starting materials in the shakeflask<br />
<strong>and</strong> HTS methods were not <strong>of</strong> the same form. Also spironolactone is known to<br />
crystallize from different solvents (e.g. methanol, acetonitrile, ethanol, ethyl acetate <strong>and</strong><br />
benzene) in different polymorphic forms (El-dalsh et al., 1983; Beckstead et al., 1993).<br />
During the HTS solubility tests no polymorphic changes were, however, observed.<br />
Although some materials had (pseudo)polymorphic changes during the solubility testing,<br />
these changes were not problematic for the interpretation <strong>of</strong> the results, because both the<br />
shake-flask <strong>and</strong> 96-well plate methods do measure the equilibrium solubility <strong>and</strong> the<br />
possible transformations are easily observed. Still, possible polymorphic changes <strong>of</strong> drug<br />
c<strong>and</strong>idates may have an influence on the solubility results in solubility screening <strong>and</strong>, later<br />
on, on the drug development (Bauer et al., 2001).<br />
48
Table 11. Starting <strong>and</strong> raw material results from the XRPD (polymorphic form) <strong>and</strong> DSC<br />
(melting temperatures) measurements (n=3). Modified from paper IV.<br />
Compound XRPD DSC Reported melting<br />
temperature<br />
Dicl<strong>of</strong>enac sodium<br />
raw material (Shake-flask)<br />
starting material (HTS)<br />
Indomethacin<br />
raw material (Shake-flask)<br />
starting material (HTS)<br />
Piroxicam<br />
raw material (Shake-flask)<br />
starting material (HTS)<br />
I, III<br />
II, III<br />
γ<br />
α, γ<br />
α, β<br />
monohydrate<br />
49<br />
288.3 o C<br />
285.5 o C<br />
160.0 o C<br />
154.0 o C, 160.5 o C<br />
202.4 o C<br />
201.5 o C<br />
a Llinas et al., 2007; b Chen et al., 2002a; c Vrecer et al., 2003<br />
5.3 Human intestinal fluid (HIF)<br />
5.3.1 Characterization <strong>of</strong> fasted state HIF (IV)<br />
II=180.5 o C a<br />
γ=160-161 o C b<br />
α=152-154 o C b<br />
α=202.6 o C c<br />
β=199.7 o C c<br />
monohydr.=202.0 o C c<br />
HIF was collected from five healthy volunteers’ duodenum. Inter- <strong>and</strong> intraindividual<br />
variations were observed in the study (Tables 12 <strong>and</strong> 13), like in the earlier studies<br />
(Lindahl et al., 1997; Clarysse et al., 2009b). The mean pH value measured from the<br />
pooled HIF was 6.24 compared to values from 6.6 to 7.5 <strong>and</strong> the mean osmolality was 205<br />
mOsmkg -1 compared to values <strong>of</strong> 200–271 mOsmkg -1 (Table 6).<br />
The mean <strong>of</strong> the total bile salt concentration in fasted HIF was 3.63 mM, which is in<br />
line with the earlier literature values from 2.0 to 3.5 mM (Table 6). Instead, the<br />
phospholipid concentration (1.8 mM, median 1.1 mM) was higher than in the earlier<br />
studies (0.2–0.6 mM) (Table 6) because <strong>of</strong> one sample, which was very concentrated. The<br />
samples vary between individuals <strong>and</strong> individual concentrations <strong>of</strong> up to 2.7 mM have<br />
been found in fasted upper small intestine earlier (Clarysse et al., 2009b). The mean<br />
cholesterol level <strong>of</strong> HIF was 1.8 mM, which was close to the level found by Mansbach et<br />
al. (1975) (1.5 mM).
Table 12. Summary <strong>of</strong> volunteers’ (V1-5) characteristics <strong>and</strong> human intestinal fluid<br />
collection from duodenum.<br />
Volunteer Gender Age BMI<br />
(kg/m 2 Volume pH Osmolality<br />
) (ml)<br />
(mOsm/kg)<br />
V1 Male 25 21.7 70 6.62 176<br />
V2 Female 39 20.8 100 5.19 219<br />
V3 Male 37 23.1 100 6.60 209<br />
V4 Female 24 21.6 40 6.27 213<br />
V5 Female 31 20.9 35 6.51 207<br />
Mean<br />
- 29 21.6 69<br />
6.24 205<br />
±s.d.<br />
8<br />
0.9 31<br />
0.60 17<br />
Table 13. Composition <strong>of</strong> fasted HIF from five healthy volunteers (V1-5). In the table<br />
DC=deoxycholate, C=cholate, CDC=chenodeoxycholate <strong>and</strong><br />
UDC=ursodeoxycholate<br />
Bile<br />
salts (mM)<br />
Phospholipids<br />
(mM)<br />
Cholesterol<br />
(mM)<br />
50<br />
DC<br />
(%)<br />
C<br />
(%)<br />
CDC<br />
(%)<br />
V1 2.45 1.35 1.48 29.5 42.5 23.8 4.2<br />
V2 2.62 1.11 1.68 24.6 56.7 18.1 0.6<br />
V3 4.54 2.18 2.27 3.4 64.6 25.3 6.7<br />
V4 1.65 0.99 1.71 0.8 57.7 35.4 6.1<br />
V5 8.60 4.69 1.48 7.1 66.6 25.6 0.7<br />
5.3.2 HIF in solubility testing (IV)<br />
UDC<br />
(%)<br />
Equilibrium HTS solubility tests were performed using the fasted state HIF (pH 6.24) for<br />
13 model drugs <strong>and</strong> the results were compared to the HTS solubility results in the aqueous<br />
buffer solution at pH 6.8 (Figure 12) – resulting in a fair correlation (R 2 =0.630). More<br />
importantly, the absolute values <strong>of</strong> compound solubility differed only modestly (less than<br />
6.9-fold in all cases), even though 1000-fold range <strong>of</strong> solubilities were studied. Slightly<br />
increased solubility values in HIF compared to the buffer solution pH 6.8 were seen for<br />
piroxicam, naproxen, rifampicin <strong>and</strong> dicl<strong>of</strong>enac sodium. Vice versa, the solubilities in HIF
compared to pH 6.8 buffer solution were decreased for ibupr<strong>of</strong>en <strong>and</strong> ketopr<strong>of</strong>en. No<br />
major differences were seen between the solubility in buffer solution pH 6.8 <strong>and</strong> HIF<br />
suggesting that buffer is an adequate medium when studying dissolution <strong>and</strong> solubility<br />
with compounds in the 100 μg/ml–100mg/ml solubility range. However, we can not rule<br />
out the possibility that more significant solubility differences would be seen for the<br />
compounds with solubility below 100 μg/ml. For example, preliminary solubility test <strong>of</strong><br />
beclomethasone dipropionate, showed 150 times higher solubility in HIF (0.6 μg/ml) than<br />
in the buffer at pH 6.8 (0.004 μg/ml). Obviously, more detailed studies are needed to<br />
determine the limits <strong>and</strong> utilization potential <strong>of</strong> HIF in drug solubility <strong>and</strong> dissolution<br />
testing, especially in the case <strong>of</strong> the compounds with very low solubility. It is worth<br />
noting, however, that the simple buffer is useful medium over surprisingly wide solubility<br />
range that extends at least to 10 2 μg/ml.<br />
HIF was, found to be a suitable solubility/dissolution medium for the 96-well plate<br />
testings. The small volumes make it possible to utilize the HIF also as a medium in drug<br />
discovery process, which may have an even more pronounced importance for very low<br />
solubility compounds. HIF is the most relevant solubility medium, <strong>and</strong> in this system it<br />
can be used instead <strong>of</strong> the buffers or simulated intestinal fluids, if needed.<br />
Figure 12 The solubilities in HIF (y-axis) <strong>and</strong> buffer solution pH 6.8 (x-axis). Modified from<br />
paper IV.<br />
51
6 Conclusions<br />
The main conclusions drawn based on the results <strong>of</strong> this study are:<br />
1. Channel flow method was introduced to the dissolution testing. The method is<br />
applicable on drug dissolution testing <strong>and</strong> enables especially kinetic follow-up<br />
at the very beginning <strong>of</strong> the dissolution process. The results from the channel<br />
flow method <strong>and</strong> the intrinsic dissolution method correlated well with each<br />
other. Later the channel flow method has been further developed to study the<br />
solvent-mediated solid-state transformation during dissolution (Aaltonen et al.,<br />
2006; Lehto et al., 2008).<br />
2. A micro dissolution method was used to determine equilibrium solubilities <strong>of</strong><br />
model compounds with smaller volumes than by the shake-flask method. The<br />
solubility results from the micro dissolution were reliable indicating utilization<br />
potential <strong>of</strong> the micro dissolution method instead <strong>of</strong> shake-flask method.<br />
<strong>Miniaturization</strong> was successful to the 96-well plates.<br />
3. A 96-well plate method was developed for measuring the equilibrium drug<br />
solubilities <strong>and</strong> dissolution pr<strong>of</strong>iles as a function <strong>of</strong> time in microlitre scale.<br />
The method is useful for screening purposes at early stages <strong>of</strong> drug<br />
development. Also polymorphic changes <strong>of</strong> compounds can be studied. The<br />
methodology developed is the first true equilibrium solubility measuring HTS<br />
system utilizing 96-well plates, meaning that in the beginning the compound<br />
was in solid form.<br />
4. A novel surface tension based microtensiometric method was presented for<br />
kinetic HTS <strong>of</strong> drug solubility properties. The method was tested by<br />
determining physiologically relevant parameters on the solubility <strong>of</strong> model<br />
compounds. The method is well suited for classifying purposes <strong>and</strong> solubility<br />
screening tests.<br />
5. HIF from duodenum was used as a medium for the first time in 96-well plate<br />
equilibrium solubility testings. Microliter volumes make it possible to utilize<br />
HIF, the most relevant solubility medium, in dissolution testing at drug<br />
discovery stage. This may be even more important in the future with very low<br />
solubility compounds. Still, the aqueous buffer solutions are the first choice<br />
due to good prediction capacity, low costs, ease <strong>of</strong> access <strong>and</strong> good correlation<br />
with model drugs.<br />
52
7 References<br />
Aaltonen, J., Heinänen, P., Peltonen, L., Kortejärvi, H., Tanninen, V.P., Christiansen, L.,<br />
Hirvonen, J., Yliruusi, J., Rantanen, J., 2006. In situ measurement <strong>of</strong> solvent-mediated<br />
phase transformation during dissolution testing. J Pharm Sci 95: 2730-2737.<br />
Aaltonen, J., Rades, T., 2009. Towards physico-relevant dissolution testing: The<br />
importance <strong>of</strong> solid-state analysis in dissolution. Dissol Tech 16: 47-54.<br />
Achanta, A.S., Gray, V.A., Cecil, T.L., Grady, L.T., 1995. Evaluation <strong>of</strong> the performance<br />
<strong>of</strong> prednisone <strong>and</strong> salicylic acid USP dissolution calibrators. <strong>Drug</strong> Dev Ind Pharm 21:<br />
1171-1182.<br />
Alelyunas, Y.W., Liu, R., Pelosi-Kilby, L., Shen, C., 2009. Application <strong>of</strong> a dried-DMSO<br />
rapid throughput 24-h equilibrium solubility in advancing discovery c<strong>and</strong>idates. Eur J<br />
Pharm Sci 37: 172-182.<br />
Allen, F.H., 2002. The Cambridge Structural Database: a quarter <strong>of</strong> a million crystal<br />
structures <strong>and</strong> rising. Acta Cryst B58: 380-388.<br />
Alsenz, J., Kansy, M., 2007. High throughput solubility measurement in drug discovery<br />
<strong>and</strong> development. Adv <strong>Drug</strong> Del Rev 59: 546-567.<br />
Alsenz, J., Meister, E., Haenel, E., 2007. Development <strong>of</strong> a partially automated solubility<br />
screening (PASS) assay for early drug development. J Pharm Sci 96: 1748-1762.<br />
Amidon, G.L., Lennernäs, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a<br />
biopharmaceutic drug classification: the correlation <strong>of</strong> in vitro drug product dissolution<br />
<strong>and</strong> in vivo bioavailability. Pharm Res 12: 413-420.<br />
Arm<strong>and</strong>, M., Borel, P., Pasquier, B., Dubois, C., Senft, M., Andre, M., Peyrot, J.,<br />
Salducci, J., Lairon, D., 1996. Physicochemical characteristics <strong>of</strong> emulsions during fat<br />
digestion in human stomach <strong>and</strong> duodenum. Am J Physiol 271: G171-G183.<br />
Artursson, P., Palm, K., Luthman, K., 2001. Caco-2 monolayers in experimental <strong>and</strong><br />
theoretical predictions <strong>of</strong> drug transport. Adv <strong>Drug</strong> Deliv Rev 46: 27-43.<br />
Ashford, M., 2007. Gastrointestinal tract – physiology <strong>and</strong> drug absorption. In: Aulton’s<br />
pharmaceutics. The design <strong>and</strong> manufacture <strong>of</strong> medicines, Aulton, M.E. (Ed),<br />
Churchill Livingstone Elsevier, pp. 270-285.<br />
Avdeef, A., 2007. <strong>Solubility</strong> <strong>of</strong> sparingly-soluble ionisable drugs. Adv <strong>Drug</strong> Del Rev 59:<br />
568-590.<br />
Avdeef, A., Berger, C.M., Brownell, C., 2000. pH-metric solubility. 2: Correlation<br />
between the acid-base titration <strong>and</strong> the saturation shake-flask solubility-pH methods.<br />
Pharm Res 17: 85-89.<br />
Back, D.J., Rogers, S.M., 1987. Review: first-pass metabolism by the gastrointestinal<br />
mucosa. Aliment Pharmacol Ther 1: 339-357.<br />
Bakatselou, V., Oppenheim, R.C., Dressman, J.B., 1991. Solubilization <strong>and</strong> wetting<br />
effects <strong>of</strong> bile salts on the dissolution <strong>of</strong> steroids. Pharm Res 8: 1461-1469.<br />
Balakin, K.V., Ivnenkov, Y.A., Skorenko, A.V., Nikolsky, Y.V., Savchuk, N.P.,<br />
Ivashchenko, A.A., 2004. In silico estimation <strong>of</strong> DMSO solubility <strong>of</strong> organic<br />
compounds for bioscreening. J Biomol Screen 9:22-31.<br />
Banaker, U.V., 1992. Pharmaceutical dissolution testing. Marcel Dekker Inc., New York.<br />
Bard, B., Martel, S., Carrupt, P.-A., 2008. High throughput UV method for the estimation<br />
<strong>of</strong> thermodynamic solubility <strong>and</strong> the determination <strong>of</strong> the solubility in biorelevant<br />
media. Eur J Pharm Sci 33: 230-40.<br />
Bauer, J., Spanton, S., Henry, R., Quick, J., Dziki, W., Porter, W., Morris, J., 2001.<br />
Ritonavir: An extraordinary example <strong>of</strong> conformational polymorphism. Pharm Res 18:<br />
859-866.<br />
53
Beckett, A.H., Quach, T.T., Kurs, G.S., 1996. Improved hydrodynamics for USP<br />
apparatus 2. Dissol Tech 3: 7-18.<br />
Beckstead, H.D., Neville, G.A., Shurvell, H.F., 1993. Differentiation <strong>of</strong> solvated<br />
spironolactone samples by FT-raman <strong>and</strong> FT-IR diffuse reflectance spectroscopy.<br />
Fresenius J Anal Chem 345: 727-732.<br />
Benet, L.Z., Amidon, G.L., Barends, D.M., Lennernäs, H., Polli, J.E., Shah, V.P.,<br />
Stavchansky, S.A., Yu, L.X., 2008. The use <strong>of</strong> BDDCS in classifying the permeability<br />
<strong>of</strong> marketed drugs. Pharm Res 52: 483-488.<br />
Berger, C.M., Tsinman, O., Voloboy, D., Lipp, D., Stones, S., Avdeef, A., 2007.<br />
Technical note: Miniaturized intrinsic dissolution rate (Mini-IDR TM ) measurement <strong>of</strong><br />
grise<strong>of</strong>ulvin <strong>and</strong> carbamazepine. Dissol Tech 14: 39-41.<br />
Bergström, C.A.S., Norinder, U., Luthman, K., Artursson, P., 2002. Experimental <strong>and</strong><br />
computational screening models for prediction <strong>of</strong> aqueous drug solubility. Pharm Res<br />
19: 182-188.<br />
Bevan, C.D., Lloyd, R.S., 2000. A high-throughput screening method for the<br />
determination <strong>of</strong> aqueous drug solubility using laser nephelometry in microtiter plates.<br />
Anal Chem 72: 1781-87.<br />
Blanquet, S., Zeijdner, E., Beyssac, E., Meunier, J.-P., Denis, S., Havenaar, R., Alric, M.,<br />
2004. A dynamic artificial gastrointestinal system for studying the behavior <strong>of</strong> orally<br />
administered drug dosage forms under various physiological conditions. Pharm Res<br />
21: 585-591.<br />
Blasko, A., Leahy-Dios, A, Nelson, W.O., Austin, S.A., Killion, R.B., Visor, G.C.,<br />
Massey, I.J., 2001. Revisiting the solubility concept <strong>of</strong> pharmaceutical compounds.<br />
Monatsh Chem 132: 789-798.<br />
Bordner, J., 1984. Piroxicam monohydrate: a zwitterionic form, C15H13N3O4S.H2O. Acta<br />
Cryst C40: 989-990.<br />
Brewster, M.E., L<strong>of</strong>tsson, T., 2007. Cyclodextrins as pharmaceutical solubilizers. Adv<br />
<strong>Drug</strong> Deliv Rev 59: 645-666.<br />
Brouwers, J., Tack, J., Lammert, F., Augustjins, P., 2006. Intraluminal drug <strong>and</strong><br />
formulation behaviour <strong>and</strong> integration in in vitro permeability estimation: a case study<br />
with amprenavir. J Pharm Sci 95: 372-383.<br />
Brown, C.A., Compton, R.G., Narramore, C.A., 1993. The kinetics <strong>of</strong> calcite<br />
dissolution/precipitation. J Colloid Interface Sci 160: 372-379.<br />
Cakiryildiz, C., Mehta, P.J., Rahmen, W., Schoenleber, D., 1975. <strong>Dissolution</strong> studies with<br />
a multichannel continuous-flow apparatus. J Pharm Sci 64: 1692-1697.<br />
Castellari, C., Ottani, S., 1997. Two monoclinic forms <strong>of</strong> dicl<strong>of</strong>enac acid. Acta Cryst C53:<br />
794-797.<br />
Charman, W.N., Rogge, M.C., Boddy, A.W., Berger, B.M., 1993. Effect <strong>of</strong> food <strong>and</strong> a<br />
monoglyceride emulsion formulation on danazol bioavailability. J Clin Pharmacol 33:<br />
381-387.<br />
Chen, X., Morris, K.R., Griesser, U.J., Byrn, S.R., Stowell, J.G., 2002a. Reactivity<br />
differences <strong>of</strong> indomethacin solid forms with ammonia gas. J Am Chem Soc 124:<br />
15012-15019.<br />
Chen, T.-M., Shen, H., Zhu, C., 2002b. Evaluation <strong>of</strong> a method for high throughput<br />
solubility determination using a multi-wavelength UV plate reader. Comb Chem High<br />
Throughput Screen 5: 575-81.<br />
Chen, X.-Q., Venkatesh, S., 2004. Miniature device for aqueous <strong>and</strong> non-aqueous<br />
solubility measurements during drug discovery. Pharm Res 21: 1758-61.<br />
Cheng, C.-L., Yu, L.X., Lee, H.-L., Yang, C.-Y., Lue, C.-S., Chou, C.-H., 2004.<br />
Biowaiver extension potential to BCS class II high solubility-low permeability drugs:<br />
54
idging evidence for metformin immediate-release tablet. Eur J Pharm Sci 22: 297-<br />
304.<br />
Clarysse, S., Psachoulias, D., Brouwers, J., Tack, J., Annaert, P., Duchateau, G., Reppas,<br />
C., Augustijns, P., 2009a. Postpr<strong>and</strong>ial changes in solubilizing capacity <strong>of</strong> human<br />
intestinal fluids for BCS class II drugs. Pharm Res 26: 1456-1466.<br />
Clarysse, S., Tack, J., Lammert, F., Duchateau, G., Reppas, C., Augustijns, P., 2009b.<br />
Postpr<strong>and</strong>ial evolution in composition <strong>and</strong> characteristics <strong>of</strong> human duodenal fluids in<br />
different nutritional states. J Pharm Sci 98: 1177-1192.<br />
Compton, R.G., Pilkington, M.B.G., Stearn, G.M., 1988. Mass transport in channel<br />
electrodes: the application <strong>of</strong> the backward implicit method to electrode reactions (EC,<br />
ECE <strong>and</strong> DISP) involving coupled homogenous kinetics. J Chem Soc Faraday Trans 1:<br />
2155-2171.<br />
Compton, R.G., Harding, M.S., Pluck, M.R., Atherton, J.H., Brennan, C.M., 1993.<br />
Mechanisms <strong>of</strong> solid/liquid interfacial reactions. The dissolution <strong>of</strong> benzoic acid in<br />
aqueous solution. J Phys Chem 97: 10416-10420.<br />
Dai, W.-G., Pollock-Dove, C., Dong, L.C., Li, S., 2008. Advanced screening assays to<br />
rapidly identify solubility-enhancing formulations: high-throughput, miniaturization<br />
<strong>and</strong> automation. Adv <strong>Drug</strong> Deliv Rev 60: 657-72.<br />
Davis, W.W., Parke, T.V., 1942. A nephelometric method for determination <strong>of</strong> solubilities<br />
<strong>of</strong> extremely low order. J Am Chem Soc 64: 101-107.<br />
Davis, J.M., Rowley, S.D., Braine, H.G., Piantadosi, S., Santos, G.W., 1990. Clinical<br />
toxicity <strong>of</strong> cryopreserved bone marrow graft infusion. Blood 75: 781-786.<br />
Dehring, K.A., Workman, H.L., Miller, K.D., M<strong>and</strong>agere, A., Pole, S.K., 2004.<br />
Automated robotic liquid h<strong>and</strong>ling/laser-based nephelometry system for high<br />
throughput measurement <strong>of</strong> kinetic aqueous solubility. J Pharm Biomed Anal 36: 447-<br />
56.<br />
Di, L., Kerns, E.H., 2006. Biological assay challenges from compound solubility:<br />
strategies for bioassay optimization. <strong>Drug</strong> Discovery Today 11:446-451.<br />
Dideborg, O., Dupont, L., 1972. La structure crystalline et moleculaire de la<br />
spironolactone (7α-acetylthio-3-oxo-17α-4-pregnene-21,17βcarbolactone). Acta Cryst<br />
B28: 3014-3022.<br />
Dokoumetzidis, A., Macheras, P., 2006. A century <strong>of</strong> dissolution research: From Noyes<br />
<strong>and</strong> Whitney to the biopharmaceutics classification system. Int J Pharm 321: 1-11.<br />
Dressman, J.B., Berardi, R.R., Dermentzoglou, L.C., Russell, T.L., Schmaltz, S.P.,<br />
Barnett, J.L., Jarvenpaa, K.M., 1990. Upper gastrointestinal (GI) pH in young, healthy<br />
men <strong>and</strong> women. Pharm Res 7: 756-761.<br />
Dressman, J.B., Amidon, G.L., Reppas, C., Shah, V., 1998. <strong>Dissolution</strong> testing as a<br />
prognostic tool for oral drug absorption: Immediate release dosage forms. Pharm Res<br />
15: 11-22.<br />
El-dalsh, S.S., El-Sayed, A.A., Badawi, A.A., 1983. Studies on spironolactone<br />
polymorphic forms. <strong>Drug</strong> Dev Ind Pharm 9: 877-894.<br />
EMEA, 2002. Note for Guidance on the Investigation <strong>of</strong> Bioavailability <strong>and</strong><br />
Bioequivalence. Committee for Medicinal Products for Human Use (CHMP).<br />
EMEA, 2008. Annex 7 to note for evaluation <strong>and</strong> recommendation <strong>of</strong> pharmacopoeial<br />
texts for use in the ICH regions on dissolution test general chapter. Committee for<br />
Medicinal Products for Human Use (CHMP).<br />
European Pharmacopoeia, 2009, 6 th edition, Druckerei C.H. Beck, Nördlingen.<br />
Fallingborg, J., Christensen, L.A., Ingeman-Nielsen, M., Jacobsen, B.A., Abildgaard, K.,<br />
Rasmussen, H.H., 1989. pH-pr<strong>of</strong>ile <strong>and</strong> regional transit times <strong>of</strong> the normal gut<br />
measured by a radiotelemetry device. Aliment Pharmacol Therap 3: 605-613.<br />
55
FDA, 1997a. Guidance for industry: <strong>Dissolution</strong> testing <strong>of</strong> immediate release solid oral<br />
dosage forms, retrieved November 2009, from http://www.fda.gov/downloads<br />
/<strong>Drug</strong>s/GuidanceComplianceRegulatoryInformation/Guidances/ucm070237.pdf<br />
FDA, 1997b. Guidance for industry: Extended release oral dosage forms: development,<br />
evaluation <strong>and</strong> application <strong>of</strong> in vitro/in vivo correlations, retrieved November 2009,<br />
from http://www.fda.gov/downloads/<strong>Drug</strong>s/GuidanceComplianceRegulatoryInformati<br />
on/Guidances/UCM070239.pdf<br />
FDA, 2000. Guidance for industry: Waiver <strong>of</strong> in vivo bioavailability <strong>and</strong> bioequivalence<br />
studies for immediate release solid oral dosage forms based on a biopharmaceutics<br />
classification system, retrieved November 2009, from http://www.fda.gov/ downloads<br />
/<strong>Drug</strong>s/GuidanceComplianceRegulatoryInformation/Guidances/ucm070246.pdf<br />
Fini, A., Fazio, G., Fern<strong>and</strong>ez-Hervas, M.J., Holgado, M.A., Rabasco, A.M., 1995.<br />
Influence <strong>of</strong> crystallization solvent <strong>and</strong> dissolution behaviour for a dicl<strong>of</strong>enac salt. Int J<br />
Pharm Sci 121: 1926.<br />
Fleischer, D., Li, C., Zhou, Y., Pao, L.-H., Karim, A., 1999. <strong>Drug</strong>, meal <strong>and</strong> formulation<br />
interactions influencing drug absorption after oral administration. Clin Pharmacokinet<br />
36: 233-254.<br />
Fligge, T.A., Schuler, A., 2006. Integration <strong>of</strong> a rapid automated solubility classification<br />
into early validation <strong>of</strong> hits obtained by high throughput screening. J Pharm Biomed<br />
Anal 42: 449-54.<br />
Galia, E., Nicolaides, E., Hörter, D., Löbenberg, R., Reppas, C., Dressman, J.B., 1998.<br />
Evaluation <strong>of</strong> various dissolution media for predicting in vivo performance <strong>of</strong> class I<br />
<strong>and</strong> II drugs. Pharm Res 15: 698-705.<br />
Garbacz, G., Wedemeyer, R.-S., Nagel, S., Giessmann, T., Mönnikes, H., Wilson, C.G.,<br />
Siegmund, W., Weitschies, W., 2008. Irregular absorption pr<strong>of</strong>iles observed from<br />
dicl<strong>of</strong>enac extended release tablets can be predicted using a dissolution test apparatus<br />
that mimics in vivo physical stresses. Eur J Pharm Sci 70: 421-423.<br />
Garbacz, G., Blume, H., Weitschies, W., 2009. Investigation <strong>of</strong> the dissolution<br />
characteristics <strong>of</strong> nifedipine extended-release formulations using USP apparatus 2 <strong>and</strong><br />
a novel dissolution apparatus. Dissol Tech 16: 7-13.<br />
Glomme, A., März, J., Dressman, J.B., 2005. Comparison <strong>of</strong> a miniaturized shake-flask<br />
solubility method with automated potentiometric acid/base titrations <strong>and</strong> calculated<br />
solubilities. J Pharm Sci 94: 1-16.<br />
Gu, L., Huynh, O., Becker, A., Peters, S., Nguyen, H., Chu, N., 1987. Preformulation<br />
selection <strong>of</strong> a proper salt for a weak acid-base (RS-82856) – a new positive inotropic<br />
acid. <strong>Drug</strong> Dev Ind Pharm 13: 437-448.<br />
Gurantz, D., Laker, M.F., H<strong>of</strong>fman, A.F., 1981. Enzymatic measurement <strong>of</strong> cholinecontaining<br />
phospholipids in bile. J Lipid Res 22: 373-376.<br />
Hauss, D.J., 2007. Oral lipid-based formulations. Adv <strong>Drug</strong> Deliv Rev 59: 667-676.<br />
Havenaar, R., Menikus, M., 1994. In vitro model <strong>of</strong> an in vivo digestive tract. JP, US,<br />
European Patent PCT/NL, 93/00225.<br />
Higuchi, T., 1961. Rate <strong>of</strong> release <strong>of</strong> medicaments from ointment bases containing drugs<br />
in suspension. J Pharm Sci 50: 874-875.<br />
Hill, J.W., Petrucci, R.H., 1999. General Chemistry, 2 nd edition, Prentice Hall, New<br />
Jersey.<br />
Ho, N.F.H., Merkle, H.P., Higuchi, W.I., 1983. Quantitative, mechanistic <strong>and</strong><br />
physiologically realistic approach to the biopharmaceutical design <strong>of</strong> oral drug<br />
delivery systems. <strong>Drug</strong> Dev Ind Pharm 9: 1111-1184.<br />
Hoelke, B., Gieringer, S., Arlt, M., Saal, C., 2009. Comparison <strong>of</strong> nephelometric, UVspectroscopic,<br />
<strong>and</strong> HPLC methods for high-throughput determination <strong>of</strong> aqueous drug<br />
solubility in microtiter plates. Anal Chem 81: 3165-3172.<br />
56
Homon, C.A., Nelson R.M., 2006. High-throughput screening: Enabling <strong>and</strong> influencing<br />
the process <strong>of</strong> drug discovery. In: The process <strong>of</strong> new drug discovery <strong>and</strong><br />
development, Smith, C.G., O’Donnell, J.T. (Eds.), 2 nd edition, Informa healthcare,<br />
New York, pp. 79-102.<br />
Int.Ph. 2009, 4 th edition, retrieved November 2009 from http://apps.who.int/phint/en/p<br />
/docf<br />
Jaiboon, N., Yos-In, K., Ruangchaithaweesuk, S., Chaichit, N., Thutivoranath, R.,<br />
Siritaedmukul, K., Hannongbua, S., 2001. New orthorhombic form <strong>of</strong> 2-((2,6dichlorophenyl)amino)<br />
benzeneacetic acid (dicl<strong>of</strong>enac acid). Anal Sci 17: 1465-1466.<br />
Jantratid, E., Janssen N., Reppas, C., Dressman, J., 2008. <strong>Dissolution</strong> media simulating<br />
conditions in the proximal human gastrointestinal tract: An update. Pharm Res 25:<br />
1663-1676.<br />
Japanese Pharmacopoeia 2009, 15 th edition, retrieved November 2009 from<br />
http://jpdb.nihs.go.jp/jp15e/<br />
Johnson, K.C., Swindell, A.C., 1996. Guidance in the setting <strong>of</strong> drug particle size<br />
specifications to minimize variability in absorption. Pharm Res 13: 1795-1798.<br />
Kalantzi, L., Goumas, K., Kalioras, V., Abrahamsson, B., Dressman, JR., Reppas, C.,<br />
2006a. Characterization <strong>of</strong> the human upper gastrointestinal contents under conditions<br />
simulating bioavailability/bioequivalence studies. Pharm Res 23: 165-176.<br />
Kalantzi, L., Persson, E., Polentarutti, B., Abrahamsson, B., Goumas, K., Dressman, J.,<br />
Reppas, C., 2006b. Canine intestinal contents vs. simulated media for the assessment<br />
<strong>of</strong> solubility <strong>of</strong> two weak bases in the human small intestinal contents. Pharm Res 23:<br />
1373-1381.<br />
Kawakami, K., 2009. Current status <strong>of</strong> amorphous formulation <strong>and</strong> other special dosage<br />
forms as formulations for early clinical phases. J Pharm Sci 98: 2875-2885.<br />
Kerns, E.H., 2001. High throughput physicochemical pr<strong>of</strong>iling for drug discovery. J<br />
Pharm Sci 90: 1838-1858.<br />
Kibbey, C.E., Poole, S.K., Robinson, B., Jackson, J.D., Durham, D., 2001. An integrated<br />
process for measuring the physicochemical properties <strong>of</strong> drug c<strong>and</strong>idates in a<br />
preclinical discovery environment. J Pharm Sci 90: 1164-1175.<br />
Kistenmacher, T.J., Marsh, R.E., 1972. Crystal <strong>and</strong> molecular structure <strong>of</strong> an<br />
antiinflammatory agent, indomethacin, 1-(p-chlorobenzoyl)-5-methoxy-2methylindole-3-acetic<br />
acid. J Am Chem Soc 94: 1340-1345.<br />
Kojic-Prodic, B., Ruzic-Toros, Z., 1982. Structure <strong>of</strong> the anti-inflammatory drug 4hydroxy-2-methyl-N-2-pyridyl-2H-1λ<br />
6 ,2-benzothiazine-3-carboxamide 1,1-dioxide<br />
(piroxicam). Acta Cryst B38: 2948-2951.<br />
Kortejärvi, H., Urtti, A., Yliperttula, M., 2007. Pharmacokinetic simulation <strong>of</strong> biowaiver<br />
criteria: The effects <strong>of</strong> gastric emptying, dissolution, absorption <strong>and</strong> elimination rates.<br />
Eur J Pharm Sci 30: 155-166.<br />
Kortejärvi, H., Shawahna, R., Koski, A., Malkki, J., Ojala, K., Yliperttula, M., 2010. Very<br />
rapid dissolution is not needed to guarantee bioequivalence for biopharmaceutics<br />
classification system (BCS) I drugs. J Pharm Sci 99: 621-625.<br />
Kortejärvi, H., Yliperttula, M., Dressman, J.B., Junginger, H.E., Midha, K.K., Shah, V.P.,<br />
Barends, D.M., 2005. Biowaiver monographs for immediate release solid oral dosage<br />
forms: ranitidine hydrochloride. J Pharm Sci 94: 1617-1625.<br />
Kramer, C., Heinisch, T., Fligge, T., Beck, B., Clark, A., 2009. A consistant dataset <strong>of</strong><br />
kinetic solubilities for early-phase drug discovery. J Med Chem 4: 1529-1536.<br />
Lehto, P., Aaltonen, J., Niemelä, P., Rantanen, J., Hirvonen, J., Tanninen, V.P., Peltonen,<br />
L., 2008. Simultaneous measurement <strong>of</strong> liquid-phase <strong>and</strong> solid-phase transformation<br />
kinetics in rotating disc <strong>and</strong> channel flow cell dissolution devices. Int J Pharm 363: 66-<br />
72.<br />
57
Lindahl, A., Ungell, A.-L., Knutson, L., Lennernäs L., 1997. Characterization <strong>of</strong> fluids<br />
from the stomach <strong>and</strong> proximal jejunum in men <strong>and</strong> women. Pharm Res 14: 497-502.<br />
Lindenberg, M., Wieg<strong>and</strong>, C., Dressman, J., 2005. Comparison <strong>of</strong> the adsorption <strong>of</strong><br />
several drugs to typical filter materials. Dissol Tech 12: 22-25.<br />
Lipinski, C.A., 2000. <strong>Drug</strong>-like properties <strong>and</strong> the causes <strong>of</strong> poor solubility <strong>and</strong> poor<br />
permeability. J Pharmacol Toxicol Methods 44: 235-249.<br />
Lipinski, C., 2003. Aqueous solubility in discovery, chemistry, <strong>and</strong> assay changes. In:<br />
<strong>Drug</strong> bioavailability; estimation <strong>of</strong> solubility, permeability, absorption <strong>and</strong><br />
bioavailability, van de Waterbeemd, H., Lennernäs, H., Artursson, P. (Eds.), Wiley-<br />
VCH Verlag GmbH, Weinheim, pp. 215-231.<br />
Lipinski, C.A., 2004. <strong>Solubility</strong> in water <strong>and</strong> DMSO: Issues <strong>and</strong> potential solutions. In:<br />
Biotechnology: Pharmaceutical aspects, vol. 1, Borchardt, R.T., Kerns, E.H., Lipinski,<br />
C.A., Thakler, D.R., Wang, B. (Eds.), American association <strong>of</strong> pharmaceutical<br />
sciences, Arlington VA, pp. 93-125.<br />
Lipinski, C.A., Lombardo, F., Dominy, B.W., Feeney, P.J., 1997. Experimental <strong>and</strong><br />
computational approaches to estimate solubility <strong>and</strong> permeability in drug discovery<br />
<strong>and</strong> development settings. Adv <strong>Drug</strong> Deliv Rev 23: 3-25.<br />
Llinas, A., Burley, J.C., Box, K.J., Glen, R.C., Goodman J.M., 2007. Dicl<strong>of</strong>enac<br />
solubility: Independent determination <strong>of</strong> the intrinsic solubility <strong>of</strong> three crystal forms. J<br />
Med Chem 50: 979-983.<br />
L<strong>of</strong>tsson, T., Hreinsdottir, D., 2006. Determination <strong>of</strong> aqueous solubility by heating <strong>and</strong><br />
equilibration. AAPS Pharm Sci Tech 7: E1-E4.<br />
Lorimer, J.W., Cohen-Adad, R., 2003. Thermodynamics <strong>of</strong> solubility. In: The<br />
experimental determination <strong>of</strong> solubilities, Hefter, G.T., Tomkins, R.P.T. (Eds.),<br />
volume 6, John Wiley <strong>and</strong> Sons Ltd, Chichester, pp. 17-76.<br />
Lütjohann, D., Hahn, C., Prange, W., Sudhop, T., Axelson, M., Sauerbruch, T., von<br />
Bergmann, K., Reichel, C., 2004. Influence <strong>of</strong> rifampin on serum markers <strong>of</strong><br />
cholesterol <strong>and</strong> bile acid synthesis in men. Int J Clin Pharmacol Ther 42: 307-313.<br />
Mansbach, M.C., Cohen, R.S., Leff, P.B., 1975. Isolation <strong>and</strong> properties <strong>of</strong> the mixed lipid<br />
micelles present in intestinal content during fat digestion in man. J Clin Invest 56: 781-<br />
791.<br />
Martinez, M.N., Amidon, G.L., 2002. A mechanistic approach to underst<strong>and</strong>ing the<br />
factors affecting drug absorption: A review <strong>of</strong> fundamentals. J Clin Pharmacol 42:<br />
620-643.<br />
Mathus-Vliegen, E.M.H., van Ierl<strong>and</strong>-van Leeuwen, M.L., Terpstra, A., 2004. Lipase<br />
inhibition by orlistat: effects on gall-bladder kinetics <strong>and</strong> cholecystokinin release<br />
obesity. Aliment Pharmacol Ther 19: 601-611.<br />
Mauger, J.W., 1996. Physicochemical <strong>and</strong> fluid mechanical factors related to dissolution<br />
testing. Dissol Tech 3: 7-11.<br />
Mauger, J., Ballard, J., Brockson, R., De, S., Gray, V., Robinson, D., 2003. Intrinsic<br />
dissolution performance testing <strong>of</strong> the USP dissolution apparatus 2 (rotating paddle)<br />
using modified salicylic acid calibrator tablets: pro<strong>of</strong> <strong>of</strong> principle. Dissol Tech 10: 6-<br />
15.<br />
Mithani, S., Bakatselou, V., TenHoor, C.N., Dressman, J.B., 1996. Estimation <strong>of</strong> the<br />
increase in solubility <strong>of</strong> drugs as a function <strong>of</strong> bile salt concentration. Pharm Res 13:<br />
163-167.<br />
Moreno de la Cruz, M.P., Oth, M., Deferme, S., Lammert, F., Tack, J., Dressman, J.,<br />
Augustijns, P., 2006. Characterization <strong>of</strong> fasted-state human intestinal fluids collected<br />
from duodenum <strong>and</strong> jejunum. J Pharm Pharmacol 58: 1079-1089.<br />
Muangsin, N., Prajuabsook, M., Chimsook, P., Chantarasiri, N., Siraleartmukul, K.,<br />
Chaichit, N., Hannongbua, S., 2004. Structure determination <strong>of</strong> dicl<strong>of</strong>enac in a<br />
58
dicl<strong>of</strong>enac-containing chitosan matrix using conventional X-ray powder diffraction<br />
data. J Appl Crystallogr 37: 288-294.<br />
Nicklasson, M., Brodin, A., 1984. The relationship between intrinsic dissolution rates <strong>and</strong><br />
solubilities in the water-ethanol binary solvent system. Int J Pharm 18: 149-156.<br />
Nicklasson, M., Brodin, A., Nyqvist, H., 1981. Studies on the relationship between<br />
solubility <strong>and</strong> intrinsic rate <strong>of</strong> dissolution as a function <strong>of</strong> pH. Acta Pharm Suec 18:<br />
119-128.<br />
Nicklasson, M., Fyhr, P., Magnusson, A.-B., Gunnvald, K. A., 1988. Preformulation study<br />
on the in vitro dissolution characteristics <strong>of</strong> the organophosphorus poisoning antidote<br />
HI-6. Int J Pharm 46: 247-254.<br />
Noyes, A.A., Whitney, W.R., 1897. The rate <strong>of</strong> solution <strong>of</strong> solid substances in their own<br />
solutions. J Am Chem Soc 19: 930-934.<br />
O’Donnell, J.R., Burnett, A.K., Sheehan, T., Tansey, P., McDonald, G.A., 1981. Safety <strong>of</strong><br />
dimethylsulphoxide. Lancet 28: 498.<br />
Omelczuk, M.O., McGinity, J.W., 1993. The influence <strong>of</strong> thermal treatment on the<br />
physical-mechanical <strong>and</strong> dissolution properties <strong>of</strong> tablets containing poly(DL-lactide<br />
acid). Pharm Res 10: 542-548.<br />
Oprea, T.I., Davis, A.M., Teague, S.J., Leeson, P.D., 2001. Is there a difference between<br />
leads <strong>and</strong> drugs? A historical perspective. J Chem Inf Comput Sci 41: 1308-1315.<br />
Pan, L., Ho, Q., Tsutsui, K., Takahashi, L., 2001. Comparison <strong>of</strong> chromatographic <strong>and</strong><br />
spectroscopic methods used to rank compounds for aqueous solubility. J Pharm Sci 90:<br />
521-29.<br />
Pernarowski, M., Woo, W., Searl, R.O., 1968. Continuous flow apparatus for the<br />
determination <strong>of</strong> the dissolution characteristics <strong>of</strong> tablets <strong>and</strong> capsules. J Pharm Sci 57:<br />
1419-1421.<br />
Persson, E.M., Gustafsson, A., Carlsson, A.S., Nilsson, R.G., Knutson, L., Forsell, P.,<br />
Hanisch, G., Lennernäs, H., Abrahamsson, B., 2005. The effects <strong>of</strong> food on the<br />
dissolution <strong>of</strong> poorly soluble drugs in human <strong>and</strong> in model small intestinal fluids.<br />
Pharm Res 22: 2141-2151.<br />
Persson, A.M., Baumann, K., Sundelöf, L.-O., Lindberg, W., Sokolowski, A., Pettersson,<br />
C., 2008. Design <strong>and</strong> characterization <strong>of</strong> a new miniaturized rotating disk equipment<br />
for in vitro dissolution rate studies. J Pharm Sci 97: 3344-3355.<br />
Porter, C.J.H., Charman, W.N., 2001. Lipid-based formulation for oral administration:<br />
opportunities for bioavailability enhancement <strong>and</strong> lipoprotein targeting <strong>of</strong> lipophilic<br />
drugs. In <strong>Drug</strong> targeting technology: physical, chemical, biological methods, vol. 115,<br />
Schreier, H., (Eds.), Marcel Dekker Inc., New York, pp. 85-130.<br />
Pudipeddi, M., Serajuddin, A.T.M., 2005. Trends in solubility <strong>of</strong> polymorphs. J Pharm Sci<br />
94: 929-939.<br />
Qureshi, S.A., McGilveray, I.J., 1999. Typical variability in drug dissolution testing: study<br />
with USP <strong>and</strong> FDA calibrator tablets <strong>and</strong> a marketed drug (glibenclamide) product.<br />
Eur J Pharm Sci 7: 249-258.<br />
Qureshi, S.A., Shabnam, J., 2003. Applications <strong>of</strong> new device (spindle) for improved<br />
characterization <strong>of</strong> drug release (dissolution) <strong>of</strong> pharmaceutical products. Eur J Pharm<br />
Sci 19: 291-297.<br />
Rapoport, A.P., Rowe, J.M., Packman, C.H., Ginsberg S.J., 1991. Cardiac arrest after<br />
autologous marrow infusion. Bone Marrow Transplant 7: 401-403.<br />
Röst, M., Quist, P.-O., 2003. <strong>Dissolution</strong> <strong>of</strong> USP prednisone calibrator tablets effects <strong>of</strong><br />
stirring conditions <strong>and</strong> particle size distribution. J Pharm Biomed Anal 31: 1129-1143.<br />
Santos, N.C., Figueira-Coelho, J., Martins-Silva, J., Saldanha, C., 2003. Multidisciplinary<br />
utilization <strong>of</strong> dimethyl sulfoxide: pharmacological, cellular <strong>and</strong> molecular aspects.<br />
Biochem Pharmacol 65: 1035-1041.<br />
59
Savolainen, M., Kogermann, K., Heinz, A., Aaltonen, J., Peltonen, L., Strachan, C.,<br />
Yliruusi, J., 2009. Better underst<strong>and</strong>ing <strong>of</strong> dissolution behaviour <strong>of</strong> amorphous drugs<br />
by in situ solid-state analysis using Raman spectroscopy. Eur J Pharm Biopharm 71:<br />
71-79.<br />
Serajuddin, A.T.M., 2007. Salt formation to improve drug solubility. Adv <strong>Drug</strong> Del Rev<br />
59: 603-616.<br />
Shah, A.C., Peot, C.B., Ochs, J.F., 1973. Design <strong>and</strong> evaluation <strong>of</strong> a rotating filterstationary<br />
basket in vitro dissolution test apparatus I: Fixed fluid volume system. J<br />
Pharm Sci 62: 671-677.<br />
Sheng, J.J., McNamara D.P., Amidon, G.L., 2009. Toward an in vivo dissolution<br />
methodology: A comparison <strong>of</strong> phosphate <strong>and</strong> bicarbonate buffers. Mol Pharm 6: 29-<br />
39.<br />
Shepherd, R.E., Price, J.C., Luzzi, L.A., 1972. <strong>Dissolution</strong> pr<strong>of</strong>iles for capsules <strong>and</strong> tablets<br />
using a magnetic basket dissolution apparatus. J Pharm Sci 61: 1152-1156.<br />
Singh, T., Dutt, J., 1985. Cyclic voltammetry at the turbular graphite electrode. Reversible<br />
process (theory). J Electroanal Chem 190: 65-73.<br />
Souliman, S., Blanquet, S., Beyssac, E., Cardot, J.-M., 2006. A level A in vitro/in vivo<br />
correlation in fasted <strong>and</strong> fed states using different methods: Applied to solid immediate<br />
release oral dosage forms. Eur J Pharm Sci 27: 72-79.<br />
Stenberg, P., Luthman, K., Artursson, P., 2000. Virtual screening <strong>of</strong> intestinal drug<br />
permeability. J Control Release 65: 231-243.<br />
Stroncek, D.F., Fautsch, S.K., Lasky, L.C., Hurd, D.D., Ramsay, N.K.C., McCullough, J.,<br />
1991. Adverse reactions in patients transfused with cryopreserved marrow.<br />
Transfusion 31: 521-526.<br />
Stuart, M., Box, K., 2005. Chasing equilibrium: Measuring the intrinsic solubility <strong>of</strong> weak<br />
acids <strong>and</strong> bases. Anal Chem 77: 983-990.<br />
Styler, M.J., Topolsky, D.L., Crilley, P.A., Covalesky, V., Bryan, R., Bulova, S., Brodsky,<br />
I., 1992. Transient high grade heart block following autologous bone marrow infusion.<br />
Bone Marrow Transplant 10: 435-438.<br />
Sunesen, V.H., Pedersen, B.L., Kristensen, H.G., Müllertz, A., 2005. In vivo in vitro<br />
correlations for a poorly soluble drug, danazol, using the flow-through dissolution<br />
method with biorelevant dissolution media. Eur J Pharm Sci 24: 305-313.<br />
Suomalainen, P., Johans, C., Söderlund, T., Kinnunen, P.K.J., 2004. Surface activity<br />
pr<strong>of</strong>iling <strong>of</strong> drugs applied to the prediction <strong>of</strong> blood-brain barrier permeability. J Med<br />
Chem 47: 1783-88.<br />
Szejtli, J., 1988. Cyclodextrin technology. Kluwer Academic Publishers, Dordrecht.<br />
Takano, R., Sugano, K., Higashida, A., Hayashi, Y., Machida, M., Aso, Y., Yamashita, S.,<br />
2006. Oral absorption <strong>of</strong> poorly water-soluble drugs: computer simulation <strong>of</strong> fraction<br />
absorbed in humans from a miniscale dissolution test. Pharm Res 23: 1144-56.<br />
Taub, M., Kristensen, L., Frokjaer, S., 2002. Optimized conditions for MDCK<br />
permeability <strong>and</strong> turbidimetric solubility studies using compounds representative <strong>of</strong><br />
BCS classes I-IV. Eur J Pharm Sci 15: 331-340.<br />
Tenho, M., Aaltonen, J., Heinänen, P., Peltonen, L., Lehto, V.-P., 2007. Effect <strong>of</strong> texture<br />
on the intrinsic dissolution behaviour <strong>of</strong> acetylsalicylic acid <strong>and</strong> tolbutamide compacts.<br />
J Appl Cryst 40: 857-864.<br />
Tsuji, A., Tamai, I., 1996. Carrier-mediated intestinal transport <strong>of</strong> drugs. Pharm Res 13:<br />
963-977.<br />
Uesu, N.Y., Pineda, E.A.G., Hechenleitner, A.A.W., 2000. Microcrystalline cellulose<br />
from soybean husk: Effects <strong>of</strong> solvent treatments on its properties as acetylsalicylic<br />
acid carrier. Int J Pharm 206: 85-96.<br />
60
USP 2003/2009, 26 th /32 nd edition, United States Pharmacopoeial Convention, Rockville<br />
MD.<br />
Vertzoni, M., Dressman, J., Butler, J., Hempenstall, J., Reppas, C., 2005. Simulation <strong>of</strong><br />
fasting gastric conditions <strong>and</strong> its importance for the in vivo dissolution <strong>of</strong> lipophilic<br />
compounds. Eur J Pharm Biopharm 60: 413-417.<br />
Vrecer, F., Vrbinc, M., Meden, A., 2003. Characterization <strong>of</strong> piroxicam crystal<br />
modifications. Int J Pharm 256: 3-15.<br />
WHO, 2006. Proposal to waive in vivo bioequivalence requirements for WHO model list<br />
<strong>of</strong> essential medicines immediate-release, solid oral dosage forms, in WHO Technical<br />
report series, N:o 937.<br />
Wood, J.H., Syarto, J.E., Letterman, H.J., 1965. Improved holder for intrinsic dissolution<br />
rate study. J Pharm Sci 54: 1068.<br />
Wu, C.-Y., Benet, L.B., 2005. Predicting drug disposition via application <strong>of</strong> BCS:<br />
Transport/Absorption/Elimination interplay <strong>and</strong> development <strong>of</strong> a biopharmaceutics<br />
drug disposition classification system. Pharm Res 22: 11-23.<br />
Yazdanian, M., Briggs, K., Yankovsky, C., Hawi, A., 2004. The ‘high solubility’<br />
definition on the current FDA guidance on biopharmaceutical classification system<br />
may be too strict for acidic drugs. Pharm Res 21: 293-299.<br />
Yu, L.X., Amidon, G.L., Polli, J.E., Zhao, H., Mehta, M.U., Conner, D.P., Shah, V.P.,<br />
Lesko, L.J., Chen, M.-L., Lee, V.H.L., Hussain, A.S., 2002. Biopharmaceutics<br />
classification system: The scientific basis for biowaiver extensions. Pharm Res 19:<br />
921-925.<br />
Yu, L.X., Carlin, A.S., Amidon, G.L., Hussain, A.S., 2004. Feasibility studies <strong>of</strong> utilizing<br />
disk intrinsic dissolution rate to classify drugs. Int J Pharm 270: 221-227.<br />
Zhou, L., Yang, L., Tilton, S., Wang, J., 2007. Development <strong>of</strong> a high throughput<br />
equilibrium solubility assay using miniaturized shake-flask method in early drug<br />
discovery. J Pharm Sci 96: 3052-71.<br />
61