Biotechnology Annual Review, Volume 12 (Biotechnology Annual Review)

ScienceDirect - Biotechnology Annual Review, Volume 12, Pages 1-405 (2006) /RJLQ + 5HJLVWHU %URXJKWWR\RXE\ 8...

Author: M. Raafat El-Gewely

176 downloads 1319 Views 5MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ScienceDirect - Biotechnology Annual Review, Volume 12, Pages 1-405 (2006) /RJLQ + 5HJLVWHU

%URXJKWWR\RXE\ 8QLYHUVLW\RI0HOERXUQH,QIRUPDWLRQ'LYQ

%URZVH

6HDUFK

0\6HWWLQJV

$OHUWV

+HOS

+RPH

4XLFN6HDUFK

7LWOHDEVWUDFWNH\ZRUGV

-RXUQDOERRNWLWOH

$XWKRU

--This Journal/Book--

HJMVVPLWK

9ROXPH

,VVXH

3DJH

%LRWHFKQRORJ\$QQXDO5HYLHZ

$OHUWPHDERXWQHZ%RRN9ROXPHV

&RS\ULJKW(OVHYLHU%9$OOULJKWVUHVHUYHG

$GGWR)DYRULWHV

6KRUWFXW85/WRWKLVSDJHKWWSZZZVFLHQFHGLUHFWFRPH]SUR[\OLEXQLPHOEHGXDXVFLHQFHERRNVHULHV

$UWLFOH)HHG

)XOOWH[WDYDLODEOH 1RQVXEVFULEHG :KDWGRHVWKLVPHDQ" 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH SS 9ROXPH3DJHV

Ż3UHYLRXVYROLVV1H[WYROLVVŹ

(GLWHGE\05DDIDW(O*HZHO\ ,6%1

$UWLFOH/LVW)XOO $EVWUDFWV

)RUHZRUG3K\VLFDO%LRORJ\7KH1H[W

(GLWRULDO%RDUG 3DJHVL[[ 6XPPDU\3OXV_)XOO7H[W/LQNV_3'). _9LHZ5HODWHG$UWLFOHV

/LVWRIFRQWULEXWRUV 3DJHV[L[LLL 3'). _9LHZ5HODWHG$UWLFOHV

file:///E|/1/ScienceDirect/Biotechnology%20Annual%20Review/12/0000.htm (1 of 3)01.03.08 18:22:07

DUWLFOHV

ScienceDirect - Biotechnology Annual Review, Volume 12, Pages 1-405 (2006)

(PHUJLQJRSWLRQVLQSURWHLQELRVHSDUDWLRQ 3DJHV .DO\DQL0RQGDO6XODNVKDQD-DLQ6XQLWD7HRWLDDQG0XQLVKZDU1*XSWD 6XPPDU\3OXV_)XOO7H[W/LQNV_3'). _9LHZ5HODWHG$UWLFOHV 5HFHQWDGYDQFHVLQDOOSURWHLQFKURPRSKRUHWHFKQRORJ\ 3DJHV 0DUN3UHVFRWW-LRQ0%DWWDG3DVFDO*:LOPDQQ-DPLH5RVVMRKQDQG5RGQH\-'HYHQLVK 6XPPDU\3OXV_)XOO7H[W/LQNV_3'). _9LHZ5HODWHG$UWLFOHV 3XWWLQJWKHµ2PH¶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

,QGH[RIDXWKRUV 3DJH 3'). _9LHZ5HODWHG$UWLFOHV

.H\ZRUGLQGH[ 3DJHV 3'). _9LHZ5HODWHG$UWLFOHV DUWLFOHV

%URZVH

6HDUFK

0\6HWWLQJV

$OHUWV

+HOS

+RPH

file:///E|/1/ScienceDirect/Biotechnology%20Annual%20Review/12/0000.htm (2 of 3)01.03.08 18:22:07

ScienceDirect - Biotechnology Annual Review, Volume 12, Pages 1-405 (2006) $ERXW6FLHQFH'LUHFW_&RQWDFW8V_7HUPV &RQGLWLRQV_3ULYDF\3ROLF\ &RS\ULJKW(OVHYLHU%9$OOULJKWVUHVHUYHG6FLHQFH'LUHFWLVDUHJLVWHUHGWUDGHPDUNRI(OVHYLHU% 9

file:///E|/1/ScienceDirect/Biotechnology%20Annual%20Review/12/0000.htm (3 of 3)01.03.08 18:22:07

v

Foreword

Physical Biology: The Next 50 Years

In bioscience and biotechnology there exist, depending on perspective, different schools of thought concerned with biological processes at different length scales – molecular biology, genetics, and cell biology. Given that biological machines operate in the non-equilibrium state, irrespective of viewpoint we have to understand how the pieces are made, how their physical forces exert control, and how feedback and feedforward of elements allow for robustness and function. Recently, there has been considerable effort to establish ‘‘systems biology’’ as a discipline covering under its umbrella all phenomena concerning function. The output in this case is the so-called ‘‘emergence.’’ Lee Hood, in a previous preface in this series, pointed out that ‘‘systems biology in its simplest terms is the identiﬁcation of the elements in a system and an analysis of their interactions while the system is functioning so as to understand the systems or emergent properties of the system.’’ Emergence is a new addition to the lexicon of biology, and other ﬁelds, but its precise deﬁnition is still amorphous. If proven that ‘‘the whole is greater than the sum of its parts,’’ the signiﬁcance would be in asking why? Elements of biological machines are deﬁned on the scale of macromolecules. With this molecular perspective one is concerned with how the elements of life cells behave and what the driving forces are for such behavior. In many ways, the complexity is similar to that of the whole cell. Elements of macromolecular systems interact, communicate, and deﬁne a nanometer-scale function (or ‘‘emergence’’). This machinery derives its power from the control it exerts, with atomic-scale precision. Protein folding and misfolding is an example. How do the elements of a polymer molecule of thousands of atoms interact to render a native structure capable of a speciﬁc function, and why, when they misbehave, do we contract diseases such as Alzheimer’s? Why do atomic-scale conformational changes of some macromolecular systems, as in prions, give rise to diseases? Only when we examine these systems as they function can we understand their landscapes in space and time and hope to understand the molecular bases of human diseases. New frontiers of research are now directed toward the observation of phenomena as they occur in space and time – in the language of biology, ‘‘visualization’’. With the power of molecular and genetic tools we should be able to visualize molecular motions, understand physical phenomena, and develop new concepts for global function and speciﬁc control. Studies at this level of the physics and chemistry of biology deﬁne the physical forces involved in biological complexity and constitute the foundation of ‘‘physical biology.’’ I believe that physical biology as a new discipline should aim to understand mechanistically how physical forces and interactions govern biological function, from the molecular to the cellular scale. In physical biology, this focus on physical forces of biological structure dynamics is distinct from the aim of mapping the engineering of information and its ﬂow; i.e., the wiring in

vi cells. For all elements, structure and dynamics are basic for understanding the function at the molecular scale, and without visualization they remain dark and elusive. On different timescales, biological structures transform through intermediates and transition states on complex energy landscapes, the surfaces of free energy. The global shape of the surface in nuclear-coordinate space reﬂects the possible conformations (entropy) and multitude of interactions (enthalpy) that could lead to the change; for both, the solvent role cannot be ignored. Unlike simple chemical structures, ﬂexible macromolecular structures have the additional complexity of changes involving numerous possible conformations, with some that are ‘‘active’’ and others that are ‘‘inactive’’ in the biological function. Such non-equilibrium structures, if determined, can provide an understanding of the origin of reduced-coordinate space for the motion, the real enthalpic and entropic contribution to the free energy, and the nature of molecular mutations that lead to diseases. Remarkably, the transformations are controlled by the balance of weak forces such as hydrogen bonding, electrostatic interactions, dispersion, and hydrophobic forces, all having energy on the order of a few kcal/mol. Progress has been made in determining equilibrium, time-averaged structures using X-ray diffraction, for systems of varying complexity. Beginning with the discovery of X-rays near the turn of the 20th century (1895), the increase in the level of complexity is impressive, from diatom salts (NaCl) to DNA, proteins, complex assemblies such as viruses, and more recently the ribosome and RNA polymerase. Of equal importance, after the discovery of the electron in 1897, was the development of electron diffraction/microscopy for structural determination of biological systems. In fact, the ﬁrst membrane protein (water insoluble) crystal structure of bacteriorhodopsin (crystal thickness of about 10 nm) was determined using a combination of electron diffraction and microscopy. Determination of the structure of the photosynthetic reaction center, a membrane protein, was achieved by X-ray diffraction techniques in 1985, and more recently the structure of ion channels was successfully completed. From the three-dimensional (3D) structures determined by electron or X-ray diffraction, one pictures the static spatial arrangements of atoms, but the mechanism for the function cannot be directly unraveled without knowledge of the dynamics of non-equilibrium transient structures. For dynamics, it has become possible in the past two decades to observe atomic motions on a femtosecond timescale, the scale of vibrational periods (femtochemistry and femtobiology). On such timescales the observed coherent nuclear motions deﬁne a fundamental transition, from ensemble-rate kinetics to single-moleculetrajectory dynamics. However, when the systems, molecules or assemblies of them, are those with thousands of atoms and the changes involve many possible conformations, one must not only resolve the temporal behavior but also determine the 3D molecular structures during the change. Such combined atomic-scale resolutions in space and time constitute the basis for a new ﬁeld of study in what we referred to elsewhere as ‘‘Four-dimensional (4D) structural dynamics.’’ In the coming years, the visualization of structural dynamics with electron or X-ray diffraction will enable the integration of structure and dynamics in elucidating

vii the function. For imaging in real time the method of choice is Ultrafast Electron Microscopy (UEM) and for an overview, see Thomas JM ‘‘A revolution in electron microscopy’’ published recently in Angew Chem Int Ed 2005;44:5563. The recent development of 4D UEM in this laboratory provides the ability to image complex structures with the spatial resolution of a TEM, but with timed (femtosecond) singleelectron coherent packets; optical imaging provides much lower spatial resolution, typically hundreds of nanometers. With this 4D microscope, biological cells were imaged and diffraction of materials (substrates) was recorded at a dosage of a few electrons per square angstrom. The major hurdle in the use of electron diffraction and imaging in microscopy was the space-charge problem, and with single-electron coherent packets this problem no longer exists. Moreover, local structures of less than 105 unit cells, as opposed to an average over 1015 unit cells, can be explored and their constituencies imaged in real space. Together with cryo-techniques we expect myriad of applications, from molecules to cells. At the end, deciphering the nature of physical forces and mechanisms at different time and length scales, including those of other building blocks of the cell (e.g., particles, aggregates, and molecular machines), should provide the needed understanding of how the cellular machines behave as functional ‘‘microscopic’’ objects. For some ‘‘macroscopic’’ properties we may not need to know some details about the elements. A crystal with 1023 atoms can be described as symmetry planes of mirrors when interpreting its diffraction of X-rays or electrons, but this says nothing as to how the atoms bond together or why they have ‘‘emergent properties’’ such as the ability of one atomic crystal to catalyze reactions on its surfaces but not on others. Thus, developing group theory for symmetry considerations is important, but the atomic description of the properties is fundamental. Knowing the critical role of the molecular elements in genetics and diseases, and in the emerging area of ‘‘synthetic biology,’’ it is not possible to ignore how these elements behave during the function. As we progress in acquiring new understanding for crossing the chasm, new questions will emerge. Some of these questions are challenging: how does welldeﬁned (classical) function emerge from ﬂuctuating quantum systems; what are the driving forces for the presence of non-equilibrium structures; do complexity and chaos make the whole greater than the sum of its parts, and if so why; how does statistics inﬂuence the behavior in a robust function; and why is the function directed. Despite the enormous progress made so far, we are back to some 50-year-old questions by Schro¨dinger: ‘‘What is [the molecular basis of] life?’’ Hopefully, in the coming 50 years physical biology will illuminate the answer to these and other questions through direct visualization of the elements of biology in space and time. Ahmed Zewail

Acknowledgements The author wishes to thank David Baltimore and Rob Phillips for stimulating discussions.

viii

Biographical Ahmed Zewail is currently the Linus Pauling Chair professor of chemistry and physics at Caltech, and Director of Physical Biology Center and Laboratory for Molecular Sciences. He received the Nobel Prize in 1999.

ix

Editorial Board* Chief Editor M. Raafat El-Gewely, Department of Molecular Biotechnology, Gene Systems Group, Institute of Medical Biology, University of Tromsø, 9037 Tromsø, Norway E-mail: [email protected]

Editors MaryAnn Foote, Abraxis BioScience Inc., 11777 San Vincente Blvd, Suite 550 Los Angeles, California 90049, USA, E-mail: [email protected] Guido Krupp, AmpTec GmbH, Koenigstr. 4A, D-22767 Hamburg, Germany E-mail: [email protected] Alfons Lawen, Monash University, Clayton Campus, Department of Biochemistry and Molecular Biology, Clayton, Victoria 3800, Australia E-mail: [email protected]

Associate Editors Marin Berovic, Faculty of Chemistry and Chemical Engineering, Department of Chemical and Biochemical Engineering, University of Ljubljana, Askercˇeva 5 1001 Ljubljana, Slovenia, E-mail: [email protected] Thomas M.S. Chang, Artiﬁcial Cells & Organs Research Centre, McGill University 3655 Drummond St., Room 1005, Montreal, Quebec, H3G 1Y6, Canada E-mail: [email protected] Thomas T. Chen, Department Molecular and Cellular Biology, University of Connecticut, 91 North Eagleville Rd, Unit 3125, Storrs, CT 06269-3149, USA E-mail: [email protected] Frank Desiere, Nestle´ Research Center, PO Box 44, CH-1000 Lausanne 26 Switzerland, E-mail: [email protected]

*Anyone wishing to publish a contribution in Biotechnology Annual Review should contact the Chief Editor, or a member of the Editorial Board. For the ‘Guide for Authors’, please visit: http://www.elsevier.com/locate/issn/13872656

x Franco Felici, Department of Microbiological, Genetic and Molecular Science University of Messina, Salita Sperone 31, 98166 Messina, Italy E-mail: [email protected] Leodevico (Vic) L. Ilag, HealthLinx Ltd, 576 Swan St, Richmond, Victoria 3121, Australia, E-mail: [email protected] Kuniyo Inouye, Laboratory of Enzyme Chemistry, Division of Food Science and Biotechnology, Graduate School of Agriculture, Kyoto University, Sakyo-ku Kyoto 606-8502, Japan, E-mail: [email protected] Jocelyn H. Ng, 2/6 Jurang Street, Balwyn, Victoria 3103, Australia E-mail: [email protected] Eric Olson, Vertex Pharmaceuticals, 130 Waverly Street, Cambridge, MA 02139 USA, E-mail: [email protected] Vincenzo Romano-Spica, University Institute of Motor Science IUSM Pizza Lauro e Bosis 15, 00194 Rome Italy, E-mail: [email protected]

xi

List of contributors Raju Adhikari, PolyNovo Biomaterials Pty Ltd, Bag 10, Clayton South Bayview Avenue, Clayton 3169, Australia Marie-Isabel Aguilar, Department of Biochemistry & Molecular Biology Monash University, Victoria 3800, Australia Carmen Alvarez-Lorenzo, Departamento de Farmacia y Tecnologı´ a Farmace´utica Facultad de Farmacia, Universidad de Santiago de Compostela, 15782-Santiago de Compostela, Spain Aradhna Anand, Thapar Centre for Industrial Research & Development, Bhadson Road, Patiala, 147 004, Punjab, India Pramod K. Bajpai, Thapar Centre for Industrial Research & Development, Bhadson Road, Patiala, 147 004, Punjab, India Pratima Bajpai, Thapar Centre for Industrial Research & Development, Bhadson Road, Patiala, 147 004, Punjab, India Jion M. Battad, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Thomas Ming Swi Chang, Artiﬁcial Cells & Organs Research, Center, Faculty of Medicine, McGill University, Montreal, Quebec, H3G 1Y6, Canada Timothy J. Cole, Department of Biochemistry and Molecular Biology, Monash University, Clayton, Victoria 3800, Australia Angel Concheiro, Departamento de Farmacia y Tecnologı´ a Farmace´utica Facultad de Farmacia, Universidad de Santiago de Compostela, 15782-Santiago de Compostela, Spain Rodney J. Devenish, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Laetitia Fauconnot, Nestle´ Research Center, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland Laurent B. Fay, Nestle´ Research Center, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland

xii MaryAnn Foote, Abraxis BioScience Inc., 11777 San Vincente Blvd, Suite 550 Los Angeles, California 90049, USA Nancy Gerits, Department of Microbiology and Virology, Institute of Medical Biology, University of Tromsø, N-9037 Tromsø, Norway Martin Grigorov, Nestle´ Research Center, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland Pathiraja Gunatillake, PolyNovo Biomaterials Pty Ltd, Bag 10, Clayton South Bayview Avenue, Clayton 3169, Australia Munishwar N. Gupta, Department of Chemistry, Indian Institute of Technology, Delhi, Hauz Khas, New Delhi 110016, India Tzong-Hsien Lee, Department of Biochemistry & Molecular Biology, Monash University, Victoria 3800, Australia Sulakshana Jain, Department of Chemistry, Indian Institute of Technology, Delhi Hauz Khas, New Delhi 110016, India Zun Chang Liu, Artiﬁcial Cells & Organs Research Center, Faculty of Medicine McGill University, Montreal, Quebec, H3G 1Y6, Canada Roshan Mayadunne, PolyNovo Biomaterials Pty Ltd, Bag 10, Clayton South Bayview Avenue, Clayton 3169, Australia Theresa Mikalsen, Department of Microbiology and Virology, Institute of Medical Biology, University of Tromsø, N-9037 Tromsø, Norway Ugo Moens, Department of Microbiology and Virology, Institute of Medical Biology, University of Tromsø, N-9037 Tromsø, Norway Kalyani Mondal, Department of Chemistry, Indian Institute of Technology, Delhi Hauz Khas, New Delhi 110016, India David M. Mutch, Department of Molecular and Experimental Medicine, The Scripps Research Institute, 10550 North Torrey Pines Road, MEM 275, La Jolla CA 92037, USA Mark Prescott, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Karen Soskin, Global Pharmaceutical Regulatory Affairs, Abbott Laboratories Abbott Park, IL, USA

xiii Jamie Rossjohn, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Sunita Teotia, Department of Chemistry, Indian Institute of Technology, Delhi Hauz Khas, New Delhi 110016, India Pascal G. Wilmann, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Ahmed H. Zewail (Nobel Laureate), NSF Laboratory for Molecular Sciences California Institute of Technology, Mail Code 127-72, 1200 East California Boulevard, Pasadena, CA 91125, USA

1

Emerging options in protein bioseparation Kalyani Mondal, Sulakshana Jain, Sunita Teotia and Munishwar N. Gupta Chemistry Department, Indian Institute of Technology, Delhi, Hauz Khas, New Delhi 110016, India

Introduction Crystallization was the ﬁrst technique that was used for protein separation. In 1926, Sumner obtained the enzyme urease in the form of crystals [1]. This was followed up by crystallization of several enzymes by Northrop and Kunitz [2,3]. About 75 crystalline enzyme preparations were obtained in the next 30 years [4]. It was gradually realized that unlike small molecular weight compounds, crystallization of a protein molecule is possible even if the protein is not pure. Next few decades witnessed the development of numerous chromatographic and electrophoretic methods that can be used for protein puriﬁcation and for establishing protein purity. In the ﬁrst phase of development, the efforts at protein separation involved using principles by which organic compounds are generally isolated and puriﬁed, such as crystallization, extraction and precipitation. Fractional precipitation with salt and organic solvents was extensively used [5]. Two precipitation methods unique to proteins were also discovered. Selective precipitation of a protein by incubation at its isoelectric point and precipitation of thermolabile contaminating proteins (by heating) to obtain relatively more thermostable desired protein in a more puriﬁed form [5]. Introduction of ion-exchange chromatography (especially cellulose-based ion exchangers) [6] and gel ﬁltration media [7] were deﬁnitely two important milestones. The next exciting discovery was that of afﬁnity chromatography. While the concept was around for quite some time, Porath’s description of a general method (by cyanogen bromide activation) of designing an afﬁnity media deﬁnitely provided the necessary boost to its wider applications [8]. As far as an enzymologist in academic sector is concerned, that is where the history of the art of protein separation ends. An enzymologist generally needs to purify an enzyme/protein for establishing its structure, structure–function relationship and mechanism. Development of tools such as HPLC, capillary electrophoresis, mass spectroscopy and microsequencing methods ensured that the amount of puriﬁed protein required by an enzymologist for these purposes is considerably less! Industrial enzymology did require starch-degrading enzymes, enzymes in detergent formulations and enzymes related to food industry in larger amounts [9]. Corresponding author: Tel.: +91-11-2659-1503. Fax: +91-11-2658-1073.

E-mail: [email protected] (M.N. Gupta). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12001-3

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

2 These enzymes were however not required in pure forms. Thus, just precipitation or crystallization was good enough after the initial step. One area wherein need for high purity was felt in these early years was proteases for sequencing purposes [10]. The demand was not much and required purity level was not very high. For example, one needed chymotrypsin free of trypsin and vice versa [11]. Ion-exchange chromatography and afﬁnity chromatography were more than adequate. Interesting enough, it was the change in the upstream scene that demanded serious rethink by downstream processing specialists. Recombinant DNA technology made it possible to create variety of proteins and in large amounts [12]. Tissue culture techniques also contributed to making upstream technology becoming more powerful [13]. Simultaneously, two trends appeared at the societal level. Much better health care was provided and called for expectation. Diagnostic reagents, biosensors, protein-based pharmaceutical formulations (e.g., clot busters) and better drug delivery systems have been the outcome of this. The second trend was increasing concerns about environment. Enzymes are preferred tools in green chemistry and any move to replace chemical processes (dubbed ‘‘ﬁre and sword’’ chemistry) [14] by enzyme-based processes is welcome and is a part of sustainable technology development [15]. The word ‘‘white biotechnology’’ is increasingly appearing in print to describe this trend/approach [14]. A conﬂuence of the two trends has motivated development of some very powerful techniques to tailor enzymes for a speciﬁc purpose. Chemical modiﬁcation [16], chemical crosslinking [17] and protein immobilization [18] have been around for quite some time and given useful results in creating more stable proteins or enzymes with different kinetic properties. However, site-directed mutagenesis [19] and directed evolution [20–22] have already surpassed these older techniques in both promise and performance. A subplot of this story has been the growth of nonaqueous enzymology [23,24]. This not only has made medium engineering a good tool to tailor the media for a speciﬁc purpose, it also created larger range of applications for existing enzymes [25]. Hence, increasingly downstream processing has become an important area of biotechnology. Often it turns out to be the rate-limiting step and most challenging part of a biotechnological process/application. An update on the status of applied enzymology is as follows Today protein/enzymes are required in larger amounts. Hence, large-

scale methods are required. While industrial grade enzymes continue to be used at low purity level,

demand for high-purity proteins is increasing for other applications [26]. Protein engineering products require separation tools that have resolution high enough to separate proteins, which differ in few amino acids!

3 Proteomics require both separation and analysis [27,28]. While earlier approach was to remove inconvenient proteins (e.g., major

proteins in blood for looking at minor proteins for medical diagnosis or antinutritional factors in food and feed), a more ambitious approach is to recover, separate and use these as well [29,30]. The present review describes the way downstream processing scientists are meeting these challenges. While earlier biochemical engineers were content enough to describe models about mass transfer constraints and role of particle size, viscosity etc. with impressive equations, today even some elite biochemical engineering journals publish papers with SDS-PAGE of the puriﬁed protein [31,32]! The ideal situation will be that such publications start appreciating the virtues of a classical concept called protein puriﬁcation table, which ultimately details in cold print the extent of success of a puriﬁcation: how pure and how much of it was obtained! Some notable paradigm shifts The last decade has seen some serious rethink on the strategies for downstream processing methods. Some subtle and others not so subtle, paradigm shifts in the approaches have been as follows. Upstream vs. downstream A simple arithmetic tells us that given any percentage recovery in a unit process, the ﬁnal recovery will be inversely proportional to the number of unit processes used in a puriﬁcation protocol. In other words, an efﬁcient puriﬁcation protocol must minimize the number of steps/unit processes. An often-cited paper in classical enzymology mentioned that ammonium sulphate precipitation-chromatography-afﬁnity chromatography is the most often used sequence of unit processes used by enzymologists while purifying an enzyme [33]. Classical enzymology, in fact, distinguishes between isolation/concentration (sometimes called primary puriﬁcation) and puriﬁcation (or called secondary puriﬁcation by some workers) [34]. Motivated by the desire to reduce the number of steps, this distinction between upstream and downstream processing has got blurred. Three approaches are notable in this regard (A) Afﬁnity tags: Recombinant DNA methodology makes it possible to obtain a desired protein in the form of a fusion protein, which has an afﬁnity tag attached to it. This afﬁnity tag makes it possible to straightaway afﬁnity capture the desired protein. Examples of such afﬁnity tags are polyhistidine, cellulose binding domain, chitin binding domain [26,35].

4 (B) In general, bringing afﬁnity steps right early in the puriﬁcation protocol even with naturally occurring proteins has been tried quite successfully [36,37]. Afﬁnity precipitation and two-phase afﬁnity extractions are two extremely promising methods in this regard. (C) An approach that deserves further development and applications is the immobilization of an afﬁnity ligand on glass beads so that disintegration of microbe and afﬁnity capture is a simultaneous process [38]. All the three approaches, of course, are afﬁnity based. Afﬁnity-based approaches offer higher selectivity and often eliminate the need for multistep protocols. Hence, these approaches bridge the upstream and downstream operations in protein/enzyme production. Dealing with unclarified broth The enzymes/proteins in their crudest form are often obtained as coloured and crude suspensions (e.g., from fermentation/tissue homogenate). Subsequent precipitation/chromatography steps take care of the soluble coloured impurities. Plant tissue homogenates with quinonoid compounds as contaminants especially pose a difﬁcult problem. If not removed or their formation suppressed (by preventing oxidation of polyphenols) right early, the quinones react with side chains of amino acids in proteins and often cause inactivation [39]. Such processes occurring in nature are responsible for the colour of the skin, hair and feathers. The centrifugation or membrane ﬁltration has to be used to obtain clear solutions. These are costly and cumbersome processes. Hence, there is a great need to develop downstream processing, which can directly deal with unclariﬁed broth (and hopefully remove colour as well). Aqueous two-phase extractions (ATPEs) have been considered most promising in this regard. Again, a later discussion in the review would show how an integration of afﬁnity precipitation and ATPE fulﬁll this promise. Expanded bed chromatography that uses stable ﬂuidized beds can also effectively deal with crude suspensions [40,41]. A recent technique called macro-afﬁnity ligand facilitated three phase partitioning (MLFTPP) may be the third candidate in the range of choices in this regard [32]. Chromatographic approaches vs. nonchromatographic techniques There are some inherent virtues of a chromatographic process, which will always remain unique to it. Given multiple theoretical plates as the basis of separation, chromatographic techniques offer very high resolution. Advances in material sciences have helped in development of new generation chromatographic media [42,43]. These offer reduced process time and higher resolution. Chromatographic processes can be automated and even put on a robotic platform [28]. As analytical methods become more sensitive, one

5 requires very small columns for handling small samples and scale up issues become irrelevant. Thus, proteomics has encouraged the development of several commercial designs based upon chromatographic separations [28]. Some of these, with separation libraries in the memory of the computer, promise ‘‘separation for dummies’’ kind of technology. As large number of scientists, well versed in molecular biology, do not wish to entangle with techniques like ammonium sulphate precipitation, these costly devices, aided by sufﬁcient advertisement budget, are becoming increasingly popular. However, apart from being costly; chromatographic protocols are difﬁcult to scale up. So, if the objective is to obtain puriﬁed protein in reasonable amount (and not merely analyze or identify), many nonchromatagraphic processes constitute attractive choices. Affinity ligands, pseudoaffinity ligands and combinatorial chemistry Classically, afﬁnity chromatography was based on the use of afﬁnity ligands to selectively capture the desired protein. These afﬁnity ligands were compounds like competitive inhibitors or coenzyme or antibodies and were chosen because such materials were known to bind to the protein/enzymes. Three discoveries led to the next step in the evolution of the concept of afﬁnity ligand. One of these was the discovery of hydrophobic interaction chromatography (HIC) wherein spacers used in design of afﬁnity chromatography were found to bind to dehydrogenases [44]. The second discovery also relates to dehydrogenases. Some textile dyes (prominent among these was cibacron blue) were found to bind to these enzymes [45]. It was rationalized that such dyes have structural motifs similar to NAD+, which is a common coenzyme for some dehydrogenases [46]. This rationalization is still considered correct but dyes are known to bind to variety of enzymes and this explanation is not valid in many such cases [47]. The third was the discovery of immobilized metal afﬁnity chromatography (IMAC) by Porath [48]. The dyes and chelated metal ions in IMAC have been sometimes referred as pseudoafﬁnity ligands. These HIC ligands and pseudoafﬁnity ligands have no biological or in vivo relationship with the enzyme to which these molecules bind. The next quantum jump in our evolution of the concept of afﬁnity ligands came out of peptide libraries [49,50] and combinatorial synthesis [51]. In both cases, randomly generated structures inherently capable of interacting with the chosen protein molecule, constitute effective afﬁnity ligands. We show in the following that one can search for afﬁnity in pre-existing naturally occurring or synthetic polymers as well. With such strategic thinking, some options have emerged as attractive choices. Most of these techniques are well established and have proved their merit in protein puriﬁcation. The remaining part of the review deals with these techniques, which may be used both at small scale as well as large scale in equally facile manner.

6 Afﬁnity precipitation Afﬁnity precipitation exploits the afﬁnity interactions in free solution by combining an afﬁnity ligand with a target protein. The afﬁnity ligand can be homobifunctional/heterobifunctional. Affinity precipitation using homobifunctional ligands Homobifunctional ligands have multiple binding sites (at least two) for the target protein. On addition, it creates a crosslinked network and gets precipitated. This involves the use of bifunctional ligands, such as nucleotides connected by a spacer (Bis-NAD+). Larsson and Mosbach [52] used BisNAD+ for precipitation of lactate dehydrogenase. This NAD-enzyme complex is dissociated by using NADH and the enzyme is recovered by gel ﬁltration. Similarly, N,N0 -bis-3-(dihydroxyborylbenzene)-adipamide, i.e., BispBA was used to isolate erythrocytes [53]. Affinity precipitation using heterobifunctional ligands Heterobifunctional ligands consist of two components. The ﬁrst is the afﬁnity ligand for the target protein and the second is usually a smart polymer, which is so designed that its solubility is controlled in a predictable way by altering some solvent parameters such as pH, temperature, ionic strength, addition of metal ions, etc. Smart polymers, which have been commonly used in afﬁnity precipitation are listed in Table 1 [54,55]. Linne et al. [56] have used alginate–soybean inhibitor complex for puriﬁcation of trypsin from bovine pancreas. Here, soybean trypsin inhibitor (STI) is afﬁnity ligand for trypsin. In some cases, polymer itself has an afﬁnity for the protein and acts as homobifunctional ligand. Since the target protein has afﬁnity for the afﬁnity macroligand (polymer), it is selectively precipitated along with the afﬁnity macroligand upon the necessary change in the controlling parameter. Target protein is recovered from polymer–protein precipitate using speciﬁc or nonspeciﬁc eluents. The process thus results in concentration and puriﬁcation of the target protein and is amenable to scale up. Fig. 1 explains the principle of afﬁnity precipitation. Afﬁnity precipitation has been applied to many systems successfully which have been listed in Table 2. Aqueous two-phase systems (ATPSs) ATPS, a partition-based downstream process has been widely recognized as an efﬁcient, scalable and a single-step technique, which can directly deal with unclariﬁed extracts/homogenates [80]. ATPSs are composed of two incompatible polymers in water solution (e.g., dextran and poly(ethylene glycol)

7 Table 1. Some commonly used smart polymers. Name

Composition

Alginate

Copolymer of CaCl2opH 2.0 EDTA>pH 2.0 galactouronic and mannuronic acid Copolymer of methacrylic opH 4.7 >pH5.8 acid and methyl methacrylate (in a molar ratio of 1:2) Copolymer of methacrylic opH 4.3 >pH 5.3 acid and methyl methacrylate (in a molar ratio of 1:1) Copolymer of >pH 6.5 opH 5.5 N-acetyl-glucosamine and glucosamine units Polymer of >32 1C o32 1C N-isopropylacrylamide

Eudragit S-100

Eudragit L-100

Chitosan Poly (N-isopropyla crylamide)

ligand

Insoluble

Soluble

polymer

+ Crude mixture of proteins

Affinity capture of target protein

Polymer based smart macro-ligand

Application of stimulus

Separation of precipitate Target protein recovered in purified form

and dissociation

Precipitated proteinaffinity macroligand complex

Fig. 1. Schematic presentation of principle of afﬁnity precipitation.

8 Table 2. Proteins/enzymes puriﬁed using afﬁnity precipitation using heterobifunctional ligand. Polymer Synthetic polymers Polyacrylamide Poly(N-isopropyl acrylamide)

Hydroxypropylmethylcellulose acetate succinate Eudragit S-100

Polymerized liposome Natural polymers Chitosan

Alginate

Galactomannan k-carragenan

Afﬁnity ligand

Protein/enzyme puriﬁed

Reference

p-Aminobezamidine p-Aminobenzamidine Biotin l-Phage DNA

Trypsin Alkaline protease Avidin HindIII restriction endonuclease Protein A

Schneider et al. [57] Pe´cs et al. [58] Morris et al. [59] Maeda et al. [60]

Cibacron blue Cibacron blue Cibacron blue STI Aminophenyl-a-dglucopyranoside Protein A None None

l-lactate dehydrogenase Pyruvate kinase Alcohol dehydrogenase Trypsin Con A

Guoquiang et al. [62] Guoquiang et al. [62] Guoquiang et al. [63] Kumar and Gupta [64] Larsson and Mattiasson [65]

Monoclonal antibody d-lactate dehydrogenase Xylanase

STI

Trypsin

Taipa et al. [66] Guoquiang et al. [67] Gupta et al. [68]; Breccia et al. [69] Sun et al. [70]

None None None None

Wheat germ agglutinin b-Glucosidase Lysozyme Lectin (from rice, tomato and potato) Chitinases Trypsin Endopolygalacturonase a-amylase (wheat germ) b-amylase (sweet potato) glucoamylase Lipase Phospholipase D Human IgG Alcohol dehydrogenase

Human IgG

None STI None None None

Protein A Cibacron blue

Taniguchi et al. [61]

Senstad and Mattiasson [71] Homma et al. [72] Tyagi et al. [73] Tyagi et al. [73] Teotia et al. [74] Linne et al. [56] Gupta et al. [75] Sharma et al. [36] Teotia et al. [76] Teotia et al. [37] Sharma and Gupta, 2001 [171] Sharma et al. [77] Bradshaw and Sturgeon [78] Mondal et al. [79]

(PEG)), or a polymer (e.g., PEG) and a high concentration of salt (e.g., phosphate). Commonly used ATPSs are listed in Table 3. Incompatibility of polymer–polymer and polymer–salt systems, and how it can be exploited for separation of cells, cell particles and proteins was described by Albertsson [81]. The physicochemical properties like protein hydrophobicity, charge and size governs the partitioning in the aqueous systems. The possibilities of denaturation of labile biomolecules are quite less in this technique as the phases contain 80–95% water; and hence, ATPSs results in the puriﬁcation of biomolecules in a native and biological active forms [82]. Some attractive applications of ATPS have been puriﬁcation of prion proteins [83], human antitrypsin (from transgenic milk, [84]), porcine insulin [85], B-phycoerythrin

9 Table 3. Frequently used ATPSs for product recovery.

PEG/dextran [81] PEG/ﬁcoll [81] Dextran/ﬁcoll [81] Methylcellulose/dextran [81] PEG/benzoil dextran [88] PEG/NaCl [89] Ucon/Reppal (hydroxypropyl starch) [81] PEG/sodium citrate [90] PEG/K2HPO4 [81] PEG/ammonium sulphate [81] PEG/starch [91] PEG/cashew nut tree gum [92]

Table 4. Additives used in ATPE for enhancing protein partitioning. 1. NaCl: Affects the ionic strength of the system, which in turns has a signiﬁcant effect on protein partitioning [94,95] 2. Na2SO4: Changes the partition coefﬁcient of the system [96] 3. NaN3/chloroacetamide: Bactericide [96] 4. Affinity ligands: An appropriate afﬁnity ligand increases the partition coefﬁcient and separation factor of the concerned protein [74,97–100]

(a red-coloured protein used as a natural dye from Porphyridium cruentum [86] and GFP (green ﬂuorescent protein overexpressed by a mutant recombinant Escherichia coli. [87]). Partition coefﬁcient in ATPSs In ATPS, phase-forming components are selected depending upon the application. Different salts in the phase system causes K values to be either increased or decreased [93]. Table 4 mentions some of the commonly used additives. In ATPSs, the partitioning of any protein [P] is deﬁned in terms of the partition coefﬁcient, . K ¼ ½Ptop ½Pbottom where [P]top and [P]bottom refer to the concentration of proteins in top and bottom phases, respectively. The partition coefﬁcient has been found empirically to depend on several factors, which act roughly independently. Albertsson et al. [101] represented

10 the contribution of these factors in logarithmic terms. K may be written as ln K ¼ ln K polymer þ ln K salt þ ln K ligand þ ln K other where Kpolymer, Ksalt, Kligand, Kother are the contributions to protein partitioning from factors originating in the phase-forming polymer, salt, afﬁnity ligands and other system factors. The effect of salts as an additive on the two-phase separation was reported very early in the development of ATPS [81]. Salts are generally classiﬁed into two groups: structure making and structure breaking to describe the effect of different ions on the structure of water. The ions (anions and cations) distribute unequally across the interface resulting into potential difference. 2+ Water structure-making ions (Li+, Na+, NH+ , Mg2+, F, 4 , Ca 2 3 SO2 4 , CO3 , PO4 , CH3COO ) favours the hydrophilic phase. Water structure-breaking ions (K+, Rb+, Cs+, Cl, Br, I, SCN, NO 3 , ClO4 ) favours the hydrophobic phase.

Construction of phase diagram Generally, phase compositions of an ATPS are described in the form of a phase diagram (Fig. 2). The curve AB called binodial curve separate a singlephase domain (below the curve) and a two-phase domain (above the curve). In the two-phase domain, when two components are mixed in the amounts corresponding to a point Y, two phases are formed with compositions represented by X and Z. The tie line XYZ represents the composition values of two phases in equilibrium. Obviously, all points on a tie line have the same compositions in individual phases. However, volumes of two phases are different and hence partitioning would change. The ratio of upper phase to lower phase volumes is equal to XY:YZ. Generally, phases separate out by gravitation. For a large scale, centrifugation is generally required. ATPS plays an important role in primary downstream processing steps. The main advantages of this technique are summarized as below 1. Scale up can be predicted easily and reliably from small laboratory experiments. 2. Little amount of mechanical energy is required while mixing, which allows rapid mass transfer and equilibrium. 3. Processing can be done with the unclariﬁed extract. 4. Separation can be made selective and rapid. 5. Separation can be carried out at room temperature due to rapid separation.

11

A PEG

Aqueous two phase region

(% w/v)

Top phase B One phase region Dextran (% w/v) Bottom phase PEG (Top phase) Dextran (Bottom phase)

Fig. 2. Two-phase diagram of PEG–dextran system. After shaking the mixture of these two polymers, phase separation is accomplished either by settling under gravity or by centrifugation. The top phase is of PEG and the bottom phase is of dextran. The systems formed at points X, Y and Z differ in the initial compositions and in the volume ratios.

6. It has proven to be more economical than other separation processes. 7. Considerable volume reduction of the protein feed. 8. The polymer stabilizes the enzyme (as the low interfacial tension avoids the detectable denaturation of proteins). Protein purification using ATPSs composed of free affinity macroligands Interfacing ATPS with the afﬁnity concept makes this technique more selective in the puriﬁcation of proteins (Table 5). In conventional format, this consists of conjugating an afﬁnity ligand to a polymer like PEG and forming PEG–afﬁnity ligand/salt or PEG–afﬁnity ligand/dextran phases [97]. Kamihara et al. [98] have described a useful strategy in which a polymeric afﬁnity ligand was incorporated in the PEG phase. In fact, the polymeric afﬁnity ligand in this case was a conjugate of the afﬁnity ligand and a smart polymer. Thus, the binding of the target protein to polymeric afﬁnity ligand favoured the former’s partitioning into the PEG phase. The complex could be separated from the PEG phase by using an appropriate stimulus. At this point, the protocol essentially follows the strategy of afﬁnity precipitation. This allows reuse of costly PEG. Equally important, the separation of the afﬁnity macroligand and the protein is quite facile. The afﬁnity macroligand can also be recycled. Later on, it has been shown in our lab that our results with afﬁnity precipitation involving inherent afﬁnity of natural and synthetic polymers could be exploited to make this ATPS strategy a still more attractive one. The proteins puriﬁed using these afﬁnity macroligands are listed in Table 5.

12 Table 5. Evolution of ATPS and its various applications. Sl. no

Strategy

Afﬁnity ligand (if any)

Enzyme/protein

Reference

1.

Simple ATPSs ATPS PEG/salt PEG/salt PEG/salt

– – –

b-Galactosidase Superoxide dismutase Aspartase

–

Thermostable aamylase

Veide et al. [102] Boland et al. [103] Paulsen and Hustedt [104] Li and Peeples [105]

2.

3.

Ethylene oxide–propylene oxide random copolymer (PEO–PPO)/dextran Polymeric phase conjugated with afﬁnity ligands ATPS PEG/dextran PEG/dextran

Dyes Cibacron blue F3G-A Procion red HE-3B

PEG/dextran

Metal ion Cu2+

Phosphofructokinase Alkaline phosphatase (calf intestine

Johansson et al. [106] Johansson et al. [107]

Haem protein

Wuenschell et al. [108]

PEG/salt Ethylene oxide–propylene oxide random copolymer (EO30PO70)/dextran Incorporation of a polymeric afﬁnity ligand in one of the phase ATPS PEG/salt

Polyaspartic tails as tags Polyarginine tails as tags Tryptophan tags Tyrosine tags

b-Galactosidase T4 lysozyme b-Galactosidase T4 lysozyme Cutinase Lactate dehydrogenase

Luther and Glatz [109] Luther and Glatz [110] Bandmann et al. [111] Fexby and Bulow [112]

Starch

Glucoamylase

Gouveia and Kilikian [113]

Incorporation of a smart macroafﬁnity ligand in one of the phases ATPS PEG/salt

Alginate

Teotia and Gupta [95]

PEG/salt

Alginate

a-amylase & bamylase phospholipase D

PEG/salt PEG/Reppal PES 200 PEG/salt

Alginate Eudragit S-100 Eudragit S-100

PEG/salt

Chitosan

Use of fusion proteins ATPS PEG/dextran PEG/dextran

4.

5.

Pullulanase Xylanase Recombinant Protein A Chitinases (from Neurospora crassa, cabbage and puffballs)

Teotia and Gupta [100] Teotia et al. [99] Teotia et al. [99] Kamihira et al. [98] Teotia et al. [74]

13 It may be added that ATPS have also been used for recovery of nonprotein products, like lutein [114] and cephalexin [115]; protein refolding [116]; quality control of protein pharmaceuticals [117]; small inclusion bodies [118] and bioconversion [96,119]. Three-phase partitioning Three-phase partitioning (TPP) is a simple but elegant nonchromatographic process used for puriﬁcation and concentration of proteins [120]. The steps are schematically outlined in Fig. 3 and several advantages are listed in Table 6. Organic cosolvent precipitation has been used for cold extraction of several proteins [121,122]. In the early 1950s, R. K. Morton used n-butanol for the extraction of several enzymes and found it as an efﬁcient solvent for removing lipids without harming the enzyme activity [123]. Later, porcine pancreatic lipase was puriﬁed using 8% (v/v) n-butanol and 20% (w/v) ammonium sulphate [124]. The puriﬁed enzyme was found to have twofold increased activity. TPP seems to have evolved out of these early results and at ﬁrst glance appears to be a hybrid of salting out and alcohol precipitation. However, there are deﬁnite indications that TPP is different from simple combinations of these steps: (a) Concentration of ammonium sulphate in TPP is much lower than that required for salting out of proteins. (b) In salt precipitation, the precipitate is a function of the precipitating salt concentration, while in TPP precipitation also depends upon the initial protein concentration [125]. (c) Organic solvent precipitations must be carried out at low temperatures, whereas TPP can be carried out at room temperature. Thus, process environments in TPP are more conducive to the maintenance of native structure. Mechanism of TPP Both ammonium sulphate and tert-butanol reinforce each other’s physicochemical effects, such as ionic strength effects, kosmotropy, osmotic stresses and exclusion-crowding effect to partition the protein as a midlayer between aqueous and organic phases [120]. Some tert-butanol also binds to the protein and increases its buoyancy against gravity and makes it ﬂoat between the aqueous and the organic layer as a gel-like disc [126]. Recently, Kiss et al. have reported that on adding ammonium sulphate to a mixture of water and tert-butanol, two distinct layers are formed with low interfacial tension (of the order of 10–0.1 mN/m), which probably provides mild condition for precipitation in TPP [127]. Also the ‘‘bulky and bushy’’ structure of tert-butanol prevents it from penetrating into the protein and unfolding it.

14 TPP

MLFTPP Optimization of TPP of macroaffinity ligand,

Protein of interest

Addition of macroaffinity ligand to crude protein mixture

Crude protein solution (NH4)2SO4 tert-butanol

Macroaffinity ligand bound protein

t-butanol layer Interfacial precipitate

(NH4)2SO4

tert-butanol

Aqueous layer

The three layers are separated using a Pasteur pipette

Interfacial precipitate of macroaffinity ligand bound protein

Interfacial precipitate of desired protein redissolved in buffer

The three layers are separated using a Pasteur pipette The protein is recovered from the complex of macroaffinity ligand bound protein

Fig. 3. Schematic presentation of three-phase partitioning (TPP) and macroafﬁnity

ligand facilitated three-phase partitionng (MLFTPP).

TPP has also been reported to have led to enhanced enzyme activity in several cases. Earlier this was attributed to the removal of inhibitors during the puriﬁcation step [128]. However, when pure Proteinase K (a serine protease) was subjected to TPP and its structure analyzed by X-ray diffraction at 1.5 A˚ resolution [129], a higher overall temperature factor (B factor) was observed, as a result of which side chains of several amino acid residues in the binding site were found to adopt more than one conformation. This resulted in the protein existing in an excited state, which explained its two-fold increased enzyme activity [129]. A similar treatment of pure a-chymotrypsin

15 Table 6. Advantages of TPP.

It is a simple procedure with short processing time Ammonium sulphate and tert-butanol are inexpensive chemicals, thus making TPP an economical protocol TPP conditions are mild and do not denature proteins It is usually carried out at room temperature TPP can be scaled down to semi-microlevels or scaled up to litre scale TPP puriﬁes as well as concentrates the protein, unlike chromatography, which dilutes the puriﬁed protein Puriﬁcation fold achieved in TPP is much higher than that achieved in simple salting out procedures TPP can be used with direct crude cultures containing cell debris; no preclariﬁcation steps such as centrifugation are required

also resulted in about two-fold increase in activity [130] but more interestingly the TPP-treated enzyme during its crystallization underwent autodigestion to produce a 14-residue peptide fragment with an exceptionally high binding constant [131]. Modification of TPP Two-step TPP It has been observed that under certain circumstances, upon subjecting the crude protein solution to TPP, three phases are formed; however, the desired protein remains predominantly in the aqueous layer. In such cases, the three layers are separated and the aqueous layer is subjected to a second TPP by adding more ammonium sulphate and tert-butanol. Three phases are formed again and this time most of the desired protein partitions into the interfacial precipitate. Such two-step TPP has been used for puriﬁcation of several enzymes, such as a-amylase [128], GFP [132], phospholipase D from carrots [133], alkaline phosphatase from chicken intestine [134] and pectinase [135]. Metal– affinity-based TPP Here, TPP was interfaced with metal–afﬁnity-based step to selectively purify proteins with surface histidine residues present at the appropriate distance. STI was chosen as the model protein. After the ﬁrst TPP, the aqueous layer (containing STI) was dialyzed and Cu/Zn salts were added in millimolar level during second TPP to purify STI (13-fold, 72% recovery) [136]. Affinity Macroligand facilitated three-phase partitioning It was soon realized that not just proteins, but water-soluble polymers too underwent TPP in the presence of ammonium sulphate and tert-butanol [137],[138]. This led to the development of a puriﬁcation protocol called macroafﬁnity ligand facilitated three-phase partitioning (MLFTPP), which

16 maybe considered as an extension of TPP [32]. The steps involved in MLFTPP have been schematically shown in Fig. 3. MLFTPP has all the advantages of TPP besides having the advantage of a more predictable design. One essential criterion for puriﬁcation of proteins using MLFTPP is that the chosen afﬁnity macroligand should partition quantitatively into the interfacial precipitate on being subjected to TPP. This was found to be true for Eudragit S-100 (a synthetic copolymer of methylmethacrylate and methylacrylic acid) [32], chitosan (a polymer obtained by deacylation of chitin) [138] and alginate (a natural polysaccharide found in brown algae) [137] and these were successfully used as afﬁnity macroligand in puriﬁcation of several enzymes (Table 7). Although only water-soluble polymers with inherent afﬁnity for certain enzymes have been tried for the puriﬁcation of the enzymes, smart polymers conjugated with afﬁnity ligands may also be tried as afﬁnity macroligand in this protocol. Incidentally, TPP as such has also been found useful for a variety of other applications as well [142–144] (Fig. 4). Expanded-bed affinity chromatography Expanded-bed chromatography [40,145] is an approach, which greatly reduces the complexity of downstream processing by eliminating ﬁltration, centrifugation and concentration steps. In this technique, the use of tailormade media allows one to operate chromatography in a ﬂuidized-bed mode. Particulate adsorbents, if packed in a column, cannot deal with crude suspensions as clogging of adsorbent beds will take place. Batch adsorption in stirred tanks is an option in such cases but shows poor resolution and capacity. A ﬂuidized bed in which inter-particle distance in adsorbent is increased, creating voids through which suspended impurities pass through, is best suited to capture the protein from crude extract. The best way to ﬂuidize adsorbent particles is by liquid ﬂow directed upwards. Such ﬂuidized beds can combine clariﬁcation, concentration and fractionation in a simple unit process. During the elution of protein, the liquid ﬂow may be reversed, leading to packing of the resin bed [41]. Such ﬂuidized beds, however, have the following limitation as chromatographic media. Adsorbent particles can move in all directions in the ﬂuidized bed. This problem is aggravated since the adsorbent particles are not monodisperse. The result is that larger particles tend to move lower and smaller particles get distributed towards the top. If, to start with, the adsorbent particles have an appropriate size distribution, ‘‘classiﬁcation’’ leads to welldeﬁned layers, gives stability to the ﬂuidized bed and shows ﬂuid dispersion characteristics closer to that of a packed bed. Such a stable ﬂuidized bed is called an expanded bed. Expanded-bed adsorption and expanded-bed chromatography are used as synonyms. Use of expanded/ﬂuidized beds is a

17 Table 7. Proteins puriﬁed by MLFTPP. Enzyme puriﬁed

Polymer used

Conditions of MLFTPP

Yield (%)

Puriﬁcation fold

Reference

Xylanase (Aspergillus niger)

Eudragit-S 100

60

95

Sharma and Gupta [32]

a-Amylase (wheat germ)

Esteriﬁed alginate

1% eudragit, 30% (w/v) ammonium sulphate, 1:1 ratio (v/v) of tbutanol to aqueous solution, 401C for 1 h 0.5% esteriﬁed alginate, 20% w/ v ammonium sulphate, 1:1 ratio of tbutanol to aqueous solution at 371C for 1 h.

77

55

Mondal et al. [139]

a-Amylase (from porcine pancreas a-Amylase (Bacillus amyloliquefaciens) Glucoamylase (Aspergillus niger)

Esteriﬁed alginate Esteriﬁed alginate Alginate

92

10

74

5.5

83

20

Mondal et al. [140]

Pullulanase (Bacillus acidopullulyticus)

Esteriﬁed alginate

89

38

Mondal et al. [140]

Pectinase

Alginate

96

13

Sharma et al. [141]

Cellulase

Chitosan

92

16

Sharma et al. [141]

1% w/v alginate, 30% ammonium sulphate, 2:1 ratio of tbutanol to aqueous solution at 371C for 1 h. 0.5% esteriﬁed alginate, 20% w/ v ammonium sulphate, 1:1 ratio of tbutanol to aqueous solution at 371C for 1 h. 1% w/v alginate, 30% ammonium sulphate, 2:1 ratio of tbutanol to aqueous solution at 371C for 1 h. 0.2% w/v chitosan, 45% ammonium sulphate, 1:1 ratio of tbutanol, 401C, 1h

18 For designing biosensors (Borbas et al., 2003 [142])

For altering biodegradability of natural polymers (Sharma et al., 2003 [138])

For extraction of oil from plant sources (Sharma et al., 2002 [144])

For preparing highly active biocatalyst Roy & Gupta., 2004 [130])

For refolding of denatured proteins (Roy et al., 2004 [143])

Fig. 4. Other applications of TPP.

promising solution to the problems of directly purifying particulate containing materials. Chromatography using ﬂuidized beds can be scaled up and there are several reports describing applications of expanded bed adsorption chromatography in large-scale protein separation [40,146,147]. The early work was carried out mostly with ion-exchange chromatography. Similar to other chromatographic formats, expanded-bed chromatography can be also carried out with an afﬁnity media and expanded-bed afﬁnity chromatography (EBAC) is a powerful tool for afﬁnity capture of the desired protein from a crude broth [148,149]. As this combines the possibility of dealing directly with crude suspensions with high selectivity of afﬁnity methods, EBAC has found several interesting applications, which are listed in Table 8. Conclusion Even a casual look at the current biotechnology literature would reveal that there is a vigorous renewal of interest in protein bioseparation. While

19 Table 8. Expanded bed afﬁnity chromatography EBAC for puriﬁcation of proteins. Matrix Reference

Afﬁnity ligand

Protein puriﬁed

Fold

puriﬁcation

Yield (%)

StreamlineTM chelating

Cu2+

His6-viral coat protein a-Amylase inhibitor His6-glutathione-Stransferase b-Galactosidase

20

71

Choe et al. [150]

23

83

3.3

80

6

86.4

–

95

12

86

Roy and Gupta [151] Clemmit and Chase [149] Clemmit and Chase [152] Tho¨mmes et al. [153] Chang et al. [154]

1.4

95

Ni2+

Streamline

rProtein A

Perﬂuorocarbon emulsion

Procion Red H-E7B

Sepharose Fast FlowTM Phosphofructokinase

CI Reactive Blue 4 Cibacron blue 5.2

39

STI

Trypsin

None None None None

Polyacrylamidemagnetite composite CELBEADS Alginate beads

Nd–Fe–B alloy-densiﬁed agarose Agarose-coated alumina

Dye Dye-IDA

Monoclonal antibody Glucose-6phosphate dehydrogenase Human serum albumin

McCreath et al. [155]

Chase and –

–

Cellulase Pectinase

36

91.2

a-Amylase Phospholipase D

58 317

90 78

Pullulanase

59

99

Pectinase Lysozyme

9.1 8.8

87.6 93.5

Alcohol dehydrogenase

Draeger [148] Cocker et al. [156] Roy et al. [157] Somers et al. [158] Roy et al. [159] Sharma et al. [160] Roy and Gupta [161] Roy et al. [162] Tong et al. [163] Hidayat et al. [164]

editorial of one journal talks of ‘‘The importance of bioseparations: giving credit where it is due’’ [165], another features ‘‘Striving for purity: advances in protein puriﬁcation’’ [28]. The one overview most relevant to the present review is entitled ‘‘Alternative bioseparation operations: life beyond packed bed chromatography’’ [166]. This last mentioned overview refers to monoliths, MLFTPP, afﬁnity precipitation, TPP as ‘‘pre-commercial’’ alternatives to chromatography. Membrane chromatography, ATPE, high-performance tangential ﬂow ﬁltration have been put together as having selected industrial applications. Interestingly, crystallization and precipitation (along with packed-bed chromatography) is described as processes with highest industrial maturity. This is one way of looking at the bioseparation scene. As this review describes, techniques like MLFTPP and afﬁnity precipitation work and deliver high yield as well as high purity. In many cases, the ﬁnal product

20 after one or two steps has purity level of single band on SDS-PAGE [36,37,136]. So these are emerging options as we say it and pre-commercial (as Przybycien et al. term it) in the sense that industry has still not caught onto them. On the basis of their simple design, scalability and sheer economics, these techniques score over other bioseparation techniques. There is one point that is often overlooked. Any afﬁnity-based approach would be less likely to resolve isoenzymes/isoproteins (except in those cases where elution protocols are so designed so as to exploit different binding constants between the matrix and the isoenzymes/isoproteins). Here, a simple ion-exchange chromatography would be more useful. This whole review in a way has looked at what maybe the bioseparation strategies of tomorrow’s industrial biotechnology. Enzyme-based bioprocesses are increasingly replacing traditional chemical processes. From producing herbicide to chiral drug intermediates, enzymes are replacing chemical catalysts. Enzymes are the powerful and important part of industrial sustainable chemistry. The need for efﬁcient downstream processing would intensify. Finally, there is a distinct trend towards simultaneous puriﬁcation and immobilization strategies [167–169]. Leads from afﬁnity precipitation have led to smart biocatalyst design which has inbuilt puriﬁcation step. Highperformance biocatalyst designs such as cross-linked enzyme aggregate (CLEA) also do not require highly puriﬁed enzymes as starting material [170]. So, ﬁrst we have had integration of upstream and downstream processes. Now, we have integration of puriﬁcation and immobilization. Thus, there are enough reasons to believe that need (for efﬁcient separation strategies) would continue to breed novel and innovative technologies. Acknowledgements The partial ﬁnancial support provided by the Department of Science and Technology (DST) and Council of Scientiﬁc and Industrial Research (CSIR) (Extramural Division & Technology Mission on Oil Seeds, Pulses and Maize), Government of India Organisations, is acknowledged. The ﬁnancial support provided by DST to ST in the form of Young Scientist (Fast Track Scheme), CSIR to SJ and IIT Delhi to KM as Senior Research Fellowships is also gratefully acknowledged. References 1. 2. 3.

Sumner JB. The isolation and crystallization of the enzyme urease. Preliminary paper. J Biol Chem 1926;69:435–441. Northrop JH. Crystallization of pepsin from alcohol. J Gen Physiol 1946;30:177–184. Kunitz M. Isolation of a crystalline compound of trypsin and of soybean trypsin inhibitor. J Gen Physiol 1947;30:311–320.

21 4. 5. 6. 7. 8. 9. 10. 11.

12. 13.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Fruton JS and Simmonds S (eds). General properties of protein. In: General Biochemistry, 2nd edn, New York, Wiley, 1958, pp. 14–44. Dixon M and Webb EC (eds). Enzyme isolation. In: Enzymes, UK, Longmans Green & Co. Ltd, 1958, pp. 27–53. Peterson EA and Sober HA. Chromatography of proteins. I. Cellulose ion exchange adsorbents. J Am Chem Soc 1956;78:751–755. Porath J and Flodin P. Gel ﬁltration: a method for desalting and group separation. Nature 1959;183:1657–1659. Axe´n R, Porath J and Ernback S. Chemical coupling of peptides and proteins to polysaccharides by means of cyanogens halides. Nature 1967;214:1302–1304. Godfrey T and West S (eds). Industrial Enzymology, Macmillan Press Ltd., London, 1988. Smyth DG. Techniques in enzymic hydrolysis. In: Methods in Enzymology, Vol. 11, New York, Academic Press, 1967, pp. 214–231. Rovery M. Chromatography of trypsin to remove chymotrypsin and of chymotrypsin to remove trypsin. In: Methods in Enzymology, Vol. 11, New York, Academic Press, 1967, pp. 231–236. Ryu DDY and Nam DH. Biomolecular engineering: a new frontier in biotechnology. J Mol Catal B: Enzym 2000;10:23–37. Sunic D, McNeil JD, Andress DL and Belford DA. Insulin-like growth factor binding protein-5 proteolytic activity in ovine articular chondrocyte culture. Biochim Biophys Acta-Gen Subjects 1998;1425:567–576. Jaeger KE. Protein technologies and commercial enzyme. White is the hype. Biocatalyst on the move. Curr Opin Biotechnol 2004;15:269–271. Jenck JF, Agterberg F and Droescher MJ. Products and processes for a sustainable chemical industry: a review of achievments and prospects. Green Chem 2004;6:544–556. Means GE and Feeny RE (eds). Chemical Modiﬁcation of Proteins, London, Holden Day Inc, 1971. Tyagi R and Gupta MN. Chemical modiﬁcation and chemical crosslinking for protein/ enzyme stabilization. Biochemistry (Moscow) 1998;63:334–344. Taylor RF (ed). Protein Immobilization, New York, Marcel Dekker Inc, 1991. Trower MK., Totowa, NJ, Humana Press Inc, 1996. Turner NJ. Directed evolution of enzymes for applied biocatalysis. Trends Biotechnol 2003;21:474–478. Labrou NE. Directed enzyme evolution: bridging the gap between natural enzymes and commercial applications. Biomol Eng 2005;22:vii–ix. Morley KL and Kazlauskas RJ. Improving enzyme properties: when are closer mutations better? Trends Biotechnol 2005;23:231–237. Gupta MN. Enzyme function in organic solvents. Eur J Biochem 1992;203:25–32. Gupta MN and Roy I. Enzymes in organic media: forms, function and applications. Eur J Biochem 2004;271:2575–2583. Gupta MN (ed). Methods in Non-aqueous Enzymology, Basel, Switzerland, Birkhauser Verlag, 2000. Lichty JJ, Malecki JL, Agnew HD, Michelson-Horowitz DJ and Tan S. Comparison of afﬁnity tags for protein puriﬁcation. Protein Expression Purif 2005;41:98–105. Lesley SA. High-throughput proteomics: protein expression and puriﬁcation in the postgenomic world. Protein Expression Purif 2001;22:159–164. Smith C. Striving for purity: advances in protein puriﬁcation. Nat Methods 2005;2:71–77.

22 29. Bajpai S, Sharma A and Gupta MN. Removal and recovery of antinutritional factors from soybean ﬂour. Food Chem 2005;89:497–501. 30. Jain S and Gupta MN. An integrated process for separation of major and minor proteins from goat serum. Appl Biochem Biotechnol 2005;125:53–62. 31. Harsa S and Furusaki S. Chromatographic separation of amyloglucosidase from the mixtures of enzymes. Biochem Eng J 2001;8:257–261. 32. Sharma A and Gupta MN. Macroafﬁnity ligand facilitated three-phase partitioning (MLFTPP) for puriﬁcation of xylanase. Biotechnol Bioeng 2002;80:228–232. 33. Bonnerjea J, Oh S, Hoare M and Dunnill P. Protein puriﬁcation: the right step at the right time. Bio/Technology 1986;4:954–958. 34. Kaul RH and Mattiasson B. Secondary puriﬁcation. Bioseparation 1992;3:1–26. 35. LaVallie ER and McCoy JM. Gene fusion expression systems in E. coli. Curr Opin Biotechnol 1995;6:501–506. 36. Sharma A, Sharma S and Gupta MN. Puriﬁcation of wheat germ amylase by precipitation. Protein Expression Purif 2000;18:111–114. 37. Teotia S, Lata R, Khare SK and Gupta MN. One step puriﬁcation of glucoamylase by afﬁnity precipitation with alginate. J Mol Recogn 2001;14:295–299. 38. Schuster M, Wasserbauer E, Ortner C, Graumann K, Jungbauer A, Hammerschmid F and Werner G. Short cut of protein puriﬁcation by integration of cell-disrupture and afﬁnity extraction. Bioseparation 2000;9:59–67. 39. Pierpoint WS. The extraction of enzymes from plant tissues rich in phenolic compounds. In: Methods in Molecular Biology, Doonan S (ed), Protein Puriﬁcation Protocols, Totowa, NJ, Humana Press Inc, 1996, pp. 58–68. 40. Chase HA. Puriﬁcation of proteins by adsorption chromatography in expanded beds. Trends Biotechnol 1994;12:296–303. 41. Anspach FB, Curbelo D, Hartmann R, Garke G and Deckwer WD. Expanded bed chromatography in primary protein puriﬁcation. J Chromatogr A 1999;865:129–144. 42. Jensen LB, Riise E, Nielsen LK, Dziegiel M, Fugger L and Engberg J. Efﬁcient puriﬁcation of unique antibodies using peptide afﬁnity-matrix columns. J Immunol Methods 2004;284:45–54. 43. Jungbauer A. Chromatographic media for bioseparation. J Chromatogr A 2005;1065:3–12. 44. Cuatrecasas P, Wilchek M and Anﬁnsen CB. Selective enzyme puriﬁcation by afﬁnity chromatography. Proc Natl Acad Sci 1968;61:636–643. 45. Thompson ST, Cass KH and Stellwagen E. Blue Dextran-Sepharose: an afﬁnity column for the dinucleotide fold in proteins. Proc Natl Acad Sci 1975;72:669–672. 46. Stellwagen E. Use of blue dextran as a probe for the nicotinamide adenine dinucleotide domain in proteins. Acc Chem Res 1977;10:92–98. 47. Denizli A and Piskin E. Dye-ligand afﬁnity systems. J Biochem Biophys Methods 2001;49:391–416. 48. Porath J, Carlsson J, Olsson I and Belfrage G. Metal chelate afﬁnity chromatography: a new approach to protein fractionation. Nature 1975;258:598–599. 49. Huang PY, Baumbach GA, Dadd CA, Buettner JA, Masecar BL, Hentsch M, Hammond DJ and Carbonell RG. Afﬁnity puriﬁcation of von Willebrand factor using ligands derived from peptide libraries. Bioorg Med Chem 1996;4: 699–708. 50. Chen B, Bestetti G, Day RM and Turner APF. The synthesis and screening of a combinatorial peptide library for afﬁnity ligands for glycosylated haemoglobin. Biosensor Bioelectron 1998;13:779–785.

23 51. Romig TS, Bell C and Drolet DW. Aptamer afﬁnity chromatography: combinatorial chemistry applied to protein puriﬁcation. J Chromatogr B, Biomed Sci Appl 1999;731:275–284. 52. Larsson PO and Mosbach K. Afﬁnity precipitation of enzymes. FEBS Lett 1979;98:333–338. 53. Burnett TJ, Peebles HC and Hageman JH. Synthesis of a ﬂuorescent boronic acid which reversibly binds to cell walls and a diboronic acid which agglutinates erythrocytes. Biochem Biophys Res Commun 1980;96:157–162. 54. Hoffmann AS. Bioconjugates of intelligent polymers and recognition proteins for use in diagnostics and afﬁnity separations. Clin Chem 2000;46:1478–1486. 55. Roy I and Gupta MN. Smart polymeric materials: emerging biochemical applications. Chem Biol 2003;10:1161–1171. 56. Linne E, Garg N, Kaul RH and Mattiasson B. Evaluation of alginate as a ligand carrier. Biotechnol Appl Biochem 1992;16:48. 57. Schneider M, Guillot C and Lamy B. The afﬁnity precipitation technique. Application of isolation and puriﬁcation of trypsin from bovine pancreas. Ann N Y Acad Sci USA 1981;369:257–263. 58. Pe´cs M, Eggert M and Schu¨gerl K. Afﬁnity precipitation of extracellular microbial enzymes. J Biotechnol 1991;21:137–142. 59. Morris JE, Hoffman AS and Fisher RR. Afﬁnity precipitation of proteins by polyligands. Biotechnol Bioeng 1993;41:991–997. 60. Maeda M, Nishimura C, Inenaga A and Takagi M. 1993 Modiﬁcation of DNA with poly(N-isopropylacrylamide) for thermally innduced afﬁnity separation. Reactive Polym 1995;21:27–35. 61. Taniguchi M, Kobayashi M, Natsui K and Fujii M. Puriﬁcation of Staphylococcal protein A by afﬁnity precipitation using a reversible soluble-insoluble polymer with human IgG as a ligand. J Ferment Bioeng 1989;68:32–36. 62. Guoquiang D, Lali A, Kaul RH and Mattiasson B. Afﬁnity thermoprecipitation of lactate dehydrogenase and pyruvate kinase from porcine muscle using Eudragit bound cibacron blue. J Biotechnol 1994;37:23–31. 63. Guoquiang D, Benhura MAN, Kaul RH and Mattiasson B. Afﬁnity thermoprecipitation of yeast alcohol-dehydrogenase through metal ion-promoted binding with Eudragit-bound cibacron-blue-3GA. Biotechnol Prog 1995;11:187–193. 64. Kumar A and Gupta MN. Afﬁnity precipitation of trypsin with soybean trypsin inhibitor linked Eudragit S-100. J Biotechnol 1994;37:185–189. 65. Larsson EL and Mattiasson B. Isolation of Con A by afﬁnity precipitation. Biotechnol Tech 1994;8:51–56. 66. Taipa MA, Kaul R, Mattiasson B and Cabral JMS. Preliminary studies on the puriﬁcation of a monoclonal antibody by afﬁnity precipitation with Eudragit S-100. J Mol Recogn 1998;11:240–242. 67. Guoquiang D, Kaul RH and Mattiasson B. Puriﬁcation of Lactobacillus bulgaricus Dlactate dehydrogenase by precipitation with an anionic polymer. Bioseparation 1993;3:333–341. 68. Gupta MN, Guoquiang D, Kaul RH and Mattiasson B. Puriﬁcation of xylanase from Trichoderma viride by precipitation with an anionic polymer Eudragit S-100. Biotechnol Techn 1993;8:117–122. 69. Breccia JD, Sin˜eriz F, Baigorı´ MD, Castro GR and Kaul RH. Puriﬁcation and characterization of a thermostable xylanase from Bacillus amyloliquefaciens. Enzyme Microb Technol 1998;22:42–49.

24 70. Sun Y, Yu K, Jin XH and Zhou XZ. Polymerized liposome as ligand carrier for afﬁnity precipitation of proteins. Biotechnol Bioeng 1995;47:20–25. 71. Senstad C and Mattiasson B. Puriﬁcation of wheat germ agglutinin using afﬁnity ﬂocculation with chitosan and a subsequent centrifugation or ﬂotation step. Biotechnol Bioeng 1989;34:387–393. 72. Homma T, Fujii M, Mori J, Kawakami T, Kuroda K and Taniguchi M. Production of cellobiose by enzymatic hydrolysis: removal of b-glucosidase from cellulase by afﬁnity precipitation using chitosan. Biotechnol Bioeng 1993;41:405–410. 73. Tyagi R, Kumar A, Sardar M, Kumar S and Gupta MN. Chitosan as an afﬁnity macroligand for precipitation of N-acetyl glucosamine binding proteins/enzymes. Iso Purif 1996;2:217–226. 74. Teotia S, Lata R and Gupta MN. Chitosan as a macroafﬁnity ligand: puriﬁcation of chitinases by afﬁnity precipitation and aqueous two-phase extractions. J Chromatogr A 2004;1052:85–91. 75. Gupta MN, Guoquiang D and Mattiasson B. Puriﬁcation of endopolygalacturonase by afﬁnity precipitation using alginate. Biotechnol Appl Biochem 1993;18:321–327. 76. Teotia S, Khare SK and Gupta MN. An efﬁcient puriﬁcation process for sweet potato b-amylase by afﬁnity precipitation with alginate. Enzyme Microb Technol 2001;28:792–795. 77. Sharma S, Sharma A and Gupta MN. 2000 One step puriﬁcation of peanut phospholipase D by precipitation with alginate. Bioseparation 2001;9:93–98. 78. Bradshaw AP and Sturgeon RJ. The synthesis of soluble polymer–ligand complexes for afﬁnity precipitation studies. Biotechnol Techn 1990;4:67–71. 79. Mondal K, Roy I and Gupta MN. k-Carrageenan as a carrier in afﬁnity precipitation of yeast alcohol dehydrogenase. Protein Expression Purif 2003;32:151–160. 80. Walter H and Johansson G. Partitioning in aqueous two-phase systems: an overview. Anal Biochem 1986;155:215–242. 81. Albertsson PA (ed). Partition of Cell Particles and Macromolecules, 3rd edn, New York, Wiley (Interscience), 1986. 82. Rito-Palomares M. Practical application of aqueous two phase partition to process development for the recovery of biological products. J Chromatogr B 2004;807:3–11. 83. Walker SJ, Dale CJ and Lyddiatt A. Aqueous two-phase partition of complex protein feedstocks derived from brain tissue homogenates. J Chromatogr B 1996;680:91–96. 84. Harris DP, Andrews AT, Wright G, Pyle DL and Asenjo JA. The application of aqueous two-phase systems to the puriﬁcation of pharmaceutical proteins from transgenic sheep milk. Bioseparation 1997;7:31–37. 85. Alves JGLF, Chumpitaz LDA, da Silva LHM, Franco TT and Meirelles AJA. Partitioning of whey proteins, bovine serum albumin and porcine insulin in aqueous two-phase systems. J Chromatogr B 2000;743:235–240. 86. Benavided J and Rito-Palomares M. Bioprocess intensiﬁcation: a potential aqueous two phase process for the primary recovery of B-phycoerythrin from Porphyridium Cruentum. J Chromatogr B 2004;807:33–38. 87. Jain S, Teotia S and Gupta MN. Puriﬁcation of green ﬂuorescent protein overexpressed by a mutant recombinant E. coli. Protein Expression Purif 2004;36:76–81. 88. Lu M, Tjerneld F, Johansson G and Albertsson PA. Preparation of benzoyl–dextran and its use in aqueous two-phase systems. Bioseparation 1991;2:247–255. 89. Stewart RM and Todd P. A poly(ethylene glycol)-sodium chloride multiphase system for extraction of acid hydrolasates. In: Frontiers in Bioprocessing II, Todd P, Sikdar SK and Bier M (eds), Washington, DC, American Chemical Society, 1992, pp. 352–356.

25 90. Snyder SM, Cole KD and Szlag CC. Phase composition viscosities, and densities for aqueous two phase system composed of polyethylene glycol and various salts at 25 1C. J Chem Eng Data 1992;37:268–274. 91. Pietruszka N, Galaev IY, Kumar A, Brzozowski ZK and Mattiasson B. 2000 New polymers forming aqueous two phase polymer systems. Biotechnol Prog 1996;16:408–415. 92. Oliverira LA, Neto BB, Porto ALF and Tambourg EB. Extractive cultivation of xylanase by Penicillin janthinellum in a poly(ethylene glycol)/cashew nut gum aqueous two phase system. Biotechnol Prog 2004;20:1880–1884. 93. Johansson G. Effects of salts on the partition of proteins in aqueous polymeric biphasic systems. Acta Chem Scand Sec B 1974;28:873–882. 94. Bamberger S, Seaman GVF, Brown JA and Brooks DV. The partition of sodium phosphate and sodium chloride in aqueous dextran poly(ethylene glycol) two phase systems. J Colloid Inter Sci 1984;99:187–193. 95. Teotia S and Gupta MN. Reversibly soluble macroafﬁnity ligand in aqueous two-phase separation of enzymes. J Chromatogr A 2001;923:275–280. 96. Hatti-Kaul R (ed). Aqueous Two Phase Systems: Methods and Protocols, Totowa, NJ, Humana Press Inc, 2000. 97. Cordes A and Kula M-R. Process design for large-scale puriﬁcation of formate dehydrogenase from Candida boidinii by afﬁnity partitioning. J Chromatogr 1986;376:375–384. 98. Kamihara M, Kaul R and Mattiasson B. Puriﬁcation of recombinant protein A by aqueous two phase extraction integrated with afﬁnity precipitation. Biotechnol Bioeng 1992;40:1381–1387. 99. Teotia S, Lata R and Gupta MN. Free polymeric bioligands in aqueous two-phase afﬁnity extractions of microbial xylanases and pullulanase. Protein Expression Purif 2001;22:484–488. 100. Teotia S and Gupta MN. Puriﬁcation of phospholipase D by two-phase afﬁnity extraction. J Chromatogr A 2004;1025:297–301. 101. Albertsson PA, Johansson G and Tjerneld F. Aqueous two phase separations. In: Separation Processes in Biotechnology, Asenjo JA (ed), New York, Marcel Dekker, 1990, pp. 287–327. 102. Veide A, Lindba¨ck T and Enfors SO. Continuous extraction of b-D-galactosidase from E. coli in an aqueous two-phase system: effects of biomass concentration on partitioning and mass transfer. Enzyme Microb Technol 1984;6:325–330. 103. Boland MJ, Hesselink PGM, Papamichael N and Hustedt H. Extractive puriﬁcation of enzymes from animal tissue using aqueous two phase systems: pilot scale studies. J Biotechnol 1991;19:19–34. 104. Paulsen J and Hustedt H. Extractive puriﬁcation of aspartase from E. coli K12. In: Methods in Enzymology, Vol. 228, Walter H and Johansson G (eds), New York, Academic Press, 1994, pp. 590–599. 105. Li M and Peeples TL. Puriﬁcation of hyperthermophilic archael amylotytic enzyme (MJAI) using thermoseparating aqueous two phase systems. J Chromatgr B 2004;807:69–74. 106. Johansson G and Andersson M. Parameters determining afﬁnity partitioning of yeast enzymes using polymer-bound triazine dye ligands. J Chromatogr 1984;303: 39–51. 107. Johansson G and Kopperschlager G. Effects of organic solvents on the partitioning of enzymes in aqueous two-phase systems. J Chromatogr 1987;388:295–305.

26 108. Wuenschell GE, Naranjo E and Arnold FH. Aqueous two-phase metal afﬁnity extraction of heme proteins. Bioprocess Eng 1990;5:199–202. 109. Luther JR and Glatz CE. Genetically engineered charge modiﬁcations to enhance protein separation in aqueous two phase systems: electrochemical partitioning. Biotechnol Bioeng 1994;44:147–153. 110. Luther JR and Glatz CE. Genetically engineered charge modiﬁcations to enhance protein separation in aqueous two phase systems: charge directed partitioning. Biotechnol Bioeng 1995;46:62–68. 111. Bandmann N, Collet E, Leijen J, Uhle´n M, Veide A and Nygren P-A. Genetic engineering of the Fusarium solani pisi lipase cutinase for enhanced partitioning in PEGphosphate aqueous two-phase systems. J. Biotechnol 2000;79:161–172. 112. Fexby S and Bulow L. Improved partitioning in aqueous two phase system of tyrosinetagged recombinant lactate dehydrogenase. Protein Expression Purif 2002;25: 263–269. 113. Gouveia TD and Kilikian BV. Bioafﬁnity extraction of glucoamylase in aqueous two phase systems using starch as free bioligand. J Chromatogr B 2000;743:241–246. 114. Cisneros M, Benavides J, Brenes C and Rito-Palomares M. Recovery in aqueous two phase systems of lutein produced by the green microalga Chlorella protothecoides. J Chromatogr A 2004;807:105–110. 115. Wei D-Z, Zhu J-H and Cao X-J. Enzymatic synthesis of cephalexin in aqueous two phase systems. Biochem Eng J 2002;11:95–99. 116. Kuboi R, Morita S, Ota H and Umakoshi H. Protein refolding using stimuli–responsive polymer-modiﬁed aqueous two-phase systems J Chromatogr B 2000;743:215–223. 117. Bermudez O and Forciniti D. Aggregation and denaturation of antibodies: a capillary electrophoresis, dynamic light scattering, and aqueous two phase partitioning study. J Chromatgr B 2004;807:17–24. 118. Walker SJ and Lyddiatt A. Aqueous two-phase systems as an alternative process route for the fractionation of small inclusion bodies. J Chromatogr B 1998;711:185. 119. Sharma S, Teotia S and Gupta MN. Bioconversion in an aqueous two phase system using a smart biocatalyst: casein hydrolysis by alpha chymotrypsin. Enzyme Microb Technol 2003;32:337–339. 120. Dennison C and Lovrien R. Three phase partitioning: concentration and puriﬁcation of proteins. Protein Expression Purif 1997;11:149–161. 121. Scopes RK (ed). Protein Puriﬁcation: Principles and Practice, New York, Springer, 1987 pp. 71–101. 122. Hatzinikolaou DG, Tsoukia C, Kekos D and Macris BJ. An efﬁcient and optimized puriﬁcation procedure for the superoxide dismutase from Aspergillus niger. Partial characterization of the puriﬁed enzyme. Bioseparation 1997;7:39–46. 123. Morton RK. Separation and puriﬁcation of enzymes associated with insoluble particles. Nature 1950;166:1092–1095. 124. Melius P and Simmons WS. Utilization of 1-butanol for the preparation of pancreatic lipase. Biochim Biophys Acta 1965;105:600–602. 125. Pike RN and Dennison C. Protein fractionation by three phase partitioning (TPP) in aqueous/t-butanol mixtures. Biotechnol Bioeng 1989;33:221–228. 126. Borbas R, Turza S, Szamos J and Kiss E. Analysis of protein gels formed by interfacial partitioning. Colloid Poly Sci 2001;279:705–713. 127. Kiss E, Szamos J, Tamas B and Borbas R. Interfacial behaviour of proteins in three phase partitioning using salt-containing water/tert-butanol systems. Colloids Surf A 1998;142:295–302.

27 128. Lovrien R, Goldensoph C, Anderson PC and Odegaard B. Three phase partitioning (TPP) via t-butanol: enzymes separation from crudes. In: Protein Puriﬁcation: Micro to Macro, Burgess R (ed), New York, NY, Alan R. Liss Inc, 1987, pp. 131–148. 129. Singh RK, Gourinath S, Sharma S, Roy I, Gupta MN, Betzel C, Srinivasan A and Singh TP. Enhancement of enzyme activity through three phase partitioning: crystal structure of a modiﬁed serine proteinase at 1.5 A˚ resolution. Protein Eng 2001;14:307–313. 130. Roy I and Gupta MN. a-Chymotrypsin shows higher activity in water as well as organic solvents after three phase partitioning. Biocatal Biotransform 2004;22:261–268. 131. Singh N, Jabeen T, Sharma S, Roy I, Gupta MN, Bilgrami S, Somvanshi RK, Dey S, Perbandt M, Betzel C and Singh TP. Detection of native peptides as potent inhibitors of enzymes: crystal structure of the complex formed between the treated bovine a-chymotrypsin and an autocatalytically produced fragment Ile–Val–Asn–Gly–Glu–Glu–Ala–Val–Pro–Gly–Ser–Trp–Pro–Trp at 2.2 A˚ resolution. FEBS J 2005;272:562–572. 132. Jain S, Singh R and Gupta MN. Puriﬁcation of recombinant green ﬂuorescent protein by three-phase partitioning. J Chromatogr A 2004;1035:83–86. 133. Sharma S and Gupta MN. Puriﬁcation of phospholipase D from Dacus carota by three phase partitioning and its characterization. Protein Expression Purif 2001;21:310–316. 134. Sharma A, Sharma S and Gupta MN. Puriﬁcation of alkaline phosphatase from chicken intestine by three phase partitioning and use of phenyl-Sepharose 6B in the batch mode. Bioseparation 2000;9:155–161. 135. Sharma A and Gupta MN. Puriﬁcation of pectinases by three phase partitioning. Biotechnol Lett 2001;23:1625–1627. 136. Roy I and Gupta MN. Three phase afﬁnity partitioning of proteins. Anal Biochem 2002;300:11–14. 137. Sharma A and Gupta MN. Three phase partitioning of carbohydrate polymers – separation and puriﬁcation of alginates. Carb Polym 2002;48:391–395. 138. Sharma A, Mondal K and Gupta MN. Some studies on characterization of three phase partitioned chitosan. Carb Polym 2003;52:433–438. 139. Mondal K, Sharma A and Gupta MN. Macroafﬁnity ligand facilitated three phase partitioning (MLFTPP) of a-amylases using modiﬁed alginate. Biotechnol Prog 2003;19:493–494. 140. Mondal K, Sharma A and Gupta MN. Macroafﬁnity ligand facilitated three phase partitioning (MLFTPP) for puriﬁcation of glucoamylase and pullulanase using alginate. Protein Expression Purif 2003;28:190–195. 141. Sharma A, Mondal K and Gupta MN. Separation of enzymes by sequential macroafﬁnity ligand-facilitated three-phase partitioning. J Chromatogr A 2003;995:127–134. 142. Borbas R, Murray BS and Kiss E. Interfacial shear rheological behaviour of proteins in three phase partitioning systems. Colloid Surf A 2003;213:93–103. 143. Roy I, Sharma A and Gupta MN. Three phase partitioning for simultaneous renaturation and partial puriﬁcation of Aspergillus niger xylanase. Biochim Biophys Acta 2004;1698:107–110. 144. Sharma A, Khare SK and Gupta MN. Three phase partitioning of extraction of oil from soybean. Bioresource Technol 2002;85:327–329. 145. Hjorth R, Leijon P, Frej AKB and Jagersten C. Expanded bed adsorption chromatography. In: Bioseparation & Bioprocessing, Vol. 1, Subramanian G (ed), Weinheim, Wiley, 1998, pp. 199–226. 146. Chase HA. Prediction of the performance of preparative afﬁnity chromatography. J Chromatogr 1984;297:179–202.

28 147. Barnﬁeld FAK, Hjorth R and Hammarstrom A. Pilot scale recovery of recombinant annexin from unclariﬁed E. coli homogenate using expanded bed adsorption. Biotechnol Bioeng 1994;44:922–929. 148. Chase HA and Draeger NM. Afﬁnity puriﬁcation of proteins using expanded beds. J Chromatogr 1992;597:129–145. 149. Clemmitt RH and Chase HA. Immobilised metal afﬁnity chromatography of b-galactosidase from unclariﬁed E. coli homogenates using expanded bed adsorption. J Chromatogr A 2000;874:27–43. 150. Choe W, Clemmit RH, Chase HA and Middelberg APJ. Comparison of histidine tag capture chemistries for puriﬁcation following chemical extraction. J Chromatogr A 2002;953:111–121. 151. Roy I and Gupta MN. Puriﬁcatioon of a ‘double headed’ inhibitor of alpha-amylase/ proteinase K from wheat germ by expanded bed chromatography. Bioseparation 2000;9:239–245. 152. Clemmitt RH and Chase HA. Facilitated downstream processing of a histidine-tagged protein from unclariﬁed E. coli homogenates using immobilized metal afﬁnity expanded bed absorption. Biotechnol Bioeng 2000;67:206–216. 153. Tho¨mmes J, Bader A, Halfar M, Karau A and Kula MR. Isolation of monoclonal antibodies from cell containing hybridoma broth using a protein A coated adsorbent in expanded beds. J Chromatogr A 1996;752:111–122. 154. Chang YK, McCreath GE and Chase HA. Puriﬁcation of G6PDH from unclariﬁed yeast cell homogenate using expanded bed adsorption (EBA) with streamlineTM Red HE7B. Adv Bioprocess Eng 1994;519–525. 155. McCreath GE, Chase HA, Purvis DR and Lowe CR. Novel afﬁnity separations based on perﬂuorocarbon emulsions. Use of perﬂuorocarbon afﬁnity emulsion for the puriﬁcation of human serum albumin from blood plasma in a ﬂuidized bed. J Chromatogr A 1992;597:189–196. 156. Cocker TM, Fee CJ and Evans RA. Preparation of magnetically susceptible polyacrylamide/magnetite beads for use in magnetically stabilized ﬂuid bed chromatography. Biotechnol Bioeng 1997;53:79–87. 157. Roy I, Pai A, Lali A and Gupta MN. Comparison of batch, packed bed and expanded bed puriﬁcation of A. niger cellulase using cellulose beads. Bioseparation 1999;8:317–326. 158. Somers W, van’t Riet K, Rozie H, Rombouts F and Visser J. Isolation and puriﬁcation of endopolygalacturonase by afﬁnity chromatography in a ﬂuidized bed reactor. Chem Eng J 1989;40:137–139. 159. Roy I, Sardar M and Gupta MN. Exploiting unusual afﬁnity of usual polysaccharides for bioseparation of enzymes on ﬂuidized beds. Enzyme Microb Technol 2000;27:53–65. 160. Sharma S, Roy I and Gupta MN. Separation of phospholipase D from peanut on a ﬂuidized bed of crosslinked alginate beads. Biochem Eng J 2001;8:235–239. 161. Roy I and Gupta MN. Puriﬁcation of a bacterial pullulanase on a ﬂuidized bed of calcium alginate beads. J Chromatogr A 2002;950:131–137. 162. Roy I, Jain S, Teotia S and Gupta MN. Evaluation of micro beads of calcium alginate for afﬁnity chromatography of Aspergillus niger pectinase. Biotechnol Prog 2004;20:1490–1495. 163. Tong X, Dong X and Sun Y. Lysozyme adsorption and puriﬁcation by expanded bed chromatography with a small sized dense adsorbent. Biochem Eng J 2002;12:117–124.

29 164. Hidayat C, Takagi MM and Yoshida T. Expanded bed adsorption for puriﬁcation of alcohol dehydrogenase using iminodiacetic acid matrix. J Biosci Bioeng 2004;97:284–287. 165. Haynes C. The importance of bioseparations: giving credit where credit is due. Biotechnol Bioeng 2004;87:257–258. 166. Przybycien TM, Pujar NS and Steele LM. Alternative bioseparation operations: life beyond packed bed chromatography. Curr Opin Biotechnol 2004;15:469–478. 167. D’Souza SF and Deshpande A. Simultaneous puriﬁcation and reversible immobilization of D-amino acid oxidase from Trigonopsis variabilis on a hydrophobic support. Appl Biochem Biotechnol 2001;5:83–92. 168. Sardar M, Roy I and Gupta MN. Simultaneous puriﬁcation and immobilization of Aspergillus niger xylanase on the reversibly soluble polymer EudragitTM L-100. Enzyme Microb Technol 2000;27:672–679. 169. Melo JS and D’Souza SF. A simple approach for the simultaneous isolation and immobilization of invertase using crude extracts of yeast and Jack bean meal. J Biochem Biophys Methods 2000;16:133–135. 170. Schoevaart R, Wolbers MW, Golubovic M, Ottens M, Kieboom APG, Van Rantwijk F, van der Wielen LAM and Sheldon RA. Preparation, optimization and structures of cross-linked enzyme aggregates (CLEAs). Biotechnol Bioeng 2004;87:754–762. 171. Sharma S and Gupta MN. Alginate as a macroafﬁnity ligand and an additive for enhanced activity and thermostability of lipases. Biotechnol Appl Biochem 2001;35:161–165.

31

Recent advances in all-protein chromophore technology Mark Prescott1,, Jion Battad1, Pascal Wilmann2, Jamie Rossjohn2 and Rodney Devenish1 1

Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia 2 Protein Crystallography Unit, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Abstract. The green ﬂuorescent protein (GFP) is the foundation of a powerful technology that has revolutionized the way in which the life scientist carries out experiments in the living cell. The technology is continually evolving and improving through the development of existing proteins and discovery of new members of the all-protein chromophore (APC) family. This review gives an overview of the more recent advances in the technology with a particular focus on APCs having optical properties that are significantly red-shifted relative to those variants derived from Aequorea victoria GFP. Keywords: green ﬂuorescent protein (GFP); chromoprotein; FRET; red ﬂuorescent protein; chromophore; all-protein chromophores.

Abbreviations avGFP A. victoria GFP APC all-protein chromophore CFP cyan ﬂuorescent protein CP chromoprotein FLIM ﬂuorescence lifetime imaging FLIP ﬂuorescence loss in photobleaching FP ﬂuorescent protein FRAP ﬂuorescence recovery after photobleaching FRET ﬂuorescence resonance energy transfer GFP green ﬂuorescent protein QY quantum yield RFP red ﬂuorescent protein YFP yellow ﬂuorescent protein

Introduction Fluorescent proteins (FPs), and in particular, the green ﬂuorescent protein (GFP) isolated from the North Atlantic jellyﬁsh, Aequorea victoria, have Corresponding author:

E-mail: [email protected] (M. Prescott). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12002-5

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

32 been the subject of scientiﬁc interest since it was recognized the lightproducing cells of this organism contained a green ﬂuorescent substance, later attributed to GFP [1]. However, it was not until the gene-encoding GFP from this jellyﬁsh was cloned, did the enormous potential of this family of proteins as genetically encoded ﬂuorescent reporters become evident [2]. FPs generate their chromophore as a result of a series of coordinated autocatalytic post-translational events that require no additional cofactors, with the exception of molecular oxygen [3]. It is this key property that enables a wide range of events in living cells to be followed in a non-invasive manner, with a variety of commonly available ﬂuorescence detection techniques [4]. A. victoria GFP (avGFP) has been extensively engineered and improved allowing it to be used in a variety of different conﬁgurations to follow a vast array of key cellular events in living cells and organisms [5,6]. Since the revelation that the vivid colours characteristic of coral reefs are a result of GFP homologues [7], in excess of 150 genes encoding all-protein chromophores (APCs) covering some 30 significant groups have been cloned from a variety of Anthozoa [8,9]. Many of the proteins they encode possess novel and biotechnologically useful properties that both extend and complement the capabilities of the original avGFP and its variants, making useful additions to the life scientist’s ‘toolbox’. Many homologues, often referred to as chromoproteins (CPs), are practically non-ﬂuorescent but intensely coloured. Although amino acid identity varies considerably across the family, CPs and FPs, share the same 11 stranded b-barrel structure enclosing the characteristic chromophore. It is for this reason that we will refer to this family of proteins as APCs. The success and wide applicability of APC technology, has resulted in a substantial volume of literature, making it necessary to restrict the scope of this review. This article will examine some developments in APC technology with a speciﬁc focus on those proteins that compared to avGFP and its variants, have significantly red-shifted optical properties. Reviews dealing primarily with avGFP are available [5,6]. We will summarize some of the ﬁndings revealed by structural biology approaches that contribute to an understanding of these novel intriguing properties of the red-shifted proteins and which underpin their further development. The quest for genetically encoded red ﬂuorescence APCs with ﬂuorescence emissions spanning a significant region (440–529 nm) of the visible spectrum have been developed [5]. Single amino acid substitutions in, or close to, the chromophore result in blue through to greenish yellow ﬂuorescence [5]. The multiple labelling of cells using such ﬂuorescent probes has proven a powerful approach that allows, more than one event to be followed in the cell simultaneously, or alternatively protein–protein interactions to be monitored using ﬂuorescence resonance

33 energy transfer (FRET) [10]. However, the blue FPs are notorious for the ease with which they photobleach. Moreover, illumination with UV light can generate autoﬂuorescence and lead to cellular damage in live-cell imaging applications. The FP pair considered for sometime to be optimal for FRET in many applications, cyan ﬂuorescent protein (CFP) and yellow ﬂuorescent protein (YFP), could be improved upon by the availability of FP pairs with more widely spaced and non-overlapping emission spectra [11]. It was clear that FPs with emissions significantly beyond YFP (529 nm) would be highly desirable. avGFP and red fluorescence The ﬁrst indication that APCs capable of producing emissions beyond 530 nm were possible, came from the serendipitous ﬁnding that avGFP and number of its variants when pre-exposed to UV light, could be excited by green light to produce a red-emitting species (lmax 600 nm) stable for over an hour [12]. This phenomenon has been exploited as an ‘optical highlighter’ for tracking spatial and temporal events in the cell. Since the phenomenon requires conditions of reduced oxygen tension, this approach is limited to cellular systems such as yeast that readily withstand periods of very low oxygen tension [12]. The molecular basis for this phenomenon remains to be determined. Non-natural amino acids in the chromophore Significant blue-shifts in the ﬂuorescence emission of avGFP can be obtained when the tyrosine, present in all naturally occurring APCs, is substituted by histidine [13], phenylalanine [14] or tryptophan [3]. It was therefore, reasonable to expect that different spectral properties might be obtained if nonnatural amino acid residues were substituted at this position. Although the standard genetic code limits the number of possible amino acids at this position to the canonical residues tryptophan, tyrosine, histidine or phenylalanine, a wide range of amino acid analogues have been incorporated by selective pressure incorporation (SPI) in Escherichia coli [15] or used in vitro translation systems [16]. Using SPI, tryptophan both in the chromophore tri-peptide and at position 57 of enhanced cyan ﬂuorescent protein (ECFP) as replaced with (4-amino)tryptophan to produce a protein, GdFP, with an emission spectrum red-shifted by 50 nm compared to EYFP (lmax 529 nm) (Fig. 1). GdFP produced a golden ﬂuorescence with a rather broad emission spectra with a maximum (lmax 574 nm) similar to DsRed (lmax 583 nm) [8]. Substitution with (4-methyl)tryptophan did not produce the same large stokes shift [17] and substitution of all 11 tyrosines in EGFP with (3-F) tyrosine produced only a 3 nm red-shift ðlmax ¼ 514 nmÞ [18].

34

Fig. 1. A selection of alternative chromophore structures found in APCs. GdFP contains the non-natural amino acid 4-amino tryptophan. The chromophores of DsRed and EqFP611 are extended by an acylimine bond and are distinguished by their respective cis- and trans-conformations with respect to the imidazole ring. Kaede requires UV illumination to generate the red-emitting chromophore by cleavage of the polypeptide backbone. asFP595 and zFP538 undergo backbone cleavage during maturation without illumination to generate their chromophore. For comparison, the chromophore structure of some GFP variants is shown in which the tri-peptide tyrosine in GFP is substituted by a tryptophan (CFP), histidine (BFP) or a phenylalanine (denoted BFPII in this review). The extent of conjugation in each chromophore is highlighted by shading. The ﬂuorescence emission maximum shown in parentheses for each chromophore type can vary depending on the protein variant. The polypeptide continuing towards the C-terminus (C) is shown. Amino acid side chains are shown (R).

In a separate study, using position-speciﬁc incorporation in cell-free translation, tyrosine of the chromophore tri-peptide was substituted, in turn by 18 different non-natural aromatic amino acids, some with multiple conjugated ring structures, (e.g., 2-anthraquinonyl-alanine) [16]. Remarkably, many of

35 these exotic structures were accepted into the correctly folded structure of the protein but, only two, those containing p-aminophenylalanine and O-methyl tyrosine were ﬂuorescent and these displayed a blue-shifted spectra relative to GFP [16]. It is possible that the non-ﬂuorescent variants evolved through a process of molecular evolution might better accommodate the larger substitutions leading to ﬂuorescence emission [16]. The reported optimization of those that were ﬂuorescent indicates that this might be a useful approach to achieving ﬂuorescence. New sources of APCS A. victoria is a bioluminescent organism. However, the vast majority of the newly cloned APCs are ﬂuorescent and coloured but, from non-bioluminescent members of the Anthozoa [9]. One particular protein drFP583, from a discosoma species, being the ﬁrst example of a commercially available red ﬂuorescent protein (DsRed), is of particular note. DsRed has a significantly red-shifted orange ﬂuorescence emission (lmax ¼ 583 nm), whose discovery has formed the basis of a range of new applications [8]. An FP, zFP538, showing true yellow ﬂuorescence emission was also isolated [8]. Subsequently, other proteins with red-shifted optical properties compared to avGFP have been isolated from natural sources (Table 1) and are detailed later in the review. Chromophore structures What is the molecular basis for the red-shifted ﬂuorescence emission of DsRed? The chromophore tri-peptide (SYG) present in avGFP leads to the formation of the characteristic conjugated double-ring structure, 5-[(4-hydroxyphenyl)methylene]-imidazolinione responsible for the optical properties of this protein [19] (Fig. 1). For comparison the conjugated chromophore structures in proteins resulting from substitution of the tri-peptide tyrosine with histidine, phenylalanine or tryptophan are depicted. The chromophore structure in DsRed derived from the tri-peptide QYG, resembles that of GFP but with the important difference that the p-bond conjugation system of the chromophore is extended by an additional double bond, namely an acylimine (Fig. 1). This additional feature is thought to contribute to the red-shifted properties of DsRed and other APCs [20–22]. It was deduced on the basis of classical protein biochemistry, supported by chemical quantum calculations [20] and conﬁrmed by the X-ray crystal structure of DsRed [21,22]. An acylimine appears to be a feature that contributes to the red-shifted ﬂuorescence of a number of proteins and was observed in the structure of the highly ﬂuorescent EqPF611 [23], the blue CP Rtms5 [24] and the far-red shifted HcRed [25]. Red-shifted optical properties can result from different

36 Table 1. Properties of some red-shifted APCs. The values for QY and extinction coefﬁcient can vary among reports in the literature, and are most likely a result of incomplete maturation of the chromoprotein under different conditions. APC

lmax excitation

lmax emission

Molar extinction coefﬁcient (M1 cm1)

QY

pKa

Maturation half-time t0.5 (1C indicated)

Reference

DsRed DsRed2 DsRed dimer2 DsRed T1 DsRed T3 DsRed T4 DsRed tdimer2 Kubisara-Orange (mKO) EqFP611 EqFP611 dimer T122R EqFP611 dimerV12

558 561 552 554 560 555 552 548

583 587 579 586 587 586 579 559

75,000 43,800 60,000 30,100 49,500 30,300 120,000 51,600

0.79 0.55 0.69 0.42 0.59 0.44 0.36 0.6

4.7 ND 4.9 4.8 ND ND 4.9 5.0

10 (37) 6 2 (37) 0.7 1.3 0.71 2 ND

[8,51] [37] [38,51] [51] [38] [38] [51] [57]

559 559 559

611 611 611

78,000 84,000 74,000

0.45 0.39 0.42

ND ND ND

4.5 (24.5) 7.5 (21) 7.5 (21)

[47] [53] [53]

HcRed HcRed dimer mRFP1 mOrange dTomato tdTomato mTangerine mStrawberry mCherry mRaspberry mPlum Rtms5H146S

598 590 584 548 554 554 568 574 587 598 590 592

645 637 607 562 581 581 585 596 610 625 649 634

125,000 160,000 50,000 71,000 69,000 138,000 38,000 90,000 72,000 86,000 ND 80,000

0.05 0.04 0.25 0.69 0.69 0.69 0.3 0.29 0.22 0.15 0.1 0.005

ND ND 4.5 6.5 4.7 4.7 5.7 o4.5 o4.5 ND ND 4.7

ND ND 4.5 2.5 h 1h 1h ND o1 50 min 100 min 55 min 3

[75] [52] [51] [54] [54] [54] [54] [54] [54] [76] [76] [24]

37 chromophore structures. Examples are the triple-ring chromophore revealed in the crystallographic structure of zFP538 [26], the imino-substituted chromophore present in asCP [27,28], or the chromophore extended through a neighbouring histidine in the photoconvertible Kaede protein [29].

Chromophore formation A scheme for the formation of the chromophore in GFP through cyclization of a tri-peptide and dehydrogenation has been documented [3]. Green ﬂuorescence predominates during early stages of DsRed maturation but diminishes as the emission develops [20–22,30,31]. Originally it was accepted that DsRed chromophore formation proceeded by an extension of the green GFP-like anionic intermediate [20]. However, based on more detailed investigations, a mechanism has been proposed where the ‘red’ chromophore in FPs and CPs forms through a common pathway that involves the neutral form of the GFP-like chromophore (B-form; Fig. 2) and not the anionic form as previously believed (G-form) [32]. The G-form once formed is trapped as it is not in equilibrium with the B-form and neither can it proceed to the predicted carbanion intermediate. This explains the persistence of green ﬂuorescence in mature DsRed and its variants, the basis of the ‘greening’ phenomenon [33] (see later). The B–G transition is presumably favoured in variants with reduced green emission. In the non-ﬂuorescent CPs, this pathway is more tightly controlled, as the G-form does not accumulate, explaining the absence of green ﬂuorescence in these proteins during maturation [25,32]. However, green ﬂuorescence can be observed if hydration of the acylimine is allowed to take place such as in the case of mild denaturation or storage of the protein. In DsRed, a proportion of the B-form is permitted to from the green ﬂuorescent anionic chromophore (G-form). The acylimine originally identiﬁed as the end product of chromophore maturation in DsRed [20] and, more recently other FPs [23] and CPs such as Rtms5 [24], may, in other cases, represent a ‘cross-road’ or intermediate leading to the formation of other chromophore structures. The yellow FP, zFP538 [8], has been to shown to contain a novel three-ring structure (Fig. 1) [27]. A scheme, yet to be conﬁrmed, for the formation of this chromophore has been proposed and includes an acylimine bond, possibly formed from a neutral GFP-like chromophore, as an intermediate that partakes in a transamination reaction of the lysine of the chromophore tri-peptide (Fig. 1). Maturation of the kindling CP, asFP595 and its variant KFP1, involves cleavage of the polypeptide backbone probably at the position of an acylimine to form a novel chromophore [28,34]. The reactivity of this acylimine together with the nature of the surrounding residues, likely explains the differing degrees to which polypeptide cleavage occurs in different acyliminecontaining proteins denatured under a range of harsh conditions [35].

38 1

2

O

O

3 O

O

N N OH

R

N

N N

N

O NH

O

O

Cyclisation -H20

+ O2

NH

R

OH

NH

R

OH

O

O

O

5

4

O

O

OH

N

O R

O

O

N N

[-]

N

NH

NH

R

OH

O

8

O

O

O

7

O

6

O

O

O

N N

N

N

R O

O

Hydrolysis H2 N

O

N HO H

N

R + H2 0

N O

O

R N O

O

Fig. 2. Proposed maturation pathway for chromophores extended by an acylimine

bond. Red-shifted chromophores containing an acylimine bond are formed by a sequence of steps that include a GFP-like chromophore intermediate. The polypeptide containing the chromophore tri-peptide (X-Tyrosine-Glycine) (Structure 1) cyclizes to form the characteristic ﬁve-membered ring (Structure 2) or C-form. An oxidation step takes place to form a conjugated chromophore (Structure 3) or B-form that corresponds to the neutral form of GFP. Depending on the protein, the further maturation proceeds by one of two routes. Formation of the green ﬂuorescent anionic form or G-form is irreversible (Structure 5). This species accounts for the green emission of DsRed. Formation of the carbanion is favoured [Structure 4] and maturation proceeds to the red or R-form (Structure 6). This route is favoured in red-shifted proteins lacking significant green emission or chromoproteins such as Rtms5. Under mild denaturing conditions, the acylimine can hydrate (Structure 6) leading to a GFP-like chromophore (Structure 7) that depending on the protein, results in a degree of fragmentation of the polypeptide backbone (Structure 8) upon boiling or treatment with acid or alkali.

Problems with the new APCS The availability of DsRed was greeted with considerable optimism. However, it was soon realized that as a ﬂuorescent reporter DsRed was far from ideal.

39 It is quite remarkable how accepting proteins are to the presence of relatively large adducts such as an APC, fused to their N- or C-terminus. The presence of the protein most often does not affect the function of the individual protein or interaction with other cellular components important to function. However, the literature does not in general reﬂect accurately the results of ‘failed’ experiments and shortcomings of the technology. Failures of FP technology can arise for a number of reasons. The protein to which an FP is fused may not fold correctly, trafﬁc to the correct location, or may prevent interactions with important partner proteins. Severe problems are seen to arise as a result of forced illegitimate interactions, driven either by nonspeciﬁc aggregation or alternatively by more ‘speciﬁc’ oligomerization of the APC. Slow maturation of the chromophore and the appearance of other contaminating ﬂuorescence signals can also be a problem. It must be remembered that GFP was not without its problems when ﬁrst introduced [11] and required considerable optimization [5]. Furthermore, the problems encountered to some extent appear common to all APCs. Below we give an overview of the stepwise optimization of an APC using DsRed as an example. Reducing aggregation Heterologous expression of DsRed, and in general Anthozoa FPs and CPs often leads to the formation of aggregates even at relatively low protein expression levels [36]. Precipitation can take place in the test tube without loss of ﬂuorescence or colour and, can be observed on pseudonative gels [37] indicating that the protein probably folded correctly. This troublesome phenomenon was eliminated in DsRed and a number of other homologous tetrameric proteins using a general strategy of introducing selected amino substitutions near the N-terminus. Aggregation appears to be driven by charge interactions between basic positively charged residues near the N-terminus and negative charges present on the rest of the molecule of neighbouring protomers within the tetramer [37]. A non-aggregating variant of DsRed is commercially available as DsRed2 (BD Biosciences), which is two-fold brighter than DsRed1. A random mutagenesis-based screen for proteins with improved maturation (see below), identiﬁed similar amino acid substitutions in the same N-terminal region of DsRed [38]. Improving chromophore maturation For DsRed maturation to the red emitting form is particularly slow, with a half-time in excess of 24 h at room temperature [30,39]. This most often increases the time required after expression before useful ﬂuorescence is observed in cells, a particular problem when expressed in an organism such as yeast that has a short doubling time (2 h). In dual-labelling experiments with GFP,

40 weak green emission resulting from the immature DsRed chromophore bleeds through into the GFP channel making quantitation of that signal unreliable. Random mutagenesis-based screening approaches have identiﬁed a number of residues important in accelerating the maturation rate [38]. A key substitution, N42Q, whose side chain faces towards and interacts with Q66 of the chromophore increased the rate of chromophore maturation by 10–15 fold [21,22]. Additional substitutions further increased the maturation rate and reduced the green emission enhanced by the N42Q substitution rate, such that variants with maturation halftimes of 0.7–1.3 h were obtained. The commercially available DsRed Express (BD Biosciences) is based on the variant DsRed-T1. A yeast-optimized form of DsRed, RedStar, has been reported to be 10–20 times brighter than DsRed1. Significant green emission is suppressed in these variants [40]. A degree of lateral thinking can often identify beneﬁts in what are otherwise considered undesirable properties. The slow maturation rate of DsRed inspired development of a ‘Fluorescent Timer’, DsRed-E5 [31]. A key amino acid substitution (V105A) close to the DsRed chromophore resulted in a protein that changed colour from green through yellow to red over time, independently of protein concentration. The ﬂuorescent timer allows timing or ‘history’ of promoter expression to be spatially imaged in developmental models, such Xenopus, C. elegans [31], or plants [41]. Oligomerization All the newly identiﬁed ACPs with red-shifted optical properties form tetramers [8,9]. This need not be a concern to the researcher, providing the application does not requires following the post-translational function of a protein-fusion. Indeed, in the case of certain APC applications such as the ﬂuorescent timer [31], or greening [42], the desirable property of the particular protein is intrinsic to its oligomeric organization. Oligomerization of APCs, in contrast to aggregation, takes place as a result of speciﬁc interactions between individual protomers that contribute to the quaternary structure of the protein [21,24,25,28] (Fig. 3). In the case of DsRed oligomerization can be considered as an arrangement of a pair of dimers. Interaction is stabilized through two chemically distinct interfaces. Many of the GFPs isolated from marine organisms are dimeric [43]. Fortunately, the commonly used avGFP behaves as a monomer in moderately concentrated solution but, can form dimers at high protein concentrations. In the vast majority of applications reported in the literature, this did not appear to be a major cause for concern. Nevertheless, in order to make avGFP variants suitable for monitoring even the weakest of protein associations, such as between membrane-anchored proteins, this weak dimer interaction has been eliminated by introducing amino acid substitutions into the two commonly used A. victoria variants, CFP and YFP [44].

41

Fig. 3. Different oligomeric states of APCs. The arrangement of protomers in tet-

rameric, dimeric and monomeric APC is shown. The position of one of the two hydrophobic interfaces (unﬁlled arrow) and one of the two hydrophobic interfaces (ﬁlled arrow) is shown for the four identical protomers in the non-ﬂuorescent Rtms5 (A) [24]. A similar arrangement of protomers was found in tetrameric proteins EqFP611 [23] and DsRed [21]. A single mutation in the hydrophobic interface of HcRed disrupts the hydrophobic interface to produce a dimer, leaving the hydrophilic interface intact (B). The hydrophilic interface in APCs is more difﬁcult to disrupt and often leads to loss of chromophore function. The structure of a monomer derived from an obligate oligomeric protein has yet to be published. The structure of the avGFP protomer is shown (C). Note the marked elliptical shape of the barrel in oligomeric proteins compared to the rounded shape for GFP. Structures are shown in ribbon form with the chromophore depicted in ball and stick format.

Overcoming oligomerization Oligomerization of the APCs in general, compared to avGFP and its variants severely limits their potential as reporters. Eliminating oligomerization has been the focus of a number of research groups. DsRed was found to be an obligate tetramer even at nanomolar concentrations [30,39,45]. Monomers of DsRed can only be formed through completed denaturation. However, ﬂuorescent dimers could be generated upon renaturation of DsRed, suggesting it might be possible to develop DsRed dimers with functional chromophores or perhaps isolate such proteins from natural sources [46] (Fig. 4). What approaches have been used to overcome this shortcoming? Perhaps, the most obvious approach is to isolate from natural sources further novel proteins in the hope that one might display reduced oligomerization. EqFP611, from the sea anemone Ectamacea quadricolor, was reported to have properties of a monomer but only at concentrations far below those practical for its use as an in vivo reporter [47]. A naturally non-oligomerizing ACP with red-shifted properties has yet to be isolated from natural sources.

42 (A)

(B)

(D)

(E)

(C)

Fig. 4. Overcoming oligomerisatin of APCs. Oligomerization of APCs can drive

inappropriate interactions between proteins to which they are fused (A). The preferred ﬂuorescent monomer (C) was sequentially engineered via a ﬂuorescent dimer (B). Strategies for the formation of pseudomonomers can suppress oligomerization of fusion proteins by co-expression of an excess of competing un-fused protein lacking a functional chromophore (unﬁlled large square) (D) or expression of dimeric APCs fused as a tandem dimer, favouring intra-molecular interactions (E).

This lack of success may well be linked to an as-yet unrecognized function in the native organism requiring oligomerization. The undesirable association of APC fusion proteins can, to some extent, be suppressed through the co-expression of a competing un-fused APC lacking a functional chromophore (i.e., non-ﬂuorescent) is with the fusion protein. Monomers from this ‘invisible’ non-ﬂuorescent pool are freely available to compete for inclusion into the forming oligomers and, if present at appropriate concentrations, will reduce significantly the level of oligomerization of the fusion protein and hopefully restore function of the fusion protein to that of the wild-type endogenous protein. This approach has the advantage that, given the availability of a non-ﬂuorescent variant, it should be applicable to any APC that forms oligomers. The oligomerization of DsRed and NM355NA, a bright variant of asFP595, have been effectively reduced using this approach [48,49]. The disadvantage of this approach is that the pseudomonomeric fusion proteins are burdened by the presence a tetrameric APC appendage, which could in some cases be deleterious to function. As an alternative approach, coexpression of the un-fused partner protein can eliminate oligomerization and restore biological function to the fusion protein [50]. The availability of the structure of DsRed has guided efforts for engineering APC with reduced tendency to oligomerize [21,22]. Nevertheless, engineering ﬂuorescent ACPs to be monomeric has proved a particularly challenging task [51]. Converting ﬂuorescent tetramers into ﬂuorescent dimers appears to been a somewhat easier task for certain proteins. A single substitution in the hydrophobic interface A/B of a number of different tetrameric proteins was sufﬁcient to generate the following dimeric APCs: HcRed [25,52], dEqFP611 [53] and Rtms5 [24] (Fig. 5). Although a significant improvement over tetrameric

43 (A) Donor

(B)

YFP CFP

Acceptor

(C) mRFP

CFP YFP

HcRed

Fig. 5. Multiple interacting partners monitored by FRET. Red-shifted proteins act as FRET acceptors to facilitate three-way FRET conﬁgurations. The commonly used, binary or two-way FRET conﬁguration of donor and acceptor is shown (A). A conﬁguration for three-chromophore (CFP, YFP, mRFP) FRET is shown (B) that is monitored using intensity-based methods. Time-resolved ﬂuorescence measurements (FLIM) allow the interaction of two donors with a single non-ﬂuorescent or weakly ﬂuorescent acceptor to be monitored (C). The tandem dimer of HcRed shown here could be replaced by a monomer variant. This approach could be readily adapted to follow multiple independent donor/acceptor pairs using the same non-ﬂuorescent acceptor.

proteins, the potential to dimerize remains. However, this can be overcome by expressing tandem fusions that behave as pseudomonomers; intra-molecular dimerization being favoured, therefore blocking inter-molecular interactions between fusion proteins. This approach has proven to be a remarkably effective stopgap solution to the problem [51–53]. However, sometimes the optical properties of the dimer might be preferable. Compared to its monomeric equivalent, mRFP1, tdTomato [54] has a high quantum yeild (QY) and extinction coefﬁcient (Table 1). The high extinction coefﬁcient of HcRed is an important pre-requisite when using APC as acceptors in ﬂuorescence lifetime imaging (FLIM) applications [55]. A single substitution in a similar position in the A/B protomer interface of DsRed at ﬁrst resulted in a variant with slow maturation, but whose properties were recovered to some extent following further mutagenesis and selection. The same substitutions were incorporated into the fast maturing DsRed.T1 variant [38] and following further rounds of mutagenesis and selection, a vastly improved dimer2 variant resulted containing a total of 17 substitutions [51]. After further rounds of mutagenesis and selection a redﬂuorescent monomer, mRFP1, was isolated bearing a total of 33 substitutions compared to the tetrameric DsRed. mRFP1 matures rapidly and has minimal emission when excited at wavelengths suitable for GFP making it suitable dual labelling studies. The 10-fold increase in maturation rate compared to DsRed helps to offset the undesirable increased bleaching rate and the lower QY and extinction coefﬁcient compared to DsRed (Table 1) probably giving it equivalent brightness in cells. Through these studies a wealth of

44 data has accumulated informing the conversion of other tetrameric proteins into ﬂuorescent monomers; for example, mAG, Azami green [56] or mKO, Kubisara-orange [57].

Improved monomeric proteins Using mRFP1 as a new starting point and the accumulated knowledge from previous studies, a palette of new monomeric FPs with emissions from honeydew yellow to cherry red have been obtained by further rounds of directed evolution [54] (see Table 1). mCherry is of particular note as it photobleaches 10-fold less easily than mRFP1, appears to mature more completely as indicated by the absence of the 510 nm emission present in previous mRFP proteins and, has similar emission maximum (610 nm) compared to mRFP1 (607 nm). In addition, the apparent sensitivity of mRFP1 to N-terminal fusions has been suppressed by replacing some of the N-terminal residues with those of EGFP, a protein, which does not suffer this drawback. The improved properties of mCherry effectively make mRFP1 obsolete. mOrange (lmax 562 nm) represents a considerable improvement in QY (0.69) and improved brightness over mRFP1 and promises to be useful for applications that involve dual-emission FRET. However, mOrange may show some sensitivity to pH depending on its application as it has a pKa (6.7) close to physiological pH. Other colours are available for multi-colour labelling.

Oligomeric FPs as in vivo cross-linking agents Although usually considered to be an undesirable property, the ability to alter or drive interactions between proteins in vivo can yield useful information. DsRed expressed fusions to individual subunits of yeast ATP synthase complexes embedded in the inner-membrane of the mitochondrion resulted in altered oligomerization of these complexes and collapse of the cristae visualized in vivo [48]. Although believed to be normally oligomeric in the membrane, it was apparent the correct organisation of complexes is essential for formation and maintenance of the cristae. Although there are other approaches for ‘cross-linking’ in vivo, none have the advantage of producing a ﬂuorescent signal upon cross-linking. The level of oligomerization can be deﬁned by substituting for other APCs such as the dimeric HcRed for DsRed (Gong and Prescott, unpublished data). It may be possible to develop this approach by engineering FPs whose oligomeric state can be regulated through interaction with speciﬁc ligands or conditions. For example, the oligomerization of the chromprotein Rtms5F162H, which is pH dependent (Prescott, unpublished observations).

45 Stability of fluorescent proteins in vivo The ﬂuorescence emission of avGFP and its variants once correctly folded, has been shown to be remarkably stable when exposed to a number of different harsh treatments including denaturants [58–60], temperature [61] and proteases [2]. This increased stability has proven to be a problem in certain applications. For example, if attempting to follow the cyclic dynamics of promoter activity, the persistence of the signal generated can mask subsequent changes in expression. In order to overcome this, masking strategies have been reported for increasing the degradation rate of these proteins in vivo, including the addition of appropriate proteolytic signals. Destabilized forms of GFP have been generated through the addition of appropriate proteolytic signals that deliver the protein to the proteasome for degradation [61,62]. The in vivo half-life could be reduced from 26 h to between 2 and 10 h depending on the application. This strategy for reducing stability also has been applied to DsRed and HcRed to reportedly yield proteins with reduced half-lives [63]. On the other hand, increased stability of an APC can result in its accumulation and increased ﬂuorescence intensity. This is useful for the tracking of long-lived cells in intact organisms. The stability of DsRed has been investigated in vivo and in vitro [61,64]. Tetrameric DsRed is inherently more stable than EGFP, showing dramatic differences in conformational stability, both in vitro and in vivo, although both were found to unfold unusually slowly taking several days to reach equilibrium [61]. It is likely that for homologous proteins the more stable variants will persist longer in vivo. This certainly appears to be the case. The estimated half-lives for stably expressed EGFP and DsRed in Drosophila S2 cells were estimated to be 1 and 4.6 days, respectively. Increased stability of DsRed transiently expressed in HEK293 cells or Drosophila embryos was also observed [61]. It has been suggested that increased stability of DsRed compared to EGFP is a result of the oligomeric nature and speciﬁc packing of the monomers in the oligomer [61]. However, it was subsequently demonstrated that monomeric mRFP1 showed comparable stability to DsRed, and that a tetrameric GFP (zFP506) was the most stable of all proteins investigated in the study, suggesting oligomerization is not the only parameter controlling stability. Increased stability would also favour certain applications in vitro. The ongoing development of ‘ﬂuorobodies’ in which diverse antibody binding loops are inserted into the loops at one end of the b-barrel of an APC, requires a stable and rigid framework on which to build [65]. The non-ﬂuorescent CPs display intense coloration and share the same increased stability as FPs lending themselves to development of non-ﬂuorescent colour readouts. pH stability An important consideration when using FPs is the pH of the environment in which they are used. This is important if the application is required to be

46 quantitative particularly when imaging live cells, where the pH alters in a dynamic and localized manner. Measurements of FRET can contain errors if the ﬂuorescence emission of the APC is significantly dependent on pH. Considerable effort has been invested in developing variants of avGFP having emissions less sensitive to pH [66], or that are better reporters of pH [67,68]. avGFP variants in which the tyrosine of the chromophore tripeptide has been substituted with histidine, phenylalanine or tryptophan are relatively insensitive to pH due to absence of the anionic form of the chromophore [5]. In contrast, many of the new red-shifted variants have ﬂuorescence emissions that are remarkably stable to pH presumably due to stabilization of the anionic ﬂuorescent form of the chromophore [20,21]. EqFP611 [47] and DsRed [20] have useful red emissions in the range pH 4–11. Above pH 11, the extended conjugation of the chromophore is lost due to hydration of the acylimine bond (Fig. 2). Below pH 4, the chromophore becomes non-ﬂuorescent due to the predominance of the neutral form of the chromophore. However, particular red-shifted proteins are not quite as stable to pH. For example, the QY of HcRed increases four-fold between pH 8 and 10 (Prescott, unpublished results), while the emission of a photoactivatable variant of mRFP1, pA-mRFP, is pH dependent [69]. pH biosensors with red-shifted emissions might be useful for imaging events in tissues but have not yet been reported. In general, there are few examples of red-shifted proteins having been tailored for sensing speciﬁc events. This is in contrast to the range of events that avGFP has been modiﬁed to monitor such as redox [70], or halide sensing [71]. Now, we have acquired the knowledge to overcome many of their undesirable properties, such as aggregation, oligomerization and slow maturation, further developments will surely follow for red-shifted proteins. The increased complexity of the chromophore and its chemistry will need careful consideration prior to any engineering attempts, but the novel properties deriving from the chromophore may lead also to new applications. For example, APCs containing the labile acylimine bond may not be suitable in multi-colour ﬂuorescence complementation assays for protein–protein interactions [72]. Far-red FPS The availability of APCs with red-shifted emissions represents a significant advance in the technology. However, it was considered highly desirable to have access to FPs with, yet further red-shifted, or even far-red shifted emissions. Far-red shifted FPs are operationally deﬁned here, and in line with their description in the literature, as those with emission maxima of 610 nm or above. Such proteins extend further, the palette of FPs from which to optimize multiple FP applications. Furthermore, their emissions being further removed from interference caused by autoﬂuoresence will be attenuated to a lesser degree in biological material allowing imaging deeper into tissue.

47 Several approaches have been taken to obtain far-red shifted proteins. These include (a) the isolation from natural sources of highly FPs; (b) engineering the QY of weakly FPs or (c) increasing the red-shift of existing highly FPs. The highly ﬂuorescent EqFP611 (QY, 0.45) isolated from the sea anemone, E. qaudricolor (Table 1) represents a moderate red-shift in emission (lmax 611 nm) [47]. Dimer variants of EqFP611 are now available making it more user-friendly [53]. The emissions observed in corals extend up into the far-red (650 nm) [73]. CPs have been identiﬁed as responsible for the intense diverse colours of coral reefs [7,74]. Although essentially non-ﬂuorescent by comparison with other FPs, upon closer examination it was apparent that certain APCs produced weak but promising emissions that extended well beyond those of DsRed [75]. Substitution of a single amino acid around the chromophore increased the QY several hundred-fold converting a number of different CPs into far-red FPs with emissions from 615 to 640 [75]. A similar substitution (histidine to serine) increased the QY 180-fold of the intensely blue CP Rtms5 [24]. A combination of random and site-directed mutagenesis approaches when applied to hcriCP produced further improvements in QY at 645 nm. Furthermore, a single substitution resulted in the conversion of the tetramer into a ﬂuorescent dimer [25,75]. Although the QY of HcRed (0.04) is low (Table 1) compared to what is considered acceptable by the standard of other FPs, this property appears not to have prevented its successful application. HcRed shows a 60 nm red-shift compared to DsRed. Tandem dimers of HcRed can be used as pseudomonomers [52]. CPs and their derivatives during the process of maturation appear to not produce the ‘green emitting’ species common to DsRed [25,32]. HcRed can be safely used in combination with other FPs emitting in the green part of the spectrum. However, we have seen that HcRed upon storage can lead to the production of green ﬂuorescence probably due to the hydration of the acylimine to form a ‘GFP-like’ chromophore (Fig. 2). Use of CPs as a starting point for engineering FPs has not yielded further proteins with a QY above that of HcRed. The reason for the low QY appears related to the conformation of the chromophore. The chromophore in Rtms5 is non-coplanar accounting for its lack of ﬂuorescence (Fig. 6) [24]. The significantly higher QY (0.04) for HcRed correlates well with the increased ﬂexibility of the chromophore and increased proportion of the ﬂuorescent coplanar chromophore observed in the crystal structure. Increasing the content of the coplanar chromophore in these proteins would presumably lead to yet higher QY. Formation of the chromophore involves a unique sequence of temporal and spatial events driven by the unique organisation of amino acid side chains in the folding protein. The task of exploring such ‘mutagenesis space’ using routine mutagenesis techniques represents an enormous task and possibly explains why mutagenesis of CP templates to produce variants having

48 1

2

1

3

3

(A)

(B)

(C)

Fig. 6. Some chromophore conformations in APCs. (A): Three distinct conforma-

tions of chromophore determined by X-ray crystallography are illustrated superposed is (DsRed), 1; trans (EqFP611), 2; and trans-non-coplanar (Rtms5), 3. (B): An orthogonal view shows more clearly the non-coplanarity of the non-ﬂuorescent Rtms5 chromophore, 3. (C): The chromophore in the moderately ﬂuorescent HcRed is inherently ﬂexible and both the cis- and trans-non-coplanar conformation are found in the population.

increased red-shift had only limited outcomes. However, this bottleneck has recently been eliminated for the evolution of FPs to longer wavelength emissions using a process of iterative somatic hypermutation. Using this approach, the monomeric DsRed variant mRFP1.2 [51] has been evolved to yield two new variants, mStrawberry and mPlum [76]. The new variant mPlum is of particular note as it has a number of remarkable properties. It is the most far-red shifted FP reported to date with an emission maximum of 649 nm, exceeding that of HcRed by 3 nm and, with an unusually large stokes shift of 59 nm. mPlum is also estimated to be 30-fold more photostable when compared to its parent mRFP1.2. Remarkably, these dramatic alterations in optical properties were achieved through a relatively small number of substitutions and only four of which have a side chain facing towards the chromophore and thereby presumably subtly affecting the chromophore environment. The underlying mechanism for the increased redshift of mPlum is at present unclear, but will most likely become clearer with information arising from structural determinations of these proteins. Near infra-red fluorescent proteins? Mammalian tissues display the lowest absorption coefﬁcient in the near infrared window (650–900 nm) [77,78]. mPlum being the most red-shifted (lmax 649 nm)

49 APC available is currently the best choice for imaging applications in vivo. However, APCs with greater redshifts would be welcome. Two independent approaches have yielded proteins, HcRed and mPlum with similar emission maxima. It is possible that 650 nm represents the limit of redshift possible with the current chromophore chemistry and that new modiﬁcations to the chromophore will be required if further redshifts are to be obtained. At present, we have insufﬁcient information concerning the additional steps required to produce APCs with extended conjugation systems having peak emissions in the near infra-red (> 650 nm). Perhaps, the inclusion of suitable non-natural amino acids in or near the chromophore may be the only way to put in place a suitable chromophore. Such an approach, if to be any real value, would require significant parallel developments in strategies for alternative codon utilization, and delivery into live cells of non-native amino acids. Nevertheless, it would be worthwhile to probe the limits of red-shift, and investigate the effects of making selected non-native amino acid substitutions into the some of the available far-red FPs. Photoconvertible proteins FPs can undergo a number of different types of photoinduced alterations of their optical properties. Probably one of the best known, and often the bane of the ﬂuorescence microscopist, is the loss of ﬂuorescence emission upon illumination or photobleaching. Intense and prolonged illumination at the excitation/absorption maximum results in alterations in the chromophore, often leading to destruction of the chromophore. However, spatial photobleaching of cells expressing FPs has been developed in such techniques as ﬂuorescence recovery after photobleaching (FRAP), and ﬂuorescence loss in photobleaching (FLIP) to follow diffusion of FP-labelled molecules [79]. A more an elegant approach, optical marking, generates new distinct ﬂuorescence signals without destruction of the chromophore. These approaches are becoming popular and are based on deﬁned alterations in the excitation and/ or emission spectra of particular FPs. A range of APCs is available for optical marking many of which are based on red-shifted optical properties (Table 2). For practical purposes, they can be grouped to illustrate their general optical properties (see below). The grouping does not reﬂect any common underlying mechanism; members of each group may have distinct mechanisms. An outline of each mechanism is discussed brieﬂy here and covered in further detail, in selected cases in a later section. 1. Essentially non-fluorescent and yielding red emissions on photoconversion. e.g., PA-mRFP and KFP1 show red and far-red emission, respectively. 2. Increase in green to red/emission on photoconversion in a process termed ‘greening’. e.g., DsRed.

50

Table 2. Some properties of photoconvertible proteins (I ¼ irreversible; R ¼ reversible). Initial excitation/ emission (nm)

Photoconverting QY pre-/postContrast ratio Reversibilty light (nm) photoconversion post-/prephotoconversion

KFP 1

600/600

Intense 568 nm

o0.001/0.07

70

DsRed

558/583

o760

0.69

DsRed 2

562/589

o760

DsRed N42Q (Greening)

560/583

PA-GFP to green 400/520480/520 wtGFP, GFP 395/ S65 T, GFPmut2

Number of subunits

Reference

Comments

R and I

4

[85]

3.3

I

4

[33]

0.62

13.8

I

4

[33,42]

o760

0.47

11.1

I

4

[42]

400 390

ND ND

100 ND

I 1 Stable for>24 h 1

[92] [12]

Undergoes reversible and irreversible photoconversion depending on duration and intensity of illumination Requires multiphoton illumination at 760 nm and single photon instrumentation. Stable for hours or days. Requires high power to photobleach Improved greening properties compared to DsRed Improved greening properties compared to DsRed Irreversible Requires reduced oxygen conditions

PA-mRFP1-1

588/602

340–380

0.001//0.08

70

R

1

[69]

PS-CFP

402/468 490/511

390

0.16/0.19

1500

I

1

[83]

Kaede

480/518 480/582

365

0.8/0.33

2000

I

4

[80]

wtEosFP d1EosFP d2EosFP mEosFP

560/516 505/526 506/516 505/516

390 390 390 390

0.70/0.55 0.68/0.62 0.66/0.60 0.64/0.62

ND ND ND ND

I I I I

4 2 2 1

[82] [86] [86] [82]

AceGFP-G222E

280

0.07/0.45

1000

N

I

[111]

Dronpa

390/460+505 480/ 505 510

488

0.85

R

1

[112]

KikGR

505/593

350–420

0.7

I

4

[81]

571/581 571/581 569/581 569/581

ND

Emission may be pH dependent. Half-life of 9 h at 37oC Dual colour emission providing similar advantages to Kaedelike proteins Uses different light for photoconverision and visualization Pseudomonomer Pseudomonomer Will not mature above 30oC

Undergoes many cycles of photoconversion More efﬁcient at low pH. Several fold brighter than Kaede

51

52 3. Alterations in excitation spectra but with a single emission. e.g., PAGFP, green emission. 4. Alteration in both excitation spectra and emission spectrum. e.g., Green to red for the Kaede [80], KikGR [81] and mEosFP [82] proteins. Cyan to green emission for PS-CFP [83], avGFP and EGFP.

Properties of some selected photoconvertible APCS The ‘greening’ of DsRed is a unique feature related to the tetrameric nature of the protein and the inherent incomplete maturation of the protein. DsRed even when mature contains a mixture of the GFP-like green light emitting intermediate and the fully mature ‘red’ emitting form [20,22]. An equal mixture of the two forms can be discerned in the high-resolution crystal structure of DsRed [22]. Highly efﬁcient FRET (68%) between the emission of the green and red form in the tetramer effectively masks the presence of the green form [30]. Obliterating each of the red-emitting chromophores (the acceptor) in each tetramer by photobleaching eliminates FRET, allowing emission of the immature green form (the donor) to be followed by excitation with blue light (488 nm) of an argon ion laser. Although, effective greening of DsRed requires the high power of input of a multi-photon laser (lo760 nm) to achieve a high level of photobleaching of the red ﬂuorophores, this approach allows for very precise highlighting of a small volume in the cell characteristic of this mode of illumination [33]. Particular DsRed variants, DsRed N42Q [38] and DsRed2 [37] have been evaluated, which show an optimized greening effect [42]. It should be noted that greening is dependent of the formation of tetramers and therefore excludes, for reasons already discussed, their use as fusion proteins. Particular APCs although classiﬁed for practical reasons as non-ﬂuorescent can give rise to traces of red ﬂuorescence, when illuminated with green light in a process called ‘kindling’. The red emission intensity increases upon prolonged illumination with intense green light [34]. The kindled state in these proteins is transient and decays once illumination has ceased. The decay of the kindled state can be accelerated by illumination with blue light. The residues around the chromophore that contribute to the phenomenon have been investigated and amino acid substitutions identiﬁed that accentuate the phenomenon and result in a degree of permanent kindling to produce irreversible photoconversion [84]. KFP1 is a commercially available variant of asCP595 optimized for kindling [75,85]. It was proposed, based on molecular modelling and structural studies, this phenomenon involves a trans (nonﬂuorescent) to a cis (ﬂuorescent) alteration in the chromophore conﬁguration [27,28,84,85]. It should be noted that KFP1 is tetrameric and therefore not suitable as a fusion partner. This mechanism is discussed in a later section in further detail in relation to the structure of these proteins.

53 Several proteins, Kaede, KikGR and EosFP, are available that irreversibly change colour from green to red when illuminated with near-UV light. The photoconverted and non-photoconverted forms of these proteins are readily tracked as the two forms of the chromophore have high QY and are spectrally distinct. Furthermore, the photoconverting light is distinct from the wavelengths of light used to excite both the green and red emission. KikGR is a new protein developed from the non-photoconverting KiKG using semi-rational mutagenesis [81]. Compared to Kaede, KikGR is more efﬁciently photoconverted and has several-fold brighter green and red emission. KikGR can be readily photoconverted using two-photon excitation at 760 nm. The existence of green- and red- emitting species in the same cell might be considered by some a disadvantage as it restricts the choice of other ﬂuorophores in multilabelling experiments. EosFP has similar properties to Kaede and KikGR with the added bonus that it exists as a monomer, while retaining its photoconversion properties [82,86,87]. However, the monomer has restricted use in that it will not express to produce a functional chromophore above 301C. The tandem-dimers, d1EosFP and d2EosFP, which behave as pseudomonomers, overcome the thermosensitivity maturing efﬁciently at 371C [86]. The main amino acid determinants of photoconversion in tetrameric proteins [84,88] have been introduced by substitution into mRFP1 [51] resulting in PA-mRFP1-1 [69], which is monomeric and therefore, suitable for labelling of proteins. Illumination with UV light 340–380 nm results in a 70-fold increase in red emission [69]. Red-shifted APCs and FRET applications FRET between APCs is revolutionizing widespread detection of protein–protein interactions as they occur in single cells. FRET is the distance-dependent transfer of energy through dipole–dipole interactions between a donor ﬂuorophore and an acceptor ﬂuorophore. It has been applied to great effect for the imaging of intra- and inter-molecular protein–protein interactions using appropriate pairs of avGFP variants [89]. In theory, many of the different avGFP variants may be used for FRET applications [90,91], but the most popular donor/acceptor pair is YFP and CFP. The availability of red FPs has extended the range of proteins suitable for FRET and provides the opportunity to reduce the undesirable spectral overlap between the emission spectra of the donor and acceptor FPs, thereby simplifying the analysis. A protease FRET biosensor based on the EYFP/HcRed pair has been reported [52]. However, the low QY of HcRed makes it impractical to follow FRET through the emission of HcRed. DsRed in its undesirable oligomeric form has been used in FRET studies [92] and strategies have been developed to compensate for the slow maturation of the chromophore [93]. The fast maturing pseudomonomeric dimer 2 [51] has been used paired with EGFP to follow intra- and inter-molecular protein interactions [94].

54 DsRed [8,54] and some new mRFP variants [54] are not excited to any significant extent with violet light (405 nm) and make a good choice of acceptor when used together with green-emitting Sapphire GFP variants as donors, which excite optimally at this wavelength. Such combinations make it possible to completely separate donor and acceptor excitation and emission in intensity-based FRET experiments. A number of optimized variants of Sapphire have been developed [95]. The mOrange-T-sapphire, when linked to a Zn2+ ﬁnger motif, was found to respond to Zn2+ in fashion similar to the equivalent and better characterized CFP–Citrine pair but a with a longer wavelength of emission. mKO, a monomeric orange-emitting APC, isolated from the stony coral Fungia concinna has been successfully used in conjunction with a new dimeric cyan-emitting APC [57] (Table 1). This pair does not suffer from crossexcitation of acceptor and the cross-detection of the donor and FRET signals suffered with the CFP/YFP pair. Most FRET applications reported in the literature using avGFP FRET pairs have involved the study of binary protein–protein interactions. However, cell biology is underpinned by multiple dynamically interacting protein molecules. The availability of red-shifted proteins has allowed strategies to be developed for monitoring the simultaneous interaction of more than one protein. In the 3-chromophore FRET protocol, parallel and sequential energy transfer between YFP, CFP and mRFP1 can be distinguished using basic methods [10] (Fig. 5). The high-extinction coefﬁcient and broad absorption spectrum of HcRed shows significant overlap with the emission spectra of a number of ﬂuorescent donors making HcRed attractive as FRET acceptor. The problem of lowacceptor QY can be overcome by using FLIM where time-resolved ﬂuorescence measurements are used to monitor FRET. In this approach, the emission of the donor alone is analyzed. Energy transfer results in a lowering of the ﬂuorescence lifetime [96]. The simultaneous interactions of EYFP and ECFP with tandem-HcRed has been imaged in the same cell and used to follow simultaneously the activation of two Ras isoforms [55,97]. Since the acceptor need not be ﬂuorescent, the high-extinction coefﬁcient and broad absorption spectra of CPs such as Rtms5 [24] are well suited to such applications. The development of monomeric variants of these CPs would facilitate this approach. The implementation of FLIM requires highly specialized and costly instrumentation [96]. However, the availability of very high intensity lightemitting diodes suitable for modulation at MHz frequencies has resulted in the development of economical systems operating in the frequency domain [98]. The availability of monomeric FPs covering much of the visible spectrum will allow multiple events to be tracked simultaneously, many of which may involve the analysis of FRET. The use of non-ﬂuorescent acceptors together with FLIM will help increase the number of events that can be simultaneously monitored in the same cell.

55 Monitoring the emissions of multiple APCs Cells are tightly integrated networks of biological processes. The power of APC technology to interrogate these networks has been demonstrated. However, to understand these networks in more depth it will become increasingly important to follow simultaneously, more than one event in the living cell. For example, one might wish to follow in the same cell Ca2+ ﬂuctuations via a Chameleon FRET biosensor [99] together with one or more other protein–protein interactions via FRET. Although the availability of an extended range of APCs provide a rainbow of spectrally distinct ﬂuorophores, the emissions in such sophisticated applications need to be resolved. Fortunately, approaches are being developed to help with such analysis. Using technology borrowed from satellite imaging and remote sensing, spectral imaging and linear un-mixing algorithms have been developed and are available as software plug-ins, or have been implemented on a number of commercially available instruments. This technology has been applied to commonly used APCs in live cell imaging (CFP, YFP, EGFP), which have strongly overlapping spectra requiring speciﬁc emission ﬁlters in order for them to be reliably detected in combination [100]. This approach allows fast, sensitive imaging and can be applied to the analysis of FRET [101]. Potentially, the full spectrum of APCs could be analyzed in the same cell. There are protocols available for the simultaneous analysis of ternary and quaternary combinations of ECFP/EGFP/EYFP/DsRed by means of ﬂow cytometry [102]. A combination of EGFP, citrine, mBanana, mOrange, mStrawberry and mCherry were analyzed by ﬂow cytometry using a single excitation at 514 nm and detection using three suitable emission channels [54].

Chromophore diversity in red-shifted APCs The chromophore in all naturally occurring APCs is formed from the concensus tri-peptide sequence, XYG. The ‘conjugated’ structure of the chromophore in engineered proteins has been altered by substitution of the tyrosine for other amino acids resulting in a relaxed concensus sequence XXG [3,13,14,54]. One of the key structural features that originally deﬁned APCs with red-shifted properties was the extension of the chromophore conjugation system by an acylimine bond (Fig. 1) [20–25]. However, structural studies have revealed the chromophore structures in APCs are considerably more diverse than we would have ﬁrst imagined. The range of different chromophore structures is a result of novel post-translational chemistry sometimes initiated by light, making them a source of interest other than for their biotechnological value. Novel features include alternative chromophore structures, such as the triple-ring structure in zFP538 [27], polypeptide backbone cleavage to form

56 Table 3. Novel post-translational features identiﬁed in all-protein chromophores. Protein

Novel feature(s)

Spectral feature

Reference

DsRed

Acylimine extending chromophore conjugation Polypeptide cleavage upon illumination.

Red emission

[20–22]

Comments

Common to EqFP611 and other chromoproteins Kaede, EosFP, Green to red emission [29,80,81,82,86,87] Leads to the formation of a novel red-emitting KikGR switch upon chromophore. photoconversion with UV light EqFP611 Trans-coplanar Far-red ﬂuorescence [47,53,113] Trans–coplanar chromophore conﬁguration is ﬂuorescent. Contains acylimine bond. Far-red ﬂuorescence [25,52,75] Cis-coplanar HcRed Cis- and trans-nonconﬁguration linked is coplanar ﬂuorescent, trans nonchromophore in coplanar is nonsame population ﬂuorescent Rtms5 Trans-non-coplanar Non-ﬂuorescent. [24] Single trans-non coplanar chromophore Intense blue colour conﬁguration [27,28] Backbone cleavage in Non-ﬂuorescent AsCP595/KFP Spontaneous native protein occurs chromophore polypeptide cleavage as part of normal undergoes trans-/ results in novel maturation leading to a cis-conformational chromophore. Transnovel chromophore change upon non-coplanar structure photoconversion chromophore. (Kindling) AceGFP-G222E Stalled maturation of Photoconersion [111] Generation may involve chromphore decarboxylation Chromophore also PA-mRFP1-1 Possible cis–trans Non-ﬂuorescent. Red [69] contains acylimine isomerisation of emission upon bond chromophore photoconversion with near UV zFP538 Triple-ring True yellow [26] chromophore, ﬂuorescence polypeptide cleavage emission GFP Y65L No chromophore [114,115] formed wtGFP, PADecarboxylation of Initiated by UV light [116–119] GFP Glu222

new chromophores and the trans-non-coplanar chromophore conﬁgurations of non-ﬂuorescent CPs [24] (Table 3). Alternative chromophore conﬁgurations The conformation of the chromophore inside the b-barrel of the protein is an important determinant of QY. The chromophore of all highly FPs for which the crystal structure is known, is coplanar (Fig. 6) while those for some nonFPs are non-coplanar. The ﬁrst suggestion that alternative conformations

57 were likely came from the observation that the highly ﬂuorescent DsRed (QY, 0.69) could be converted into a non-FP NF-DsRed (QY, o0.001) having a high-extinction coefﬁcient [103]. The position of the amino acid substitutions in NF-DsRed were modelled and suggested that the chromophore occupies a trans but coplanar conﬁguration. A trans- to cis-isomerization was predicted to be the basis of photoconversion in these proteins [84,85], the molecular basis for which has now been conﬁrmed [104]. This concept of the chromophore being able to adopt conformations other than cis-coplanar has been supported in a number of recent structural analyses and is discussed below. A detailed understanding of the factors controlling chromophore formation and alterations of chromophore conﬁguration will inform future engineering of APCs with novel properties. Cis-coplanar configurations. The tyrosine moiety of the chromophore in GFP is held in a coplanar conﬁguration and cis to the imidazolinone ring [105,106] (Fig. 1). The same is true of all the highly ﬂuorescent A. victoria variants for which the structure has been deposited in the Protein Data bank including BFP (Y66H substitution) [107], CFP (Y66W substitution) and GdFP (Y66 4-aminotryptophan substitution) [17]. The highly ﬂuorescent red FPs DsRed [21,22] and EqFP611 [23] also have coplanar chromophores. Trans-coplanar configuration. The importance of a coplanar chromophore is emphasized by the ﬁnding that EqFP611 (QY, 0.45; lmax em 611 nm) from the sea anemone E. quadricolor contains a coplanar conformation, but in the trans-conﬁguration. In comparison to DsRed, the 4-hydroxyphenyl group in EqFP611 was rotated 1801 about the Ca2–Cb2 bond (Fig. 6). Trans-non-coplanar configuration. Other chromophore conﬁgurations are possible. The crystal structure of the non-ﬂuorescent Rtms5, showed the tyrosine ring of the chromophore tri-peptide QYG is in a non-coplanar and trans-conformation despite having an identical chromophore sequence (Gln–Tyr–Gly) to DsRed. The tyrosyl moiety was rotated 431 with respect to the plane of the imidazolinone ring [24] (Fig. 6). A similar ﬁnding was true for non-FPs isolated from the sea anemone, Anemonia sulcata [27,28]. Cis-coplanar and trans-non-coplanar configuration in the same protein. The QY of certain CPs may be increased several hundred-fold or more by making key amino acid substitutions in regions surrounding the chromophore [24,34,75]. The moderately ﬂuorescent dimeric HcRed (QY, 0.05; lmax em 645 nm) derived by mutagenesis from the tetrameric hcriCP contains both trans-non-coplanar (non-ﬂuorescent) and a cis-coplanar (ﬂuorescent) chromophore conﬁgurations. Chromophore configuration and photoconversion It has been suggested that alterations in chromophore conﬁguration are the basis of the kindling phenomenon. Rtms5 when illuminated with light near its absorbance maximum ðlmax abs ¼ 592 nmÞ undergoes a remarkable transition,

58 termed ‘kindling’, to a long-lived ﬂuorescent state ðlmax em ¼ 632 nmÞ (Prescott, unpublished observations). A similar phenomenon has been observed for the CP asFP595 isolated from A. sulcata [84,85]. The kindling phenomenon was enhanced through the inclusion of the substitution A143G to produce the commercially available variant KFP. The crystal structure of KFP has been determined in different crystal forms to 2.1 and 1.38 A˚ resolution [27,28]. The structure revealed the chromophore, like Rtms5, to be in a trans and noncoplanar conﬁguration. Since the highly ﬂuorescent EqFP611 contained a trans-coplanar chromophore, and HcRed contains a proportion of the chromophore in the cis-coplanar conformation, it was concluded that the nonplanarity of the chromophore was responsible for the non-ﬂuorescent property of the non-kindled state. Conversely, it has been argued that the ‘kindled state’ can be attributed to a planar chromophore and that kindling involves a cis– trans-isomerization. Cavity volume calculations for KFP indicate sufﬁcient space for the chromophore to adopt a cis-conformation. The major barrier to such as transition is the presence of His197. Since this residue was observed to adopt two conformations in the 1.38 A˚ crystal structure it was proposed to act as a conformational switch. In this case, trans– cis-isomeriation would require rearrangement between the two conformations. The fragmentation of the polypeptide between Cys62 and Met63 gives the chromophore a degree of freedom that may facilitate the isomerization of the chromophore. This mode for the photoswitch is supported by X-ray crystal structures of an asCP variant in the non-ﬂuorescent form and photoconverted ﬂuorescent form that contain the chromophore in trans- and cis-conformation, respectively. Molecular dynamic calculations suggested that cis– trans-isomerization occurred via a ‘hula-twist’ of the bonds bridging the tyrosyl and imidazole moieties of the chromophore.

Alternative chromophore structures Triple ring chromophore The yellow FP, zFP538, isolated from a Zoanthus coral polyp [8] compared to the best greenish-yellow variant derived from A. victoria GFP ðlmax em ¼ ¼ 538 nmÞ. 529 nmÞ [66], displays ‘true’ yellow ﬂuorescence emission ðlmax em The protein contains a novel three-ring structure (Fig. 1) [26] compared to the double ring of GFP. Evidently, yellow emission is the direct result of an increase in the physical extent of the chromophore by an additional double bond (C ¼ N) compared to GFP. The formation of this chromophore is postulated to proceed through an acylimine intermediate followed by reaction with the lysine side chain of the chromophore tri-peptide (KYG), cleavage of the main backbone of the protein and formation of a sixmembered-ring.

59 It has been suggested the acylimine may represent an intermediate in the formation of chromphores in other proteins [26]. AsCP isolated from A. sulcata undergoes a polypeptide fragmentation during the process of maturation to yield a novel chromophore (Fig. 1). Fragmentation ocurrs between the Cys62C and Met63N1 of the MYG chromophore tripeptide. Light-induced alteration in chromophore structure Maturation of DsRed from the green-emitting form to the red-emitting form is driven by a process of chemical oxidation. There are a number of proteins in which the maturation can be considered to have been naturally suspended at the green-emitting form and, require the energy provided by UV light to continue to the red-emitting form. Examples of such proteins are Kaede [80], EosFP [82,86,87], DendFP, mcavRFP [108] and rﬂoFP [109]. A key feature of all these proteins is a histidine in the chromophore tripeptide, HYG. The best-studied FPs of this type are Kaede and EosFP from the stony coral Trachphyllia geoffroyi and Lobophyllia hemprichii, respectively. The HYG tripeptide initially forms the 4-(p-hydroxybenzylidene)-5 imidazalinone chromophore similar to that found in GFP and responsible for green ﬂuorescence (Fig. 1). The ‘green’ form shows two absorption peaks at 380 nm and 508 nm. Photoconversion with UV involves an unconventional cleavage within the native protein between the amide nitrogen and the a carbon at the histidine of the tripeptide via a formal b-elimination reaction. The double bond created in this process results in the formation of a novel chromophore, 2-[(1E)-2-(5-imidazolyl)ethenyl]-4-(p-hydroxybenzylidene)-5 imidazolinone responsible for the red emission in which the p conjugation is extended (Fig. 1). The X-ray structures of the green and red forms of EosFP have been determined and show in the intact protein, the UV-induced polypeptide cleavage and the resulting chromophore structures [87]. Concluding comments The ever-increasing expectations of the users of FP technology are the driving force for continuing improvements. Reduced toxicity and far-red shifted variants such as mPlum will prove popular for use in live animal studies. However, FPs with red shift that emit beyond the lmax of 649 nm for mPlum (648 nm) are desirable. It is not clear whether such proteins exist in nature [73], or indeed if they can be engineered using the existing FP framework as a starting point. Reports of organisms with novel bioluminescent properties are only now being investigated [110]. The approach of iterative somatic hypermutation would appear to be able to produce outcomes not yet identiﬁed in nature [76]. Certainly, structuredirected engineering of such proteins would appear to be a formidable task,

60 given the additional chemistry that may be needed to form such chromophores with far-red shifted spectra. Some applications of FPs are founded on exploiting, to best advantage, properties initially considered to be highly undesirable. Excellent examples of this are the kindling phenomenon of cpAs595, or the slow maturation of DsRed that inspired the ‘ﬂuorescent timer’. Undoubtedly, new applications will be found for the often intriguing properties of other APCs. For example, a novel range of FP sensors for use in monitoring mechanical force at the molecular level may be possible, based on observations that certain CPs can increase their ﬂuorescence emission 100-fold when exposed to relatively mild denaturing conditions (Prescott, unpublished results). References 1. 2. 3. 4. 5. 6. 7. 8.

9.

10. 11.

12. 13.

14. 15.

Davenport D and Nicol JAC. Luminescence in hydromedusae. Proc R Soc London Ser B 1955;144:399–411. Chalﬁe M, Tu Y, Euskirchen G, Ward WW and Prasher DC. Green ﬂuorescent protein as a marker for gene expression. Science 1994;263:802–805. Heim R, Prasher DC and Tsien RY. Wavelength mutations and posttranslational autoxidation of green ﬂuorescent protein. Proc Natl Acad Sci USA 1994;91:12501–12504. Tsien, RY Imagining imaging’s future. Nat Rev Mol Cell Biol 2003; Suppl SS16–SS21. Tsien RY. The green ﬂuorescent protein. Annu Rev Biochem 1998;67:509–544. Zimmer M. Green ﬂuorescent protein (GFP): applications, structure, and related photophysical behavior. Chem Rev 2002;102:759–781. Dove SG, Hoegh-Guldberg O and Ranganathan S. Major colour patterns of reefbuilding corals are due to a family of GFP-like proteins. Coral Reefs 2001;19:197–204. Matz MV, Fradkov AF, Labas YA, Savitsky AP, Zaraisky AG, Markelov ML and Lukyanov SA. Fluorescent proteins from nonbioluminescent Anthozoa species. Nat Biotechnol 1999;17:969–973. Labas YA, Gurskaya NG, Yanushevich YG, Fradkov AF, Lukyanov KA, Lukyanov SA and Matz MV. Diversity and evolution of the green ﬂuorescent protein family. Proc Natl Acad Sci USA 2002;99:4256–4261. Galperin E, Verkhusha VV and Sorkin A. Three-chromophore FRET microscopy to analyze multiprotein interactions in living cells. Nat Methods 2004;1:209–217. Heim R and Tsien RY. Engineering green ﬂuorescent protein for improved brightness, longer wavelengths and ﬂuorescence resonance energy transfer. Curr Biol 1996;6:178–182. Elowitz MB, Surette MG, Wolf PE, Stock J and Leibler S. Photoactivation turns green ﬂuorescent protein red. Curr Biol 1997;7:809–812. Patterson GH, Knobel SM, Sharif WD, Kain SR and Piston DW. Use of the green ﬂuorescent protein and its mutants in quantitative ﬂuorescence microscopy. Biophys J 1997;73:2782–2790. Cubitt AB, Heim R, Adams SR, Boyd AE, Gross LA and Tsien RY. Understanding, improving and using green ﬂuorescent proteins. Trends Biochem Sci 1995;20:448–455. Pal PP, Bae JH, Azim MK, Hess P, Friedrich R, Huber R, Moroder L and Budisa N. Structural and spectral response of Aequorea victoria green ﬂuorescent proteins to chromophore ﬂuorination. Biochemistry 2005;44:3663–3672.

61 16. Kajihara D, Hohsaka T and Sisido M. Synthesis and sequence optimization of GFP mutants containing aromatic non-natural amino acids at the Tyr66 position. Protein Eng Des Sel 2005;18:273–278. 17. Bae HJ, Rubini M, Jung G, Wiegand G, Seifert MH, Azim MK, Kim JS, Zumbusch A, Holak TA, Moroder L, Huber R and Budisa N. Expansion of the genetic code enables design of a novel ‘‘gold’’ class of green ﬂuorescent proteins. J Mol Biol 2003;328:1071–1081. 18. Budisa N, Pal PP, Alefelder S, Birle P, Krywcun T, Rubini M, Wenger W, Bae JH and Steiner T. Probing the role of tryptophans in Aequorea victoria green ﬂuorescent proteins with an expanded genetic code. Biol Chem 2004;385:191–202. 19. Niwa H, Inouye S, Hirano T, Matsuno T, Kojima S, Kubota M, Ohashi M and Tsuji FI. Chemical nature of the light emitter of the Aequorea green ﬂuorescent protein. Proc Natl Acad Sci USA 1996;93:13617–13622. 20. Gross LA, Baird GS, Hoffman RC, Baldridge KK and Tsien RY. The structure of the chromophore within DsRed, a red ﬂuorescent protein from coral. Proc Natl Acad Sci USA 2000;97:11990–11995. 21. Wall MA, Socolich M and Ranganathan R. The structural basis for red ﬂuorescence in the tetrameric GFP homolog DsRed. Nat Struct Biol 2000;7:1133–1138. 22. Yarbrough D, Wachter RM, Kallio K, Matz MV and Remington SJ. Reﬁned crystal structure of DsRed, a red ﬂuorescent protein from coral, at 2.0-A resolution. Proc Natl Acad Sci USA 2001;98:462–467. 23. Petersen J, Wilmann PG, Beddoe T, Oakley AJ, Devenish RJ, Prescott M and Rossjohn J. The 2.0-A crystal structure of eqFP611, a far-red ﬂuorescent protein from the sea anemone Entacmaea quadricolor. J Biol Chem 2003;278:44626–44631. 24. Prescott M, Ling M, Beddoe T, Oakley AJ, Dove S, Hoegh-Guldberg O, Devenish RJ and Rossjohn J. The 2.2 A crystal structure of a pocilloporin pigment reveals a nonplanar chromophore conformation. Structure 2003;11:275–284. 25. Wilmann PG, Petersen J, Pettikiriarachchi A, Buckle AM, Smith SC, Olsen S, Perugini MA, Devenish RJ, Prescott M and Rossjohn J. The 2.1 A crystal structure of the far-red ﬂuorescent protein HcRed: inherent conformational ﬂexibility of the chromophore. J Mol Biol 2005;349:223–237. 26. Remington SJ, Wachter RM, Yarbrough DK, Branchaud B, Anderson DC, Kallio K and Lukyanov KA. zFP538, a yellow-ﬂuorescent protein from Zoanthus, contains a novel three-ring chromophore. Biochemistry 2005;44:202–212. 27. Quillin ML, Anstrom DM, Shu X, O’Leary S, Kallio K, Chudakov DM and Remington SJ. Kindling ﬂuorescent protein from Anemonia sulcata: dark-state structure at 1.38 A resolution. Biochemistry 2005;44:5774–5787. 28. Wilmann PG, Petersen J, Devenish RJ, Prescott M and Rossjohn J. Variations on the GFP chromophore: a polypeptide fragmentation within the chromophore revealed in the 2.1-A crystal structure of a nonﬂuorescent chromoprotein from Anemonia sulcata. J Biol Chem 2005;280:2401–2404. 29. Mizuno H, Mal TK, Tong KI, Ando R, Furuta T, Ikura M and Miyawaki A. Photoinduced peptide cleavage in the green-to-red conversion of a ﬂuorescent protein. Mol Cell 2003;12:1051–1058. 30. Baird GS, Zacharias DA and Tsien RY. Biochemistry, mutagenesis, and oligomerization of DsRed, a red ﬂuorescent protein from coral. Proc Natl Acad Sci USA 2000;97:11984–11989. 31. Terskikh A, Fradkov A, Ermakova G, Zaraisky A, Tan P, Kajava AV, Zhao X, Lukyanov S, Matz M, Kim S, Weissman I and Siebert P. ‘‘Fluorescent timer’’: protein that changes color with time. Science 2000;290:1585–1588.

62 32. Verkhusha VV, Chudakov DM, Gurskaya NG, Lukyanov S and Lukyanov KA. Common pathway for the red chromophore formation in ﬂuorescent proteins and chromoproteins. Chem Biol 2004;11:845–854. 33. Marchant JS, Stutzmann GE, Leissring MA, LaFerla FM and Parker I. Multiphotonevoked color change of DsRed as an optical highlighter for cellular and subcellular labeling. Nat Biotechnol 2001;19:645–649. 34. Lukyanov KA, Fradkov AF, Gurskaya NG, Matz MV, Labas YA, Savitsky AP, Markelov ML, Zaraisky AG, Zhao X, Fang Y, Tan W and Lukyanov SA. Natural animal coloration can be determined by a nonﬂuorescent green ﬂuorescent protein homolog. J Biol Chem 2000;275:25879–25882. 35. Martynov VI, Maksimov BI, Martynova NY, Pakhomov AA, Gurskaya NG and Lukyanov SAA. Purple-blue chromoprotein from Goniopora tenuidens belongs to the DsRed subfamily of GFP-like proteins. J Biol Chem 2003;278:46288–46292. 36. Jakobs S, Subramaniam V, Schonle A, Jovin TM and Hell SW. EFGP and DsRed expressing cultures of Escherichia coli imaged by confocal, two-photon and ﬂuorescence lifetime microscopy. FEBS Lett 2000;479:131–135. 37. Yanushevich YG, Staroverov DB, Savitsky AP, Fradkov AF, Gurskaya NG, Bulina ME, Lukyanov KA and Lukyanov SA. A strategy for the generation of nonaggregating mutants of Anthozoa ﬂuorescent proteins. FEBS Lett 2002;511:11–14. 38. Bevis BJ and Glick BS. Rapidly maturing variants of the Discosoma red ﬂuorescent protein (DsRed). Nat Biotechnol 2002;20:83–87. 39. Wiehler J, von Hummel J and Steipe B. Mutants of Discosoma red ﬂuorescent protein with a GFP-like chromophore. FEBS Lett 2001;487:384–389. 40. Knop M, Barr F, Riedel CG, Heckel T, and Reichel C. Improved version of the red ﬂuorescent protein (drFP583/DsRed/RFP). Biotechniques 2002;33:592–594; 596–598 passim. 41. Mirabella R, Franken C, van der Krogt GN, Bisseling T and Geurts R. Use of the ﬂuorescent timer DsRed-E5 as reporter to monitor dynamics of gene activity in plants. Plant Physiol 2004;135:1879–1887. 42. Robinson LC and Marchant JS. Improved ‘‘optical highlighter’’ probes derived from Discosoma red ﬂuorescent protein. Biophys J 2005;88:1444–1457. 43. Ward WW. Biochemical and physical properties of green ﬂuorescent protein. In: Green Fluorescent Protein. Properties, Applications and Protocols, Chalﬁe M and Kain SR (eds), New York, Wiley-Liss, 1998, pp. 45–75. 44. Zacharias DA, Violin JD, Newton AC and Tsien RY. Partitioning of lipid-modiﬁed monomeric GFPs into membrane microdomains of live cells. Science 2002;296:913–916. 45. Vrzheshch PV, Akovbian NA, Varfolomeyev SD and Verkhusha VV. Denaturation and partial renaturation of a tightly tetramerized DsRed protein under mildly acidic conditions. FEBS Lett 2000;487:203–208. 46. Sacchetti A, Subramaniam V, Jovin TM and Alberti S. Oligomerization of DsRed is required for the generation of a functional red ﬂuorescent chromophore. FEBS Lett 2002;525:13–19. 47. Wiedenmann J, Schenk A, Rocker C, Girod A, Spindler KD and Nienhaus GU. A farred ﬂuorescent protein with fast maturation and reduced oligomerization tendency from Entacmaea quadricolor (Anthozoa, Actinaria). Proc Natl Acad Sci USA 2002;99:11646–11651. 48. Gavin P, Devenish RJ and Prescott M. An approach for reducing unwanted oligomerisation of DsRed fusion proteins. Biochem Biophys Res Commun 2002; 298:707–713.

63 49. Bulina ME, Verkhusha VV, Staroverov DB, Chudakov DM and Lukyanov KA. Hetero-oligomeric tagging diminishes non-speciﬁc aggregation of target proteins fused with Anthozoa ﬂuorescent proteins. Biochem J 2003;371:109–114. 50. Lauf U, Lopez P and Falk MM. Expression of ﬂuorescently tagged connexins: a novel approach to rescue function of oligomeric DsRed-tagged proteins. FEBS Lett 2001;498:11–15. 51. Campbell RE, Tour O, Palmer AE, Steinbach PA, Baird GS, Zacharias DA and Tsien RY. A monomeric red ﬂuorescent protein. Proc Natl Acad Sci USA 2002;99:7877–7882. 52. Fradkov AF, Verkhusha VV, Staroverov DB, Bulina ME, Yanushevich YG, Martynov VI, Lukyanov S and Lukyanov KA. Far-red ﬂuorescent tag for protein labelling. Biochem J 2002;368:17–21. 53. Wiedenmann J, Vallone B, Renzi F, Nienhaus K, Ivanchenko S, Rocker C and Nienhaus GU. Red ﬂuorescent protein eqFP611 and its genetically engineered dimeric variants. J Biomed Opt 2005;10:14003. 54. Shaner NC, Campbell RE, Steinbach PA, Giepmans BN, Palmer AE and Tsien RY. Improved monomeric red, orange and yellow ﬂuorescent proteins derived from Discosoma sp. red ﬂuorescent protein. Nat Biotechnol 2004;22:1567–1572. 55. Peyker A, Rocks O and Bastiaens PI. Imaging activation of two Ras isoforms simultaneously in a single cell. Chembiochem 2005;6:78–85. 56. Karasawa S, Araki T, Yamamoto-Hino M and Miyawaki A. A green-emitting ﬂuorescent protein from Galaxeidae coral and its monomeric version for use in ﬂuorescent labeling. J Biol Chem 2003;278:34167–34171. 57. Karasawa S, Araki T, Nagai T, Mizuno H and Miyawaki A. Cyan-emitting and orangeemitting ﬂuorescent proteins as a donor/acceptor pair for ﬂuorescence resonance energy transfer. Biochem J 2004;381:307–312. 58. Bokman SH and Ward WW. Renaturation of Aequorea green-ﬂuorescent protein. Biochem Biophys Res Commun 1981;101:1372–1380. 59. Ward WW and Bokman SH. Reversible denaturation of Aequorea green-ﬂuorescent protein: physical separation and characterization of the renatured protein. Biochemistry 1982;21:4535–4540. 60. Fukuda H, Arai M and Kuwajima K. Folding of green ﬂuorescent protein and the cycle3 mutant. Biochemistry 2000;39:12025–12032. 61. Verkhusha VV, Kuznetsova IM, Stepanenko OV, Zaraisky AG, Shavlovsky MM, Turoverov KK and Uversky VN. High stability of Discosoma DsRed as compared to Aequorea EGFP. Biochemistry 2003;42:7879–7884. 62. Li X, Zhao X, Fang Y, Jiang X, Duong T, Fan C, Huang CC and Kain SR. Generation of destabilized green ﬂuorescent protein as a transcription reporter. J Biol Chem 1998;273:34970–34975. 63. Destabilized DsRed-express and HcRed vectors. Clontechniques 2002;17(4):14–15. 64. Stepanenko OV, Verkhusha VV, Kazakov VI, Shavlovsky MM, Kuznetsova IM, Uversky VN and Turoverov KK. Comparative studies on the structure and stability of ﬂuorescent proteins EGFP, zFP506, mRFP1, ‘‘dimer2’’, and DsRed1. Biochemistry 2004;43:14913–14923. 65. Zeytun A, Jeromin A, Scalettar BA, Waldo GS and Bradbury AR. Fluorobodies combine GFP ﬂuorescence with the binding characteristics of antibodies. Nat Biotechnol 2003;21:1473–1479. 66. Griesbeck O, Baird GS, Campbell RE, Zacharias DA and Tsien RY. Reducing the environmental sensitivity of yellow ﬂuorescent protein. Mechanism and applications. J Biol Chem 2001;276:29188–29194.

64 67. Miesenbock G, De Angelis DA and Rothman JE. Visualizing secretion and synaptic transmission with pH-sensitive green ﬂuorescent proteins. Nature 1998;394:192–195. 68. Hanson GT, McAnaney TB, Park ES, Rendell ME, Yarbrough DK, Chu S, Xi L, Boxer SG, Montrose MH and Remingtone SJ. Green ﬂuorescent protein variants as ratiometric dual emission pH sensors. 1. Structural characterization and preliminary application. Biochemistry 2002;41:15477–15488. 69. Verkhusha VV and Sorkin A. Conversion of the monomeric red ﬂuorescent protein into a photoactivatable probe. Chem Biol 2005;12:279–285. 70. Ostergaard H, Henriksen A, Hansen FG and Winther JR. Shedding light on disulﬁde bond formation: engineering a redox switch in green ﬂuorescent protein. EMBO J 2001;20:5853–5862. 71. Jayaraman S, Haggie P, Wachter RM, Remington SJ and Verkman AS. Mechanism and cellular applications of a green ﬂuorescent protein-based halide sensor. J Biol Chem 2000;275:6047–6050. 72. Hu CD and Kerppola TK. Simultaneous visualization of multiple protein interactions in living cells using multicolor ﬂuorescence complementation analysis. Nat Biotechnol 2003;21:539–545. 73. Salih A, Larkum A, Cox G, Kuhl M and Hoegh-Guldberg O. Fluorescent pigments in corals are hotoprotective. Nature 2000;408:850–853. 74. Dove SG, Takabayashi M and Hoegh-Guldberg O. Isolation and partial characterization of the pink and blue pigments of Pocilloporid and Acroporid corals. Biol Bull 1995;189:288–297. 75. Gurskaya NG, Fradkov AF, Terskikh A, Matz MV, Labas YA, Martynov VI, Yanushevich YG, Lukyanov KA and Lukyanov SA. GFP-like chromoproteins as a source of far-red ﬂuorescent proteins. FEBS Lett 2001;507:16–20. 76. Wang L, Jackson WC, Steinbach PA and Tsien RY. Evolution of new nonantibody proteins via iterative somatic hypermutation. Proc Natl Acad Sci USA 2004;101:16745–16749. 77. Weissleder R. A clearer vision for in vivo imaging. Nat Biotechnol 2001;19:316–317. 78. Weissleder R and Ntziachristos V. Shedding light onto live molecular targets. Nat Med 2003;9:123–128. 79. Klonis N, Rug M, Harper I, Wickham M, Cowman A and Tilley L. Fluorescence photobleaching analysis for the study of cellular dynamics. Eur Biophys J 2002;31:36–51. 80. Ando R, Hama H, Yamamoto-Hino M, Mizuno H and Miyawaki A. An optical marker based on the UV-induced green-to-red photoconversion of a ﬂuorescent protein. Proc Natl Acad Sci USA 2002;99:12651–12656. 81. Tsutsui H, Karasawa S, Shimizu H, Nukina N and Miyawaki A. Semi-rational engineering of a coral ﬂuorescent protein into an efﬁcient highlighter. EMBO Rep 2005;6:233–238. 82. Wiedenmann J, Ivanchenko S, Oswald F, Schmitt F, Rocker C, Salih A, Spindler KD and Nienhaus GU. EosFP, a ﬂuorescent marker protein with UV-inducible green-tored ﬂuorescence conversion. Proc Natl Acad Sci USA 2004;101:15905–15910. 83. Chudakov DM, Verkhusha VV, Staroverov DB, Souslova EA, Lukyanov S and Lukyanov KA. Photoswitchable cyan ﬂuorescent protein for protein tracking. Nat Biotechnol 2004;22:1435–1439. 84. Chudakov DM, Feofanov AV, Mudrik NN, Lukyanov S and Lukyanov KA. Chromophore environment provides clue to ‘‘kindling ﬂuorescent protein’’ riddle. J Biol Chem 2003;278:7215–7219.

65 85. Chudakov DM, Belousov VV, Zaraisky AG, Novoselov VV, Staroverov DB, Zorov DB, Lukyanov S and Lukyanov KA. Kindling ﬂuorescent proteins for precise in vivo photolabeling. Nat Biotechnol 2003;21:191–194. 86. Nienhaus GU, Nienhaus K, Holzle A, Ivanchenko S, Renzi F, Oswald F, Wolff M, Schmitt F, Rocker C, Vallone B, Weidemann W, Heilker R, Nar H and Wiedenmann J. Photoconvertible ﬂuorescent protein EosFP-biophysical properties and cell biology applications. Photochem Photobiol 2005. (Epub. Ahead of print). 87. Nienhaus K, Nienhaus GU, Wiedenmann J and Nar H. Structural basis for photoinduced protein cleavage and green-to-red conversion of ﬂuorescent protein EosFP. Proc Natl Acad Sci USA 2005;102:9156–9159. 88. Verkhusha VV and Lukyanov KA. The molecular properties and applications of Anthozoa ﬂuorescent proteins and chromoproteins. Nat Biotechnol 2004;22:289–296. 89. Zhang J, Campbell RE, Ting AY and Tsien RY. Creating new ﬂuorescent probes for cell biology. Nat Rev Mol Cell Biol 2002;3:906–918. 90. Heim R. Green ﬂuorescent protein forms for energy transfer. In: Green Fluorescent Protein, Conn pm (ed), San Diego, Academic Press, 1999, pp. 408–423. 91. Pollok BA and Heim R. Using GFP in FRET-based applications. Trends Cell Biol 1999;9:57–60. 92. Mizuno H, Sawano A, Eli P, Hama H and Miyawaki A. Red ﬂuorescent protein from Discosoma as a fusion tag and a partner for ﬂuorescence resonance energy transfer. Biochemistry 2001;40:2502–2510. 93. Erickson MG, Moon DL and Yue DT. DsRed as a potential FRET partner with CFP and GFP. Biophys J 2003;85:599–611. 94. Yang X, Xu P and Xu T. A new pair for inter- and intra-molecular FRET measurement. Biochem Biophys Res Commun 2005;330:914–920. 95. Zapata-Hommer O and Griesbeck O. Efﬁciently folding and circularly permuted variants of the Sapphire mutant of GFP. BMC Biotechnol 2003;3:5. 96. Bastiaens PI and Squire A. Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell. Trends Cell Biol 1999;9:48–52. 97. Rocks O, Peyker A, Kahms M, Verveer PJ, Koerner C, Lumbierres M, Kuhlmann J, Waldmann H, Wittinghofer A and Bastiaens PI. An acylation cycle regulates localization and activity of palmitoylated Ras isoforms. Science 2005;307:1746–1752. 98. van Geest LK and Stoop K W J. FLIM on a wide ﬁeld ﬂuorescence microscope. Lett Pept Sci 2003;10:501–510. 99. Miyawaki A, Griesbeck O, Heim R and Tsien RY. Dynamic and quantitative Ca2+ measurements using improved cameleons. Proc Natl Acad Sci USA 1999;96: 2135–2140. 100. Zimmermann T, Rietdorf J and Pepperkok R. Spectral imaging and its applications in live cell microscopy. FEBS Lett 2003;546:87–92. 101. Zimmermann T, Rietdorf J, Girod A, Georget V and Pepperkok R. Spectral imaging and linear un-mixing enables improved FRET efﬁciency with a novel GFP2–YFP FRET pair. FEBS Lett 2002;531:245–249. 102. Hawley TS, Herbert DJ, Eaker SS and Hawley RG. Multiparameter ﬂow cytometry of ﬂuorescent protein reporters. Methods Mol Biol 2004;263:219–238. 103. Bulina ME, Chudakov DM, Mudrik NN and Lukyanov KA. Interconversion of Anthozoa GFP-like ﬂuorescent and non-ﬂuorescent proteins by mutagenesis. BMC Biochem 2002;3:7. 104. Andersen M, Wahl MC, Stiel AC, Grater F, Schafer LV, Trowitzsch S, Weber G, Eggeling C, Grubmuller H, Hell SW and Jakobs S. Structure and mechanism of the

66

105.

106. 107.

108.

109.

110. 111.

112. 113.

114.

115.

116.

117.

118. 119.

reversible photoswitch of a ﬂuorescent protein. Proc Natl Acad Sci USA 2005;102:13070–13074. Ormo M, Cubitt AB, Kallio K, Gross LA, Tsien RY and Remington SJ. Crystal structure of the Aequorea victoria green ﬂuorescent protein. Science 1996;273:1392–1395. Yang F, Moss LG and Phillips Jr GN. The molecular structure of green ﬂuorescent protein. Nat Biotechnol 1996;14:1246–1251. Wachter RM, King BA, Heim R, Kallio K, Tsien RY, Boxer SG and Remington SJ. Crystal structure and photodynamic behavior of the blue emission variant Y66H/ Y145F of green ﬂuorescent protein. Biochemistry 1997;36:9759–9765. Pakhomov AA, Martynova NY, Gurskaya NG, Balashova TA and Martynov VI. Photoconversion of the chromophore of a ﬂuorescent protein from Dendronephthya sp. Biochemistry (Mosc) 2004;69:901–908. Shagin DA, Barsova EV, Yanushevich YG, Fradkov AF, Lukyanov KA, Labas YA, Semenova TN, Ugalde JA, Meyers A, Nunez JM, Widder EA, Lukyanov SA and Matz M. GFP-like proteins as ubiquitous metazoan superfamily: evolution of functional features and structural complexity. Mol Biol Evol 2004;21:841–850. Haddock SH, Dunn CW, Pugh PR and Schnitzler CE. Bioluminescent and redﬂuorescent lures in a deep-sea siphonophore. Science 2005;309:263. Gurskaya NG, Fradkov AF, Pounkova NI, Staroverov DB, Bulina ME, Yanushevich YG, Labas YA, Lukyanov S and Lukyanov KA. A colourless green ﬂuorescent protein homologue from the non-ﬂuorescent hydromedusa Aequorea coerulescens and its ﬂuorescent mutants. Biochem J 2003;373:403–408. Ando R, Mizuno H and Miyawaki A. Regulated fast nucleocytoplasmic shuttling observed by reversible protein highlighting. Science 2004;306:1370–1373. Nienhaus K, Vallone B, Renzi F, Wiedenmann J and Nienhaus GU. Crystallization and preliminary X-ray diffraction analysis of the red ﬂuorescent protein eqFP611. Acta Crystallogr D Biol Crystallogr 2003;59:1253–1255. Rosenow MA, Huffman HA, Phail ME and Wachter RM. The crystal structure of the Y66L variant of green ﬂuorescent protein supports a cyclization–oxidation–dehydration mechanism for chromophore maturation. Biochemistry 2004;43:4464–4472. Rosenow MA, Patel HN and Wachter RM. Oxidative chemistry in the GFP active site leads to covalent cross-linking of a modiﬁed leucine side chain with a histidine imidazole: implications for the mechanism of chromophore formation. Biochemistry 2005;44:8303–8311. Brejc K, Sixma TK, Kitts PA, Kain SR, Tsien RY, Ormo M and Remington SJ. Structural basis for dual excitation and photoisomerization of the Aequorea victoria green ﬂuorescent protein. Proc Natl Acad Sci USA 1997;94:2306–2311. van Thor JJ, Gensch T, Hellingwerf KJ and Johnson LN. Phototransformation of green ﬂuorescent protein with UV and visible light leads to decarboxylation of glutamate 222. Nat Struct Biol 2002;9:37–41. Bell AF, Stoner-Ma D, Wachter RM and Tonge PJ. Light-driven decarboxylation of wild-type green ﬂuorescent protein. J Am Chem Soc 2003;125:6919–6926. van Thor JJ, Georgiev GY, Towrie M and Sage JT. Ultrafast and low barrier motions in the photoreactions of the green ﬂuorescent protein. J Biol Chem 2005; 280:33652–33659.

67

Putting the ‘Ome’ in lipid metabolism David M. Mutch1,2,y, Laetitia Fauconnot1, Martin Grigorov1 and Laurent B. Fay1, 1

Nestle´ Research Center, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland Center for Integrative Genomics, University of Lausanne, CH-1015 Lausanne, Switzerland

2

Abstract. The recognition that altered lipid metabolism underlies many metabolic disorders challenging Western society highlights the importance of this metabolomic subset, herein referred to as the lipidome. Although comprehensive lipid analyses are not a recent concept, the novelty of a lipidomic approach lies with the application of robust statistical algorithms to highlight subtle, yet significant, changes in a population of lipid molecules. First-generation lipidomic studies have demonstrated the sensitivity of interpreting quantitative datasets with computational software; however, the innate power of comprehensive lipid proﬁling is often not exploited, as robust statistical models are not routinely utilized. Therefore, the current review aims to brieﬂy describe the current technologies suitable for comprehensive lipid analysis, outline innovative mathematical models that have the ability to reveal subtle changes in metabolism, which will ameliorate our understanding of lipid biochemistry, and demonstrate the biological revelations found through lipidomic approaches and their potential implications for health management. Keywords: Bioinformatics; Biomathematics; Computational lipidomics; Drug discovery; Functional genomics; Gas chromatography; Health management; Lipidomics; Mass spectrometry; Nuclear magnetic resonance; Nutrition; Principal component analysis.

Introduction The comprehensive analysis of genes, proteins, and metabolites has led to fundamental changes in our approach in characterizing health and disease states. Using extensive molecular catalogues and powerful mathematical software to objectively probe massive biological datasets, various analytical platforms will provide requisite and complementary information implicit in unraveling the mechanisms underlying metabolic disorders; however, these platforms are advancing at distinctly different rates. Whereas transcriptomics has beneﬁted from the complete sequencing and annotation of several mammalian genomes [1–5], standardized methods for archiving data [6,7] and bioinformatic tools to ease the interpretation of these massive datasets, proteomics and metabolomics are not yet routine and standardized procedures, and continue to face challenges such as sample preparation, technological sensitivity, lack of standardized biomathematical analysis, and publicly interpretable databases, etc. [8–12]. Nevertheless, their potential beneﬁts for Corresponding author: Tel: +41-21-785-8609. Fax: +41-21-785-8544.

E-mail: [email protected] (L.B. Fay). Current address: INSERM, U755 Nutriomique, Paris, F-75004 France; Pierre and Marie Curie-Paris 6 University, Faculty of Medicine, Les Cordeliers, 75004 Paris, France.

y

BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12003-7

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

68 health management are recognized and fuel the current effort to utilize and integrate these global technologies with the goal to more accurately deﬁne a phenotype characterizing health status [13,14]. Proﬁling metabolites, which depict an organism’s acute cellular function, offers an attractive platform to identify disease biomarkers [15]. Several of the current epidemics challenging the Western world, such as atherosclerosis, hypertension, diabetes, and obesity are characterized by altered lipid metabolism [16–18]; prompting an emerging group of scientists to focus on the lipid fraction of the metabolome [19]. This subset of the metabolome is comprised of a great variety of different lipid species that have been experimentally demonstrated to have key roles in such diverse biological processes as signal transduction, membrane trafﬁcking and sorting, morphogenesis, and proliferation [20]. Indeed, lipids constitute a fundamentally important subset of the metabolome, as they have both a structural role in cell membranes and functional role as signaling molecules that are capable of responding to the cellular environment [21]. This subset of the metabolome has been deﬁned as the lipidome, and was ﬁrst used by Kishimoto and colleagues [22] to deﬁne the complete lipid composition of an organism; however, the classical definition of lipids as a diverse group of molecules soluble in organic solutions may underestimate the complexity of the lipidome, as many compounds considered as lipids are structurally unrelated and more soluble in water than organic solvents [23]. Furthermore, it is highly possible that a given lipid species will not have a single function within an organism, as these bioactive compounds are highly sensitive to their cellular environment and their molecular interactions within it. If the goal is to understand the contribution of the lipidome to a physiological state, then it becomes imperative to consider that the location of lipids and their associations within the cell (i.e., lipid–lipid and lipid–protein interactions) will have downstream effects on gene expression, protein abundance and location, and cell signaling events, which will ultimately modulate biological function. In other words, the comprehensive proﬁling and/or quantiﬁcation of lipids without knowledge of their cellular location will not conclusively deﬁne their role in a given biological function. Indeed, for this to be accomplished, lipids must be considered as integral players in a complex biological system, and therefore analyzed in a system-wide context (Fig. 1). While this remains ambitious, both from a technological and analytical perspective, targeted studies are emerging in which the tools required for such a task are being created and validated. The few studies that have proﬁled lipids in a biological sample have used an approach aimed at either quantifying and/or identifying a subset of the lipidome without examining such aspects as lipid–protein interactions and spatial distribution [24]; nevertheless their importance in mediating the lipidome’s inﬂuence on an organism’s health status is unquestionable. For example, lipoproteins, which are primary targets for combating cardiovascular disease, are macromolecular complexes composed of a lipid-rich core

69

Fig. 1. The spatial distribution of lipids in a biological system. To date, most targeted lipidomic studies concentrate on proﬁling or quantifying common lipid species in a biological system (circled) without examining their spatial distribution between particles (e.g., lipoproteins), cell organelles (e.g., mitochondrial, nuclear, or plasmatic membranes), or within an organelle following lipid–protein partitioning (e.g., lipid rafts). If the goal is to understand how lipids contribute to health and disease in a biological organism, then the aforementioned points will need to be incorporated in future lipidomic studies. (Images are from the following sources: http://www.netterimages.com/internalmed/image7, http://www.moreenergy.com/info, http:// www.astrosurf.com/lombry/bioastro-originevie3).

surrounded by a surface monolayer containing phospholipids (PLs), unesteriﬁed cholesterol, and speciﬁc proteins; however, the distribution of lipids across the various lipoproteins have distinct biological implications, as exempliﬁed with HDL- and LDL-cholesterol [25,26]. Furthermore, the discovery of lipid rafts, i.e., membrane microdomains enriched with sphingolipids and cholesterol, which function as platforms in which membrane-bound proteins associate to trigger signaling cascades proffers an important structural aspect of lipids capable of modulating cellular function [27–29] . Thus,

70 unraveling the inﬂuence of the thousands of species comprising the lipidome on biological function is a complex and ambitious goal, as such challenges as lipid–protein interactions and spatial distribution comprise clear hurdles that need to be overcome. Although studying the lipidome in its entirety is currently fraught with analytical obstacles, even assaying a subset of the lipidome will yield an overwhelmingly large quantity of data. Whether performing a targeted or comprehensive analysis of lipids, researchers are beginning to explore the applicability of robust statistical algorithms for the interpretation of these datasets (referred to as computational lipidomics [30]). In this regard, it is expected that both the analysis and interpretation of comprehensive lipid datasets will beneﬁt from the creative use of statistical algorithms in a manner similar to that seen with microarray experiments, i.e., rather than comparing individual lipid species between experimental conditions, the analyst will exploit the intrinsic properties of the entire dataset. The examples presented herein demonstrate that unraveling the effect of drugs, nutrients, toxins, etc. on lipid metabolism and, ultimately, on health can beneﬁt from the inherent power of a computational lipidomic approach, and thereby enhance our understanding of lipid biochemistry and metabolic disorders characterized by dyslipidemia. Exploring the lipidome Historically, much of our knowledge in lipid biochemistry has been obtained using standard gas chromatography (GC); however, technologies such as mass spectrometry (MS) and nuclear magnetic resonance (NMR) have proven proﬁcient analytical platforms to study lipids. Gas chromatography Although GC methods yield accurate results in lipid analysis, their use in high-throughput lipidome analyses is limited due to the time constraints for sample preparation [31]; however, GC technology is adapting and evolving. For example, a recent report using a ‘fast’ GC technique demonstrated its ability to identify lipids from complex mixtures to a comparable degree as standard GC methods; but the experimental time was reduced by more than 95% [31]! Thus, ‘fast’ GC and ‘fast’ GC–MS have the potential to become routine methods for lipidomic analysis by increasing sample throughput and improving laboratory efﬁciency [32,33]. Alternatively, a review by Dallu¨gue [34] outlined the enhanced resolving power of comprehensive two-dimensional GC (GC GC) vs. standard GC methods. Interestingly, coupling this technique with fatty acid methyl ester (FAME) analysis demonstrated its ability to predict compound identiﬁcation in the absence of commercially available standards by using linear retention

71 indices, MS detection, and information reported in the literature. Although this analytical approach has resolved previous technical complications, such as the co-elution of sterols and stanols following derivatization [35], this technique remains labor intensive due to the lack of appropriate and commercially available software suitable for quantitative analysis [36]. Indeed, as standards are not commercially available for the thousands of molecules present in the lipidome, the predictive power of this approach in the absence of standards is extremely attractive [37] and underlines the importance of developing analytical software to compliment this emerging technology. Mass spectrometry MS, using atmospheric pressure ionization systems (atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI)), generates the appropriate ions for the characterization of lipids, thereby providing information concerning the molecular weight, and fatty acyl composition and distribution of a given lipid. Multiple ESI–MS techniques have been developed and used for the analysis of various classes, subclasses, and individual molecular species of lipids from biological sources (such as plasma, blood, tissues, etc.), and its applicability for lipid proﬁling has been recently described [21,38]. By exploiting the advantages inherent with ESI–MS/MS techniques (such as sensitivity, high selectivity, and high efﬁciency), one can identify alterations in the lipidome directly from biological samples [38,39]; however, quantifying lipids with MS/MS must be made with caution, as the fragmentation kinetics are highly dependent on the collisional activation energy employed and the molecular structure, thus substantial differences between lipid species can be expected [40]. A two-dimensional ESI–MS approach has been recently utilized to ﬁngerprint the major classes, subclasses, and individual molecular species in the murine hepatic cellular lipidome directly from their chloroform extract [40]. Furthermore, ESI–MS/MSbased lipid proﬁling has been adapted to yield high-throughput analytical techniques; however, it still remains a targeted strategy as instrumental conditions have to be adapted to each class of analytes and requires pre-characterization of the response of the targeted lipids [21,39]. Nuclear magnetic resonance NMR represents a popular and alternate technique for proﬁling lipids in cells, tissues, and body ﬂuids [41–43]. As most lipids have a structure-speciﬁc resonance or set of resonances, NMR enables both their identiﬁcation and quantiﬁcation [42]. The relative amounts and partial assignment of the major lipid classes (diacylglycerols (DAG), TG, PL, ether lipids, sphingolipids, and steroids) can be directly determined from the 1H spectra of unfractionated lipid extracts [42]. Furthermore, NMR can be adapted to evaluate the degree

72

Fig. 2. Principal lipidomic platforms and their deliverables. As no single analytical

platform is capable of quantitatively assessing the lipidome because of such issues as sensitivity, speciﬁcity, and comprehensiveness, the lipidome is currently attainable only through the amalgamation of various analytical platforms. Through this amalgamation, one will be able to identify lipidomic ﬁngerprints and potential biomarkers.

of lipid saturation (13C NMR) and the fractionation of mixtures prior to NMR analysis is able to identify lipid components derived from complex solutions [43], as demonstrated in studies exploring the composition of lipoproteins [44]. To date, NMR is considered the most suitable platform for comprehensive metabolite ﬁngerprinting [45]; however, the similar magnetic resonance of lipids suggests that this technique is limited for identifying individual components of the lipidome [46,47]. In contrast, ESI–MS and ESI–MS/MS platforms, both targeted strategy-based approaches, have proven to be the most inclusive for lipid molecular species [21,30,38]. Nevertheless, it is recognized that issues regarding sensitivity, selectivity, and comprehensiveness currently render the lipidome assessable only through the amalgamation of various analytical platforms (Fig. 2). Biomathematics and the lipidome Although comprehensive lipid proﬁling is not a recent concept, it is the interpretation of these analyses with robust statistical algorithms that deﬁnes the essence and novelty of lipidomics. These algorithms couple important aspects of biology and technology with robust statistics, and have thus procured the label ‘biomathematical models’ [48]. In order to exploit and interpret the multivariate

73 datasets stemming from global studies, an array of data-mining techniques has been developed and can be generalized into two groups: supervised and unsupervised [9]. Unsupervised methods mine through data and extract relevant information without the presence of a teacher. Therefore, in situations lacking a phenotypic endpoint or other observable effect for a given perturbation on a biological system, one typically applies unsupervised methods. In contrast, supervised methods use a teacher to extract information concerning a perturbed system and relate it to an already characterized system. Thus, when the effect of the perturbation is either known or measurable, the preferred supervised methods enable the addition of external information during the data processing and learning phases. Although few comprehensive lipid studies to date have been biomathematically interpreted, we have identiﬁed several algorithms that have great potential for future lipidomic datasets. Unsupervised methods A popular method for analyzing multivariate datasets is principal component analysis (PCA) [49]. This procedure extracts the most relevant information in a biological study while simultaneously ﬁltering out as much noise as possible. Thus, the predominant relationships in complex datasets are extracted by transforming the number of possibly correlated variables into a smaller number of uncorrelated variables called principal components. In other words, the objective is to reduce the dimensionality of the dataset but to retain most of the information content. Therefore, the appropriateness of applying PCA in situations where the lipid proﬁles span widely different concentration ranges is questionable; however, as subsequently demonstrated, data mining with PCA appears suitable in situations where metabolite concentrations span a similar dynamic range [50,51]. An emerging, unsupervised learning algorithm based on Kohonen neural networks (a self-organizing map technique) offers an exciting new method to interpret lipidomic datasets. This method falls within the class of connectionist learning algorithms and was developed by simulating how the brain maps sensory input onto different regions of the cerebral cortex [52]. While its application to lipids is limited to date, Kohonen networks have been successfully used to distinguish the lipoprotein proﬁles from various disease groups, i.e., individuals suffering from cancer and coronary heart disease, which were heterogeneous for age, sex, and medication [53,54]. Despite lipoproteins not being routinely considered in lipidomic analyses, Kohonen neural networks are well positioned to become a useful biomathematical method in the future of the lipidome. Recently, Ivanova and colleagues [55] have described a new tool, termed lipid arrays, for the analysis of comprehensive lipid datasets. Not to be confused with the tangible ‘chips’ used for studying gene and protein expression, lipid arrays were used by the authors to describe a means to normalize,

74 tabulate, and computationally analyze large sets of MS data. The goal of this approach is to rapidly analyze large lipid datasets and identify statistically significant differences in mass spectra between two biological conditions. By ﬁrst normalizing the spectra to reduce the inﬂuence of different ionization efﬁciencies and the environment in which the compounds were analyzed, and then using Shewhart control charts to examine the variation in the system, a classical statistical test (e.g., a student’s test, ANOVA, etc.) can then be used to identify differences between two conditions. As biologically demonstrated, the authors were able to identify changes in the phospholipid cell membrane composition of mast cells following an allergic inﬂammatory response. This analytical approach seems well suited for lipids present in a biological sample at comparable concentrations; however, considering lipids spanning a wide range of concentrations will undoubtedly require that the ﬁnal statistical test in this model be adapted to address such points as analytical background noise and signal interference from abundant metabolites. Supervised methods Supervised machine learning algorithms (SMLA) are methods aimed at transforming multivariate data into a biological hypothesis under the guidance of a ‘teacher’. Indeed, the principle behind SMLAs are that, in cases where the underlying molecular mechanisms are sufﬁciently characterized, a ‘teacher’ is empowered to predict a response based on already characterized situations ranging from biological functions to diseased states [56]. Generally, two types of data are deﬁned for a given sample (e.g., a lipid proﬁle and the related biological response) and are referred to as inputs and the target, respectively. Therefore, the goal is to analyze a new dataset and ﬁnd a model mapping that will correctly associate an input with a target. A variety of different algorithms currently exist, such as artiﬁcial neural networks and decision tree learning; however, as few biological systems are sufﬁciently characterized, these algorithms are only now beginning to be applied to complex molecular biology datasets [57,58]. Nevertheless, enthusiasm to use these algorithms to analyze complex medical data to ultimately facilitate disease management suggests that individual’s with dyslipidemic diseases may beneﬁt from having a ‘teacher’ analyze their lipid proﬁles and predict disease outcomes [59,60]. In summary, such data-mining methods are available to assess the lipid proﬁles produced by current technologies, and their power to highlight biologically relevant changes can be further enhanced by coupling these methods. As we are currently lacking extensive databases of lipid proﬁles associated with precise physiological states, supervised data-mining techniques cannot be employed at their full potential. Hence, the majority of the analyses are possible through discovery-oriented, unsupervised methods aimed at identifying the molecular components leading to distinct clustering patterns.

75 Lipidomics in drug discovery, nutrition and metabolism, and functional genomics To date, the great majority of studies focused on lipid metabolism have used traditional experimental approaches and common biomarkers (such as plasma cholesterol or TG) to advance our understanding of various metabolic disorders. Inasmuch as these studies are constantly improving the validity and accuracy of our knowledge of lipid biochemistry, they all suffer from a similar inherent quandary; namely that accepted and characterized biomarkers are examined, thus their ability to identify novel mechanisms of action and/or biomarkers is greatly diminished. Clearly, lipidomic technologies provide a solution to resolve this tunnel vision [61]. Indeed, only a handful of studies have quantitatively measured the composition of lipid classes at a comprehensive scale; however, these studies, when coupled with biomathematical algorithms, reinforce the additional power inherent in these global technologies and demonstrate their effectiveness in further understanding a wide range of biological phenomena, such as toxicity [62], drug discovery [63,64], disease characterization [65], nutrient metabolism [51,66], and functional genomics [67,68]. In this regard, lipidomics offers a means to diagnose and predict disease, identify population subsets that are responsive (be it positively or negatively) to medicinal and/or dietary intervention, and determine the efﬁcacy of therapeutic intervention [69]. To our knowledge, no computational lipidomic experiments have been performed on lipid data stemming from NMR technology, thus the following examples describe significant advances made using GC/MS, LC/MS, or ESI/MS approaches; but extending such computational approaches to comprehensive lipid datasets derived from alternative analytical platforms has similar promise. Drug discovery Drug toxicity and adverse effects are of concern for the pharmaceutical industry; therefore, measuring metabolites offers a complete characterization of the biochemical outcomes following therapeutic intervention. Work involving the type-2 diabetes drug rosiglitazone demonstrated the appropriateness of lipid proﬁling in drug discovery [63]. As diabetes is characterized by metabolic changes, the goal is to address these changes while minimizing undesirable side effects. With an appropriate mouse model, quantitative lipid data for the liver, adipose, and heart tissues were obtained by FAME–capillary GC. Using proprietary data-mining software (Lipomics SurveyorTM), the authors revealed that fatty acid synthesis was markedly increased in the aforementioned tissues following treatment with rosiglitazone, leading to tissue toxicity; however, this increased synthesis was not reﬂected in the plasma lipid proﬁle, where cholesterol esters and TGs were decreased. Therefore, examining standard plasma biomarkers would have falsely indicated that rosiglitazone decreases de novo

76 lipid synthesis. In reality, lipid synthesis was increased and the processes regulating lipid import/export into tissues were apparently modiﬁed; concealing the side effects stemming from chronic rosiglitazone therapy. Interestingly, the acyl composition (namely 16:1n7 and 18:1n7) of plasma lipids reﬂected this increased lipid synthesis, suggesting that their use as potential biomarkers should be further explored. This highlights the ability of coupling quantitative lipid analysis with analytical software for the identiﬁcation of possible biomarkers and their potential use in clinical diagnostics. Nutrition and metabolism Nutrition is rapidly being deﬁned as a science of integrative metabolism [70]. Indeed, unraveling the metabolic events following the consumption of nutritional compounds via an integrated approach has the potential to characterize molecular targets that will lead to improved personalized dietary recommendations [12,71]. One of the most widely debated issues in nutrition regards the addition of arachidonic acid (AA; 20:4n-6) and docosahexaenoic acid (DHA; 22:6n-3) into infant formula; despite the beneﬁcial roles of longchain polyunsaturated fatty acids (LC-PUFA) on health status having been repeatedly demonstrated [72]. Based on previous ﬁndings, it could be assumed that a predominant role of DHA is to simply antagonize AA-mediated events [73,74]; however, a recent study demonstrated that hepatic stearoylCoA desaturase (SCD) is a speciﬁc target for AA, and that this novel mechanism of action is neither synergized nor attenuated by the addition of a DHA-enriched ﬁsh oil, indicating unique actions for each LC-PUFA [66]. Comprehensive PL proﬁles, produced by FAME–GC analysis, of four dietary groups were ﬁrst analyzed by singular value decomposition (SVD), an unsupervised method similar to PCA, and then subjected to single linkage hierarchical clustering analysis (a supervised method), leading to the identiﬁcation of four distinct clusters within the lipid classes analyzed (Fig. 3). Reinforced with transcriptomic data, the authors concluded that the AAmediated inhibition of SCD activity might have important health implications, as high SCD activity has been correlated with increased TGs and an increased risk of cardiovascular disease [75]. Furthermore, because both LCPUFA have unique actions, the range of beneﬁts attributed with LC-PUFA consumption can be achieved if both AA and DHA are present in the diet. Hence, integrating lipidomics, transcriptomics, and biomathematics offers an attractive means to unravel nutrient-speciﬁc contributions to lipid metabolism and, ultimately, to health status. In a second example, insight into orotic acid-induced fatty liver disease was investigated through a combined metabolomic and transcriptomic approach [51]. The authors ﬁrst proﬁled a wide range of metabolites by 1HNMR spectroscopy and analyzed the spectra using commercially available software (XWINNMR software). Interestingly, they revealed using PCA

77

Fig. 3. SVD analysis of hepatic phospholipid classes following dietary intervention.

The three-dimensional graph is plotted in a space deﬁned by the three axis along which appears the most important fractions of variability in the dataset. On the right-hand side of the ﬁgure, peak diagrams account for the weight of the different fatty acid species in separating the biological samples. The clusters corresponding to cardiolipin (CL), phosphatidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylserine/phosphatidyl-inositol (PS/PI) are delineated. The representative points correspond to the replicates (n ¼ 5) fed with various diets such as control diet (very light grey), ﬁsh oil diet (light grey), fungal oil diet (darkest grey), and fungal ﬁsh mixed diet (dark grey). A large shift between the control and the other groups is seen in PC species, and to a lesser extent in other PLs. No differences in diet were observed in CL species.

analysis that the consumption of 1% orotic acid resulted in an accumulation of mobile lipids in the liver. More specifically, increased resonances were observed for several lipid categories, such as cholesterol esters and PC. The quantiﬁcation of these ﬁndings was determined by GC analysis and revealed significant increases in several fatty acid species (16:0, 16:1, 18:0, and 18:1). Coupled with transcriptional data, the authors concluded that SCD expression and activity are modulated following the consumption of orotic acid. Although a comprehensive and quantitative lipid analysis was not performed, this study eloquently demonstrates the complementary nature of

78 these analytical platforms and the ability of one technology to produce a hypothesis that can be specifically examined by another. Functional genomics Several technologies are available for the functional characterization of genes, such as knock-out animal models, transcriptomics, small inhibitory RNAs, etc.; however, metabolite proﬁling is an attractive and complementary technique for understanding the mechanism by which a dysfunctional gene contributes to phenotype. The integrative biological analysis of the APOE*3Leiden transgenic mouse model revealed the power of a systems biology approach and the inherent need to implement and analyze the vast quantity of information with biomathematical models [68]. Using cDNA microarray technology to quantify gene expression, and LC/MS to qualitatively proﬁle proteins and lipids, the authors examined the hepatic lipid metabolism of wildtype mice and mice fed with a standard chow diet, which have been characterized with mild type I and type II lesions and an increased predisposition to developing atherosclerosis. An analysis of the lipid metabolites using PCA clearly separated the two pools of mice into two distinct clusters. The clustering pattern stemmed from the increased hepatic TG levels and speciﬁc increases in lysoPC C16:0 and lysoPC C18:0 in mice fed with the standard chow diet. Furthermore, changes seen in the lipidome agreed with those observed in the underlying regulatory genes and proteins. Thus, lipidomic proﬁling technologies are powerful new instruments within the functional genomics toolbox, and offers an innovative method for identifying potential lipid biomarkers characteristic of a speciﬁc dysfunctional molecular candidate. The lipid metabolome: The next frontier The future of comprehensive lipid proﬁling will further beneﬁt from a chemogenomics approach, which is an emerging paradigm in pharmaceutical chemistry [76]. This new technology can be simply deﬁned as the mapping of all small bioactive molecules onto all macromolecular targets, thereby permitting a system-wide analysis of the effects of disparate molecules via receptor binding (in vitro) or phenotypic endpoint (in vivo) proﬁles that characterize precise physiological states [77]. The chemogenomics strategy is generally two fold. The generated small molecular compounds can be used as both indicators for drug and/or nutrient optimization and, more interestingly, to probe molecular mechanisms. Moreover, the current challenge is to no longer apply this approach to single receptor molecules, but to biological networks on the whole. Indeed, this goal stems from the demonstrated concept of integrated metabolism, i.e., molecular mechanisms are not isolated objects, but irregularly interconnected networks with universal properties. Illustrating fatty acid molecules as a highly ordered structure on a

79 three-dimensional lattice indicating chain length, the number of double bonds, and the position of the ﬁrst double bond demonstrates that these molecules are interconnected through a small number of enzymes regulating desaturation and elongation (Fig. 4). Lipids are not autonomously acting molecules in the biological system; rather, they contribute to deﬁning health status via their interactions with genes and proteins. Therefore, inasmuch as the targeted analysis of lipids can further our understanding of lipid biochemistry, unraveling the molecular mechanisms of those diseases currently challenging Western society requires that this subset of the metabolome be integrated into the biological network. Indeed, a dysfunctional network underlying disease may entail changes in gene expression, protein abundance and location, lipid–protein interactions, and metabolite production and abundance; thus, it is critical to consider the functional significance of such lipid-related structures as lipoprotein particles and membrane microdomains. Current technologies are able to assess these structures [78]; however, they have not been routinely examined in past studies and comprise a clear challenge for future lipidomic studies [25]. Although lipoproteins are chain length position of first double bond

12

10

4

6

8

2

2 4 14:0

14:1 ω5 16:0 16:1 ω7

16:2 ω7

6 number of double bonds

18:0

18:1 ω7 18:1 ω9

20:0 22:0 24:0

24:1 ω9

Fig. 4. Lipid network. Visualizing a metabolic network of known fatty acid species

and the their relationship via enzymatic reactions. Lipid transformations were retrieved from different metabolic databases (e.g., http://www.lipidmaps.org/, http:// lipidbank.jp/, y).

80 routinely analyzed, they still require a method to separate/isolate these particles prior to determining their lipid content (e.g., NMR coupled with ESI–MS). The same holds true for lipid rafts, in that these membrane microdomains must be isolated prior to examining their lipid composition; however, it is currently unclear whether the protein and lipid composition of lipid rafts are biologically faithful to the in vivo situation or simply an artifact of the isolation processes [79,80]. A recent report by Pike and colleagues [81] has demonstrated that differences in lipid rafts can be identiﬁed with an ESI/MS approach and, although the authors did not use biomathematics to analyze their results, demonstrates that that once the issue of isolation procedures has been definitely resolved, lipidomic datasets derived from lipid rafts are ready to be analyzed by computational lipidomics. Indeed, the complexity of lipid metabolism, and its regulation by exogenous compounds, suggests that a chemogenomic approach will prove of particular interest for the interpretation of lipidomic proﬁling experiments, as this will consider lipids not as independent molecules, but rather as part of the underlying biological system deﬁning health. With the advent of comprehensive technologies and powerful biomathematical algorithms to tease out relevant information from perturbed biological systems, the future promises to be exciting and rich for lipid biochemists. The use of biomathematics proffers a powerful means to analyze lipid datasets and extract highly significant results. In this regard, computational models will become a fundamentally important aspect of lipidomics just as they did with transcriptomics. Only then will we be able to unravel the complex and interacting roles of lipids with genetic polymorphisms, nutritional molecules, therapeutic intervention and lifestyle, and their corresponding inﬂuence on the predisposition to dyslipidemic diseases.

Acknowledgments The authors thank Dr Katherine Mace´, Dr Marco Turini, and Dr J. Bruce German for their critical review and discussion of this manuscript.

References 1. 2. 3. 4. 5.

McPherson JD, et al. A physical map of the human genome. Nature 2001;409:934–941. Venter JC, et al. The sequence of the human genome. Science 2001;291:1304–1351. Okazaki Y, et al. Analysis of the mouse transcriptome based on functional annotation of 60,770 full-length cDNAs. Nature 2002;420:563–573. Waterston RH, et al. Initial sequencing and comparative analysis of the mouse genome. Nature 2002;420:520–562. Gibbs RA, et al. Genome sequence of the Brown Norway rat yields insights into mammalian evolution. Nature 2004;428:493–521.

81 6.

7. 8. 9. 10. 11. 12.

13. 14. 15.

16. 17. 18. 19.

20.

21.

22.

23. 24.

Brazma A, Hingamp P, Quackenbush J, Sherlock G, Spellman P, Stoeckert C, Aach J, Ansorge W, Ball CA, Causton HC, Gaasterland T, Glenisson P, Holstege FC, Kim IF, Markowitz V, Matese JC, Parkinson H, Robinson A, Sarkans U, Schulze-Kremer S, Stewart J, Taylor R, Vilo J and Vingron M. Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet 2001;29:365–371. Edgar R, Domrachev M and Lash AE. Gene expression omnibus: NCBI gene expression and hybridization array data repository. Nucl. Acids Res 2002;30:207–210. Corthals GL, Wasinger VC, Hochstrasser DF and Sanchez JC. The dynamic range of protein expression: a challenge for proteomic research. Electrophoresis 2000;21:1104–1115. Mendes P. Emerging bioinformatics for the metabolome. Brief Bioinform 2002;3:134–145. Rabilloud T. Two-dimensional gel electrophoresis in proteomics: old, old fashioned, but it still climbs up the mountains. Proteomics 2002;2:3–10. Cahill DJ and Nordhoff E. Protein arrays and their role in proteomics. Adv Biochem Eng Biotechnol 2003;83:177–187. Arab L. Individualized nutritional recommendations: do we have the measurements needed to assess risk and make dietary recommendations? Proc Nutr Soc 2004;63:167–172. Watkins SM. Comprehensive lipid analysis: a powerful metanomic tool for predictive and diagnostic medicine. Isr Med Assoc J 2000;2:722–724. Sumner LW, Mendes P and Dixon RA. Plant metabolomics: large-scale phytochemistry in the functional genomics era. Phytochemistry 2003;62:817–836. Goodacre R, Vaidyanathan S, Dunn WB, Harrigan GG and Kell DB. Metabolomics by numbers: acquiring and understanding global metabolite data. Trends Biotechnol 2004;22:245–252. Watson KE, Horowitz BN and Matson G. Lipid abnormalities in insulin resistant states. Rev Cardiovasc Med 2003;4:228–236. Faraj M, Lu HL and Cianflone K. Diabetes, lipids, and adipocyte secretagogues. Biochem Cell Biol 2004;82:170–190. Ordovas JM. The quest for cardiovascular health in the genomic era: nutrigenetics and plasma lipoproteins. Proc Nutr Soc 2004;63:145–152. Fahy E, Subramaniam S, Brown HA, Glass CK, Merrill AHJr, Murphy RC, Raetz CR, Russell DW, Seyama Y, Shaw W, Shimizu T, Spener F, van MG, VanNieuwenhze MS, White SH, Witztum JL and Dennis EA. A comprehensive classiﬁcation system for lipids. J Lipid Res 2005;46:839–862. Hermansson M, Uphoff A, Kakela R and Somerharju P. Automated quantitative analysis of complex lipidomes by liquid chromatography/mass spectrometry. Anal Chem 2005;77:2166–2175. Welti R and Wang X. Lipid species proﬁling: a high-throughput approach to identify lipid compositional changes and determine the function of genes involved in lipid metabolism and signaling. Curr Opin Plant Biol 2004;7:337–344. Kishimoto K, Urade R, Ogawa T and Moriyama T. Nondestructive quantiﬁcation of neutral lipids by thin-layer chromatography and laser-ﬂuorescent scanning: suitable methods for ‘‘lipidome’’ analysis. Biochem Biophys Res Commun 2001;281:657–662. Christie WW. Lipid Analysis: Isolation, Separation, Identiﬁcation and Structural Analysis of Lipids, 3rd edn, Bridgwater, England, The Oily Press, 2003. Balazy M. Eicosanomics: targeted lipidomics of eicosanoids in biological systems. Prostaglandins Other Lipid Mediat 2004;73:173–180.

82 25. Asztalos BF. High-density lipoprotein metabolism and progression of atherosclerosis: new insights from the HDL atherosclerosis treatment study. Curr Opin Cardiol 2004;19:385–391. 26. Lada AT and Rudel LL. Associations of low density lipoprotein particle composition with atherogenicity. Curr Opin Lipidol 2004;15:19–24. 27. Simons K and Ikonen E. Functional rafts in cell membranes. Nature 1997;387:569–572. 28. Simons K and Ehehalt R. Cholesterol, lipid rafts, and disease. J Clin Invest 2002;110:597–603. 29. Cohen AW, Hnasko R, Schubert W and Lisanti MP. Role of caveolae and caveolins in health and disease. Physiol Rev 2004;84:1341–1379. 30. Forrester JS, Milne SB, Ivanova PT and Brown HA. Computational lipidomics: a multiplexed analysis of dynamic changes in membrane lipid composition during signal transduction. Mol Pharmacol 2004;65:813–821. 31. Mondello L, Casilli A, Tranchida PQ, Costa R, Chiofalo B, Dugo P and Dugo G. Evaluation of fast gas chromatography and gas chromatography–mass spectrometry in the analysis of lipids. J Chromatogr A 2004;1035:237–247. 32. Mastovska K and Lehotay SJ. Practical approaches to fast gas chromatography–mass spectrometry. J Chromatogr A 2003;1000:153–180. 33. Matisova E and Domotorova M. Fast gas chromatography and its use in trace analysis. J Chromatogr A 2003;1000:199–221. 34. Dalluge J, Beens J and Brinkman UA. Comprehensive two-dimensional gas chromatography: a powerful and versatile analytical tool. J Chromatogr A 2003;1000:69–108. 35. Truong TT, Marriott PJ, Porter NA and Leeming R. Application of comprehensive twodimensional gas chromatography to the quantiﬁcation of overlapping faecal sterols. J Chromatogr A 2003;1019:197–210. 36. van Mispelaar VG, Tas AC, Smilde AK, Schoenmakers PJ and van Asten AC. Quantitative analysis of target components by comprehensive two-dimensional gas chromatography. J Chromatogr A 2003;1019:15–29. 37. Mondello L, Casilli A, Tranchida PQ, Dugo P and Dugo G. Detailed analysis and group-type separation of natural fats and oils using comprehensive two-dimensional gas chromatography. J Chromatogr A 2003;1019:187–196. 38. Han X and Gross RW. Global analyses of cellular lipidomes directly from crude extracts of biological samples by ESI mass spectrometry: a bridge to lipidomics. J Lipid Res 2003;44:1071–1079. 39. Taguchi R, Houjou T, Nakanishi H, Yamazaki T, Ishida M, Imagawa M and Shimizu T. Focused lipidomics by tandem mass spectrometry. J Chromatogr B Anal Technol Biomed Life Sci 2005;823(1):26–36. 40. Han X, Yang J, Cheng H, Ye H and Gross RW. Toward ﬁngerprinting cellular lipidomes directly from biological samples by two-dimensional electrospray ionization mass spectrometry. Anal Biochem 2004;330:317–331. 41. Adosraku RK, Choi GT, Constantinou-Kokotos V, Anderson MM and Gibbons WA. NMR lipid proﬁles of cells, tissues, and body ﬂuids: proton NMR analysis of human erythrocyte lipids. J Lipid Res 1994;35:1925–1931. 42. Bonzom PMA, Nicolaou A, Zloh M, Baldeo W and Gibbons WA. NMR lipid proﬁle of Agaricus bisporus. Phytochemistry 1999;50:1311–1321. 43. Tugnoli V, Bottura G, Fini G, Reggiani A, Tinti A, Trinchero A and Tosi MR. H-NMR and 13C-NMR lipid proﬁles of human renal tissues. Biopolymers 2003;72:86–95. 44. Cushley RJ and Okon M. NMR studies of lipoprotein structure. Annu Rev Biophys Biomol Struct 2002;31:177–206.

83 45. Nicholson JK, Connelly J, Lindon JC and Holmes E. Metabonomics: a platform for studying drug toxicity and gene function. Nat Rev Drug Discov 2002;1:153–161. 46. Weckwerth W. Metabolomics in systems biology. Annu Rev Plant Biol 2003;54: 669–689. 47. Watkins SM. Lipomic proﬁling in drug discovery, development and clinical trial evaluation. Curr Opin Drug Discov Dev 2004;7:112–117. 48. Chytil MK. On the concept of biomathematics. Acta Biotheor 1977;26:137–150. 49. Eriksson L, Antti H, Gottfries J, Holmes E, Johansson E, Lindgren F, Long I, Lundstedt T, Trygg J and Wold S. Using chemometrics for navigating in the large data sets of genomics, proteomics, and metabonomics (gpm). Anal Bioanal Chem 2004; 380(3):419–429. 50. Fievez V, Vlaeminck B, Dhanoa MS and Dewhurst RJ. Use of principal component analysis to investigate the origin of heptadecenoic and conjugated linoleic acids in milk. J Dairy Sci 2003;86:4047–4053. 51. Grifﬁn JL, Bonney SA, Mann C, Hebbachi AM, Gibbons GF, Nicholson JK, Shoulders CC and Scott J. An integrated reverse functional genomic and metabolic approach to understanding orotic acid-induced fatty liver. Physiol Genom 2004;17:140–149. 52. Kohonen T. Analysis of a simple self-organizing process. Biol Cybern 1982;44:135–140. 53. Ala-Korpela M. HiltunenY Bell JD. Artiﬁcial neural network analysis of 1H nuclear magnetic resonance spectroscopic data from human plasma. Anticancer Res 1996;16:1473–1477. 54. Bathen TF, Engan T, Krane J and Axelson D. Analysis and classiﬁcation of proton NMR spectra of lipoprotein fractions from healthy volunteers and patients with cancer or CHD. Anticancer Res 2000;20:2393–2408. 55. Ivanova PT, Milne SB, Forrester JS and Brown HA. LIPID arrays: new tools in the understanding of membrane dynamics and lipid signaling. Mol Interv 2004;4:86–96. 56. Azuaje F. Genomic data sampling and its effect on classiﬁcation performance assessment. BMC Bioinform 2003;4:5. 57. Narayanan A, Keedwell EC and Olsson B. Artiﬁcial intelligence techniques for bioinformatics. Appl Bioinform 2002;1:191–222. 58. Kuo WP, Kim EY, Trimarchi J, Jenssen TK, Vinterbo SA and Ohno-Machado L. A primer on gene expression and microarrays for machine learning researchers. J Biomed Inform 2004;37:293–303. 59. Ramesh AN, Kambhampati C, Monson JR and Drew PJ. Artiﬁcial intelligence in medicine. Ann R Coll Surg Engl 2004;86:334–338. 60. McCarthy JF, Marx KA, Hoffman PE, Gee AG, O’Neil P, Ujwal ML and Hotchkiss J. Applications of machine learning and high-dimensional visualization in cancer detection, diagnosis, and management. Ann N Y Acad Sci 2004;1020:239–262. 61. Mooser V and Ordovas JM. ‘Omic’ approaches and lipid metabolism: are these new technologies holding their promises? Curr Opin Lipidol 2003;14:115–119. 62. Coen M, Ruepp SU, Lindon JC, Nicholson JK, Pognan F, Lenz EM and Wilson ID. Integrated application of transcriptomics and metabonomics yields new insight into the toxicity due to paracetamol in the mouse. J Pharm Biomed Anal 2004;35:93–105. 63. Watkins SM, Reifsnyder PR, Pan HJ, German JB and Leiter EH. Lipid metabolomewide effects of the PPARgamma agonist rosiglitazone. J Lipid Res 2002;43:1809–1817. 64. Verhoeckx KC, Bijlsma S, Jespersen S, Ramaker R, Verheij ER, Witkamp RF, van der Greef J and Rodenburg RJ. Characterization of anti-inﬂammatory compounds using transcriptomics, proteomics, and metabolomics in combination with multivariate data analysis. Int Immunopharmacol 2004;4:1499–1514.

84 65. Esch SW, Williams TD, Biswas S, Chakrabarty A and Levine SM. Sphingolipid proﬁle in the CNS of the twitcher (globoid cell leukodystrophy) mouse: a lipidomics approach. Cell Mol Biol (Noisy-le-grand) 2003;49:779–787. 66. Mutch DM, Grigorov M, Berger A, Fay LB, Roberts MA, Watkins SM, Williamson G and German JB. An integrative metabolism approach identiﬁes stearoyl-CoA desaturase as a target for an arachidonate-enriched diet. FASEB J 2005;19:599–601. 67. Watkins SM, Zhu X and Zeisel SH. Phosphatidylethanolamine-N-methyltransferase activity and dietary choline regulate liver-plasma lipid ﬂux and essential fatty acid metabolism in mice. J Nutr 2003;133:3386–3391. 68. Clish CB, Davidov E, Oresic M, Plasterer TN, Lavine G, Londo T, Meys M, Snell P, Stochaj W, Adourian A, Zhang X, Morel N, Neumann E, Verheij E, Vogels JT, Havekes LM, Afeyan N, Regnier F, van der Greef J and Naylor S. Integrative biological analysis of the APOE*3-leiden transgenic mouse. Omics 2004;8:3–13. 69. Lindon JC, Holmes E and Nicholson JK. Metabonomics: systems biology in pharmaceutical research and development. Curr Opin Mol Ther 2004;6:265–272. 70. Zeisel SH, Allen LH, Coburn SP, Erdman JW, Failla ML, Freake HC, King JC and Storch J. Nutrition: a reservoir for integrative science. J Nutr 2001;131:1319–1321. 71. Watkins SM, Hammock BD, Newman JW and German JB. Individual metabolism should guide agriculture toward foods for improved health and nutrition. Am J Clin Nutr 2001;74:283–286. 72. Anderle P, Farmer P, Berger A and Roberts MA. Nutrigenomic approach to understanding the mechanisms by which dietary long-chain fatty acids induce gene signals and control mechanisms involved in carcinogenesis. Nutrition 2004;20:103–108. 73. Watanabe S, Doshi M, Akimoto K, Kiso Y and Hamazaki T. Suppression of plateletactivating factor generation and modulation of arachidonate metabolism by dietary enrichment with (n-9) eicosatrienoic acid or docosahexaenoic acid in mouse peritoneal cells. Prostaglandins Other Lipid Mediat 2001;66:109–120. 74. Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Willett WC and Rimm EB. Habitual dietary intake of n-3 and n-6 fatty acids in relation to inﬂammatory markers among US men and women. Circulation 2003;108:155–160. 75. Attie AD, Krauss RM, Gray-Keller MP, Brownlie A, Miyazaki M, Kastelein JJ, Lusis AJ, Stalenhoef AF, Stoehr JP, Hayden MR and Ntambi JM. Relationship between stearoyl-CoA desaturase activity and plasma triglycerides in human and mouse hypertriglyceridemia. J Lipid Res 2002;43:1899–1907. 76. Caron PR, Mullican MD, Mashal RD, Wilson KP, Su MS and Murcko MA. Chemogenomic approaches to drug discovery. Curr Opin Chem Biol 2001;5:464–470. 77. Kopka J, Fernie A, Weckwerth W, Gibon Y and Stitt M. Metabolite proﬁling in plant biology: platforms and destinations. Genome Biol 2004;5:109. 78. Van Laethem F, Liang X, Andris F, Urbain J, Vandenbranden M, Ruysschaert JM, Resh MD, Stulnig TM and Leo O. Glucocorticoids alter the lipid and protein composition of membrane rafts of a murine T cell hybridoma. J Immunol 2003;170:2932–2939. 79. Magee AI and Parmryd I. Detergent-resistant membranes and the protein composition of lipid rafts. Genome Biol 2003;4:234. 80. Pike LJ. Lipid rafts: heterogeneity on the high seas. Biochem J 2004;378:281–292. 81. Pike LJ, Han X and Gross RW. EGF receptors are localized to lipid rafts that contain a balance of inner and outer leaflet lipids: a shotgun lipidomics study. J Biol Chem 2005;280(29):26796–26804.

85

Trends in the development and application of functional biomembrane surfaces Tzong-Hsien Lee and Marie-Isabel Aguilar Department of Biochemistry and Molecular Biology, Monash University, Victoria 3800, Australia Keywords: bilayer lipid membranes; supported lipid monolayer; tethered lipid bilayers; immobilised phospholipid monolayer; gel-entrapped liposomes; membrane microarray; nanoarray

Introduction Biological membranes form the interface between a cell and its local environment and thus play a central role in all biochemical processes centred around the cell. The principal role of lipid biomembranes involves maintaining a concentration gradient of compounds inside the cells of living organisms, which are different from those outside together with maintaining an electric potential difference between the interior of the cell and the external ﬂuid. Membrane proteins are involved in energy production, signal transduction, cellular recognition, immunomodulation and ion/molecular transport. Functional membrane proteins are vital to maintain cellular homeostasis and speciﬁc defects are associated with many known disease states. Membrane proteins are also the targets of a large number of active pharmacological and toxic substances and are also responsible for their uptake, metabolism and clearance. Consequently, there is an enormous effort in developing systems to characterise the unique molecular structures of all membrane components. Our current understanding of the structure and function of biomembranes is based on the ﬂuid mosaic model proposed by Singer and Nicholson [1] which continues as the framework for the dynamic structure of biomembranes. It is now becoming increasingly recognised that biomembranes are organised laterally into compositionally and functionally speciﬁc domains or compartments [2–4]. The conﬁnement of membrane lipids and proteins into specialised domains involves hydrophobic and electrostatic interactions, van der Waals dispersion, hydrogen bonding hydration forces and steric elastic strain between the speciﬁc lipid–lipid and lipid–protein components. Interactions among lipids and between lipids and proteins are also selectively modulated and regulated by a number of membrane binding ligands as well as environmental changes. Corresponding author: Tel: +61-3-9905-3723. Fax: +61-3-9905-5726.

E-mail: [email protected] (M.-I. Aguilar). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12004-9

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

86 The function of biomembranes is therefore determined by (i) chemical composition, (ii) physical state and (iii) membrane organisation, all of which are interdependent. Investigating the interplay between the chemical composition and the physical properties of membrane systems in response to the interior and exterior stimuli and the environmental changes will lead to an understanding of the functions of the cell surfaces. However, the large number of molecular components in a biomembrane provides an enormous challenge both conceptually and experimentally to develop systems that can be used to characterise membrane structure and to delineate membranemediated processes at the molecular level. Due to the molecular diversity of biological membranes, different techniques and model membrane systems have been explored to delineate the biological processes, such as energy production, permeability barrier of ions and molecules, nerve conduction, molecular transport, immune recognition, cytolysis, etc. in physical and chemical terms. The most widely adopted experimental approach is through reconstitution of model membranes from a few lipids and membrane-associated proteins which has certainly facilitated the elucidation of fundamental structure–function relationships of protein– membrane interactions at the molecular level. Coupled with recent advances in biophysical techniques, the conformation, dynamics and thermodynamics of a peptide/protein can now be investigated in a chemically deﬁned lipid membrane environment. This review provides an overview of the current developments in the design of various model biomembrane systems. Membrane model systems mimicking the membrane physicochemical properties, exhibiting long-term stability are used widely for investigating a variety of membrane-related physical, chemical and biological phenomena. The information gained from these studies offers direct molecular insights into the physicochemical basis of membrane-related processes. Integrating the available information on the physical properties, chemical structure, and biochemical function of the protein–membrane system and lipid organisation will provide the basis to understand membrane function at the molecular level and for the development of biomembrane-based sensing devices. Further integration of the membrane model systems with various detection devices can also offer a new dimension in developing biosensors and molecular devices. Artiﬁcial lipid membrane models The investigation of membrane-related processes requires the development of model membrane systems that allow screening on a reasonably highthroughput scale. Biological membranes directly isolated from cells of living organisms are the major consideration of most membranologists in studying the structure–function relationship of biomembranes. However, due to the heterogeneous and complex components found in natural biomembranes,

87 valuable complementary information on the physical properties and functional roles of individual lipid species in membranes and the effects of the membrane on the structure and activity of peptides/proteins can be obtained through the use of artiﬁcial lipid membrane model systems containing the lipids, proteins, carbohydrates, other organic and inorganic species of interest. The preparation of model membrane systems generally involves several steps, namely (i) isolation of lipid molecules extracted from the cell source of interest or the chemical synthesis of a given lipid, (ii) construction of an appropriate model system containing that lipid molecule and (iii) subsequent incorporation of a particular peptide/protein. Pure synthetic lipids provide great advantages in understanding the relationships between chemical structure, polymorphic behaviour and dynamic properties of the lipid membrane and the inﬂuence of environmental factors, such as temperature, pH, ionic strength, hydration level and organic solvent. Owing to the variety of cell membranes, investigators combine different techniques and artiﬁcial membrane models to explore biological processes. These membrane systems can range from simple models consisting of a single phospholipid to very complex platforms in which lipids, proteins, carbohydrates and their complexes can interact. In general, the two most widely used model membranes are lipid membranes of planar or spherical conﬁguration (Fig. 1). Spherical vesicles and planar lipid membranes are complementary to each other in their structure and organisation since both types can be derived from common amphipathic lipids and other related compounds. Both the planar and spherical conﬁguration allows for the application of a number of powerful surface- or interface-sensitive experimental techniques that can give a very detailed picture of the structural and dynamic organisation and order in these systems. The formation and application of artiﬁcial biomembrane models in various formats are described in the following sections. Free-suspended lipid membranes A bilayer lipid membrane (BLM) is formed from two layers of amphiphilic lipid molecules (phospholipids, sterols, glycolipids), in which one part of the molecule is hydrophilic and the other lipophilic, in such a way that hydrocarbon chains of the lipids are oriented away from water whereas the polar groups of lipids are in contact with water. Hydrophobic parts of the lipid molecules aggregate, which is the driving force of the self-assembly of such structures in aqueous solutions. Such structures can fuse with one another and exhibit self-sealing behaviour. A modiﬁed or reconstituted BLM is viewed as a two-dimensional (2D) dynamic ﬂuid system that changes in response to environmental stimuli. To impart relevant biofunctions in BLMs, a variety of compounds, such as ionophores, enzymes, receptors, pigments, etc. can be incorporated. The freesuspended BLMs are formed at small Teﬂon oriﬁces (typically with diameters

88

Fig. 1. General conﬁguration of artiﬁcial lipid membrane models. The lipid molecules can self-assemble into spherical vesicles in solution or form a planar monolayer/bilayer structure on solid substrates.

of 2.0 mm or smaller) separating two compartments containing aqueous solutions. Separation of two aqueous solutions by the BLMs allows electrical measurements using macroscopic electrodes. At equilibrium, the planar lipid membrane consists of an oriented bimolecular ﬁlm in which the electrical properties of surface potential, capacitance and conductance provide information related to the bilayer thickness and long-term dynamics. Subsequent to understanding these fundamental properties, the BLMs become useful for the investigation of biological transport mechanisms. Precise capacitance, conductance and impedance measurements, both in the absence and in the presence of ionophores, have contributed to the understanding of impulse and ion-transfer mechanisms. Additionally, the BLMs provide a unique system that allows the assay of the functional activity, and the measurement of biophysical properties of various membrane-associated peptides/proteins, carriers and channels translocating ions across the membrane [5]. Since the pioneering work by Mueller et al. [6,7] several methods have been developed for the artiﬁcial formation of planar BLMs across a small window (diameter 1 mm or less) in polyethylene ﬁlms, or other types of hydrophobic ﬁlm, between two aqueous phases. In the earliest work, the lipid membrane

89 was formed by spreading a solution of lipids with a thin brush across the window immersed in aqueous solutions [6–8]. When solvent diffuses into the bulk of an aqueous phase, a lipid bilayer is formed on the window; interference colours are observed initially, then the layer becomes black. The types of solvent and lipids used and the size and geometry of the pinhole, as well as the methods of preparation determine the amount of solvent molecules which remain in the BLMs. This procedure can be performed with commercially available devices. A lipid solution can also be introduced into the hole by means of small syringes or pipettes with plastic tips. Comparably simple is the ‘‘tip–dip’’ technique, in which the 1–2-mm diameter capillary tip of a glass pipette is immersed in an aqueous solution with a lipid monolayer on the surface, is then taken out, and is then immersed again [9]. In the course of these operations a planar BLM is formed on the tip of a capillary ﬁlled with aqueous solution. This method has been successfully used, for instance, for the formation of ionophore-incorporated BLMs [10]. The formation of the lipid bilayer can be followed by microscopic observation of the black lipid membrane, by measurement of the capacitive current passing the BLM, or its resistance; it can also be characterised electrochemically by use of channel peptides [5] on the basis of the transport of monovalent cations through the membrane. The ability of lipids to form highly oriented monolayer structures at air–water interfaces has also been used to prepare BLMs [11]. This method of BLM folding from monolayers can be significantly improved by appropriate use of the Langmuir–Blodgett (LB) casting technique. In the early set-up, a Teﬂon ﬁlm 25–500 mm thick with an aperture of 0.25–1 mm was placed in the middle of a vessel with a hole above the level of the aqueous solution. After introduction of a lipid layer onto the surface of the aqueous solution and evaporation of the solvent, a Teﬂon partition was lowered in which two monolayers are joined as a result of interaction of the hydrophobic parts of lipid molecules. As showed in later experiments, a surface area of membranes >0.1 mm2 formed in this manner will greatly reduce the stability of the bilayer structure [12]. The bilayer formed by the monolayer apposition method is free from the destabilising effects caused by organic solvent and these membranes are thinner, have more stable capacitance and are better targets for the recognition of channel-forming toxic proteins. The solventfree bilayer can also be prepared from a solution of unilamellar vesicles [13]. The equipment set-up for the preparation of free-suspended planar BLMs [10,14] is shown in Fig. 2. Such a cell allows a simultaneous measurement of the transmembrane current. The BLMs were formed across a small, smooth, circular aperture produced in 7.5 mm polyimide ﬁlm by passage of an electrical spark. The aperture diameters used were 120 and 20 mm, respectively, for multi- and single-channel sensing BLMs [15]. In another version of this method, a lipid solution was placed on the surface of an aqueous solution from one side of the partition and the solution level was then lowered by

90

Fig. 2. Experimental arrangement for membrane formation using the folding

method, and electrochemical measurements with an ionophore-incorporated BLM [10]. The Teﬂon chamber is placed on a magnetic stirrer, which is placed in a Faraday cage on a vibration-free table. The planar BLM is formed across a small aperture in the Teﬂon ﬁlm.

pumping out below the aperture; it was then again elevated above the aperture. When the aperture diameter was 0.32 mm, the BLM system obtained was stable for more than 4 h, with lifetimes up to 24 h. BLMs prepared from solutions of oxidised cholesterol are stable even after storage for 1 month [16]. Lipids commonly used for formation of artiﬁcial BLMs include components extracted from a variety of tissues from living organisms, as well as synthetic lipids, and also other cationic, anionic and non-ionic amphiphilic compounds, porphyrins and pigments. Free-suspended BLMs have been employed for several analytical purposes and when unmodiﬁed also have high resistivity, which makes them excellent insulators. However, several shortcomings are apparent for BLMs. The geometric shape of the chamber, conformation of the BLMs, the presence of a variable unstirred layer directly in front of the BLMs, variability in the BLM surface area and the presence of residual organic solvents often affect the reproducibility in the assay for channel activity in BLMs. The poor stability of BLMs in response to hydrostatic pressure-induced bending in a lipid bilayer is also a limiting factor [17]. This bilayer bending in a BLM system is accompanied by a dynamic exchange of lipid and solvent molecules and is manifested by the reorientation of lipid molecules in the bilayer membrane. A consequence of these two molecular events is increased penetration of the solvent into the hydrophobic parts of the BLMs. The mechanical stability of BLMs can be improved by use of a variety of chemical additives in the solutions used for formation of the membranes. For example, cholesterol has been found to stabilise free-suspended BLMs [18]. This effect has been attributed to the modiﬁcation by cholesterol of the short-range repulsive interactions between phosphatidylcholine (PC)

91 membranes [19]. Stearylamine has been found to have a pronounced stabilizing effect on BLMs over a range of negative potentials [20]. Destruction of BLMs as a result of dehydration can be prevented by use of some saccharides in the bilayers; this is attributed to replacement of water bound to the polar heads in membrane phospholipids. Such effects were observed for polysaccharide derivatives bearing hydrophobic anchor groups [21], for sucrose [22] and trehalose [23]. It has also been shown recently that planar BLMs formed by the folding of monolayers of PC containing bolaamphiphilic steroid dimer are more stable than those without the dimer [24]. Because electrocompressibility is reduced by a factor of approximately 20, the membrane formation is shorter; these membranes can be used to construct capacitance sensors. Other methods used to improve the stability of planar BLMs include blending with hydrophobic polymers [25], use of polymerisable [26] or photoactivable [27] lipids, or the use of phospholipids with highly branched hydrophobic chains [28]. High stability is observed for BLMs with in situgenerated microcrystalline inorganic semiconductors; the possibility of preparing arrays of BLM supported semiconductors might ﬁnd application in the design of analytical microdevices [29]. Inconvenient high resistivity of planar, free-suspended BLMs can be eliminated by incorporating conducting electron mediators, such as 7,7,8,8-tetracyanoquinodimethane or tetrathiofulvalene [30–32] or conducting polymers, e.g., polypyrrole formed in BLM by chemical oxidation of monomer [33], into the membranes. The polypyrrole–lecithin BLM is electrically conductive and its mechanical stability is good. Despite the generally limited stability of free-suspended BLM systems, several analytical applications have been published in the literature. They have, for example, been used as transducers of immunochemical and enzymatic reactions [34,35], in ion-channel-based sensors for glutamic acid [15], adenosine 50 -triphosphate [36], and D-glucose, and for evaluation of the chemical selectivity of agonists towards receptor ion-channel proteins. Potentiometric sensing with ionophore-incorporated BLMs is used to detect selected insecticides, the herbicide atrazine and aﬂatoxin metabolites [37]. The use of protein–lipid vesicles provides an easy way to incorporate proteins into the planar lipid bilayer and to screen the effect of various drugs on the activity of membrane proteins. The well-deﬁned set of components including water, salts, buffers, lipids and organic solvent in the set-up of free-suspended BLMs provides a feasible way to examine the effects new proteins and peptides incorporated into BLMs. Additionally, a wide variety of ionic compositions, lipids, temperatures and pH conditions can be manipulated to study the biophysical properties of protein channels in question at various physiological conditions. The molecular events of the transition between the open and closed state in ion-channel proteins have been investigated via the development of simultaneous electrical and spectroscopic characterisation of BLMs. Reﬂectivity [38], laser-intracavity absorption [39], steady-state and time-resolved ﬂuorescence

92 [17], holography [40], and interferometry [41] have all been utilised for simultaneous characterisation of BLMs with electrical measurements. Formation of a planar lipid bilayer on a microchip device has also been developed for automated multichannel studies of membrane proteins. The bilayer membrane is formed at the etched aperture with a diameter of 50–200 nm on a silicon substrate using either a microﬂuidic system or a liquid droplet injection system [42]. It consists of two ﬂuidic channels on both sides of the silicon substrate and they are connected with apertures (Fig. 3). A lipid bilayer is formed at the aperture by ﬂowing lipid solution and buffer alternatively. After the bilayer is formed, membrane proteins can be incorporated through for example proteoliposome fusion, and the activity can be monitored with the integrated electrode. In this microchip set-up, the thickness of the initial layer of phospholipids on the aperture is crucial for the bilayer formation. A high amount of lipid solution, which results in a thick layer of lipid on the aperture, should be avoided. The wetting property of the substrate is another crucial factor during the spreading of the lipid solution which is wettable, such as SiO2 and Parylene, for the lipid solution spreading out easily. The substrate has to be electrically insulated to diminish the noise for the electrophysiological recording. Supported lipid membranes The formation of biomembranes on various types of planar solid supports, either unmodiﬁed or chemically modiﬁed, can be divided into three major

Fig. 3. Microﬂuidic chip devices of a protein-incorporated planar lipid bilayer [42]. The channel width and depth are 500 and 100 mm, respectively. The aperture size is 50–200 mm. The bottom channel is ﬁlled with buffer. The surface of the buffer stays at the front of the aperture because of the surface tension. The lipid solution is introduced to the upper channel and ﬂushed with air and a thin layer of lipid monolayer is formed over the buffer surface at the aperture. The bilayer is then formed by passing either the same or different lipid solution through the upper channel at low ﬂow rates of 0.1 mL/min.

93

Fig. 4. Various types of membranes formed on solid supports.

categories according to the ﬁnal structure of the lipid on the solid surfaces which are illustrated schematically in Fig. 4. In the ﬁrst type of solid supported membrane, lipid bilayers are adsorbed onto a hydrophilic surface, which can be prepared by the transfer of a lipid monolayer formed at the air/ water interface to solid supports using the Langmuir technique or by the adsorption of lipid vesicles (Fig. 4A–C). The second type of membrane system is composed of a phospholipid monolayer on a hydrophobic support, which is modiﬁed with either a short or long alkyl chain (Fig. 4D and E). The third type of membrane system is the covalently immobilised/anchored lipid monolayer or bilayer on chemically modiﬁed hydrophilic surfaces. Since the initial preparation of molecular monolayers or multilayers on solid substrates, the so-called ‘‘LB ﬁlms’’ have been used to study the physicochemical properties of various molecules on the surfaces of solid materials, to fabricate new materials with novel surface properties [43,44], and as model systems for biological membranes, as well as for their application in electrochemical sensors and biosensors [45]. The preparation of LB ﬁlms with the LB technique involves moving a clean plate in a vertical or horizontal orientation (the solid substrate) through an aqueous solution–air interface resulting in the transfer of the lipid monolayer onto the planar solid support during immersion or withdrawal and the lipid ﬁlms can be removed with organic solvents or aqueous detergent solutions. The orientation of lipid molecules in LB ﬁlms is dependent on the surface properties of the solid

94 substrate. For hydrophobic substrates, the tails of lipids are preferentially adsorbed onto the surfaces of solid substrate and the monolayer is transferred during dipping. In contrast, the head groups of lipid molecules are adsorbed onto polar substrate surfaces and monolayer transfer preferentially occurs during the withdrawal process. The organisation, surface density and number of lipid layers can be controlled by repeated withdrawal and dipping of a hydrophilic substrate through the monolayer and leads to the formation of an alternative head–tail–tail–head multilayer LB ﬁlms. Such tail-to-tail hydrophobic interactions and head-to-head electrostatic interactions greatly increase the stability of the LB ﬁlm. Thus, different lipid compositions in each leaflet of a multibilayer can be constructed by alternatively changing the lipid composition during the preparation of the Langmuir monolayer. However, difﬁculties have been reported in the deposition of the second monolayer in which the ﬁrst monolayer is retransferred back onto the surface of the Langmuir trough when the covered substrates are immersed in the subphase [46,47]. The effectiveness of monolayer adsorption and assembly onto a substrate relies on the preparation of a stable, well-compressed and homogeneous monolayer on the Langmuir trough. Additionally, a constant, steady and slow (in the order of millimetres per second) speed is required to achieve a stable monolayer transfer to the solid substrate and above a critical transfer velocity no monolayer can be formed. Controlled evaporation of organic solvent from phospholipids in organic solution and adsorption of small unilamellar vesicles (SUVs) and medium unilamellar vesicles (MLVs) onto solid substrates are commonly used as alternative methods to LB ﬁlms for the formation of ultrathin ﬁlms which retain the regular bilayer structure of biomembranes. For the multibilayer membrane, each of these stacked bilayers on the solid surface is separated by a water layer about 1–2.5 nm in thickness. Such oriented lipid multibilayers have been frequently used in polarised transmission and attenuated total reﬂection (ATR)-IR studies, which provide information on the structure and organisation of pure lipid bilayers in response [48]. The formation of single bilayers on hydrophilic supports can be achieved by incubation of the lipid vesicle dispersions on substrates [49,50] or by dialysis of a detergent solution of the lipid in the presence of the substrate materials [51]. The success and ease in forming a bilayer structure from liposomes on a solid support is dependent on several factors including type, structure (including roughness) and cleaning of the support, composition and size of the liposomes, composition of the surrounding medium, temperature for the bilayer formation and liposome preparation, the protein content and the geometry of and ﬂow dynamics in the fusion cell. Fully ﬂuid supported bilayers are known to form by vesicle fusion on highly hydrophilic mica, silica solid surfaces (i.e., glass, quartz, thermally grown silicon dioxide and sputtered silicon dioxide), and oxidised poly(dimethylsiloxane) (PDMS), which displays a hydrophilic SiOx surface. However,

95 intact vesicles adsorbed onto a SiO2 surface were found when the amount of liposomes is below the critical surface coverage [52]. In contrast, exposure of lipid vesicles to other hydrophilic solid surfaces results in lipid structures other than a fully ﬂuid bilayer in which an intact vesicular layer was formed on oxidised gold, oxidised platinum or titanium oxide. In addition to the planar solid substrate, a membrane bilayer can also form on spherical substrates, such as polystyrene-divinylbenzene, glass and silica spherical beads [53–55]. The dynamic properties of the lipid molecules were retained in a single bilayer of either a single species or a binary mixture of phospholipids adsorbed onto silica particles (pore size 4,000 A˚, diameter 1071.5 mm) [56]. Transition endotherms measured by differential scanning microcalorimetry revealed that the lipid bilayer systems that consisted of two lipid species on the spherical support exhibited a rather broad and complex feature consistent with partial demixing of two lipids in the gel phase. In contrast, the endotherm for adsorbed bilayers with a single lipid species showed speciﬁc heat changes consistent with the presence of a single phospholipid species. In addition, the transition temperatures for the single and binary lipid bilayers were also different. These unique membrane-like thermodynamic properties were then exploited to provide a selective puriﬁcation system of the peripheral membrane protein, myristoylated alanine-rich C kinase substrate (MARCKS)-related protein [57]. This bilayer-adsorbed spherical solid support thus provides charge-selective protein separation under biocompatible conditions solely by changes in the column temperature without changes in ionic strength, which allowed the native conformation of the protein to be maintained during the puriﬁcation. In addition to the adsorbed bilayer, the perfusion of phospholipids dissolved in organic solvent (85% isopropanol) into reversed phase chromatographic supports resulted in a hybrid membrane bilayer in which a monolayer of lipids is non-covalently adsorbed onto the hydrophobic alkyl chains [58]. The amount of adsorbed phospholipids was found to be similar to that of lipid vesicles, thereby conﬁrming the formation of a bilayer-like structure. This system was stable in solvents containing less than 35% acetonitrile and was used to study peptide–lipid binding properties. For lipid monolayer covered substrates, the pre-deposited lipid monolayer provides a very stable surface when subjected to harsh conditions. The liposomes intended to form the outer lipid layer must make contact with the alkyl chains of the deposited monolayer. The underlying solid support is important for the process, with platinum (Pt) generally resulting in better transfers than Si/SiO2 [59]. Liposome fusion to bare and lipid-coated Pt results in membranes with similar characteristics [60]. Liposomes prepared from unsaturated PCs, such as POPC or DOPC are most frequently used for forming supported planar bilayer structures, as they are in the liquid crystalline state at room temperature, facilitating rapid spreading and formation of a ﬂuid membrane on different types of support. The unsaturated PCs are

96 inferior to DPPC for obtaining membranes with the required qualities. High ﬂuidity in the hydrocarbon chains per se appears not to be the reason for this effect, since cardiolipin is polyunsaturated but had no significant effect on the responses. A mismatch due to the differences in area occupied per molecule in the densely packed monolayer and the much greater area occupied by each unsaturated PC in the liposomes might be a reason for an inferior hydrophobic interaction, with consequences for the transfer. Divalent cations were found to promote attachment of mixed PC/ phosphatidylglycerol vesicles to form supported vesicle layers and bilayer formation from pure PC vesicles. Adding a high concentration of calcium ions promoted lipid transfer from the liposomes to plain hydrophilic supports and simpliﬁed the adsorption pattern. Other ‘‘intrinsic’’ fusogenic agents such as protein and a high DPPE/cholesterol ratio also often promoted the transfer of lipids and bilayer formation to the solid surfaces. The combination of LB ﬁlms and vesicle fusion provides an extra dimension for the preparation of a ﬂuid bilayer on a hydrophilic support [54]. In addition to the planar bilayer systems, a single phospholipid bilayer adsorbed onto spherical particles, such as microglass beads (0.3–10 mm in diameter) and polysaccharide nanoparticles (30–75 nm in diameter), has also been developed and the dynamic properties of the bilayer also characterised by 31 P-, 2H-NMR and ﬂuorescence spectroscopy [54,61]. The single phospholipid bilayers on hydrophilic supports can be fully immersed in water or physiological buffers. The adsorption or fusion approaches are advantageous for their easy vesicle preparation and the incorporation of transmembrane proteins from proteoliposomes. The thin ﬁlm of a bilayer membrane adsorbed on optically transparent substrates, such as glass, mica and quartz is ideally suited for detailed structural studies of membrane structure, phase changes and speciﬁc membraneassociated proteins and peptides via neutron/X-ray diffraction, electron microscopy, circular dichroism and FT-IR spectroscopy. Covalently immobilised lipid membranes Membrane lipids can be covalently attached onto solid substrates either directly by the reactive groups coupled to the head groups or via the o-end of acyl chains. Phospholipid molecules can be immobilised onto the solid supports via the reactive groups coupled at the end of acyl chain (Fig. 4F and G). Chemisorption of PC consisting of thiol groups or heterocyclic disulﬁde groups at the terminus of fatty acyl chains formed a stable phospholipid monolayer on gold with the choline head groups extended outward from the surface [62]. As studied by a combination of ellipsometry, contact angle measurements, heterogenous electron-transfer properties and X-ray photoelectron spectroscopy, a monolayer consisting of thiol-coupled PC molecules showed very similar surface properties to biomembranes. However, the

97 liquid–crystalline properties of the monolayer have not been demonstrated in such a lipid monolayer system. Membrane lipids have been covalently attached onto silica-based chromatographic supports via the reactive groups coupled at the o-end of the acyl chain. A wide range of phospholipids modiﬁed with reactive groups at the end of acyl chains are shown in Fig. 4. The head groups exposed to the aqueous environment in these phospholipids covalently immobilised chromatographic supports exhibit similar physicochemical properties found at the biomembrane surfaces. The reaction of carboxyl groups attached at the o-end of phospholipids with an aminosilane on the silica surface resulted in the formation of an immobilised lipid monolayer on the surface of chromatographic supports [63] (Fig. 4F). In addition to the use of carboxyl groups as a functionalised moiety, the synthesis of a new class of immobilisable glycerophospholipids containing amino groups at the o-terminus of each of the sn-1 and sn-2 acyl chains has also been reported [64,65]. In addition to the immobilisable PC, other lipid analogues, such as phosphatidylglycerol, phosphatidylethanolamine, phosphatidylserine and phosphatidic acids with different charges, size and hydration properties have also been synthesised [66,67]. The methods used for the preparation of the immobilised glycerophospholipid monolayer generally consist of three sequential steps: (i) the activation step in which the silica is pre-treated and subsequently reacted with an organosilane with deﬁned functionality whereby it is covalently bound to the silica, (ii) the coupling step in which the activated silica is subjected to reaction with the biomimetic ligands under gentle conditions to couple the ligands, and (iii) the end-blocking step in which the residual activating groups from the coupling steps are removed by means of appropriate reactions. An example of an immobilised PC monolayer on an activated silica support is shown in Fig. 5. Thus, immobilisation of lipid analogues either as a single species or a mixture in different proportions onto the activated spherical chromatographic supports therefore provide a series of model biomembranes which can closely mimic the lipid compositions found in naturally occurring membranes. The covalently immobilised lipid monolayers on a silica substrate exhibit long-term stability under aqueous conditions and also in the presence of organic solvent [63–65]. In addition, the types of salts, ionic strength, pH and organic solvent composition in the mobile phase can be optimised to provide baseline separation of drugs and membrane protein mixtures [68]. As the residual reactive groups on the surfaces are blocked with small chemical compounds, the leaching of lipids from the column is greatly diminished under slightly acidic conditions and when stored in detergent solution for long periods of time. However, perfusion of an immobilised biomembrane column with 0.1% triﬂuoroacetic acid (TFA)/H2O, 0.1% TFA/acetonitrile, 35 mM citric acid, 35 mM NH4H2PO4 and 35 mM NH4Cl resulted in

98

Fig. 5. The preparation of immobilised phosphatidylcholine (PC) monolayer on ac-

tivated silica supports. The pre-treatment of silica at 1801C forms fully hydroxylated silica, which is reacted with 3-isothiocyanato-propyltriethoxysilane (ITCPS) to provide the speciﬁc functional groups (–NCS) on the activated silica. The immobilisable PC derivatives consisting of either one or two amino groups at the o-end of the acyl chains are covalently attached to the –NCS groups on the activated silica with the formation of a thiourea derivative. The residual –NCS groups are further reacted with octylamine to form the ﬁnal immobilised PC monolayer.

approximately 2% of the immobilised lipid molecules leaching from the column. Although the lipid density on the silica particles is similar to that found in natural biomembranes, an important disadvantage associated with the immobilised lipid monolayer is the lack of lipid molecular dynamics, such as lateral diffusion, ﬂip–ﬂop and axial displacement, due to the covalent linkage of the lipid molecules to the silica surface. However, the lack of lipid dynamics associated with the ﬂuid biomembrane surfaces is compensated by the increased stability derived from the covalent immobilisation of lipids onto solid surfaces. Thus, the immobilised phospholipid monolayers on the solid supports have unique applications compared to the immobilised liposomes. In particular, organic solvents and detergents can be included in the buffer for screening the drug binding, examining peptide/protein–lipid interactions and separating membrane-associated proteins. Immobilisation of liposomes and proteoliposomes Encapsulation in soft gels Non-covalent immobilisation of erythrocytes, proteoliposomes and lipid vesicles into a matrix of chromatographic gel particles represents another

99 class of immobilised biomembrane system and has been used for membrane protein puriﬁcation in detergent solution [68]. Such covalently immobilised liposomes exhibited high stability at high salt concentration, together with desirable membrane permeability properties and solute-partition properties and have been used for the measurement of drug–membrane partitioning and chromatographic refolding of denatured proteins. Liposomes and proteoliposomes have been immobilised into various types of soft gel beads in a similar way to the entrapment of whole cells into soft gel supports [69], as illustrated in Fig. 6. The gel matrices used for the immobilisation of liposomes and proteoliposomes require several considerations: (1) the size distribution of the gel beads, (2) the pore size and surface area of the gel matrices and (3) the stability of the gel support. The relatively large agarose (Sepharose) and allyldextran-bisacrylamide (Sephacryl) gel beads were used in the early development of this class of biomembrane

Fig. 6. The general outline of preparation of a proteoliposome-entrapped gel chromatographic support. (1) The proteoliposomes are prepared in the presence of detergent with high CMC and separated with gel ﬁltration such as Sephadex G50 to remove detergents and other small molecules. (2) The prepared proteoliposomes are fractionated according to size on, for example, Sepharose 4B. (3) Immobilisation of fractionated proteoliposomes via hydrophobic interaction into a large pore gel, such as Sepharose 2B or Sephacryl S-1000.

100 chromatography [70]. However, their broad size distribution and large pores limited the use of high ﬂow rate and poor resolution was obtained during the separation process. In contrast, Superdex 200 PG and TSK 6000 PW provide a superior sorbent as a result of their relatively homogeneous size distribution and larger surface area. In addition, the relative rigidity of these gel beads allows higher ﬂow rates to be employed resulting in superior separation [71]. Several immobilisation methods have been developed including steric entrapment [72,73], hydrophobic binding [74] and avidin–biotin afﬁnity binding [75]. Liposomes consisting of PC, phosphatidylethanolamine and phosphatidylglcerol have also been covalently immobilised onto chloroformate-activated gels [76,77]. The steric entrapment methodology was used to immobilise liposomes/ proteoliposomes in gel beads of Sepharose 6B and Sephacryl S-1000 [70]. The non-trapped lipid vesicles are then removed by chromatographic procedures and by centrifugation. The amount of entrapped liposomes can be increased by increasing the initial lipid concentration. The amount of entrapped liposomes also depends on the size of prepared unilamellar lipid vesicles whereby higher amounts of vesicles could be entrapped in gel beads when large unilamellar vesicles with average diameter of 230 nm were used. However, this preparative method resulted in a material which exhibited short-term stability. Freeze–thawing a mixture of suspended biomembranes and dried gel beads was developed as an alternative way to induce fusion and entrapment of liposomes in gel beads, such as Sephacryl S-1000 and Superdex 200 [73,78]. The amount of immobilised liposomes was dependent on the temperature of the freeze–thawing cycle and the dimensions of the tube. Other methods involve freeze drying the liposomes together with gel beads followed by the rehydration of the dried liposome–gel beads or reverse-phase evaporation of organic solvent in an aqueous emulsion of lipids with gel beads [73] or by dialysis preparation of (proteo)liposomes in beads [70,79]. Non-covalent adsorption of liposomes has also been achieved on the hydrophobic alkyl chains, such as butyl (C4), octyl (C8) and octadecyl (C18) derivatised gel beads such as Sephacryl S-1000 [74,80]. The alkyl derivatised Sephacryl S-1000 exhibited higher ligand density than the Sepharose adsorbents. Small, medium and large vesicles of radii of approximately 20, 50 and 100 nm were immobilised onto octylsulﬁde-Sephacryl S-1000 in amounts corresponding to 110, 40 and 20 mmol of phospholipids per millilitre gel [74]. Although high lipid loading was obtained with the use of small unilamellar vesicles, partial loss of vesicle integrity was found during the adsorption process, as evident from the release of calcein entrapped in the vesicles during the adsorption process. From the determination of entrapped calcein, radii of 20 and 50 nm were estimated for the adsorbed small and medium vesicles, respectively. Phospholipid analogues functionalised with carboxyl groups at the o-end of the acyl chain have also been immobilised onto modiﬁed substrates, such

101 as AH-Sepharose 6B and AH-Sepharose 4B [81–83]. However, the lipid density after the covalent immobilisation onto the soft gel matrices was low and unlikely to form a bonded monolayer. Nevertheless, this type of immobilised lipid on a soft gel provided an afﬁnity matrix for the puriﬁcation of various lipases and enzymes involved in the biosynthesis of lipids. Modiﬁcation of the lipid head groups with various functional groups can be used to immobilise lipid vesicles either non-covalently or covalently onto the activated surface of gel beads. The binding of biotin to avidin or streptavidin is one of the alternative approaches to non-covalently adsorb liposomes onto solid surfaces. In this method, the introduction of biotin into the head group of phospholipids provides an anchor group to gel beads modiﬁed with avidin or streptaavidin [75]. Liposomes bearing photoreactive lipids have been bound to gel beads by the freeze–thawing/entrapment method followed by irradiation. The resulting liposomes were more stable than the entrapped liposomes during chromatography of basic proteins under gradient elution conditions [73]. The covalent immobilisation of liposomes onto the surfaces of gel beads has also been achieved through the incorporation of amino-containing lipids, such as phosphatidylethanolamine into PC-based liposomes. In this approach, the surfaces of gel beads were activated with amine-reactive groups, such as chloroformate, CNBr, hydroxysuccinimide and tresyl groups [77]. However, the liposomes without the amino-containing lipids, such as PC and phosphatidylglycerol can also be covalently immobilised onto the chloroformate-activated gel through a phosphoester linkage between the phosphate moiety of the phospholipid and the carbonate moiety of the active group by nucleophilic phosphate catalysis. The covalently immobilised liposomes were highly stable against osmotic effects such as the presence of guanidinium chloride at concentrations up to 5 M. However, the covalently immobilised liposomes were permeable to the entrapped hydrophilic molecules upon storage as a result of membrane defects caused by the change in packing geometry of the coupled lipids in the membranes. Nevertheless, this immobilisation method provided a simple, economic and less time-consuming process than the preparation of the avidin–biotin immobilised liposomes. Encapsulation in sol– gels An emerging method for the immobilisation of liposomes and proteoliposomes is their entrapment within inorganic matrixes formed by the sol–gelprocessing method [84–90]. A schematic outline is shown in Fig. 7. Sol–gel is a chemical synthesis technology employed in preparing gels, glasses and ceramic powders. Synthesis of materials by the sol–gel process generally involves the use of metal alkoxides, mostly but not exclusively the SiO2 materials, which undergo hydrolysis and condensation polymerisation reactions to produce the gel. The method involves the formation of a colloidal sol solution resulting from the hydrolysis of suitable silane precursors, such

102 Bilayer Membrane

Membrane Receptor Expression and Isolation

5 nm

Receptor proteins

Self-Assembly 5-100 µm

Phospholipid molecules

Proteoliposome

Sol-Gel Process

Tetramethylorthosilicate (TMOS) Sol

Rotary evaporation Entrapment into Sol-gel substrate

Water

PEG

Fig. 7. Schematic illustration of the entrapment of receptor proteoliposomes into

nanoporous sol–gel matrices for the development of nano/microarray and microﬂuidic sensor devices.

as tetramethylorthosilicate (TMOS), tetraethyl orthosilicate (TEOS) and diglycerylsilane (DGS) [87]. A buffered aqueous solution containing the liposomes or proteoliposomes is added to the sol to initiate a rapid polycondensation reaction, which produces a hydrated gel that effectively entraps the liposomes without the need for tethering to a solid surface. The sol–gel porous matrices in general and doped matrices provide unique properties such as (1) optical transparency with minimal quenching of ﬂuorescence

103 reagents, (2) long-term mechanical, thermal and chemical stability therefore enhancing the stability of the encapsulated liposomes and proteoliposomes up to months and preventing leaching of the entrapped molecules inside the liposomes, (3) tailorable properties, such as surface functionalisation, thin ﬁlms, bulk materials and (4) controllable surface area and relatively uniform pore size and distribution. The porosity of the sol–gel glass allows small molecules to be diffused in, whereas the large vesicles and protein-incorporated liposomes remain physically entrapped in the matrix. New classes of precursors, based on polyol silicates and polyol siloxanes and especially those derived from glycerol, have been under intense investigation [91,92]. These precursors can be distinguished from the traditional ones by properties such as high biocompatibility and ability to encapsulate proteins, liposomes/proteoliposomes and cells under mild conditions in a reproducible manner. Such characteristics enable the sol–gel carrier to address most of the problems encountered with traditional encapsulation methods. The addition of polymers, such as PEG, as part of the solvent phase of the porous gel matrix further stabilises the liposomes and increases the porosity and ﬂexibility of the gel. It has been demonstrated that transmembrane proteins such as gramicidin A, bacteriorhodopsin, F0F1-ATP synthase, ligand-gated ion-channel and G protein-coupled receptors retain their activities over a period of at least 1 month [88–90]. Such liposome or proteoliposome-encapsulated sol–gel materials are useful in the construction of nanodimensional sensor devices, reaction chambers and may be extended to develop artiﬁcial cells/tissues for disease treatment. However, the chemical structure of sol–gel derived biomaterials changes over several months as a result of continued hydrolysis and condensation of silica. This leads to slow changes in pore morphology and silica surface chemistry, which can reduce the membrane protein activity. Further advances in sol–gel materials is necessary to better control the stability of the encapsulated liposomes and proteoliposomes. Polymer supported/tethered lipid bilayers The adsorption of lipid bilayers directly onto a hydrophilic substrate provides the simplest approach to mimicking a biomembrane. The ultrathin water layer of 12.5 nm acts as a lubricant between the bilayer and the substrate which enables the bilayer to exhibit thermodynamic properties and long-range lateral mobility similar to the natural membrane. However, the close vicinity of the solid substrate has some limitations. The functional coupling between the phospholipid head groups and the hydrophilic substrate results in slowing lateral diffusion and breakdown of the 2D ﬂuidic nature of the membrane. The incorporated transmembrane proteins usually show no 2D mobility due to the strong interaction with the substrate. The asymmetric character of the bilayer substrate assembly results in different

104 lateral mobility in the inner and outer leaflets. Although the adsorbed bilayer and hybrid bilayer systems have been widely used in optical biosensor applications, these drawbacks have limited such system for several electrical biosensor applications. The limitations mentioned above can be overcome by increasing the thickness of the aqueous lubricant layer by introducing a soft hydrophilic polymer network or polyelectrolyte ﬁlm (10100 nm) between the membrane bilayer and the solid substrate to maintain both the dynamic properties and the integrity of the functional protein–membrane assembly. The polymer ﬁlm, which acts as a spacer between solid substrate and the lipid bilayer, minimises negative substrate effects, such as defect formation, decreases lateral mobility and limited self-annealing of the membrane, thus providing important biophysical properties [93]. Furthermore, the polymer interface allows the incorporation of transmembrane proteins in their native form with the surface domains protruding into the hydrophilic polymer network. The dynamic properties of lipids and the incorporated molecules were also demonstrated by the remarkable long-range lateral mobility of lipopeptides embedded into a polymer-supported bilayer [94]. Various approaches to creating the composite polymer–lipid ﬁlms have been established and generally include the design and adsorption/immobilisation of the polymer support itself onto a solid substrate and the deposition and ﬁxation of a membrane onto the polymer support. To form a thermodynamically and mechanically stable polymer–lipid composite ﬁlm on a solid substrate (exposed to air or water) and to avoid the formation of polymer blisters, the wetting conditions have to be carefully controlled and the spreading pressure needs to be maintained positively for the bilayer on the polymer and the polymer on the solid substrate. In particular, the interactive forces between membranes and substrate have to be only either weakly attractive or repulsive to avoid dewetting of the soft layered ﬁlms following bilayer deposition [95,96]. A useful strategy to avoid dewetting is to chemically graft the polymer cushions to the surface. After the appropriate wetting conditions have been found for a given system, the composite ﬁlm forms by self-assembly. Another important aspect of the polymer-supported bilayer is the relationship between the strength of the adsorption at the polymer–lipid interface and the lateral mobility of the bilayer. In addition to the hydration level and stability of the polymer ﬁlms, high ﬂexibility and exclusion of highly charged polyelectrolytes are also major concerns in the design the polymer–lipid composite ﬁlms. The methods for the preparation of stable polymer–membrane composite ﬁlms include: (1) the chemical grafting of a ﬁlm of a highly water-soluble natural polymer, such as dextran, hyaluronic acid or poly(acrylamide), to the solid surface [94,97] and subsequent deposition of a lipid bilayer which is stabilised via attractive electrostatic interaction; (2) the deposition of soft hydrophilic or hydrophobic multilayers of rod-like molecules with alkyl side

105 chains and the subsequent transfer of lipid monolayers or bilayers [98]; (3) the reconstitution of lipopolymers (lipids with macromolecular groups coupled to the head group as shown in Fig. 8) that form stealths, thus separating the bilayer from the solid surface [99,100] and (4) the displacement of predeposited lipids by adding poly(ethylenimine) polymers to a solid-supported bilayer, thus forming a polymer-supported bilayer to which the attractive

Fig. 8. The chemical structure of lipopolymers where the lipid head groups or fatty

acids are coupled with polymer forms stealth at the membrane surface. This polymeric layer provides additional space and act as a cushion for the acts membrane separated from the solid surface.

106 polymer–substrate interaction appears to be stronger than the polymer– bilayer interaction [101,102]. The whole membrane system can be stabilised by a balance of different superimposed molecular interactions of mostly electrostatic, van der Waals and hydrophobic origins. The stability of the entire system depends on attractive potentials at both the polymer–substrate and the polymer–bilayer interfaces and their strength relative to each other. A weaker attraction at the polymer–substrate interface than the lipid–polymer was found for the case of dextran and poly(acrylamide)-based system. On the other hand, the attractive polymer–substrate interaction is stronger than the polymer–bilayer interaction in the case of the poly(ethylenimine)-based systems. An alternative approach to stabilisation of a polymer-supported lipid bilayer is based on controlled covalent tethering between the polymer cushion and the solid substrate and between the lipid bilayer and the polymer cushion. The tethering density can be controlled in a manner by adjusting the density of cross-linker molecules at the substrate through different selfassembly conditions and the density of actual tethering points through different reaction times. There are several simple but versatile techniques for covalently tethering polymers to solids (Fig. 9). The linkage is mediated by monolayers of alkyl silanes (for Si/SiO2 or indium-tin-oxide surfaces) or alkylmercaptanes (for gold and GaAs surfaces) carrying functional groups at the opposite ends, which can be covalently coupled to polymer chains. These functional groups can be: (1) epoxy groups, which are covalently bound to carboxyl groups; (2) amines, which bind covalently to carboxyl groups of chains (activated via coupling of succinimide or imidazole esters) and (3) photocross-linking groups (benzophenone silane-functionalised glass), which can be photochemically linked to any polymer segment. The density of the active anchoring sites of the grafted monolayer can be subsequently adjusted by controlled partial deactivation of the epoxy- and photocross-linkers through hydrolysis. Because this yields polar groups, the solid surface is simultaneously rendered highly hydrophilic, resulting in a significant reduction in non-speciﬁc binding of biological macromolecules. The deposition of ultrathin polymer ﬁlms by electrochemical polymerisation is another procedure used for the formation of soft functionalised polymer ﬁlms (or polymer cushions for membranes). Synthetic polymer chains (e.g., polyethyleneglycol) or oligopeptides (e.g., epitopes of antigens) are coupled to phenol derivatives, which can be polymerised at the anode electrode surface [103]. This enables selective and directed functionalisation of individual electrodes of multielectrode arrays. Bilayer deposition can occur by three methods: (1) monolayer transfer from the air–water interface with LB techniques which enables the deposition of asymmetric bilayers; (2) vesicle fusion in which lipid vesicles are deposited onto the substrate from vesicle suspensions. By reconstitution of 10% of the charged lipids, the vesicles open and form adherent bilayer patches that fuse into continuous bilayers after annealing at

107

Fig. 9. Schematic of the different approaches taken to assemble tethered bilayer architectures on solid substrates. Method 1 showed the direct adsorption of vesicles onto the quasi-dried polymer on a solid support. Method 2 shows the polymer adsorption on the lipid bilayer formed on the bare solid support. Method 3 is referred to as the top-down concept with a lipid–lipocopolymer layer being transferred from the water–air interface to a solid support, pre-coated with a reactive monolayer capable of covalently binding to some of the polymer units. Method 4 illustrates the bottom-up layer-by-layer approach with the substrate being ﬁrst coated by a reactive monolayer to which a polymer ‘‘cushion’’ binds after adsorption from solution. The ﬁnal monolayer, also pre-organised at the water–air interface contains some reactive ‘‘anchor’’ lipids, able to bind to the tethering polymer. Both methods 3 and 4 yield a polymer-supported monolayer from which the tethered bilayer is obtained by a Langmuir–Scha¨fer transfer of the distal lipid monolayer (method 3) or by vesicle fusion (method 4).

108 elevated temperatures (501C) and (3) single bilayer spreading which is achieved simply by depositing a lipid reservoir from an organic solution onto the solid. Following the addition of water, a single bilayer is spontaneously pulled over the surface by adhesion forces (if it is hydrophilic and sufﬁciently attractive). The bilayer is continuous and self-healing because local pores heal rapidly owing to the strong spreading pressure provided by the lipid reservoir [104]. The polymer-supported bilayer membrane in combination with microelectrophoresis provides a tool to study the migration of molecules within a bilayer in an applied electric ﬁeld and for the local enrichment of charged molecules including membrane proteins in 2D systems. Such local enrichment of receptor proteins using microelectrophoresis not only allows the separation of the membrane proteins in their native form but also leads to signal ampliﬁcation for the detection of speciﬁc binding reactions [105,106]. Tethered bilayer membranes In addition to the polymerically tethered membrane systems, direct coupling lipid molecules modiﬁed with long spacer groups to the solid substrate have also been developed to form tethered bilayer systems. The spacer introduced at the lipid head group link the membrane to the substrate in a mechanically and chemically stable way by establishing covalent bonds between some of the lipid molecules in the proximal monolayer of the membrane and the tethering units and to speciﬁc reactive groups on the modiﬁed substrate surface [117–112]. Chemical stabilisation of the whole complex architecture results in long-term stability and this tethering system leads to the required structural spatial and functional decoupling of membrane and substrate. As a result, a sufﬁcient submembraneous space can serve both as an ionic reservoir as well as provide adequate space for the unperturbed incorporation of bulky proteins. The lipids covalently attached to the solid surfaces through modiﬁed head groups form a monolayer with hydrocarbon chains exposed to the solvent. A thiolipid comprising a hydrophilic oligoethoxy spacer between the lipid head group and the reactive sulphur group has been used to form lipid monolayers on gold surfaces via chemisorption [99]. Bilayers can then be formed by the phospholipid self-assembling processes or by the vesicle fusion approach. Cross-linking the thiolipids onto the titanium oxide surfaces of optical waveguides provides bilayer membranes of similar design [107]. This strategy provides the basis for the more sophisticated design of protein–membrane systems for biosensor development. Others have synthesised lipid disulphide anchors on polymer backbones to mediate the organisation of lipid bilayer formation on solid supports [108]. In a different preparation method, the reaction between vesicles containing DODA-Suc-NHS molecules as anchors and cysteamine-modiﬁed gold surface resulted in the self-assembly of a

109 phospholipid bilayer covalently anchored onto the gold surface [109]. Cholesterol modiﬁed with oligoethoxythiol has also been tethered onto gold surfaces, which provides an anchor for the fusion of lipid vesicle to form a supported bilayer [110]. Effects of the tethered cholesterol density on the kinetics of membrane formation on the gold surfaces were also studied using surface plasmon resonance (SPR) techniques. The resulting tethered bilayers provide long-term stability and the spacer between the surfaces and the head groups results in the elastic properties of the bilayers. Additionally, the water molecules are retained in the hydrophilic spacer region, which serves as a medium to maintain the hydration and ionic properties of the membrane surface. The elastic properties of the spacer region also allow the incorporation of transmembrane proteins in their native functional form. Peptide spacers coupled to the lipid head groups have also been used to form tethered lipid monolayers on gold surfaces. This peptide-based lipid monolayer was used for the bilayer deposition and incorporation of the membrane proteins F0F1–ATPase from chloroplasts and E. coli [111]. The controlled construction of protein–membrane systems has also been reported through the development of tethered membranes. The combination of thiolipid and modiﬁed gramicidin tethered on a gold substrate forms the basis of a biomembrane-based biosensor [112]. Hydrophilic spacer groups were coupled to a fraction of the tethered lipids that forms a reservoir and improves the conductivity of the membrane. A tethered membrane spanning lipid was also incorporated to increase the stability of the membrane. Such an immobilised gramicidin–membrane system has also provided a useful system for the investigation of molecular antigen–antibody recognition processes and peptide–membrane interactions. In an analogous approach, a rhodopsin–transducin complex has been coupled to patterned membranes on a gold surface [113]. The self-assembled carboxyl-exposing thiols were ﬁrst tethered onto the gold surfaces and subsequently removed within the designed patterning using UV lithography methods. The resulting spaces were ﬁlled with thiolipids consisting of a long spacer group coupled to the head group. A micropatterned array of ﬂuid bilayers and incorporation of the rhodopsin receptor was then assembled onto this patterned tethered membrane. This type of patterned supported membrane system offers a novel approach to screen membrane receptor agonists and antagonists. An alternative approach to construct a tethered bilayer membrane, particularly for membrane proteins, is based on the initial immobilisation of proteins onto the surface followed by bilayer deposition. Several protein immobilisation strategies can be used based on high afﬁnity interactions between oligohistidine sequences and nitrilotriacetic acid (NTA), glutathioneS-transferase and glutathione, antibody and antigen, streptavidin and biotin or complementary oligonucleotides. In such membrane protein-tethered bilayer systems, the lipid molecules have a high degree of lateral mobility and ﬂip–ﬂop motion, which provides the environment for protein conformational

110 dynamics necessary for biological activity. In addition, the orientation and the surface concentration of the membrane proteins can be specifically controlled on the surface, which is crucial for the analysis of protein function and applications in biosensing. The generation of a protein-tethered bilayer lipid membrane (ptBLM) involves the chemical modiﬁcation of the support, speciﬁc afﬁnity adsorption of recombinant proteins and reconstitution into the lipid environment by detergent substitution. The speciﬁc functional groups can be either attached at the N- or C-terminus of target proteins. Alternatively, functional groups can be added with speciﬁc enzymes that recognise particular sequences introduced at a speciﬁc region of the protein. ptBLMs have been developed on surfaces modiﬁed with biotinylated bovine serum albumin or NTA [114–116]. For example, a biotinylated neurokinin-1 receptor (a G protein-coupled receptor (GPCR)) was immobilised in a uniform orientation on streptavidincoated quartz sensor surfaces [115]. In another example, the His-tagged membrane protein cytochrome c oxidase was ﬁrst attached to the surface in its detergent solubilised form. In the second step, the detergent molecules were substituted by lipid molecules thus forming a lipid bilayer that is tethered to the support by the protein itself. The coupling His-tag provides sufﬁcient intramolecular ﬂexibility to allow for activity in the reconstituted protein. Above all, it renders the method universally applicable to all His-tagged membrane proteins. Both in situ afﬁnity coupling of proteins and in situ dialysis were combined to form a membrane protein-tethered lipid bilayer [116]. Micro- and nano-arrayed lipid bilayers As the mechanism by which the formation of lipid bilayer membranes on solid surfaces is elucidated, this understanding has evolved into the basis of the technology of patterning membrane arrays and membrane protein microarrays. Microarrays consist of spatially indexed microspots (‘‘probe’’ microspots) comprising immobilised molecules of biological interest. When exposed to a sample of interest (the ‘‘target’’), binding of molecules in the sample to the microspots occurs to an extent determined by the concentration of that molecule and its afﬁnity for a particular probe microspot. In principle, if the target concentrations are known, the afﬁnity of the target for the different probe microspots can be estimated simultaneously. Conversely, in principle, given the known afﬁnities of the different molecules in the target for each probe microspot, the amounts of binding observed at each microspot may be used to simultaneously estimate the concentrations of multiple analytes in the sample. The attractiveness of microarray technology lies in the ability to obtain highly multiplexed information using small amounts of sample. Microarrays containing components of the cell membrane provide an attractive platform for efﬁciently studying fundamental aspects of molecular

111 recognition at the cell surface. These studies are of immediate relevance to diverse applications ranging from drug discovery to the detection of pathogens. The ability to pattern arrays of lipid bilayers has become almost essential for performing multiplexed, high information-content assays. To form a patterned biomembrane microarray, it is essential to control the formation and composition of the supported membrane in a well-deﬁned geometry. The geometric arrangement of barriers to lipid diffusion constrains the membrane and determines its topology. Different lipid compositions and proteins can be deposited or ﬂowed into individual microspots in the array, thus producing a mosaic pattern of ﬂuid membrane each with a different composition. The charged membrane components can be rearranged by introducing an electric ﬁeld across the membrane array. This provides further information about the clustering state of molecules in the membrane and quantitative analysis of transient interaction between membrane components. A variety of techniques have been developed for micropatterning-supported bilayers, four of which are described here and are shown in Fig. 10. (1) Several multistep, indirect methods have been reported that use prepatterned substrate surfaces which present physical, chemical and/or electrostatic barriers to membrane lateral diffusion and formation [117]. Many of these methods circumvent the requirements of wet ambient by employing pre-patterned surfaces. Typically, patterns of barrier materials are deposited onto substrate surfaces using controlled deposition techniques, such as photolithography [118,119], e-beam lithography and microcontact printing [120–122] which modiﬁes the solid substrate. Lipid vesicles readily adsorb and fuse into a continuous supported membrane over the pre-patterned surface, where they do not form a ﬂuid membrane on the barriers. Barrier materials have included metals (Au, Al, Cr, Ti) and metal oxides (Al2O3, TiO2) [123], photoresists [124], proteins [125], polymers [126] and even photopolymerisable lipids [127] but simple mechanical scratches [117] have also proven useful. (2) Soft lithography applications of PDMS stamps for direct patterned deposition/stamping of preformed membrane onto substrate. Alternatively, polymer lift-off [128] or blotting away material [129,130] from the lipid membrane-coated surface have also proven successful in generating a wide range of membrane composition on very high spatial densities (>106 corrals/cm2). (3) Alternative methods using a spatially directed illumination of preformed supported lipid bilayers by short-wavelength ultraviolet radiation [131]. Patterns of hydrophilic voids were formed within a ﬂuid membrane as well as isolated membrane corrals over large substrate areas. These voids can be reﬁlled with secondary lipid vesicles with

112

Fig. 10. Strategies for fabricating supported membrane microarray. Supported

membranes can be partitioned into precisely deﬁned arrays of ﬂuid membrane corrals. Each corral can be ﬁlled with different components as a way of displaying a library of membrane proteins or other molecules in a membrane environment. Such microarrays hold great potential for the study of cell membranes as well as the production of integrated biological-solid-state devices. Method A: photolithographic patterning. A photomask is used to selectively expose each corral in an array with a speciﬁed dose of light. The light drives a chemical transformation: in this case, photobleaching of a ﬂuorescent probe. Diffusive mixing leads to homogeneous compositions within each corral. A range of compositions can be generated in a single photolithographic step. Method B: microprinting is combined with pre-existing ﬁxed barriers on the surface. Variably sized patches of supported membrane

113 different lipid and protein compositions by vesicle fusion from solution, thereby providing a synthetic means for probing 2D reaction–diffusion processes, manipulating membrane compositions, and designing membrane protein arrays generating functional microdomains in well-deﬁned patterns. (4) Using electron microscopy grids to laterally control the extent of plasma oxidation, the substrate can be partitioned into regions of different hydrophilicities [132]. Addition of vesicles then results in supported ﬂuid lipid monolayers, adsorbs intact lipid vesicles or supports ﬂuid bilayers separated by regions that contain no lipid. In all cases, a ﬂuid membrane is partitioned into compartments, which are separated by barriers restricting the lateral diffusion of the lipids. The methods requiring substrate pre-patterning depend on the prior deposition of exogenous materials on the substrate surface and form single, permanent patterns [133]. Moreover, the mechanisms by which barrier materials compartmentalise membrane patterns remain poorly understood. For instance, while aluminium oxide resisted vesicle spreading, deposition of lipids, albeit immobile, was noted for chrome, gold, and indium-tin oxide surfaces. Methods based on PDMS stamps, on the other hand, require optimisation of the contact time and associated contact pressure for different lipid compositions. Arraying membrane proteins has been achieved by printing mixtures of the protein and associated lipids, which warrants appropriate surface chemistry for the immobilisation of lipids. Cell membrane preparations containing GPCRs from a cell line overexpressing the receptor was used directly for fabrication of GPCR microarrays. Multiple arrays of GPCR-containing cell membranes were printed on a g-aminopropylsilane (GAPS)-derivatised surface with rapid kinetics while maintaining the lateral ﬂuidity and significant mechanical stability after multiple withdraws through an air–water interface without lipid desorption [134]. The advent of DNA microarray technology and the development of several printing technologies also greatly facilitates can be deposited into an array of ﬁxed corrals using a PDMS stamp inked with membranes by vesicle fusion directly to the oxidised PDMS surface. The unﬁlled portions of each corral are then ﬁlled with a second membrane, thus leading to a composition array. In this example, various amounts of red membrane are printed, and a blue membrane is used to ﬁll in. Method C: irreversible vesicle fusion. Two or more solutions of vesicles containing different molecules of interest are ﬂowed over the surface of a pre-patterned substrate. Vesicles adsorb irreversibly to the surface, fuse and mix. This gives rise to variable composition arrays (central panel). Fluidity is demonstrated by applying a lateral electric ﬁeld (right panel). The dye-labelled lipids have opposite charges and are separated by application of the electric ﬁeld. (Adapted from Groves and Boxer [133].)

114 the fabrication of membrane microarrays. Quill- and solid pin printing have been used for fabricating GPCR–membrane microarray. The quill-pin printing method offers advantages for its high-quality and low sample usage of 0.5 nL or less per data point from a single insertion of the pin into the GPCR sample to yield several hundred microspots. The functional activity of the GPCRs such as the human neurotensin receptor, adrenergic receptor and dopamine receptor was further demonstrated by the speciﬁc binding to their corresponding ﬂuorescence-labelled cognate ligands. The GPCR microarray can therefore contribute to lead identiﬁcation and validation stages of drug discovery. In addition, the GPCR–G protein complex is largely preserved in the microspot which may provide potential in parallel studies of ligand binding and the downstream activation of G proteins. Nanoparticles covered with supported lipid bilayers which combine the intrinsic properties of metal oxide particles and of lipid bilayers constitute potential interest both in basic and in applied science. By uncoupling the inorganic surface from the surrounding aqueous phase, nanoSLBs provide a natural environment for biomolecules, eliminating or reducing possible problems of non-speciﬁc adsorption or protein denaturation, known to be critical in the development of biomaterials. SLB-coated nanoparticles open numerous possibilities of functionalisation for the development of nanovectors with speciﬁc molecular targets as well as for carrying and delivering molecules to a selected site. In addition, strategies developed for silica nanoparticles are directly applicable to core-shell silica nanoparticles, such as lipid-modiﬁed quantum dots [135,136], extending the range of physical properties that could be used for addressing, activation or detection.

Reconstitution of membrane proteins Oriented lipid membranes deposited on solid substrates offer unique experimental models for biophysical studies on lipid structure and binding of biomolecules, such as peptides, proteins and drugs to lipid membranes. Such membrane models also provide a suitable system for studying the structure and function of membrane proteins, and ligand binding to membrane proteins. Membrane proteins play a major role in every living cell. With the advances in genome mapping it was revealed that 20–30% of the genes of an organism encode for membrane proteins [137]. These proteins are the key factors in the cellular metabolism, for example in cell–cell interaction, immunoregulation, signal transduction and transport of ions and nutrients. Owing to the wide range of critical functions, membrane proteins are preferred targets for most pharmaceuticals, which cover currently 60% of consumed drugs [138] but there are enormous challenges in achieving their isolation in sufﬁcient amount and purity for extensive biophysical studies in model biomembranes systems.

115 Nevertheless, various methods have been developed to incorporate membrane proteins into liposomes and supported membranes. In general, the functional reconstitution of integral membrane proteins into lipid membranes is based upon detergent removal from the protein solution in the presence of a lipid phase. The efﬁciency of reconstitution depends on both the type of detergent and the method of removal. Among the different techniques, dialysis, gel ﬁltration and dilution are the most commonly used for their reconstitution. The reconstitution of membrane proteins into lipid bilayers ﬁrst involves solubilisation of the original membrane with detergent and commonly requires several subsequent separation steps. The ﬁnal pure proteins are obtained in a detergent solution, often mixed with residual lipids. The nature of the detergent in both protein separation and subsequent incorporation is a critical determinant of success. Most frequently, detergents with a high critical micellar concentration (CMC) such as octyl-b-glucoside (OG) have been used because of the ease with which they can be removed by dialysis. The short chain phospholipids diheptanoyl phosphatidylcholine is also used as detergent in the puriﬁcation membrane proteins, which are sensitive to harsh detergent such as OG. Other lysophospholipids in combination with various detergents have also been applied to increase the solubility of membrane proteins. The reconstitution of membrane proteins into bilayers can be achieved by diluting the detergent concentration in the presence of lipids. For example, upon dilution of a mixed micellar suspension containing detergent OG and either eggPC or lipid and the membrane protein porin OmpF, at least two populations of structures with different sizes were formed during the transition from micelles to bilayer as monitored by dynamic light scattering [139]. In the case with only eggPC present, further dilution results in a homogeneous population of vesicles. With both eggPC and OmpF present, micellar aggregation started earlier during dilution. On further dilution, the heterogeneous intermediate structure formed tubes and sheets of variable size as well as vesicles. The micelle–bilayer transition stage is known to be the key to reconstitution. Neutron scattering, dynamic light scattering and cryoelectron microscopy have shown for several lipid–detergent systems that this transition involves the formation of tube-like extended lipid micelles probably capped by detergents that must convert into vesicles on detergent removal. Another important parameter is therefore the lipid/protein ratio, which should be optimised to prevent aggregation. At the start of a typical reconstitution experiment, an excess of detergent ensures a homogeneous distribution of protein and lipid in micelles. As detergent concentration decreased, lipid and protein interact due to the exposure of their hydrophobic surfaces. With an excess of lipid over protein, the protein is mainly incorporated into bilayers, similar to its native state. In an excess of protein over lipid, the protein largely ends up in amorphous aggregates and possibly denatured. Reconstitution is closely linked to the properties of the detergents

116 used during both puriﬁcation and the reconstitution itself (the detergents used for puriﬁcation can be exchanged for a different detergent used for reconstitution). The manner in which the detergent concentration is decreased is an important consideration. The commonly used techniques for detergent removal are dilution [139], dialysis [140] and selective adsorption of the detergent [141]. Diluting a solution of protein, lipid and detergent decreases the concentration of all components by equal factors, until the free detergent concentration drops below saturation. As the protein is significantly diluted during this process, rather high initial concentrations are required. On the other hand, the dilution method allows the process to be arrested when the saturation point is reached, extending the time in which an ordered assembly of the component can take place. Dialysis is the most widely used technique in proteoliposomes, usually in the form of small sample compartments dialysed against large buffer volumes. A temperature-controlled continuous ﬂow dialysis apparatus can be used to improve the reproducibility of the reconstitution process. The major advantage of this system is the precise control of the temperature proﬁle that was found to be critical in some cases. Additionally, a maximal gradient of detergent concentration is maintained across the dialysis membrane, which improves reproducibility. A drawback of the dialysis method is the long dialysis times needed to remove low CMC detergents, making it only practical for medium to high CMC detergents (typically CMC >1 mM). The reconstitution of membrane proteins into liposomes can also be achieved by the selective adsorption of the detergent to hydrophobic substrates of Bio-Beads [141]. This has been employed with success in incorporating the photosystem I complex, cytochrome bc1 and Ca2+–ATPase, the melibose transporter, the cytochrome b6f complex and cytochrome c oxidase into lipid bilayers. Because adsorption of lipid together with detergent is undesirable, the Bio-Beads have been thoroughly characterised with radiolabeled detergents and lipids. It was found that the adsorption of lipids was very limited, so that minimal removal of lipid is achievable with amounts of beads that remove essentially all detergent. The difﬁculty in quantitation of the Bio-Beads is perhaps the biggest drawback of this approach. Of these various reconstitution procedures, detergent dilution is the most common technique with preformed lipid ﬁlms such as the self-assembled solidsupported lipid bilayer. Integrated analysis of functional biomembranes Analysis of the functionality of the model biomembrane systems described in the previous sections is generally performed by the application of speciﬁc surface-based interrogation by spectroscopic means. Binding of organic molecules or proteins to the membrane surface on a solid support changes the physical properties such as the refractive index and the mass at the interface.

117 These changes can be detected by various surface-sensitive techniques: optical detection methods, such as SPR, grating coupler (GCP), resonant mirror (RM), ellipsometry and interferometry [142], acoustic devices such as quartz crystal microbalances [143] and acoustic shear devices, as well as piezoelectric mass-sensing devices [144]. These techniques allow time-resolved monitoring of adsorption and desorption of molecules in aqueous solution to the lipid membranes with detection limits in the range of 1–10 pg/mm2, which corresponds to less than 0.1% of a protein monolayer. Thus, they provide an excellent tool for in situ analysis of the deposition of molecular layers on a surface, as well as for the functional properties of immobilised proteins in terms of ligand binding. Techniques such as IR spectroscopy, contact angle measurements, X-ray and neutron reﬂectometry as well as electrochemical techniques, such as impedance spectroscopy, cyclic voltametry and chronoamperometric studies, have all been applied to characterizing functional tethered membranes [145]. However, for characterizing the binding of single molecules, the lateral resolution as well as sensitivity is not sufﬁcient. While all these techniques have been utilised in the interrogation of functional biomembranes systems, optical techniques together with scanning probe microscopy (SPM) have recently provided the most detailed information on the interrelationship between protein and lipid structure and function and will be reviewed here. Optical biosensors Surface plasmon resonance spectroscopy In the last few years, SPR spectroscopy has become a widely used technique to study antibody–antigen, DNA–DNA, DNA–protein, protein–protein and receptor–ligand interactions [146,147]. More recently, SPR spectroscopy has also been applied to the study of biomembrane-based systems, interactions that involve planar mono- or bilayers or liposomes [148–157] and demonstrate the enormous potential of SPR to enhance our molecular understanding of membrane-mediated events. SPR spectroscopy relies on the SPR phenomenon, which allows the realtime measurement of biomolecules binding to biomimetic surfaces without the application of a speciﬁc label because the SPR method is dependent on the change in adsorbed mass at the sensor surface [146]. SPR is a surfacesensitive technique where the ligand is immobilised onto a solid support and the solute is in solution and the binding event can be readily detected and analysed. A typical SPR system consists of an SPR detector, light source, ﬂow channel and sensor surface, comprising a conducting surface such as gold or silver. P-polarised light is emitted by the light source and reﬂected on the gold-coated sensor surface and detected by the diode array detector (Fig. 11). The SPR phenomenon causes a change in the intensity of reﬂected light at a speciﬁc angle and the SPR detector detects these changes in optical

118

Fig. 11. Schematic representation of the commercial Biacore system and the change in incidence angle during a binding process. (a) When there is no analyte bound to the sensor surface the reﬂective intensity plot shows the default incidence angle indicated by a sharp dip. (b) When the analyte binds to the sensor surface the change in the absorbed mass on the surface causes the dip representing the incident angle to

119 properties at the sensor surface following adsorption and desorption of a solute bound to the sensor surface. The change in optical properties depends on a number of factors including the thickness of the gold surface, the wavelength of the light and most importantly the adsorbed mass on the sensor surface. The SPR technique can also be fully automated using the commercially available instruments and large numbers of samples can be rapidly and conveniently analysed. Typically, ligands are immobilised on the surface of a sensor chip, which is covered by a thin gold layer. When the analyte is injected over the surface in a continuous ﬂow, it adsorbs onto the immobilised ligand and so changes the incidence angle by modifying the refractive index at the surface of the sensor chip. The resulting sensorgram is a plot of the change in SPR incidence angle against time, which allows the binding event between the analyte and the ligand to be visualised and can be used to gain information on the binding kinetics of the interaction (Fig. 11). There are two commercially available chips that are suitable for studying membrane-based systems. The hydrophobic association (HPA) sensor chip consists of self-assembled alkanethiol molecules covalently attached to the gold surface of the chip which can be used to prepare hybrid BLMs by the fusion of liposomes onto the hydrophobic surface [158]. The availability of the HPA sensor chip significantly improved the preparation of solid supported lipid membranes to investigate membrane-mediated interactions. This sensor chip provides a hydrophobic surface on which liposomes fuse to form a hybrid bilayer membrane system similar to that prepared earlier by Plant et al. [159,160] by modiﬁcation of a dextran-coated chip. There have since been a number of examples where this sensor chip has been applied to the study of peptide- and protein–membrane interactions. These applications range from the analysis of protein–protein and protein–ligand interactions in a membrane environment to the study of the direct binding of peptides and proteins to a speciﬁc phospholipid surface [161]. For many applications, the covalent attachment of the hybrid bilayer formed with the HPA chip restricts the insertion of integral peptides and proteins. In order to provide a more appropriate experimental model for the characterisation of these interactions, the vesicle capture (L1) sensor chip was introduced for analysis of model membrane systems by SPR. The L1 sensor chip is composed of a thin dextran matrix modiﬁed by lipophilic compounds on a gold surface, where the lipid bilayer system can be prepared through the shift also shown in (b) as angle 2. The resulting sensorgram monitors the change in angle as shown in (c). (c) A schematic representation of the real-time sensorgram of a binding event. During the association phase, the analyte is present in the buffer ﬂow and binds to the sensor surface. This is followed by the dissociation phase after removal of the analyte from the buffer ﬂow. Analysis of these sections of the curves provides response units (RU).

120 capture of liposomes by the lipophilic compounds as shown in Fig. 11 [150]. The immobilisation of the biomimetic lipid surface onto the sensor chips is generally a fast and reproducible process. Both the HPA and L1 sensor chips can be conveniently applied to the study of membrane-based biomolecular interactions and to measure the binding afﬁnity related to these interactions and an increasing number of examples of the use of this membrane surface in a wide range of biological applications has been reported including analysis of cytolytic peptide action, membrane-mediated cell signalling and neurodegeneration [161–163]. Coupled plasmon-waveguide resonance A variant of SPR has been developed that involves a coupling of plasmon resonances in a thin metal ﬁlm and waveguide modes in a dielectric overcoating. This technique is referred to as coupled plasmon-waveguide resonance (CPWR) spectroscopy which combines characteristic features of both waveguide spectroscopy and SPR spectroscopy (Fig. 12). The CPWR directly measures anisotropies in refractive index and optical absorption coefﬁcient in a dielectric ﬁlm adsorbed onto the surface of the overcoating. Spectroscopic measurements with CPWR devices are based on the resonant excitation of electromagnetic modes of the structure by both transverse magnetic (TM) and transverse electric (TE) polarised components of continuous wave He–Ne laser light within the available incident angles under total internal reﬂection conditions [164–168]. The increased electromagnetic ﬁeld intensities at the dielectric surface enhance the sensitivity and the decreased resonance line widths greatly improve the spectral resolution. The anisotropies in both refractive index and optical absorption coefﬁcient can therefore be determined with high precision and sensitivity. The optical properties of thin-ﬁlm materials, including lipid–protein systems, are characterised by their thickness (t) and by a complex dielectric constant, which includes the refractive index (n) and the extinction coefﬁcient (k) [164–168]. The CPWR phenomenon is a straightforward consequence of the application of thin-ﬁlm electromagnetic theory. The inﬂuence of ﬁlm optical parameters (t, n and k) on the characteristics of the CPWR spectrum can be obtained for any dielectric layer, including lipid and proteolipid membranes. The number of measured CPWR curve parameters (i.e., position, width and depth of the resonance curve) equals the number of unknown optical parameters, and thus the latter can be readily evaluated. The thickness, refractive and extinction coefﬁcient can be uniquely evaluated from the spectra using a non-linear least-squares technique to ﬁt a theoretical resonance curve to an experimental one. In addition, because the refractive index (n) directly reﬂects a mass density (deﬁned as mass per unit volume of deposited material), one can obtain the total deposited mass from the t and n parameters [164–168]. These three optical parameters (t, n and k) can be evaluated for both polarisations, thereby providing the measurement of

121

Fig. 12. Illustration of a CPWR apparatus. The device contains a glass prism coated

with a silver layer overcoated with a SiO2 ﬁlm. A free-suspended lipid bilayer with inserted proteins, such as GPCR, held in place by a Teﬂon spacer via a plateauGibbs border is immobilised on the resonator surface. The prism, the detector and the aqueous compartment of the device are mounted onto a rotating table allowing the angle to be varied from 351 to 701 with 1-millidegree steps.

several structural parameters of the thin ﬁlms including the thickness, the orientation of molecules (by measuring the anisotropy in n) and the orientation of chromophores attached to the molecules within the sensing layer (by measuring the anisotropy of k). This technology allows direct measurements of the anisotropic optical properties of biomembranes and real-time characterisation of changes in the mass density and molecular orientation of molecules contained therein, and thus it can be used to monitor the thermodynamics and kinetics of binding processes and the accompanying structural changes. When applied to membrane-bound receptors, CPWR allows changes in the structure of the receptor, both parallel and perpendicular to the lipid membrane, to be examined in response to the binding of ligand [169–175]. For example, it was found that binding of agonist to the human d-opioid receptor caused an increase in the thickness and molecular packing density of the membrane, while antagonist binding did not show this effect. The results suggested differences in the degree of transmembrane helical reorientation upon ligand binding and provided direct evidence for the differential effects

122 of agonists and antagonists upon membrane structure. These studies clearly provide important information on the structural changes that occur in both the membrane and the receptor during the early stages of a signalling cascade. Dual polarisation interferometry In alternative evanescent wave techniques, a dielectric sensor surface is provided and excitation of bound or partially bound optical modes provides the measurement principle [176]. Unlike the resonance methods such as SPR, optical waveguide interferometers detect the change in optical path length experienced by a light wave passing through the sensing path of the interferometer. Sensitivity is governed, among other things, by the length of the path of interaction and resolution by the signal-to-noise ratio of the detection scheme. The technology of dual polarisation interferometry (DPI) uses dual polarisation excitation of the modes of an optical slab waveguide sensor structure with face-normal incidence end-ﬁre of the modes [177,178] (Fig. 13). Such polarisation switching is readily achieved using a fast liquid crystal switch acting as a switchable half-wave plate and involves no mechanical realignment. In these structures, both TE and TM responses are obtained (one from each polarisation) alternately at a rate of up to 20 kHz and typically at 50 Hz, thus giving data updates at least every 20 ms. The two pieces of information provide enough information to determine the thickness and refractive index (density) of ultra-thin ﬁlms growing at the surface, such as protein adsorption onto a lipid bilayer without any speciﬁc labelled probes. Using classical optical theory, it is possible to interpret the two measurements in terms of thickness and density for the protein layer [179–181]. Using this technique, measurements can be made to subatomic resolutions (typically to 0.01 nm) in protein layer thicknesses in real time [182,183]. More recently, DPI has been applied to the analysis of membrane surfaces [184,185] and provides structural measurements of orientation, insertion and aggregation thereby further extending the degree of molecular insight of these complex systems. It is important to recognise that the DPI measurement embodies a quantitative analytical technique rather than a simple ‘‘sensor’’ response function, providing absolute measurements that can be related directly to the conformation, orientation and kinetics of peptide/protein binding to the biomembranes immobilised in close proximity to the measurement surface. Scanning probe microscopy/atomic force microscopy During the past decade, the atomic force microscope (AFM) has become a key technique in biochemistry and biophysics to characterise supported lipid ﬁlms. The unique capabilities of AFM are: (i) capacity to probe, in real time and in aqueous environment, the surface structure of lipid ﬁlms; (ii) ability to

123

Fig. 13. Schematic of the DPI. The sensor chip comprises ﬁve layers of deposited

silicon oxynitride. A window is opened in the ﬁnal layer to expose the sensing waveguide. This technique allows the quantitative measurement of thickness, refractive index, mass and density of the added biomolecule such as phospholipid bilayers and peptides/proteins on the planar solid support.

directly measure physical properties at high spatial resolution and (iii) possibility to modify the ﬁlm structure and biophysical processes in a controlled way. The common principle of the different adaptations of SPM is the scanning of the sample by a very ﬁne probe (cantilever tip with a radius of curvature below 100 nm and a cantilever length of 100 and 200 mm, respectively) at a very small distance. A vacuum is not required and biological objects can be imaged in their physiological environment. Various physicochemical probes have been adapted for SPM. Atomic or scanning force microscopy is the most widely used technique for analysing the immobilisation of biomolecules at surfaces. Using this technique, the topography of the surface is probed by interaction forces between the ﬁne tip and the molecules at the surface. With optimised scanning modes and tips, even soft matter such as hydrated polymers or proteins can be imaged with a lateral resolution in the nanometer range, i.e., with molecular resolution. Thus, this technique is extremely valuable for analysing the formation, the molecular arrangements, and the

124 destruction of surface architectures [186–190]. Application of AFM in analytical single-molecule research ranges from monitoring individual interactions [191] to conformational dynamic processes [192,193]. Early pioneering work on measurements of the magnitude of the force required to create indentations with deﬁned depth on their surfaces and to separate interacting pairs of avidin–biotin, antigen–antibody and complementary DNA pairs formed the basis of manipulating the cell membrane with AFM. The method has subsequently been applied to map the presence of cell-surface receptors and polysaccharides on live cell membranes by force measurement. Attempts to extract phospholipids and proteins from lipid bilayers and live cell surfaces have been reported, providing a new tool for the manipulation of cellular activities and biochemical analysis at the singlecell level. An increasing awareness of the effect of the pulling speed (nm/s or Mm/s), or more accurately, the force-loading rate (pN/s or nN/s) on the magnitude of the rupture force, has led researchers to construct energy diagrams of rupture events based on the parameters available from such studies. Information on the nature of the interplay of force and loading rate is vital for nanomanipulation of living cells and cell membranes. The AFM has been extensively used not only to image nanometer-sized biological samples but also to measure their mechanical properties by using the force curve mode of the instrument. It is possible to functionalise the AFM tip with speciﬁc ligands so that one can target the adhesive interaction to speciﬁc pairs of ligands and receptors. The presence of speciﬁc receptors on the living cell surface has been mapped by this method. The force to break the co-operative 3D structure of globular proteins or to separate a double stranded DNA into single strands has been measured. The method for harvesting functional molecules from the cytosol or the cell surface for biochemical analysis has been developed and extended to the biochemical nano-analysis based on AFM technology. The study of phospholipid membranes is one of the ﬁelds where the application of AFM provides enormous advantages. Through the preparation of supported membranes, large areas of uniform membrane are achieved in a reproducible way. The deposition of bilayers or multilayers using a LB trough (equipped for this effect) is the method that leads to better quality samples. The AFM studies on planar membranes involve two major ﬁelds: those which focus on the membrane itself and those which deal with the insertion of other molecules in the membrane. Lipid membranes can present a complex behaviour, with phase changes and phase coexistence. They present distinct phospholipid organisations, leading to the existence of several physical parameters that can be resolved by AFM (e.g., bilayer thickness and its spacing, viscosity, formation of bidimensional arrays of phospholipids, dimensions of membrane domains and their shape). On the study of the surface organisation of some molecules, the AFM can reach the resolution of individual phospholipid polar groups and fatty acid salts. However, this resolution has

125 not been obtained yet for the majority of the lipid systems. Other studies developed in this ﬁeld include the imaging of the membrane lesions resultant from the incorporation of an antibiotic [194], degradation of lipid bilayers by phospholipase A2 [195], preparation of planar membranes from high-density lipoproteins [196], domain interdigitation caused by ethanol [197] and phase transition in fatty acid monolayers [198]. More recently, the mechanical properties of membranes have also been studied [199–202] while the combined analysis with AFM and SPR has provided insight into the two-state binding of beta2-glycoprotein with membranes [203]. Future perspectives Efforts to study membrane structure and function have historically been the domain of a few research groups highly specialised in a speciﬁc area of membrane biophysics. However, the emergence of a wide range of model systems together with an increasing number of analytical techniques to probe the structure and function of the model membrane systems now places the study of lipid bilayer membranes within reach of a much larger number of researchers. This has come about partly because of the powerful combination of the ﬁelds of surfaces and interfaces together with membrane biology [204,205]. As a consequence, what has arguably been the significant challenge in structural biology in terms of the structural elucidation of the molecular components of membranes together with their interactions now appears to be more amenable to experimental interrogation. In this review, we have focused specifically on concepts and design criteria that have underpinned the development of the supported membrane surfaces together with a selection of techniques used to analyse those surfaces. What is apparent from these studies is the increasing structural information currently available for both membrane proteins and the membrane bilayer, which is now providing us with unprecedented insight into the exquisitely complex architecture of the cell membrane. As with many areas in biomedical research, this increased knowledge of the fundamental processes is also driving the development of new technologies based on artiﬁcial model membranes such as novel drug-screening tools and receptor-based diagnostics. Clearly, as the deposition and patterning technologies continue to develop, the capacity to create artiﬁcial cells with speciﬁc membrane-mediated cell–cell communication opens up exciting new opportunities for exploring new avenues for target identiﬁcation and drug design and screening. References 1.

Singer SJ and Nicholson GC. The ﬂuid mosaic model of the structure of cell membranes. Science 1972;175:720–731.

126 2. 3. 4. 5. 6.

7. 8. 9. 10.

11.

12. 13.

14.

15. 16. 17. 18.

19. 20. 21.

22.

Kinnunen PKJ. On the principles of functional ordering in biological membranes. Chem Phys Lipids 1991;57:375–399. Welti R and Glaser M. Lipid domains in model and biological membranes. Chem Phys Lipids 1994;73:121–137. Jacobson K, Sheets ED and Simon R. Revisiting the ﬂuid mosaic model of membranes. Science 1995;268:1441–1442. Kagan BL and Sokolov Yu. Use of lipid bilayer membranes to detect pore formation by toxins. Methods Enzymol 1994;235:691–705. Mueller D, Rudin O, Tien HT and Wescott WC. Reconstitution of cell membrane structure in vitro and its transformation into an excitable system. Nature 1962;194: 979–980. Mueller D, Rudin O, Tien HT and Wescott WC. Methods for the formation of single bimolecular lipid membranes in aqueous solution. J Phys Chem 1963;67:534–535. Del Castillo J, Rodriguez A, Romero CA and Sanchez V. Lipid ﬁlms as transducers for detection of antigen–antibody and enzyme–substrate reactions. Science 1966;153:185–188. Coronado R and Lattore R. Detection of K+ and CI channels from calf cardiac sarcolema in planar lipid bilayer membranes. Nature 1982;298:849–852. Sato H, Hakamada H, Yamazaki Y, Uto M, Sugawara M and Umezawa Y. Ionophore incorporated bilayer lipid membranes that selectively respond to metal ions and induce membrane permeability changes. Biosensors Bioelectron 1998;13:1035–1046. Montal M and Mueller P. Formation of biomolecular membranes from lipid monolayers and a study of their electrical properties. Proc Nat Acad Sci USA 1972;69:3561–3566. Vodyanoy V and Murphy RB. Solvent-free lipid bimolecular membranes of large surface area. Biochim Biophys Acta 1982;687:189–194. Schindler H and Rosenbusch JP. Matrix protein from Escherichia coli outer membranes forms voltage-controlled channels in lipid bilayers. Proc Natl Acad Sci USA 1978;75:3751–3755. Sato H, Wakabayashi M, Ito T, Sugawara M and Umezawa Y. Potentiometric responses of ionophore-incorporated bilayer lipid membranes with and without added anionic sites. Anal Sci 1997;13:437–446. Minami H, Sugawara M, Odashima K, Umezawa Y, Uto M, Michaelis EK and Kuwana T. Ion channel sensors for glutamic acid. Anal Chem 1991;63:2787–2795. Robinson RL and Strickholm A. Oxidized cholesterol bilayers. Dependence of electrical properties on degree of oxidation and aging. Biochim Biophys Acta 1978;509:9–20. Yogev D, Todorov AT and Fendler JH. Fluorescence lifetimes of diphenylhexatriene in ﬂat and bent bilayer lipid membranes. J Phys Chem 1991;95:3892–3894. Papahadjopoulos D, Cowden M and Kimelberg HK. Role of cholesterol in membranes. Effects on phospholipid–protein interactions, membrane permeability and enzymatic activity. Biochim Biophys Acta 1973;330:8–26. McIntosh TJ, Magid AD and Simon SA. Cholesterol modiﬁes the short-range repulsive interactions between phosphatidylcholine membranes. Biochemistry 1989;28:17–25. Otto M, Snejdarkova M and Rehak M. Hydrogen peroxide/oxygen biosensor based on supported phospholipid bilayer. Anal Lett 1992;25:653–662. Moellerfeld J, Prass W, Ringsdorf H, Hamazaki M and Sunamoto J. Improved stability of black lipid membranes by coating with polysaccharide derivatives bearing hydrophobic anchor groups. Biochim Biophys Acta 1986;857:265–270. Kotowski J and Tien HT. Sucrose inﬂuence on lecithin and polypyrrole lecithin bilayer membranes. Bioelectrochem Bioenerg 1989;22:69–74.

127 23. Hianik T, Dlugopolsky J, Gyeppessova M, Sivak B, Tien HT and Ottova-Leitmanova A. Stabilization of bilayer lipid membranes on solid supports by trehalose. Bioelectrochem Bioenerg 1996;39:299–302. 24. Kalinowski S, Lotowski Z and Morzycki JW. The inﬂuence of bolaamphiphilic steroid dimer on the formation and structure of bilayer lipid membranes. Cell Mol Biol Lett 2000;5:107–118. 25. Ash PS, Bunce AS, Dawson CR and Hider RC. The effect of synthetic polymers on the electrical and permeability properties of lipid membranes. Biochim Biophys Acta 1978;510:216–229. 26. Benz R, Prass W and Ringsdorf H. Black lipid membranes from polymerizable lipids. Angew Chem 1982;21(Suppl):869–880. 27. Borle F, Sa¨nger M and Sigrist H. Planar bilayer membranes from photoactivable phospholipids. Biochim Biophys Acta 1991;1066:144–150. 28. Redwood WR, Pfeiffer FR, Weisbach JA and Thompson TE. Physical properties of bilayer membranes formed from a synthetic saturated phospholipid in n-decane. Biochim Biophys Acta 1971;233:1–6. 29. Zhao XK, Baral S, Rolandi R and Fendler JH. Semiconductor particles in bilayer lipid membranes. Formation, characterization, and photoelectrochemistry. J Am Chem Soc 1988;110:1012–1024. 30. Tien HT. Cyclic voltammetry of bilayer lipid membranes. J Phys Chem 1984;88: 3172–3174. 31. Krysinski P and Tien HT. Some electrochemical studies on charge transfer of TCNQ modiﬁed BLMs. Bioelectrochem Bioenerg 1988;19:227–233. 32. Yamada H, Shiku H, Matsue T and Uchida I. Electron transfer at a planar bilayer lipid membrane incorporated with 7,7,8,8-tetracyanoquinodimethane studied by a.c. impedance spectroscopy. J Phys Chem 1993;97:9547–9549. 33. Kotowski J, Janas T and Tien HT. Electrical oscillations in polypyrrole–lecithin bilayer lipid membranes. Bioelectrochem Bioenerg 1988;19:283–289. 34. Hianik T and Passechnik VI. Bilayer Lipid Membranes: Structure and Mechanical Properties, Dordrecht, Kluwer Academic Publishers, 1995. 35. Van Wie SRBJ, Sutsina H, Moffett DF, Koch AR, Silber M and Davies WC. Bispecific antibody modiﬁcation of nicotinic acetylcholine receptors for biosensing. Biosensors Bioelectron 1996;11:91–102. 36. Adachi Y, Sugawara M, Taniguchi K and Umazawa Y. Na+, K+–ATPase-based bilayer lipid membrane sensor for adenosine-50 -triphosphate. Anal Chim Acta 1993;281:577–584. 37. Trojanowicz M. Miniaturized biochemical sensing devices based on planar bilayer lipid membranes. Fresenius J Anal Chem 2001;371:246–260. 38. Rolandi R, Flom SR, Dillon I and Fendler JH. Electrical properties of polymerised, planar, bimolecular membranes. Prog Colloid Polymer Sci 1987;73:131–141. 39. Zhao XK and Fendler JH. Simultaneous intracavity laser absorption spectroscopic and electrical measurements of bilayer lipid membranes. J Phys Chem 1986;90:3886–3890. 40. Zhao XK, Picard G and Fendler JH. Holographic interferometry on bilayer lipid membranes. J Phys Chem 1988;92:7161–7163. 41. Picard G, Denicourt N and Fendler JH. Simultaneous electrical and optical interferometric measurements of pressure- and applied-potential-induced bilayer lipid membrane deformation. J Phys Chem 1991;95:3705–3715. 42. Suzuki H, Tabata K, Kato-Yamada Y, Noji H and Takeuchi S. Planar lipid bilayer reconstitution with a micro-ﬂuidic system. Lab Chip 2004;4:502–505.

128 43. Fendler JH. Membrane Mimetic Chemistry, New York, Wiley, 1982. 44. Ulman A. An Introduction to Ultrathin Organic Films from Langmuir–Blodgett to Self-assembly, Boston, Academic Press, 1991. 45. Saﬁnya CR. Biomolecular materials: structure, interactions and higher order self-assembly. Colloids Surf A 1997;128:183–195. 46. Cui DF, Howarth VA, Petty MC, Ancelin H and Yarwood J. The deposition and characterisation of phosphatidic acid Langmuir–Blodgett ﬁlms. Thin Solid Films 1990; 192:291–296. 47. Lotta TI, Laakkonen LJ, Virtanen JA and Kinnunen PKJ. Characterisation of Langmuir–Blodgett ﬁlms of 1,2-dipalmitoyl-sn-glycero-3-phosphatylcholine and 1-palmitoyl-2-[10-(pyren-1-yl)decanoyl]-sn-glycero-3-phosphatidylcholine by FTIR-ATR. Chem Phys Lipids 1988;46:1–12. 48. Fringeli UP and Gunthard HH. Infrared membrane spectroscopy. Mol Biol Biochem Biophys 1981;31:270–332. 49. Brian A and McConnell HM. Allogenic stimulation of cytotoxic T cells by supported planar membranes. Proc Natl Acad Sci USA 1984;81:6159–6163. 50. Tamm LK and McConnell HM. Supported phospholipid bilayers. Biophys J 1985;47: 105–113. 51. Singh S and Keller DJ. Atomic force microscopy of supported planar membrane bilayers. Biophys J 1991;60:1401–1410. 52. Reimhult E, Ho¨o¨k F and Kasemo B. Intact vesicle adsorption and supported biomembrane formation from vesicles in solution: inﬂuence of surface chemistry, vesicle size, temperature, and osmotic pressure. Langmuir 2003;19:1681–1691. 53. Retzinger GS, Meredith SC, Lau SH, Kaiser ET and Kezdy FJ. A method for probing the afﬁnity of peptides for amphiphilic surfaces. Anal Biochem 1985;150:131–140. 54. Bayerl TM and Bloom M. Physical properties of single phospholipid bilayers adsorbed to micro glass beads. A new vesicular model system studied by 2H-nuclear magnetic resonance. Biophys J 1990;58:357–362. 55. Loidl-Stahlhofen A, Eckert A, Hartmann T and Schottner M. Solid-supported lipid membranes as a tool for determination of membrane afﬁnity: high-throughput screening of a physicochemical parameter. J Pharm Sci 2001;90:599–606. 56. Loidl-Stahlhofen A, Kaufmann S, Braunschweig T and Bayerl T. The thermodynamic control of protein binding to lipid bilayers for protein chromatography. Nat Biotechnol 1996;14:999–1002. 57. Schmitz AAP, Schleiff E, Ro¨hring C, Loidl-Stahlhofen A and Verge`res G. Interactions of myristoylated alanine-rich C kinase substrate (MARCKS)-related protein with a novel solid-supported lipid membrane system (TRANSIL). Anal Biochem 1999;268: 343–353. 58. Krausse E, Dathe M, Wieprecht T and Beinert M. Noncovalent immobilized artiﬁcial membrane chromatography, an improved method for describing peptide–lipid bilayer interactions. J Chromatogr A 1999;849:125–133. 59. Puu G, Gustafson I, Artursson E and Ohlsson P-A˚. Retained activities of some membrane proteins in stable lipid bilayers on a solid support. Biosens Bioelectron 1995;10: 463–476. 60. Puu G and Gustafson I. Planar lipid bilayers on solid supports from liposomes – factors of importance for kinetics and stability. Biochim Biophys Acta 1997;1327:149–161. 61. Major M, Prieur E, Tocanne JF, Betbeder D and Sautereau AM. Characterisation and phase behaviour of phospholipid bilayers adsorbed on spherical polysaccharidic nanoparticles. Biochim Biophys Acta 1997;1327:32–40.

129 62. Fabianowski W, Coyle LC, Weber BA, Granata RD, Castner DG, Sadownik A and Regen SL. Spontaneous assembly of phosphatidylcholine monolayers via chemisorption onto gold. Langmuir 1989;5:35–41. 63. Pigeon C and Venkataram UV. Immobilized artiﬁcial membrane chromatography: supports composed of membrane lipids. Anal Biochem 1989;176:36–47. 64. Mozsolits H, Lee TH, Wirth HJ, Perlmutter P and Aguilar MI. The interaction of bioactive peptides with an immobilized phosphatidylcholine monolayer. Biophys J 1999;77:1428–1444. 65. Lee TH, Rivett D, Werkmeister J, Hewish D and Aguilar MI. Lett Pept Sci 1999;5:1. 66. Qiu X, Ong S, Bernal C, Rhee D and Pidgeon C. A general synthetic route for preparing ether phospholipids suitable for immobilization: a phosphotriester approach. J Org Chem 1994;59:537–543. 67. Lee TH, Mozsolits H and Aguilar MI. Measurement of the afﬁnity of melittin for zwitterionic and anionic membranes using immobilized lipid biosensors. J Peptide Res 2001;58:464–476. 68. Pidgeon C, Cai SJ and Bernal C. Mobile phase effects on membrane protein elution during immobilized artiﬁcial membrane chromatography. J Chromatogr A 1996;72: 213–230. 69. Lundqvist A and Lundahl P. Chromatography on cells and biomolecular assemblies. J Chromatogr B 1997;699:209–220. 70. Lundahl P and Yang Q. Liposome chromatography: liposomes immobilized in gel beads as stationary phase for aqueous column chromatography. J Chromatogr 1991;544:283–304. 71. Brekkan E, Lu L and Lundahl P. Properties of immobilized-liposome-chromatographic supports for interaction analysis. J Chromatogr A 1995;711:33–42. 72. Yang Q and Lundahl P. Binding of lysozyme on the surface of entrapped phosphatidylserine, phosphatidylcholine vesicles and an example of high-performance lipid vesicle surface chromatography. J Chromatogr 1990;512:377–386. 73. Yang Q and Lundahl P. Steric immobilisation of liposomes in chromatographic gel beads and incorporation of integral membrane proteins into their lipid bilayers. Anal Biochem 1994;218:210–221. 74. Yang Q, Wallste´n M and Lundahl P. Immobilisation of phospholipid vesicles and protein–lipid vesicles containing red cell membrane proteins on octyl derivatives of large-pore gels. Biochim Biophys Acta 1988;938:243–356. 75. Yang Q, Liu XY, Ajiki SI, Hara M, Lundahl P and Miyake J. Avidin–biotin immobilisation of unilamellar liposomes in gel beads for chromatographic analysis of drug–membrane partitioning. J Chromatogr B 1998;707:131–141. 76. Yoshimoto M, Kuboi R, Yang Q and Miyake J. Immobilised liposome chromatography for studies of protein–membrane interactions and refolding of denatured bovine carbonic anhydrase. J Chromatogr B 1998;712:59–71. 77. Yang Q, Liu XY, Yoshimoto M, Kuboi R and Miyake J. Covalent immobilisation of unilamellar liposomes in gel beads for chromatography. Anal Biochem 1999;268: 354–362. 78. Lundqvist A, Ocklind G, Haneskog L and Lundahl P. Freeze–thaw immobilization of liposomes in chromatographic gel beads: evaluation by confocal microscopy and effects of freezing rate. J Mol Recog 1998;11:52–57. 79. Yang Q, Wallsten M and Lundahl P. Lipid-vesicle-surface chromatography. J Chromatogr 1990;506:379–389. 80. Zhang Y, Zeng CM, Li YM, Hjerten S and Lundahl P. Immobilized liposome chromatography of drugs on capillary continuous beds for model analysis of drug–membrane interactions. J Chromatogr A 1996;749:13–18.

130 81. Rock CO and Snyder F. Rapid puriﬁcation of phospholipase A2 from Crotalus adamanteus venom by afﬁnity chromatography. J Biol Chem 1975;250:6564–6566. 82. Barsukov LI, Dam CW, Bergelson LD, Muzja GI and Wirtz KWA. Afﬁnity chromatography of the phosphatidylcholine exchange protein from bovine liver. Biochim Biophys Acta 1978;513:198–204. 83. Dijkman R, Beiboer SH and Verheij HM. An afﬁnity column for phospholipase A2 based on immobilised acylaminophospholipid analogues. Biochim Biophys Acta 1997; 1347:1–8. 84. Ishiwatari T, Shimizu I and Mitsuishi M. Formation of silicone coated vesicle by sol–gel method. TEM observation. Chem Lett 1996;25:33–34. 85. Yamanaka SA, Charych DH, Loy DA and Sasaki DY. Solid phase immobilization of optically responsive liposomes in sol–gel materials for chemical and biological sensing. Langmuir 1997;13:5049–5053. 86. Nguyen T, McNamara KP and Rosenzweig Z. Optochemical sensing by immobilizing ﬂuorophore-encapsulating liposomes in sol–gel thin ﬁlms. Anal Chim Acta 1999;400: 45–54. 87. Besanger TR, Zhang Y and Brennan JD. Characterization of ﬂuorescent phospholipid liposomes entrapped in sol–gel derived silica. J Phys Chem B 2002;106:10535–10542. 88. Besanger TR and Brennan JD. Ion sensing and inhibition studies using the transmembrane ion channel peptide gramicidin A entrapped in sol–gel-derived silica. Anal Chem 2003;75:1094–1101. 89. Besanger TR, Easwaramoorthy B and Brennan JD. Entrapment of highly active membrane-bound receptors in macroporous sol–gel derived silica. Anal Chem 2004;76:6470–6475. 90. Luo TJ, Soong R, Lan E, Dunn B and Montemagno C. Photo-induced proton gradients and ATP biosynthesis produced by vesicles encapsulated in a silica matrix. Nat Mater 2005;4:220–224. 91. Gill I and Ballesteros A. Encapsulation of biologicals within silicate, siloxane, and hybrid solgel polymers: an efﬁcient and generic approach. J Am Chem Soc 1998;120: 8587–8598. 92. Brook MA, Chen Y, Guo K, Zhang Z and Brennan JD. Sugar-modiﬁed silanes: precursors for silica monoliths. J Mater Chem 2004;14:1469–1479. 93. Stelzle M and Sackmann E. Sensitive detection of protein adsorption to supported bilayers by frequency dependent capacitance measurements and microelectrophoresis. Biochim Biophys Acta 1989;981:135–142. 94. Ku¨hner M, Tampe´ R and Sackmann E. Lipid mono- and bilayer supported on polymer ﬁlms–composite polymer–lipid ﬁlms on solid substrates. Biophys J 1994;67:217–226. 95. Elender G and Sackmann E. Wetting and dewetting of Si/SiO2-wafers by free and lipidmonolayer covered aqueous solutions under controlled humidity. J Phys II 1994;4: 455–479. 96. Ku¨hner M and Sackmann E. Ultrathin hydrated dextran ﬁlms grafted on glass: preparation and characterization of structural, viscous, and elastic properties by quantitative microinterferometry. Langmuir 1996;12:4866–4876. 97. Elender G, Kuhner M and Sackmann E. Functionalisation of Si/SiO2 and glass surfaces with ultrathin dextran ﬁlms and deposition of lipid bilayers. Biosens Bioelectron 1996;11:565–577. 98. Wiegand G, Jaworek T, Wegner G and Sackmann E. Studies of structure and local wetting properties on heterogeneous, micropatterned solid surfaces by microinterferometry. J Colloid Interface Sci 1998;196:299–312.

131 99. Lang H, Duschl C and Vogel H. A new class of thiolipids for the attachment of lipid bilayers on gold surfaces. Langmuir 1994;10:197–210. 100. Sackmann E. Supported membranes: scientiﬁc and practical applications. Science 1996;271:43–48. 101. Wong JY, Majewski J, Seitz M, Park CK, Israelachvili JN and Smith GS. Polymercushioned bilayers. I. A structural study of various preparation methods using neutron reﬂectometry. Biophys J 1999;77:1445–1457. 102. Wong JY, Park CK, Seitz M and Israelachvili JN. Polymer-cushioned bilayers. II. An investigation of interaction forces and fusion using the surface forces apparatus. Biophys J 1999;77:1458–1468. 103. Heiduschka P, Go¨pel W, Beck W, Kraas W, Kienle S and Jung G. Microstructured peptide-functionalised surfaces by electrochemical polymerization. Chem Eur J 1996;2: 667–672. 104. Ra¨dler J, Strey H and Sackmann E. On the phenomenology and kinetics of lipid bilayer spreading on hydrophilic surfaces. Langmuir 1995;11:4539–4548. 105. Stelzle M, Miehlich R and Sackmann E. Two dimensional microelectrophoresis in supported lipid bilayers. Biophys J 1992;63:1346–1354. 106. Dietrich C and Tampe´ R. Charge determination of membrane molecules in polymersupported lipid layers. Biochim Biophys Acta 1995;1238:183–191. 107. Heyse S, Vogel H, Sa¨nger M and Sigrist H. Covalent attachment of functionalised lipid bilayers to planar waveguides for measuring protein binding to biomimetic membranes. Protein Sci 1995;4:2532–2544. 108. Spinke J, Yang J, Wolf H, Liley M, Ringdorf H and Knoll W. Polymer-supported bilayer on a solid substrate. Biophys J 1992;63:1667–1671. 109. Brink G, Schmitt L, Tampe´ R and Sackman E. Self-assembly of covalently anchored phospholipid supported membranes by use of DODA-Suc-NHS-lipids. Biochim Biophys Acta 1994;1196:227–230. 110. Williams LM, Evans SD, Flynn TM, Marsh A, Knowles PF, Bushby RJ and Boden N. Kinetics of the unrolling of small unilamellar phospholipid vesicles onto selfassembled monolayers. Langmuir 1997;13:751–757. 111. Naumann R, Jonczyk A, Kopp R, van Esch J, Ringdorf H, Knoll W and Gra¨ber P. Incorporation of membrane proteins in solid-supported lipid layers. Angew Chem Int Ed Engl 1995;34:2056–2058. 112. Cornell BA, Braach-Maksvytis VLB, King LG, Osman PDJ, Raguse B, Wieczorek L and Pace RJ. A biosensor that use ion-channel switches. Nature 1997;387:580–583. 113. Heyse S, Ernst OP, Dienes Z, Hofmann KP and Vogel H. Incorporation of rhodopsin in laterally structured supported membranes: observation of transducin activation with spatially and time-resolved surface plasmon resonance. Biochemistry 1998;37:507–522. 114. Bieri C, Ernst OP, Heyse S, Hofmann KP and Vogel H. Micropatterned immobilization of a G protein-coupled receptor and direct detection of G protein activation. Nat Biotechnol 1999;17:1105–1108. 115. Martinez KL, Meyer BH, Hovius R, Lundstrom K and Vogel H. Ligand binding to G protein-coupled receptors in tethered cell membranes. Langmuir 2003;19:10925–10929. 116. Giess F, Fiedrich MG, Heberle J, Naumann RL and Knoll W. The protein-tethered lipid bilayer: a novel mimic of the biological membrane. Biophys J 2004;87:3213–3220. 117. Groves JT, Ulman N and Boxer SG. Micropatterning ﬂuid lipid bilayers on solid supports. Science 1997;275:651–653. 118. van Oudenaarden A and Boxer SG. Brownian ratchets: molecular separations in lipid bilayers supported on patterned arrays. Science 1999;285:1046–1048.

132 119. Morigaki K, Baumgart T, Offenhausser A and Knoll W. Patterning solid-supported lipid bilayer membranes by lithographic polymerization of a diacetylene lipid. Angew Chem Int Ed 2001;40:172–174. 120. Jenkins ATA, Bushby RJ, Boden N, Evans SD, Knowles PF, Liu QY, Miles RE and Ogier SD. Ion-selective lipid bilayers tethered to microcontact printed self-assembled monolayers containing cholesterol derivatives. Langmuir 1998;14:4675–4678. 121. Kung L, Groves JT, Ulman N and Boxer SG. Printing via photolithography on micropartitioned ﬂuid lipid membranes. Adv Mater 2000;12:731–734. 122. Srinivasan MP, Ratto TV, Stroeve P and Longo ML. Patterned supported bilayers on self-assembled monolayers: conﬁnement of adjacent mobile bilayers. Langmuir 2001;17:7951–7954. 123. Groves JT, Ulman N, Cremer PS and Boxer SG. Substrate–membrane interactions: mechanisms for imposing patterns on a ﬂuid bilayer membrane. Langmuir 1998;14: 3347–3350. 124. Glasma¨star K, Gold J, Andersson AS, Sutherland DS and Kasemo B. Silicone transfer during microcontact printing. Langmuir 2003;19:5475–5483. 125. Kung LA, Kam L, Hovis JS and Boxer SG. Patterning hybrid surfaces of proteins and supported lipid bilayers. Langmuir 2000;16:6773–6776. 126. Sackmann E and Tanaka M. Supported membranes on soft polymer cushions: fabrication, characterization and applications. Trends Biotechnol 2000;18:58–64. 127. Morigaki K, Baumgart T, Jonas U, Offenhausser A and Knoll W. Photopolymerization of diacetylene lipid bilayers and its application to the construction of micropatterned biomimetic membranes. Langmuir 2002;18:4082–4089. 128. Orth RN, Kameoka J, Zipfel WR, Ilic B, Webb WW, Clark TG and Craighead HG. Creating biological membranes on the micron scale: forming patterned lipid bilayers using a polymer lift-off technique. Biophys J 2003;85:3066–3073. 129. Hovis JS and Boxer SG. Patterned barriers to lateral diffusion in supported lipid bilayer membranes by blotting and stamping. Langmuir 2000;16:894–897. 130. Hovis JS and Boxer SG. Patterning and composition arrays of supported lipid bilayers by microcontact printing. Langmuir 2001;17:3400–3405. 131. Yee CK, Amweg ML and Parikh AN. Direct photochemical patterning and refunctionalization of supported phospholipids bilayers. J Am Chem Soc 2004;126: 13962–13972. 132. Lenz P, Ajo-Franklin CM and Boxer SG. Patterned supported lipid bilayers and monolayers on poly(dimethylsiloxane). Langmuir 2004;20:11092–11099. 133. Groves JT and Boxer SG. Micropattern formation in supported lipid membranes. Acc Chem Res 2002;35:149–157. 134. Fang Y, Frutos AG and Lahiri J. Membrane protein microarrays. J Am Chem Soc 2002;124:2394–2395. 135. Voura EB, Jaiswal JK, Mattoussi H and Simon SM. Tracking early metastatic progression with quantum dots and emission scanning microscopy. Nat Med 2004;10:993–998. 136. Medintzi IL, Uyeda TH, Goldmani ER and Mattoussi H. Quantum dot bioconjugates for imaging, labelling and sensing. Nat Mater 2005;4:435–446. 137. Gerstein M and Hegyi H. Comparing genomes in terms of protein structure: surveys of a ﬁnite parts list. FEMS Microbiol Rev 1998;22:277–304. 138. Hopkins AL and Groom CR. The druggable genome. Nat Rev Drug Discov 2002;1:727–730.

133 139. Dolder M, Engel A and Zulauf M. The micelle to vesicle transition of lipids and detergents in the presence of a membrane protein: towards a rationale for 2D crystallization. FEBS Lett 1996;382:203–208. 140. Jap BK, Zulauf M, Scheybani T, Hefti A, Baumeister W, Aebi U and Engel A. 2D crystallization: from art to science. Ultramicroscopy 1992;46:45–84. 141. Rigaud JL, Mosser G, Lacapere JJ, Olofsson A, Levy D and Ranck JL. Bio-Beads: an efﬁcient strategy for two-dimensional crystallization of membrane proteins. J Struct Biol 1997;118:226–235. 142. Striebel C, Brecht A and Gauglitz G. Characterization of biomembranes by spectral ellipsometry, surface plasmon resonance and interferometry with regard to biosensor application. Biosens Bioelectron 1994;9:139–146. 143. Pignataro B, Steinem C, Galla HJ, Fuchs H and Janshoff A. Speciﬁc adhesion of vesicles monitored by scanning force microscopy and quartz crystal microbalance. Biophys J 2000;78:487–498. 144. Janshoff A, Galla HJ and Steinem C. Piezoelectric mass-sensing devices as biosensors – an alternative to optical biosensors? Angew Chem Int Ed Engl 2000;39:4004–4032. 145. Knoll W, Frank CW, Heibel C, Naumann R, Offenhausser A, Ruhe J, Schmidt EK, Shen WW and Sinner A. Functional tethered lipid bilayers. J Biotechnol 2000;74: 137–158. 146. Cooper MA. Optical biosensors in drug discovery. Nat Rev Drug Discov 2002;1:515–528. 147. Malmqvist M and Karlsson R. Curr Opin Chem Biol 1997;1:378–383. 148. Kuziemko GM, Stroh M and Stevens RC. Biochemistry 1996;35:6375–6384. 149. Heyse S, Stora T, Schmid E, Lakey JH and Vogel H. Biochim Biophys Acta 1998;1376:319–338. 150. Cooper MA, Hansson A, Lofas S and Williams DH. Anal Biochem 2000;277:196–205. 151. Wang W, Smith DK, Moulding K and Chen HM. J Biol Chem 1998;273:27438–27448. 152. Saenko E, Sarafanov A, Ananyeva N, Behre E, Shima M, Schwinn H and Josic D. J Chromatogr A 2001;921:49–56. 153. Mozsolits H, Wirth HJ, Werkmeister J and Aguilar MI. Biochim Biophys Acta 2001;1512:64–76. 154. Ananthanarayanan B, Das S, Rhee SG, Murray D and Cho W. J Biol Chem 2002;277:3568–3575. 155. Danelian E, Karlen A, Karlsson R, Winiwarter S, Hansson A, Lofas S, Lennernas H and Hamalainen MD. J Med Chem 2000;43:2083–2086. 156. Subasinghe S, Unabia S, Barrow CJ, Mok SS, Aguilar MI and Small DH. Cholesterol is necessary both for the toxic effect of Ab peptides on vascular smooth muscle cells and for Ab binding to vascular smooth muscle cell membranes. J Neurochem 2003;84: 471–479. 157. Jin Y, Mozsolits H, Hammer J, Zmuda E, Zhu J, Zhang Y, Aguilar MI and Blazyk J. Inﬂuence of tryptophan on lipid-binding of linear amphipathic cationic antimicrobial peptides. Biochemistry 2003;42:9395–9405. 158. Cooper MA, Try AC, Carroll J, Ellar DJ and Williams DH. Surface plasmon resonance analysis at a supported lipid monolayer. Biochim Biophys Acta 1998;1373:101–111. 159. Plant AL, Brigham-Burke M, Petrella EC and O’Shannessy DJ. Phospholipid/alkanethiol bilayers for cell-surface receptor studies by surface plasmon resonance. Anal Biochem 1995;226:342–348. 160. Hubbard JB, Silin V and Plant AL. Self assembly driven by hydrophobic interactions at alkanethiol monolayers: mechanisms of formation of hybrid bilayer membranes. Biophys Chem 1998;75:163–176.

134 161. Mozsolits H and Aguilar MI. Surface plasmon resonance spectroscopy: an emerging tool for the study of peptide–membrane interactions. Biopolym – Pept Sci 2002;66:3–18. 162. Karlsson OP and Lofas S. Flow-mediated on-surface reconstitution of G-protein coupled receptors for applications in surface plasmon resonance biosensors. Anal Biochem 2002;300:132–138. 163. Mozsolits H, Thomas WG and Aguilar M-I. Surface plasmon resonance spectroscopy in the study of membrane-mediated cell signalling. J Pept Sci 2002;66:3–18. 164. Salamon Z, Macleod HA and Tollin G. Coupled plasmon-waveguide resonators: a new spectroscopic tool for probing proteolipid ﬁlm structure and properties. Biophys J 1997;73:2791–2797. 165. Salamon Z and Tollin G. Graphical analysis of mass and anisotropy changes observed by plasmon-waveguide resonance spectroscopy can provide useful insights into membrane protein function. Biophys J 2004;86:2508–2516. 166. Salamon Z, Cowell S, Varga E, Yamamura HI, Hruby VJ and Tollin G. Plasmon resonance studies of agonist/antagonist binding to the human delta-opioid receptor: new structural insights into receptor–ligand interactions. Biophys J 2000;79:2463–2474. 167. Salamon Z, Macleod HA and Tollin G. Surface plasmon resonance spectroscopy as a tool for investigating the biochemical and biophysical properties of membrane protein systems. I: Theoretical principles. Biochim Biophys Acta 1997;1331:117–129. 168. Salamon Z, Macleod HA and Tollin G. Surface plasmon resonance spectroscopy as a tool for investigating the biochemical and biophysical properties of membrane protein systems. II: Applications to biological systems. Biochim Biophys Acta 1997;1331: 131–152. 169. Alves ID, Salamon Z, Varga E, Yamamura HI, Tollin G and Hruby VJ. Direct observation of G-protein binding to the human delta-opioid receptor using plasmonwaveguide resonance spectroscopy. J Biol Chem 2003;278:48890–48897. 170. Alves ID, Ciano KA, Boguslavski V, Varga E, Salamon Z, Yamamura HI, Hruby VJ and Tollin G. Selectivity, cooperativity, and reciprocity in the interactions between the delta-opioid receptor, its ligands, and G-proteins. J Biol Chem 2004;279:44673–44682. 171. Alves ID, Cowell SM, Salamon Z, Devanathan S, Tollin G and Hruby VJ. Different structural states of the proteolipid membrane are produced by ligand binding to the human delta-opioid receptor as shown by plasmon-waveguide resonance spectroscopy. Mol Pharmacol 2004;65:1248–1257. 172. Salamon Z, Devanathan S, Alves ID and Tollin G. Plasmon-waveguide resonance studies of lateral segregation of lipids and proteins into microdomains (rafts) in solidsupported bilayers. J Biol Chem 2005;280:11175–11184. 173. Alves ID, Salgado GF, Salamon Z, Brown MF, Tollin G and Hruby VJ. Phosphatidylethanolamine enhances rhodopsin photoactivation and transducin binding in a solid supported lipid bilayer as determined using plasmon-waveguide resonance spectroscopy. Biophys J 2005;88:198–210. 174. Alves ID, Salamon Z, Hruby VJ and Tollin G. Ligand modulation of lateral segregation of a G-protein-coupled receptor into lipid microdomains in sphingomyelin/ phosphatidyl-choline solid-supported bilayers. Biochemistry 2005;44:9168–9178. 175. Subramaniam V, Alves ID, Salgado GF, Lau PW, Wysocki Jr RJ, Salamon Z, Tollin G, Hruby VJ, Brown MF and Saavedra SS. Rhodopsin reconstituted into a planarsupported lipid bilayer retains photoactivity after cross-linking polymerization of lipid monomers. J Am Chem Soc 2005;127:5320–5321. 176. Cush R, Cronin JM, Stewart WJ, Maule CH, Molloy J and Goddard NJ. The resonant mirror: a novel optical biosensor for direct sensing of biomolecular interactions: Part I.

135

177. 178.

179.

180.

181.

182.

183.

184.

185.

186. 187.

188. 189. 190.

191. 192. 193.

Principle of operation and associated instrumentation. Biosens Bioelectron 1993;8: 347–353. Cross GH, Ren Y and Freeman NJ. Young’s fringes from vertically integrated slab waveguides: applications to humidity sensing. J Appl Phys 1999;86:6483–6499. Cross GH, Reeves AA, Brand S, Popplewell JF, Peel LL, Swann MJ and Freeman NJ. A new quantitative optical biosensor for protein characterisation. Biosens Bioelectron 2003;19:383–390. Cross GH, Reeves AA, Brand S, Swann MJ, Peel LL, Freeman NJ and Lu JR. The metrics of surface adsorbed small molecules on the Young’s fringe dual-slab waveguide interferometer. J Phys D Appl Phys 2004;37:74–80. Swann MJ, Peel LL, Carrington S and Freeman NJ. Dual-polarization interferometry: an analytical technique to measure changes in protein structure in real time, to determine the stoichiometry of binding events, and to differentiate between speciﬁc and nonspeciﬁc interactions. Anal Biochem 2004;329:190–198. Biehle SJ, Carrozzella J, Shukla R, Popplewell J, Swann M, Freeman N and Clark JF. Apolipoprotein E isoprotein-speciﬁc interactions with tissue plasminogen activator. Biochim Biophys Acta 2004;1689:244–251. Lu JR, Swann MJ, Peel LL and Freeman NJ. Lysozyme adsorption studies at the silica/water interface using dual polarization interferometry. Langmuir 2004;20: 1827–1832. West DC, Rees CG, Duchesne L, Patey SJ, Terry CJ, Turnbull JE, Delehedde M, Heegaard CW, Allain F, Vanpouille C, Ron D and Fernig DG. Interactions of multiple heparin binding growth factors with neuropilin-1 and potentiation of the activity of ﬁbroblast growth factor-2. J Biol Chem 2005;280:13457–13464. Oikawa T, Yamaguchi H, Itoh T, Kato M, Ijuin T, Yamazaki D, Suetsugu S and Takenawa T. PtdIns(3,4,5)P3 binding is necessary for WAVE2-induced formation of lamellipodia. Nat Cell Biol 2004;6:420–426. Popplewell J, Freeman N, Carrington S, Ronan G, McDonnell C and Ford RC. Quantiﬁcation of the effects of melittin on liposome structure. Biochem Soc Trans 2005;33:931–933. Radmacher M, Tillmann RW, Fritz M and Gaub HE. From molecules to cells: imaging soft samples with the atomic force microscope. Science 1992;257:1900–1905. Jass J, Tjarnhage T and Puu G. From liposomes to supported, planar bilayer structures on hydrophilic and hydrophobic surfaces: an atomic force microscopy study. Biophys J 2000;79:3153–3163. Dufreˆne YF and Lee GU. Advances in the characterization of supported lipid ﬁlms with the atomic force microscope. Biochim Biophys Acta 2000;1509:14–41. Shao Z and Yang J. Progress in high resolution atomic force microscopy in biology. Quart Rev Biophys 1995;28:195–251. Wielert-Badt S, Hinterdorfer P, Gruber HJ, Lin JT, Badt D, Wimmer B, Schindler H and Kinne RK. Single molecule recognition of protein binding epitopes in brush border membranes by force microscopy. Biophys J 2002;82:2767–2774. Danker T and Oberleithner H. Nuclear pore function viewed with atomic force microscopy. Pﬂugers Arch 2000;439:671–681. Oesterhelt F, Oesterhelt D, Pfeiffer M, Engel A, Gaub HE and Mu¨ller DJ. Unfolding pathways of individual bacteriorhodopsins. Science 2000;288:143–146. Deleu M, Nott K, Brasseur R, Jacques P, Thonart P and Dufreˆne YF. Imaging mixed lipid monolayers by dynamic atomic force microscopy. Biochim Biophys Acta 2001;1513:55–62.

136 194. Santos NC, Ter-Ovanesyan E, Zasadzinski JA, Prieto M and Castanho MARB. Filipininduced lesions in planar phospholipid bilayers imaged by atomic force microscopy. Biophys J 1998;75:1869–1873. 195. Grandbois M, Clausen-Schaumann H and Gaub HE. Atomic force microscope imaging of phospholipid bilayer degradation by phospholipase A2. Biophys J 1998;74: 2398–2404. 196. Carlson JW, Jonas A and Sligar SG. Imaging and manipulation of high-density lipoproteins. Biophys J 1997;73:1184–1189. 197. Mou J, Yang J, Huang C and Shao Z. Alcohol induces interdigitated domains in unilamellar phosphatidylcholine bilayers. Biochemistry 1994;33:9981–9985. 198. Sikes HD and Schwartz DK. A temperature-dependent two-dimensional condensation transition during Langmuir–Blodgett deposition. Langmuir 1997;13:4704–4709. 199. Santos NC, Ter-Ovanesyan E, Zasadzinski JA and Castanho MARB. Reconstitution of phospholipids bilayer by an atomic force microscope tip. Biophys J 1998;75:2119–2120. 200. Nielsen LK, Bjørnholm T and Mouritsen OG. Fluctuations caught in the act. Nature 2000;404:352. 201. Schneider J, Dufreˆne YF, Barger Jr WR and Lee GU. Atomic force microscope image contrast mechanisms on supported lipid bilayers. Biophys J 2000;79:1107–1118. 202. Oishi Y, Umeda T, Kuramori M and Suehiro K. Mechanical properties of a Langmuir–Blodgett ﬁlm measured by atomic force microscopy. Langmuir 2002;18: 945–947. 203. Gamsjaeger R, Johs A, Gries A, Gruber HJ, Romanin C, Prassl R and Hinterdorfer P. Membrane binding of beta2-glycoprotein I can be described by a two-state reaction model: an atomic force microscopy and surface plasmon resonance study. Biochem J 2005;389:665–673. 204. Tanaka M and Sackmann. Polymer-supported membranes as models of the cell surface. Nature 2005;437:656–663. 205. Engelman DM. Membranes are more mosaic than ﬂuid. Nature 2005;438:578–580.

137

Coencapsulation of hepatocytes and bone marrow cells: In vitro and in vivo studies Zun Chang Liu and Thomas Ming Swi Chang Artificial Cells & Organs Research Center, Faculty of Medicine, McGill University, Montreal, Quebec, Canada H3G 1Y6 Abstract. Bioencapsulation of cells is one of the many areas of artiﬁcial cells being extensively investigated by centers around the world. This includes the bioencapsulation of hepatocytes. A number of methods have been developed to maintain the speciﬁc function and phenotype of the bioencapsulated hepatocytes for in vitro and in vivo applications. These include supplementation of factors in the culture medium; use of appropriate substrates and the co-cultivation of hepatocytes with other type of cells, the so called ‘‘feeder cells’’. These feeder cells can be of liver origin or non-liver origin. We have recently studied the role of bone marrow cells in the maintenance of hepatocytes viability and phenotype by using the coculture of hepatocytes with bone marrow cells (nucleated cells including stem cells), and the coencapsulation of hepatocytes with bone marrow stem cells. This way, the hepatocytes viability and speciﬁc function can be maintained signiﬁcantly longer. In vivo studies of both syngeneic and xenogeneic transplantation show that the hepatocytes viability can be maintained longer when coencapsulated with bone marrow cells. Transplantation of coencapsulated hepatocytes and bone marrow cells enhances the ability of the hepatocytes in correcting congenital hyperbilirubinmia in Gunn rats. Both in vitro and in vivo studies show that bone marrow cells can enhance the viability and phenotype maintenance of hepatocytes. Thus, bone marrow cells play an important role as a new type of feeder cells for bioencapsulated hepatocytes for the cellular therapy of liver diseases. Keywords: hepatocytes; viability maintenance; bone marrow cells; artiﬁcial cells; microencapsulation; bioencapsulation; transplantation; syngeneic; xenogeneic; coculture; coencapsulation; stem cells; Gunn rat; hyperbilirubinemia.

Introduction The preparation and use of ultrathin semipermeable polymer membrane microcapsules as artiﬁcial cells was ﬁrst reported in 1964 [1]. This now includes artiﬁcial cells containing hemoglobin, enzymes, living cells, genetically engineered cells, stem cells, adsorbent, hormones and other biologically active materials [1–4]. Ever since it has been extensively developed for use in different areas of biotechnology and medicine [4]. In particular, one of the original purposes in using them for the immunoprotection of many types of transplanted cells [1–3] has now been extensively developed [4–6]. This review discusses the bioencapsulation of hepatocytes and the coencapsulation of hepatocytes with bone marrow stem cells. Corresponding author:

E-mail: [email protected] (T.M.S. Chang). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12005-0

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

138 Transplantation of bioencapsulated hepatocytes for liver diseases Orthotopic liver transplantation is now an accepted and effective way for the treatment of life-threatening liver diseases. However, the growing disparity between the number of donated liver and the disproportionately large number of patients awaiting transplantation has provided an impetus for developing alternative therapies for the treatment of liver failure [7]. Cellbased therapy is one possible approach that can be carried out using one of the two methods. One is free hepatocyte transplantation and the other is extracorporeal bioartiﬁcial liver support device. Of these two approaches, free hepatocytes transplantation is a less invasive treatment for liver diseases. Free hepatocytes transplantation in animal liver failure shows that it can improve the survival of animals with both chemically and surgically induced acute liver failure [8,9]. However, free hepatocytes transplantation requires immunosuppressive agents, which may cause severe adverse effects. Thus, bioencapsulated hepatocytes transplantation has been studied extensively [10–32]. Transplantation of bioencapsulated hepatocytes prevents the xenogeneic hepatocytes from immunorejection [12]. Intraperitoneal injection of bioencapsulated hepatocytes increases the survival time of galactosamineinduced liver failure rats [11]. Survival rates of 90% hepatectomized rats can be improved by the intraperitoneal transplantation of encapsulated hepatocyte spheroids [30]. Gunn rats have a congenital dridine diphosphate glucuronyltransferase (UDPGT) deﬁciency, which is necessary for the conjugation of bilirubin to bilirubin diglucuronide [32]. After birth, the blood bilirubin levels increase and stay at an abnormally high level. In both congeneic and xenogeneic transplantation, free hepatocytes transplantation cannot lower the bilirubin levels [24]. However, transplantation of bioencapsulated hepatocytes into hyperbilirubinemia Gunn rats, lowers blood bilirubin levels [15,24,25]. Although animal studies show promises for actual clinical uses in liver failure patients, it will be important to improve the viability of hepatocytes. Approaches to prolong free hepatocytes viability and phenotype in vitro Adult rat hepatocytes can be easily prepared by liver perfusion method. Isolated mature hepatocytes are well differentiated and have very poor proliferation ability once being isolated. Conventional culture conditions are sufﬁcient for supporting hepatocytes viability for up to 1–2 weeks [33]. However, free hepatocytes are not stable in vitro or in vivo. Researchers have considered several approaches to prolong the viability. The ﬁrst is the selective supplementation of the medium with various factors, such as hormones and growth factors. Thus, primary human hepatocytes cultured in keratinocyte-stimulating factor medium resulted in prolonged culture of primary human hepatocytes with preservation of hepatocytes differentiation,

139 some cells were maintained for more than 56 days [34]. The second is the use of appropriate substrates, such as collagens and ﬁbronectins, liver connective biomatrix [34], however, with these substrates, as is the case with plastic, hepatocytes exhibited phenotypic changes leading to partial or complete loss of highly differentiated functions [34–40]. The third is to cocultivate hepatocytes with other types of cells, the latter is called ‘‘feeder’’ cells, most commonly used feeder cells are epithelial- and ﬁbroblast-like cells [41]. In a study with human hepatocytes, the hepatocytes could survive 2–3 weeks in a coculture with a rat liver epithelial cell [42]. Human hepatocytes cocultured with human biliary epithelial cells can restore the synthetic and metabolic liver function, the metabolic potential of ammonia detoxiﬁcation to urea, cytochrome P450-dependent lignocaine conversion into mono-ethyl-glycinexylidide (MEGX), and protein expression and secretion [35,36,43–49]. Coculture of hepatocytes with ﬁbroblasts improves the hepatocytes function and life span [47,50,51]. In addition, coculture of rat heaptocytes with sinusoidal cells (endothelium and Kupffer cells) permitted cultures to survive longer with greater retention of the ultrastructural markers distinctive for all of these cell types, heaptocytes, endothelium and Kupffer cells [39]. The use of thawed rat hepatocytes cocultured with 3T3-J2 murine ﬁbroblasts could retain their synthetic and detoxiﬁcation functions on a long-term basis [34,42,52–54]. Table 1 shows the cell types used in coculture with hepatocytes. Cell coculture was ﬁrst described by Puck and Marcus in 1955 in the studies with HeLa cells [45]. One of the most successful methods for maintaining functional primary cultures is to use irradiated or mitomycin-Ctreated ‘‘feeder layers’’ of another cell type (often ﬁbroblasts), which are thought to ‘‘feed’’ the primary cells with essential nutrients or other factors. The studies of bone marrow cells (BMs, bone marrow nucleated cells)as a new type of feeder cells in coculture with hepatocytes were ﬁrst initiated a few years ago [60]. Rat hepatocytes and BMs were cultured in two models. One model was to mix the hepatocytes and BMs in one plate well, the other was to coculture the hepatocytes and BMs in a plate well but these two kinds of cells were separated by a well insert chamber, the chamber membrane pore size was 1 mm. This way there would be no direct cell–cell contact between the BMs and the hepatocytes, however the medium and factors could pass through freely between the two compartments. In both the coculture models,

Table 1. Some cell types employed as feeder cells cocultured with hepatocytes. (1) (2) (3) (4) (5)

Liver epithelial cells [34,42,52–54] Kupffer cells [55–59] Bone marrow cells [60–62] Fibroblast [39,63–66] Other cell types

140 the hepatocytes viability were maintained as long as 21 days, while the hepatocytes culture alone could only be maintained for 7–10 days (Fig. 1). The precise mechanism of longer term hepatocytes phenotype maintenance in the coculture systems has not been elucidated yet. This may involve the cell–cell direct interactions between hepatocytes and feeder cells, and the soluble factors from the feeder cells. The cell–cell contact interaction is important in the maintenance of hepatoctyes [37]. In vivo, the primary cell–cell relationship for hepatocytes is with the endothelium as all hepatocytes are bound to a single layer of endothelium which lines sinusoids and is the primary barrier between hepatocytes and blood. This cellular corporation appeared to be tissue speciﬁc, since the hepatic epithelial cells could improve hepatocytes survival in coculture [37]. The use of lung ﬁbroblasts [51] or C3H/10T1/2 mouse embryonic cells [35] as feeder cells did not signiﬁcantly improve cell survival or quantitative maintenance of speciﬁc hepatic function. BMs as feeder cells enhance the hepatocytes maintenance. BMs has various types of cells, including stromal cells and blood cells, and their respective progenitor cells or stem cells. Whether the stromal cell types or the blood cell

Fig. 1. Hepatocytes coculture with bone marrow cells or cultured alone. (A) He-

patocytes cocultured with BMs at day 14, most hepatocytes remained attached, and showing normal growth. (B) Hepatocytes cultured alone at day 14, most cells detached and died [60].

141 types or both play key roles needs to be further investigated. In the twochamber culture system, the BMs and hepatocytes did not contact directly, the hypothesis for the effect of bone marrow cells is that the soluble secretary factors from the BMs must also contribute to this outcome. Coencapsulated BMs and hepatocytes in culture and transplantation In the coculture models, free BMs could improve the viability of free hepatocytes. We have carried out studies to see whether this phenomenon still exists when the hepatocytes and BMs are coencapsulated inside artiﬁcial cells.Brieﬂy, the procedure we have used is as follows. One ml of 20 106 /ml hepatocytes is mixed with 1 ml of 10 106 /ml BMs, then the 2 ml of mixed cell suspension is added to 2 ml of 4% stock sodium alginate. The suspension is extruded through a droplet generating apparatus with controlled rate of airﬂow. The rest of the procedures follow that of the preparation of alginate–polysine–alginate membrane microcapsules (APA microcapsules) [60]. The microcapsules containing hepatocytes and BMs are cultured using the mixed medium of IMDM and William’s E. The viability of hepatocytes when coencapsulated with BMs is signiﬁcantly higher at day 28 of culture than that of bioencapsulated hepatocytes [60] (Fig. 2A). This viability is longer than in the free cells coculture system (28 days vs. 14 days). Previous study shows that encapsulated hepatocytes retain the HGFlike factors inside the microcapsules and this helps to improve the growth of hepatocytes in the microcapsules [17]. Thus in the coencapsulation culture, in addition to the cell–cell interactions of hepatocytes and BMs, the retained HGF improves the hepatocytes viability maintenance even further.

Fig. 2. Hepatocytes viability when coencapsulated with BMs in vivo and in vitro. (A) Hepatocytes (H) viability in culture of coencapsulated hepatocytes with BMs. (B) hepatocytes viability in the transplantation of coencapsulated hepatocytes and BMs. Both in vitro and in vivo results showed improved hepatocytes viability when coencapsulated with bone marrow cells [60,61].

142

hepatocytes viability (%)

Viability of recovered Hepatocytes 100

Encap H Encap H+BM

80 60 40 20 0 0

10

20

30

40 50 Time (day)

60

70

80

90

Fig. 3. Xenotransplantation of coencapsulated rat BMs and hepatocytes into mouse.

There were signiﬁcantly difference of hepatocytes viability between the two groups from week 3 to week 11 post-transplantation. At week 11, the hepatocytes viability was only 10% in single hepatocytes encapsulaton transplantation (black line), whereas in the coencapusltate hepatocytes with BMs (grey lines), the hepatocytes viability was 26% (Po0.04).

Coencapsulated hepatocytes and BMs from Wistar rats are transplanted into normal Wistar rats, syngeneic transplantation. The implanted microcapsules are explanted at different time intervals and thereby the hepatocytes viability determined. In the coencapsulated cells group, the hepatocytes viability from the explanted microcapsules maintains at a high level and for a longer time. The microcapsules can be recovered 4 months after transplantation; at this time point the hepatocytes still maintain a high-level viability. From week 7 after transplantation, the coencapsulated hepatocytes viability is signiﬁcantly higher than that of hepatocytes encapsulation alone [61] (Fig. 2B). For xenotransplantation, coencapsulated hepatocytes and BMs from Wistar rats are transplanted into the male CD-1 Swiss mice. The hepatocytes viability from the explanted microcapsules also showed higher levels throughout the study period (Fig. 3). Thus, like in the in vitro studies, both the syngeneic and xenogeneic transplantation show that coencapsulated hepatocytes and BMs improves the hepatocytes viability maintenance after transplantation.

Speciﬁc function maintenance of coencapsulated hepatocytes with BMs The coencapsulation of hepatocytes and BMs improves the viability of hepatocytes both in culture and in normal rat transplantation [62], further studies has been carried out to investigate the maintenance of the hepatocytes speciﬁc liver function. The ammonia removal ability and the bilirubin conjugation in vivo are determined (Fig. 4).

143

Fig. 4. Hepatocytes-speciﬁc function tests. The comicroencapsulated hepatocytes and bone marrow cells were cultured in ammonia supplemented media, or transplanted into hyperbilirubinemia Gunn rats.

Fig. 5. The ammonia removal test in culture. From day 12, the ammonia levels in

coencapsulated hepatocytes and bone marrow cells culture (encap H+BM) were signiﬁcantly lower than those in the single hepatocytes encapsulation culture (encap H).

The coencapsulated cells are cultured in the mixed medium of IMDM and William’s E supplemented with different concentration of ammonium chloride solution. During the culture, the ammonium chloride concentration is much lower from week 2 culture in the coencapsulation group than in the single hepatocytes encapsulation group (Fig. 5) [62]. Transplantation of coencapsulated hepatocytes and BMs into Gunn rats lowers the bilirubin levels signiﬁcantly more in the conencapsulated cells transplantated rats than those in the single hepatocytes encapsulation group from week 3 to week 8 post-transplantation, but in the ﬁrst 2 weeks

144

Fig. 6. Plasma total bilirubin levels in Gunn rats of transplanted with encapsulated

hepatocytes decreased from week 1, and gradually increased from week 3 posttransplantation, in the coencapsulated cells transplanted Gunn rats, the bilirubin decreased immediately after transplantation, and maintained signiﬁcantly lower levels than those in the encapsulated hepatocytes transplantation from day 21 to day 56.

post-transplantation, there were no signiﬁcant difference in these two groups. After 10 weeks post-transplantation, the bilirubin levels in the coencapsulation group increased and reached the same levels as in the single hepatocytes encapsulation group. The bilirubin levels in the encapsulated BMs did not differ from those of the control group [62], i.e., the single BMs encapsulation transplantation did not have the effect of lowering blood bilirubin in Gunn rats (Fig. 6). Novel cell encapsulation techniques The widely used encapsulation method is one-step alginate-poly L-lysinealginate (APA) encapsulation method, the cell-alginate mixture drops into CaCl2 solution and forms the alginate beads, which are further processed in poly L-lysine and sodium citrate, ﬁnally yield the APA membrane. The microcapsules are spherical in diameter around 100–600 mm. Sometimes this one-step method has drawbacks; the microcapsules prepared by this method were often found some cells entrapped in the capsular membrane (Fig. 7) [61], this results in membrane defects. After transplantation, the defected membrane caused the host immunorejection and inﬁltration by leukocytes and macrophages [67]. To reduce the cell entrapping in the membrane, one can reduce the cell concentration and the number of cells when making cellalginate beads, but sometimes this may cause the lack of enough cells in the microcapsules and lack of cell–cell attachment, which is critical for the survival of isolated cells. To solve this problem, Wong and Chang designed a two-step encapsulation method [16]. The principle of this novel encapsulation

145

Fig. 7. Hepatocytes microcapsules made by single-step and two-step encapsulation

methods. The imperfections of microcapsule membrane by single-step method and the uniform membrane by two-step method were clearly observed [61].

146 method is that the alginate gel beads (small beads) containing cells were mixed with alginate and passed through the droplet generator again to produce large beads, and after the large beads were allowed to cure in the calcium solution, the beads were immersed in poly L-lysine and 0.2% sodium alginate, ﬁnally these large beads were placed in sodium citrate to dissolve the small gel beads inside. Microcapsules made by this two-step method did not show cells entrapped in the outer membrane (Fig. 7). When the microcapsules made by two-step method were transplanted, the capsular membrane kept intact without host leukocytes and microphages inﬁltration [61]. In the previous xenotransplantation study, the encapsulated rat hepatocytes, using the novel two-step method, was more effective on preventing immunorejection when transplanted into mice [16]. Other novel encapsulation method has been developed in a modiﬁed encapsulation method, the microcapsules consist of an inner core of modiﬁed collagen and an outer shell of terpolymer of methyl methacrylate, methacrylate and hydroxyethyl methacrylate, the hepatocytes encapsulated in these microcapsules demonstrated enhanced functions [68]. Microcapsule membrane prepared using alginate and polymethylene-co-guanidine (PMCG) presented increased stability compared to classical Ca2+/alginate microcapsules [69]. In addition to the APA microencapsulation, the commercial hollow ﬁber system is also a choice for encapsulation [70]. Issues to be solved prior to the clinical application Although a vast numbers of lab studies on the microencapsulated hepatocytes have been carried out and demonstrated promising results, clinical trials were very rare. Before this approach goes to clinical use, there are still some critical issues need to be solved. Careful selection of cell lines is most important; the ideal cells should survive long enough to provide biological function. The reported maximum duration of hepatocytes maintenance is months; this is long enough for use in acute liver failure, but not enough for the chronic liver diseases such as cirrhosis. While using xenotransplants there should be no pathological agents transmitted during transplantation, for example, when using porcine hepatocytes transplanted into human, the porcine endogenous retrovirus (PERV) may endanger the procedure [71]. In addition, for xenotransplantation, the ethical concern may arise [72]. Another important issue is to develop high biocompatible polymeric membrane, with sufﬁcient durability and appropriate permeability. Puriﬁcation of the alginate used in APA capsular membrane not only reduces the total amount of impurities, but also avoids antibody response when microcapsules are transplanted. For a particular cell encapsulation and clinical application, the encapsulation procedure should be standardized among the research centers, in order that the results could be evaluated and compared.

147 Conclusions Artiﬁcial cell technique is a practical and ideal approach for cellular therapy, especially for allogeneic and xenotransplantation. Bioencapsulated heptocytes transplantation discussed in this review shows promising application for this approach. The long-term hepatocytes viability maintenance is a key factor in this procedure. Coculture of hepatocytes and BMs prolongs the hepatocytes viability in vitro, and maintains the liver speciﬁc function longer. When coencapsulated hepatocytes with BMs are transplanted, the hepatocytes viability is maintained longer, and could lower the blood bilirubin levels in the hyperbilirubinemia Gunn rats. This new method of prolonging hepatocytes viability maintenance provides a new approach in the cellular therapy of liver diseases. The liver is a complex organ with multiple functions; the main functions of detoxiﬁcation, metabolism and synthesis must work sufﬁciently. So in liver failure, the liver support treatment will provide all these liver functions. Acknowledgments TMSC gratefully acknowledges the operating grant support of the Canadian Institute of Health Research and the Virage Centre of Excellence in Biotechnology award. References 1. 2.

Chang TM. Semipermeable microcapsules. Science 1964;146:524–525. Chang TMS and Mason SG. Semipermeable aqueous microcapsules: I. preparation and properties. Can J Physiol Pharmacol 1966;44:115–128. 3. Chang TMS. Artiﬁcial Cells, Monograph, Sprinﬁeld, Charles C Thomas Publisher, IL, 1972 (out of print, but available free on www.artcell.mcgill.ca). 4. Chang TM. Therapeutic applications of polymeric artiﬁcial cells. Nat Rev Drug Discov 2005;4:221–235. 5. Lim F and Sun AM. Microencapsulated islets as bioartiﬁcial endocrine pancreas. Science 1980;210:908–910. 6. Orive G, Hernandez RM, Gascon AR, Calaﬁore R, Chang TM, De Vos P, Hortelano G, Hunkeler I. Lacik D, Shapiro AM and Pedraz JL. Cell encapsulation: promise and progress. Nat Med 2003;9:104–107. 7. Fox IJ and Chowdhury JR. Hepatocyte transplantation. Am J Transplant 2004;4(6 Suppl):7–13. 8. Gupta S and Chowdhary JR. Hepatocyte transplantation: back to the future. Hepatology 1992;15:156–162. 9. Makowka L, Rotstein LE, Falk RE, Falk JA, Langer B, Nossal NA, Blendis LM and Phillips MJ. Reversal of toxic and anoxic induced hepatic failure by syngeneic, allogeneic, and xenogeneic hepatocyte transplantation. Surgery 1980;88:244–253. 10. Chang TM. Bioencapsulated hepatocytes for experimental liver support. J Hepatol 2001;34:148–149.

148 11. Wong H and Chang TM. Bioartiﬁcial liver: implanted artiﬁcial cells microencapsulated living hepatocytes increases survival of liver failure rats. Int J Artif Organs 1986;9:335–336. 12. Bruni S and Chang TM. Hepatocytes immobilised by microencapsulation in artiﬁcial cells: effects on hyperbilirubinemia in Gunn rats. Biomater Artif Cells Artif Organs 1989;17:403–411. 13. Dixit V and Chang TM. Brain edema and the blood brain barrier in galactosamineinduced fulminant hepatic failure rats. An animal model for evaluation of liver support systems. ASAIO Trans 1990;36:21–27. 14. Dixit V, Darvasi R, Arthur M, Brezina M, Lewin K and Gitnick G. Restoration of liver function in Gunn rats without immunosuppression using transplanted microencapsulated hepatocytes. Hepatology 1990;12:1342–1349. 15. Bruni S and Chang TM. Encapsulated hepatocytes for controlling hyperbilirubinemia in Gunn rats. Int J Artif Organs 1991;14:239–241. 16. Wong H and Chang TM. A novel two step procedure for immobilizing living cells in microcapsules for improving xenograft survival. Biomater Artif Cells Immobil Biotechnol 1991;19:687–697. 17. Kashani SA and Chang TM. Effects of hepatic stimulatory factor released from free or microencapsulated hepatocytes on galactosamine induced fulminant hepatic failure animal model. Biomater Artif Cells Immobil Biotechnol 1991;19:565–577. 18. Ito Y and Chang TM. In vitro study of multicellular hepatocyte spheroids formed in microcapsules. Artif Organs 1992;16:422–427. 19. Dixit V, Darvasi R, Arthur M, Lewin K and Gitnick G. Cryopreserved microencapsulated hepatocytes – transplantation studies in Gunn rats. Transplantation 1993;55:616–622. 20. Dixit V and Gitnick G. Transplantation of microencapsulated hepatocytes for liver function replacement. J Biomater Sci Polym Ed 1995;7:343–357. 21. Balladur P, Crema E, Honiger J, Calmus Y, Baudrimont M, Delelo R, Capeau J and Nordlinger B. Transplantation of allogeneic hepatocytes without immunosuppression: long-term survival. Surgery 1995;117:189–194. 22. Wells GD, Fisher MM and Sefton MV. Microencapsulation of viable hepatocytes in HEMA-MMA microcapsules: a preliminary study. Biomaterials 1993;14:615–620. 23. Honiger J, Balladur P, Mariani Calmus P, Vaubourdolle YM, Delelo R, Capeau J and Nordlinger B. Permeability and biocompatibility of a new hydrogel used for encapsulation of hepatocytes. Biomaterials 1995;16:753–759. 24. Gomez N, Balladur P, Calmus Y, Baudrimont M, Honiger J, Delelo R, Myara A, Crema E, Trivin F, Capeau J and Nordlinger B. Evidence for survival and metabolic activity of encapsulated xenogeneic hepatocytes transplanted without immunosuppression in Gunn rats. Transplantation 1997;63:1718–1723. 25. Bruni S and Chang TM. Comparison of bilirubin conjugation in encapsulated hepatocytes, hepatocyte homogenate and intact hepatocytes. Artif Cells Blood Substit Immobil Biotechnol 1999;27:357–365. 26. Benoist S, Sarkis R, Chafai N, Barbu V, Honiger J, Lakehal F, Becquemont L, Baudrimont M, Capeau J, Housset C and Nordlinger B. Survival and differentiation of porcine hepatocytes encapsulated by semiautomatic device and allotransplanted in large number without immunosuppression. J Hepatol 2001;35:208–216. 27. Ambrosino G, Varotto S, M Basso S, Cecchetto A, Carraro P, Naso A, De Silvestro G, Plebani M, Abatangelo G, Donato D, Cestrone A, Giron G and D’Amico DF. Hepatocyte transplantation in the treatment of acute liver failure: microencapsulated hepatocytes versus hepatocytes attached to an autologous biomatrix. Cell Transplant 2003;12:43–49.

149 28. Mai G, Huy NT, Morel P, Mei J, Bosco D, Berney T, Majno P, Mentha G, Trono D and Buhler LH. Treatment of fulminant liver failure by transplantation of microencapsulated primary or immortalized xenogeneic hepatocytes. Transplant Proc 2005;37:527–529. 29. Gao Y, Xu J, Sun B and Jiang HC. Microencapsulated hepatocytes and islets as in vivo bioartiﬁcial liver support system. World J Gastroenterol 2004;10:2067–2071. 30. Hamazaki K, Doi Y and Koide N. Microencapsulated multicellular spheroid of rat hepatocytes transplanted intraperitoneally after 90% hepatectomy. Hepatogastroenterology 2002;49:1514–1516. 31. Aoki T, Jin Z, Nishino N, Kato H, Shimizu Y, Niiya T, Murai N, Enami Y, Mitamura K, Koizumi T, Yasuda Y. Izumida D, Avital I, Umehara Y, Demetriou AA, Rozga J and Kusano M. Intrasplenic transplantation of encapsulated hepatocytes decreases mortality and improves liver functions in fulminant hepatic failure from 90% partial hepatectomy in rats. Transplantation 2005;79:783–790. 32. Haque T, Chen H, Ouyang W, Martoni C, Lawuyi B, Urbanska AM and Prakash S. In vitro study of alginate-chitosan microcapsules: an alternative to liver cell transplants for the treatment of liver failure. Biotechnol Lett 2005;27:317–322. 33. Katsura N, Ikai I, Mitaka T, Shiotani T, Yamanokuchi S, Sugimoto S, Kanazawa A, Terajima H, Mochizuki Y and Yamaoka Y. Long-term culture of primary human hepatocytes with preservation of proliferative capacity and differentiated functions. J Surg Res 2002;106:115–123. 34. Guguen-Guillouzo C, Clement B, Baffet G, Beaumont C, Morel-Chany E, Glaise D and Guillouzo A. Maintenance and reversibility of active albumin secretion by adult rat hepatocytes co-cultured with another liver epithelial cell type. Exp Cell Res 1983;143:47–54. 35. Langenbach R, Malick L, Tompa A, Kuszynski C, Freed H and Huberman E. Maintenance of adult rat hepatocytes on C3H/10T1/2 cells. Cancer Res 1979;39:3509–3514. 36. Michalopoulos G, Russell F and Biles C. Primary cultures of hepatocytes on human ﬁbroblasts. In Vitro 1979;15:796–806. 37. Miyoshi H, Yanagik H, Furukwa K and Ohshima N. Continuous culture of hepatocytes under different medium conditions using a packed bed reactor loaded with porus resins. Jpn J Artif Organs 1993;22:147–152. 38. Sakai y and Suzuki m. Functional expressions by hepatocytes spheroids entrapped in collagen gel in hollow ﬁber modules. Jpn J Artif Organs 1994:473–478. 39. Sugimachi K, Sosef MN, Baust JM, Fowler A, Tompkins RG and Toner M. Long-term function of cryopreserved rat hepatocytes in a coculture system. Cell Transplant 2004;13:187–195. 40. Taguchi K, Matsushita M, Takahashi M and Uchino J. Development of a bioartiﬁcial liver with sandwiched-cultured hepatocytes between two collagen gel layers. Artif Organs 1996;20:178–185. 41. Clement B, Guguen-Guillouzo C, Campion JP, Glaise D, Bourel M and Guillouzo A. Long-term co-cultures of adult human hepatocytes with rat liver epithelial cells: modulation of albumin secretion and accumulation of extracellular material. Hepatology 1984;4:373–380. 42. Auth MK, Woitaschek D, Beste M, Schreiter T, Kim HS, Oppermann E, Joplin RE, Baumann U, Hilgard P, Nadalin S, Markus BH and Blaheta RA. Preservation of the synthetic and metabolic capacity of isolated human hepatocytes by coculture with human biliary epithelial cells. Liver Transpl 2005;11:410–419. 43. Jurima-Romet M, Casley WL, Neu JM and Huang HS. Induction of CYP3A and associated terfenadine N-dealkylation in rat hepatocytes cocultured with 3T3 cells. Cell Biol Toxicol 1995;11:313–327.

150 44. Malick LE and Langenbach R. Interaction of adult rat liver cells and C3H/10T1/2 mouse embryo ﬁbroblasts in co-culture. Scan Electron Microsc 1979;12:153–160. 45. Puck TT and Marcus PI. A rapid method for viable cell titration and clone production with Hela cells in tissue culture: the use of X-irradiated cells to supply conditioning factors. Proc Natl Acad Sci USA 1955;41:432–437. 46. Reid LM and Jefferson DM. Culturing hepatocytes and other differentiated cells. Hepatology 1984;4:548–559. 47. Ries K, Krause P, Solsbacher M, Schwartz P, Unthan-Fechner K, Christ B, Markus PM and Probst I. Elevated expression of hormone-regulated rat hepatocyte functions in a new serum-free hepatocyte-stromal cell coculture model. In Vitro Cell Dev Biol Anim 2000;36:502–512. 48. Strom S and Michalopoulos G. Mutagenesis and DNA binding of benzo(a)pyrene in cocultures of rat hepatocytes and human ﬁbroblasts. Cancer Res 1982;42:4519–4524. 49. Tilles AW, Baskaran H, Roy P, Yarmush ML and Toner M. Effects of oxygenation and ﬂow on the viability and function of rat hepatocytes cocultured in a microchannel ﬂatplate bioreactor. Biotechnol Bioeng 2001;73:379–389. 50. Kaihara S, Kim S, Kim BS, Mooney DJ, Tanaka K and Vacanti JP. Survival and function of rat hepatocytes cocultured with nonparenchymal cells or sinusoidal endothelial cells on biodegradable polymers under ﬂow conditions. J Pediatr Surg 2000;35:1287–1290. 51. Wanson JC and Mosselmans R. Coculture of adult rat hepatocytes and sinusoidal cells: a new experimental model for the study of ultrastructural and functional properties of liver cells. In: Communications of Liver Cells, Popper H (ed), MTP press, Lancaster, 1980, pp. 229–239. 52. Auth MK, Joplin RE, Okamoto M, Ishida Y, McMaster P, Neuberger JM, Blaheta RA, Voit T and Strain AJ. Morphogenesis of primary human biliary epithelial cells: induction in high-density culture or by coculture with autologous human hepatocytes. Hepatology 2001;33:519–529. 53. Fourel I, Gripon P, Hantz O, Cova L, Lambert V, Jacquet C, Watanabe K, Fox J, Guillouzo C and Trepo C. Prolonged duck hepatitis B virus replication in duck hepatocytes cocultivated with rat epithelial cells: a useful system for antiviral testing. Hepatology 1989;10:186–191. 54. Lerche C, Fautrel A, Shaw PM, Glaise D, Ballet F, Guillouzo A and Corcos L. Regulation of the major detoxication functions by phenobarbital and 3-methylcholanthrene in co-cultures of rat hepatocytes and liver epithelial cells. Eur J Biochem 1997;244:98–106. 55. Billiar TR, Lysz TW, Curran RD, Bentz BG, Machiedo GW and Simmons RL. Hepatocyte modulation of Kupffer cell prostaglandin E2 production in vitro. J Leukoc Biol 1990;47:305–311. 56. Billiar TR, Maddaus MA, West MA, Dunn DL and Simmons RL. The role of intestinal ﬂora on the interactions between nonparenchymal cells and hepatocytes in coculture. J Surg Res 1988;44:397–403. 57. Laurens M, Defamie V, Scozzari G, Schmid-Alliana A, Gugenheim J and Crenesse D. Hypoxia-reoxygenation-induced chemokine transcription is not prevented by preconditioning or intermittent hypoxia in mice hepatocytes. Transpl Int 2005;18:444–452. 58. Matsuo R, Ukida M, Nishikawa Y, Omori N and Tsuji T. The role of Kupffer cells in complement activation in D-Galactosamine/lipopolysaccharide-induced hepatic injury of rats. Acta Med Okayama 1992;46:345–354.

151 59. Rikiyama T, Suzuki M, Unno M, Fukuhara K, Uchiyama T and Matsuno S. Kupffer cell-mediated inhibition of liver regeneration after combined hepatectomy and pancreatectomy. J Gastrointest Surg 1999;3:654–661. 60. Liu Z and Chang TM. Effects of bone marrow cells on hepatocytes: when co-cultured or co-encapsulated together. Artif Cells Blood Substit Immobil Biotechnol 2000;28:365–374. 61. Liu ZC and Chang TM. Increased viability of transplanted hepatocytes when hepatocytes are co-encapsulated with bone marrow stem cells using a novel method. Artif Cells Blood Substit Immobil Biotechnol 2002;30:99–112. 62. Liu ZC and Chang TM. Coencapsulation of hepatocytes and bone marrow stem cells: in vitro conversion of ammonia and in vivo lowering of bilirubin in hyperbilirubemia Gunn rats. Int J Artif Organs 2003;26:491–497. 63. Azzaroli F, Mehal W, Soroka CJ, Wang L, Lee J, Crispe N and Boyer JL. Ursodeoxycholic acid diminishes Fas-ligand-induced apoptosis in mouse hepatocytes. Hepatology 2002;36:49–54. 64. Bhatia SN, Balis UJ, Yarmush ML and Toner M. Microfabrication of hepatocyte/ ﬁbroblast co-cultures: role of homotypic cell interactions. Biotechnol Prog 1998;14:378–387. 65. Khetani SR, Szulgit G, Del Rio JA, Barlow C and Bhatia SN. Exploring interactions between rat hepatocytes and nonparenchymal cells using gene expression proﬁling. Hepatology 2004;40:545–554. 66. Schlosser SF, Azzaroli F, Dao T, Hingorani R, Nicholas Crispe I and Boyer JL. Induction of murine hepatocyte death by membrane-bound CD95 (Fas/APO-1)-ligand: characterization of an in vitro system. Hepatology 2000;32:779–785. 67. Wong H and Chang TM. The microencapsulation of cells within alginate poly L-lysine microcapsules prepared with the standard single step drop technique: histologically identiﬁed membrane imperfections and the associated graft rejection. Biomater Artif Cells Immobil Biotechnol 1991;19:675–686. 68. Yin C, Mien Chia S, Hoon Quek C, Yu H, Zhuo RX, Leong KW and Mao HQ. Microcapsules with improved mechanical stability for hepatocyte culture. Biomaterials 2003;24:1771–1780. 69. Orive G, Hernandez AR, Gascon Igartua M and Pedraz JL. Development and optimisation of alginate–PMCG–alginate microcapsules for cell immobilization. Int J Pharm 2003;259:57–68. 70. Yang MB, Vacanti JP and .Ingber DE. Hollow ﬁbers for hepatocyte encapsulation and transplantation: studies of survival and function in rats. Cell Transplant 1994;3:373–385. 71. Martin U, Kiessig V, Blusch JH, Haverich A, von der Helm K, Herden T and Steinhoff G. Expression of pig endogenous retrovirus by primary porcine endothelial cells and infection of human cells. Lancet 1998;352:692–694. 72. Murphy FA. The public health risk of animal organ and tissue transplantation into humans. Science 1996;273:746–747.

153

Inhibitors of signal transduction protein kinases as targets for cancer therapy Theresa Mikalsen, Nancy Gerits and Ugo Moens Department of Microbiology and Virology, Institute of Medical Biology, University of Tromsø, N-9037 Tromsø, Norway Abstract. Cancer development requires that tumour cells attain several capabilities, including increased replicative potentials, anchorage and growth-factor independency, evasion of apoptosis, angiogenesis and metastasis. Many of these processes involve the actions of protein kinases, which have emerged as key regulators of all aspects of neoplasia. Perturbed protein kinase activity is repeatedly found to be associated with human malignancies, making these proteins attractive targets for anti-cancer therapy. The last decade has witnessed an exponential increase in the development of speciﬁc small protein kinase inhibitors. Many of them are in clinical trials in patients with different types of cancer and some are successfully used in clinic. This review describes different approaches that are currently applied to develop such speciﬁc protein kinase inhibitors and provides an overview of protein kinase inhibitors that are currently in clinical trials or are administered in the clinic. Focus is directed on inhibitors against receptor tyrosine kinases and protein kinases participating in the signalling cascades. Keywords: anaplastic lymphoma kinase; C-Abl; clinical trials; EGFR; FLT-3; integrin-linked kinase; C-Kit; MAP kinase; MTOR; non-receptor tyrosine kinase; PDGFR; PKC; VEGFR.

Introduction Post-translational modiﬁcations are important processes in the regulation of protein activity. One of the major reversible protein modiﬁcations is phosphorylation as it is estimated that 30% of all the proteins in the cell are transiently phosphorylated. The human genome project has led to the identiﬁcation of 520 protein kinases, underscoring the importance of phosphorylation. The phosphorylation pattern of proteins within a cell is determined by the antagonistic action of protein kinases that phosphorylate proteins and protein phosphatases that dephosphorylate proteins [1]. Protein kinases can covalently link a phosphate group from the ATP donor molecule to serine, threonine and/or tyrosine residues on their substrates. This phosphotransfering reaction requires the presence of three speciﬁc sites within the protein kinase: an ATP-binding site, a domain catalysing the transfer of a phosphate group of ATP, and a substrate-binding site in the protein kinase. Of these protein kinases, 90 are tyrosine kinases, 43 are tyrosine kinase-like, while the remaining majority comprises serine/ threonine kinases. Tyrosine kinases are divided into receptor tyrosine kinases (RTKs) and non-receptor tyrosine kinases (non-RTKs). The former are Corresponding author: Tel: +47-77644622. Fax: +47-77645350.

E-mail: [email protected] (U. Moens). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12006-2

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

154 transmembrane proteins that contain a ligand-binding extracellular domain and an intracellular catalytic domain, whereas non-RTKs are intracellular proteins that generally function downstream of the RTK. The intrinsic tyrosine kinase activity of RTK is activated upon binding of the ligand and subsequent oligomerization of the receptor. The intracellular catalytic domain of the activated receptor autophosphorylates tyrosine residues. These phosphotyrosine residues form binding sites for proteins containing src-homology 2 (SH-2) domains. The SH-2 containing proteins further transmit the signal often via non-RTK or serine/threonine protein kinases. The cascade of phosphorylation events results in ampliﬁcation and intracellular transmission of the signal [2,3]. Protein kinases possess a wide variety of substrates including structural proteins, metabolic enzymes, protein kinases controlling the cell cycle and transcription factors. Hence, signal transduction pathways play a pivotal role in the regulation of fundamental cellular processes such as metabolism, proliferation, differentiation, survival, migration and angiogenesis. The activity state of these proteins determines the fate of the cell and aberrant expression and activities of these functional classes of enzymes result in abnormal signal transmission. Perturbed signalling transduction provokes dysregulation of processes involved in angiogenesis, apoptosis, cell migration and cell cycle control and can therefore lead to a malignant phenotype. As such, protein kinases have merged as key regulators of all aspects of neoplasia, including proliferation, invasion, angiogenesis and metastasis, hence making cancer fundamentally a disease of aberrant protein kinase activity and signal transmission. In fact, perturbed activity of more than 50% of the RTK and of several serine/ threonine protein kinases have been repeatedly found to be associated with human malignancies (Table S2 in [4], [5]). Dysregulation of normal protein kinase activity may occur by genomic rearrangements, including chromosomal translocations that generate hybrid proteins containing the catalytic domain of a certain protein kinase and an unrelated protein. A well-known example is the fusion of the tyrosine kinase c-Abl to the breakpoint cluster region (BCR) protein in chronic myelogenic leukaemia (CML). A second important mechanism that disrupts the normal function of a protein kinase is a mutation that renders the kinase constitutive active, as exempliﬁed by mutated c-Kit in gastrointestinal tumours. A third mechanism involves increased or aberrant expression of the protein kinase or the ligand for RTK ampliﬁcation such as the RTK HER-2 ( ¼ ErbB2 ¼ EGFR2) in breast cancer or aberrant expression of vascular endothelial growth factor (VEGF) and its receptor (VEGFR), which enhances angiogenesis and tumour growth. Finally, deregulation of kinase activity by activation of oncogenes or loss of tumour suppressor functions can also contribute to tumourigenesis. For example, mutation in the GTPase Ras leads to deregulation of c-Raf, while loss of the protein phosphatase PTEN activates the serine/ threonine kinase AKT [6–8].

155 Conventional management strategies in cancer therapy have been relying on surgery, radiation and chemotherapy. However, since the underlying molecular basis of many cancers lies in the dysfunction of protein kinases, these proteins have become attractive targets for novel anti-cancer drug design. Here, we review different therapeutic approaches in the design of speciﬁc inhibitors against protein kinases that are involved in cancer. We highlight recently developed drugs that target protein kinases engaged in signal transduction, while inhibitors against cyclin-dependent kinases are beyond the scope of this review. Furthermore, we provide an overview of the status of protein kinase inhibitors that have entered the clinic or that are in clinical trials and we also focus on novel promising inhibitors. The protein kinase inhibitors such as Imatinib (STI571/Gleevec), Cetuximab (C225/Erbitux), Erlotinib (OSI-774/Tarceva), Geﬁtinib (ZD1839/Iressa) and Trastuzumab (Herceptin) that have entered the clinic have been extensively and excellently reviewed elsewhere and will only be brieﬂy discussed here. Protein kinase inhibitors and therapeutic strategies In order to carry out its function, a protein kinase has to bind ATP and its substrate before the catalytic activity can transfer the phosphate group. Preventing ATP binding, substrate binding or inhibiting the kinase activity or a combination may therefore form the basis for drug design, as this will block the function of the protein kinase. Interdisciplinary research has tremendously increased our knowledge on the structure, function and regulation of protein kinases. Although the 3D structure of less than 10% of the human protein kinases is known so far, X-ray structure information has facilitated the prediction and comparison of the binding sites of ATP- and substrate-related molecules to protein kinases [9]. However, the development of protein kinase drugs as anti-cancer therapy was originally hampered by the assumption that these proteins were no suitable targets. The reason for this assumption was that the high intracellular ATP concentrations would compete with the ATP-binding site inhibitors so that no ATP site-directed inhibition was obtained. Moreover, the common catalytic mechanism of the kinase domain made it unlikely to develop speciﬁc inhibitors, because blocking the activity of these kinases prevented additional physiological processes unrelated to proliferation, which resulted in severe side effects [7]. ATP-mimics In order to develop chemical compounds that exert an inhibitory effect on protein kinases, chemical libraries that consist of hundreds of thousands of synthetic compounds are screened. The composition of the library can be

156 based on the actual structure of the ATP-binding pocket of the protein kinase or a family member (structure-based library design) or on the structure of compounds already known to bind to the ATP pocket (ligand-based library design) [10]. In this way, ATP-mimics have been identiﬁed and modiﬁed to increase their potential to compete with ATP binding at the ATP-binding site of the protein kinase, thereby inhibiting the kinase activity of the protein. The degree of conservation of the ATP-binding sites in the distinct protein kinases is not absolute, so it is possible to develop ATPmimics with relatively high selectivity. Besides high selectivity, protein kinase inhibitors must bind with an extremely high afﬁnity. The afﬁnity of the inhibitor for the kinase should be several orders of magnitude higher than that of ATP, because the inhibitor will be present in micromolar–nanomolar range concentrations, whereas the intracellular concentration of ATP is millimolar [10,11]. Recently, a new approach that uses ‘‘selectivity ﬁlters’’ to increase the speciﬁcity of the drug has been described in the literature. This method selectively targets two determinants in the ATP-binding pocket. One is the socalled ‘‘gatekeeper’’, which is the residue that ﬂanks a highly variable hydrophobic pocket at the rear of the ATP-binding site. The other site is a reactive cysteine within the active site. Of the 491 protein kinases, only 19 had cysteine residues within the conserved glycine loop, a region that forms close contacts with bound ATP. By combining a drug that occupied both the hydrophobic pocket and irreversibly alkylated the cysteine residue, a drug that selectively inhibited the p90 ribosomal S6 kinases RSK1 and RSK2 was designed. Moreover, this dual targeting drug may reduce the risk of drugresistant mutations in the kinases, as two mutations must occur in order to induce resistance [12,13]. Substrate mimics Another approach to interfere with the function of protein kinases is the use of substrate-related molecules that compete with the genuine substrate to become phosphorylated. The rational behind the use of this class of inhibitors was that phosphorylation of the substrate mimics, but not of the actual substrate will terminate the signalling transduction event that contributes to the neoplastic state of the cells. These inhibitors are likely to be less toxic than ATP-mimics since they bind to regions in the kinase domain that are less conserved than the ATP-binding site and therefore less likely to hit many other targets (very stringent speciﬁcity). The substrate mimic ThymectacinTM (NB1011) that targets thymidylate synthase, an enzyme overexpressed in tumours, has entered clinical trials. This illustrates that substrate mimicry, as a means to inhibit enzymatic activity, possesses potential clinical applications [14]. However, substrate competitive inhibitors targeting protein kinases await to enter clinical trials.

157 Monoclonal antibodies against protein kinases Monoclonal antibodies against protein kinases or in the case of RTK against their ligand have been developed and can be used to shut-off receptormediated signalling as a means of blocking (tumour) cell growth. The complexity of the immunoglobulin molecule, its murine origin and the selection of appropriate target structures on the target antigen initially were obstacles for the production and the utilization. The replacement of most of the mouse sequences with equivalent human sequences has helped to turn monoclonal antibodies into valuable tools in cancer therapy [15,16]. Another class of monoclonal antibodies consists of bispeciﬁc antibodies that bind an RTK and immunologic effector cell. Examples are MDX-447 and H22-EGF, which will be discussed below. The major disadvantage of therapeutic monoclonal antibodies is their production costs, leading inevitably to expensive medication. To circumvent this, researchers have developed nanobodies, which both reduces production costs and increases speciﬁcity (see section on future perspectives). Anti-sense RNA and RNA interference Since anti-cancer drugs are dependent on a time and dose schedule, cancer cells might soon develop resistance towards treatment, thereby rendering the small molecule protein kinase inhibitors ineffective. Therefore, other therapeutic strategies have been considered. One of them uses anti-sense RNA (asRNA) or RNA interference (siRNA/RNAi) to prevent translation of proteins implicated in cancer. The advantage of siRNA molecules is their high speciﬁcity, i.e., siRNA can be used to target unique mRNA transcripts that are only present in cancer cells, sometimes only differing from the wildtype transcript by a single point mutation. This may circumvent unwanted side effects in healthy cells because the siRNA will not affect the expression of the healthy gene ([17] and references therein). Since cancer cells are characterized by chromosomal rearrangements, deletions and/or alternative splicing of pre-mRNA, they can produce unique mRNA (see e.g., Table 1 in [17]). These cancer-speciﬁc aberrant transcripts can be used as a target for highly speciﬁc hybridization with a long asRNA. For example, persistently active truncated epidermal growth factor receptor (EGFR) (A2-7 EGFR) is expressed in advanced glioblastoma multiform. Long dsRNA (>30 bp), which only binds A2-7 EGFR but not wild-type EGFR, will activate the dsRNA-dependent protein kinase PKR, which subsequently induces cell death. Lentiviral vector-mediated delivery and expression of 39 nucleotides asRNA against EGFR in glioblastoma xenograft strongly inhibited the growth of the tumour in the mouse brain [18]. Several problems that may occur with RNA interference can be imagined. The asRNA may function as a classical antisense, i.e., it will inhibit the expression

158 of the target gene, but fail to induce activation of PKR, while in other tumours, the activity of PKR is impaired. Another pitfall could be that secondary structures in long dsRNA prevent its binding to target sequences. One drawback with siRNA is incomplete inhibition of expression of the target. Another disadvantage is that, when the siRNA is directed against a region of the target mRNA that does not include the mutation, the inhibitory effect will also occur in healthy cells, leading to undesirable side effects. siRNA vectors from which expression is induced speciﬁcally in cancer cells, e.g., by a promoter that is activated by hypoxia or glucose deprivation or by localized irradiation or drug application can solve this problem. Although it was originally assumed that the substrate speciﬁcity of individual siRNAs was very high, recent studies have indicated that siRNA can tolerate single mutations in the centre of the molecule, and up to four mutations are necessary for complete inactivation. Non-speciﬁcity of the siRNA may cause difﬁculties in knocking down the mutated transcript in the neoplastic cell. As protein kinases often carry single point mutations, it may be difﬁcult to design siRNA that speciﬁcally reduced the expression of the mutated protein kinase in tumour cells and does not affect the wild-type transcripts in normal cells. Probably, relatively high concentrations and continuous administration of siRNA must be given over extended periods to be beneﬁcial for the patients. At concentration of 100 nM, siRNA non-speciﬁcally induced the expression of a signiﬁcant number of genes, many of them involved in apoptosis and stress response. Reduction to 20 nM eliminated this non-speciﬁc response. In addition, siRNA can bind to cellular proteins and induce non-speciﬁc changes in gene expression. Several siRNA sequences have been shown to activate the innate immunity response genes in freshly isolated human monocytes, resulting in the production of TNFa and activation of the NFKB and mitogen-activated protein kinase (MAPK) pathways. Delivery forms another challenge in the use of siRNA in clinical treatment. Exogenous delivery by liposome-based transfection or viral-based vectors may give unwanted side effects [17,19]. Bisubstrate analogue inhibitor Bisubstrate inhibitors compete for both ATP and substrate binding simultaneously. The rational for bisubstrate inhibitors is promising, but their use in clinics is hampered by the difﬁculty of delivery. Transforming these peptides into cell-permeable molecules is absolutely required to make them successful therapeutic drugs [6]. Other strategies to inhibit protein kinases Other approaches have been tested in the search for strategies to counteract the kinase activity of protein kinases. One method relies on inhibiting the expression of the regulatory subunits of certain protein kinases. The cyclic

159 AMP (cAMP)-dependent protein kinase or protein kinase A (PKA) is a heterotetramer that consists of two regulatory (R) and two catalytic subunits (C). The regulatory subunits contain two binding sites for the secondary messenger cAMP, while the catalytic subunits include the kinase domain. Both the R(RIa, RIb, RIIa, RIIb) and the C subunits (Ca, Cb, Cg) are differentially expressed. The R2C2 holoenzyme will dissociate into the R and C subunits upon binding of cAMP, resulting in activation of PKA [20]. Perturbed PKA activity has been implicated in tumours. For example, PKAI and/or its regulatory subunit RI are generally overexpressed in human cancer cell lines and in primary tumours and inactivating mutations of RIa are involved in Carney complex (CNC), a multiple endocrine neoplasia syndrome [21–23]. The RIa subunit forms an attractive target in human cancers with aberrant PKA activity. Phase I studies with the anti-sense oligonucleotide GEMs231 against RIa mRNA alone or combined with docetaxel in patients with solid tumours showed that GEMs231 could be safely used, but no objective responses were noted. Stable disease (up to 16 weeks) was observed in few patients. On the basis of the results, additional clinical trials and phase II trials have been recommended [24–26]. An alternative strategy relies on different conformations that activated and inactive protein kinase can acquire. Antagonists for protein kinase inhibitors can be selected that exclusively bind to the inactive form of the kinase, so as to sequester the molecule in a state that cannot participate in signal transduction. Certain classes of BCR–ABL inhibitors bind differentially to kinases in active and inactive states [7]. Another way to restrain protein kinases is by altering their stability. Heat shock protein 90 (Hsp90) functions as a molecular chaperone by binding to various cellular proteins and regulates the folding, stability and function of the target protein. Bona ﬁde Hsp90 target proteins include protein kinases. Thus, inhibiting the normal function of Hsp90 may obstruct the proliferation of cancer cells [27]. Encouraging in vitro results with the histone deacetylase inhibitor LBH589 and an analogue of geldanamycin (17-allyl-amino-demethoxy geldanamycin or 17-AAG) support this assumption. Both LBH589 and 17-AAG disrupt the chaperone association of Hsp90 with its target proteins BCR–ABL (see section on c-Abl) and mutant FLT-3 (see section on FLT-3), resulting in polyubiquitination and proteasomal degradation of both proteins. The combination of LBH589 and 17-AAG provoked more apoptosis of imatinib mesylate-resistant primary CML in blast crisis cells expressing BCR–ABL and acute myeloid leukaemia cells with an activated mutant of FLT-3 than either compound alone. These data should encourage the use of both agents in clinical trials [28]. Phosphotyrosines on RTK function as docking sites for other proteins that will transmit the signal received by the RTK. Small molecules that inhibit this binding and prevent downstream cell signalling can also be used as therapeutic application. Finally, research is aimed towards the development of

160 inhibitors that prevent dimerization of RTK or that interfere with appropriate subcellular localization of protein kinases. However, none of these strategies have entered clinical trials so far [6]. Clinical trials and problems facing the use of protein kinase inhibitors Before a drug can enter the clinic, it has to be thoroughly tested. Preclinical tests include biochemical analyses in vitro, studies in cell cultures of cancer cell lines and intact animals such as xenograft studies or animal models for cancers and ﬁnally clinical trials. In vitro analysis, cell culture and most of the xenograft and/or syngeneic systems used for drug screening suffer from a poor predictability with respect to the history of the molecular pathophysiology of human malignancy. In that regard, transgenic and knockout animal models may provide more realistic approaches before clinical trials are initiated [6]. Clinical trials are usually divided into three phases. In phase I clinical trials, the new drug or treatment is tested in a small group of people (20–80) for the ﬁrst time to evaluate its safety, determine a safe dosage range and identify side effects. These studies also determine how the compound is absorbed, distributed, metabolized and excreted, as well as the duration of its action. Drugs that are used in phase I clinical trials have not been used in humans before and thus adverse reactions speciﬁc to humans are possible. That is why phase I trials are conducted stepwise starting at relatively low doses. After each dose step the trial temporarily closes and a toxicity evaluation is performed. The study only continues at a higher dose level if no serious drug-related toxicities have occurred. In phase II clinical trials, the studied drug or treatment is given to a larger group of people to see if it is effective. Phase II trials further evaluate the safety and effectiveness of an agent or intervention, and evaluate how it affects the human body. Phase II studies usually focus on a particular type of cancer, and include fewer than 100 patients. In phase III trials, the drug or treatment is given to large groups of people to conﬁrm its effectiveness, monitor side effects, compare it to commonly used treatments and collect information that will allow the drug or treatment to be used safely. These phase III trials compare a new agent or intervention (or new use of a standard one) with the current standard therapy (a currently accepted and widely used treatment for a certain type of cancer, based on the results of past research). Participants are randomly assigned to the standard group or the new group, usually by a computer. This method, called randomization, helps to avoid bias and ensures that human choices or other factors do not affect the study’s results. In most cases, 10 studies move into phase III testing only after they have shown promise in phases I and II. Phase III trials may include hundreds of people across the country. Although promising results with speciﬁc protein kinase inhibitors have been obtained in clinical trials (see section on Status of speciﬁc protein kinase inhibitors in clinical trials or in the clinic), several problems are facing the use

161 of these drugs. One obstacle is when to apply the drug during the treatment. Experiences with imatinib mesylate show that the drug is most effective in the pre-fully malignant chronic phase, indicating that this compound is probably most beneﬁcial in the early process of tumourigenesis. However, in early stages, patients may not demonstrate clinical evidence for the disease and the clinical trials are mostly done on patients in terminal stages of the disease. Routine screening of patients may enable the detection of early stages of cancer, but will also meet ethical constrictions as the patients get to know very early that they have cancer. Another problem that faces the use of protein kinase inhibitors is the ubiquitous establishment that the particular protein kinase is indeed implicated in tumourigenesis. It is not sufﬁcient simply to know whether it is activated (or inhibited) in a speciﬁc disease state, because perturbed activity can be an irrelevant consequence of the disease rather than a major contributing factor to disease pathology. Moreover, tumours rarely have a single genetic alteration driving the neoplastic process. With likely few exceptions, a single protein kinase inhibitor may not be sufﬁcient to shutdown cancer cell activity because several protein kinases may exert perturbed activity in a speciﬁc tumour so that combinations of inhibitors may be required for positive effects. [7,10]. Furthermore, protein kinases do not exclusively regulate cell proliferation or apoptosis, but participate in other processes too. For example, Akt/PKB plays a crucial role in nutrient metabolism so that inhibition of this kinase may have such severe unwanted side effects that the drug is toxic for humans [29]. An additional problem facing the use of an appropriate inhibitor is deﬁning the suitable patient population in which to test these compounds. For example, CML patients with the lymphoid subtype of blast crisis derive very little beneﬁt from imatinib mesylate, while those with the myeloid subtype derive more limited beneﬁt in comparison with chronic phase patients. This is probably because blast phase CML patients have more heterogenous karyotypes compared to chronic phase CML patients. Thus, it is not merely the expression of a mutated protein kinase in a tumour that determines success for protein kinase-directed therapies, the context in which the mutation occurs is also important [7]. Recently, mutation analyses of all protein kinases were performed in breast cancer and colon cancer tumour samples [30]. This strategy allowed mapping the mutation spectrum in single cancer samples and the design of a protein kinase inhibitor cocktail therapy adapted to individual patients. In this regard it is important to consider that some protein kinase inhibitors have an additive or synergistic effect when combined with chemotherapy, while others fail to work or have severe side effects. For example, the combination of UCN-01, which prevents AKT activation, and the MEK inhibitor CI-1044 (PD184352) worked synergistically to induce apoptosis in human leukaemia cells. However, VEGFR inhibitors in combination with cytotoxic agents have been associated with thrombosis or haemorrhagic side effects [7]. Other examples of combined clinical trials are given in section Status of speciﬁc protein kinase inhibitors section.

162 Status of speciﬁc protein kinase inhibitors in clinical trials or in the clinic More than 40 speciﬁc protein kinase inhibitor drugs are currently in clinic or have entered clinical trials and their number is still expanding. The drugs discussed in this review are summarized in Table 1. Inhibitors against RTK Epidermal growth factor receptor The family of EGFRs consists of four family members known as EGFR (HER-1, ErbB-1), HER-2 (ErbB-2, neu), HER-3 (ErbB-3) and HER-4 (ErbB-4). Activation of the receptor requires binding of growth factors, which normally leads to oligomerization and tyrosine-autophosphorylation by the intracellular kinase domain of the receptor (Fig. 1). At least 12 different ligands that bind to HER-1, HER-3 and HER-4 receptors have been identiﬁed, whereas there is no known ligand for HER-2 so far. Instead, this protein functions as a co-receptor through heterodimerization with the other members in the EGFR family [31]. All members of the EGFR family are implicated in the development and progression of numerous human tumours, including breast, cervical, lung, colorectal, ovarian, glioma, non-small cell lung cancer (NSCLC), prostate, oesophageal, bladder, endometrial and head and neck cancer [32,33]. RTK are frequently overactive in cancer cells, even in the absence of ligand. The presence of such an overactive EGFR, due to overexpression or truncated forms, is often correlated with advanced disease or poor prognosis. The most common mutant form of EGFR is EGFRvIII, which lacks important parts of the extracellular domain and is unable to bind a ligand, but displays constitutive kinase activity [8,34,35]. Anti-EGFR therapy includes small ATP-mimicking molecules and monoclonal antibodies against the receptor and several of them have reached clinical trials or the clinics. However, results from early phase clinical trials suggest that monotherapy targeting EGFR alone may not be sufﬁcient to effectively ﬁght established tumours. Combining EGFR inhibitors with chemotherapy or incorporating EGFR inhibition into cancer prevention have been proven much more efﬁcient [35]. Erlotinib, TarcevaTM. Erlotinib (TarcevaTM) is a small molecule ATP-competitive inhibitor of HER1/EGFR. Its inhibiting effect has been studied in a whole range of cancers, including NSCLC, pancreatic cancer, ovarian cancer, cancer of head and neck. The Food and Drug Administration (FDA) recently approved Erlotinib for treatment of patients with locally advanced or metastatic NSCLC after failure of at least one chemotherapy regime [3]. Gefitinib, Iressa. Geﬁtinib (Iressa or ZD1839) is also an ATP-competitive inhibitor of EGFR that has been studied in NSCLC, colon cancer, breast

Table 1. Inhibitors of signalling protein kinases in ongoing clinical trials. Product

Synonyms

Structures

Inhibits kinase

Clinical

ABX-EGF AG-013736 AMN-107

Panitumumab

MAb

EGFR VEGF, PDGFR Bcr-Abl

II II II

VEGFRl mTOR

II II

MEK1/2

I

NVP-AMN107

CF3 O N

NH

NH

N N

N H3C

CH3 N

Angiozyme AP23573

RPI4610

Enzyme CH3

O

CH3

O

CH3

O

N O HO CH3

O

O

O

O

CH3

O

CH3

AZD6244

OH

CH3 O

ARRY-142886

O CH3 CH3

CH3

163

164

Table 1 (Continued ) Product

Synonyms

Structures

Avastin AZD2171

Bevacizumab

MAb N

Inhibits kinase

Clinical

VEGFR VEGFR1, 2, 3

I–II–III I

Bcr-Abl and Src inhib

I

PKC

I–II

O O

F N

N

N

O

BMS-354825

CH3 CH3 N S

N

N

NH NH

N

O

N

Cl OH

Bryostatin 1

H COCH3

H

CH3 CH3 OH

H O

O

OAc

H O

CH3

OH O

H OH

O

CH3 C3H7

O O

O OH COCH3

CH3

CCI-779

Cell cycle inhibitor, Temsirolimus

HO O

OH

mTOR

II–III

CH3 O O

O

N

CH3

O HO

O O

CH3

CH3 O

O CH3

OH CH3 O

CH3 CH3 CH3 O

O CH3 CH3

CEP701

KT-5555

FLT3, FMS, PDGFR

II–III

CEP751

KT-6587 (prodrug ¼ CEP2563)

RETkinase, FLT3

I–II

165

166

Table 1 (Continued ) Product

Synonyms

CEP5214

CEP7055 ¼ prodrug (see side group)

Structures NH

Inhibits kinase

Clinical

VEGFR

I

PKC

II

EGFR, PDGFR

III

O O

N O OH

CGP41251

PKC412

NH

N CH3 CH3

N

O CH3

CGP-53716

O

O

NMe2

N

O

CI-1040

PD0325901

CP547632

CI-1033

PD183805, Canertinib dihydrochloride

F

MEK1

II

VEGFR2, fGFR

II

EGFR

I–II–III

EGFR

I

O Cl

NH NH N N

O

N

O

EKB-569

EGFR kinase inhibitor 86

167

168

Table 1 (Continued ) Product

Synonyms

Structures

Inhibits kinase

Clinical

EMD-72000 Enzastaurim Erlotinib

Matuzumab LY317615.HCl TARCEVATM, NSC718781, CP358774, OSI774, R-1415

MAb

EGFR PKCb EGFR

II I–II III

Geﬁtinib

Iressa, ZD1839

EGFR

II–III

Non-speciﬁc TK inhibitor

I

Genistein

OH O OH

OH

O

HKI-272

Compound 25O

ErbB2/4

I

BCR-Abl

IV

EGFR PKC

III II

Raf1 VEGFR

II I

N O CH3 N

O HN

CH3

HN

CN

EtO

Imatinib

ST-1571; Gleevec; CGP57148B

N

N N

H N

N CH3

H N O

N N CH3

IMC-C225 ISIS-3521

Erbitux, Cetuximab LY900003, Afﬁnitak, Aprinocarsen, CGP64128A, ISI641A

MAb As

ISIS-5132 KRN-951

CGP69846A

As CH3

O

N

CH3 O O NH

N NH

O CH3

Cl

169

170

Table 1 (Continued ) Product

Synonyms

Lapatinib

GW2016, GW-572016, 572016, Lapatinib ditosylate

Structures Cl

Inhibits kinase

Clinical

ErbB1 and 2

II

c-RAF EGFR KIT, PDGFR, FLT3

I II I–II

BCR–ABL

III

O

O

O

NH

S NH

O

N

F

N

LErafAON MDX-447 MLN-518

ON-012380

As MAb CT53518

Perifosine

Pertuzumab PKI166 RAD001

KRX0401, NSC639966, D21266

Omnitarg, rhu-Mab2C4 CGP-75166 Everolimus

O O P O O

MAb O

OH

O CH3

N O CH3

SKI-606

ErbB2 ErbBr2, Met, Kit, TLK mTOR

II I–II II–III

VEGFR1/2, bFGF, PDGFR

I–stop

Src

I

N

OH O

O

O

O O OH O

O

SU5416

II

CH3 CH3 CH3

Semaxanib

PI3K-Akt/PKB

CH3

CH3

O CH3

CH3

171

172

Table 1 (Continued ) Product

Synonyms

Sorafenib

Bay 43-9006

Structures

CF3

O

Cl

O

O N

Inhibits kinase

Clinical

Raf, PDGFR, VEGFR

III

PDGFR, VEGFR2, FGFR

I–stop

N N

N

SU6668

COOH

NH O NH Sutent

SU11248, Sunitib malate

VEGFR, PDGFRa, cKit, FLT3

II–III

TAK165

Mubritinib

HER2

I

ThermaCIM h-R3 Trastuzumab UCN-01

Nimotuzumab Herceptin

MAb MAb NH

O

OH

EGFR ErbB2 PKC

II I–II–III II

VEGFR1, 2

II–III

VEGFR2, Ret, EGFR

II–III

N

N O

CH3 OMe NHCH3

Vatalanib

PTK787, ZK222584

N N N NH

Cl ZD-6474

Zactima

TM

F

Br

NH MeO N N

N

O

173

174

Fig. 1. Schematic presentation of the epidermal growth factor receptor (EGFR)

signalling pathway. The EGFR family consists of the family members ErbB1/ EGFR, ErbB2/HER-2, ErbB3/HER-3 and ErbB4/HER-4, which can form homoand heterodimers. After activation, the intrinsic tyrosine kinase activity will phosphorylate and thereby activate different signalling pathways, including the RasRaf-MAP kinase, the PI3K/Akt, the Crk/c-Abl and the PKC pathways. This leads to activation of transcription factors and modulation of cellular processes such as proliferation, cell survival and protein synthesis. Perturbed activation of the EGFR signalling pathway may contribute to tumourigenesis. Several drugs and strategies have been developed to speciﬁcally inhibit protein kinases in this pathway. The compounds discussed in this article are boxed. The red box represents antisense oligonucleotides, the blue box contains small inhibitors that interfere with the kinase activity of the enzyme, while the drugs in the green box encompass antibodies against members of the EGFR family.

cancer, cancer of head and neck and skin cancer. Recently, a phase III trial revealed that compared to placebo, IressaTM showed no improved survival in NSCLC patients and the drug was retained for new patients [36]. CI-1033. CI-1033, an abbreviation for Canertinib dihydrochloride belongs to the anilinoquinazoline class compound. This inhibitor, also known as

175 PD183805, is a water-soluble analogue of PD-169414. It not only irreversibly inhibits pan-erbB kinase, but its activity extends to blocking all four members of the erbB family. ErbB1, ErbB2 and ErbB4 are directly blocked, whereas ErbB3 is indirectly affected because of the lack of catalytically active heterodimeric partners. Additionally, the highly tumourigenic constitutively activated variants of the ErbB1, like EGFRvIII, can be blocked as well. This inhibition prevents further activation of the downstream signalling pathways through PI3kinase/Akt and Ras/MAP kinase (see sections PI3K/Akt and Mitogen-activated protein kinases, respectively). The activity of CI-1033 lies in the formation of a covalent bond between the acrylamide side chain at position six and the cysteine 773 (784 and 778 in erbB2 and erbB4, respectively), thereby permanently preventing autophosphorylation and thus activity of the tyrosine kinase. Since the interaction with the kinase is irreversible, the effects last longer, making the receptor more prone to ubiquitin tagging and subsequent degradation. In this way, new receptor synthesis is required rather than pharmacological clearance of the drug. Other pharmacological properties of the highly protein bound CI-1033 include rapid absorption, billiary excretion and good tissue distribution, except for the brain. However, toxicity seemed to be a problem due to the highly reactive acryloyl group [11]. By reducing this reactivity and increasing its afﬁnity for the receptor, annoying side effects could be decreased [11,37,38]. In preclinical xenograft models of breast, colon, lung, pancreatic, ovarian cancers and cancers of glial origin, CI-1033 delayed tumour growth or induces tumour stasis. Especially in human vulvae carcinoma xenografts, near maximum suppression of ErbB1 phosphorylation (even at the lowest dose level) was acquired. The disadvantages of this treatment are related to toxicities that concern primarily the gastrointestinal tract in rats and monkeys and dermatological toxicities in rats only. Subsequent phase I studies have been carried out in advanced stages of colon cancer, NSCLC, head and neck and breast cancers. Here, mild to moderate adverse effects were noted (thrombocytopenia and rare hypersensitivity reactions). The overall result of the heavy CI-1033 treatment schedule indicated a 40–50% decrease in the phosphorylated receptor after 7 days of treatment. Therefore, a less intensive treatment schedule was evaluated in another phase I study with patients suffering from advanced solid malignancies. The patients were submitted to treatment for 7 days, and were then untreated for the subsequent seven days. This schedule seemed to be better tolerated by the patients, yet with the same type of toxicities. However, a disappointing 6 out of 31 cases responded to the treatment with stable disease, and no better results were obtained at higher doses. Therefore, a third phase I was conducted to evaluate the responses of patients treated for 14 days combined with 1 week without treatment. Although the patients in this study could tolerate higher doses, the side effects remained. Some combination studies have also been conducted and these learned that the combination of CI-1033 with topotecan, gemcitabine

176 and SN38 (active metabolite of irinotecan) and cisplatin lead to synergistic activities against the ErbB receptors. CI-1033 will be further evaluated in subsequent phase II/III trials and as in many anti-cancer drugs combinatorial studies with other agents are on the programme as well [37,39,40]. TheraCIMh-R3. TheraCIM h-R3 is a humanized monoclonal antibody with speciﬁcity against EGFR. Phase I studies have shown that it is well tolerated and that it may enhance radiotherapy. Phase I/II trials in 24 patients with advanced carcinomas of the head and neck showed that 14 of 16 evaluable patients had an objective response and 7 had complete response [41]. Phase I/ II trials in pancreatic cancer, breast cancer and paediatric gliomas have been initiated or are planned. Derivatives of h-R3 labelled with radioisotopes are also being investigated [42]. A phase II trial including 17 children with gliomas demonstrated that of six patients that were fully evaluable all demonstrated either stable disease or partial response [43]. In another study with 24 evaluable glioma tumour patients 4 had complete response, while 5 had partial response. In total, 20 patients had stable disease [44]. HKI-272/Compound25o. HKI-272 (compound 25o) is an EKB-569 derivative with improved efﬁcacy against HER-2, but it also inhibits EGFR and blocks HER4. It binds within the ATP-binding site of these protein kinases [45,46]. HKI-272 is currently being tested in phase I and II clinical trials with breast cancer and NSCLC patients [38,47–49]. MDX-447. MDX-477 represents an example of a bispeciﬁc antibody. This monoclonal antibody was constructed by cross-linking the F/ab0 fragment of monoclonal antibody H22 (directed against CD64) and monoclonal antibody H425 ( ¼ anti-EGFR). The CD64 receptor (or FcgRI), which is present on key cytotoxic effector cells, has been implicated in the destruction of tumour cells, while overexpression of EGFR correlates with a worse clinical outcome in several cancers, including NSCLC and cancer of the prostate, breast, stomach, colon, ovary and head and neck [50,51]. Phase I studies have been or are being performed with patients with solid tumours and adult glioblastoma [52,53]. Of the 62 evaluable patients with solid tumours, 21 experienced a stable disease for 12 weeks, while one patient with mucoepidermoid parotid tumour had tumour reduction [52]. A phaseI/II trial with renal cell carcinoma, head and neck, bladder, ovarian, prostate and skin cancer showed that MDX-447 was well tolerated and that of the 36 evaluable patients, 9 had a stable disease of 3–6 months [50]. Phase II studies with squamous cell carcinoma of the head and neck are ongoing [54]. Panitumumab, ABX-EGF. This monoclonal antibody blocks activation of EGFR and prevents tumour growth and perhaps shrinks tumours. It has reached phase II clinical trial in the treatment of advanced NSCLC, and

177 metastatic colorectal cancer with relapsed or progressive disease following treatment with ﬂuoropyrimidine, irinotecan and oxaliplatin chemotherapy [53,55]. Cetuximab. Cetuximab, also known as Erbitux or IMC-C225, is a chimaeric EGFR inhibitor that was developed from a panel of antibodies by immunizing mice with human A431 epidermoid carcinoma cells that express high levels of EGFR. Subsequently, the antibodies that bound speciﬁcally to the extracellular part of the EGFR were selected and screened for those that inhibit EGF binding and receptor phosphorylation. One of those antibodies was MAb225, for which a human:murine chimaeric version was produced. The antibody binds via its heavy and light chains CDR region (rich in tyrosine and tryptophane) to a site exclusively on domain III that covers an epitope that partially overlaps the ligand-binding site. Thereby, it not only prevents the substrate from binding to the receptor but it also prevents the EGFR extracellular region from adopting the dimerization-competent extended conﬁguration [56]. The inhibition of the EGFR activity leads to a decreased cell proliferation, angiogenesis, metastasis and increased apopotosis. Furthermore an indirect anti-angiogenic effect of Cetuximab has been recorded due to dose-dependent downregulation of the expression of important angiogenic factors (VEGF, IL8, bFGF). Cetuximab has gone through phase III trials in stage III colon cancer patients [53]. Since the outcome was promising, it has recently been appointed as a second- and third-line treatment for the metastatic type of colon cancer of patients that showed disease progression while they were under Irinotecan treatment. Currently, clinical trials are being set up to evaluate the combination of Cetuximab with other anti-cancer agents [3,40,57]. PKI166. The pyrolo-pyrimidine compound PKI166 (CGP-75166) potently inhibits the EGFR and HER-2 tyrosine kinase activities, but possesses good selectivity against the tyrosine kinases Met, Kit and TCK as well [34]. Phase I studies in patients with solid tumours, including colorectal, renal, thyroid, head and neck and NSCLC demonstrated a stable disease in some of the patients and a partial remission in one of the NSCLC patients [58,59]. Hence, PKI166 reached phase II trial, but Novartis announced in October 2002 that the development of PKI166 was discontinued [34]. AEE788. AEE788 is a potent EGFR, HER2 and VEGFR inhibitor that is currently, alone or combined with RAD001, in phase I clinical trials with patients with glioblastoma or solid tumours [53]. A phase II trial in NSCLC patients has been initiated in 2005 [60]. EMD72000. EMD72000 (or Matuzumab) is a monoclonal antibody directed against the EGFR. It is the human descendant of the murine precursor

178 antibody EMD55900. It consists of the human IgG1 heavy and light chains with some remaining murine amino acids within the CDRs. The antibody binds with high afﬁnity and speciﬁcity (Kd ¼ 3.4 10 mM) and competes with the natural ligand by blocking the receptor’s binding site, which abrogates receptor-mediated downstream signalling. The anti-tumour activities of both compounds, EMD55900 and EMD72000 have been illustrated in xenographic rodent models. However, the former elicited the human anti-mouse antibody (HAMA) response and was therefore developed into a human variant. Matuzumab displays signiﬁcant anti-tumour activity in several solid tumours. Several phase I studies were carried out in advanced EGFR-positive solid malignancies unresponsive to therapy, and although there was no clear dose response relationship, sufﬁcient inhibition of the EGFR effector network was achieved at doses way below the maximum tolerated dose (MTD). The overall response rate was 23%, and the disease control rate 50%, demonstrating that single agent activity can occur in these patients. A similar study in combination with Paclitaxel in heavily pre-treated NSCLC patients indicated that the treatment was well tolerated, with moderate toxicity of Matuzumab, but with a partial clinical response (one complete response, several partial and stable disease cases). A phase II clinical trial was carried out in heavily pre-treated subjects (in primary peritoneal and platinum resistant carcinoma cases), now showing the absence of signiﬁcant clinical activity. This could be due to the far advanced stages of the cancers. Therefore, currently other phase II studies are ongoing in e.g., recurrent ovarian cancer [61–63]. Trastuzumab, Herceptin. One of the success stories of protein kinase inhibitors is Trastuzumab ( ¼ Herceptin) and its approval in 1998 was considered a milestone in EGFR-directed therapy. The history and clinical applications of Trastuzumab have been extensively discussed in excellent reviews and will therefore be brieﬂy summarized here [64,65]. Trastuzumab is a monoclonal antibody against HER2/ErbB2. This member of the EGFR family is overexpressed in 30% of invasive human breast cancers. Trastuzumab inhibits HER2 activity by prevention of ligand-induced ErbB receptor activation. Patients with HER2-overexpressing tumours are measured using a scale from 0 (negative) to 3+ (strongly positive), and the stronger the overexpression, the more likely the patient is to beneﬁt from the drug. The FDA approved the use of Trastuzumab to treat metastatic breast cancer. Trastuzumab is now also being studied in clinical trials phase II for osteosarcoma and endometrium cancer [53]. TAK165. Another HER-2/ErbB-2 inhibitor is TAK165, also called Mubritinib. Phase I trials with TAK165 are in progress in breast cancer patients [66]. Lapatinib, GW2016. Lapatinib is a dual inhibitor of the receptors ErbB1 and ErbB2, which have been implicated in various tumours such as breast cancer

179 and lung cancer. A phase II trial is currently testing Lapatinib as a ﬁrst-line therapy for breast cancer patients with tumours expressing large amounts of ErbB2 [67]. Pertuzumab. Pertuzumab, also known as Omnitarg, is a recombinant humanized monoclonal antibody that also binds to the dimerization domain of the HER2 receptor. This occlusion prevents the contact of a hairpin from the dimerization partner and consequently blocks downstream propagation of the signal. A phase I clinical study in patients with advanced malignancies revealed its clinical activity and indicated that it was well tolerated on a threeweek dosing schedule. Currently, Pertuzumab undergoes evaluation in several phase II studies for the treatment of prostate, ovarian, NSCLC and metastatic breast cancer. It has been suggested that this antibody may also be effective against androgen-independent prostate cancer [68–70]. EKB-569. EKB-569 or EGFR kinase inhibitor 86 irreversible inhibits ErB1 and ErbB2 by forming a covalent bond with Cys773 of the ATP-pocket. Since EKB-569 is very speciﬁc and water soluble, it has a good bioavailability and speciﬁc reactivity towards its target, and therefore exerts potent antitumour effects and causes few side effects. Apart from inhibiting the EGFR kinase, it also seems to contain, although less efﬁciently, activity towards HER2 in BT474 cells. Clinical trials with this component are steadily growing. Reports of its use in solid tumours indicated that EKB-569 is well tolerated and has an acceptable pharmacokinetic safety proﬁle. Toxicities associated with EKB-569 treatment were of gastrointestinal and sometimes of dermatological origin. A phase I–II dose-escalation study of EKB-569 in combination with chemotherapy FOLFOX4 and FOLFIRI pointed to some additional toxicities, including thrombocytopenia, and in cases of high doses haematological toxicities and neuropathy. But the overall responses were good as in the majority of the cases either complete or partial responses or stable disease was noted, although a minority showed signs of progressive disease as well. Currently, phase II studies in advanced colorectal cancers and combination studies of CCI-779 and Celecoxib in combination with EKB-569 are being set up to evaluate the potency of EKB569 at a larger scale [38,46,53,71–73]. Vascular endothelial growth factor receptor Angiogenesis is the process of new capillary blood vessels formation. Solid tumours, regardless of their type and origin, cannot grow beyond a certain size and therefore depend on the establishment of new blood vessels to ensure the survival and growth of the tumours. These vessels extend from existing blood vessels and they facilitate the delivery of nutrients to the tumour, as well as the removal of waste products. The VEGF, mitogens speciﬁc for vascular endothelial cells, play a pivotal role in the angiogenic process.

180 VEGF are secreted by tumour cells and macrophages and bind the VEGFR, a family of RTK [8]. There are three members of VEGFR (Fig. 2): VEGFR1 (also known as fms-like tyrosine kinase, FLT-1), VEGFR2 (kinase-insert domain receptor KDR or FLK-1) and VEGFR3 (also known as fms-like tyrosine kinase 4, FLT-4). A naturally occurring splice variant of VEGFR1, sVEGFR1 (sﬂt1) exists, while two isoforms that differ in their C-termini have been described in humans. The soluble sVEGFR1 variant lacks the intracellular tyrosine kinase domain. VEGFR1 and VEGFR2 mediate angiogenesis, while VEGFR2 and VEGFR3 are involved in lymphangiogenesis, i.e., the growth of new lymphatic vessels. Lymphangiogenesis often accompanies

Fig. 2. Signalling through class III receptor tyrosine kinases (RTKs). These RTKs

are characterized by an insertion between the two conserved elements of the tyrosine kinase domain and include the vascular endothelial growth factor receptor family (VEGFR), c-KIT and FLT-3. The VEGFR family consists VEGFR1 ( ¼ FLT-1), VEGFR2 ( ¼ KDR ¼ FLK-1) and VEGFR3 ( ¼ FLT-4). The class III RTK pathways modulate processes such as proliferation, cell migration and survival. The compounds discussed in this article are boxed. The red box represents antisense oligonucleotides, the blue box contains small inhibitors that interfere with the kinase activity of the enzyme, while the drugs in the green box encompass antibodies against members of the class III RTK family.

181 angiogenesis and may contribute to tumour metastasis. Expression of the VEGFR is cell-speciﬁc. VEGFR1 and VEGFR2 are located on e.g., activated vascular endothelial cells, dendritic cells, osteoblasts and some tumour cells, while VEGFR1 can also be present on haematopoietic stem cells. VEGFR2 is in addition expressed on circulating endothelial cells, while VEGFR3 is exclusively found on lymphatic endothelial cells in adults [74–77]. The VEGF ligands are excessively expressed in virtually all types of cancer and not only stimulate angiogenesis, but they can inhibit the anti-tumour immune response as well, thereby decreasing the host’s ability to eradicate tumour cells. This makes anti-angiogenesis an attractive strategy to treat cancer. Antiangiogenesis has a number of potential advantages over conventional cancer treatment. The risk of resistance is lower presumably because tumour-associated endothelial cells are more genetically stable than cancer cells. Moreover, inhibition of angiogenesis may prove to be more speciﬁc and less toxic because the vasculature is normally quiescent in adults. Several therapeutic strategies targeting angiogenesis in cancer are being developed. They include drugs that inhibit the VEGF ligand or the tyrosine kinase activity of their receptors, but also antibodies that inhibit the action of VEGFR, ribozymes that degrade VEGFR mRNA and soluble decoy receptors that trap and inactivate VEGF are under development or have entered clinical trials or the clinic. Problems that may arise with anti-angiogenesis therapy are unexpected haemorrhage and that this strategy is only useful in combinational therapy, but is insufﬁcient alone [75,77]. Some of the anti-angiogenic drugs are discussed below. Avastin. Avastin or Bevacizumab is a monoclonal antibody against the VEGF. Treatment with Avastin blocks the VEGF and prevents angiogenesis [78]. Phase I clinical trials in refractory solid tumours (renal cell carcinoma, breast cancer, sarcoma and lung cancer) showed safe Avastin administration without dose-limiting toxicities even in combination with chemotherapy. However, patients run an increased risk of encountering thromboembolic events. Therefore, Avastin should be administered in combination with anticoagulatory agents [79,80]. Phase II studies with Avastin were carried out in several solid tumours as well (androgen-independent prostate cancer, in peritoneal and epithelial ovary carcinoma, in non-metastatic hepatocellular cancer, relapsed metastatic breast cancer and IL-2 refractory renal cell carcinoma and studies in combination with ﬁrst-line treatment in NSCLC, metastatic colorectal cancer). The most encouraging results were obtained in renal cell cancer, NSCLC and colorectal cancers. Consequently, the effect of Avastin treatment in colorectal cancers and NSCLC (with paclitaxel and carboplatin) was further evaluated in a phase III trial. Since treatment of the former cancers lead to increased survival, the FDA approved Avastin as ﬁrstline treatment for colorectal cancers. This has also been done for metastatic breast cancer and advanced renal cell carcinomas in combination with

182 Paclitaxel and interferon, respectively [81]. However, in breast cancer patients with poor prognosis, the combined activity of Avastin with capecitabine resulted in a less effective activity, indicating that Avastin should not be used in more advanced stages of breast cancers (e.g., third-line treatment) [80]. Currently several combined clinical trials, including paclitaxel (phases I and II), Sorafenib (phase I), Rituximab (phase II), Tarceva (phase II), Erlotinib (phase II trials) are ongoing. Additionally, some trials investigate the use of Avastin as an adjuvant in combination with Irinotecan (also called Camptosar, a topoisomerase inhibitor often used in colon and rectal cancers) and capecitabine [53]. As indicated by this wide range of studies, the applicability of Avastin in several cancers seems to expand rapidly. However, for these different cancers, other optimal doses are of concern. This needs to be carefully investigated [81]. IMC-1C11. IMC-1C11 is a chimaeric monoclonal antibody that prevents VEGFR2 activation by speciﬁc binding to the extracellular domain of this RTK. The conclusions from a phase I study in 14 patients with colorectal carcinoma and hepatic metastasis are that IMC-1C11 is both safe and well tolerated and one patient showed prolonged stable disease for 28 weeks [82]. CP547632. CP547632 (3-(4-bromo-2,6-diﬂuorobenzyloxy)-5-[4-(pyrrolidin1-yl) butylaminocarbonyl-amino] isothiazole-4-carboxamide hydrobromide salt) is a potent, selective inhibitor of the tyrosine kinase activity of VEGFR2 [83]. This drug has been used in phase I clinical trials in patients with advanced NSCLC and patients with advanced solid tumours. Positive effects were observed on evaluable patients [84,85]. This has advanced CP547632 into phase II trials with ovarian cancer patients with minimal disease and NSCLC [86,87]. ZD6474/ZactimaTM. ZD6474 is an anilinoquinazoline derivative that possesses potent inhibitory activity against VEGFR2, but also EGFR and Ret. The VEGF pathway is a key mediator of tumour angiogenesis, while the EGF pathway plays a pivotal role in cell proliferation. Hence, combined targeting of both pathways may provide additive or even synergistic beneﬁt on inhibiting tumour growth [88,89]. The drug can be orally administered and phase I studies have shown that ZD6474 is well tolerated by patients. One study showed that four out of nine NSCLC patients exhibited a positive response [88,89]. Monotherapy and combined phase II studies have been evaluated in patients with NSCLC and breast cancer patients. A phase II trial in 44 evaluable breast cancer patients with ZD6474 alone reported no objective response, although one patients has stable disease for Z24 weeks [90]. In another study with breast cancer patients, a 6–7% response rate was observed [91]. A combined phase II study in NSCLC patients with ZD6474 and docetaxel indicated that combined therapy improved time to cancer

183 progression more effectively than docetaxel alone [89]. ZD6474 plus paclitaxel/carboplatin regimes showed partial responses in 7 out of 18 evaluated patients with NSCLC. Another two patients had stable disease for >12 weeks [92]. Phase II comparative studies between ZD6474 and Geﬁtinib are in progress [89], while phase II studies with small cell lung cancer (SCLC) patients, multiple myeloid and thyroid cancer patients are planned/ongoing [53]. No objective responses, however, were measured in multiple myeloma patients thus far [93]. Recruitment for phase III trials in NSCLC patients has begun. AZD2171. The indole–ether quinazoline AZD2171 is a highly potent inhibitor of the tyrosine kinase activity of VEGFR1, -2 and -3, but it does not possess activity against EGFR. A phase I trial is studying the side effects and best dose of AZD2171 when given together with chemotherapy in treating patients with advanced NSCLC or colorectal cancer. This compound has also been tested in a phase I clinical trials including patients with advanced solid tumours or liver metastases [53,88,94]. AG-013736. The inhibitor AG-013736 targets both the VEGFR and plateletderived growth factor receptor (PDGFR). This inhibition prevents downstream signalling and receptor activation, which leads to inhibition of proliferation and angiogenesis. Therefore, AG-013736 could be used in metastatic melanomas, renal cell cancer, thyroid carcinoma and NSCLC. A phase I study in advanced solid tumours indicated that there was a considerable difference between administrations to patients in a variable nutritional state. In the fasting state, two patients who had suggestions of a tumour response, suffered from an episode of haemoptysis. This reminded the researchers to the same type of episodes under treatment of Bevacizumab/ Carboplatin/Paclitaxel, which resulted in four fatal cases. In the study with AG-013736, however, there was less interpatient variability in the fasted state and within the dose range evaluated and kinetics was linear. The dose-limiting toxicities like hypertension, seizures, stomatitis, pancreatitis and abnormal liver function were noted as well. In a phase II study with kidney cancer patients, 46% of the patients showed a partial response [95–97]. CEP5214 and CEP7055. CEP5214, alias compound 21, consists of a C3 (isopropylmethoxy) part fused to a pyrrolocarbazole part and functions as a potent low-nanomolar pan-inhibitor of all three human VEGF receptors by preventing autophosphorylation. CEP5214 also seems to contain a limited selectivity against PKC, Tie2, TrkA, CDK1, p38 and IRK. CEP5214 demonstrated in vitro activity in human umbilical vein endothelial cells, seemingly in the absence of toxicity. Furthermore, it establishes a signiﬁcant in vivo antitumour activity in xenographic tumour models (prostate carcinomas and renal cell carcinomas) where it reduced metastasis by increasing tumour

184 apoptosis and decreasing microvessel density. Therefore, it was advanced into phase I clinical trials as the water soluble N,N-dimethylglycine ester prodrug 40 (CEP-u) for the treatment of a variety of solid tumours. First signs of possible clinical activity in a sarcoma and prostate cancer patient were obtained, but no treatment-related toxicities were established although at higher doses CEP5214 seemed to delay wound angiogenesis. Currently, other phase I studies are ongoing to evaluate increased dosage [98–100]. KRN-951. KRN-951 is a urea-based compound that speciﬁcally inhibits VEGF RTKs of class III and V. It potently suppresses VEGF-driven mitogenesis and capillary tube formation of the endothelial cells in vitro. However, no suppression of proliferation was discovered even at 1 mM. On the other hand, xenograft models have indicated growth inhibition during KRN951 treatment and thus currently phase I clinical trials are in progress to explore the biological activity of this compound [101]. SU-type inhibitors. Since often ATP-competitive inhibitors favour mutation in a speciﬁc residue of the ATP-binding site, conferring resistance of the tumour to therapy, researchers sought to develop drugs that simultaneously target multiple protein kinases. The ﬁrst such compound synthesized was SU5416 or Semaxanib. This highly lipophilic protein-bound small compound inhibited primarily autophosphorylation by blocking the conserved ATPbinding site within the kinase domain and thus activation of the VEGFR2 (KDR). Additionally, it had some afﬁnity for the bFGFR, PDGFR and the VEGFR1 (Flt1) as well. The compound’s high lipophilicity made it difﬁcult to solubilize and therefore it had to be administered with Cremophor via i.v. injection. In preclinical studies of xenografts (malignant melanoma, glioma, ﬁbrosarcoma, carcinomas of the lung, breast, prostate, colon and skin) tumour associated microvessel formation and proliferation were inhibited. Furthermore, SU5416 did not seem to exert cytotoxic properties, although a signiﬁcant dose-dependent tumour regression occurred. Phase I clinical studies in head and neck cancer yielded a beneﬁt for 36% of the patients, but increased the risk of thromboembolic events. Other phase I–II clinical trials with SU5416 alone or combined with radiotherapy were initiated. Unfortunately, only minimal signs of clinical activity could be detected. Owing to the formulation with cremophor and i.v. administration, patients suffered from skin sensitivities, thromboembolic events as well as gastrointestinal toxicities. Finally, in a phase III study with metastatic colorectal cancer patients in combination with 50 ﬂuorouracil and leucovorin failed to show survival beneﬁt. Therefore, the development of this inhibitor was terminated [96,102,103]. The follow-up drug of SU5416 was SU6668. This indolino type of inhibitor had increased pharmacological properties and an increased potency as was shown by its simultaneous inhibition of three receptor-tyrosine kinases (Flk1, PDGFR, basic ﬁbroblast growth factor receptor) involved in

185 neovascularization. Xenograft models of primary patient tumours showed inhibition of phosphorylation of the targeted receptors and subsequent tumour regression. However, again clinical phase I trials indicated problems with toxicity and no objective responses were observed (although four patients acquired stable disease). Despite these toxicities, hints of anti-neoplastic activity included stable disease in patients with melanomas, soft-tissue sarcomas, renal cell carcinoma and breast cancer, but these were generally weaker activities than those indicated by the xenograft models [95,104–106]. SU11248 or Sunitib Malate (Sutent) represents the new generation of drugs. This broad spectrum of competitive ATP inhibitor prevents activation of the VEGFR, PDGFRa, c-KIT and FLT-3 by targeting signalling through the class III RTK. In cell culture, SU11248 showed a rather weak activity, although in combination with Cytarabine synergistic inhibition of proliferation of primary acute myelogenous leukaemia (AML) myoblasts with Flt3ITD mutations was obtained. In contrast to cell cultures, very strong and broad potency was observed in xenograft models where this compound inhibited both tumour growth and angiogenesis. A phase study in AML patients showed that Sutent was generally well tolerated and partial remission for short durations in time was noticed. Patients with metastatic kidney cancer, breast cancer and gastrointestinal stromal tumours (GIST), that were previously unresponsive to therapy manifested partial responses and stable disease in a series of phase II studies. Phase III studies in renal cell cancer, imatinib-resistant GIST and others are ongoing. In some of those studies, tumour growth delayed considerably, and the studies were unblinded for the control group so they could beneﬁt from the treatment as well [38,78,103,107–110]. SU014813, which inhibits the members of the VEGFR family, PDGFR, KIT, RET and FLT3, is in phase I clinical trials [111]. Vatalanib. Vatalanib, also called ZK-222584, ZK-22854 and PTK-787, selectively, potently and reversibly inhibits the VEGFR2 (and the VEGFR1 in a less efﬁcient way). When applied in higher concentrations, other tyrosine kinases like PDGFR-b, c-kit and c-FMS are inhibited as well. However, Vatalanib contains no afﬁnity towards the EGFR, the bFGFR-1, c-Met, Tie2 or intracellular kinases like c-Src, c-Abl and PKCa. Xenographic models treated with Vatalanib showed a reduced microvessel growth, density and vascular permeability. In multiple myeloma cells and bone marrow cells it inhibits cell growth, survival and drug resistance. The combination of Vatalanib with the histone deacetylase inhibitor NVP-LAQ824 resulted in synergistic activity against VEGFR in prostate, breast and renal cell carcinoma cell lines. On the basis of these ﬁndings phase I studies were set up. These clinical trials indicated that Vatalanib was well tolerated and caused mild but frequent side effects. Minor regressions in renal cell cancer and stable disease in colorectal cancer, prostate and renal cell cancer and liver

186 metastasis were observed. In haematological malignancies, the drug can be tolerated at higher doses, however in that case patients suffer from more side effects. Subsequently, phase III studies were carried out in patients with solid tumours. In general, few side effects (hypertension) are noted, which is in contrast to other broad range VEGFR inhibitors that produce rashes and other allergic reactions. Ongoing combination studies indicate that Vatalanib can be safely administered with 50 FU and leucovorin/oxiplatin [38,103,112–115]. Angiozyme (RPI4610). Angiozyme or RPI4610 is a speciﬁc ribozyme that binds to and cleaves the mRNA encoding VEGFR1. A phase II trial in 83 patients with metastatic colorectal cancer with angiozyme combined with chemotherapy (Irinotecan/5-ﬂuorouracil/leucovorin) reported objective responses in some of the patients [116]. A phase Ib combined study of angiozyme with carboplatin and paclitaxel in 12 patients with solid tumours revealed a complete response for more than 7 months in one bladder cancer patient and a partial response lasting longer than 3 months in an oesophageal cancer patient. Stable disease lasting from 2 up to more than 12 months was observed in three other patients [117]. In another phase I study with 28 evaluable patients with solid tumours, two patients (nasopharyngeal carcinoma and melanoma) showed minor responses, while seven patients had stable disease for at least 6 months, with the longest 16 months [118]. Phase II studies with breast, lung, colorectal, melanoma and renal cancer patients have begun or are planned. Platelet-derived growth factor receptors PDGFRs form another family of RTK. They are the so-called class III RTK and are characterized by an insertion between the two conserved elements of the tyrosine kinase domain. This family consists of two members, PDGFRa and PDGFRb. Four ligands have been described, PDGF-A, PDGF-B, PDGF-C and PDGF-D. PDGF-A and PDGF-B can form homo- and heterodimers, while little is known about PDGF-C and PDGF-D. PDGFRa binds PDGF-AA, -BB, -AB and -CC ligand dimers with similar afﬁnity, while PDGFRb possesses the highest afﬁnity for PDGF-BB and -DD homodimers, but does not seem to bind PDGF-AA or -AB dimers. The phosphorylated intracellular domain of PDGFR can bind more than 10 different SH2-domain containing proteins, including the signal transduction molecules tyrosine kinase Src, PI3-K, PLCg, SHP-2 and GAP, and the transcription factor STAT. The diversity of the PDGFR-interacting proteins indicates that activated PDGFR can stimulate a wide variety of signalling pathways and participate in several cellular processes (Fig. 3). Indeed, PDGFR and their ligands play critical roles in mesenchymal cell migration and proliferation. Abnormalities of PDGF/PDGFR are thought to contribute to a number of human diseases, especially malignancy. Abnormalities of PDGFR activity or

187

Fig. 3. Signalling through the platelet-derived growth factor receptors (PDGFR) controls cellular processes such as proliferation, cell survival and migration. The PDGFR family consists of two members, PDGFRa and PDGFRb that bind particular homo- and heterodimers combinations of the four ligands PDGF-A, PDGFB, PDGF-C and PDGF-D described so far. PDGFR signalling can be mediated by the PI3K/Akt, the PKC and the Ras/Raf/MAPK pathways. The PDFGR inhibitors discussed in this article are shown. Activation and overexpression of the non-RTK Src is strongly associated with cancer progression. The Src inhibitor AZD0530 has recently entered phase I clinical trials. The colour code of the inhibitors is as described in the legend of Fig. 1.

expression is implicated in myeloid leukaemia and many solid tumours, including breast cancer, prostate cancer, ovarian cancer, lung cancer, melanoma, meningioma, osteoblastoma, glioblastoma, medullablastoma and astrocytoma. For example, PDGFRb is constitutively activated by enforced dimerization mediated by fusion with the TEL transcription factor (TELPDGFRb) in chronic myelogenous monocytic leukaemia patients. Approximately, half of the KIT (see section on c-Kit) mutation negative cases in GIST patients have activating mutations of PDGFRa (reviewed in [119]). Several RTK inhibitors not only target PDGFR, but also other members of the class III RTK. They will be discussed below.

188 CEP-751 and CEP-2563. CEP-751 or KT-6587 is the biologically active compound of the prodrug CEP-2563, a soluble lysinyl-g-alanyl ester of CEP751. It belongs to the same class of molecules as CEP-701 and can be converted into the latter by O-desmethylation. CEP-751 inhibits tyrosine kinase receptors like the PDGFR, by competing with ATP for binding to the kinase domain of the receptor, thereby preventing autophosphorylation and reducing downstream signalling. Apparently, CEP-751 can also inhibit Trka, Trkb and Trkc to some extent. When administered in preclinical xenograft models both CEP-2563 and CEP-751 possess inhibitory activity against a variety of tumours (e.g., medullary thyroid carcinoma, which harbours a mutation activating the RET allele). Even extended exposures to the inhibitor did not elicit neurological damage or side effects. However, application of doses above the maximum tolerated dose resulted in cardiovascular dysfunction, gastrointestinal effects, hypersensitivity and neurological reactions in rats and dogs. CEP-2563 has been evaluated in a phase I clinical trial in patients with advanced solid tumours refractory to standard therapy. The study indicated that CEP-2563 was reliably converted into CEP-751 and that even with rapid dose-escalation studies in single patient cohorts, toxicity levels remained acceptable. Generally CEP-2563 and CEP-751 seem to be safe and establish the same efﬁciency proﬁles with most common dose-limiting toxicities being hypotension and urticaria in grade III. Since the majority of the responses occurred within 80–120% of the recommended phase II dose, no further dose escalations were explored for reasons of patient’s safety [120,121]. MLN-518. MLN-518, formerly known as CT-53518 is a small molecule inhibitor developed to inhibit type III RTKs (Flt3, PDGFR and c-Kit). These RTKs are often mutated in cases of GIST, AML and SM (systemic mastocytosis) and contain mutations in the KIT juxtamembrane region and the kinase domain. Imatinib can only inhibit the juxtaform but not the kinase mutations. Therefore, MLN-518 and PD-180970 were developed, the latter compound also contains activity against Src, Abl and Kit and is used in many imatinib-resistant active site mutations. However, it is not active against malignancies expressing KIT active site mutations. MLN-518 inhibits both the juxtaform variant and the kinase mutant forms. The results of this inhibition cause inactivation of c-KIT and STAT3, which translates in the repression of cell proliferation and the induction of apoptosis. Phase I clinical trials have proven that MLN-518 effectively inhibits activation loop mutants of c-KIT and that different c-KIT mutants showed different sensitivities towards treatment with this compound. Furthermore, MLN-518 elicited low toxicity in ALM cases, which indicates a possible new treatment replacing Imatinib in these cancers. Currently, MLN-518 undergoes preclinical evaluation in gliomas, in which PDGF–PDGFR are believed to play a role as well. It is also being evaluated in phaseI/II trial for refractory AML [38,122].

189 AMG706. AMG706 selectively targets the intrinsic tyrosine kinase activity of all known VEGF, PDGF, Kit and Ret receptors. Results of a phase I trial showed that this drug was well tolerated and at least half of the subjects with advanced metastatic cancer experienced disease stabilization while on AMG706 therapy. Partial/minor responses were measured in 6 of the 34 evaluable patients [123]. Phase II trial to determine the safety and effectiveness of AMG706 in patients with advanced GIST is ongoing. c-Kit The viral oncogene v-kit was originally isolated from the Hardy–Zuckerman 4 feline sarcoma virus. This RTK is structurally similar to PDGFR and Flt-3 and therefore classiﬁed as a class III RTK (Fig. 2). Gain-of-function mutations of KIT result in development of tumours from mast cells, germ cells and GIST. It took only 3 years from the ﬁrst successful administration of Gleevec to a GIST patient and the identiﬁcation of the mutated c-kit gene as a cause of GIST [124]. Another inhibitor of the tyrosine kinase activity of KIT is E-7080, a compound similar to KRN-951 (see under section on Vascular endothelial growth factor receptor). This molecule inhibits speciﬁcally c-kit in SCLC lines. Xenograft models also seem to conﬁrm its biological activity towards c-Kit [101]. The compound PKC412 (see following section) was also used in a patient with mast cell leukaemia with the imatinib-resistant D816V mutation in RTK KIT. The patient showed a partial response, suggesting that resistance of the D816V KIT mutant to Imatinib can be circumvented by another inhibitor (e.g., PKC412), which targets the same protein kinase [125]. FLT-3 (Fig. 2) The FMS-like tyrosine kinase-3 or FLT-3 (the v-fms oncogene was originally isolated from the feline sarcoma virus SM-FeSV) belongs to the class III RTK and plays a role in normal haematopoiesis. Moreover, FLT-3 is involved in regulating the cell proliferation and survival because FLT-3 transduces signals through the phosphotidylinositol-3 kinase/AKT, the Ras/ MEK/ERK and the STAT-5 pathways. Furthermore, FLT-3 participates in the induction of the expression of the anti-apoptotic bcl-2 gene and inhibition of the pro-apoptotic bax gene. Mutations in FLT-3 resulting in constitutive active tyrosine kinase activity have been detected in AML and acute lymphoblastic leukaemia (ALL). In fact, flt-3 is the most commonly mutated gene in AML, where an internal tandem duplication in the juxtamembrane region of FLT-3 is most frequent. This internal tandem duplication can vary in length, but it causes ligand-independent dimerization and constitutive activation of the receptor. This insertion occurs in 30–40% of all AMLs and in 1–3% of ALL cases, and is associated with a more aggressive form of the disease. In 5–10% of AML patients, the gain-of-function Asp-835 to Tyr substitution in the catalytic domain of FLT-3 is detected. Other substitutions have been identiﬁed as well. Mutations in FLT-3 have rarely been observed

190 in adult patients with ALL, but seem to be more common in paediatric ALL patients [8,126–128]. CEP-701. Because of the close structural homology between FLT-3 and other class III RTK (KIT, FMS and PDGFR), several inhibitors also target these tyrosine kinases. The indolocarbazole alkaloid CEP-701 or KT-5555, which resembles CEP-751 (see section on Platelet-derived growth factor receptors), is a derivative of K252a and a fermentation product of Nonomurea longicatena. This small molecule inhibitor of the ﬂt3 gene product also contains selectivity towards the VEGFR2 and to a lesser extent to the PDGFR-b. CEP-701 competes with ATP in the kinase domain of its target receptor, and thereby inhibits autophosphorylation and downstream signalling. In 5 out of 6 preclinical xenograft mouse models of the human pancreatic ductal adenocarcinoma tested, CEP-701 inhibited both tumour growth and invasiveness. Other reports (e.g., androgen-(in)dependent prostate carcinomas, FLT3/ITD leukaemia) have indicated the same potency for CEP-701 as a single agent or in combination with other anti-cancer compounds. Consequently, clinical phase I studies have been carried out in patients with advanced malignancies. These patients tolerated CEP-701 well and showed no particular toxicities. Although a recommended dose level for phase II studies was given, CEP-701 elicited no convincing clinical activity. Another study carried out in patients with AML-expressing FLT3-activating mutations indicated clinical activity in older and heavily pre-treated patients. In this study, some cases of drug resistance were found, suggesting that CEP-701 should be combined with other kinds of treatments to eliminate this problem. In phase II studies carried out thus far, the same type of cancer has been used, and some clinical activity was detected in a few patients. Furthermore, due to its low water solubility, a lot of interindividual pharmacokinetic variability was demonstrated. Other phase II studies should help determine the clinical activity and chemotherapeutic utility of the oral administration of CEP-701 with a focus on prostate and pancreatic malignancies. Currently, phase I combinatorial studies are ongoing with CEP-701 and Gemcitabine [129–131]. PKC412. PKC412 (40 -N-benzoyl-staurosporine or midostaurin; CGP41251) is not only a broad-spectrum PKC inhibitor, it also inhibits the tyrosine kinase activities of FLT3. Phase I trials in AML patients with PKC412 combined with 5-ﬂuorouracil or with cisplatin and gemcitabine showed partial responses and tumour stabilization [132,133]. A phase II trial in 20 AML patients with FLT3 mutations achieved seven signiﬁcant responders. One patient remained stable for 11 months, while one patient attained a near complete remission [134]. CT53518. CT53518 is a selective inhibitor for FLT3, PDGFR and KIT in vitro and was found to inhibit ectopically expressed mutated FLT-3, as well

191 as to induce apoptosis in human AML cell lines with gain-of-function mutations in the FLT3 gene. Positive responses and increase in survival were observed in two mouse models of mutant FLT3-mediated AML when CT53518 was administered [135]. These encouraging results have opened for phase I and II clinical studies in AML patients [53,136]. ALK The RTK anaplastic lymphoma kinase (ALK) was originally identiﬁed as part of the chimaeric nucleophosmin (NPM)–ALK protein associated with anaplastic large cell lymphoma (ALCL). Although its expression in normal tissues is fairly restricted, ALK expression has been described in cell lines derived from neuroblastomas and in neuroectodermal tumours [137]. Like many other RTKs, ALK is implicated in oncogenesis due to genetic abnormalities. Translocation of the ALK gene at 2p23 is shown to produce oncogenic ALK fusion proteins, resulting in ALCL and inﬂammatory myoﬁbroblastic tumour (IMT). This makes ALK one of the few examples of an RTK to be involved in both non-haematopoietic and haematopoietic oncogenesis (reviewed in [138] and [139]). NPM–ALK has been linked to several apoptotic and cell proliferation signalling pathways, thus making it a potential target for alternative therapeutical approaches [138]. CD30 is a member of the tumour necrosis factor (TNF) receptor family and found to physically interact with NPM–ALK. Although the exact functional relevance of CD30 in ALK-positive ALCL is still unclear, both conjugated to saporin and unconjugated anti-CD30 antibodies possess in vivo anti-tumour activity in human ALCL cells. AntiCD30mAb (or SGN-30) have reached phase I/II clinical trials [53,140]. The 7-hydroxystaurosporine (UCN-01) is an ATP-competitive small molecule inhibitor that blocks the activity of numerous kinases and a phase I study suggested an effect in treatment of ALK-positive malignancies [141]. Herbimycin A inhibits NPM–ALK kinase activity in cell models, but so far this tyrosine kinase inhibitor has not tested clinically on ALCL patients [142]. Ribozyme-mediated therapeutic approaches have been tried, but cleavage of the NPM– ALK fusion transcript by anti-ALK ribozymes failed in human ALCL cell lines due to the prolonged half-life of NPM–ALK [138]. However, ribozyme-mediated degradation against full-length ALK is more likely to succeed due to the frequent low levels of ALK expression in tumours. Inhibitors against non-RTK c-Abl The proto-oncogene c-abl was originally identiﬁed because of its homology with the viral oncogene v-abl, whose gene product was able to induce acute neoplastic transformation in the mouse. The cellular c-abl encodes a non-RTK that plays key functions in the signalling pathways regulating

192 growth factor-induced proliferation and in the regulation of cell growth and cell cycle (Fig. 1). Moreover, this protein is engaged in the mechanisms that regulate the variations of the cellular morphology and the intercellular adhesion. In addition, c-ABL has been shown to inhibit migration of ﬁbroblasts. Finally, c-Abl may affect gene expression as it possesses a DNAbinding domain in its C-terminus and c-Abl can interact with the transcription factor cAMP-response element-binding protein [143,144]. c-Abl is strongly implicated in CML, which is tightly associated with a chromosomal abnormality known as the Philadelphia chromosome. This reciprocal translocation involves the long arms of chromosomes 9 and 22 and results in the juxtaposition of parts of the BCR and c-abl genes to form a hybrid bcr– abl gene that encodes a 210 kDa fusion protein BCR–ABL. This protein has a causative role in the neoplastic transformation of stem cells leading to CML. The oncogenic action of the BCR–ABL chimaeric protein remains elusive, but the tyrosine kinase activity of the fusion protein is substantially increased compared to wild-type ABL [144,145]. Imatinib mesylate (STI571, Glivec, Gleevecs). Imatinib mesylate was among the ﬁrst anti-cancer agents to be developed and is probably the best proof of the therapeutic potentials of rational anti-cancer drug design. The history, mode of action and resistance problems of imatinib mesylate have been extensively and excellently reviewed by others. Therefore, imatinib mesylate and its follow-up drugs are only brieﬂy discussed here. For further information, the reader is referred to recent reviews [3,146]. Imatinib mesylate inhibits the tyrosine kinase activity of the fusion protein BCR–ABL, an oncogenic hybrid found in 90% of all patients with CML and 5–25% of all patients with ALL. Imatinib also inhibits the tyrosine kinase activities of PDGFR-a, PDGFR-b and the c-KIT receptor, and is therefore used in patients with metastatic GIST [3,147]. AMN-107. AMN-107 was designed as a follow-up of Gleevec as 95% of all patients treated with Gleevec achieve remission. The rational was to develop a drug that binds tighter to BCR–ABL to increase its potency and overcome resistance due to mutations in BCR–ABL. AMN-107 is approximately 20-fold more potent than Imatinib and displays improved inhibitory activity against most of the common BCR–ABL mutations. It has reached clinical trial phase II for patients with CML and other bloodrelated cancers [148]. BMS-354825. BMS-354825 was designed for Philadelphia chromosome CML (Ph+CML) and CML patients that were resistant or intolerant to Gleevec. This is an oral agent that inhibits ﬁve tyrosine kinases, including BCR–ABL and the non-RTK Scr. The latter kinase may be activated in rare

193 patients with Imatinib resistance and play a role in signal transduction downstream of BCR–ABL. The drug has undergone phase I trial [149]. ON-012380. A reoccurring problem with Imatinib treatment is that a signiﬁcant portion of the treated patients develops mutations in the kinase domain of BCR–ABL resulting in resistance to the drug. New drugs that bind outside the ATP-binding site and thereby circumvent the problem of resistance are therefore needed. The small-molecule inhibitor ON-012380 has this potential. It inhibits BCR–ABL by substrate competition and induces cell death in Ph+ CML cells, and induces apoptosis in Imatinib-resistant mutants in vivo. In mice expressing T315I, the most common BCR–ABL mutant, ON-012380 was able to induce a regression in the leukaemia [150]. Clinical trials are being conducted [53]. SKI-606. SKI-606 is an inhibitor of the non-RTK Abl and Src that can be orally administered. Phase I studies in subjects with advanced malignant solid tumours, including breast, colorectal, pancreatic, CML and NSCLCs are in progress [47,151,152]. Src Src was originally identiﬁed as the transforming gene in the genome of the Rous sarcoma virus (v-Src) causing sarcoma in chickens. The v-src gene was a transduced form of the cellular gene c-src. The Src protein is the founding member of a family comprised of eight other members: Fyn, Yes, Lck, Hck, Blk, Fgr, Lyn and Yrk. [153]. The non-RTK Src has been found both overexpressed and highly activated in a number of human cancers, and the relationship between c-Src activation and cancer progression is signiﬁcant. Furthermore, c-Src may play a role in the acquisition of the invasive and metastatic phenotype. The Src inhibitor AZD0530 has entered phase I clinical trials [154]. General non-RTK inhibitors Genistein Genistein, an isoﬂavone that is present at high levels in soy, has been shown to alter the cellular levels of tyrosine phosphorylation of proteins. At high concentrations, genistein acts as a non-speciﬁc protein-tyrosine kinase inhibitor, whereas at lower concentrations, inhibition of tyrosine-speciﬁc protein phosphatases appears to predominate. A phase I analysis in 13 cancer patients (11 prostate and 2 colon) showed that of 12 evaluable patients, all except one, experienced disease progression [155,156]. Phase II trials in metastatic breast cancer, bladder cancer and localized prostate cancer have been approved or are in progress [53,157]. However, it has been reported that genistein can cause genetic mutation by chromosomal aberrations in human lymphoblastoid cells and peripheral blood lymphocytes in in vitro experiments [158,159].

194 Inhibitors against serine/threonine kinases PI3K/Akt AKT/PKB (protein kinase B) is a family of serine/threonine kinases. In humans, three Akt/PKB genes have been identiﬁed encoding PKBa/AKT1, PKBb/AKT2 and PKBg/AKT3, respectively. A splice variant of PKBg, referred to as PKBg1 has been identiﬁed. AKT/PKB proteins are central mediators of signal transduction pathways in response to growth factors and other extracellular stimuli and they contribute to several cellular functions such as nutrient metabolism, cell growth, transcriptional regulation and cell survival. AKT/PKB act as downstream effectors of phosphoinositide-3 kinase (PI3-K). PI3-K can be activated by RTK and G-protein-coupled receptors and activated PI3-K will generate phosphatidyl inositol-3,4,5-triphosphate (PIP3) from phosphatidyl inositol-3,4-diphosphate (PIP2). PIP3 does not activate AKT/PKB directly, but recruits AKT/PKB to the plasma membrane where it is subsequently phosphorylated (Fig. 1). The PKBa, b and g isoenzymes, except the PKBg1 splice variant, contain two regulatory phosphorylation sites, one in the activation loop within the kinase domain (Thr-308, Thr-309 and Thr-305, respectively) and one in the C-terminal regulatory domain (Ser-473, Ser-474 and Ser-472, respectively). Phosphorylation of threonine residue alone partially activates AKT, while phosphorylation of serine alone has little effect on AKT activity. For complete activation, AKT requires phosphorylation at both sites. Phosphorylation of the threonine phosphoacceptor site is mediated by the 63 kDa serine/threonine kinase phosphoinositide-dependent kinase (PDK1). The mechanism of phosphorylation of the serine residue in the Cterminal regulatory domain remains controversial. Autophosphorylation, indirect phosphorylation by PDK1 and phosphorylation by other kinases such as PDK2 and integrin-linked kinase have been suggested. Recently, it was demonstrated that mTOR in complex with rictor can phosphorylate Ser-473 in vitro and in vivo (see section on mTOR; [160]). Furthermore, phosphorylation of tyrosine residues and PI3-K-independent phosphorylation (by cAMP-dependent protein kinase or by Ca2+/calmodulin-dependent kinase) have been suggested as alternative mechanism to activate AKT/PKB, but the biological signiﬁcance remains to be determined [29,161,162]. AKT/PKB seems to play a causative role in several human cancers. Overexpression of AKT2 due to gene ampliﬁcation has been observed in 30–40% of ovarian carcinomas and pancreatic cancers, while ampliﬁcation of the Akt1 gene was reported in gastric cancers. Increased AKT1 kinase activity has been described in prostate cancer (in >50% of the tumours), breast and ovarian cancers (40% of the cancers), while AKT2 activity was especially upregulated in hepatocellular carcinomas and in colorectal cancers. Increased AKT3 activity was monitored in oestrogen receptor-deﬁcient breast cancer and in androgen-insensitive prostate cancer cell lines [161]. The implication of AKT/PKB in human cancer has urged the development of inhibitors of the PI3-K/AKT signalling pathway.

195 Perifosine. Perifosine (KRX0401; NSC639966; D-21266) is a synthetic heterocyclic alkylphosphocholine analogue derived from the miltefosine type of molecules. It differs from miltefosine in that it contains a longer alkyl chain and a piperidine head group. These features improve the inhibitors bioavailability and compared to miltefosine, Perifosine causes less nausea. In vitro experiments with cancer cell lines from squamous carcinoma, leukaemia, lung, prostate, larynx and others indicated inhibitory activity of the PI3KAKT/PKB pathway, which is often associated with tumour survival and growth. Preclinical xenograft models with breast cancer showed activity upon administration of high doses (and the high doses being correlated with the onset of the response). Three phase I studies in patients with solid tumours resulted in only minor positive responses, but mainly stable diseases were recognized. Toxicities associated with this therapy comprise gastroinstestinal complaints, which increased with increasing doses. Seemingly, variations in metabolism also contributed to the fact that patients only showed stable disease. Additionally, prophylactic 5-HT3 and steroid anti-emetics need to be supplied to the patient while under treatment. Perifosine is now being explored in phase II and in combinational studies with other compounds on the market [163,164]. mTOR Growth and proliferation of cancer cells are dependent on external signals like growth factors and signals indicating the availability of sufﬁcient nutrient and blood supply. Many of these signals are conveyed by pathways engaging mammalian target of rapamycin (mTOR). The serine/threonine kinase mTOR is a 289 kDa protein that belongs to the family of PI3K and related kinases (PIKK). The kinase domain of these members resembles the catalytic domain of phosphoinositide 3-kinase (PI3K). The PIKKs include the subfamilies: TOR (target of rapamycin), ataxia telangiectasia mutated (ATM), ataxia telangiectasia and Rad3-related (ATR) and DNA-dependent protein kinase. The kinase mTOR seems to act as a master switch of cellular metabolism, signalling cells to expand, grow and proliferate. Moreover, a role for mTOR in cell survival has also been suggested. mTOR, which forms part of the PI3K/AKT/mTOR signalling pathway, regulates the response of tumour cells to nutrients and growth factors, and controls tumour blood supply and angiogenesis through effects on VEGF in tumour and endothelial cells (Fig. 1). This pathway is activated in many cancers, especially those with elevated PI3-K signalling or those harbouring mutations in the tumour suppressor PTEN (phosphatase and tensin homologue deleted on chromosome 10). It is estimated that the mTOR signalling pathway is activated in 30–50% of prostatic cancer, 30–60% of malignant gliomas, 30–50% of endometrial carcinoma, >50% of melanoma, >30% of renal cell carcinoma and 10% of breast cancer. This makes components of this pathway very attractive targets for therapeutic inhibitors. [165–167]. However, recent ﬁndings have

196 shown that mTOR exists at least in two distinct complexes: a rapamycinsensitive complex deﬁned by its interaction with the raptor protein (regulatory-associated protein of mTOR) and a rapamycin-insensitive complex including the rictor protein (rapamycin-insensitive companion of mTOR). The latter complex cannot bind the rapamycin-FKBP12 and is therefore refractory to rapamycin. The rapamycin–rictor complex is suggested to regulate the cytoskeleton through PKCa and was recently shown to phosphorylate Ser-473 of AKT/PKB (see section on PI3K/Akt) [160]. Despite this rapamycin-insensitive complex, rapamycin analogues have entered clinical trials to inhibit the hyperactivated rapamycin-sensitive mTOR pathway in human cancers. The natural antibiotic rapamycin (or sirolimus) produced by Streptomyces hygroscopicus interacts with the cytoplasmic receptor FK506binding protein-12 (FKBP12) and the rapamycin–FKBP12 complex speciﬁcally interacts with mTOR to inhibit mTOR signalling. Although originally used as immunosuppressant drug for organ transplantation, rapamycin derivates are now entering clinical trial in cancer therapy [165,168]. The development of speciﬁc mTOR-rictor pathway inhibitors may form a novel anti-cancer therapeutic strategy in cancer with elevated AKT/PKB phosphorylation. AP23573. The rapamycin analogue AP23573 was obtained by modiﬁcation of the C-43 secondary alcohol moiety of the cyclohexyl group of rapamycin with substituted phosphonate and phosphinate groups [167]. AP23573 starves cancer cells and shrinks tumours by inhibiting the critical cell-signalling protein and is currently used in clinical trials. Phase I clinical trials indicated that AP23573 was well tolerated with no serious adverse events. In a phase I clinical trial, AP23573 completely blocked in vivo mTOR activity in peripheral blood mononuclear cells. Of eight evaluable patients, one partial response has been observed in a patient with metastatic renal cell cancer and one patient with metastatic sarcoma obtained stable disease for more than 6 months [169]. In another report of a phase I clinical trial, one out of ﬁve evaluable patients had stable medullary thyroid cancer for more than 2 months [170]. In another study, 19 out of 52 patients (37%) of patients treated with AP23573 and evaluable for at least 4 months demonstrated sustained anti-tumour responses including three patients with partial responses (conﬁrmed tumour regression >30%) and 16 patients with stable disease for at least 4 months. Five of these patients continue on trial with stable disease for at least 6 months. Twenty-three of the 32 (72%) sarcoma patients who entered the trial with tumour-related symptoms (e.g., pain, shortness of breath and cough) demonstrated clinically beneﬁcial symptomatic relief early during AP23573 treatment. Another phase II study (12.5 mg i.v. daily for 5 days every 2 weeks) in 12 patients with haematological malignancies (leukaemia and lymphomas) revealed that 6 out of 11 evaluable patients had minor improvement or stable disease [171]. Chawla and

197 co-workers [172] reported their observations of a phase II clinical trial on advanced sarcoma patients. Of the 23 evaluable patients, 13 gave symptomatic clinical improvement. Phase II study and phase Ib clinical trials with AP23573 as single agent or in combination with other anti-cancer therapies in solid tumours (sarcomas, breast, ovarian, prostate cancer, NSCLC and glioblastoma multiforme) are in progress [173]. Interestingly, AP23573 may also be used to prevent reblockage of injured vessels following stent-assisted angioplasty, a common non-surgical procedure for dilating or opening narrowed arteries because AP23573 has also been shown to potently block the growth, proliferation and migration of vascular smooth muscle cells, the primary cause of narrowing and blockage of injured vessels. CCI-779 or temsirolimus. Cell cycle inhibitor-779 (CCI-779), developed by Wyeth, is a more water-soluble ester derivative of rapamycin and can be administered i.v. Minor to major responses were obtained in phase I clinical trials involving patients with lung, renal cell, breast carcinoma, neuroendocrine tumours, soft-tissue sarcoma, endometrial and cervical carcinomas. A phase I combined trial with CCI-779 and 5-ﬂuorouracil increased the cytotoxic effects in patients with advanced solid tumours [167]. Subsequently, phase II studies have been initiated in patients with advanced breast cancer or advanced renal cell carcinoma. Clinical beneﬁt was noted in 26–37% of the breast cancer patients. However, none of the 32 Her-2-negative tumours showed any signiﬁcant response to the treatment with CCI-779. Combining CCI-779 with letrozole did not seem to have an additional effect. The overall response in renal cell carcinoma patients was 7%. Another phase II study in recurrent glioblastoma multiforme showed that the time of tumour progression had signiﬁcantly increased in 36% of the patients [174,175]. A 40% response was observed in a phase II clinical trial in patients with mantle cell lymphoma [176]. A phase II trial with CCI-7999 plus the antibody Rituximab is in progress [53]. Phase III trials with renal cell carcinoma and mantle cell lymphoma patients have been initiated [165,167,177,178]. RAD001 or everolimus. RAD001 (40-O-(2-hydroxyrthyl)-rapamycin; Everolimus) is another analogue of rapamycin that can be administered orally. This mTOR inhibitor has been/is being tested in several phase I trials in patients with solid tumours, brain tumours and leukaemia. Out of 16 evaluable patients with solid tumours, four (one with hepatocellular cancer, one with ﬁbrosarcoma and two with NSCLC) had a stable disease for more than 16 weeks [179]. Phase I combination studies with imatinib mesylate (patients with GIST [180]), gemcitabine (patients with solid tumours, [181]), letrozole (Femaras, breast cancer patients, [182]), Bevacizumab and Terlotinib [183]), Geﬁtinib (NSCLC patients, [184]) or Terlotinib (NSCLC patients, [185]) are ongoing. Of 14 evaluable patients with GIST treated with a combination of RAD001 and imatinib mesylate, eight had stable disease for more than

198 4 months, while two patients displayed partial response [180]. Phase II studies in patients with chronic lymphocytic leukaemia, mantle cell lymphoma and endometrial cancers are in progress. The rational for the latter study is that 40–50% of endometrial tumours have mutations in PTEN, which leads to constitutive activation of AKT and hence of mTOR [186]. Patients with refractory rhabdomyosarcoma or non-rhabdomyosarcomatous soft-tissue sarcoma (aged 3–21 years) are being enrolled for a phase I trial at the St. Jude Children’s Research Hospital [187], while phase III studies in breast cancer patients are planned [53]. Mitogen-activated protein kinases The MAPKs regulate diverse cellular activities including gene expression, mitosis, metabolism, motility, cell survival, apoptosis and differentiation. The MAPK pathways consist basically of a three-component module in which a cascade of phosphorylation events successively activates a MAP kinase kinase kinase (MKKK), a MAP kinase kinase (MKK) and a MAP kinase (MAPK). MAPK can phosphorylate substrates directly, or may phosphorylate yet another kinase (MAPK-activated protein kinase or MAPKAPK), which then phosphorylates substrates [188]. The role of MAPK in cancer is well documented and has been recently reviewed [189,190]. The Raf kinases and the mitogen-activated or extracellular signal-regulated protein kinases 1 and 2 have been the main focus for anti-cancer therapy therapeutic and will be discussed further. The serine/threonine kinase Raf. The Raf kinase family consists of the members c-Raf-1, B-Raf and A-Raf in humans. c-Raf-1 is the human homologue of viral Raf, which was isolated more than 20 years ago from the acutely transforming murine sarcoma virus 3311-MSV. The name Raf is derived from rapidly transforming ﬁbrosarcoma and this protein turned out to be the ﬁrst oncoprotein with serine/threonine kinase activity. Activation of the Raf members, especially c-Raf-1 is complex, incompletely understood and beyond the scope of this review. In brief, binding of extracellular ligands to their cognate receptors activates the small G protein Ras, which in its GTP-bound form recruits and activates Raf. Activated Raf, in turn activates the dual speciﬁcity protein kinases MEK1 and MEK2 (MAPK and extracellular signal-regulated kinases 1 and 2). Activated MEK1 and MEK2 phosphorylate and activate the extracellular signal-regulated kinases ERK1 and ERK2 (Fig. 1). B-Raf, rather than the more studied c-Raf-1 seems the main isoform that couples Ras to ERK [191,192]. The Ras-Raf-MEK-ERK pathway has long been associated with human cancers because oncogenic mutations in Ras occur in 15% of the cancers and ERK is hyperactivated in 30% of human tumours. Mutations in c-Raf-1 are seldom, but B-Raf is mutated with high frequency in several cancers, including melanoma (30–60%), thyroid cancer (30–50%), colorectal cancer (5–20%) and ovarian

199 cancer (30%). Over 45 gain-of-functions have been described in B-Raf, but the substitution of valine residue 599 for glutamic acid accounts for ca 90% of the B-raf mutations seen in human tumours. Mutations that render Ras constitutively active seem to be responsible for hyperactivated c-Raf-1 in tumours. The A-RAF gene seems to be rarely mutated in human cancers [191,193–195]. Despite the dominant role of B-Raf in Ras-Raf-MEK-ERK signalling and its high frequency of gain-of-function mutations, most anticancer therapies are directed at counteracting the hyperactivity of the c-Raf-1 protein kinase. ISIS-5132. ISIS-5132 (CGP69846A) is a 20-mer phosphorothionate antisense oligonucleotide that targets the 30 -untranslated region of c-raf-1 mRNA. Signiﬁcant reductions in c-raf-1 mRNA expression were observed in most of the patients within 48 h of initial ISIS-5132 dosing. The results from three phase I and one phase II studies, including patients with advanced cancers, revealed prolonged stable disease which lasted more than 7 months in a few patients. Except for one patient with ovarian carcinoma, no major responses were noticed [7,196]. This anti-sense oligonucleotide augments the cytotoxic effect of standard cytotoxic agents, suggesting more beneﬁcial effects in combined therapies. A phase II study with SCLC and NSCLC patients gave a 20% response rate in NSCLC patients, while the results in SCLC patients were inconclusive due to the limited number of patients included in this study [197]. Another phase II study in patients with locally advanced or metastatic colorectal cancer showed no evidence of clinical activity of ISIS-5132 in 15 evaluable patients, although ﬁve of them had stable disease (median duration of 3.5 months) [198]. No objective responses were observed in a phase II study in patients (n ¼ 16) with hormone-refractory prostate cancer. One patient had a stable disease for 6 months [199]. No evidence of clinical activity was observed in a phase II study in 16 evaluable patients with recurrent ovarian cancer. Four of them had a stable disease for a median of 3.8 months [200]. LErafAON. LErafAON is an anti-sense oligonucleotide against human craf that is incorporated within a cationic liposome 400 nm in diameter. The cationic lipid bilayer that entraps the negatively charged oligonucleotide stimulates cellular uptake and prevents enzymatic degradation by extracellular nucleases present in plasma. In a phase I study in 22 patients with advanced solid tumours, no objective responses were observed after treatment with LErafAON. Two patients had evidence of stable disease for >8 weeks [201]. NeoPharm developed a modiﬁed version of LErafAON, LErafAON-ETU. The lipid component in LErafAON-ETU consists of a novel, positively charged, synthetic cardiolipin (PCL-2). A phase I study with LErafAON-ETU in patients with advanced cancer is in progress [202]. Sorafenib/Bay 43-9006. The progression of renal cell carcinoma involves alternations in serine/threonine kinase c-Raf, while up to 70% of melanoma tissues carry mutations in B-Raf. In a screen for c-Raf inhibitors Bay 43-9006

200 was discovered. This drug also inhibits wild-type B-Raf and the B-RafV599E mutant, although 5–10 times less potently. Moreover, BAY 43-9006 has been shown to inhibit RTKs VEGFR2, FLT-3, PDGFR and c-Kit [101,203]. Several phase I trials showed that BAY 43-9006, either alone or in combination with traditional chemotherapy, was well tolerated. These studies included patients with colon, rectum, liver, nasopharyngeal, ovary cancer and head and neck squamous cell carcinoma. However, most studies, which have focused on renal cell carcinoma and melanoma patients as mutations in c-Raf and B-Raf, respectively, have been implicated in these tumours. In most studies, stable disease and partial responses were observed in the patients. The positive effect of BAY 43-9006 was more profound in renal cell carcinoma patients than in melanoma patients or patients with other solid tumours. This is probably because BAY 43-9006 is a better inhibitor of c-Raf [204–208]. Phase I trials with BAY 43-9006 combined with Geﬁtinib (NSCLC patients; [209]), Dacarbazine (melanoma patients; [210]), Irinotecan (solid tumour patients; [211]), Bevacizumab (solid tumour patients; [53]) and carboplatin plus paclitaxel (melanoma patients; [212]) are in progress. Phase II trials in patients with androgen-independent prostate cancer, pancreatic cancer, NSCLC, breast cancer, renal cell cancer and advanced melanoma with BAY43-9006 alone or combined with doxorubicin or with interferon have been initiated [213–218]. Phase III studies with BAY 43-9006 alone or in combination with carboplatin and paclitaxel in patients with advanced metastatic melanoma and advanced kidney cancer are underway [219,220]. Mitogen-activated or extracellular signal-regulated protein kinases 1 and 2 (MEK1/2). MEK1 and MEK2 are dual-speciﬁcity protein kinases that function in the Ras-Raf signal transduction cascade (Fig. 1). The only known MEK1/2 substrates to date are the extracellular signal-regulated kinases ERK1 and ERK2, which become phosphorylated on speciﬁc tyrosine and threonine residues by activated MEK1/2. The human MEK1 and MEK2 share 80% amino acid sequence homology, their overall structure is highly homologous and they are equally competent to phosphorylate ERK1/2 substrates. MEK1/2 are activated by a wide variety of growth factors and cytokines and also by membrane depolarization and calcium inﬂux. Dysregulation of the Ras>Raf>MEK>ERK pathway with hyperactivated MEK1/2 has been detected in more than 30% of the human tumours, yet mutations in the MEK1 and MEK2 genes are seldom, such that hyperactivation of MEK1/2 usually results from gain-of-function mutations in Ras and/or B-Raf. Increased ERK activity is found in nearly 50% of human breast tumours and is often associated with a poor prognosis [221–223]. ARRY-142886. The non ATP-competitive inhibitor ARRY-142886 or AZD6244, potently inhibits MEK1/2 by abrogating basal phosphorylation of ERK in human tumour cell lines in low nanomolar range. Preclinical

201 mouse models of pancreatic tumours, breast, colon, lung and skin cancer seem to conﬁrm its potency. Regression of the tumours occurred in all of the tested animal models (melanoma, pancreatic, colon, lung and breast cancers). The drug has recently entered phase I clinical trials in patients with advanced solid malignancies [224–226]. CI-1040 and PD0325901. CI-1040, also known as PD184352, was developed in order to inhibit MEK1 in a non-competitive reversible way. Although its most important target is MEK1, this compound seems to contain some low activity against other kinases as well. CI-1040 binds an allosteric pocket that is adjacent, but not overlapping with the ATP-binding site and inactivates the kinase activity as a result of stabilization of an inactive conformation of the activation loop and a deformation of the catalytic site [222]. The downstream effects of this compound were clear: it completely inhibited ERK phosphorylation in cells. In the subsequent phase I doseescalation study in advanced cancers indicated that this compound can be biologically active since levels of phosphorylated ERK were roughly reduced by 50% or greater in the biopsies taken. This feature was also apparent at physiological levels since some patients (pancreatic cancer) acquired a partial response and 25% of the patients showed stable disease, although CI-1040 showed poor metabolic stability and bioavailability. All dose levels were tolerated in almost all patients and mild toxicities (like diarrhoea, fatigue, rash, vomiting) were noted. Next, a phase II study trial indicated that CI-1040 was well tolerated, but despite stable disease lasting 4.4 months (range 4–18 months, the results showed insufﬁcient tumour activity to warrant further development in advanced colorectal cancer, NSCLC, breast and pancreatic cancer (only 12% responses). These effects may have been due to the pharmacological properties described during the phase I studies [227]. Therefore, by improving these properties, another component, that structurally resembled CI-1040, was developed. This compound, PD0325901, also inhibits MEK1/MEK2 in a non-ATP competitive manner, but seems to be much more potent at subnanomolar scale in cell culture. In preclinical models, tumour growth was inhibited in six out of the seven xenograft models tested. Furthermore, the reduction of phosphorylation of ERKs also seemed to last longer compared to the CI-1040 component. Additional studies to evaluate this component are currently being carried out [226,228,229]. Protein kinase C (PKC) The serine/threonine protein kinase C family consists of several members that are divided into three major groups: the classical PKC (a, b and g), the novel PKC (d, e, Z and y) and atypical PKC (l, z and t). PKC activation occurs in response to various growth factors and results in distinct cellular responses, including differentiation, physiological processes, proliferation, apoptosis and migration (Fig. 1). A role for PKC in carcinogenesis has been recognized

202 in different cancer types. Mutations in the PKC genes resulting in constitutive active kinases are very rare. Rather, increased expression or increased activities due to activation of upstream targets have been reported in tumours. Upregulated or downregulated protein levels of PKC isoenzymes or increased or decreased activity of PKC isoforms have been reported in different cancers. Elevated PKCa and PKCb activities are associated with increased mobility and invasion of the tumour cells and favours a role as inducer of proliferation and suppressor of apoptosis. Moreover, PKCb seems to be important in angiogenesis. In malignancies, downregulation of PKCd activity has been shown to lead to inhibition of apoptosis [230,231]. Because of their implication in cancer, inhibitors against PKC have been developed and tested in clinical trials. Some of them are discussed below. CGP-4152/PKC412. PKC412 (40 -N-benzoyl-staurosporine) is a broad-spectrum PKC inhibitor that also inhibits the tyrosine kinase activities of FLT3, KIT, PDGFR and VEGFR-2. PKC412 has been used in phase I clinical trials in patients with advanced NSCLC. These trials have shown partial responses and tumour stabilization. PKC412 was combined with 5-ﬂuorouracil or with cisplatin and gemcitabine [132,133]. In a phase II study with patients with advanced solid cancers, one patient with cholangiocarcinoma had stable disease lasting 4.5 months, while another had partial response lasting 4 months [232]. Reduction in tumour load in patients with chronic B cell malignancies was also observed in a phase II clinical trial with this drug [233]. A problem with PKC412 is that up to 98% of the drug binds to human plasma proteins, especially alfa-1-acide glycoprotein [232]. Bryostatin-1. Bryostatin-1 is a macrocyclic lactone that is produced by symbiont bacteria in the marine invertebrate Bugula neritina to protect the bryozoan larva from predation. Bryostatin-1 binds to the regulatory domain of protein kinase C and it was demonstrated that short-term exposure to bryostatin-1 promoted activation of PKC, whereas prolonged exposure signiﬁcantly downregulated PKC activity. In numerous haematological and solid tumour cell lines, bryostatin-1 inhibited proliferation, induced differentiation and promoted apoptosis. Furthermore, preclinical studies indicated that bryostatin-1 potently enhanced the effect of chemotherapy [234]. Therefore, a phase I study in subjects with metastatic cancer was performed. This study indicated that bryostatin-1 was safely administered and that the dose responses appeared to correlate with PKC inhibition. However, the compound showed limited clinical activity as a single agent. Other phase I and II clinical studies indicated that bryostatin-1 alone did not show clinical activity (stable disease at most in a few cases) in B-cell malignancies or leukaemia, in metastatic cervical cancer, platinum sensitive ovary carcinomas, primary peritoneal carcinomas, NSCLC and advanced renal cell cancer. Studies in

203 combination with other cytotoxic-targeted therapies are in progress. [225,235–237]. ISIS-3521. ISIS-3521, also called LY900003, Afﬁnitak, Aprinocarsen, CGP 64128A or ISI-641A, is a phosphorothioate anti-sense oligonucleotide directed against PKCa. In preclinical cell systems, this inhibitor reduces PKCa in human glioblastoma tumour cell lines. The effect of ISIS-3521 as anticancer drug has been validated in a phase I–II study alone or in combination with cisplatin and gemcitabine in patients with advanced NSCLC and other solid tumours. These studies indicated anti-tumour activity via several partial responses, mainly due to insufﬁcient characterization of patients. Currently, phase II studies in patients with metastatic colorectal cancer, non-Hodgkin’s lymphoma, ovarian carcinoma, breast cancer and other solid tumours are being explored. A recent study in recurrent high-grade gliomas showed neither tumour response nor clinical beneﬁt. Unfortunately, the patients under treatment also experienced increased intracranial pressure and oedema. However, again there was a signiﬁcant interpatient variability in e.g., plasma concentrations, which needs to be taken into account in future clinical trials. [238–240]. UCN-01. UCN-01 is a 7-hydroxy staurosporin that was isolated as a selective inhibitor of Ca2+- and phospholipid-dependent protein kinase C. Recent reports indicated that in cells depleted of PKC, UCN-01 exerted cytotoxicity. The mechanism is not completely understood, but it is assumed that UCN-01 acts as a non-speciﬁc modulator of cell cycle-dependent kinases via direct and indirect mechanisms. Thus far, this inhibitor has been tested in several phase I clinical trials and in combination studies. One study indicated that this compound modulated the effects of chemotherapeutic agents at non-toxic levels and when applied at higher concentrations it became tumouricidic. Indeed, reports about synergistic activity of UCN-01 in combination with tiotepa, cisplatin, melphalan, topotecan, gemcitabine, ﬂudarabine, ﬂuorouracil and radiation therapy in preclinical models have been noted. Recent phase I study showed no tumour responses, but only stable disease on a short infusion time schedule. The results of these studies foresee therefore that UCN-01 will be more beneﬁcial when combined with conventional chemotherapy and should be tested in phase II trials [38,163,241–244]. Enzastaurin or LY317615.HCl. Enzastaurin (LY317615.HCl), a selective inhibitor of protein kinase Cb with anti-angiogenic activity, has been tested in a phase I trial with patients with solid tumours. Four of 27 patients had stable disease [245]. LY317615 is currently in phase II trials in treating patients with gliomas and lymphomas. Of 79 evaluable patients, 13 had stable disease for more than 3 months. Fourteen patients had objective responses, one of which had a complete response [53,246].

204 Integrin-linked kinase Integrin-linked kinase (ILK) is an ankyrin repeat-containing serine/threonine kinase that interacts with the cytoplasmic domains of b1 and b3 integrins (Fig. 4). Integrins are a family of cell surface receptors that mediate the interaction of cells with the extracellular matrix (ECM). Integrins act as the bridge between ECM components and the cytoskeleton and other proteins regulating cell survival, proliferation, differentiation and migration. This kinase is widely expressed in tissues throughout the body, with the highest expression in pancreas and in cardiac and skeletal muscles. ILK functions as the effector of PI3K/protein kinase B (PKB/Akt) signalling pathway. ILK directly phosphorylates PKB/Akt on serine-473. ILK is also believed to play a role in signalling pathways that control the activity of NFKB, and in the Wnt and growth factor signalling pathways. The tumour suppressor PTEN is

Fig. 4. Signalling pathways activated by integrin-linked kinase (ILK). ILK interacts

with integrin b receptor subunit. The activity of ILK is regulated in a PI3K-dependent manner. Activated ILK phosphorylates the ser/thr kinase Akt/PKB, which forms part of the mTOR-signalling pathway. ILK also stimulates the phosphorylation of glycogen synthase kinase 3b (GSK-3b), leading to its inhibition and relieving negatively regulated pathways. This results in the stimulation of transcription factor AP-1 and the b-catenin/TCF complex. AP-1 regulates expression of MMP9, a matrix metalloproteinase implicated in invasion, while both AP-1 and the b-catenin/ TCF complex affect proliferation through upregulating the expression of cyclin D1. ILK can also activate the small GTPases Rac and CDC42, both of which can inﬂuence cell migration. Thus, ILK forms a key regulator of several cellular processes that, when aberrantly triggered, promote tumourigenesis. The inhibitors discussed in this review are boxed. The colour code is as mentioned in the legend of Fig. 1.

205 a 30 inositol lipid phosphatase that can regulate ILK activity. Mutational loss or inactivation of PTEN is commonly seen in prostate, brain and breast cancers. PTEN-null prostate carcinoma cells have constitutively increased levels of ILK activity. Inhibition of ILK activity might be an effective strategy in the treatment of PTEN-mutated cancers [247]. The importance of ILK in carcinogenesis is supported by several observations. Over-expression of ILK in epithelial cells results in anchorage-independent cell growth with increased cell cycle progression and constitutive up-regulation of cyclin D, cyclin A1, CDK4 and VEGF, and reduced the inhibitory activity of p27Kip1. Inoculation of nude mice with ILK over-expressing cells leads to tumour formation. Over-expressing of ILK also results in increased MMP-9 expression, consistent with the involvement of ILK in tumour invasion and angiogenesis. ILK gives rise to mammary tumours in transgenic mice [248]. ILK expression and activity have been correlated with malignancy in several human tumour types, including melanoma, Ewing’s sarcoma and cancers of breast, prostate, brain, stomach, ovary and colon. Higher-grade tumours express higher levels of ILK protein. Moreover, the ILK gene is localized on human chromosome 11p15.5-p15.4. This part of chromosome 11 is strongly associated with tumourigenesis (for recent reviews see [249,250]). Small molecule ILK inhibitors (e.g., KP-SD1 or KP-392, KP-SD2, KP307) that act as ATP antagonists, which have been designed by Kinetek Pharmaceutical Inc., have been shown to reduce tumour growth in a xenograft model of human colon cancer (LS-180 cells) in SCID mice. These inhibitors have an IC50 in the low micromolar range (for recent reviews see [247,250]). QLT0254, an analogue of KP-SD, inhibits the kinase activity of ILK in a cell-free assay at 185 nM, and it possesses 100-fold selectivity over other protein kinases, including CDK2, CDK5, CK2, CSK, ERK1, GSK3b, LCK, PIM1, PKA, DNA-PK and PKB/Akt. QLT0254 exerted an anti-cancer effect on a xenograft model of pancreatic cancer. Daily i.p. injection of QLT0254 for 3 weeks produced signiﬁcant tumour growth inhibition compared to vehicle control. There was also an increase, although not statistically signiﬁcant, in apoptosis and a non-signiﬁcant decrease in cell proliferation. QLT0254 in combination with gemcitabine gave a signiﬁcant increase in apoptosis compared to mock treated controls [251]. So far, no clinical trials with these inhibitors have been reported. Another ILK inhibitor is QLT0267. In vivo studies in murine models of glioblastoma are ongoing [247,252]. Another approach is the use of anti-sense oligonucleotides against ILK (ILKAS). These anti-sense oligonucleotides inhibit glioblastoma cell growth in vitro and in vivo xenograft models. Tumour growth in IKLAS-treated animals was less than 7% in animals i.p. injected once a day with 5 mg ILKAS/ kg for a 3-weeks period, while it was >100% in mock-treated animals. These doses of ILKAS did not cause toxicity to the animals [247]. Several clinical trials are underway [53].

206 Future perspectives The development of speciﬁc protein kinase inhibitors has opened a new and promising approach in anti-cancer therapy research. The multidisciplinary scientiﬁc research has considerably increased our knowledge on the structure, function and regulation of protein kinases and their role in human cancers. This has allowed the rational design of protein kinase inhibitors, with imatinib mesylate the prototype of a successful story. The implication of more protein kinases in human tumours is being appreciated and several novel inhibitors are in the pipeline or are in preclinical or clinical trials, while alternative therapeutic strategies are being developed. One promising approach is the use of nanobodies. A nanobody is the smallest available intact antigen-binding fragment harbouring the full antigen-binding capacity of the original naturally occurring heavy-chain antibodies. Nanobodies are easy to manufacture, stable, highly soluble and highly speciﬁc. This makes them excellent cancer therapeutic agents [253,254]. Despite rational drug design, the use of many protein kinase inhibitors have been hampered by problems of toxicity, non-speciﬁcity and resistance to the drug and patient-to-patient divergence in the protein kinase mutation proﬁle in a speciﬁc cancer. Numerous clinical studies have shown that protein kinase inhibitors alone are not sufﬁcient, but combined with classical chemotherapy, they may elucidate an additive or even synergistic response [78]. A beniﬁcial treatment will also require precise pinpointing of molecular defects (rapid mutation screening) in the tumour cell and an individual protein kinase activity proﬁle in each patient [3]. Recently, determining the coding sequence of 518 protein kinases in 25 breast cancer patients revealed that no common point-mutated and activated kinase gene was found in invasive ductal breast cancer [30]. In a similar study, mutation analysis was performed on the genes of 340 different serine/threonine kinases in 24 colorectal cancers. Mutations affecting eight different kinases (MKK4/JNKK1, MYLK2, PDK1, PAK4, AKT2, MARK3, CDC7 and PDIK1L) were detected. These studies demonstrate that nearly 40% of colorectal tumours had alternations in proteins that are members of the PI3K signalling pathway, making these attractive targets for therapeutic intervention [255]. Alternatively, phosphoproteomic proﬁling may offer an impression of which protein kinases are abnormally active as this will result in different phosphoprotein proﬁles in malignant tissue compared to normal tissue [186]. On the basis of this information, a particular cocktail of appropriate protein kinases can be administered to individual patients. As often, natural occurring compounds may also provide a helping hand in the ﬁght against cancer. It was recently shown that SL0101, a kaempferol glycoside isolated from the plant Fosteronia refracta, speciﬁcally inhibits the MAPK p90 ribosomal S6 kinase (RSK). As RSK is involved in proﬂiferation of prostate and breast cancer cell lines, SL0101 or a chemically modiﬁed version may be a drug target for these cancers [256].

207 References 1. 2. 3. 4. 5. 6.

7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21. 22.

Cohen P. The origins of protein phosphorylation. Nat Cell Biol 2002;4:E127–E130. Krause DS and Van Etten RA. Tyrosine kinases as targets for cancer therapy. N Engl J Med 2005;353:172–187. Tibes R, Trent J and Kurzrock R. Tyrosine kinase inhibitors and the dawn of molecular cancer therapeutics. Annu Rev Pharmacol Toxicol 2005;45:357–384. Manning G, Whyte DB, Martinez R, Hunter T and Sudarsanam S. The protein kinase complement of the human genome. Science 2002;298:1912–1934. Bode AM and Dong Z. Signal transduction pathways in cancer development and as targets for cancer prevention. Prog Nucl Acid Res Mol Biol 2005;79:237–297. Fabbro D, Ruetz S, Buchdunger E, Cowan-Jacob SW, Fendrich G, Liebetanz J, Mestan J, O’Reilly T, Traxler P, Chaudhuri B, Fretz H, Zimmermann J, Meyer T, Caravatti G, Furet P and Manley PW. Protein kinases as targets for anticancer agents: from inhibitors to useful drugs. Pharmacol Ther 2002;93:79–98. Dancey J and Sausville EA. Issues and progress with protein kinase inhibitors for cancer treatment. Nat Rev Drug Discov 2003;2:296–313. Traxler P. Tyrosine kinases as targets in cancer therapy – successes and failures. Expert Opin Ther Targets 2003;7:215–234. Sawyer TK. Novel oncogenic protein kinase inhibitors for cancer therapy. Curr Med Chem Anti-Cancer Agents 2004;4:449–455. Force T, Kuida K, Namchuk M, Parang K and Kyriakis JM. Inhibitors of protein kinase signaling pathways: emerging therapies for cardiovascular disease. Circulation 2004;109:1196–1205. Levitzki A. Protein kinase inhibitors as a therapeutic modality. Acc Chem Res 2003;36:462–469. Ahn NG and Resing KA. Cell biology. Lessons in rational drug design for protein kinases. Science 2005;308:1266–1267. Cohen MS, Zhang C, Shokat KM and Taunton J. Structural bioinformatics-based design of selective, irreversible kinase inhibitors. Science 2005;308:1318–1321. Celmed Biosciences, 2005. http://www.celmedbio.com/english/products/NB1011.html Groner B, Hartmann C and Wels W. Therapeutic antibodies. Curr Mol Med 2004;4:539–547. Hinoda Y, Sasaki S, Ishida T and Imai K. Monoclonal antibodies as effective therapeutic agents for solid tumors. Cancer Sci 2004;95:621–625. Friedrich I, Shir A, Klein S and Levitzki A. RNA molecules as anti-cancer agents. Semin Cancer Biol 2004;14:223–230. Shir A and Levitzki A. Inhibition of glioma growth by tumor-speciﬁc activation of double-stranded RNA-dependent protein kinase PKR. Nat Biotechnol 2002;20:895–900. Sioud M. Therapeutic siRNAs. Trends Pharmacol Sci 2004;25:22–28. Taske´n K and Aandahl EM. Localized effects of cAMP mediated by distinct routes of protein kinase A. Physiol Rev 2004;84:137–167. Miller WR. Regulatory subunits of PKA and breast cancer. Ann N Y Acad Sci 2002;968:37–48. Tortora G and Ciardiello F. Antisense targeting protein kinase A type I as a drug for integrated strategies of cancer therapy. Ann N Y Acad Sci 2003;1002:236–243.

We apologize to those authors whose work was not referred due to space limitations.

208 23. Bossis I and Stratakis CA. Minireview: PRKAR1A: normal and abnormal functions. Endocrinology 2004;145:5452–5458. 24. Chen HX, Marshall JL, Ness E, Martin RR, Dvorchik B, Rizvi N, Marquis J, McKinlay M, Dahut W and Hawkins MJ. A safety and pharmacokinetic study of a mixed-backbone oligonucleotide (GEM231) targeting the type I protein kinase A by two-hour infusions in patients with refractory solid tumors. Clin Cancer Res 2000;6:1259–1266. 25. Goel S, Desai K, Bulgaru A, Fields A, Goldberg G, Agrawal S, Martin R, Grindel M and Mani SA. A safety study of a mixed-backbone oligonucleotide (GEM231) targeting the type I regulatory subunit alpha of protein kinase A using a continuous infusion schedule in patients with refractory solid tumors. Clin Cancer Res 2003;9:4069–4076. 26. Goel S, Desai K, Macapinlac M, Wadler S, Goldberg G, Fields A, Einstein M, Volterra F, Wong B, Martin R and Mani S. A phase I safety and dose escalation trial of docetaxel combined with GEMs231, a second generation antisense oligonucleotide targeting protein kinase A R1 alpha in patients with advanced solid cancers. Invest New Drugs 2005 (July 18) (Epub. ahead of print). 27. Miyata Y. Hsp90 inhibitor geldanamycin and its derivatives as novel cancer chemotherapeutic agents. Curr Pharm Des 2005;11:1131–1138. 28. George P, Bali P, Annavarapu S, Scuto A, Fiskus W, Guo F, Sigua C, Sondarva G, Moscinski L, Atadja P and Bhalla K. Combination of histone deacetylase inhibitor LBH589 and the hsp90 inhibitor 17-AAG is highly active against human CML–BC cells and AML cells with activating mutation of FLT-3. Blood 2005;105:1768–1776. 29. Hanada M, Feng J and Hemmings BA. Structure, regulation and function of PKB/ AKT – a major therapeutic target. Biochim Biophys Acta 2004;1697:3–16. 30. Stephens P, Edkins S, Davies H, Greenman C, Cox C, Hunter C, Bignell G, Teague J, Smith R, Stevens C, O’Meara S, Parker A, Tarpey P, Avis T, Barthorpe A, Brackenbury L, Buck G, Butler A, Clements J, Cole J, Dicks E, Edwards K, Forbes S, Gorton M, Gray K, Halliday K, Harrison R, Hills K, Hinton J, Jones D, Kosmidou V, Laman R, Lugg R, Menzies A, Perry J, Petty R, Raine K, Shepherd R, Small A, Solomon H, Stephens Y, Tofts C, Varian J, Webb A, West S, Widaa S, Yates A, Brasseur F, Cooper CS, Flanagan AM, Green A, Knowles M, Leung SY, Looijenga LH, Malkowicz B, Pierotti MA, Teh B, Yuen ST, Nicholson AG, Lakhani S, Easton DF, Weber BL, Stratton MR, Futreal PA and Wooster R. A screen of the complete protein kinase gene family identiﬁes diverse patterns of somatic mutations in human breast cancer. Nat Genet 2005;37:590–592. 31. Leahy DJ. Structure and function of the epidermal growth factor (EGF/ErbB) family of receptors. Adv Protein Chem 2004;68:1–27. 32. Salomon DS, Brandt R, Ciardiello F and Normanno N. Epidermal growth factorrelated peptides and their receptors in human malignancies. Crit Rev Oncol Hematol 1995;19:183–232. 33. Menard S, Casalini P, Campiglio M, Pupa SM and Tagliabue E. Role of HER2/neu in tumor progression and therapy. Cell Mol Life Sci 2004;61:2965–2978. 34. Garcia-Echeverria C and Fabbro D. Therapeutically targeted anticancer agents: inhibitors of receptor tyrosine kinases. Mini Rev Med Chem 2004;4:273–283. 35. Grandis JR and Sok JC. Signaling through the epidermal growth factor receptor during the development of malignancy. Pharmacol Ther 2004;102:37–46. 36. http://www.iressa-us.com/news.asp. 37. Dewji MR. Early phase I data on an irreversible pan-reb inhibitor: CI-1033. What did we learn? J Chemother 2004;16:44–48.

209 38. Fabian MA, Biggs III WH, Treiber DK, Atteridge CE, Mihai D, Benedetti MG, Carter TA, Pietro Ciceri P, Edeen PT, Floyd M, Ford JM, Galvin M, Gerlach JL, Grotzfeld RM, Herrgard S, Insko DE, Insko MA, Lai AG, Lelias JM, Mehta SA, Milanov ZV, Velasco AM, Wodicka LM, Patel HK, Zarrinkar PP and Lockhart DJ. A small molecule–kinase interaction map for clinical kinase inhibitors. Nature 2005;23:329–336. 39. Calvo E, Tolcher AW, Hammond LA, Patnaik A, deBono JS, Eiseman IA, Olson SC, Lenehan PF, McCreery H, LoRusso P and Rowinsky EK. Administration of CI-1033, an irreversible pan-erbB tyrosine kinase inhibitor, is feasible on a 7-day on, 7-day off schadule: a phase I pharmacokinetic and food effect study. Clin Cancer Res 2004;10:7112–7120. 40. Nemunaitis J, Eiseman I, Cunningham C, Senzer N, Williams A, Lenehan PF, Olson SC, Bycott P, Schlicht M, Zentgraff R, Shin DM, Ralph G and Zinner RG. Phase 1 clinical and pharmacokinetics evaluation of oral ci-1033 in patients with refractory cancer. Clin Cancer Res 2005;11:3846–3853. 41. Crombet T, Osorio M, Cruz T, Roca C, del Castillo R, Mon R, Iznaga-Escobar N, Figueredo R, Koropatnick J, Renginfo E, Fernandez E, Torres DO, Ramos M, Leonard I, Perez R and Lage A. Use of the humanized anti-epidermal growth factor receptor monoclonal antibody h-R3 in combination with radiotherapy in the treatment of locally advanced head and neck cancer patients. J Clin Oncol 2004;22:1646–1654. 42. Spicer J. Technology evaluation: nimotuzumab, the Center of Molecular Immunology/ YMBioSciences/Oncoscience. Curr Opin Mol Ther 2005;7:182–191. 43. Bode U. Presentation at the European High-grade Glioma Meeting, Rensburg, Germany, February 25, 2005. 44. Crombet TR, Figueredo J, Catala M, Gonzalez S, Selva JC, Toledo C, Torres O, Perez R and Lage A. Treatment of high-grade astrocytic tumors with the humanized antiEGFR antibody h-R3 and radiotherapy. J Clin Oncol 2005;23:2554. 45. Rabindran SK, Discafani CM, Rosfjord EC, Baxter M, loyd MB, Golas J, Hallett WA, Johnson BD, Nilakantan R, Overbeek E, Reich MF, Shen R, Shi X, Tsou HR and Wissner A. Antitumor activity of HKI-272, an orally active, irreversible inhibitor of the HER-2 tyrosine kinase. Cancer Res 2004;64:3958–3965. 46. Tsou HR, Overbeek-Klumpers EG, Hallett WA, Reich MF, Floyd MB, Johnson BD, Michalak RS, Nilakantan R, Discafani C, Golas J, Rabindran SK, Shen R, Shi X, Wang YF, Upeslacis J and Wissner A. Optimization of 6,7-disubstituted-4-(arylamino)quinoline-3-carbonitriles as orally active, irreversible inhibitors of human epidermal growth factor receptor-2 kinase activity. J Med Chem 2005;48:1107–1131. 47. Anasetti C. High priority studies at Mofﬁtt. Phase I and I/II. Clin Trials Update 2005;6:1–6. 48. The Cleveland Clinic Taussig Cancer Center; http://www.clevelandclinic.org; 2005. 49. PhRMA; http://www.phrma.org/newmedicines; 2005. 50. Curnow RT. Clinical experience with CD64-directed immunotherapy. An overview. Cancer Immunol Immunother 1997;45:210–215. 51. Sridhar SS, Seymour L and Shepherd FA. Inhibitors of epidermal-growth-factor receptors: a review of clinical research with a focus on non-small-cell lung cancer. Lancet Oncol 2003;4:397–406. 52. Pﬁster D, Alla L, Robert B, Motzer R, Corinn W, Metz E, Sherman E and Curnow R. A phase I trial of the epidermal growth factor receptor (EGFR)-directed bispeciﬁc antibody (BsAB) MDX-447 in patients with solid tumors. 1999 ASCO Annual Meeting, Abstract 1667. 53. Clinicaltrials.gov. http://clinicaltrials.gov/, 2005.

210 54. http://www.healthsystem.Virginia.edu. 55. Tyagi P. Recent results and ongoing trials with panitumumab (ABX-EGF), a fully human anti-epidermal growth factor receptor antibody, in metastatic colorectal cancer. Clin Colorectal Cancer 2005;5:21–23. 56. Shiqing L, Schmitz KR, Jeffrey PD, Wiltzius JJW, Kussie P and Ferguson KM. Structural basis for inhibition of the epidermal growth factor receptor by cetuximab. Cancer Cell 2005;7:301–311. 57. Hoff PM. Future directions in the use of antiangiogenic agents in patients with colorectal cancer. Semin Oncol 2004;31:17–21. 58. Hoekstra R, Dumez H, van Oosterom AT, Sizer KC, Ravera C, Vaidyanathan S, Verweij J and Eskens FA. A phase I and pharmacological study of PKI166, an epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, administered orally in a two weeks on, two weeks off scheme to patients with advanced cancer. Proc Am Soc Clin Oncol 2002;21:86a. 59. Murren JR, Papadimitrakopoulou VA, Sizer K, Vaidyanathan S, Ravera C and Abbruzzese JL. A phase I dose-escalating study to evaluate the biological activity and pharmacokinetics of PKI166, a novel tyrosine kinase inhibitor, in patients with advanced cancers. Proc Am Soc Clin Oncol 2002;21:95a. 60. Giaccone G. Epidermal growth factor receptor inhibitors in the treatment of non-smallcell lung cancer. J Clin Oncol 2005;23:3235–3242. 61. Kim T. Technology evaluation: Matuzumab. Curr Opin Mol Therap 2004;6:1–8. 62. Vanhoefer U, Tewes M, Rojo F, Dirsh O, Schleucher N, Rosen O, Tillner J, Kovar A, Braun AH, Trarbach T, Seeber S, Harstrick A and Baselga J. Phase I study of the humanized anti-epidermal grwoth factor receptor monoclonal antibody EMD72000 in patients with advanced solid tumors that express the epidermal growth factor receptor. J Clin Oncol 2004;22:175–184. 63. Sieden M, Burris HA, Matulonis U, Hall J, Armstrong D, Speyer J, Tillner J, Weber D and Muggia F. A phase II trial of EMD72000 (matuzumab), a humanized anti-EGFR monoclonal antibody in subjects with heavily treated and platinum-resistant advanced muellerian malignancies. J Clin Oncol 2005;23:3151. 64. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, Fleming T, Eierman W, Wolter J, Pegram M, Baselga J and Norton L. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344:783–792. 65. Stern M and Herrmann R. Overview of monoclonal antibodies in cancer therapy: present and promise. Crit Rev Oncol Hematol 2005;54:11–29. 66. Robert NJ. A patient with Her2-overexpressing metastatic breast cancer treated with paclitaxel/carboplatin plus trastuzumab. Case Studies Breast Cancer 2004;3:1–8. 67. Gomez HL, Chavez MA, Dova DCl, Chow LWC, Wood BA, Berger MS and Sledge GW. A phase II, randomized trial using the small molecule tyrosine kinase inhibitor lapatinib as a ﬁrst-line treatment in patients with FISH positive advanced or metastatic breast cancer. J Clin Oncol 2005;23:3046. 68. Agus BD, Gordon MS, Taylor C, Natale RB, Karian B, Mendelson DS, Press MF, Allison DE, Sliwkowski MX, Lieberman G, Kelsey SM and Fyfe G. Phase I clinical study of pertuzumab, a novel HER dimerization inhibitor, in patients with advanced cancer. J Clin Oncol 2005;23:2534–2543. 69. Franklin MC, Carey KD, Vajdos FF, Leahy DJ, de Vos AM and Sliwkoski MX. Insights into ErbB signaling from the structure of the ErbB2-pertuzumab complex. Cancer Cell 2004;5:317–328.

211 70. Genentech. Omnitarg (Pertuzumab). http://www.biooncology.com/bioonc/ic_o.jsp, 2005. 71. Hidalgo M, Erlichman C, Rowinsky EK, Koepp-Norris J, Jensen K, Boni J, KorthBradley J, Quinn S and Zacharchuk C. Phase I trial of EKB-569, an irreversible inhibitor of the epidermal growth factor receptor (EGFR), patients with advanced solid tumors, ASCO Annual Meeting, 2002, Abstract 65. 72. Bonomi P. Clinical studies with non-iressa EGFR tyrosine kinase inhibitors. Lung Cancer 2003;41:S43–S48. 73. Tejpar S and Casado E. Toxicity proﬁle of the epidermal growth factor receptor inhibitor EKB-569 combined with ﬂuoroacil-based chemotherapy in patients with advanced colorectal cancer. Cancer Abstr Summaries 2004;5. 74. Cross MJ, Dixellus J, Matsumoto T and Claesson-Welsh L. VEGF-receptor signal transduction. Trends Biochem Sci 2003;28:488–494. 75. Bergsland EK. Vascular endothelial growth factor as a therapeutic target in cancer. Am J Health Syst Pharm 2004;61:S4–S11. 76. Bicknell R and Harris AL. Novel angiogenic signalling pathways and vascular targets. Annu Rev Pharmacol Toxicol 2004;44:219–238. 77. Tammella T, Enholm B, Alitalo K and Paavonen K. The biology of vascular endothelial growth factors. Cardiovasc Res 2005;65:550–563. 78. Marx J. Encouraging results for second-generation antiangiogenesis drugs. Science 2005;308:1248–1249. 79. Bioseeker. Analytical tool: Avastin and the competitive landscape. http://www.piribo.com/publications/diseases_conditions/cancer/BSK087.html, 2004. 80. Rugo HS. Bevacizumab in the treatment of breast cancer: rationale and current data. The Oncologist 2004;9:43–49. 81. Midgley R and Kerr K. Bevacizumab – current status and future directions. Gan Bunshi-Hyoteki Chiryo 2005;3:124–132. 82. Posey JA, Ng TC, Yang B, Khazaeli MB, Carpenter MD, Fox F, Needle M, aksal H and LoBuglio AF. A phase I study of anti-kinase insert domain-containing receptor antibody, IMC-1C11, in patients with liver metastases from colorectal carcinoma. Clin Cancer Res 2003;9:1323–1332. 83. Beebe JS, Jani JP, Knauth E, Goodwin P, Higdon C, Rossi AM, Emerson E, Finkelstein M, Floyd E, Harriman S, Atherton J, Hillerman S, Soderstrom C, Kou K, Gant T, Noe MC, Foster B, Rastinejad F, Marx MA, Schaeffer T, Whalen PM and Roberts WG. Pharmacological characterization of CP-547, 632, a novel vascular endothelial growth factor receptor-2 tyrosine kinase inhibitor for cancer therapy. Cancer Res 2003;63:7301–7309. 84. Cohen RB, Simon G, Langer CJ, Schol JR, McHale J, Eisenberg P, Hainsworth JD, Liau KF and Healey D. Phase I trial of CP-547, 632 (VEGFR-2) in combination with paclitaxel (P) and carboplatin (C) in advanced non-small cell lung cancer (NSCLC). J Clin Oncol 2004;22:3014. 85. Tolcher A, O’Leary JJ, DeBono J, Molpus K, Woodard C, Warnat M, Liau K, Noe D, Healy D and Huberman M. A phase I study of an oral vascular endothelial growth factor receptor2 (VEGFR2) tyrosine kinase inhibitor, CP547, 632, in patients with advanced solid tumors. AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics, Geneva, Switzerland, September 28–30, 2004. 86. Polawski S, 2005. http://www.g-o-c.org/patient/Advocacy/trials/Hamilton/PFILZER.asp.

212 87. OSI Annual report 2004. http://www.media.corporate-ir.net/mediaﬁles/NSD/OSIP/reports/OSIAR04.pfd 88. Gridelli C, Massarelli E, Maione P, Rossi A, Herbst RS, Onn A and Ciardiello F. Potential role of molecularly targeted therapy in the management of advanced nonsmall cell lung carcinoma in the elderly. Cancer 2004;101:1733–1744. 89. Heymach JV. ZD6474 – clinical experience to date. Br J Cancer 2005;92:S14–S20. 90. Miller KD, Trigo JM, Wheeler C, Barge A, Rowbottom J, Sledge G and Baselga J. A multicenter phase II trial of ZD6474, a vascular endothelial growth factor receptor-2 and epidermal growth factor receptor tyrosine kinase inhibitor, in patients with previously treated metastatic breast cancer. Clin Cancer Res 2005;11:3369–3376. 91. Cobleigh MA, Langmuir VK, Sledge GW, Miller KD, Haney L, Novotny WF, Reimann JD and Vassel A. A phase I/II dose-escalation trial of bevacizumab in previously treated metastatic breast cancer. Semin Oncol 2003;30:117–124. 92. Johnsen BE, Ma P, West H, Kerr R, Prager D, Sandler A, Herbst RS, Stewart DJ, Dimery IW and Heymach JV. Preliminary phase II safety evaluation of ZD6474, in combination with carboplatin and paclitaxel, as 1st-line treatment in patients with NSCLC. J Clin Oncol 2005;23:7102. 93. Kovacs MJ, Reece DE, Marcellus D, Meyer R, Matthews S, Dong RP and Eisenhauer EA. A phase II study of ZD6474, a vascular endothelial growth factor receptor (VEGFR) and epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) in patients with relapsed multiple myeloma (MM). Blood 2004;104(11 Suppl):Abstract 346. 94. Drevs J, Medinger M, Mross K, Zirrgiebel U, Strecker R, Unger C, Puchalski TA, Fernandes N, Roberston J and Siegert P. Phase I clinical evaluation of AZD2171, a highly potent VEGF receptor tyrosine kinase inhibitor, in patients with advanced tumors. J Clin Oncol 2005;23:3002. 95. Hwang JH. Inhibition of vascular endothelial growth factor receptor tyrosine kinase activity by small molecules. http://www.cancerpublications.com/newsletter/angiogenesis/VEGF/v1n2/articles3.html, 2004a. 96. Hwang JH. Vasc Endothel Growth Factor Oncol 2004;1:11–14. 97. Inai T, Mancuso M, Hashizume H, Baffert F, Haskell A, Baluk P, Hu-Lowe DD, Shalinsky DR, Thurston G, Yancopoulos GD and McDonald DM. Inhibition of vascular endothelial growth factor (VEGF) signaling in cancer causes loss of endothelial fenestrations, regression of tumor vessels, and appearance of basement membrane ghosts. Am J Pathol 2004;165:35–52. 98. Gingrich DE, Reddy DR, Iqbal MA, Singh J, Aimone LD, Angeles TS, Albom M, Yang S, Ator MA, Meyer SL, Robinson C, Ruggeri BA, Dionne CA, Vaught JL, Mallamo JP and Hudkins RL. A new class of potent vascular endothelial growth factor receptor tyrosine kinase inhibitors: structure–activity relationships for a series of 9alkoxymethyl-12-(3-hdyroxypropyl)indeno[2,1-a]pyrrolo[3,4-c]carbazole-5-ones and the identiﬁcation of CEP-5214 and uts dimethylglycerine ester prodrug clinical candidate CEP-7055. J Med Chem 2003;46:5375–5388. 99. Pili R, Carducci MA, Brown P, Russel L and Hurwitz H. A phase I study of the panVEGFR tyrosine kinase inhibitor, CEP-7055, in patients with advanced malignancy. Proc Am Soc Clin Oncol 2003;22:207, Abstract 831. 100. Ruggeri R, Singh J, Gingrich D, Angeles T, Albom M, Chang H, Robinson C, Hunter K, Dobrzanski P, Jones-Bolin S, Aimone L, Klein-Szanto A, Herbert JM, Bono F, Schaeffer P, Casellas P, Bourie B, Pili R, Isaacs J, Ator M, Hudkins R, Vaught J, Mallamo J and Dionne C. CEP-7055: a novel, orally pan inhibitor of vascular end-

213

101. 102.

103. 104.

105. 106.

107.

108.

109.

110.

111. 112.

othelial growth factor receptor tyrosine kinases with potent antiangiogenic activity and antitumor efﬁcacy in preclinical models. Cancer Res 2003;63:5978–5991. Dumas J, Smith RA and Lowinger TB. Recent development in the discovery of protein kinase inhibitors from the urea class. Curr Opin Drug Discov Dev 2004;7:600–616. Zangari M, Anaissie E, Stopeck A, Morimoto A, Tan N, Lancet J, Cooper M, Hanah A, Garcia-Manero G, Faderl S, Kantarjian H, Cherrinton J, Albitar M and Giles FJ. Phase II study of SU5416, a small molecule vascular endothelial growth factor tyrosine kinase receptor inhibitor, in patients with refractory multiple myeloma. Clin Cancer Res 2004;10:88–95. Arora A and Scholar EM. Role of tyrosine kinase inhibitors in cancer therapy. J Pharmacol Exp Ther 2005;315:971–979. Britten CD, Rosen LS, Kabbinavar F, Rosen P, Mulay M, Hernandez L, Brown J, Bello C, Kelsey SM and Scigalla P. Phase I trial of SU6668, a small molecule receptor tyrosine kinase inhibitor, given twice daily in patients with advanced cancers. American Society of Clinical Oncology Annual Meeting 2002, Abstract 1922. Shepherd FA and Sridhar SS. Angiogenesis inhibitors under study for the treatment of lung cancer. Lung Cancer 2003;41:563–572. Davis DW, Takamori R, Raut CP, Xiong HQ, Herbst RS, Stadler WM, Heymach JV, Demetri GD, Rashid A, Shen Y, Wen S, Abbruzzese JL and McConkey DJ. Pharmacodynamic analysis of target inhibition and endothelial cell death in tumors treated with the vascular endothelial growth factor receptor antagonists SU5416 and SU6668. Clin Cancer Res 2005;11:678–689. Abrams TJ, Murray LJ, Presenti E, Holway VW, Colombo T, Lee LB, Cherrington JM and Pryer NK. Preclinical evaluation of the tyrosine kinase inhibitor su11248 as a single agent and in combination with ‘‘standard of care’’ therapeutic agents for the therapy of breast cancer. Mol Cancer Ther 2003;2:1011–1021. Mendel DB, Laird AD, Xin X, Louie SG, Christensen JG, Li G, Schreck RE, Abrams TJ, Ngai TJ, Lee LB, Murray LJ, Carver J, Chan E, Moss KG, Haznedar OJ, Sukbuntherng J, Blake RA, Tang LSAC, Miller T, Shirazian S, McMahon G and Cherrington JM. In vivo antitumor activity of SU11248, a novel tyrosine kinase inhibitor targeting vascular endothelial growth factor and platelet-derived growth factor receptors: determination of a pharmacokinetic/pharmacodynamic relationship. Clin Cancer Res 2003;9:327–337. Yee KWH, Schittenheim M, O’Farrel AM, Town AR, McGreevey L, Bainbridge T, Cherrington JM and Heinrich MC. Synergistic effect of SU11248 with cytarabine or daunorubicin on FLT3 ITD-positive leukemic cells. Blood 2004;104:4202–4209. Fiedler W, Serve H, Dohner H, Schwittay M, Ottmann OG, O’Farrell AM, Bel lo CL, Allred R, Manning WC, Cherrington JM, Louie SG, Hong W, Brega NM, Massimini G, Scigalla P, Berdel WE and Hossfeld DK. A phase I study of SU11248 in the treatment of patients with refractory or resistent acute myeloid leukemia (AML) or not amenable to conventional therapy for disease. Blood 2005;105:986–993. http://www.multi-targetedtherapy.com. Morgan B, Thomas AL, Drevs J, Hennig J, Buchert M, Jivan A, Horsﬁled MA, Mross K, Ball HA, Lucy Lee L, Mietlowski W, Fuxius S, Unger C, O’Byrne K, Henry A, Cherryman GR, Laurent D, Dugan M, Marme D and Steward WP. Dynamic contrastenhanced magnetic resonance imaging as a biomartker for the pharmacological response of PTK787/ZK222584, an inhibito of the vascular endothelial growth factor receptor tyrosine kinases, in patients with advanced colorectal cancer and liver metastasis: results from two phase I studies. J Clin Oncol 2003;21:3955–3964.

214 113. Qian DZ, Wang X, Kachhap SK, Kato Y, Wei Y, Zhang L, Atadja P and Pili R. The histone deacetylase inhibitor NVP-LAQ824 inhibits angiogenesis and has a greater antitumor effect in combination with the vascular endothelial growth factor receptor tyrosine kinase inhibitor PTK787/ZK222584. Cancer Res 2004;64:6626–6634. 114. Drevs J, Zirrgiebel U, Schmidt-Gersbach CIM, Mross K, Medinger M, Lee L, Pinheiro J, Wood J, Thomas AL, Unger C, Henry A, Steward WP, Laurent D, Lebwohl D, Dugan M and Marme D. Soluble markers for the assessment of biological activity with PTK787/ZK222584 (PTK/ZK, a vascular endothelial growth factor receptor (VEGR)) tyrosine kinase inhibitor in patients with advanced colorectal cancer from two phase I trials. Ann Oncol 2005;16:558–565. 115. Hess-Stumpp H, Haberey M and Thierauch KH. PTK787/ZK222584, a tyrosine kinase inhibitor of all known VEGF receptors, represses tumor growth with high efﬁcacy. Chem Biochem 2005;6:550–557. 116. Venook A, Hurwitz H, Cunningham C, Burris HA, Aitchison R, Radka S, Pavco P, Capra W, Wolin M and Usman N. Relationship of clinical outcome in metastatic colorectal carcinoma to levels of soluble VEGFR-1: results of a phase II trial of a ribozyme targeting the pre-mRNA of VEGFR-1 (angiozyme), in combination with chemotherapy. Proc Am Soc Clin Oncol 2003;22:256. 117. Kobayashi H, Eckhardt SG, Lockridge JA, Rothenberg ML, Sandler AB, O’Bryant CL, Cooper W, Holden SN, Aitchison RD, Usman N, Woil M and Basche ML. Safety and pharmacokinetic study of RPI.4610 (ANGIOZYME), an anti-VEGFR-1 ribozyme, in combination with carboplatin and paclitaxel in patients with advanced solid tumors. Cancer Chemother Pharmacol 2005;56:329–336. 118. Weng DE, Masci PA, Radka SF, Jackson TE, Weiss PA, Ganapathi R, Elson PJ, Capra WB, Parker VP, Lockridge JA, Cowens JW, Usman N and Borde EC. A phase I clinical trial of a ribozyme-based angiogenesis inhibitor targeting vascular endothelial growth factor receptor-1 for patients with refractory solid tumors. Mol Cancer Ther 2005;4:948–955. 119. Jones AV and Cross NCP. Oncogenic derivatives of platelet-derived growth factor receptors. Cell Mol Life Sci 2004;61:2912–2923. 120. Strock CJ, Park JI, Rosen M, Dionne C, Ruggeri B, Jone-Bolin S, Denmeade SR, Ball DW and Nelkin BD. CEP-701 and CEP-751 inhibit constitutively activated RET tyrosine kinase activity and block medullary thyroid carcinoma cell growth. Cancer Res 2003;63:5559–5563. 121. Undevia SD, Vogelzang NJ, Mauer AM, Janisch L, Mani S and Ratain MJ. Phase I clinical trial of CEP-2563 dichloride, a receptor tyrosine kinase inhibitor, in patients with refractory solid tumors. Invest New Drugs 2004;22:449–458. 122. Corbin AS, Grisold IJ, LaRosee P, Yee KWH, Heinrich MC, Reimer CL, Druker BJ and Deininger MWN. Sensitivity of oncogenic KIT mutants to the kinase inhibitors MLN518 and PD180970. Blood 2004;104:3754–3757. 123. Herbst R, Kurzrock R, Parson M, Benjamin R, Chen L, Ng C, Ingram M, Wong S, Chang D and Rosen L. AMG 706 ﬁrst in human, open-label, dose-ﬁnding study evaluating the safety and pharmacokinetics (PK) in subjects with advanced solid tumors. Poster is found at http://www.gistsupport.org/amg%20706.html. 124. Kitamura Y and Hirota S. KIT as a human oncogenic tyrosine kinase. Cell Mol Life Sci 2004;61:2924–2931. 125. Gotlib J, Berube´ C, Growney JD, Chen CC, George TI, Williams C, Kajiguchi T, Ruan J, Lilleberg SL, Durocher JA, Lichy JH, Wang Y, Cohen PS, Arber D, Heinrich MC, Neckers L, Galli SJ, Gilliland DG and Coutre´ SE. Activity of the tyrosine kinase

215

126. 127. 128. 129.

130.

131.

132.

133.

134.

135.

136. 137.

138.

139. 140.

inhibitor PKC412 in a patient with mast cell leukemia with the D816V KIT mutation. Blood 2005;106:2867–2870. Laird AD and Cherrington JM. Small molecule tyrosine kinase inhibitors: clinical development of anticancer agents. Expert Opin Invest Drugs 2003;12:51–64. Stirewalt DL and Radich JP. The role of FLT3 in haematopoietic malignancies. Nat Rev Cancer 2003;3:650–665. Markovic A, MacKenzie KL and Lock RB. FLT-3: a new focus in the understanding of acute leukemia. Int J Biochem Cell Biol 2005;37:1168–1172. Miknyoczki SJ, Chang H, Klein-Szanto A, Gionne CA and Ruggeri BA. The Trk tyrosine kinase inhibitor CEP-701 (KT-5555) exhibits signiﬁcant antitumor efﬁcacy in preclinical xenograft models of human pancreatic ductal adenocarcinoma. Clin Cancer Res 1999;5:2205–2212. Smith BD, Levis M, Beran M, Gilles F, Kantarjian H, Berg K, Murphy KM, Dauses T, Allebach J and Small D. Single-agent CEP-701, a novel FLT3 inhibitor, shows biological and clinical acitivity in patients with relapsed or refractory acute myeloid leukemia. Blood 2004;103:3669–3878. Marshall JL, Kindler H, Deeken J, Bhargava P, Vogelzang NJ, Rizvi N, Luhtala T, Boylan S, Dordal M, Robertson P, Hawkins MJ and Ratain MJ. Phase I trial of orally administered cep-701, a novel neurotrophin receptor-linked tyrosine kinase inhibitor. Invest New Drugs 2005;23:31–37. Eder JP, Garcia-Carbonero R, Clark JW, Supko JG, Puchalski TA, Ryan DP, Deluca P, Wozniak A, Campbell A, Rothermel J and LoRusso P. A phase I trial of daily oral 40 -N-benzoyl-staurosporine in combination with protracted continuous infusion 5-ﬂuorouracil in patients with advanced solid malignancies. Invest New Drugs 2004;22:139–150. Monnerat C, Henriksson R, Le Cevalier T, Novello S, Berthaud P, Faivre S and Raymond E. Phase I study of PKC412 (N-benzoyl-staurosporine), a novel oral protein kinase C inhibitor, combined with gemcitabine and cisplatin in patients with non-smallcell lung cancer. Ann Oncol 2004;15:316–323. Stone RM, DeAngelo DJ, Klimek V, Galinsky I, Estey E, Nimer SD, Grandin W, Lebwohl D, Wang Y, Cohen P, Fox EA, Neuberg D, Clark J, Gilliland DG and Grifﬁn JD. Patients with acute myeloid leukemia and an activating mutation in FLT3 respond to a small-molecule FLT3 tyrosine kinase inhibitor, PKC412. Blood 2005;105:54–60. Kelly LM, Yu JC, Boulton CL, Apatira M, Li J, Sullivan CM, Williams I, Amaral SM, Curley DP, Duclos N, Neuberg D, Scarborough RM, Pandey A, Hollenbach S, Abe K, Lokker NA, Gilliland DG and Giese NA. CT53518, a novel selective FLT3 antagonist for the treatment of acute myelogenous leukemia. Cancer Cell 2002;1:421–432. http://www.dana-faber.org. Lamant L, Pulford K, Bischof D, Morris SW, Mason DY, Delsol G and Mariame B. Expression of the ALK tyrosine kinase gene in neuroblastoma. Am J Pathol 2000;156:1711–1721. Pulford K, Lamant L, Espinos E, Jiang Q, Xue L, Turturro F, Delsol G and Morris SW. The emerging normal and disease-related roles of anaplastic lymphoma kinase. Cell Mol Life Sci 2004;61:2939–2953. Pulford K, Morris SW and Turturro F. Anaplastic lymphoma kinase proteins in growth control and cancer. J Cell Physiol 2004;199:330–358. A Phase II multi-dose study of SGN-30 (anti-CD30 mAb) in patients with refractory or recurrent Hodgkin’s disease or anaplastic large cell ymphoma. http://data.umms.org/ scripts/trial.cfm?ID=GCC%200425.

216 141. Sausville EA, Arbuck SG, Messmann R, Headlee D, Bauer KS, Lush RM, Murgo A, Figg WD, Lahusen T, Jaken S, Jing X, Roberge M, Fuse E, Kuwabara T and Senderowicz AM. Phase I trial of 72-hour continuous infusion UCN-01 in patients with refractory neoplasms. J Clin Oncol 2001;19:2319–2333. 142. Turturro F, Arnold MD, Frist AY and Pulford K. Model of inhibition of the NPM–ALK kinase activity by herbimycin A. Clin Cancer Res 2002;8:240–245. 143. Johannessen M, Delghandi MP and Moens U. What turns CREB on? 2004;16:1211–1227. 144. Saglio G and Cillioni D. Abl: the prototype of oncogenic fusion proteins. Cell Mol Life Sci 2004;61:2897–2911. 145. Ren R. Mechanisms of BCR–ABL in the pathogenesis of chronic myelogenous leukaemia. Nat Rev Cancer 2005;5:172–183. 146. Deininger M, Buchdunger E and Druker BJ. The development of imatinib as a therapeutic agent for chronic myeloid leukemia. Blood 2005;105:2640–2653. 147. Duffaud F and Blay JY. Gastrointestinal stromal tumors: biology and treatment. Oncology 2003;65:187–197. 148. O’Hare T, Walters DK, Deininger MWN and Druker BJ. AMN107: tightening the grip of imatinib. Cancer Cell 2005;7:117–119. 149. Doggrell SA. BMS-354825: a novel drug with potential for the treatment of imatinib-resistant chronic myeloid leukaemia. Expert Opin Invest Drugs 2005; 14:89–91. 150. Gumireddy K, Baker SJ, Cosenza SC, John P, Kang AD, Robell KA, Reddy MV and Reddy EP. A non-ATP-competitive inhibitor of BCR–ABL overrides imatinib resistance. Proc Natl Acad Sci USA 2005;102:1992–1997. (Erratum in Proc Natl Acad Sci USA 2005;102:5635.) 151. http://www.irelandcancercenter.org. 152. http://www.pharma.org. 153. Playford MP and Schaller MD. The interplay between Src and integrins in normal and tumor biology. Oncogene 2004;23:7928–7946. 154. Johnston SRD. Clinical trials of intracellular signal transductions inhibitors for breast cancer – a strategy to overcome endocrine resistance. Endocr Relat Cancer 2005;12:S145–S157. 155. Poisson BA, Takimoto CH, Shapiro A, Gallot L, Nabhan C, Lieberman R and Bergan R. Pharmacokinetic analysis of the putative protstrate cancer chemopreventive agent, genistein. ASCO Meeting, 2001, Aabstract 334. 156. Takimoto CH, Glover K, Huang X, Hayes SA, Gallot L, Quinn M, Jovanovic BD, Shapiro A, Hernandez L, Goetz A, Llorens V, Lieberman R, Crowell JA, Poisson BA and Bergan RC. Phase I pharmacokinetic and pharmacodynamic analysis of unconjugated soy isoﬂavones administered to individuals with cancer. Cancer Epidemiol Biomarkers Prev 2003;12:1213–1221. 157. Karmanos Cancer Institute; http://www.karamanos.org. 158. Morris SM, Chen JJ, Domon OE, McGarrity LJ, Bishop ME, Manjanatha MG and Casciano DA. p53, mutations, and apoptosis in geni stein-exposed human lymphoblastoid cells. Mutat Res 1998;405:41–56. 159. Abe T. Infantile leukemia and soybeans – a hypothesis. Leukemia 1999;13:317–320. 160. Sarbassov DD, Guertin DA, Ali SM and Sabatini DM. Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 2005;307:1098–1101. 161. Bellacosa A, Kumar CC, Cristofano AD and Testa JR. Activation of AKT kinases in cancer: implications for therapeutic targeting. Adv Cancer Res 2005;94:29–86.

217 162. Song G, Ouyang G and Bao S. The activation of Akt/PKB signaling pathways and cell survival. J Cell Mol Med 2005;9:59–71. 163. Senderowicz AM. Novel small molecule cyclin-dependent kinases modulators in human clinical trials. Cancer Biol Ther 2003;2:S084–S095. 164. Van Ummersen L, Binger K, Volkman J, Marnocha R, Tutsch K, Kolesar J, Arzoomanian R, Alberti D and Wilding G. A phase I trials of perifosine (NSC 639966) on a loading dose/maintenance dose schedule in patients with advanced cancer. Clin Cancer Res 2004;10:7450–7456. 165. Chan S. Targeting the mammalian target of rapamycin (mTOR): a new approach to treating cancer. Br J Cancer 2004;91:1420–1424. 166. Guertin DA and Sabatini DM. An expanding role for mTOR in cancer. Trends Mol Med 2005;11:353–361. 167. Vignot S, Faivre S, Aguirre D and Raymond E. mTOR-targeted therapy of cancer with rapamycin derivatives. Ann Oncol 2005;16:525–537. 168. Bjornsti MA and Houghton PJ. The TOR pathway: a target for cancer therapy. Nat Rev Cancer 2004;4:335–348. 169. Mita MM, Rowinsky EK, Goldston ML, Mita AC, Chu Q, Syed S, Knowles HL, Rivera VM and Tolcher AW. Phase I, pharmacokinetic (PK), and pharmacodynamic (PD) study of AP23573, an mTOR inhibitor, administered iv daily x5 every other week in patients (pts) with refractory or advanced malignancies. J Clin Oncol 2004;14S:3076. 170. Desai AA, Janisch L, Berk LR, Knowles HL, Rivera VM, Bedrosian CL and Ratain MJ. A phase I trial of a novel mTOR inhibitor AP23573 administered weekly (wkly) in patients (pts) with refractory or advanced malignancies: a pharmacokinetic (PK) and pharmacodynamic (PD) analysis. J Clin Oncol 2004;14S:3150. 171. Feldman E, Giles F, Roboz G, Yee K, Curcio T, Rivera VM, Albitar M, Laliberte R and Bedrosian CL. A phase 2 clinical trial of AP23573, an mTOR inhibitor, in patients with relapsed or refractory hematologic malignancies. ASCO Annual Meeting 2005, Abstract 6631. 172. Chawla SP, Sankhala KK, Chua V, Meendez LR, Eilber FC, Eckhardt JJ, Daly ST, Rana GS, Bedrosian CL and Demetri GD. A phase II study of AP23573 (an mTOR inhibitor) in patients (pts) with advanced sarcomas. ASCO Annual Meeting 2005, Abstract 9068. 173. http://www.ariad.com. 174. Galanis E, Buckner JC, Maurer MJ, Kreisberg JI, Ballman K, Boni J, Peralba JM, Jenkins RB, Dakhill SR, Morton RF, Jaeckle KA, Scheithauer BW, Dancey J, Hidalgo M and Walsh DJ. Phase II trial of temsirolimus (CCI-779) in recurrent glioblastoma multiforme: North Central Cancer Treatment Group. J Clin Oncol 2005;23:5294–5304. 175. Galanis E, Buckner JC, Maurer MJ, Hidalgo M, Kreisberg JI, Peralba JM, Jenkins RB and Walsh DJ. N997B: phase II trial of CCI-779 in recurrent glioblastoma multiforme (GBM): updated results and correlative laboratory analysis. J Clin Oncol 2005;23:1505. 176. Witzig TE, Geyer SM, Ghobrial I, Inwards DJ, Fonseca R, Kurtin P, Ansell SM, Luyun R, Flynn PJ, Morton RFS, Dakhil R, Gross H and Kaufmann SH. Phase II trial of single-agent temsirolimus (CCI-779) for relapsed mantle cell lymphoma. J Clin Oncol 2005;23:5347–5356. 177. Atkins MB, Hidalgo M, Stadler WM, Logan TF, Dutcher JP, Hudes GR, Park Y, Liou SH, Marshall B, Boni JP, Dukart G and Sherman ML. Randomized phase II study of multiple dose levels of CCI-779, a novel mammalian target of rapamycin kinase inhibitor, in patients with advanced refractory renal cell carcinoma. J Clin Oncol 2004;22:909–918.

218 178. Prescott LM. Meeting highlights: the 40th annual meeting of the American society of clinical oncology. Pharm Ther 2004;29:496–499. 179. O’Donnell A, Faivre S, Judson I, Delbado C, Brock C, Lane H, Shand N, Hazell K, Armand JP and Raymond E. A phase I study of the oral mTOR inhibitor RAD001 as monotherapy to identify the optimal biologically effective dose using toxicity, pharmacokinetics (PK) and pharmacodynamic (PD) endpoints in patients with solid tumours. Proc Am Soc Clin Oncol 2003;22:200. 180. van Oosterom A, Reichardt P, Blay J, Dumez H, Fletcher J, Debiec-Rychter M, Shand N, Drimitrijevic S, Yap A and Demetri G. A phase I/II trial of the oral mTOR-inhibitor Everolimus (E) and Imatinib Mesylate (IM) in patients (pts) with gastrointestinal stromal tumor (GIST) refractory to IM: study update. J Clin Oncol 2005;23:9033. 181. Pacey S, Rea D, Steven N, Brock C, Knowlton N, Shand N, Hazell K, Zoellner U, O’Donnell A and Judson I. Results of a phase I clinical trial investigating a combination of the oral mTOR-inhibitor Everolimus (E, RAD001) and gemcitabine (GEM) in patients (pts) with advanced cancer. J Clin Oncol 2004;22:3120. 182. Awada A, Cardoso F, Fontaine C, Dirix L, De Gre´ve J, Sotiriou C, Steinseifer J, Wouters C, Tanaka C, Ressayre-Djaffer C and Piccart M. A phase Ib study of the mTOR inhibitor RAD001 (everolimus) in combination with letrozole (Femarad s), investigating safety and pharmacokinetics in patients with advanced breast cancer stable or slowly progressing on letrozole. 28th Annual San Antonio Breast Cancer Symposium, 2005. 183. Duke Comprehensive Cancer Center; http://www.cancer.duke.edu. 184. Milton DT, Kris MG, Azzoli CG, Gomez JE, Heelan R, Krug LM, Pao W, Pizzo B, Rizvi NA and Miller VA. Phase I/II trial of geﬁtinib and RAD001 (everolimus) in patients (pts) with advanced non-small cell lung cancer (NSCLC). J Clin Oncol 2005;23:7104. 185. MD Anderson Cancer Center. 2005. http://www.utm-ext01a.mdacc.tmc.edu. 186. Broaddus RR and Lu KH. Future challenges in clinical and translational research for endometrial cancer. Int J Gynecol Cancer 2005;15:398–411. 187. St. Jude Children’s Research Hospital; http://www.stjude.org. 188. Roux PP and Blenis J. ERK and p38 MAPK-activated protein kinases: a family of protein kinases with diverse biological functions. Microbiol Mol Biol Rev 2004;68:320–344. 189. Morin CI and Huot J. Recent advances in stress signaling in cancer. Cancer Res 2004;64:1893–1898. 190. Viala E and Pouyssegur J. Regulation of tumor cell motility by ERK mitogen-activated protein kinases. Ann N Y Acad Sci 2004;1030:208–218. 191. Wellbrock C, Karasarides M and Marais R. The RAF proteins take centre stage. Nat Rev Mol Cell Biol 2004;5:875–885. 192. Baccarini M. Second nature: biological functions of the Raf-1 ‘‘kinase’’. FEBS Lett 2005;579:3271–3277. 193. Storm SM and Rapp UR. Oncogene activation: c-raf-1 gene mutations in experimental and naturally occurring tumors. Toxicol Lett 1993;67:201–210. 194. Franse´n K, Klintena¨s M, Osterstrom A, Dimberg J, Monstein HJ and So¨derkvist P. Mutation analysis of the BRAF, ARAF and RAF-1 genes in human colorectal adenocarcinomas. Carcinogenesis 2004;25:527–533. 195. Lee JW, Soung YH, Kim SY, Park WS, Nam SW, Min WS, Kim SH, Lee JY, Yoo NJ and Lee SH. Mutational analysis of the ARAF gene in human cancers. APMIS 2005;113:54–57.

219 196. Strumberg D and Seeber S. Raf kinase inhibitors in oncology. Onkologie 2005;28:101–107. 197. Coudert B, Anthoney A, Fiedler W, Droz JP, Dieras V, Borner M, Smyth JF, Morant R, de Vires MJ, Roelvink M and Fumoleau P. Phase II trial with ISIS 5132 in patients with small-cell (SCLC) and non-small cell (NSCLC) lung cancer. A European Organization for Research and Treatment of Cancer (EORTC) Early Clinical Studies Group report. Eur J Cancer 2001;37:2194–2198. 198. Cripps MC, Figueredo AT, Oza AM, Taylor MJ, Fields AL, Holmlund JT, McIntosh LW, Geary RS and Eisenhauer EA. Phase II randomized study of ISIS 3521 and ISIS 5132 in patients with locally advanced or metastatic colorectal cancer: a National Cancer Institute of Canada clinical trials group study. Clin Cancer Res 2002;8:2188–2192. 199. Tolcher AW, Reyno L, Venner PM, Ernst SD, Moore M, Geary RS, Chi K, Hall S, Walsh W, Dorr A and Eisenhauer E. A randomized phase II and pharmacokinetic study of the antisense oligonucleotides ISIS 3521 and ISIS 5132 in patients with hormone-refractory prostate cancer. Clin Cancer Res 2002;8:2530–2535. 200. Oza AM, Elit L, Swenerton K, Faught W, Ghatage P, Carey M, McIntosh L, Dorr A, Holmlund JT and Eisenhauer E. NCIC clinical trials group study (NCIC IND.116). Phase II study of CGP 69846A (ISIS 5132) in recurrent epithelial ovarian cancer: an NCIC clinical trials group study (NCIC IND.116). Gynecol Oncol 2003;89: 129–133. 201. Rudin CM, Marshall JL, Huang CH, Kindler HL, Zhang C, Kumar D, Gokhale PC, Steinberg J, Wanaski S, Kasid UN and Ratain MJ. Delivery of a liposomal c-raf-1 antisense oligonucleotide by weekly bolus dosing in patients with advanced solid tumors: a phase I study. Clin Cancer Res 2004;10:7244–7251. 202. Steinberg JL, Mendelson DS, Block H, Green SB, Shu VS, Parker K, Cullinan P, Dul JL, von Hoff DD and Gordon MS. Phase I study of LErafAON-ETU, an easy-to-use formulation of liposome entrapped c-raf antisense oligonucleotide, in advanced cancer patients. J Clin Oncol 2005;23:3214. 203. Ahmad T and Eisen T. Kinase inhibition with BAY 43-9006 in renal cell carcinoma. Clin Cancer Res 2004;10:6388s–6392s. 204. Awada A, Hendlisz A, Gil T, Bartholomeus S, Mano M, de Valeriola D, Strumberg D, Brendel E, Haase CG, Schwartz B and Piccart M. Phase I safety and pharmacokinetics of BAY 43-9006 administered for 21 days on/7 days off in patients with advanced, refractory solid tumours. Br J Cancer 2005;92:1855–1861. 205. Clark JW, Eder JP, Ryan D, Lathia C and Lenz HJ. Safety and pharmacokinetics of the dual action raf kinase and vascular endothelial growth factor receptor inhibitor, BAY 43-9006, in patients with advanced, refractory solid tumors. Clin Cancer Res 2005;11:5472–5480. 206. Chudnovsky Y, Adams AE, Robbins PB, Lin Q and Khavari PA. Melanoma genetics and the development of rational therapeutics. J Clin Invest 2005;1(15):813–824. 207. Minami H, Kawada K, Ebi H, Kitagawa K, Kim YI, Araki K, Mukai H, Tahara M, Nakajima H and Nakajima K. A phase I study of BAY 43-9006, a dual inhibitor of Raf and VEGFR kinases, in Japanese patients with solid cancers. J Clin Oncol 2005;23:3062. 208. Siu LL, Winquist E, Agulnik M, Chin SF, Pond GR, Cheiken R, Francis P, Petrenciuc O and Chen EX. A phase II study of BAY 43-9006 in patients with recurrent and/or metastatic head and neck squamous cell carcinoma (HNSCC) and nasopharyngeal cancer (NPC). J Clin Oncol 2005;23:5566.

220 209. Adjei AA, Mandrekar S, Marks RS, Hanson LJ, Aranguren D, Jett JR, Simantov R, Schwartz B and Croghan GA. A phase I study of BAY 43-9006 and geﬁtinib in patients with refractory or recurrent non-small-cell lung cancer (NSCLC). J Clin Oncol 2005;23:3067. 210. Eisen T, Ahmad T, Gore ME, Marais R, Gibbens I, James MG, Schwartz B and Bergamini L. Phase I trial of BAY 43-9006 (sorafenib) combined with dacarbazine (DTIC) in metastatic melanoma patients. J Clin Oncol 2005;23:7508. 211. Steinbild S, Baas F, Gmehling D, Brendel E, Christensen O, Schwartz B and Mross K. Phase I study of BAY 43-9006 (sorafenib), a Raf kinase and VEGFR inhibitor, combined with irinotecan (CPT-11) in advanced solid tumors. J Clin Oncol 2005;23:3115. 212. Flaherty KT, Brose M, Schuchter L, Tuveson D, Lee R, Schwartz B, Lathia B, Weber B and O’Dwyer P. Phase I/II trial of BAY 43-9006, carboplatin (C) and paclitaxel (P) demonstrates preliminary antitumor activity in the expansion cohort of patients with metastatic melanoma. J Clin Oncol 2004;22:7507. 213. Pharmacy Choice; http://www.pharmacychoice.com. 214. Pharmaceutical news; http://www.usc.edu. 215. Posadas EM, Gulley J, Arlen PM, Harold N, Fioravanti S, Meltzer P, Scripture CD, Figg WD, Kohn EC and Dahut WL. A phase II study of BAY 43-9006 in patients with androgen-independent prostate cancer (AIPC) with proteomic proﬁling. J Clin Oncol 2005;23:4762. 216. Virginia Piper Cancer Institute; http://www.allina.com. 217. Ratain MJ, Eisen T, Stadle WM, Flaherty KT, Gore M, Desai A, Patnaik A, Xiong HQ, Schwartz B and O’Dwyer P. Final ﬁndings from a phase II, placebo-controlled, randomized discontinuation trial (RDT) of sorafenib (BAY 43-9006) in patients with advanced renal cell carcinoma (RCC). J Clin Oncol 2005;23:4544. 218. Toledo Community Hospital. http://www.tchop.com. 219. Onyx Pharmaceuticals; http://www.onyx-pharm.com. 220. Escudier B, Szczylik C, Eisen T, Stadler WM, Schwartz B, Shan M and Bukowski RM. Randomized phase III trial of the Raf kinase and VEGFR inhibitor sorafenib (BAY 439006) in patients with advanced renal cell carcinoma (RCC). J Clin Oncol 2005;232005:LBA4510. 221. Bansal A, Ramirez RD and Minna JD. Mutation analysis of the coding sequences of MEK-1 and MEK-2 genes in human lung cancer cell lines. Oncogene 1997;14:1231–1234. 222. Ohren JF, Chen H, Pavlovsky A, Whitehead C, Zhang E, Kuffa P, Yan C, McConnell P, Spessard C, Baotai C, Mueller WT, Delaney A, Omer C, Sebolt-Leopold J, Dudley DT, Leung IK, Flamme C, Warmus J, Kaufman M, Barrett S, Tecle H and Hasemann CA. Structures of human MAP kinase kinase 1 (MEK1) and MEK2 describe novel noncompetitive kinase inhibition. Nat Struct Mol Biol 2004;11:1192–1197. (Erratum in Nat Struct Mol Biol 2005;12:278.) 223. Fang JY and Richardson BC. The MAPK signalling pathways and colorectal cancer. Lancet Oncol 2005;6:322–327. 224. Bioexchange. Array Biopharma achieves milestone for initiation of phase I clinical trial for anticancer compound ARRY-4886.http://www.bioexchange.com/news/new_page.cfm?id=20501. 225. Nezhat F, Wadler S, Muggia F, Mandeli J, Goldberg G, Rahaman J, Runowicz C, Murgo AJ and Gardner GJ. Phase II trial of the combination of bryostatin-1 and cisplatin in advanced or recurrent carcinoma of the cervix: a New York gynecologic oncology group study. Gyn Oncol 2004;93:144–148.

221 226. Thompson N and Lyons J. Recent progress in targeting the Raf/MEK/ERK path way with inhibitors in cancer drug discovery. Curr Opin Pharmacol 2005;5: 350–356. 227. Rinehart J, Adjei AA, Lorusso PM, Waterhouse D, Hecht JR, Natale RB, Hamid O, Varterasian M, Asbury P, Kaldjian EP, Gulyas S, Mitchell DY, Herrera R, Seboltleopold JS and Meyer MB. Multicenter phase II study of oral MEK inhibitor, CI-1040, in patients with advanced non-small-cell lung, breast, colon, and pancreatic patients. J Clin Oncol 2004;22:4442–4445. 228. Allen LF, Sebolt-Leopold J and Meyer MB. CI-1040 (PD184352), a targeted signal transduction inhibitor of MEK (MAPK). Sem Oncol 2003;30:105–116. 229. Sebolt-Leopold JS and Herrera R. Targeting the mitogen-activated protein kinase cascade to treat cancer. Nat Rev Cancer 2004;4:937–947. 230. Hofmann J. Protein kinase C isozymes as potential targets for anticancer therapy. Curr Cancer Drug Targets 2004;4:125–146. 231. Koivunen J, Aaltonen V and Peltonen J. Protein kinase C (PKC) family in cancer progression. Cancer Lett 2005;235:1–10. 232. Propper DJ, McDonald AC, Thavasau AM, Balkwill F, Braybrooke JP, Caponigro F, Graf P, Dutreix C, Blackie R, Kaye SB, Ganesan TS, Talbot DC, Harris AL and Twelves C. Phase I and pharmacokinetic study of PKC412, an inhibitor of protein kinase C. J Clin Oncol 2001;19:1485–1492. 233. Virchis A, Ganeshaguru K, Hart S, Jones D, Fletcher L, Wright F, Wickremasinghe R, Man A, Csermak K, Meyer T, Fabbro D, Champain K, Yap A, Prentice HG and Mehta A. A novel treatment approach for low grade lymphoproliferative disorders using PKC412 (CGP41251), an inhibitor of protein kinase C. Hematol J 2002;3:131–136. 234. Kortmansky J and Schwartz GK. Bryostatin-1: a novel PKC inhibitor in clinical development. Cancer Invest 2003;21:924–936. 235. Marshall JL, Bangalore N, El-Ashry D, Fuxman Y, Johnson M, Norris B, Oberst M, Ness E, Wojtowicz-Praga S, Bhargava P, Rizvi N, Baidas S and Michael J. Phase I study of prolonged infusion Bryostatin-1 in patients. Cancer Biol Ther 2002;1:409–416. 236. Haas NB, Smith M, Lewis N, Littman L, Yeslow G, Joshi ID, Murgo A, Bradley J, Gordon R, Wang H, Rogatko A and Hudes GR. Weekly bryostatin-1 in metastatic renal cell carcinoma: a phase II study. Clin Cancer Res 2003;9:109–114. 237. Madhusudan S, Protheroe A, Propper D, Han C, Corrie P, Earl H, Hancock B, Vasey P, Turner A, Balkwill F, Hoare S and Harris AL. A multicentre phase II trial of bryostatin-1 in patients with advanced renal cancer. Br J Cancer 2003;89: 1418–1422. 238. Villalona-Calero MA, Ritch P, Figueroa JA, Otterson GA, Belt R, Dow E, George S, Leonardo J, McCachren SG, Miller GL, Modiano M, Valdivieso M, Geary R, Oliver JW and Holmlund JA. A phase I/II study of LY900003, an antisense inhibitor of protein kinase Ca, in combination with cisplatin and gemcitabine in patients with advanced non-small cell lung cancer. Clin Cancer Res 2004;10:6086–6093. 239. Advani R, Lum BL, Fisher GA, Halsey J, Geary RS, Holmlund JT, Kwoh TJ, Dorr FA and Sikic BI. A phase I trial of aprinocarsen (ISIS 3521/LY900003), an antisense inhibitor of protein kinase C-alpha administered as a 24-hour weekly infusion schedule in patients with advanced cancer. Invest New Drugs 2005;23:467–477. 240. Grossman SA, Alavi JB, Supko JG, Carson KA, Priet R, Dorr FA, Grundy JS and Holmlund JT. Efﬁcacy and toxicity of the antisense oligonucleotide Aprinocarsen

222

241.

242.

243.

244.

245.

246.

247.

248.

249.

250. 251.

252.

253.

directed aginst protein kinase Ca delivered as a 21-day continuous intravenous infusion in patients with recurrent high-grade astrocytomas. Neuro-Oncology 2005;7:32–40. Bhonde MR, Hanski ML, Margrini R, Moorthy D, Mueller A, Sausville EA, Kohno K, Wiegland P, Daniel PT, Zeitz M and Hanski C. The broad-range cyclin-dependent kinase inhibitor UNC-01 induces apoptosis in colon carcinoma cells through transcriptional suppression of the bcl-x protein. Oncogene 2005;24: 148–156. Dees EC, Baker SD, O’Reilly S, Rudek MA, Davidson SB, Aylesworth C, Elza-Brown K, Carducci MA and Donehower RC. A phase I and pharmacokinetics study of short infusions of UCN-01 in patients with refractory solid tumors. Clin Cancer Res 2005;11:664–671. Kortmansky J, Shah MA, Kaubisch A, Weyerbacher A, Yi S, Tong W, Sowers R, Gonen M, O’Reilly E, Kemeny N, Ilson DI, Saltz LB, Maki RG, Kelsen DP and Schwartz GK. Phase I trial of the cyclin-dependent kinase inhibitor and protein kinase C inhibitor 7-hydroxystaurosporine in combination with Fluorouracil in patients with advanced solid tumors. J Clin Oncol 2005;23:1875–1884. Lara Jr. PN, Mack PC, Synold T, Frankel P, Longmate J, Gumerlock PH, Doroshow JH and Gandara DR. The cyclin-dependent kinase inhibitor UCN-01 plus cisplatin in advanced solid tumors: a California cancer consortium phase I pharmacokinetic and molecular correlative trial. Clin Cancer Res 2005;11:4444–4450. Herbst RS, Thornton DE, Kies MS, Sinha V, Flanagan S, Cassidy CA, Carducci MA. Phase 1 study of LY317615, a protein kinase Cb inhibitor. American Society of Clinical Oncology Annual Meeting 2002, Abstract 326. Fine HA, Kim L, Royce C, Draper D, Haggarty I, Ellinzano H, Albert P, Kinney P, Musib L and Thornton D. Results from phase II trial of enzastaurin (LY317615) in patients with recurrent high grade gliomas. J Clin Oncol 2005;23:1504. Edwards LA, Thiessen B, Dragowska WH, Daynard T, Bally MB and Dedhar S. Inhibition of ILK in PTEN-mutant human glioblastomas inhibits PKB/Akt activation, induces apoptosis, and delays tumor growth. Oncogene 2005;24: 3596–3605. White DE, Cardiff RD, Dedhar S and Muller WJ. Mammary epithelial-speciﬁc expression of the integrin-linked kinase (ILK) results in the induction of mammary gland hyperplasias and tumors in transgenic mice. Oncogene 2001;20:7064–7072. Yoganathan N, Yee A, Zhang Z, Leung D, Yan J, Fazli L, Kojic DL, Costello PC, Jabali M, Dedhar S and Sanghera J. Integrin-linked kinase, a promising cancer therapeutic target: biochemical and biological properties. Pharmacol Ther 2002;93:233–242. Hannigan G, Troussard AA and Dedhar S. Integrin-linked kinase: a cancer therapeutic target unique among its ILK. Nat Rev Cancer 2005;5:51–63. Yau CYF, Wheeler JJ, Sutton KL and Hedley DW. Inhibition of integrin-linked kinase by a selective small molecule inhibitor, QLT0254, inhibits the PI3K/PKB/mTOR, Stat3, and FKHR pathways and tumor growth, and enhances gemcitabine-induced apoptosis in human orthotopic primary pancreatic cancer xenografts. Cancer Res 2005;65:1497–1504. Tan C, Cruet-Hennequart S, Troussard A, Fazli L, Costello P, Sutton K, Wheeler J, Gleave M, Sanghera J and Dedhar S. Regulation of tumor angiogenesis by integrinlinked kinase (ILK). Cancer Cell 2004;5:79–90. Revets H, De Baetselier P and Muyldermans S. Nanobodies as novel agents for cancer therapy. Expert Opin Biol Ther 2005;5:111–124.

223 254. Cortez-Retamozo V, Backmann N, Senter PD, Wernery U, De Baetselier P, Muyldermans S and Revets H. Efﬁcient cancer therapy with a nanobody-based conjugate. Cancer Res 2004;64:2853–2857. 255. Parsons DW, Wang TL, Samuels Y, Bardelli A, Cummins JM, DeLong L, Silliman N, Ptak J, Szabo S, Willson JKV, Markowitz S, Kinzler KW, Vogelstein B, Lengauer C and Velculescu VE. Colorectal cancer: mutations in a signalling pathway. Nature 2005;436:792. 256. Clark DE, Errington TM, Smith JA, Frieson Jr. HF, Weber MJ and Lannigan DA. The serine/threonine protein kinase, p90 ribosomal S6 kinase, is an important regulator of prostate cancer cell proliferation. Cancer Res 2005;65:3108–3116.

225

Molecularly imprinted materials as advanced excipients for drug delivery systems Carmen Alvarez-Lorenzo and Angel Concheiro Departamento de Farmacia y Tecnologı´a Farmace´utica, Facultad de Farmacia, Universidad de Santiago de Compostela, 15782-Santiago de Compostela, Spain Abstract. The application of the molecular imprinting technology in the design of new drug delivery systems (DDS) and devices useful in closely related ﬁelds, such as diagnostic sensors or biological traps, is receiving increasing attention. Molecular imprinting technology can provide polymeric materials with the ability to recognize speciﬁc bioactive molecules and with a sorption/release behaviour that can be made sensitive to the properties of the surrounding medium. In this review, an introduction to the imprinting technology presenting the different approaches in preparing selective polymers of different formats is given, and the key factors involved in obtaining of imprinted binding sites in materials useful for pharmaceutical applications are analysed. Examples of DDS based on molecularly imprinted polymers (MIPs) can be found for the three main approaches developed to control the moment at which delivery should begin and/or the drug release rate; i.e., rate-programmed, activation-modulated or feedback-regulated drug delivery. This review seeks to highlight the most remarkable advantages of the imprinting technique in the development of new efﬁcient DDS as well as to point out some possibilities of adapting the synthesis procedures to create systems compatible with both the relative instable drug molecules, especially of peptide nature, and the sensitive physiological tissues with which MIP-based DDS would enter into contact when administered. The prospects for future development are also analysed. Keywords: drug delivery, controlled release, pharmaceutical excipients, hydrogels, stimulisensitive polymers, smart materials, imprinted particles, imprinted beads, molecular imprinting, rational design, imprinted gels, covalent imprinting, non-covalent imprinting, imprinting without template, contact lenses, trap systems.

Introduction Over the past few years a number of signiﬁcant advances have been made in the development of new technologies for optimizing drug delivery [1]. Drug delivery systems (DDS) must be capable of regulating the rate of release (delayed- or extended-release systems) and/or targeting the drug to a speciﬁc site, to maximize the efﬁcacy and safety of medicines. Efﬁcient DDS should provide a desired rate of delivery of the therapeutic dose at the most appropriate place in the body, in order to prolong the duration of pharmacological action and reduce the adverse effects, minimize the dosing frequency and enhance patient compliance. The three following approaches have been developed to control the drug release rate and the moment at which delivery should begin [2] (Fig. 1): (i) rate-programmed drug Corresponding author: Tel: +34-981-563100. Fax: +34-981-547148.

E-mail: [email protected] (C. Alvarez-Lorenzo). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12007-4

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

226

Fig. 1. The three major approaches to control the release from DDS. (Reproduced from Chien and Lin [2] with permission from Adis International Ltd.)

delivery: drug diffusion from the system has to follow a speciﬁc rate proﬁle;(ii) activation-modulated drug delivery: the release is activated by some physical, chemical or biochemical processes; and (iii) feedback-regulated drug delivery: the rate of drug release is regulated by the concentration of a triggering agent, such as a biochemical substance, concentration of which is itself dependent on the drug concentration in the body. When the triggering agent is above a certain level, the release is activated. This induces a decrease in the level of the triggering agent and, ﬁnally, the drug release is stopped. The sensor embedded in the DDS tries to imitate the recognition role of enzymes, membrane receptors and antibodies in living organisms for regulation of chemical reactions and for maintenance of the homeostatic equilibrium. Artiﬁcial polymer-based systems can be provided with such a capability to recognize speciﬁc molecules by using the molecular imprinting approach. This technology consists of preparing synthetic networks in the presence of the substance of interest with the aim of this acting as a template to create tailored cavities. Removal of template molecules results in sites (called cavities or pockets) within the polymeric network that are complementary, both sterically and chemically, to the original template molecule. As a consequence, once the polymer system again enters into contact with the target molecules, these speciﬁcally interact with the empty cavities [3,4]. The intensity of the interaction can be modulated by external factors such as pH, temperature or

227 concentration of competitive species. This provides materials able to imitate the behaviour of biological actuators (enzymes, antibodies, etc.), and that can be particularly useful when developing new drug delivery platforms with a performance tailored to the characteristics of certain diseases [5–8]. Continuous development of molecular imprinting technology makes it possible to obtain materials of various formats, with different physical and chemical properties, widening the scope of applications into quite diverse ﬁelds. Nowadays, molecular imprinting is a well-developed tool in the analytical ﬁeld, mainly for separating and quantifying a wide range of substances contained in relatively complex matrices [9–12]. The ability of the molecularly imprinted polymers (MIPs) to selectively absorb and retain target analytes has expanded their use as stationary phases in chromatographic columns [12] and as ﬁllers of solid phase extraction cartridges [13]. MIPs can also be used as traps of environmental contaminants [14]. A long list of the advantages of MIPs as replacements for biological antibodies in immunoassays has also been reported [15]. In the pharmaceutical ﬁeld, imprinted materials that mime biological receptors for the screening of new substances with potential pharmacological activity, or that detect speciﬁc drugs in biological ﬂuids in screening assays for drugs of abuse, are receiving enormous attention [16–18]. There is also a progressive increase in the number of papers and patents devoted to the application of MIPs in the design of new DDS [7,8,19] and also of devices useful in closely related ﬁelds, such as diagnostic sensors or chemical traps to remove undesirable substances from the body [20–22]. Although the possibilities of this technology for the pharmaceutical and biomedical applications are still in an exploratory stage, the enhancement of the afﬁnity of the MIPs for speciﬁc drugs and physiological substances has already been shown as an exciting and versatile way to increase drug loading and controlled release properties [5–8]. This review starts with a brief introduction to the concepts behind the molecular imprinting with the aim of explaining the basis of this technology. Readers interested in gaining an insight into this speciﬁc topic are recommended to consult some of the excellent, general overviews published recently [23–26]. The following sections provide up-to-date information about the different attempts to prepare MIPs, able to actuate under in vivo conditions, as delivery systems of drugs and other biologically active substances and, also, as diagnostic sensors or traps of undesirable substances. These approaches range from ‘‘classical imprinting’’, which refers to imprinted systems prepared in organic solvents with a high cross-linked proportion, to weakly cross-linked imprinted, and even stimuli-sensitive, hydrogels. Approaches to molecular imprinting In undertaking to create tailor-made cavities with high speciﬁcity and afﬁnity for a target molecule, synthetic polymer networks can be prepared, in the

228 presence of the substance of interest, using ‘‘functional’’ monomers, i.e., reactant monomers able to associate with the template. The functional monomer/template association can be achieved through: (i) covalent bonds or (ii) non-covalent interactions, such as ionic, hydrogen bond, hydrophobic or charge-transfer interactions (Fig. 2). In the pre-organized or covalent approach, introduced in the 1970s by Wulff and co-workers [3,27], the template is covalently bound to the monomers prior to polymerization. These bonds are reversibly broken after the synthesis of the networks to form the imprinted cavities. In the self-assembly or non-covalent approach, proposed in the 1980s by Arshady, Mosbach and co-workers [28,29], the template molecules and functional monomers arrange themselves, prior to polymerization, to form stable and soluble complexes of appropriate stoichiometry by noncovalent or metal coordination interactions. In the latter approach, multiplepoint interactions between a template molecule and various functional monomers are required to form strong complexes. In general, the non-covalent imprinting protocol allows more versatile combinations of templates and monomers, and provides faster bond association and dissociation kinetics

OH HO

O

OH O

HO

OH OH

O

O

O

Template

Polymerizable structure

Imprinted cavity

(A)

OH

hb

hb OH

hb

hb

HO

HO

O

Template

ii

OH ii

OH

hi

hi

O

Monomers arrangement

hi

hi

Imprinted cavity

(B)

Fig. 2. Schematic view of the imprinting process: (A) covalent approach, in which the

template is covalently bound to polymerizable binding groups that are reversible broken after polymerization; and (B) non-covalent approach, in which the template interacts with functional monomers through non-covalent interactions (e.g., ionic ii, hydrophobic hi or hydrogen bond hb) before and during polymerization.

229 than the covalent imprinting approach [5]. In either case, the preparation of MIPs requires the co-polymerization of the functional monomer-template complexes with high proportions of cross-linking agents and the subsequent removal of the template molecules, in order to create recognition cavities complementary in shape and functionality to the template molecules. An interesting alternative to the classical approaches is the imprinting without template. When the pure target analyte is expensive or difﬁcult to synthesize in sufﬁcient quantities for preparing signiﬁcant amounts of MIP, the replacement of the target molecule as the template by an analogue can be considered. For analytical purposes, the use of MIPs moulded against a structural analogue of the target analyte avoids problems related to incomplete template removal, which sometimes prevents the use of MIPs in quantitative trace analysis [30]. Another way to avoid the use of a template consists of using pairs of functional monomers directly bonded to each other. The functional groups are separated after polymerization to provide ionic groups of the same charge that are close together and are able to bind target molecules through multiple-point ionic interactions [31,32].

Key issues for a rational design Several factors have to be taken into account in order to achieve good performing imprinted pockets. First, it is essential to ensure that the template does not bear any polymerizable group that could attach itself irreversibly to the polymer network. It should not interfere in the polymerization process and should be stable at the moderately elevated temperatures or upon exposure to UV radiation used to synthesize the polymer network [33]. The functional monomers are responsible for creating cavities with high afﬁnity for the template molecules and, therefore, the chemical groups of these monomers should be matched to those of the template. Among the monomers used in imprinting protocols, the acrylate-, styrene- and silanebased ones are the most common. The acrylate networks can be rapidly formed by free-radical polymerization, and the template molecules usually do not interfere in the reaction [23,33]. For a rational design of the MIPs, a preliminary analysis of the interactions of the template molecules with the functional monomers, under the conditions at which the polymer synthesis will occur, should be carried out. Analytical techniques, such as Raman and NMR spectroscopies, UV spectrophotometry or microcalorimetry, can provide information about the efﬁciency of the complexation process and its stoichiometry [34–37]. Combinatorial and computational simulations are also useful as faster screening techniques [34,38–41]. These investigations help to shorten the time required to develop the MIPs, compared to the traditional trial-and-error approaches. The use of factorial designs or chemometric models for optimizing the synthesis procedure also contributes to the

230 simpliﬁcation of the screening of the most adequate compositions [42,43]. Knowledge about the functional monomers–template interactions is crucial to maximize the yield of the selective cavities, minimizing the excess of functional monomer that leads to the formation of nonselective sites on the polymer [36,44]. If a low number of stable complexes between the template and the functional monomer is formed or if a large excess of functional monomer is added (as traditionally recommended for the non-covalent approach), the number of cavities with medium-to-high binding afﬁnity becomes extremely low (0.5–1% of the theoretical binding sites) [45–47]. These cavities show a great heterogeneity in their binding properties (Fig. 3), which complicates the characterization of the MIP performance. A high binding afﬁnity is expected when the template–functional monomer complexes involve several functional monomer units, at the most adequate stoichiometric proportion, and the MIP is synthesized at lowest possible temperature [48,49]. Table 1 summarizes some analytical techniques helpful for the design of MIPs together with those useful to characterize their binding properties. Other two important factors with a strong inﬂuence on the conformation of the imprinted cavities are the solvent and the cross-linker proportion used to synthesize the polymer. The ﬁrst criterion in selecting the reaction medium is its capability to totally dissolve the products that participate in the process, such as the template, the functional monomers or the cross-linkers and the initiators, in order to achieve a complete polymerization. The solvent also determines the intensity of the interactions between the template and the functional monomers [29,50]. If a covalent imprinting method is applied, the 1.2

Log N (µmol/g)

1.0 0.8

0.36 mmol

0.6

0.14 mmol

0.4 0.07 mmol

0.2

0 mmol 0.0 2.5

3.0

3.5 Log K (L /mol)

4.0

4.5

Fig. 3. Afﬁnity distribution (AD) in number of binding sites (N) vs. afﬁnity constant (log K) for MIPs prepared with MAA (8.6 mmol) and EGDMA (62 mmol) and synthesized in the presence of different ethyl adenine-9-acetate amounts. (Adapted from Rampey et al. [47] with permission from the American Chemical Society.)

231 Table 1. Some analytical techniques useful for the rational design (at the pre-polymerization stage) and for characterizing the afﬁnity of the MIPs for the target molecules (after polymerization). Stage

Technique

Information

Pre-polymerization

NMR spectroscopy

Hydrogen bonding, selfassociation, p–p stacking, ionic interactions Complexation events Screening of suitable functional monomer for a given template Binding energy and stoichiometry of the complexes Polymer–template interactions Polymer–template interactions Rebinding energy Distribution of the imprinted cavities

UV spectroscopy Molecular modeling Isothermal calorimetry Post-polymerization

FT-IR spectroscopy Solid state NMR Isothermal calorimetry Transmission electron microscopy

covalent bonds between the template and the monomer/s are often highly stable in any medium. In contrast, the non-covalent imprinting requires much more care to be taken. In this case, competition between the solvent and the monomers for bonding to the template will, at best, compromise the imprinting effect and, at worst, eliminate it entirely. As a general rule, apolar solvents are used to increase the likelihood of formation of hydrogen bonds, while water or other polar solvents are preferred to cause the complexation by hydrophobic forces [51]. In practice, the formation of receptor sites is the result of a combination of hydrogen-bonding, ion-pairing and hydrophobic interactions between template and functional monomer/s. Therefore, the effect of solvents on the efﬁciency of non-covalent imprinting is not easy to predict. During the formation of the network, the solvent can work as a ‘‘porogen’’, ﬁlling spaces and creating pores in the bulk of the network. The nature and proportion of the solvent in the monomer soup, condition the shape of the pores, the pore size distribution and the total porosity of the network [50,52]. The use of organic solvents is limited by compatibility problems with some templates such as peptides, oligonucleotides or sugars; therefore, molecular imprinting in water is gaining more and more attention in the creation of MIPs for pharmaceutical applications. Difﬁculties arise in materializing the methodology because of the weakness of electrostatic and hydrogen-bonding interactions in this polar medium, which decrease the afﬁnity and selectivity of MIPs for the ligands [53]. A combination of hydrophobic interactions

232 (e.g., using cyclodextrins as functional monomers) and metal coordination can enhance the template–functional monomers association in water [53–55]. To obtain successfully imprinted networks, the cavities should have a structure stable enough to maintain the conformation in the absence of the template and, at the same time, be sufﬁciently ﬂexible to facilitate the attainment of a fast equilibrium between the release and re-uptake of the template in the cavity [27]. The conformation and the stability of these pockets are related to the mechanical properties of the network and depend to a great extent on the crosslinker proportion. Most imprinted systems require around 50–90% of crosslinker agent to resist mechanical stress and the chemical and enzymatic attacks, when they enter into contact with biological media [56]. The most usual crosslinkers, ethylene glycol dimethacrylate (EGDMA) and other related molecules, provide stable networks in vitro, in a wide range of pHs and temperatures [57]. However, additional research should be carried out since esterases and extreme pHs could catalyse its hydrolysis in vivo [58]. Reactivities should be similar to ensure the smooth incorporation of all comonomers (functional, non-functional and cross-linkers) [59]. These crosslinking levels increase the hydrophobicity of the network and prevent the polymer from changing the conformation adopted during synthesis. In consequence, the afﬁnity for the template is not dependent on external variables and it is not foreseen that the device will have regulatory or switching capabilities. The lack of response, through a change in polymer conformation, to the alterations of the physico-chemical properties of the medium or to the presence of a speciﬁc substance limits their potential uses as activationmodulated or feedback-modulated DDS. A high cross-linker proportion also considerably increases the stiffness of the network making the adjustment of the shape to the administration site difﬁcult and causing mechanical friction with the surrounding tissues (especially when administered by topical and ocular routes or as implant). The mechanical properties and the morphology of the polymers are determined by both the level of cross-linker and the nature and volume of the solvent used in the synthesis (Fig. 4) [51]. Macroporous monoliths are obtained with relatively high cross-linking ratios and in poorer solvents. Under these conditions, the polymer chains separate from the liquid phase as they grow, and produce a network of aggregated polymer chains with void spaces (pores) interconnected between themselves, which are permanent even in the dry state. If the high cross-linking ratio is maintained and the volume of the solvent is further increased, discrete polymer particles can be formed and recovered [60]. In contrast, gel-type polymers are obtained at relatively low cross-linker ratios in solvents compatible with the resulting polymer network [61]. Only recently, the molecular imprinting technology was adapted to create imprinted hydrogels [62]. The adaptability of the molecular imprinting technology to the drug delivery ﬁeld also requires the consideration of safety and toxicological concerns. The device will enter into contact with sensitive tissues; therefore, it should

233

Solvent or porogen (%)

Polymer particles

Macroporous monoliths Gel-type polymers

Cross-link ratio

Fig. 4. Inﬂuence of the volume of the solvent and amount of cross-linker on the mor-

phology of the MIPs: gel-type polymers, macroporous monoliths, polymer particles. (Adapted from Cormack and Elorza [51] with permission from Elsevier Science.)

not be toxic, neither should its components, residual monomers, impurities or possible products of degradation [63,64]. Therefore, to ensure biocompatibility it might be more appropriate to attempt to adapt the imprinting technique to already tested materials instead of creating a completely new polymeric system. The presence of residual organic solvents may cause cellular damage and should be the object of precise control. In consequence, hydrophilic polymer networks that can be synthesized and puriﬁed in water are preferable to those that require organic solvents. A hydrophilic surface also enhances biocompatibility and avoids adsorption of proteins and microorganisms [65]. Depending on the speciﬁc application of the device, an adequate balance between the performance such as imprinted systems – that determines the efﬁciency as DDS or biological sensor – and safety when administered should be reached. Only for applications in which the physiological aspects play a less important role, would it be possible to prepare the networks considering mostly their performance as imprinted devices. It is clear that the polymer composition and solvent are key parameters in the achievement of a good imprinting and that, as a consequence, a compromise between functionality and biocompatibility is needed. On the basis of these determining factors, a wide range of imprinted networks have been developed with the aim to be used as drug delivery platforms, as will be referred in the following sections. MIPs as basis of drug delivery systems Rate-programmed drug delivery Imprinted particles as DDS excipients MIPs prepared in organic solvents with a high cross-linker proportion have been proposed as base excipients for controlled-release devices of drugs with

234

Theophylline released (dose fraction)

1.0

0.8

0.6

0.4 50 mg/g 10 mg/g 0.2

2.0 mg/g 0.1 mg/g

0.0 0

100

200

300 400 Time (min)

500

600

Fig. 5. Theophylline release proﬁles in phosphate buffer pH 7 from imprinted pol-

ymers (theophylline:MAA:EGDMA 5.22: 20.9: 94.3 mmol in chloroform) loaded with different amounts of drug. (Reproduced from Norell et al. [67] with permission from Wiley.)

a narrow therapeutic index; i.e., with a small difference between the minimum blood level to be active and the level at which collateral effects advise against their use. Therefore, drugs such as theophylline have to be administered in devices able to control their release precisely [66]. With a view to oral administration of theophylline, Norell et al. [67] prepared non-covalent imprinted particles (65 mm) in chloroform, using the method of Vlatakis et al. [68]. These particles were able to sustain the release in pH 7.0 phosphate buffer for several hours (Fig. 5). The increase in release rate observed at high loadings is explained by a partial weak attachment to non-speciﬁc binding points. The slightly faster release showed by the reference (non-imprinted) systems also support this hypothesis. Avoiding the decrease in the strength of the interaction between the MIPs and the ligands in aqueous medium and to enhance their performance as sustaining release excipients of transdermal DDS, Allender et al. [69,70] proposed preventing water from associating with the imprinted binding site, by embedding the MIP and the drug within a secondary polymer matrix made of a commercially available non-polar transdermal adhesive. The adhesive material – freely diffusible for drug molecules, but relatively hydrophobic – was able to create an environment within which selective binding could occur. Transdermal devices were prepared dispersing in chloroform propranolol (19.1 mg) together with imprinted or non-imprinted polymer (100, 300 and 500 mg), and then mixing with the self-curing acrylic copolymer adhesive. The viscous dispersions were left to cure overnight and,

235

Propranolol diffused (µg/cm2)

100

Control 100mg NIP 100mg MIP 300mg NIP 300mg MIP 500mg NIP 500mg MIP

80

60

40

20

0 0

5

10 15 Time (hours)

20

25

Fig. 6. Inﬂuence of polymer content and imprinting effect on propranolol release

from 1-cm diameter discs constituted by imprinted (MIP) or non-imprinted (NIP) polymers (MAA:EGDMA 6:30 mmol in chloroform) embedded in a non-polar transdermal adhesive. (Reproduced from Allender et al. [70] with permission from Elsevier Science.)

then, cut in 1 cm-diameter discs containing 0.5 mg propranolol. Propranolol diffusion studies carried out in water:ethanol (50:50) mixture showed that the devices containing MIPs can control the release rate (Fig. 6), which indicates that the speciﬁc binding characteristic of these systems can provide a useful means of sustaining the delivery proﬁle. MIPs for tetracycline have also been developed using methacrylic acid (MAA) as functional monomer [71]. The MIP particles prepared with 80 mol% cross-linker loaded substantially greater amounts of drug and sustained the release for longer than the corresponding non-imprinted particles. Theophylline nanospheres with modulated loading and controlled release capacities can be obtained choosing adequate MAA/methylmethacrylate (MMA) ratios [72]. The imprinted nanospheres were incorporated in the bulk or on the surface of poly(MMA-co-AA) membranes to increase the loading of theophylline in aqueous media and the selectivity of the binding [73]. The observed increase in the recognition factor (from 1.02 to 6.32) suggests that the incorporation of the particles to the membranes enhances the quality of microenvironment inside the cavities and, therefore, their rebinding ability. MIPs are also promising as enantioselective release excipients. A growing awareness of the often profoundly different pharmacokinetic parameters–oral bioavailability, clearance, protein binding, elimination half-life–and also pharmacodynamic and toxicological properties of the enantiomers of chiral substances, has motivated strict regulations for their use as active components in drug products [74]. Large differences in activity and toxicity

236 justify the necessity of administering pure single enantiomers. However, their production is still difﬁcult and time-consuming, and enantiopure drugs may suffer racemization during pharmaceutical processing, storage and in vivo dissolution [75]. In consequence, racemates are still being widely used. The ability of MIPs to recognize, in organic media, subtle structural differences that occur in enantiomers is well known in the analytical ﬁeld [76,77], although the efﬁciency seems to be more limited in water. Suedee and coworkers [78,79] have carried out a detailed analysis of the potential of MIPs in the enantioselective-controlled delivery of a b-blocker and two nonsteroidic antiinﬂammatory drugs (NSAIDs), when incorporated in granules and tablets. The enantiomers R-propranolol, S-ibuprofen and S-ketoprofen were used as templates. 4-Vinylpyridine (VPy) was the functional monomer for NSAIDs, and MAA in the case of propranolol. EGDMA was used as cross-linker. Spherical polymer particles were prepared from a microemulsion of the monomers by multistep swelling of polystyrene seeds, followed by polymerization for 24 h at 501C. The particles (10–30 mm) were washed and dried under vacuum. Granules, based on either imprinted or non-imprinted polymers, were obtained by mixing racemic drug and polymer particles with a polyvinylpyrrolidone (PVP) solution (20% PVP in ethanol). The release in water of the enantiomer used as the template was, in all cases, slower than that of the non-imprinted one. The granules prepared with the non-imprinted polymers showed that whatever the drugs used, similar release rates were observed for both enantiomers. The imprinting effect on the enantioselective release was especially evident when high MIP/drug ratios (15–25) were chosen to prepare the granules. This means that the ability of the system to selectively release the non-imprint molecule increases as more imprinted cavities are available, since the imprint molecule ﬁnds it more difﬁcult to escape from them. When the enantioselectivity values were estimated, as enantiomeric excess release percentage, using the equation: %ee ¼ ½ðS isomer Risomer Þ=ðRisomer þ Sisomer Þ þ 100 in which Sisomer and Risomer represent the amounts of each isomer released at a given time, the results obtained indicated that the granulation of a mix of S-ibuprofen MIP and S-ketoprofen MIP particles (combined MIP granules) provided a higher release rate of R-ibuprofen and R-ketoprofen, and a lower release rate of S-ibuprofen and S-ketoprofen. In consequence, these granules showed a greater enantioselectivity (%ee for ibuprofen ¼ 63.6%; %ee for ketoprofen ¼ 13.5) than that shown by those prepared with each single MIP granule (%ee for ibuprofen ¼ 43.5; %ee for ketoprofen ¼ 10.0). The procedure used to prepare tablets containing MIPs as main excipients, strongly determines the release rate. Tablets prepared after wet granulation of a mixture of racemic propranolol and polymer particles, with an ethanol

237 solution of PVP and hydrogenated vegetable oil, released the print molecule faster than the non-imprint enantiomer. The low retention of the enantiomers is explained by two causes: a high rate of water uptake by the matrix and an inadequate restoring of the imprinted cavities. When the tablet enters into contact with the dissolution medium, water dissolves the drug at the same time as the polymer is hydrated. If drug dissolution occurs before the polymer is ready to recognize the drug molecules, no delay in the release is observed. Furthermore, since these MIP granules swelled considerably in water, the conformation of the imprinted cavities can be altered, decreasing their afﬁnity for the imprint molecule. Tablets with a core of racemic propranolol and an outer layer of granules of MIP/PVP covered by a ﬁlm of hydrogenated vegetable oil were also assayed [80]. Only once the shell of these tablets becomes completely hydrated, does the medium penetrate into the core and start to dissolve the drug. The release of the drug starts only after this lag time when the drug can diffuse outward. The drug has to pass through the hydrated polymer before leaving the matrix, and this process may enhance any possible interaction of the enantiomer with the MIP. Therefore, the tablets released faster the nonprint enantiomer, for which the polymer had a lesser afﬁnity. After some days in pH 7.4 phosphate buffer, the differences in release rate decreased and even reversed and as a consequence of that the ﬁnal swelling of the tablets causes a distortion of the cavities, promoting the release of the imprint enantiomer (Fig. 7). 1.6 S-imprinted

R/S ratio

1.4

1.2

1.0

0.8 R-imprinted 0.6

0

24

48

72 96 120 Time (hours)

144

168

192

Fig. 7. Time-course of the relative percentage of R-propranolol and S-propranolol

enantiomers released from tablets made of a racemic propranolol core and a lowswelling imprinted polymer shell (template:metacrylic acid:EGDMA 3:12:310 mmol in chloroform). (Reproduced from Suedee et al. [80] with permission from Marcel Dekker, Inc.)

238 The ﬁndings regarding the use of MIPs as enantioselective excipients are quite encouraging and suggest that the transition of the original polymer from the dry to the swollen state and the possible loss of conformation at the active sites, are key factors for efﬁcient drug release behaviour. Nevertheless, to achieve systems with optimum enantioselectivity more work should be done. Polymers imprinted in water as DDS Imprinting of peptides and proteins. In molecular imprinting, relatively low molecular weight compounds are generally used as templates while the imprinting of macromolecules is less common. The synthesis of MIPs selective to big structures, such as proteins, is mainly hindered by steric (bulky protein cannot easily move in and out through the mesh of a polymer network [81]) and thermodynamic reasons. The ﬁrst attempts to overcome these limitations have been focused on synthesizing macroporous MIPs [82] or creating imprinted cavities at the surface of the network using metal (Cu2+)-ligand monomers [83]. Hydrogels imprinted for haemoglobin (Hb), mainly intended for analytical purposes, have been recently developed into four different formats: (i) polyacrylamide hydrogels prepared after dissolving the protein in an acetic acid (10% v/v)/sodium dodecylsulphate (10% w/v) solution [84]; (ii) chitosan beads immersed in acrylamide and Hb solutions in pH 6.8 phosphate buffer that were then polymerized [85]; (iii) semi-interpenetrated polymer networks based on polyacrylamide and chitosan [86]; (iv) covalently imprinted silica surfaces [87]. Taking into account that the use of large non-rigid templates, such as polypeptides and proteins, yields less well-deﬁned recognition sites [88], an approach focussed on imprinting only a short part of the macromolecule (called epitope) has been developed [89] (Fig. 8). The MIPs are intended to recognize such a portion of amino acids in any protein, as the antibodies recognize speciﬁc sequences in macromolecular antigens. For example, a sequence of four amino acids (Tyr-Pro-Leu-Gly) can be chosen as the template of oxytocin (Cys-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly-NH2) using MAA as functional monomer, EGDMA as cross-linker (monomer to cross-linker ratios ranged from 1:30 to 1:7.5), and water/acetonitrile mixture as polymerization medium. The imprinted macroporous polymer efﬁciently recognized both the template and the whole protein, even in pH 6.5 aqueous medium. Using a similar methodology, MIPs for the octapeptide [Sar1, Ala8] angiotensin II, a competitive inhibitor of the peptide hormone angiotensin II [90], and MIPs for a 15-mer peptide of dengue virus protein [91] have been successfully developed. In general, the high selectivity and afﬁnity of the MIPs for peptides and proteins obtained make them potentially useful for the development of DDS with a high loading capacity and able to control the release of these macromolecules in speciﬁc physiological environments.

239 A Polymerization mixture

Functional monomers

B

+ Cross-linking agent C

D

Porogen, initiator Polymerization

Formation of a recognition site around a short peptide

A

B C

D

Removal of template

Free recognition site

Rebinding

A Recognition of a larger peptide

B C D

Fig. 8. Schematic representation of the imprinting process of a peptide using the

epitope approach. (Reproduced from Rachov and Minoura [89] with permission from Elsevier Science.)

Imprinting with cyclodextrins. A new molecular imprinting technique based on the ability of cyclodextrins to form, in water, inclusion complexes with relatively hydrophobic drugs has been introduced and optimized over the last few years by Asanuma, Komiyama and co-workers [53,92]. Cyclodextrins are non-toxic cyclic oligomers of 6–9 glucopyranose units, which present a hydrophilic outer surface and a non-polar cavity of 5–8 A˚ inner diameter and 7 A˚ depth [93]. Complex formation depends on the size of the hydrophobic region of the drug, on the dimensions of the cavity, and on the intensity of the interactions between the chemical groups of both. The strategy consists of assembling several host cyclodextrins in a polymer network, with a spatial

240 distribution that allows each to ﬁt a designated portion of the target molecule and, at the same time, the assembly as a whole to recognize the guest molecules – sized in a range from angstroms to nanometers – exclusively (Fig. 9). Receptors for various peptides having two or more hydrophobic residues and polymers for large antibiotics have been successfully obtained applying this procedure [94]. The size and the complex structure of vancomycin and cefazolin made the simultaneous interaction with several cyclodextrins possible and, as a consequence, the imprinted polymers showed a considerably greater binding constant than the non-imprinted ones (630/M vs. 240/M for vancomycin; 320/M vs. 140/M in the case of cefazolin). MIPs for cefazolin were not able to bind vancomycin, which emphasizes the speciﬁcity of the imprinting procedure. The afﬁnity for the network should also regulate drug release. It has been reported that although stronger the drug-cyclodextrin binding constant, the slower the dissociation kinetics, the release of the drug from common cyclodextrin complexes is practically instantaneous when the complexes are diluted in aqueous ﬂuids [93,95]. Therefore, cyclodextrin solutions do not provide, in general, a sustained release. However, in the case of formulations based on cyclodextrin hydrogels, the dilution phenomena are minimal since the cyclodextrins are covalently attached to each other and the volume of water that can enter the hydrogel is limited by its own network.

Fig. 9. Molecular imprinting of cyclodextrins as receptors of nanometer-scaled tem-

plates. (a) The cyclodextrin is cross-linked by diisocyanate, in dimethylsulfoxide, in the presence of the template. (b) The vinyl monomer of cyclodextrin is copolymerized with methylenebisacrylamide, in water, in the presence of the template. (Reproduced from Asanuma et al. [92] with permission from Wiley-VCH Verlag.)

241 This should provide a microenvironment rich in cavities available to interact with the surrounding drug molecules and, as a consequence, cyclodextrin hydrogels can be useful as a basis of delivery systems of peptides and drugs for different routes [96–98]. The combination of cyclodextrins and ionic monomers could open the application scope of these materials to amphiphilic molecules containing hydrophobic and ionic groups [54]. Weakly cross-linked MIPs prepared without solvents as drug delivery soft contact lenses Some drugs may be directly dissolved in monomer mixtures when at least one is at liquid state, without the need of using additional solvents. The feasibility of using hydrocortisone-imprinted polyhydroxyethylmethacrylate (PHEMA), prepared by dissolving the drug (250 mg) in a mixture of HEMA (3 mL) and cross-linker EGDMA (3.5 mL), as sustained-release material has been recently proved by Sreenivasan [99]. The loading capacity of the MIP was almost 40 times greater than that of the non-imprinted polymer (9.20 mg/g vs. 0.23 mg/g).Furthermore, in methanol/water medium, this latter system released all loaded hydrocortisone in less than one day, while the MIP released it in a considerably slower way; after one month, nearly 96% of hydrocortisone was still inside the network. This high afﬁnity, which prevents the release of an important fraction of loaded dose, limits the interest of this MIP with a view of their use as therapeutic device. Nevertheless, imprinted PHEMA networks without using solvents offer interesting possibilities for the development of DDS, if prepared with a less degree of cross-linking. Additionally, from the point of view of the biocompatibility, HEMA is a very adequate monomer when preparing hydrogels for long stays in contact with biological tissues, such as implants and contact lenses [100]. Our group has developed an imprinted technique to create soft contact lenses able to load and release drugs in a controlled way for the treatment of ocular pathologies [101–103]. Soft contact lenses are made of thin ﬂexible sheets of polymer hydrogels; the composition and format are chosen to make them highly compatible with the ocular surface in addition to correcting impaired vision. The possibility of both correcting the vision and acting as a DDS could remarkably enhance the beneﬁts of soft contact lenses. Conventional ophthalmic drug products generally show extremely low local bioavailability because of the ocular protective mechanisms, such as tear drainage and blinking. The efﬁcacy of drops and ointments can be improved by viscosity enhancers, in situ gelling systems and mucoadhesive polymers, although, even applying these approaches, the permanence time of drug on the eye is still quite limited [104]. Medicated contact lenses may be particularly useful for increasing drug bioavailability [105,106]. The feasibility of using drug-loaded soft contact lenses depends on whether the drug and the hydrogel material can be matched so that the lens uptakes a sufﬁcient quantity of drug and releases it in a controlled fashion. In general, drug-loading capacity of

242 conventional soft contact lenses is insufﬁcient and, therefore, they have rarely been employed for ophthalmic drug delivery [106,107]. The application of the molecular imprinting technology has been shown to be particularly useful in overcoming this drawback. Timolol, a drug used in glaucoma therapy, is a suitable molecule for providing imprinted systems since it offers multiple sites for the interaction with the functional monomer MAA. This monomer can interact through ionic and hydrogen bonds with timolol before polymerization in the HEMA solution. Since the lenses are assayed and used after swelling in aqueous medium, a signiﬁcant deformation of the cavities may be expected when compared to the anhydrous state upon polymerization. In the presence of timolol, only the high afﬁnity cavities would be able to recover the same conformation as that before swelling (‘‘induced ﬁt’’). These cavities will be able to load the drug and, subsequently, to sustain the release process. This behaviour was experimentally evidenced with imprinted lenses based on HEMA or N,N-diethylacrylamide (DEAA), and synthesized with different proportions of MAA (1.28–5.12 mol%) and cross-linker EGDMA (0.32–8.34 mol%), which were able to take up more timolol than the corresponding non-imprinted ones (Fig. 10). Some of these lenses, in particular those with the lowest MAA proportion and the highest cross-linking degree, could load a therapeutic dose of timolol, sustain its release in lachrymal ﬂuid for more than 12 h and reload another dose overnight, ready for use the next day [101,102,108]. Owing to the requirements of optical clarity, ﬂexibility and oxygen permeability, the main composition of soft contact lenses is restricted to some approved monomers that differ in water afﬁnity and hydrogen-bonding capacity [100]. Furthermore, the proportions of functional monomer and the cross-linking agent have to be relatively low and, in consequence, the physical

Fig. 10. Timolol uptake from water by imprinted and non-imprinted N,N-diethylacryl-

amide-based contact lenses made with MAA (100 mM) and different proportions of EGDMA. (Reproduced from Hiratani et al. [102] with permission Elsevier Science.)

243 stability of the binding sites is a main concern. Therefore, to try to generalize the applicability of the molecular imprinting technology to manufacture therapeutic contact lenses, the inﬂuence of the backbone monomers was analysed keeping the proportions of the functional monomer and cross-linker constant [109]. Four types of lenses were prepared by UV irradiation of DEAA, HEMA, 1-(tristrimethyl-siloxysilylpropyl)-methacrylate (SiMA) and N,N0 -dimethylacrylamide (DMAA) (50:50 v/v), or MMA and DMAA (50:50 v/v) solutions, to which the functional monomer MAA (100 mM), the crosslinker EGDMA (140 mM) and timolol maleate (25 mM) were previously added. Non-imprinted systems were also synthesized in the same way, but without timolol. Regarding the timolol overall afﬁnity, SK, the lenses ranked in the following order HEMA>SiMA–DMAA>MMA–DMAA>DEAA. The highest imprinting effect (i.e., the greatest relative increase in SK with regard to non-imprinted systems) was obtained for the MMA–DMAA and DEAA lenses. These results may be explained by both the interaction capability with timolol and the conformational features of the lenses. Regarding potential interactions with timolol molecules, HEMA monomers have an important hydrogen-bonding capability (greater than that of the other backbone monomers) to interact with timolol. Although HEMA lenses (without MAA) do not signiﬁcantly load this drug, the presence of HEMA around the MAA mers may contribute to create an adequate microenvironment to enhance the interactions with timolol, chieﬂy in the imprinted systems. This explains the greatest afﬁnity of both non-imprinted and, especially, imprinted HEMA-based lenses for timolol compared to the other lenses. On the other hand, the slightly greater cross-linker molar ratio of SiMA–DMAA lenses provides a greater stability of the conformation of the imprinted sites. As SiMA–DMAA lenses are also quite hydrophilic, the access of timolol molecules to the binding sites is relatively easy. In consequence, these lenses ranked second with the more loading capacity. MMA–DMAA hydrogels showed the highest swelling capacity. This causes the network to have large pores, making timolol diffusion easier, but also disrupting the initial conformation of the network. In these hydrated non-imprinted lenses, MAA groups are too far apart to gather to form binding sites for timolol. In contrast, the imprinting procedure provides cavities made of groups of MAA close together in space, considerably increasing the loading capacity of the lenses. Finally, DEAA-based lenses are quite hydrophobic at 37oC and have, % in general, a slight afﬁnity for timolol. For these lenses the imprinting effect on drug loading capacity becomes particularly relevant. Although all lenses studied showed sustained release in 0.9% NaCl solution, the values of diffusion coefﬁcients conﬁrmed that timolol molecules move out easily from hydrophilic networks with low afﬁnity for the drug (i.e., MMA–DMAA and SiMA–DMAA) (Fig. 11). The control-release ability of these imprinted hydrogels has been proved in vivo, by monitoring timolol levels in lachrymal ﬂuid after the insertion of

244 Imprinted

Non-imprinted

Timolol released (µg)

200

150 DEAA MMA-DMAA SiMA-DMAA HEMA

100

50

0

0

2

4 6 Time (hours)

8

10

0

2

4 6 Time (hours)

8

10

Fig. 11. Timolol release in 0.9% NaCl solution at 371C from reload imprinted and

non-imprinted lenses made with different backbone monomers. Drug diffusion coefﬁcients through the imprinted lenses were 2.2 109 cm2/s for N,N-diethylacrylamide (DEAA) systems, 9.9 109 cm2/s for HEMA systems, 66.5 109 cm2/s for MMA-co-N,N0 -DMAA (MMA-DMAA) systems and 71.3 109 cm2/s for 1(tristrimethyl-siloxysilylpropyl)-methacrylate-co-N,N0 -DMAA (SiMA-DMAA) systems. (Reproduced from Hiratani and Alvarez-Lorenzo [109] with permission from Elsevier Science.)

imprinted and non-imprinted loaded lenses and instillation of eyedrops [1 1 0] (Fig. 12). A second relevant factor to optimize the performance of the imprinted hydrogels as DDS is the template/functional monomer molar ratio, which determines the afﬁnity of the cavities for the drug. In a recent paper, it was shown that timolol release rate, from N,N0 -DMAA and tris(trimethylsiloxy)sililpropyl methacrylate (TRIS) hydrogels, strongly decreased by increasing the MAA/timolol ratio in the gel recipe [1 1 1]. Hydrogels prepared with 400 mM MAA, 600 mM EGDMA and a timolol/MAA mole ratio of 1/16–1/32 showed drug diffusion coefﬁcients in two orders of magnitude below those from non-imprinted hydrogels. This strong inﬂuence of the timolol/MAA ratio is related to differences in conformation of the imprinted cavities [46,47]. If a large amount of timolol is present during polymerization, binding sites for timolol can be created but these cavities may not have all the MAA units needed to fulﬁl the interaction capacity of the drug molecule. Therefore, the binding sites for timolol show a weaker afﬁnity for the drug and it can be easily released. In contrast, with a smaller amount of timolol in the medium, more MAA units are available to gather during synthesis to form efﬁcient imprinted cavities, which will have great multiple-point binding constants; that is to say, each binding site is more perfectly constructed and it can more effectively retain the drug, hindering the release process (Fig. 13).

245 500

Timolol conc. (mM)

400 300 200 100 0 0

30

60 90 120 Time (min)

150

180

Fig. 12. Timolol tear ﬂuid concentration-time proﬁles, on rabbits’ eyes, after appli-

cation of timolol loaded imprinted and non-imprinted contact lenses (’ and &, respectively), and instillation of timolol eyedrops (0.068% J; 0.25% K). The doses were 34 mg for imprinted contact lens, 21 mg for non-imprinted contact lens, 34 mg for 0.068% timolol eyedrop and 125 mg for 0.25% timolol eyedrop. Each point represents the mean7S.D. (n ¼ 3–5). (Reproduced from Hiratani et al. [110] with permission from Elsevier Science.)

1.0

Timolol released

0.8 Non-imprinted 1/4 1/8 1/16 1/32

0.6

0.4

0.2

0.0

0

12

24

36 48 Time (hours)

60

72

Fig. 13. Timolol release proﬁles, in 0.9% NaCl at 251C, from non-imprinted and

imprinted hydrogels prepared with MAA (400 mM) and EGDMA (600 mM) in the presence of different proportions of timolol in order to achieve the template: functional monomer molar ratios shown in the ﬁgure. Data are expressed as mean 7 S.D (n ¼ 5). (Reproduced from Hiratani et al. [111] with permission from Wiley Interscience.)

246 As a whole, the investigations carried out until now show that drug-loading capacity and release proﬁles of the soft contact lenses can be modulated by incorporating adequate functional monomers and by applying the molecular imprinting technology. This provides interesting possibilities for the improvement of ocular drug delivery. Additional studies with drugs belonging to various therapeutic groups and different chemical structures are being carried out in our laboratory to extend the scope of application of the imprinted lenses as therapeutic devices. Activation-modulated drug delivery This section is devoted to a group of systems potentially useful to develop DDS able to release drugs in response to environmental changes. These changes can affect the binding of the drug to the network in a direct way (by competitive binding or by hydrolysis of labile bounds) or indirectly through a change in the swelling state of the polymer, such a volume phase transition being induced by an external stimulus. Competitive binding Activation-modulated delivery may be achieved with an imprinted gel that releases the drug because of the competitive binding of another substance to the polymer. If a non-imprint drug is loaded in the network and the imprint molecule appears in the medium, the network exchanges the drug for the print molecule. When a fall in the concentration of the free imprint substance in the medium occurs, the release stops. A system of these characteristics able to release testosterone at a rate depending on the concentration of hydrocortisone in the medium was described by Sreenivasan [112]. To prepare such a system, HEMA (1 g) was cross-linked with EGDMA (4 g) and imprinted for hydrocortisone (100 mg) in chloroform (6–8 mL). This MIP absorbed, after removing hydrocortisone, a considerable amount of testosterone (175 mg/ 100 mg vs. 36 mg/100 mg control polymer). The release of testosterone to the aqueous medium was considerably enhanced in the presence of hydrocortisone (Fig. 14). Competitive binding behaviour in an aqueous medium was also achieved with particles imprinted for bupivacaine, when loaded with other local anaesthetic drugs [113]; with particles imprinted for theophylline loaded with caffeine or theobromine; and with particles imprinted for 17-b-estradiol, loaded with other structurally related sterols (17-aestradiol, 17-a-ethynylestradiol) [114]. Following this same principle, imprinted thin layers with potential applications in drug delivery have been recently developed for separation and sensing technologies. Basically, these layers have the ability to change their permeability when the template is in the medium, resembling their behaviour to that of a cell membrane with receptors and channels. Nanometer-ordered layer of theophylline-imprinted copolymer of EGDMA and MAA grafted

247 180

Testosterone released (µg)

150 With hydrocortisone Without hydrocortisone

120 90 60 30 0 0

2

4

6 8 Time (hours)

10

24

Fig. 14. Effect of the presence of hydrocortisone in the aqueous medium (50 mg/L)

on testosterone release rate. The initial load of testosterone was 175 mg in 100 mg of hydrocortisone-imprinted polymers (HEMA: ethylenglycoldimethacrylate 1:4 g). (Data taken from Sreenivasan [112]; reproduced from Alvarez-Lorenzo and Concheiro [6] with permission from Elsevier Science.)

onto an indium-tin oxide electrode [115] or cellulosic dialysis membrane [116] showed a faster release of creatinine in the presence of theophylline. The most interesting phenomenon observed in these studies is the ability of the MIP membranes to change its diffusive permeability automatically by responding to the template. Therefore, they could be applicable in developing delivery systems with molecular recognition that exclusively release the drug when a speciﬁc substance appears in the medium. Hydrolytically induced drug release A particularly useful approach to modulate drug delivery consists of creating erodible systems in which the drug cannot be released unless the polymer degrades or polymer/drug bonds break. The external conditions that induce these processes can be, for example, an extreme physiological pH or the catalytic activity of an enzyme. For example, drugs that are unstable under the gastric conditions may be selectively released in the colon by incorporating them into polymers that serve as substrates of the enzymes of this intestinal region or that degrade at slightly alkaline environments [117]. In other cases, the rate of hydrolysis of the drug linkage to the polymer network controls the release rate. However, ester bonds are usually broken at pH values 10–11 that are much more alkaline than those found in the physiological environment. Additionally, the presence of electron donating groups in the drug molecule (p-metoxy or p-amino) considerably suppresses the rate

248 of hydrolysis [118]. To enhance the hydrolysis of polymer/drug ester or amide bonds under mild pH conditions, Karmalkar et al. [119] proposed incorporating imidazole groups (which are nucleophilic catalyst) near the drug linkage, applying the molecular imprinting technology. The hydrogels, designed for the release of p-amino benzoic acid, were prepared dissolving HEMA, N-vinylimidazole (NVIm) and 2-methacryloylethyl p-aminobenzoate (PAP) in methanol, in the proportions indicated in Table 2. Imprinted hydrogels were only obtained when Co2+ ions were added to the PAP and NVIm solution. The metallic ions bring together both monomers, forming a coordination complex as shown in Fig. 15. Polymerization of such a complex and subsequent removal of the metal ion would lead to polymers having the labile bond and imidazole located in contiguous positions on the same chain. In

Table 2. Composition (in weight) of hydrogels used for selective release p-amino benzoic acid, in slightly alkaline pH medium, by a hydrolytically induced mechanism. To enhance the release rate, imprinted cavities containing imidazole groups and the monomeric drug were created using Co2+ coordination interactions [119]. Polymer

HEMA

PAP

NVIm

CoCl2 6H2O

PAP-1 PAP-2 PAP-3

4.22 4.30 4.50

0.5 0.5 0.5

0.17 0.17 0

0.106 0 0

EGDMA, AIBN O

O

O

65°C

N

O

O

N

O O

O N

N O

O

HO

O

O

HO

Co2+ NH2

NH2

Fig. 15. Synthetic route proposed by Karlmakar et al. [119] for creating imprinted

cavities in which the imidazole catalytic group is positioned close to the drug-polymer bound. The procedure consisted on the preorganization of the monomeric drug with a metal-complexing monomer and subsequent polymerization. (Reproduced from Alvarez-Lorenzo and Concheiro [6] with permission from Elsevier Science.)

249 ethanol/pH 8 phosphate buffer medium, the release of p-amino benzoic acid from the imprinted system (PAP-1) was considerably easier (rate constant 39.8 103/day) than from the other two gels (PAP-2: 8.1 103/day; PAP3: 6.1 103/day) and even faster than from PAP-3 gels in 0.01N NaOH (pH 11). The cross-linking was essential for the NVIm to behave as catalyst; the release rates from linear imprinted and non-imprinted polymers being similar. To bring into close proximity imidazol and drugs that cannot form a complex with the metal ion, another technique was developed [119,120]. The polymerization method, based on that previously proposed by Leonhardt and Mosbach [121], consisted of using template molecules of size and structure similar to those of the drug, but able to form the complexes. After polymerization and removal of the template, the drug was loaded into the hydrogel and immobilized by polymerization. Since the cavity contained an imidazol group, the release rate was considerably enhanced. Additionally, these polymers had the ability to switch on and off the release in response to a change in pH from 3 to 6.8 [120]. This is because the catalytic activity is lost at pH below 3.5 and, consequently, the release does not occur while the system is kept in that environment. The switch on and off effect can also be achieved by a change in the distance between the imidazole group and the substrate molecule induced by a temperature-phase transition of the polymer [122]. In the DDS ﬁeld, the main interest of this group of MIPs lies in their potential utility for targeting drugs, when orally administered, to speciﬁc regions of the gastrointestinal tract. Stimuli-sensitive imprinted hydrogels Stimuli-sensitive polymer gels that modify their structure and, as a consequence, their properties in response to changes in the physicochemical characteristics of the physiological medium are excellent candidates in achieving an optimum control of the moment and the rate of drug release [123,124]. These materials undergo enormous volume changes triggered by small modiﬁcations in temperature, pH, solvent composition, ionic strength, electric ﬁeld, light or the presence of speciﬁc molecules [125]. The combination of stimuli-sensitivity and imprinting may have considerable practical advantages; the imprinting provides a high loading capacity of speciﬁc molecules, while the ability to respond to external stimuli contributes to modulating the afﬁnity of the network for the target molecules, providing a regulatory or switching capability of the loading/release processes. Poly(N-isopropylacrylamide) (PNIPA) is a hydrogel that undergoes a temperature-controlled volume phase transition at 331C and that has been widely studied with pharmaceutical aims [125,126]. Tanaka and co-workers [62,127–130] proposed the creation of stimuli sensitive gels able to recognize and capture target molecules using polymer networks consisting of at least two species of monomers, each having a different role. One forms a complex

250 Template

Swollen state

Collapsed state

Fig. 16. Diagram of the recognition process of a template by a stimuli sensitive

imprinted hydrogel as proposed by Tanaka et al. [127]. The volume phase transition of the hydrogel – induced by an external stimuli such as a change in pH, temperature or electrical ﬁeld – modify the relative distance of the functional groups inside the imprinted cavities. This alters their afﬁnity for the template. (Reproduced from Alvarez-Lorenzo and Concheiro [6] with permission from Elsevier Science.)

with the template (functional monomer), and the other allows the polymers to swell and shrink reversibly in response to environmental changes. The hydrogel is synthesized in the collapsed state (temperature>341C) and, after polymerization, swelled in a washing medium. When the functional monomers come into proximity, the imprinted cavities develop afﬁnity for the template molecules, which diminishes when they are separated. The distance between functional monomers is controlled by the reversible phase transition that, consequently, controls the adsorption/release of the template (Fig. 16). Imprinted and non-imprinted copolymers of N-isopropylacrylamide (NIPA) and different methacrylic monomers with carboxyl groups – which form complexes with divalent ions in the relation 2:1 – were prepared in order to obtain a system with the capacity to recognize calcium ions. The effect of temperature on the adsorption capacity of the copolymers synthesized in different organic solvents and imprinted in the presence of different templates was analysed. Successful imprinting was obtained with NIPA-lead dimethacrylate monomers in dioxane. After washing lead out and swelling in water at room temperature, the afﬁnity for divalent ions disappeared. When the hydrogels were shrunken by an increase in temperature, the afﬁnity was recovered and the original relative position of the carboxylic groups was recalled. Control hydrogels with randomly distributed MAA experienced difﬁculty in forming pairs (‘‘frustration’’) and their afﬁnity for divalent ions decreased exponentially as a function of the cross-linker concentration (Fig. 17). In contrast, the topological constraints were completely absent in the imprinted gels, showing that molecular memorization had been achieved. The success of the imprinting can be attributed to

251

1000

Overall affinity (SK)

800

600

400 Imprinted Non-imprinted 0

50

100 150 Cross-linker (mM)

200

250

Fig. 17. Inﬂuence of the cross-linker (methylenebis-acrylamide) proportion on the

overall afﬁnity for calcium ions of the imprinted and non-imprinted NIPA (6 lM) gels in the shrunken state in water. The amount of functional monomers (MAA) was ﬁxed at 32 mM. (Reproduced from Alvarez-Lorenzo et al. [62] with permission from American Chemical Society.)

the fact that, during polymerization, the dissociation between the lead and two methacrylate molecules is negligible and, therefore, lead is responsible for ﬁxing the position of pairs of this monomer [62,127]. From a theoretical point of view, the study of the ability of the polymer network to memorize a speciﬁc conformation after a dramatic change in swelling degree has a strong interest, and is the aim of recent reviews [129–131]. PNIPA-based imprinted hydrogels have also been prepared using organic molecules as templates. Liu et al. [132,133] obtained temperature-sensitive hydrogels imprinted for 4-aminopyridine and l-pyroglutamic, which showed the same ability to sorb and release the drug after several shrunken–swollen cycles (Fig. 18). These hydrogels had signiﬁcantly larger saturation and afﬁnity constants than the non-imprinted ones, and were also highly selective. These results indicate that temperature-sensitive imprinted gels have a great potential in drug delivery. The reversibility of drug release and re-uptake as a function of temperature may be useful for the treatment of some pathological events that are accompanied by local changes in temperature, or to stop the delivery by externally applied local hyperthermia. pH-sensitive imprinted gels The precipitation polymerization method can be applied to prepare homogenous spherical microparticles imprinted for sulfasalazine, a prodrug used in

252

Fig. 18. Adsorption of l-pyroglutamic acid (Pga) by the imprinted and non-imprinted

hydrogels in the shrunken state at 551C. MIPs were prepared with N-isopropylacrylamide (40 mmol), MAA (2 mmol), EGDMA (6 mmol) and Pga (0.5 mmol). (Reproduced from Liu et al. [133] with permission from Wiley Interscience.)

Sulfasalazine released (%)

100

Non-imprinted

80 Imprinted 60 40 20 0 0

2

4 6 Time (hours)

8

10

Fig. 19. Release proﬁle of sulfasalazine at pH 1 from 0 to 2 h, and at pH 6.8 from 2 to

20 h, from non-imprinted and imprinted (0.23 mmol of template) particles (in the photography) made via a precipitation polymerization method using MAA (12 mmol) and EGDMA (16 mmol). (Adapted from Puoci et al. [134] with permission from Wiley Interscience).

colon diseases [134]. Drug-loaded microparticles showed a pH-dependent release (Fig. 19). At pH 1.0, the carboxylic groups of the functional monomer are not ionized and there is a good interaction with sulfasalazine, which is strongly associated to the imprinted cavities. When the pH of the medium is adjusted to 6.8, simulating intestinal conditions, the carboxylic groups become ionized and the drug is released. Disregarding the pH, the

253 non-imprinted particles released sulfasalazine more rapidly than the imprinted ones, which clearly prove the greater afﬁnity of the imprinted cavities for the drug. A strategy of preparing pH-sensitive imprinted hydrogels, using amylose as a host matrix, has been proposed by Kanekiyo et al. [135,136]. Amylose chains modiﬁed with acryloyl groups (acryloylamylose) form, in aqueous medium, helical inclusion-complexes with bisphenol-A that acts as template. The complexes are copolymerized with a monomer that has ionizable groups (acrylic acid, AA) and BIS as cross-linker (Fig. 20). Similarly, other MIPs were prepared with acrylamide instead of AA. In all cases, the polymers were directly obtained as particles. The rebinding ability of MIPs prepared with acrylic acid showed a strong dependence on pH; the higher the pH, the lower the binding of bisphenol-A. This is due to that the increase in pH promoting the electrostatic repulsion between the anionic groups of AA, which causes conformational changes in the amylase chains. As a consequence, the binding cavities created through the imprinting process are disrupted. A decrease in pH restores the cavities and the binding afﬁnity.

Fig. 20. Synthesis of amylose-based imprinted polymer and its pH-responsive structural change. (Reproduced from Kanekiyo et al. [136] with permission from Wiley Interscience.)

254 Feedback-regulated DDS Frequently, the levels of some substances are indirect indicators of the degree of dysfunction of certain physiological mechanisms. The availability of systems capable of selective recognition of these substances is an essential step to create feedback-regulated DDS able to modulate drug release as a function of their concentration in the body [137]. The prevalence of diabetes in developed countries is quite considerable [138] and several attempts in achieving glucose-regulated insulin delivery systems have been made [139]. Molecular imprinting technology is an adequate tool to develop systems with receptors for glucose, which may be the ﬁrst step in preparing a feedback-regulated delivery system. A response that consists of a change in pH proportional to the glucose concentration in the medium can be obtained with a device prepared using a-methyl-D-glucoside imprinted polymer, through metal coordination [140]. New functional monomers to enhance the recognition of glucose at 5.5opHo7.5, such as[(4-(N-vinylbenzyl)diethylenetriamine) copper(II)] diformate or [(diethylenetriamine) copper(II)] dinitrate, which can form 1:1 complexes with carbohydrates in this range of pH, were synthesized [141–143] (Fig. 21). These materials can load other carbohydrates by using cross-linkers and monomers that may establish hydrogenbonding interactions. As the matrix polarity increased, the polymer preference for the large and polar disaccharide lactose also rose, in detriment of the less polar monosaccharide glucose [142]. Thus, depending on the speciﬁc carbo-

Carbohydrate sorption (µmol/ g)

150 Imprinted Non-imprinted

125 100 75 50 25 0 Glucose

Galactose

Mannose

Fig. 21. Rebinding capability of glucose-imprinted and non-imprinted polymers for

glucose, galactose and mannose in water. The polymers consisted of [(4-(N-vinylbenzyl)diethylenetriamine) copper(II)] diformate (5 mol%) cross-linked with pentaerythritol tetraacrylate (95 mol%). (Reproduced from Striegler [141] with permission from Elsevier.)

255 hydrate to be recognized, it may be practical to use polar or nonpolar crosslinking monomers. Mayes et al. [144] and Oral, Peppas and co-workers [145] demonstrated the feasibility of hydrogen-bonding interactions between carbohydrates and methacrylic monomers, to produce high afﬁnity binding polymer networks. These sensors may serve as a basis for the obtaining of glucoseresponsive DDS. Kataoka et al. [146] have developed a glucose-responsive gel able to regulate insulin release, incorporating 3-acrylamidophenylboronic acid as functional monomer. The apparent pKa of phenylboronic acid is lowered from 8.9 to 7.9 when interacted with hydroxyl groups of glucose. In consequence, when glucose concentration in a pH 9 rises, the solubility of the polymer is abruptly increased. This provokes a rapid release of insulin that is effectively shut off by decreasing glucose concentration below a critical value. Although this approach was made using non-imprinted polymers, sugar imprinting of boronic groups in polymer networks is achievable. In fact, imprinted systems prepared using boronate esters – formed between pairs of hydroxyl groups of the sugar molecule and p-vinylphenylboronic acid – may even distinguish racemates of glycosides and free carbohydrates [147]. Neighbouring amino functionalities in the boronic acids greatly enhance the formation of boronic acid–sugar complexes, making it possible to create carbohydrate sensors at neutral pH [148]. Therefore, the application of the molecular imprinting technology to the development of this type of systems could contribute to making them more selective, improving their performance signiﬁcantly. MIPs as trap systems There is a very long list of food components that may be related to the apparition/worsening of certain illness (e.g., sugars, hormones, cholesterol, some amino acids and ions). Imprinted traps able to bind such substances in the gastrointestinal tract, hindering their absorption by the body, may be used instead of or as a complement of the pharmacological therapy, for the efﬁcient treatment of these illnesses. The information obtained in the development of these systems might also serve as a base for the design of feedbackregulated DDS. Glucose traps Starting from poly(allylamine hydrochloride), a commercially available high molecular weight polymer (15,000 Da), and glucose phosphate monosodium salt, Wizeman and Koﬁnas [149] prepared a non-covalent imprinted network by ionic association, with the aim of binding glucose in the stomach and small intestine and, then, passing unabsorbed through the body. The networks were prepared with three different cross-linkers using, both for the synthesis and the testing, aqueous solutions. Dried polymers added to either

256 Table 3. Sugar binding ability of poly(allylamine hydrochloride) hydrogels, which were imprinted with glucose phosphate monosodium salt (1%) or non-imprinted [149]. Cross-linker

Imprinting

Glucose binding (g/g)

Fructose binding (g/g)

Epichlorohydrin

Yes No Yes No Yes No

0.54 0.20 0.53 0.18 0.34 0.13

0 0 0.21 0.11 0.18 0.05

Ethyleneglycol diglycidylether Glycerol diglycidylether

(0.03) (0.01) (0.03) (0.01) (0.06) (0.01)

(0.03) (0.02) (0.03) (0.03)

Mean values (standard deviations).

glucose or fructose solution (50 mg/mL) showed very different binding capability depending on the cross-linker used. The higher degree of speciﬁcity for glucose was achieved with epichlorohydrin, the smallest molecular size cross-linker, after 4 h of immersion in the sugar solutions (Table 3). The MIP system would simply be ingested along with foods that are high in glucose with the aim of reducing the sharp rise in blood sugar associated with the ingestion of signiﬁcant amounts of monosaccharides. These MIP systems may also have applications in the preparation of glucose sensors. Cholesterol traps Cholesterol is receiving increasing attention as a template to create sequestrant devices, owing to the inﬂuence of high blood cholesterol levels on the onset of atherosclerosis and myocardial infarction. Comprehensive reviews about MIPs for steroids and, in particular, for cholesterol have been carried out by Davidson and Hayes [150] and Gore et al. [151]. Polymer networks containing imprinted cyclodextrins were found particularly useful in enhancing their capability to form inclusion complexes with cholesterol [152]. Free cyclodextrins are commonly used to precipitate cholesterol from concentrated solutions. However, the yield of the reaction is limited by the fact that cholesterol is too large to be accommodated within the cavity of a single b-cyclodextrin. To be effectively bound, two or more b-cyclodextrins must interact simultaneously with one cholesterol molecule. Komiyama et al. [53] prepared, in dimethylsulfoxide, polymer networks that hold groups of two cyclodextrins at the adequate distance to complex cholesterol, by cross-linking vinyl monomers of bcyclodextrins with hexamethylene diisocyanate or toluene-2,4-diisocyanate in the presence of cholesterol (see Fig. 9). The cholesterol-MIP systems showed strong and selective rebinding of the substrate in the aqueous medium; the hydroxyl group at the 3-position and the alkyl chain at the 17-position playing an important role in the recognition. More recently, Egawa et al. [153] have prepared microspheres of cyclodextrins in a dimethylsulfoxide/poly

257 (dimethylsiloxane) emulsion, using cholesterol as the template molecule. These imprinted microspheres also showed enhanced binding ability to other steroids such as progesterone or testosterone. Also taking advantage of van der Waals interactions, Sellergren et al. [22] obtained cholesterol-MIPs using amphiphilic monomers under conditions favouring apolar association between the monomers and the template. These amphiphilic monomers were polymerizable derivatives of cholesterol or bile acids (3 mmol), which were mixed with EGDMA (30 mmol) and MAA (6 mmol) and dissolved in ethanol. To obtain an imprinting effect, cholesterol (1.5 mmol) was added before polymerization. In intestine-mimicking medium, cholesterol-imprinted polymers exhibited a considerably greater adsorption capacity (ca. 45 mmol/g) than the non-imprinted ones (ca. 33 mmol/ g). Zhong et al. [154] observed that combining the use of polymerizable cholesterol and cyclodextrins, the MIPs take up cholesterol more efﬁciently. Other two covalent approaches, using cholesterol bound to a monomer, have been proposed to obtain (i) surface imprinted core-shell nanoparticles using pyridinium 12-(cholesteryloxycarbonyloxy)dodecane-sulphate as template [155] and (ii) imprinted bulk monoliths using cholesteryl-2-hydroxyethyl methacrylate carbonate as template [156]. Non-covalent imprinting using conventional monomers, such as MAA or 4-VP, usually results in lower binding afﬁnity [157,158].

Other traps Another way to decrease the dietary uptake of cholesterol consists of using imprinted polymers for bile acids, which participate in its emulsiﬁcation in the gastrointestinal tract, prior to its absorption. A strong decrease in bile acids concentration was achieved with (i) deoxycholic acid-imprinted poly(N,N0 -diethyl(4-vinylphenyl)amide) networks [159] and (ii) cholateimprinted poly(allylammonium chloride) systems [160]. Iron is an essential trace element but at high concentrations it may be responsible for severe toxic effects. The human body has no physiological route for elimination of excess iron and, therefore, chelating and absorbent agents are used for its removal. Blood detoxiﬁcation by extracorporeal afﬁnity adsorption has been recently proposed as an adequate approach but the already developed adsorbents for this purpose lack good selectivity. A high selectivity is essential to avoid the extraction of other trace elements. Imprinted beads for Fe3+, prepared by suspension polymerization, has shown a great afﬁnity and selectivity for this ion in plasma human [161] (Table 4). Although at a lower interest in the therapeutic ﬁeld, MIPs useful as caffeine traps have also been developed using cross-linker (divinylbenzene) and functional monomers, to which caffeine is adsorbed through hydrophobic p–p interactions [162].

258 Table 4. Distribution coefﬁcient (Kd, mL/g) and selectivity coefﬁcient (k) of Zn2+ and Cr3+ with respect to Fe3+ in fresh human plasma. Metal ion

3+

Fe Zn2+ Cr3+

Non-imprinted beads

Imprinted beads

Kd

K

Kd

K

82.4 90.6 120.5

— 0.72 0.29

612 45.7 51.8

— 12.5 14.1

k0

17.3 48.6

Kd ¼ [(CiCf)/Cf] V/m, where Ci and Cf are initial and ﬁnal concentrations of metal ions, respectively; V is the volume of the solution and m is the mass of beads. k ¼ Ktemplate/Kcompetitor; being Ktemplate and Kcompetitor the corresponding equilibrium binding constants. The relative selectivity coefﬁcient was calculated as k0 ¼ kimprinted/knon-imprinted. Reproduced from Yavuz et al. [161] with permission form Elsevier Science.

A view to the future Despite the already developed interesting applications of MIPs, commented on previous sections, the incorporation of the molecular imprinting approach for the development of DDS is just at its incipient stage. Nevertheless, it can be foreseen that in the next few years signiﬁcant progress will occur in this ﬁeld, taking advantage of the improvements of this technology in other areas. The use of predictive tools for a rational design of imprinted systems and the development of molecular imprinting in water may be highlighted among the evolution lines that should contribute more to the enhancement of their applicability in drug delivery. The optimization, using predictive tools, of the nature and amount of functional monomers should overcome the time- and material-consuming method of ‘‘trial and error’’ and improve the speciﬁcity and the afﬁnity for the template. This may be achieved with the use of rapid experimental screening processes [163], combinatorial libraries [164] and with the application of isothermal titration calorimetry analysis, which has proved to be a suitable method to elucidate the stoichiometry of the functional monomer:template complexes, to investigate the thermodynamics of molecular recognition and to assess the efﬁciency of the molecular imprinting process [49,165,166]. A biomolecular binding database, which includes binding afﬁnities for proteins, nucleic acids, carbohydrates, drugs and synthetic host and guest molecules obtained from literature, is in a growing process [167]. Additionally, new ways of tackling the challenge of preparing MIPs that work efﬁciently in aqueous solutions are required. Although MIPs made in organic solvents can often recognize the template in water, much more work should be devoted towards making good imprints in aqueous media. This may open the possibility of obtaining imprinted systems for very labile molecules such as some peptidic drugs and oligonucleotides, making it possible even to use them in gene therapy [168,169]. For example, MIPs able to recognize speciﬁc compounds of cell

259 surface, especially oligosaccharides or lectins, may be suitable for targeting the DDS to speciﬁc tissues or cells, increasing the residence time at the absorption site and providing an intimate contact to the absorptive tissue. Lectinimprinted polymers have been prepared in aqueous solutions using the ability of this glycoprotein to reversibly bind certain saccharidic functional monomers [170]. The afﬁnity of each lectin for speciﬁc mucosal surfaces of the gastrointestinal, respiratory and urogenital tract may serve as a basis for the development of lectin-mediated DDS of poorly absorbable drugs by opening endocytic pathways [171]. The advances in the preparation of MIPs as gels [131], ﬁlms [172,173] and nano- and micro-particles [60,165] should also enhance the possibilities of application of these polymers for sustained and targeting drug delivery by different routes, including the parenteral one. The molecular imprinting technology allows advanced excipients to be obtained, each tailored for each speciﬁc drug. Therefore, the development of each MIP should be approached in an individualized way to ﬁt the speciﬁc drug. This makes the preparation of these new and highly efﬁcient materials more complex than that of conventional excipients, which are of general and unspeciﬁc use. The adaptation of the molecular imprinting technology to the safety requirements (i.e., toxicity of the unreacted substances and biocompatibility of the polymer) of the pharmaceutical products should also be the aim of the near future work. Acknowledgements This work was ﬁnanced by the Xunta de Galicia (PGIDT 02BTF20302PR; PGIDT 03PXIC20303PN) and the Ministerio de Ciencia y Tecnologı´ a, Spain (RYC2001-8; SAF2005-01930). References 1. 2. 3.

4.

5.

Rathbone MJ, Hadgraft J and Roberts MS (eds). Modiﬁed-release drug delivery technology, New York, Marcel Dekker, Inc., 2003. Chien YW and Lin S. Optimisation of treatment by applying programmable ratecontrolled drug delivery technology. Clin Pharmacokinet 2002;41:1267–1299. Wulff G. Molecular imprinting in cross-linked materials with the aid of molecular templates – a way towards artiﬁcial antibodies. Angew Chem Int Ed Engl 1995;34:1812–1832. Maeda M and Bartsch RA. Molecular and ionic recognition with imprinted polymers: a brief overview. In: Molecular and Ionic Recognition with Imprinted Polymers, Barstch RA and Maeda M (eds), Washington DC, ACS Symposium Series 703, American Chemical Society, 1998, pp. 1–8. Allender CJ, Brain KR and Heard CM. Molecularly imprinted polymers – preparation, biomedical applications and technical challenges. In: Progress in Medicinal Chemistry, Vol. 36, King FD and Oxford AW (eds), Elsevier, Amsterdam, 1999, pp. 235–291.

260 6. 7.

8. 9. 10.

11.

12.

13. 14.

15. 16.

17.

18.

19. 20. 21. 22. 23. 24.

Alvarez-Lorenzo C and Concheiro A. Molecularly imprinted polymers for drug delivery. J Chromatogr B 2004;804:231–245. Ciardelli G, Cioni B, Cristallini C, Barbani N, Silvestri D and Giusti P. Acrylic polymeric nanospheres for the release and recognition of molecules of clinical interest. Biosens Bioelectron 2004;20:1083–1090. van Nostrum CF. Molecular imprinting: a new toll for drug innovation. Drug Discov Today 2005;2:119–124. Owens PK, Karlsson L, Lutz ESM and Andersson LI. Molecular imprinting for bioand pharmaceutical analysis. Trends Anal Chem 1999;18:146–154. Andersson LI. Molecular imprinting for drug bioanalysis. A review on the application of imprinted polymers to solid-phase extraction and binding assay. J Chromatogr B 2000;739:163–173. Martin PD, Jones GR, Stringer F and Wilson ID. Comparison of normal and reversedphase solid phase extraction methods for extraction of beta-blockers from plasma using molecularly imprinted polymers. Analyst 2003;128:345–350. Prasad BB and Banerjee S. Preparation, characterization and performance of a silica gel bonded molecularly imprinted polymer for selective recognition and enrichment of beta-lactam antibiotics. React Funct Polym 2003;55:159–169. Masque´ N, Marce´ RM and Borrull F. Molecularly imprinted polymers: new tailor-made materials for selective solid-phase extraction. Trends Anal Chem 2001;20:477–486. Sanbe H, Hosoya K and Haginaka J. Preparation of uniformly sized molecularly imprinted polymers for phenolic compounds and their application to the assay of bisphenol A in river water. Anal Sci 2003;19:715–719. Ansell RJ. Molecularly imprinted polymers in pseudoimmunoassay. J Chromatogr B 2004;804:151–165. Chassing C, Stokes J, Venn RF, Lanza F, Sellergren B, Holmberg A and Berggren C. Molecularly imprinted polymers for the determination of a pharmaceutical development compound in plasma using 96-well MISPE technology. J Chromatogr B 2004;804:71–81. Piletska EV, Romero-Guerra M, Chianella I, Karim K, Turner APF and Piletsky SA. Towards the development of multisensor for drugs of abuse based on molecular imprinted polymers. Anal Chim Acta 2005;542:111–117. Nakamura N, Ono M, Nakajima T, Ito Y, Aketo T and Haginaka J. Uniformly sized molecularly imprinted polymer for atropine and its application to the determination of atropine and scopolamine in pharmaceutical preparations containing scopolia extract. J Pharm Biomed Anal 2005;37:213–237. Hilt JZ and Byrne ME. Conﬁgurational biomimesis in drug delivery: molecular imprinting of biologically signiﬁcant molecules. Adv Drug Deliv Rev 2004;56:1599–1620. Yano K and Karube I. Molecularly imprinted polymers for biosensor applications. TRAC 1999;18:199–204. Xu X, Zhu L and Chen L. Separation and screening of compounds of biological origin using molecularly imprinted polymers. J Chromatogr B 2004;804:61–69. Sellergren B, Wieschemeyer J, Boos KS and Seidel D. Imprinted polymers for selective adsorption of cholesterol from gastrointestinal ﬂuids. Chem Mater 1998;10:4037–4046. Ramstro¨m O and Ansell RJ. Molecular imprinting technology: challenges and prospects for the future. Chirality 1998;10:195–209. Alexander C, Davidson L and Hayes W. Imprinted polymers: artiﬁcial molecular recognition materials with applications in synthesis and catalysis. Tetrahedron 2003;59:2025–2057.

261 25. Marty JD and Mauzac M. Molecular imprinting: state of art and perspectives. Adv Polym Sci 2005;172:1–35. 26. Yan M and Ramstro¨m O. Molecular imprinting – an introduction. In: Molecularly Imprinted Materials, Yan M and Ramstro¨m O (eds), New York, Marcel Dekker Inc., 2005, pp. 1–12. 27. Wulff G and Bifﬁs A. Molecularly imprinting with covalent or stoichiometric non-covalent interactions. In: Molecularly Imprinted Polymers, Sellergren B (ed), Amsterdam, Elsevier, 2001, pp. 71–111. 28. Arshady R and Mosbach K. Synthesis of substrate-selective polymers by host–guest polymerization. Makromol Chem 1981;182:687–692. 29. Sellergren B. The non-covalent approach to molecular imprinting. In: Molecularly Imprinted Polymers, Sellergren B (ed), Amsterdam, Elsevier, 2001, pp. 113–184. 30. Andersson LI, Hardenborg E, Sandberg-Stall M, Moller K, Henriksson J, BramsbySjostrom I, Olsson LI and Abdel-Rehim M. Development of a molecularly imprinted polymer based solid-phase extraction of local anaesthetics from human plasma. Anal Chim Acta 2004;526:147–154. 31. D’Oleo R, Alvarez-Lorenzo C and Sun G. A new approach to design imprinted polymer gels without using a template. Macromolecules 2001;34:4965–4971. 32. Moritani T and Alvarez-Lorenzo C. Conformational imprinting effect on stimulisensitive gels made with an imprinter monomer. Macromolecules 2001;34:7796–7803. 33. Oral E and Peppas NA. Dynamic studies of molecular imprinting polymerizations. Polymer 2004;45:6163–6173. 34. Molinelli A, O’Mahony J, Nolan K, Smyth MR, Jakusch M and Mizaikoff B. Analyzing the mechanisms of selectivity in biomimetic self-assemblies via IR and NMR spectroscopy of prepolymerization solutions and molecular dynamics simulations. Anal Chem 2005;77:5196–5204. 35. McStay D, Al-Obaidi AH, Hoskins R and Quinn PJ. Raman spectroscopy of molecularly imprinted polymers. J Opt A: Pure Appl Opt 2005;7:S340–S345. 36. Fish WP, Ferreira J, Sheardy RD, Snow NH and O’Brien TP. Rational design of an imprinted polymer: maximizing selectivity by optimizing the monomer-template ratio of a cinchonidine MIP, prior to polymerization, using microcalorimetry. J Liq Chromatogr Relat Techn 2005;28:1–15. 37. O’Mahony J, Molinelli A, Nolan K, Smyth MR and Mizaikoff B. Towards the rational development of molecularly imprinted polymers: 1 H NMR studies on hydrophobicity and ion-pair interactions as driving forces for selectivity. Biosens Bioelectron 2005;20:1884–1893. 38. Batra D and Shea KJ. Combinatorial methods in molecular imprinting. Curr Opin Chem Biol 2003;7:434–442. 39. Dong WG, Yan M, Zhang ML, Liu Z and Li YM. A computational and experimental investigation of the interaction between the template molecule and the functional monomer used in the molecularly imprinted polymer. Anal Chim Acta 2005;542:186–192. 40. Pavel D and Lagowski J. Computationally designed monomers and polymers for molecular imprinting of theophylline and its derivatives. Part I. Polymer 2005;46:7528–7542. 41. Piletska EV, Turner NW, Turner APF and Piletsky SA. Controlled release of the herbicide simazine from computationally designed molecularly imprinted polymers. J Controlled Rel 2005;108:132–139. 42. Davies MP, Biasi VD and Perrett D. Approaches to the rational design of molecularly imprinted polymers. Anal Chim Acta 2004;504:7–14.

262 43. Rosengren AM, Karlsson JG, Andersson PO and Nicholls IA. Chemometric models of template-molecularly imprinted polymer binding. Anal Chem 2005;77:5700–5705. 44. Umpleby II RJ, Baxter SC, Rampey AM, Rushton GT, Chen Y and Shimizu KD. Characterization of the heterogeneous binding site afﬁnity distributions in molecularly imprinted polymers. J Chromatogr B 2004;804:141–149. 45. Andersson HS and Nicholls IA. Spectroscopic evaluation of molecular imprinting polymerisation systems. Bioorg Chem 1997;25:203–211. 46. Andersson HS, Karlsson JG, Piletsky SA, Koch-Schmidt AC, Mosbach K and Nicholls I. Study of the nature of recognition in molecularly imprinted polymers. Inﬂuence of monomer-template ratio and sample load on retention and selectivity. J Chromatogr A 1999;848:39–49. 47. Rampey AM, Umpleby II RJ, Rushton GT, Iseman JC, Shah RN and Shimizu KD. Characterization of the imprinting effect and the inﬂuence of imprinting conditions on afﬁnity, capacity, and heterogeneity in molecularly imprinted polymers using the Freundlich isotherm-afﬁnity distribution analysis. Anal Chem 2004;76:1123–1133. 48. Nicholls IA, Adbo K, Andersson HS, Andersson PO, Ankarloo J, Hedin-Dahlstrom J, Jokela P, Karlsson JG, Olofsson L, Rosengren J, Shoravi S, Svenson J and Wikman S. Can we rationally design molecularly imprinted polymers? Anal Chim Acta 2001;435:9–18. 49. Weber A, Dettling M, Brunner H and Tovar GEM. Isothermal titration calorimetry of molecularly imprinted polymer nanospheres. Macromol Rapid Commun 2002;23:824–828. 50. Schmidt RH, Belmont AS and Haupt K. Porogen formulations for obtaining molecularly imprinted polymers with optimized binding properties. Anal Chim Acta 2005;542:118–124. 51. Cormack PAG and Elorza AZ. Molecularly imprinted polymers: synthesis and characterization. J Chromatogr B 2004;804:173–178. 52. Wu LQ, Zhu KC, Zhao WP and Li YZ. Theoretical and experimental study of nicotinamide molecularly imprinted polymers with different porogens. Anal Chim Acta 2005;549:39–44. 53. Komiyama M, Takeuchi T, Mukawa T and Asanuma H. Molecular imprinting, Weinheim, Wiley-VCH, 2003. 54. Piletsky SA, Andersson HS and Nicholls IA. Combined hydrophobic and electrostatic interaction-based recognition in molecularly imprinted polymers. Macromolecules 1999;32:633–636. 55. Kubo T, Hosoya K, Nomachi M, Tanaka N and Kaya K. Preparation of a novel molecularly imprinted polymer using a water-soluble crosslinking agent. Anal Bioanal Chem 2005;382:1698–1701. 56. Sibrian-Vazquez M and Spivak DA. Improving the strategy and performance of molecularly imprinted polymers using cross-linking functional monomers. J Org Chem 2003;68:9604–9611. 57. Svenson J and Nicholls IA. On the thermal and chemical stability of molecularly imprinted polymers. Anal Chim Acta 2001;435:19–24. 58. Yourtee DM, Smith RE, Russo KA, Burmaster S, Cannon JM, Eick JD and Kostoryz EL. The stability of methacrylate biomaterials when enzyme challenged: kinetic and systematic evaluations. J Biomed Mat Res 2001;57:522–531. 59. Greenley RZ. Free radical copolymerization reactivity ratios. In: Polymer Handbook, 3rd edn, Brandrup J and Immergut EH (eds), New York, Wiley, 1989, pp. II153–II251.

263 60. Mayes AG. Polymerisation techniques for the formation of imprinted beads. In: Molecularly Imprinted Polymers, Sellergren B (ed), Amsterdam, Elsevier, 2001, pp. 305–324. 61. Peppas NA, Bures P, Leobandung W and Ichikawa H. Hydrogels in pharmaceutical formulations. Eur J Pharm Sci 2000;50:27–46. 62. Alvarez-Lorenzo C, Guney O, Oya T, Sakai Y, Kobayashi M, Enoki T, Takeoka Y, Ishibashi T, Kuroda K, Tanaka K, Wang G, Grosberg A Yu, Masamune S and Tanaka T. Polymer gels that memorize elements of molecular conformation. Macromolecules 2000;33:8693–8697. 63. Aydin O, Attila G, Dogan A, Aydin MV, Canacankatan N and Kanik A. The effects of methyl methacrylate on nasal cavity, lung, and antioxidant system (an experimental inhalation study). Toxicol Pathol 2002;30:350–356. 64. Fiume MZ. Final report on the safety assessment of acrylates copolymer and 33 related cosmetic ingredients. Int J Toxicol 2002;21:1–50. 65. Garrett Q, Laycock B and Garrett RW. Hydrogen lens monomer constituents modulate protein sorption. Invest Ophthalmol Vis Sci 2000;41:1687–1695. 66. Al-Habet S, Bashaw ED, Lesko L, Williams R and Balian J. Narrow therapeutic index (NTI) drugs: preliminary deﬁnition and criteria. Clin Pharmacol Ther 2000;67:PIII23. 67. Norell MC, Andersson HS and Nicholls IA. Towards molecularly imprinted polymer based controlled release materials. J Mol Recog 1998;11:98–102. 68. Vlatakis G, Andersson LI, Mu¨ller R and Mosbach K. Drug assay using antibody mimics made by molecular imprinting. Nature 1993;361:645–647. 69. Allender CJ, Heard CM and Brain KR. Molecularly imprinted polymers as novel excipients for selective (trans) dermal delivery. In: Perspectives in Percutaneous Penetration, Vol. 5b, Brain KR, James VJ and Walters KA (eds), STS Publishing, Cardiff, 1998, pp. 183–185. 70. Allender CJ, Richardson C, Woodhouse B, Heard CM and Brain KR. Pharmaceutical applications for molecularly imprinted polymers. Int J Pharm 2000;195:39–43. 71. Cai W and Gupta RB. Molecular-imprinted polymers selective for tetracycline binding. Sep Purif Technol 2004;35:215–221. 72. Ciardelli G, Cioni B, Cristallini C, Barbani N, Silvestri D and Giusti P. Acrylic polymeric nanospheres for the release and recognition of molecules of clinical interest. Biosens Bioelectr 2004;20:1083–1090. 73. Silvestri D, Borrelli C, Giusti P, Cristallini C and Ciardelli G. Polymeric devices containing imprinted nanospheres: a novel approach to improve recognition in water for clinical uses. Anal Chim Acta 2005;542:3–13. 74. Srinivas NR, Barbhaiya RH and Midha KK. Enantiomeric drug development: issues, considerations, and regulatory requirements. J Pharm Sci 2001;90:1205–1215. 75. Hillier MC and Reider PJ. Stereoselective synthesis from a process research perspective. Drug Discover Today 2002;7:303–314. 76. Ansell RJ and Kuah KL. Imprinted polymers for chiral resolution of (7)-ephedrine: understanding the pre-polymerisation equilibrium and the action of different mobile phase modiﬁers. Analyst 2005;130:179–187. 77. Sun R, Yu H, Luo H and Shen Z. Construction and application of a stoichiometric displacement model for retention in chiral recognition of molecular imprinting. J Chromatogr A 2004;1055:1–9. 78. Suedee R, Srichana T and Rattananont T. Enantioselective release of controlled delivery granules based on molecularly imprinted polymers. Drug Del 2002;9:19–30.

264 79. Suedee R, Srichana T and Martin GP. Evaluation of matrices containing molecularly imprinted polymers in the enantioselective-controlled delivery of beta-blockers. J Controlled Release 2000;66:135–147. 80. Suedee R, Srichana T, Chotivatesin R and Martin GP. Stereoselective release behaviors of imprinted bead matrices. Drug Dev Ind Pharm 2002;28:545–554. 81. Refojo MF and Leong FL. Microscopic determination of the penetration of proteins and polysaccharides into poly(hydroxyethyl methacrylate) and similar hydrogels. J Appl Polym Sci 1979;66:227–237. 82. Shea K. Molecular imprinting of synthetic network polymers: the de novo synthesis of macromolecular binding and catalytic sites. Trends Polym Sci 1994;2:166–173. 83. Shnek DR, Pack DW, Sasaki DY and Arnold FH. Speciﬁc protein attachment to artiﬁcial membranes via coordination to lipid-bound copper(II). Langmuir 1994;10:2382–2388. 84. Hawkins DM, Stevenson D and Reddy SM. Investigation of protein imprinting in hydrogel-based molecularly imprinted polymers (HydroMIPs). Anal Chim Acta 2005;542:61–65. 85. Guo TY, Xia YQ, Wang J, Song MD and Zhang BH. Chitosan beads as molecularly imprinted polymer matrix for selective separation of proteins. Biomaterials 2005;26:5737–5745. 86. Xia YQ, Guo TY, Song MD, Zhang BH and Zhang BL. Hemoglobin recognition by imprinting in semi-interpenetrating polymer network hydrogel based on polyacrylamide and chitosan. Biomacromolecules 2005;26:2601–2606. 87. Shiomi T, Matsui M, Mizukami F and Sakaguchi K. A method for the molecular imprinting of hemoglobin on silica surfaces using silanes. Biomaterials 2005;26:5564–5571. 88. Nicholls IA. Thermodynamic considerations for the design of and ligand recognition by molecularly imprinted polymers. Chem Lett 1995;11:1035–1036. 89. Rachkov A and Minoura N. Towards molecularly imprinted polymers selective to peptides and proteins. The epitope approach. Biochim Biophys Acta 2001;1544:255–266. 90. Rachkov A, Hu MJ, Bulgarevich E, Matsumoto T and Minoura N. Molecularly imprinted polymers prepared in aqueous solution selective for [Sar(1),Ala(8)]angiotensin II. Anal Chim Acta 2004;504:191–197. 91. Tai DF, Lin CY, Wu TZ and Chen LK. Recognition of dengue virus protein using epitope-mediated molecularly imprinted ﬁlm. Anal Chem 2005;77:5140–5143. 92. Asanuma H, Hishiya T and Komiyama M. Tailor-made receptors by molecular imprinting. Adv Mater 2000;12:1019–1030. 93. Uekama K. Design and evaluation of cyclodextrin-based drug formulation. Chem Pharm Bull 2004;52:900–915. 94. Asanuma H, Akiyama T, Kajiya K, Hishiya T and Komiyama M. Molecular imprinting of cyclodextrin in water for the recognition of nanometer-scaled guests. Anal Chim Acta 2001;435:25–33. 95. Stella VJ, Rao VM, Zannou EA and Zia V. Mechanisms of drug release from cyclodextrin complexes. Adv Drug Del Rev 1999;36:3–16. 96. Liu YY, Fan XD, Hu H and Tang ZH. Release of chlorambucil from (poly(N-isopropylacrylamide) hydrogels with b-cyclodextrin moieties. Macromol Biosci 2004;4:729–736. 97. Zhang JT, Huang SW, Gao FZ and Zhuo RX. Novel temperature-sensitive, b-cyclodextrin-incorporated poly(N-isopropylacrylamide) hydrogels for slow release of drug. Colloid Polym Sci 2005;283:461–464.

265 98. Rodriguez-Tenreiro C, Rodriguez-Perez AI, Alvarez-Lorenzo C, Concheiro A and Torres-Labandeira JJ. New cyclodextrin hydrogels cross-linked with diglycidylethers with a high drug loading and controlled release ability. Pharm Res 2006;23:121–130. 99. Sreenivasan K. Application of molecularly imprinted polymer as a drug retaining matrix. Angew Makromol Chem 1997;246:65–67. 100. Nicolson PC and Vogt J. Soft contact lens polymers: an evolution. Biomaterials 2001;22:3273–3283. 101. Alvarez-Lorenzo C, Hiratani H, Go´mez-Amoza JL, Martı´ nez-Pacheco R, Souto C and Concheiro A. Soft contact lenses capable of sustained delivery of timolol. J Pharm Sci 2002;91:2182–2192. 102. Hiratani H and Alvarez-Lorenzo C. Timolol uptake and release by imprinted soft contact lenses made of N,N-diethylacrylamide and methacrylic acid. J Controlled Rel 2002;83:223–230. 103. Hiratani H, Alvarez-Lorenzo C. Process for production of hydrogel material enhanced in the intake of drugs and permitting sustained release of drugs. WO03090805, November 6, 2003. 104. Davis JL, Gilger BC and Robinson MR. Novel approaches to ocular drug delivery. Curr Opin Mol Ther 2004;6:195–205. 105. Heyrman TP, McDermott ML, Ubels JL and Edelhauser HF. Drug uptake and release by a hydrogel intraocular lens and the human crystalline lens. J Cataract Refract Surg 1989;15:169–175. 106. Karlgard CCS, Wong NS, Jones LW and Moresoli C. In vitro uptake and release studies of ocular pharmaceutical agents by silicon-containing and p-HEMA hydrogel contact lens materials. Int J Pharm 2003;257:141–151. 107. Sato T, Uchida R, Tanigawa H, Uno K and Murakami A. Application of polymer gels containing side-chain phosphate groups to drug-delivery contact lenses. J Appl Polym Sci 2005;98:731–735. 108. Hiratani H and Alvarez-Lorenzo C. Timolol uptake and release by imprinted soft contact lenses made of N,N-diethylacrylamide and methacrylic acid. J Controlled Rel 2002;83:223–230. 109. Hiratani H and Alvarez-Lorenzo C. The nature of backbone monomers determines the performance of imprinted soft contact lenses as timolol drug delivery systems. Biomaterials 2003;25:1105–1113. 110. Hiratani H, Fujiwara A, Tamiya Y, Mizutani Y and Alvarez-Lorenzo C. Ocular release of timolol from molecularly imprinted soft contact lenses. Biomaterials 2005;26:1293–1298. 111. Hiratani H, Mizutani Y and Alvarez-Lorenzo C. Controlling drug release from imprinted hydrogels by modifying the characteristics of the imprinted cavities. Macromol Biosci 2005;5:728–733. 112. Sreenivasan K. On the application of molecularly imprinted poly(HEMA) as a template responsive release system. J Appl Polym Sci 1999;71:1819–1821. 113. Karlsson JG, Andersson LI and Nicholls IA. Probing the molecular basis for ligandselective recognition in molecularly imprinted polymers selective for the local anaesthetic bupivacaine. Anal Chim Acta 2001;435:57–61. 114. Ye L, Cormack PAG and Mosbach K. Molecular imprinting on microgel spheres. Anal Chim Acta 2001;435:187. 115. Yoshimi Y, Ohdaira R, Iiyama C and Sakai K. ‘‘Gate effect’’ of thin layer of molecularly imprinted poly(methacrylic acid-co-ethyleneglycol dimethacrylate). Sensors Actuators B 2001;73:49–53.

266 116. Hattori K, Yoshimi Y and Sakai K. Gate effect of cellulosic dialysis membrane grafted with molecularly imprinted polymer. J Chem Eng Japan 2001;34:1466–1469. 117. Bourgeois S, Harvey R and Fattal E. Polymer colon drug delivery systems and their application to peptides, proteins, and nucleic acids. Am J Drug Del 2005;3:171–204. 118. Shah SS, Kulkarni MG and Mashelkar RA. Release kinetics of pendant substituted bioactive molecules from swellable hydrogels – role of chemical-reaction and diffusive transport. J Membr Sci 1990;51:83–104. 119. Karmalkar RN, Kulkarni MG and Mashelkar RA. Pendent chain linked delivery systems.2. Facile hydrolysis through molecular imprinting effects. J Controlled Release 1997;43:235–243. 120. Mashelkar RA, Kulkarni MG and Karmalkar RN. Polymer composition for controlled release of active ingredients in response to pH, and a process of preparing the same. US Patent 5,851,546, December 22, 1998. 121. Leonhardt A and Mosbach K. Enzyme-mimicking polymers exhibiting speciﬁc substrate binding and catalytic functions. React Polym 1987;6:285–290. 122. Wang G, Kuroda K, Enoki T, Grosberg A, Masamune S, Oya T, Takeoka Y and Tanaka T. Gel catalysts that switch on and off. Proc Nat Acad Sci 2000;97:9861–9864. 123. Yuk SH and Bae YH. Phase-transition polymers for drug delivery. Crit Rev Ther Drug Carrier Systems 1999;16:385–423. 124. Okano T, Bae YH, Jacobs H and Kim SW. Thermally on off switching polymers for drug permeation and release. J Control Release 1990;11:255–265. 125. Hirokawa Y and Tanaka T. Volume phase-transition in a nonionic gel. J Chem Phys 1984;81:6379–6380. 126. Schild HG. Poly (N-isopropylacrylamide) – experiment, theory and application. Prog Polym Sci 1992;17:163–249. 127. Tanaka T, Wang C, Pande V, Grosberg A, English A, Masamune S, Gold H, Levy R and King K. Polymer gels that can recognize and recover molecules. Faraday Discuss 1996;102:201–206. 128. Alvarez-Lorenzo C, Guney O, Oya T, Sakai Y, Kobayashi M, Enoki T, Takeoka Y, Ishibashi T, Kuroda K, Tanaka K, Wang G, Grosberg AYu, Masamune S and Tanaka T. Reversible adsorption of calcium ions by imprinted temperature sensitive gels. J Chem Phys 2001;114:2812–2816. 129. Ito K, Chuang J, Alvarez-Lorenzo C, Watanabe T, Ando N and Grosberg A Yu. Multiple point adsorption in a heteropolymer gel and the Tanaka approach to imprinting: experiment and theory. Prog Polym Sci 2003;28:1489–1515. 130. Ito K, Chuang J, Alvarez-Lorenzo C, Watanabe T, Ando N and Grosberg A Yu. Multiple-contact adsorption of target molecules by heteropolymer gels. Macromol Symp 2004;207:1–16. 131. Alvarez-Lorenzo C and Concheiro A. Molecularly imprinted gels and nano- and microparticles. Manufacture and applications. In: Smart Assemblies and Particulates, Vol. 7, Arshady R and Kono K (eds), London, Kentus Books, 2006, pp. 275–336. 132. Liu XY, Ding XB, Guan Y, Peng YX, Long XP, Wang XC, Chang K and Zhang Y. Fabrication of temperature-sensitive imprinted polymer hydrogel. Macromol Biosci 2004;4:412–415. 133. Liu XY, Guan Y, Ding XB, Peng YX, Long XP, Wang XC and Chang K. Design of temperature sensitive imprinted polymer hydrogels based on multiple-point hydrogen bonding. Macromol Biosci 2004;4:680–688.

267 134. Puoci F, Iemma F, Muzzalupo R, Spizzirri UG, Trombino S, Cassano R and Picci N. Spherical molecularly imprinted polymers (SMIPs) via a novel precipitation polymerization in the controlled delivery of salazine. Macromol Biosci 2004;4:22–26. 135. Kanekiyo Y, Naganawa R and Tao H. Molecular imprinting of bisphenol-A and alkylphenols using amylose as a host matrix. Chem Commun 2002;22:2698–2699. 136. Kanekiyo Y, Naganawa R and Tao H. pH-responsive molecularly imprinted polymers. Angew Chem 2003;42:3014–3016. 137. Miyata T, Uragami T and Nakamae K. Biomolecule-sensitive hydrogels. Adv Drug Del Rev 2002;54:79–98. 138. http://www.diabetes.org, accessed on June 2005. 139. Chu LY. Controlled release systems for insulin delivery. Expert Opin Ther Pat 2005;15:1147–1155. 140. Chen G, Guan Z, Chen CT, Fu L, Sundaresan V and Arnold FH. A glucose-sensing polymer. Nat Biotechnol 1997;15:354–357. 141. Striegler S. Selective discrimination of closely related monosaccharides at physiological pH by a polymeric receptor. Tetrahedron 2001;57:2349–2354. 142. Striegler S. Investigation of disaccharide recognition by molecularly imprinted polymers. Bioseparation 2001;10:307–314. 143. Striegler S and Dittel M. Evaluation of new strategies to prepare templated polymers with sufﬁcient oligosaccharide recognition capacity. Anal Chim Acta 2003;484:53–62. 144. Mayes AG, Andersson LI and Mosbach K. Sugar binding polymers showing high anomeric and epimeric discrimination obtained by non-covalent molecular imprinting. Anal Biochem 1994;222:483–488. 145. Byrne ME, Oral E, Hilt JZ and Peppas NA. Networks for recognition of biomolecules: molecular imprinting and micropatterning poly(ethylene glycol)-containing ﬁlms. Polym Advan Technol 2002;13:798–816. 146. Kataoka K, Miyazaki H, Bunya M, Okano T and Sakurai Y. Totally synthetic polymer gels responding to external glucose concentration: their preparation and application to on–off regulation of insulin release. J Am Chem Soc 1999;120:12694–12695. 147. Wulff G and Schauhoff S. Enzyme-analog-built polymers. 27. Racemic-resolution of free sugars with macroporous polymers prepared by molecular imprinting – selectivity dependence on the arrangement of functional-groups versus spatial requirements. J Org Chem 1991;56:395–400. 148. Norrild JC and Sotofte I. Crystal structures of 2-(N,N-dimethylaminoalkyl) ferroceneboronic acids and their diol derivatives. The quest for a B–N intramolecular bond in the solid state. J Chem Soc Perkin Trans 2001;2:727–732. 149. Wizeman WJ and Koﬁnas P. Molecularly imprinted polymer hydrogels displaying isomerically resolved glucose binding. Biomaterials 2001;22:1485–1491. 150. Davidson L and Hayes W. Molecular imprinting of biologically active steroidal systems. Curr Organic Chem 2002;6:265–281. 151. Gore MA, Karmalkar RN and Kulkarni MG. Enhanced capacities and selectivities for cholesterol in aqueous media by molecular imprinting: role of novel crosslinkers. J Chromatrogr B 2004;804:211–221. 152. Hishiya T, Shibata M, Kakazu M, Asanuma H and Komiyama M. Molecularly imprinted cyclodextrins as selective receptors for steroids. Macromolecules 1999;32:2265–2269. 153. Egawa Y, Shimura Y, Nowatari Y, Aiba D and Juni K. Preparation of molecularly imprinted cyclodextrin microspheres. Int J Pharm 2005;293:165–170.

268 154. Zhong N, Byun HS and Bittman R. Hydrophilic cholesterol-binding molecular imprinted polymers. Tetrahedron Lett 2001;42:1839–1841. 155. Pe´rez N, Whitcombe MJ and Vulfson EN. Surface imprinting of cholesterol on submicrometer core–shell emulsion particles. Macromolecules 2001;34:830–836. 156. Wang SF, Xu JW, Tong YJ, Wang L and He CB. Cholesterol-imprinted polymer receptor prepared by a hybrid imprinting method. Polym Int 2005;54:1268–1274. 157. Hwang CC and Lee WC. Chromatographic characteristics of cholesterol-imprinted polymers prepared by covalent and non-covalent imprinting methods. J Chromatogr A 2002;962:69–78. 158. Sreenivasan K. Chromatographic characteristics of cholesterol-imprinted polymers prepared by covalent and non-covalent imprinting methods. J Appl Polym Sci 2001;82:889–893. 159. Green BS and Priwler M. Molecularly imprinted polymers for the treatment and diagnosis of medical conditions. US Patent 2002/0015690 A1, February 7, 2001. 160. Huval CC, Bailey MJ, Braunlin WH, Holmes-Farley SR, Mandeville WH, Petersen JS, Polomoscanik SV, Sacchiro RJ, Chen X and Dhal PK. Novel cholesterol lowering polymeric drugs obtained by molecular imprinting. Macromolecules 2001;34:1548–1550. 161. Yavuz H, Arica Y and Denizli A. Therapeutic afﬁnity adsorption of iron(III) with dyeand ferritin-immobilized pHEMA adsorbent. Mat Sci Eng C-Bio S 2005;25:521–528. 162. Villamena FA and De La Cruz AA. Caffeine selectivity of divinylbenzene crosslinked polymers in aqueous media. J Appl Polym Sci 2001;82:195–205. 163. Lanza F and Sellergren B. Method for synthesis and screening of large groups of molecularly imprinted polymers. Anal Chem 1999;71:2092–2096. 164. Takeuchi T, Fukuma D and Matsui J. Combinatorial molecular imprinting: an approach to synthetic polymer receptors. Anal Chem 1999;71:285–290. 165. Tovar GEM, Krauter I and Gruber C. Molecularly imprinted polymer nanospheres as fully synthetic afﬁnity receptors. Top Curr Chem 2003;227:125–144. 166. Kirchner R, Seidel J, Wolf G and Wulff G. Calorimetric investigation of chiral recognition processes in a molecularly imprinted polymer. J Incl Phenom Macro 2002;43:279–283. 167. Chen X, Lin Y, Liu M and Gilson MK. The binding database: data management and interface design. Bioinformatics 2002;18:130–139. 168. Spivak DA and Shea KJ. Investigation into the scope and limitations of molecular imprinting with DNA molecules. Anal Chim Acta 2001;435:65–74. 169. Slinchenko O, Rachkov A, Miyachi H, Ogiso M and Minoura N. Imprinted polymer layer for recognizing double-stranded DNA. Biosens Biolectron 2004;20:1091–1097. 170. Nagahori N and Nishimura SI. Tailored glycopolymers: controlling the carbohydrate–protein interaction based on template effect. Biomacromolecules 2001;2:22–24. 171. Gabor F, Bogner E, Weissenboeck A and Wirth M. The lectin–cell interaction and its implications to intestinal lectin-mediated drug delivery. Adv Drug Del Rev 2004;56:459–480. 172. Duffy DJ, Das K, Hsu SL, Penelle J, Rotello VM and Stidham HD. Binding efﬁciency and transport properties of molecularly imprinted polymer thin ﬁlms. J Am Chem Soc 2002;124:8290–8296. 173. Schneider F, Piletsky S, Piletska E, Guerreiro A and Ulbricht M. Comparison of thinlayer and bulk MIPs synthesized by photoinitiated in situ crosslinking polymerization from the same reaction mixtures. J Appl Polym Sci 2005;98:362–372.

269

Glucocorticoid action and the development of selective glucocorticoid receptor ligands Timothy J. Cole Department of Biochemistry and Molecular Biology, Monash University, Clayton, Victoria 3800, Australia Abstract. Glucocorticoids are important endocrine regulators of a wide range of physiological systems ranging from respiratory development, immune function to responses to stress. Glucocorticoids in cells activate the cytoplasmic glucocorticoid receptor (GR) that dimerizes, translocates to the nucleus and functions as a ligand-dependent transcriptional regulator. Synthetic glucocorticoids such as dexamethasone and prednisolone have for decades been the cornerstone for the clinical treatment of inﬂammatory diseases, such as rheumatoid arthritis and asthma, and in some lymphoid cancers, yet its prolonged use has undesirable side effects such as obesity, diabetes, immune suppression and osteoporosis. Detailed knowledge on the mechanism of GR action has led to the development of novel selective glucocorticoid receptor modulators (SGRMs) that show promise of being efﬁcacious for speciﬁc treatments of disease but with fewer side effects. SGRMs promote speciﬁc recruitment of transcriptional co-regulators that elicit speciﬁc gene responses and show promise of greater efﬁcacy and speciﬁcity in treatment of inﬂammatory diseases and type-2 diabetes. Keywords: glucocorticoids; glucocorticoid receptor; immune function; inﬂammatory disease; type-2 diabetes; steroid ligand; selective glucocorticoid receptor modulator; ligand-binding domain.

Introduction Glucocorticoid steroids Glucocorticoids are members of the family of steroid hormones that are synthesized and secreted by specialized endocrine cells in the adrenal gland. The principal physiological glucocorticoid in humans is cortisol whereas corticosterone is in rodents. Glucocorticoids are derived from the four-ring steroid cholesterol and exert effects on embryonic development and in the coordinated maintenance of general physiological homeostasis [1,2]. Glucocorticoids have effects in many tissues, helping to regulate processes such as metabolic homeostasis, central nervous system (CNS) function and modulation of the immune response. Glucocorticoids are secreted by the adrenal cortex, together with another important class of steroid, the mineralocorticoids whose major role is to help regulate the balance of systemic electrolytes, such as sodium and potassium [1].

Corresponding author:

E-mail: [email protected] (T.J. Cole). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12008-6

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

270 Cortisol (4-pegnen-11b, 17a, 21-triols-3, 20-dione) is comprised of three 6-carbon hexane rings and a 5-carbon pentane ring backbone, which is characteristic of all steroid hormones [3]. The presence of the keto-oxygen on carbon number 3 and the hydroxylation of carbon numbers 19 and 21 are responsible for the glucocorticoid activity of cortisol [1]. The naturally occurring glucocorticoids such as cortisol and corticosterone as well as synthetic glucocorticoids such as dexamethasone, methylprednisolone and prednisolone have been developed and utilized for decades as therapeutic drugs in the clinic. Structures for the physiological glucocorticoids and some synthetic agonists and antagonists are shown in Fig. 1. A commonly used glucocorticoid antagonist is the synthetic compound RU486 which is also an antagonist of the progesterone receptor and is an active ingredient in a widely available oral abortion pill [4]. Recently, this compound has been tested for the treatment of depression [5]. Glucocorticoids are synthesized primarily from cholesterol that is taken up directly from circulating blood, although small amounts of cholesterol can also be synthesized within adrenal cortical cells from acetyl coenzyme A. Cholesterol is taken up by the mitochondria of adrenocortical cells and cortisol is synthesized by a series of speciﬁc steroidogenic enzymes in the zona fasciculate and zona reticularis of the adrenal cortex. The primary positive regulator of glucocorticoids synthesis and secretion is the adrenocorticotropic hormone (ACTH) that is released from the anterior pituitary of the brain. ACTH stimulates adrenocortical cells to synthesize glucocorticoids by binding to the cell surface of ACTH receptor and activating adenyl cyclase in the cell membrane. This results in the formation of cyclic-adenosine monophosphate (cAMP), which then increases the activity of intracellular biosynthetic enzymes, such as the steroidogenic cytochrome P-450scc(side-chain cleavage) to help catalyse the formation of the adrenal steroids [1]. Plasma levels of glucocorticoid hormones are controlled via negative feedback mediated by the occupancy of glucocorticoid receptors (GR) at the level of the hypothalamus and pituitary, which regulates ACTH release, rather than from feedback resulting from target cells. ACTH also enhances the production of adrenal androgens [6]. The release of ACTH from the pituitary is controlled by corticotropin-releasing hormone (CRH) produced and secreted by the hypothalamus [1]. Cortisol circulates in the blood in three main forms: protein-bound, ‘‘free’’ cortisol or as cortisol conjugates. ‘‘Free’’ cortisol is the form of the physiologically active hormone and is able to act directly at target tissues. Only 5% of cortisol is in this ‘‘free’’ form. Nearly 90–95% are protein bound. The normal concentration of cortisol in the blood, on an average, is 12 mg dL1 whereas the cortisol exists in the protein-bound form. There are two known cortisol-binding factors in plasma. These are the high-afﬁnity, low-capacity alpha2 globulins called transcortin or cortisol-binding globulin (CBG) and the low-afﬁnity, high-capacity protein albumin [7]. Cortisol is degraded in the liver where it is

271

Fig. 1. Chemical structures of common human physiological glucocorticoids (corti-

sol and cortisone) and synthetic glucocorticoid agonists (prednesilone, prednisone, methylprednesilone, dexamethasone, triamcinolone and ﬂudrocortisone) used in clinical treatment of disease.

272 conjugated to form glucuronide and sulphate derivatives [8]. This conjugated form of cortisol is inactive and is excreted via the bile, faeces and urine. Glucocorticoid receptors Target cells respond to glucocorticoids via speciﬁc intracellular cytoplasmic protein receptors that mediate the action of glucocorticoids to produce functional changes within cells. Glucocorticoids have binding afﬁnities for two types of intracellular nuclear receptors, initially termed the type I (now known as the mineralocorticoid receptor (MR)) and type II (now termed the GR). The GR is expressed to varying degrees in all cell types so far examined and glucocorticoid steroids are its major physiological ligand. The type I or MR also binds mineralocorticoids such as the adrenal steroid aldosterone with high afﬁnity, and the spatial expression of MR is much more restricted [9]. The MR is present at high levels in tissues containing epithelial layers such as the kidney, colon, the parotid gland, salivary glands, ocular tissues, lung, osteoblasts, haemopoetic cells and some regions of the small intestine [10,11] where its physiological ligand is aldosterone. The presence of MR can also be detected in non-epithelial tissues such as in the heart and brain where their functional role is less well understood. A potential role of aldosterone in the pathogenesis of heart failure has recently been suggested from the results of the Randomized Aldactone Evaluation Study (RALES) [12]. In the RALES trial, patients with moderately severe heart failure were given (on top of normal therapy) a low dose of spironolactone, an MR antagonist. Over 2 years of the trial, this produced a 30% reduction in mortality and a 35% reduction in morbidity, and caused the trial to be stopped prematurely. The implied pathology of aldosterone via MR in heart failure, either direct or indirect, is poorly understood. The primary role of aldosterone in epithelia such as in the kidney collecting duct or distal convoluted tubule is to facilitate the retention of sodium and water, thereby helping to maintain blood volume and pressure. In spite of circulating levels of plasma cortisol at 2–3 orders of magnitude higher than that of aldosterone, it has been clearly demonstrated that MR in epithelia will bind and respond to aldosterone in mineralocorticoid target tissues in vivo. The mechanism providing protection of MR in epithelia from higher levels of glucocorticoids remained unknown until it was demonstrated that MR-containing epithelia also express an enzyme that is able to effectively modify and inactivate glucocorticoids. This enzyme was shown to be 11b-hydroxysteroid dehydrogenase type II, which is necessary for conferring aldosterone speciﬁcity on MR in mineralocorticoid target tissues, by converting biologically active cortisol into inactive cortisone in humans, and corticosterone to 11-dehydrocorticosterone in rodents [13,14]. GR and MR are members of the steroid receptor subfamily of the nuclear receptor superfamily (Table 1) and share a common domain structure

273 Table 1. Members of the human nuclear receptor superfamily of receptors. ‘‘Known ligand’’ receptor

‘‘Orphan’’ Receptor

TR a, b

RevErb a, b

Reverse ErbA

ROR/RZR a, b, g UR COUP-TF a, b, g

Retinoid orphan/Z receptor Ubiquitous receptor COUP-transcription factor

HNF-4

Hepatocyte nuclear factor 4 Tailess-related receptor Photoreceptor nuclear receptor Testis receptor Estrogen-related receptor Nurr77/Nor/NGFI receptors Steroidogenic factor 1 Germ cell nuclear factor Small heterodimeric partner Dosage-sensitive sex reversal

RAR a, b, g VDR PPAR a, b, g, d PXR CAR a, b

Thyroid hormone receptor Retinoic acid receptor Vitamin D receptor Peroxisome proliferator activated receptor Pregnane X receptor

TLX

LXR a

Constitutive androstane receptor Liver X receptor

FXR RXR a, b, g

Farnesoid X receptor Retinoid X receptor

TR2 a, b ERR a, b, g

GR

Glucocorticoid receptor Androgen receptor Progesterone receptor

NGFI-B a, b, g SF1 a, b GCNF

Mineralocorticoid receptor Estrogen receptor

SHP

AR PR MR ER a, b

PNR

DAX-1

(Fig. 2). Both receptors are characterized by a central DNA-binding domain (DBD), which functions to target the receptor to speciﬁc DNA sequences known as hormone response elements (HREs). The DBD of nuclear receptors is made up of two highly conserved zinc ﬁnger motifs. The GR and MR have high almost identical DBDs and therefore are able to bind to identical HREs. The HRE is comprised of two DNA half sites organized as inverted repeats and mutation of these repeats prevents GR binding [15,16]. The C-terminal end of steroid hormone receptors has a ligand-binding domain (LBD) and this provides the essential property of binding hormone. The N-terminal part of the receptor, also termed the A/B domain, is involved in transactivation of target genes after the receptor complex is bound to DNA. In the unliganded state, GR forms an inactive cytoplasmic multi-protein complex with heat-shock proteins, such as HSP 90, HSP 70 and HSP 56. These chaperones play a role in assisting the proper folding of GR into a conformation optimal for hormone binding and act as molecular chaperones

274 1

NTD 100%

DBD 100%

LBD 100%

777 GR

<15%

94%

60%

MR

<15%

92%

50%

PR

<15%

86%

50%

AR

<15%

52%

24%

ERα

<15%

52%

23%

ERβ

Fig. 2. Structural comparison for members of the steroid receptor subfamily from the nuclear receptor superfamily. NTD, N-terminal domain; DBD, DNA-binding domain; LBD, ligand-binding domain; GR, glucocorticoid receptor; MR, mineralocorticoid receptor; PR, progesterone receptor; AR, androgen receptor; ERa, estrogen receptor alpha; and ERb, estrogen receptor beta. Percentage identity at the amino acid level is related to GR (set at 100%).

CYTOPLASM

Functional change

NUCLEUS Protein mRNA

GR Binds Dimerize

GRE

GENE

Glucocorticoid

Fig. 3. General mechanism of action for glucocorticoids in target cells. Cortisol

enters target cells by passive diffusion across the cell membrane where it binds the ligand-binding domain of the GR initiating dimerization and translocation to the nucleus. The activated GR dimer binds to speciﬁc glucocorticoid response elements (GRE) and increases the rate of transcription. This produces an elevation of protein level within the cell and eventual physiological change.

preventing the unoccupied receptor from translocating to the nucleus [17,18]. Upon ligand binding, the activated GR or MR is released, dimerizes and translocates to the nucleus where it binds to speciﬁc DNA HREs in the promoter control regions of target genes and alters the rate of gene transcription (see Fig. 3). Altered expression of these gene products then brings

275 about an alteration in cell and eventually tissue function. Speciﬁcity of DNA binding is achieved by interaction of the P-box in the ﬁrst zinc ﬁnger of the GR DBD with the palindromic hormone or GRE of the target DNA (Fig. 4). GR and MR may also have the ability to heterodimerize with each other in tissues co-expressing both receptors [18]. This may prove to be an important mechanism as heterodimerization of both corticosteroid receptors expands the potential of these steroid hormones to regulate a larger range of responsive target genes. GR also has several other modes of actions. First, GR can carry out DNA binding-dependent repression of several genes, such as proopiomelanocortin (POMC) and octeocalcin, by binding to speciﬁc negative GREs [19]. Second, GR can also modulate the transcriptional activity of certain genes via non-DNA binding protein–protein interactions, such as with the AP-1 [20] and NF-kB transcription factor complexes [21]. Such regulation involves direct protein–protein interaction between GR and speciﬁc subunits of these complexes and is commonly known as ‘‘transrepression’’. Finally, several other mechanisms have been proposed for the possible further cross-talk of GR with other transcription factors such as c-Jun and Stat-5 (extensively reviewed in [22]). Low cortisol

SRC-1 GRIP-1

GR

GR

(co-activators) SRC-2/3

(basal only)

PGC-1

(cortisol)

CBP/300

RNAPolII TATAA Basal transcription of target gene

GRE

High cortisol GR

GR SRC-1 SRC-2/3 GRIP-1 PGC-1

GR GRE

GR

CBP/300

(co-activators)

(basal + induced)

(+ve) RNAPolII TATAA Transcription of target gene

Fig. 4. Mechanism of action for activation of target genes by GR in the cell nucleus. In the cytoplasm, cortisol binds the LBD of the GR initiating dimerization and translocation to the nucleus. Nuclear cortisol-bound GR binds to speciﬁc GRE sites in DNA upstream of target genes that allow recruitment of co-activator proteins (such as CBP/300, GRIP-1, PGC-1 or SRC1/2/3), which promote an increase in initiation rates of transcription by RNA polymerase II at the target gene promoter. (a) At low cortisol levels, GR is not activated sufﬁciently to promote DNA binding and cortisol-induced transcription of target genes. (b) At high cortisol levels, GR is activated, GR dimers bind at GRE DNA-binding sites and recruit co-activators that increase rates of transcription at target gene promoters.

276 Some steroids can use their intracellular receptors as well as act in certain tissues by the use of a membrane-bound receptor (reviewed in [23]). While the genomic effects of intracellular steroid receptors are mediated at the levels of gene transcription and protein expression over a course of hours, membrane receptors can alter intracellular, non-genomic, signalling pathways in a matter of minutes. The ﬁeld of rapid non-genomic effects of steroids mediated by putative membrane-bound steroid receptors is relatively new and remains somewhat controversial [24]. A number of different studies have reported potential roles for membrane receptor signalling in response to progesterone, estrogen, androgens and glucocorticoids, but it is not certain if the receptors mediating these effects are different forms of the intracellular nuclear receptors or a different distinct family of steroid-binding proteins [23,25]. A recent study isolated and characterized a membrane progesterone receptor in the most biologically relevant model of membrane receptor action studied so far, the steroid-induced maturation of oocytes in the freshwater trout [26]. This receptor was found to be a novel membrane-associated steroid-binding protein, genetically distinct from the intracellular nuclear progesterone receptor, and importantly, was able to rapidly activate the MAP kinase ERK. Studies examining the non-genomic actions of glucocorticoids have described membrane GR activity in frog neuronal cells, rat pituitary cells and T cells [25,27,28]. The nature of the glucocorticoid-binding membrane receptor remains controversial, however as several studies suggested that the classical intracellular GR can also function as a membrane receptor [27], yet another describes it as a separate G-protein coupled receptor [25]. Here the activation of MAP kinase activity, mediated by membrane receptors, was reported within 30 min of glucocorticoid treatment. The role of glucocorticoids in maintaining normal systemic physiology In mammals, gluococorticoids regulate a range of physiological systems and in the face of regular and sometimes sudden changes to the external environment help to maintain physiological homeostasis. They play an important role during embryonic development, particularly in the ﬁnal stages of lung maturation prior to birth. In an adult, they help to regulate metabolism, strongly repress the activated immune system and regulate brain function during the response to stress [2]. These and other physiological systems that glucocorticoids modulate are depicted in Fig. 5. Embryonic lung development The development and growth of the mammalian embryo is a complex and highly organized process involving a combination of intrinsic growth and differentiation factors, and concerted actions from circulating factors and hormones. Glucocorticoids are one of a number of circulating hormones that

277

Adipose - differentiation - metabolism

Embryonic Development - lung, CNS, gut

Endocrine Systems - HPA axis - gonadal axis

Cardiovascular - salt/water flux - blood pressure

Immune suppression - anti-inflammatory - T cell death

Metabolism - gluconeogenesis - protein turnover - lipolysis

Bone - formation - Ca2+ metabolism

Brain - stress responses - memory consolidation

Fig. 5. The major physiological functions of cortisol in the human body. CNS, cen-

tral nervous system; HPA, hypothalamus–pituitary–adrenal.

also include retinoids, thyroid hormone and other cAMP-mediated factors that are important during the ﬁnal embryonic stages of lung development. These hormones play key roles during late gestation, particularly in the differentiation and development of terminal respiratory alveoli and the stimulation of lung surfactant production. Endogenous glucocorticoid levels increase in the foetus rapidly just prior to birth and are well known to participate in the regulation of biochemical and cytoarchitectural changes in the developing foetal lung. Yet little is known of the underlying molecular and cellular events that underpin these glucocorticoid-mediated effects [29]. To investigate the role of glucocorticoid action via GR signalling during embryonic development, the GR gene has been ablated using gene targeting in mice, which caused a number of phenotypic effects [30]. The most striking observation was seen in the lung where GR-null mice at birth displayed severe lung atelectasis with little to no inﬂation of lung tissue. On a C57Bl6/ 129sv genetic background, greater than 90% of GR-null mice die at birth from respiratory dysfunction, whereas all GR-null mice die at birth on a 129sv isogenic genetic background. An identical phenotype of perinatal death and lung dysfunction has been described in another GR-null mouse constructed by gene-targeted deletion of exon 2 and the proximal promoter of the GR gene [31]. A more recent detailed analysis of the developing lung in the ﬁrst GR-null mouse has shown that glucocorticoid signalling via GR is not essential for surfactant production and an analysis of alveolar epithelial cell (AEC) types using electron microscopy and cell-speciﬁc markers, revealed a marked reduction in differentiated type-I AECs prior to birth [32]. This indicated that GR signalling is an important epithelial cell differentiation signal, which in its absence leads to respiratory dysfunction due to a

278 severe reduction in the ability of the lung to mount and mediate appropriate gas exchange. An almost identical lung phenotype is seen in CRH-deﬁcient mice (of CRH-deﬁcient mothers) in which there is a clear impairment in glucocorticoid production both in the foetus and mother [33]. CRH-null mice have delayed induction of surfactant protein-A and surfactant protein-B with reduced pulmonary septal thinning and airway formation. Interestingly, mice with a gene-targeted point mutation in GR, which prevents dimerization and DNA binding, develop normally with no overt lung defect, indicating that the actions of glucocorticoids in the developing lung may be mediated by DNA-binding independent actions [34]. A recent study using GR-null mice with deletions of exon 2 and the proximal promoter has found a marked reduction in the growth factor midkine, 3 days before birth, which may contribute to the immature lung phenotype detected at birth [35]. It is not then surprising that synthetic glucocorticoids (particularly betamethasone or dexamethasone) are widely used antenatally to reduce the severity of the respiratory distress syndrome (RDS) suffered by very preterm infants, and act to accelerate foetal lung maturation and increase lung surfactant production [36]. Antenatal glucocorticoid treatment has had a major beneﬁt in reducing the incidence of neonatal RDS and intraventricular haemorrhage, leading to decreased neonatal mortality. Its use, however, remains controversial, particularly the administration of multiple doses, due to the reported side effects of glucocorticoids on lung and body growth, and the postnatal development of the CNS [37,38]. Maintaining systemic homeostasis Glucocorticoids are important components of the endocrine response to systemic stress that also includes the catecholamines, CRH and a large number of cytokines [2,39]. Glucocorticoids are central in priming systems for the ﬁght or ﬂight response and crucial in mediating a return to homeostasis. The brain is essential for orchestrating our response to stress and glucocorticoid hormones are important mediators of neural stress response systems [40,41]. In response to stress, the brain initiates physiological and behavioural responses with many of these mediated by glucocorticoid effects in the hippocampus and amygdala [41]. Short-term responses include promoting and improving memory, and increasing acuity. Long-term chronic stress can produce atrophy and damage to the hippocampus, a response mimicked by high glucocorticoid administration in animals. Although GRs are expressed in most cells of the brain, both GR and MR are co-expressed at the highest levels in limbic neurones that include the hippocampal CA1 and dentate gyrus as well as the nuclei of the amygdala and medial prefrontal cortex [42,43]. In these areas, they have important roles in the control of emotion and cognition. The hippocampus is very important

279 in memory formation and consolidation. In neurons of the hippocampus, MR acts principally as a receptor for glucocorticoids, having effects on memory, behaviour and stress responses. Blockade of GR by selective antagonists has shown that GR is important for memory consolidation [44]. These results have been conﬁrmed with gene-targeted GR knockout mice where memory formation was impaired [45]. Recent over-expression of GR in the pre-frontal cortex of mice produced increased anxiety-like behaviour akin to bipolar disorder [46]. Clearly changes in the speciﬁc levels of GR in regions of the brain, either pharmacologically or via increase or decrease of gene expression will affect stress-induced brain function. Glucocortcioids can affect a range of processes within the CNS such as cellular metabolism, electrophysiological properties, neuroendocrine secretion, mood, memory, mental performance, cell maturation and survival within different regions of the brain [47–51]. Glucocorticoids have important inﬂuences on brain development and are considered to be a signiﬁcant factor in brain damage and aging [47]. Glucocorticoids also affect gliogenesis, and glial cell differentiation and maturation [52]. For example, when glucocorticoids are given during the period of rapid gliogenesis, brain growth, the acquisition of DNA and myelination are inhibited [52]. Additional roles of glucocorticoids in the function of hippocampal neuronal excitability, memory and modulation of cognitive function in rats and humans have been described in [53]. Cortisol is released from the adrenal gland in response to a range of stressors that include trauma, infection, intense heat or cold, injection of drugs (norepinephrine, necrotizing substances and other sympathomimetic drugs), surgery and physical restraint [54]. The HPA axis is the neuroendocrine system that responds to stress and the consequences of chronic or acute stress are dependent on the modulated response of the HPA axis [4]. The responsiveness of the HPA axis to stress is dependent on the ability of corticosteroids to regulate CRH and ACTH release [55]. Much evidence suggests that glucocorticoids inhibit HPA activity, not only at the hypothalamus and pituitary, but at higher CNS centers. Lesions of the hippocampus, which contain the highest density of both MR and GR, were associated with failed attenuation of the HPA stress response, circadian rhythm and increased CRH and vasopressin expression [56]. The importance of the GR in negative feedback sensitivity of the HPA axis has been demonstrated in tissue-speciﬁc GR knockout mice that lack the GR throughout the CNS [57]. A lack of GR function in the nervous system impaired the regulation of the HPA axis in these mice, resulting in increased plasma glucocorticoid levels that lead to similar clinical symptoms observed in Cushing’s syndrome. Apart from GR, the enzyme 11b-hydroxysteroid dehydrogenase 1 (11bHSD1) can also modulate the regulation of the HPA axis as demonstrated in 11b-HSD1-null mice. 11b-HSD1 converts inactive cortisone into cortisol, thereby amplifying glucocorticoid-mediated effects within cells. 11b-HSD1-null

280 mice have altered HPA axis function both basally and in response to stress [58]. In 11b-HSD1-null mice, there is no regeneration of active glucocorticoids in the CNS and hence it would result in a compensatory increase in basal corticosterone secretion. This was apparent in the 11b-HSD1-null mice as they have adrenocortical hyperplasia and increased sensitivity to ACTH. However, basal plasma corticosterone should be unaltered if HPA sensitivity to glucocorticoids is unchanged. In 11b-HSD1-null mice, basal hypersecretion of glucocorticoids and increased basal HPA axis activity was observed. This implied that the increased basal HPA axis activity might be due to either increased drive and/or attenuated feedback control. The 11b-HSD1-null mice have reduced sensitivity to glucocorticoid negative feedback upon the HPA axis; however, these mice can maintain circadian rhythmicity and good HPA responses to stress, suggesting that stimulatory pathways are intact. Glucocorticoids and integrated metabolism In general, glucocorticoids are catabolic and have signiﬁcant effects on carbohydrate, protein and lipid metabolism. In the metabolic response to stress or after chronic glucocorticoid administration, there is a rapid increase in blood glucose levels. This is promoted in a number of ways. Glucocorticoids alter glucose metabolism by both decreasing glucose utilization in the periphery and increasing glucose production and availability via stimulating hepatic gluconeogenesis [2]. In the liver, glucocorticoids stimulate gluconeogenesis via increasing the transcription rates of genes such as phosphoenolpyruvate carboxykinase, tyrosine amino transferase and glucose-6-phosphatase all of which are key enzymes in the gluconeogenic pathway [59]. Glucocorticoids also mobilize amino acids from peripheral tissues such as muscle and adipose, and also reduce the rate of glucose utilization in cells by decreasing the oxidation of NADH, thereby lowering the rate of glycolysis. Glucocorticoids are also known to reduce protein synthesis and increase catabolism of protein stores in many cells of the body by repressing the formation of RNA and the transport of amino acids in many peripheral tissues such as muscle and lymphoid tissue. Glucocorticoids also stimulate the mobilization of fatty acids from adipose tissue. The dramatic and counter-regulatory effects that glucocorticoids have on insulin action suggest that blockade of glucocorticoid effects via antagonizing the GR may be a potential treatment for diabetic hyperglyceamia. Glucocorticoids and the immune system An immune defence against a variety of harmful pathogens is critical to survival in an ever-changing biological environment. Mammals have a highly developed and exquisitely ornate immune system that is able to mount a robustly activated response to pathogenic challenge. Activated immune

281 responses, however, if left to continue unchecked can be detrimental to the host. Therefore, there are important molecular mechanisms that are activated to ﬁnally suppress immune responses. Glucocorticoids are key players in switching off immune responses, particularly in aspects driven by proinﬂammatory cytokines. From moderate to high systemic levels, glucocorticoids have a potent immunosuppressive effect [17,39]. This is achieved not only by blocking the synthesis of pro-inﬂammatory cytokines but also by blocking their systemic effects. Glucocorticoids inhibit the synthesis of certain cytokine receptors such as the IL-2 receptor [60]. Also, they can inhibit the cellular effects of certain cytokines such as IL-2 and IL-6, by repressing the transcription factors AP-1 and NF-kB, hence repressing inﬂammation [61]. Cortisol decreases the number of eosinophils and lymphocytes in the blood and a major effect of cortisol is the inhibition of T-cell proliferation [62]. Glucocorticoids also induce apoptosis of immature thymocytes (immature thymic T cells) and some B cells via apoptosis [63]. One of the most common clinical uses of glucocorticoids is for the treatment of rheumatoid arthritis, a discovery and treatment that led to the awarding of the 1950 Nobel Prize for Physiology and Medicine to Hench, Kendall and Reichstein [64]. Synthetic glucocorticoids are potent inhibitors of inﬂammation, allergy and endotoxic shock [65]. They inhibit the synthesis of cytokines, such as tumour necrosis factor (TNF) and many cytokines (i.e., IL-1 and IL-6) in macrophages, which fuel immune and inﬂammatory responses, and also interact with lipocortin-1, a member of the annexin family of calcium and phospholipid-binding proteins [66]. Studies on the inﬂuence of glucocorticoids on the immune system have focussed on their role during stress or following pharmacological administration (reviewed in [17]). There are clear differences in potency between natural and synthetic glucocorticoids. For example, the most commonly used glucocorticoid in studies of the immune system, dexamethasone, is a synthetic hormone that exerts 25 times greater suppression of IL-6 or IL-1 levels than a comparable amount of the naturally occurring glucocorticoid, cortisol [67,68]. Studies that have examined normal physiological concentrations of glucocorticoids demonstrate that glucocorticoids prime the immune system and aid in its function during the early phases of the immune response. These studies have relied upon adrenalectomized animals to examine T cell responses in the absence of adrenal hormones. T cells from adrenalectomized rats were found to respond weakly to concanavalin A; however, the replacement of low-level glucocorticoids restored T-cell responses to that observed in sham-operated rats [69]. Similarly, a low level of glucocorticoid administered before endotoxin injection actually increased the subsequent secretion of TNF-a and IL-6 [70]. IL-1, IL-6 and TNF-a are pro-inﬂammatory cytokines that are released at high levels by multiple cell types during a stress response. These molecules are the most potent stimulators of CRH production and can induce high levels of glucocorticoid production within hours

282 [71,72]. These stress-induced levels of glucocorticoids cause a profound atrophy of the thymus, and also diminish the size of peripheral immune tissues; yet all other tissues remain intact [2,73]. At high stress-induced levels, glucocorticoids act to dampen and reduce the immune response at all levels, including preventing the further driving of cytokine activation, the priming of lymphocytes by APCs and reducing the responsiveness of already activated cells. Glucocorticoids have a uniform inhibitory effect on the synthesis, release and efﬁcacy of the pro-inﬂammatory cytokines IL-1, IL-2, IL-6, IL-8, IL-11, IL12, IFN-g and TNF-a [74]. Glucocorticoids also inhibit cytokine effects in target cells, and in the case of IL-2, IL-4, IL-7 and IL-15, do so by acting directly on the JAK–STAT pathway component STAT5 [60]. Separate studies have also reported that the GR can interact with STAT3 and STAT5 to reduce IL-6 signalling [75,76]. It is currently not known if diurnal levels of GCs inﬂuence cytokine signalling during T-cell development and homeostasis, or if these effects are conﬁned to the stress response. Glucocorticoids also inhibit antigen processing and expression of MHC class II by dendritic cells as well as prevent their migration by reducing adhesion molecule expression [77,78]. The interaction of GR with two particular transcription factor complexes, activator protein-1 (AP-1) and NF-kB, mediates most of the antiinﬂammatory effects of glucocorticoids and also explains the effects of glucocorticoids on genes that do not contain speciﬁc GREs in their promoters. AP-1 is primarily composed of two subunits named c-Fos and c-Jun, and is a proinﬂammatory transcription factor inducible by a number of cytokines [79]. The activated GR can interact with c-Jun, in a protein–protein interaction, termed transrepression, which prevents its interaction with c-Fos [80]. Furthermore, the c-Jun/GR heterodimer can still bind to AP-1 response elements in promoter regions to prevent the transcription of AP-1 target genes. This was demonstrated by the fact that glucocorticoid-mediated repression of cJun expression, whose promoter does not contain any GREs, required intact AP-1-binding sites [81]. A second study showed a similar process in the POMC promoter, which also lacks a GRE but is inhibited by glucocorticoid treatment. Here it was demonstrated that the GR could directly bind to Nur77, another member of the nuclear receptor superfamily, and subsequently to its response elements, to prevent POMC transcription [82]. A separate mechanism for gene suppression utilizes a composite GRE, which contains a binding site for the GR closely associated with that of a different transcription factor. The CRH promoter contains an AP-1 site closely located to a GRE and binding to this composite GRE by an AP-1/GR complex is a potential mechanism for the suppression of CRH expression and thus the production of glucocorticoids [83]. Glucocorticoids can also inhibit the MAP kinase JNK by an unknown mechanism, which prevents the phosphorylation and thus activation of c-Fos/c-Jun dimers already present in the cell [84]. The inhibition of cytokine production is mediated mainly through the interaction of

283 the GR with NF-kB [85]. The GR can inhibit NF-kB actions by two mechanisms, the transcription of a family of speciﬁc NF-kB inhibitors (IkBs) or inhibition of NF-kB activity by direct protein–protein binding [78,86,87]. Interestingly, the exact residues in the GR DBD necessary for this direct interaction with NF-kB are not the same as those required to repress AP-1 [88]. The transcription factors NF-kB, CREB and NFAT are essential for T-cell activation, and at least in the case of NFAT, positive selection of doublepositive thymocytes in the thymus [89]. These factors are dependent on direct interaction with the transcriptional co-activators such as CBP/p300 to mediate their effects [90,91]. These co-activators are also essential for the transcriptional activation of the GR [92], leading to the suggestion that competition for co-activator binding may be a method of GR-mediated repression [93]. However, recent ﬁndings favour a model where GR directly binds to the RelA sub-unit of NF-kB, without preventing its binding to its speciﬁc DNA element, but represses the transcription of its target genes by direct protein interactions with the basal transcriptional machinery [94,95]. However, multiple recent studies involving overexpression of various transcriptional co-activators or mutation of co-activator binding sites in the GR demonstrated no loss of transrepression of AP-1 or NF-kB [96–98]. In addition, despite being unable to recruit co-activators, antagonist (RU486) bound GR has showed to maintain the ability to transrepress NF-kB [96]. Thus, despite their role as integrators of GR signalling and that of a variety of other transcription factors, present evidence leans strongly against a role for transcriptional co-activators as limiting factors in transrepression. It is currently not clear if glucocorticoids mediate these effects at their diurnal levels, or only during the stress response. Glucocorticoids, the thymus and T cells Apart from the ability of glucocorticoids to modulate many signalling pathways that activate T cells, there have been several separate observations suggesting a role for glucocorticoid signalling in the selection of immature T cells in the thymus. First, high glucocorticoid levels can cause pronounced atrophy of the thymus [99] and second, glucocorticoid signalling could rescue T cell hybridomas from the apoptosis caused by T-cell receptor stimulation and vice versa [100]. The major selection events that occur during thymocyte development are dependent on the level of TCR stimulation received. Thus, the fact that glucocorticoids could antagonize signalling from the TCR suggests that it is possible that glucocorticoids are involved in T-cell development. This was supported by studies that the cloned glucocorticoid-inducible genes from T cells, GILZ and GITR prevented TCR-induced cell death when expressed at high levels in T cells [101]. It was also observed by immunohistochemistry [102],or RT–PCR [103] that steroid biosynthetic enzymes appeared to be expressed in stromal cells of the thymus. Further studies in the

284 mouse thymus demonstrated using RT–PCR that every enzyme necessary to produce glucocorticoids was present, albeit at much lower levels compared to the adrenal [103]. A more recent report has investigated the effect of a conditional rat GR transgene that has been speciﬁcally expressed in mouse T cells using the tetracycline-inducible expression system with a human T cell-speciﬁc CD2 promoter [104]. They showed in adrenalectomized (to exclude systemic glucocorticoids) transgenic mice that there is a dramatic increase in thymocyte death by apoptosis after doxycycline-induction of the GR transgene, implicating induction of T-cell death from endogenously produced glucocorticoids. This was replicated in vitro using a thymic organ culture system and the increased death in vivo and in vitro could be at least partially blocked by the GR antagonist RU486. These results give compelling evidence for the presence of thymic-produced glucocorticoids that, at least in this transgenic system, are able to trigger signiﬁcant thymocyte apoptosis. While it is difﬁcult to imagine these low levels of enzymes producing substantial amounts of glucocorticoids, it was proposed that the thymic stroma might only need to produce glucocorticoids in speciﬁc microenvironments where the local concentrations created may be sufﬁciently high to antagonize TCR signalling. Initial experiments which directly tested the role of glucocorticoids in thymocyte development, reduced GR signalling in vitro by using metyrapone, a speciﬁc antagonist of 11b hydroxylase, the ﬁnal steroidogenic enzyme required to synthesize GC from its precursor deoxycorticosterone. An addition of metyrapone, or RU486 an antagonist of the GR and PR, to foetal thymic organ culture caused large reductions in double-positive thymocyte number with development of very few mature thymocytes, and suggested defective positive selection [102]. To further investigate a possible role of glucocorticoids in T cell selection, transgenic mouse lines were created that expressed the 30 untranslated region of the rat GR mRNA molecule in an antisense orientation under the control of the T cell-speciﬁc Lck promoter [105]. Thymocytes from these mice possessed a 50% reduction in GR mRNA and displayed defective thymocyte development in the adult, which could be detected from as early as day 16 of gestation. Both TCR-b selection (a proliferative stage between the doublenegative and double-positive thymocytes) as well as positive selection appeared to be impaired, leading to a 10-fold reduction in double-positive thymocyte number, with very few mature thymocytes or peripheral T cells. To explain these effects on thymocyte development, the mutual antagonism hypothesis was proposed (reviewed in [106]). This hypothesis postulated that TCR signalling in developing thymocytes caused a deletion-inducing signal, which could be antagonized by GR signalling. Only those thymocytes whose TCR signalling was balanced with GR signalling would be positively selected, while less or more TCR signalling than this positive selection threshold would cause glucocorticoid-mediated death by neglect or

285 TCR-mediated negative selection, respectively. A prediction of this hypothesis is that thymocytes that would normally be positively selected would be deleted if GR signalling was reduced. This prediction was conﬁrmed in two different models that found increased thymocyte sensitivity to deletion following inhibition of GR signalling. When antisense GR mice were crossed to the MRL background, the characteristic autoimmunity of this line was prevented [107]. Additionally, the pigeon cytochrome-c peptide antigen-speciﬁc T cells normally found in wild-type mice were deleted when crossed to antisense GR mice [108]. Despite this strong evidence of the role of glucocorticoids during the positive selection of thymocytes, there was contrasting data reported in a separate line of GR-antisense transgenic mice. Here, the same antisense construct was expressed under the control of the neuroﬁlament promoter, which also caused a 50% decrease in GR levels, yet a small increase in double-positive thymocytes was observed [109]. A separate group compared these transgenic antisense GR mice, and found that both of the promoters resulted in increased thymocyte numbers [110]. Separate studies also reported decreased thymocyte sensitivity to deletion when GR signalling was reduced by the GR antagonist RU486 [111,112]. Collectively, these experiments suggest a role for GR signalling in thymocyte selection and development. However, a number of recent experiments with mice lacking GR have shown these conclusions to be most likely incorrect. Studies using GR knockout or mutant mice have failed to show an inﬂuence of glucocorticoid signalling on thymocyte development, selection or survival [31,34,113–116]. The observations of normal thymocyte development have been conﬁrmed in separate studies that have used either a global GR-deﬁcient mouse model [31,113], or a mouse in which the GR has been speciﬁcally made deﬁcient in only T cells [115,116]. In the global GR-deﬁcient mouse model, normal negative selection was further conﬁrmed with in vitro models of this process involving challenge by staphylococcal enterotoxin B or CD3/CD28 stimulation, and demonstrated that GR signalling is most likely not required for this process. Taken together, these data from GR-null mice demonstrate that TCR-based thymocyte selection steps are not reliant on GR signalling. Clearly, this remains an area of ongoing controversy that requires further investigation [117–121]. While the precise inﬂuence of glucocorticoid hormones in the thymus remains unclear, the exquisite sensitivity of thymocytes to glucocorticoids, and the constitutive production of glucocorticoid hormones in this tissue, strongly suggests that this signalling pathway can at least potentially play a role in T-cell development. A role is clear at times of stress when glucocorticoid levels increase leading to massive thymocyte apoptosis and it is possible that the steady-state glucocorticoids might normally play a more subtle role in the thymus, which is not always detectable or reproducible across different mouse models of T-cell development. For example, synthetic glucocorticoids have recently been showed to strongly activate the IL7-R-a-chain on T cells,

286 indicating that endogenous glucocorticoids may play a positive regulatory role in T-cell survival via activation of this receptor [122]. Novel glucocorticoid receptor ligands Recent advances in the nuclear mechanisms of GR action, together with the determination of the tertiary structure of the GR LBD [123], have promoted great interest recently in the development of novel tissue selective ligands that include GR antagonists and differential GR agonists, to selectively modulate GR actions to treat speciﬁc diseases (Fig. 6) [124,125]. This has been driven by the ever-present clinical problems associated with chronic treatment of patients with the currently available synthetic glucocorticoids, such as prednisolone and dexamethasone. The myriad of effects glucocorticoids have in many systems in the body makes direct treatment for a speciﬁc disease difﬁcult without eliciting major side effects that include osteoporosis, weight gain, glaucoma and neuropsychiatric symptoms [125]. This has driven interest in the development of better glucocorticoid agonists and antagonists of glucocorticoid action, which have no or reduced side-effect proﬁles. A list of currently used glucocorticoid agonists, antagonists and those in development are shown in Table 2. A number of studies have screened large numbers of non-steroidal compounds and used modiﬁcation chemistries of lead structures to reﬁne novel selective glucocorticoid receptor modulators (SGRMs) [126,127]. The following section summarizes a number of classes of compounds that have been published, and outlines their properties as well as their potential use in treating glucocorticoid-related conditions. Respiratory Distress Syndrome Central Obesity

Reproductive dysfunction agonists/antagonists

Hypertension & Other Cardiovascular Disorders

Osteoporosis

Anti-inflammatory & Immune suppression

Type-2 Diabetes

Brain Disorders - Psychosis - Depression

Fig. 6. The primary diseases and medical conditions that could potentially be treated

by tissue selective glucocorticoid receptor agonists or antagonists.

287 Table 2. Glucocorticoid ligands currently in clinical use or under development as SGRMs. Ligand

GR Ki (nM)

Effect of ligand

Clinical disease for treatment

Dexamethasone

1.2

Agonist

Prednisolone

2.4

Agonist

RU486

0.3

Antagonist

AL-438a

2.5

Modulator

20.3 210 4.8

Modulator Antagonist Antagonist

Arthritis, antiinﬂammatory Arthritis, antiinﬂammatory Psychiatric disorders Speciﬁc antiinﬂammatory Antiinﬂammatory ND Type-2 diabetes

1.6

Modulator

Antiinﬂammatory

ZK216348b AL082D06c Methanesulphonamidesd Benzopyrano-quinolinese

Note: SGRMs, selective glucocorticoid receptor modulators; ND, not determined. a

[132]; [131]; c [127]; d [130]; e [134]. b

Glucocorticoid receptor agonists Since the availability of cortisone in the late 1950s, a number of useful synthetic glucocorticoid agonists have been developed for use in clinical practice. These include prednisolone, dexamethasone and betamethasone, and have been primarily used for their powerful antiinﬂammatory and immunosuppressive effects. All are more potent than cortisol, have varied half-lives, and have differences in antiinﬂammatory activity and side-effect proﬁles [1]. They also have less cross-reactivity with MR and therefore, less salt retention and hypertensive side effects. These agonists can be given orally, parenterally or by various topical routes (skin, inhalation, rectal). Most synthetic steroids have low afﬁnity for circulating carrier proteins (i.e., corticosteroid-binding globulin) and circulate as free steroid (up to 30%) or bind to albumin. Chronic administration of synthetic agonists is common in respiratory and arthritic diseases and it has been estimated that up to 0.5% of the western population are now prescribed chronic glucocorticoids [1]. Some common synthetic glucocorticoid agonists are Prednisolone – Similar in structure to cortisol and has four times more potent antiinﬂammatory effect. Widely used in clinical practise. Dexamethasone – An addition of a 9a-ﬂuoro group, a 1,2 saturated bond and a 16a-methyl group to cortisol produces dexamethasone, which is 25

288 times more potent glucocorticoid than cortisol. It is widely used as an antiinﬂammatory agent and in laboratory research for studying the mechanism of action of glucocorticoids, particularly in immune suppression. Betamethasone – This steroid is similar in structure to dexamethasone but instead has a substituted 16b-methyl group. It is widely used in nasal aerosols for asthma and in the treatment of RDS in very preterm infants. Glucocorticoid receptor antagonists Ligand antagonists of GR are of clinical use in the treatment of hypercortisolemia such as in Cushing’s disease and have recently been tested in the treatment of depression. RU486 – This steroidal antagonist, also called mifepristone, has been available for many years [128] and is currently being used in the treatment of some psychiatric disorders [5]. It does, however, cross-react with progesterone receptors, preventing widespread use. This has, however, led to its use as an abortion pill for the termination of early pregnancy by antagonizing PR in the ovary and uterus. AL082D06 – A non-steroidal compound, AL082D06 was recently developed as a more selective GR antagonist [127]. It is characterized by a tri-aryl methane core, binds to GR with a nanomolar afﬁnity (see Table 2) and does not cross-react with other related steroid receptors such as MR or PR. AL082D06 was shown to antagonize both glucocorticoid-mediated transcriptional activation and repression. AL082D06 failed to induce formation of GR–GRE complexes or promote binding to DNA, and ligand binding initiated much reduced GR nuclear translocation [127]. AL082D06 exhibited very little agonist activity and therefore is an example of a pure GR antagonist. This compound may be of great clinical use for the treatment of a number of cortisol-related endocrine diseases. Sulphonamides – Another class of non-steroidal compounds recently developed as a passive GR antagonist has been sulphonamide derivatives [126,129,130]. A series of non-steroidal passive N-(3-dibenylamino-2-alkyl phenyl)-methanesulphonamide GR modulators have been described. A number of these compounds have been recently analysed for beneﬁcial effects in liver selectivity and in rodent models of type-2 diabetes [130]. These compounds had a nanomolar afﬁnity for human GR and did not bind signiﬁcantly to other members of the steroid receptor family. Two compounds effectively lowered plasma glucose, cholesterol and free fatty acid levels, and also reduced weight gain in the ob/ob mouse model of type-2 diabetes. These compounds did not signiﬁcantly activate the HPA axis in unstressed mice or have abortive effects on pregnant mice, indicating that these passive GR antagonists may have utility for treatment of type-2 diabetes and other aspects of the metabolic syndrome.

289 Selective glucocorticoid receptor modulators Successful treatment of acute and chronic inﬂammatory disease is compounded by many unwanted side effects such as obesity, diabetes, osteoporosis and psychosis. Many of these effects are mediated by the transactivating function of GR on gene targets while most of the antiinﬂammatory effects are repressive in mechanism. This prompted a number of research groups to develop GR ligands with dissociated proﬁles for transactivation and transrepression, that would lead to a much reduced side-effect proﬁle in vivo. A number of recently described GR modulators currently under analysis are listed below. ZK 216348 – This is a non-steroidal selective GR agonist that shows signiﬁcant dissociation between transactivation and transrepression in cellbased assays and in animals in vivo [131]. ZK 216348 has a nanomolar afﬁnity for GR but also bind PR and MR as well. It shows antiinﬂammatory activity comparable to prednisolone in a mouse model of systemic and topical inﬂammation. It has a better side-effect proﬁle of increase in blood glucose, activation of liver gluconeogenesis (as measured by tyrosine aminotransferase activity), splenic involution and to a degree skin atrophy. There is, however, an activation of the HPA axis. This compound is therefore a promising SGRM candidate for treatment of acute inﬂammatory conditions with an improved therapeutic index and reduced side-effect proﬁle. AL-438–AL-438 was derived from a synthetic progestin scaffold compound and ﬁnally selected for its superior properties from a series of related compounds [132]. It binds to GR with a very low nanomolar afﬁnity (Table 2), but also binds MR at relatively high afﬁnity. When tested in vivo AL438 had full antiinﬂammatory potency (similar to prednisolone) but had negligible effects on bone metabolism and did not produce hyperglycaemia in orally treated rats. Interestingly, when given together, AL-438 was able to antagonize the effect of prednisolone in this assay. The mechanism driving these selective effects was shown to be differential recruitment of nuclear cofactors with reduced interaction with PGC-1 (an important co-activator for upregulation of hepatic glucose production) and normal interaction with another nuclear GR co-activator GRIP1. AL-438 may represent a selective non-steroidal ligand for speciﬁc treatment of inﬂammatory diseases. Benzopyrano-quinolines – These compounds are non-steroidal GR ligands and were derived from a tetracyclic (benzopyrano[3,4-f]quinoline) scaffold designed generally for nuclear hormone receptors [133]. C-10 substitutions conferred GR selectivity and C-5 substitutions conferred modulation of transcriptional activity [134]. Many C-5-substituted compounds showed nanomolar afﬁnity for GR with high selectivity over PR. Comparisons of repression of E-selectin vs. activation of a GRE in transfection assays showed that a number of compounds had dissociated activities for transrepression and transactivation [135]. In cell-based assays, these non-steroidal

290 modulators had signiﬁcantly reduced transcriptional activation via GREs of aromatase and tyrosine aminotransferase, but maintained almost 100% efﬁcacy with the transcriptional repression of interleukin-6, collagenase and PGE2. These compounds also show promise as selective antiinﬂammatory drugs but require more testing in vivo to fully assess their dose response and effectiveness for the reduction of GR transactivation-mediated side effects. Conclusion There are potential beneﬁts in the development of the novel glucocorticoids with better efﬁcacy for treatment of inﬂammatory disease and diabetes. Reduction of side effects would greatly enhance their speciﬁcity and allow more long-term systemic use for the treatment of inﬂammatory conditions such as rheumatoid arthritis. Selective modulators have been developed for other nuclear receptors. The best example is the development of the mixed agonists/antagonists for the estrogen receptor [136]. The selective estrogen receptor modulators tamoxifen and raloxifene have different agonist/ antagonist proﬁles depending on the speciﬁc tissue. Such properties have led to the use of raloxifene as a selective estrogenic drug in the clinic for the treatment of osteoporosis. Similarly, selective thyroid hormone receptor modulators are being developed to treat various thyroid disorders [137], with one of these, termed NH-3, showing promise as a selective thyroid hormone antagonist [138]. Approaches to ﬁnd novel selective GR ligands have included the screening of very large chemical libraries and the design of speciﬁcally modiﬁed ligands from the known corticosteroid steroid chemical structure [127,139]. The isolation of the pure GR antagonist AL082D06 came after the screening of many compounds using a GR-based co-transfection transcriptional assay [127]. Further characterization showed that this antagonist inhibited glucocorticoid-mediated transcriptional regulation and competed with other steroids for binding to the LBD of GR, where it failed to induce the correct conformational change to allow recruitment of co-factors for agonist activation. Recent determination of the structure of the GR LBD by Bledsoe and co-workers [123] has provided further insight into the basic mechanisms of GR function, ligand selectivity and co-activator recruitment. The human GR LBD was crystallized in a complex with dexamethasone and a TIF2 coactivator peptide (LxxLL). The three-dimensional structure was similar to other nuclear receptor LBDs, but contained some unique features including a distinct dimerization interface and a unique steroid-binding pocket. Similar to PR [140] and unlike ER, the dimerization interface does not involve contact with the LBD helix 10 but involves, via b-turns of strands 3 and 4, in formation of a central hydrophobic intermolecular b-sheet. This novel dimerization interface is critical for both transactivation and repression functions of the receptor. The GR-LBD structure also reveals a distinct ligand-binding

291 pocket that contains an additional side pocket due to the different positioning of helix 6 and 7 [123]. The high sequence homology of MR in these regions suggests that MR may also contain this side pocket in its LBD. This side pocket in the GR LBD allows the binding of glucocorticoids over other steroids while its absence in ER, PR and AR explains why these receptors do not. The difference in GR and MR selectivity may be due to differences in hydrogen-bond formation perhaps involving residues Q642 of GR that interacts with the 17a-hydroxy group of dexamethasone. The corresponding leucine residue in MR would not promote this interaction and may explain the difference in ligand speciﬁcity. Crystallization of the MR LBD is eagerly awaited. Determination of GR-LBD structures with other ligands, both agonist and antagonists will shed important light on the conformational changes that these ligands induce and the effect on co-activator/co-repressor recruitment. A promising selective GR modulator is the non-steroidal compound AL438 that has full antiinﬂammatory effects but is a partial agonist [132]. Co-activator recruitment studies showed that AL438 was able to induce an interaction with GRIP-1 with an equal efﬁcacy of prednisolone but recruited the key hepatic co-activator PGC-1 with an efﬁcacy that was much reduced relative to prednisolone. This reduced interaction was also demonstrated in GST-pull down interaction assays and indicates that the structural conformational changes induced by AL438 are different to those of prednisolone, and differentiates interactions between speciﬁc co-activators. This also helps to explain the strong antiinﬂammatory activity yet a weak hyperglycaemic effect seen in animal studies. In line with these results, the importance of speciﬁc GR surfaces for the activation of particular sets of target genes has recently been analysed using domain-disrupted GR in ostoesarcoma cells and gene microarrays [141]. Activation of different sets of genes was shown to require different domains or surfaces of the GR further implicating coactivator context-speciﬁc transcriptional regulation. Other avenues for the design of selective ligands need to be considered. These may be extended to interactions with GR surfaces outside the LBD. The co-activator-binding surface that contains a groove for interaction with the co-activator, may be utilized for the design of small molecule inhibitors. Small differences in nuclear receptor co-activator-binding surfaces could be used as targets for selective ligands [142]. For example desethylamiodarone, a metabolite of amiodarone, an antiarrhythmic drug, may regulate TR activity by interacting at a site distinct to the TR LBD [143]. The dimerization interface may also be used as a target for inhibiting receptor function. In fact, studies in a gene-targeted mouse mutant of GR, where dimerization of GR is blocked, indicates that interruption of GR homodimerization may be effective in separating the activation and repression functions of GR in vivo [144]. In summary, further characterization of SGRMs is required, but the development of a number of safer GR ligands, particularly for the treatment of

292 inﬂammatory diseases is more promising. These compounds should have a better side-effect proﬁle and will allow safer therapeutic use for a range of common medical conditions that includes type-2 diabetes and rheumatoid arthritis. Abbreviations ACTH Adrenocorticotropic hormone CRH Corticotropin-releasing hormone DBD DNA-binding domain GR Glucocorticoid receptor LBD Ligand-binding domain MR Mineralocorticoid receptor SGRM Selective glucocorticoid receptor modulator

Acknowledgements This work was supported by the National Health and Medical Research Council of Australia and I thank Mr. Douglas Liddicoat for his critical reading of the manuscript. References 1. 2.

3.

4. 5.

6.

7. 8.

Stewart PM. The adrenal cortex. In: Williams Textbook of Endocrinology, 10th edn, Larsen PRKHM, Melmed S and Polonsky KS (eds), Saunders, 2003, pp. 491–551. Sapolsky RM, Romero LM and Munck AU. How do glucocorticoids inﬂuence stress responses? integrating permissive, suppressive, stimulatory, and preparative actions. Endocr Rev 2000;21:55–89. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schu¨tz G, Umesono K, Blumberg B, Kastner P, Manuel M, Chambon P and Evans RM. The nuclear receptor superfamily: the second decade. Cell 1995;83:835–839. Dallman MF. Adaptation of the hypothalamic–pituitary–adrenal axis to chronic stress. Trends Endocrinol Metab 1993;4:62–69. Flores BH, Kenna H, Keller J, Solvason HB and Schatzberg AF. Clinical and biological effects of mifepristone treatment for psychotic depression. Neuropsychopharmacology 2006;31:628–636. Dallman MF, Akana SF, Levin N, Walker CD, Bradbury MJ, Suemaru S and Scribner KS. Corticosteroids and the control of function in the hypothalamo–pituitary–adrenal (HPA) axis. Ann NY Acad Sci 1994;746:22–32. Funder JW. Target tissue speciﬁcity of mineralocorticoids. Trends Endocrinol Metab 1990:145–148. Guyton, AR. The adrenocortical hormones. In: Textbook of Medical Physiology, 8th edn, Saunders, WB (ed), W.B. Saunders Company, Philadelphia, USA, 1991, pp. 842–854.

293 9.

10. 11. 12.

13. 14.

15.

16. 17. 18.

19. 20. 21.

22.

23. 24. 25.

26.

27.

Kalinyak JE, Grifﬁn CA, Hamilton RW, Bradshaw JG, Perlman AJ and Hoffman AR. Developmental and hormonal regulation of glucocorticoid receptor messenger RNA in the rat. Clin Invest 1989;84:1843–1848. Reichardt HM and Schutz G. Glucocorticoid signalling-multiple variations of a common theme. Mol Cell Endocrinol 1998;146:1–6. Agarwal MK and Mirshahi M. General overview of mineralocorticoid hormone action. Pharmacol Ther 1999;84:273–326. Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, Palensky J and Wittes J. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized aldactone evaluation study investigators. N Engl J Med 1999;341:709–717. Funder JW, Pearce PT, Smith R and Smith AI. Mineralocorticoid action: target tissue speciﬁcity is enzyme not receptor, mediated. Science 1988;242:583–585. Edwards CR, Stewart PM, Burt D, Brett L, McIntyre MA, Sutanto WS, de Kloet ER and Monder C. Localisation of 11 beta-hydroxysteroid dehydrogenase-tissue speciﬁc protector of the mineralocorticoid receptor. Lancet 1988;2:986–989. Lanz RB and Rusconi S. A conserved carboxy-terminal subdomain is important for ligand interpretation and transactivation by nuclear receptors. Endocrinology 1994;135:2183–2195. Lanz RB, Wieland S, Hug M and Rusconi S. A transcriptional repressor obtained by alternative translation of a trinucleotide repeat. Nucl Acids Res 1995;23:138–145. Wilckens T. Glucocorticoids and immune function: physiological relevance and pathogenic potential of hormonal dysfunction. Trends Pharmacol Sci 1995;16:193–197. Trapp T, Rupprecht R, Castren M, Reul JM and Holsboer F. Heterodimerization between mineralocorticoid and glucocorticoid receptor: a new principle of glucocorticoid action in the CNS. Neuron 1994;13:1457–1462. Beato M. Gene regulation by steroid hormones. Cell 1989;56:335–344. Pearce D. A mechanistic basis for distinct mineralocorticoid and glucocorticoid receptor transcriptional speciﬁcities. Steroids 1994;59:153–159. Nissen RM and Yamamoto KR. The glucocorticoid receptor inhibits NFkappaB by interfering with serine-2-phosphorylation of the RNA polymerase II carboxy terminal domain. Genes Dev 2000;14:2314–2329. De Bosscher K, Vanden Berghe W and Haegeman G. The interplay between the glucocorticoid receptor and nuclear factor-kappaB or activator protein-1: molecular mechanisms for gene repression. Endocrinol Rev 2003;24:488–522. Hammes SR. The further redeﬁning of steroid-mediated signaling. Proc Natl Acad Sci USA 2003;100:2168–2170. Falkenstein E, Tillmann HC, Christ M, Feuring M and Wehling M. Multiple actions of steroid hormones – a focus on rapid, nongenomic effects. Pharmacol Rev 2000;52:513–556. Evans SJ, Murray TF and Moore FL. Partial puriﬁcation and biochemical characterization of a membrane glucocorticoid receptor from an amphibian brain. J Steroid Biochem Mol Biol 2000;72:209–221. Zhu Y, Rice CD, Pang Y, Pace M and Thomas P. Cloning, expression, and characterization of a membrane progestin receptor and evidence it is an intermediary in meiotic maturation of ﬁsh oocytes. Proc Natl Acad Sci USA 2003;100:2231–2236. Solito E, Mulla A, Morris JF, Christian HC, Flower RJ and Buckingham JC. Dexamethasone induces rapid serine-phosphorylation and membrane translocation of annexin 1 in a human folliculostellate cell line via a novel nongenomic mechanism involving the glucocorticoid receptor, protein kinase C, phosphatidylinositol 3-kinase, and mitogen-activated protein kinase. Endocrinology 2003;144:1164–1174.

294 28. Bartholome B, Spies CM, Gaber T, Schuchmann S, Berki T, Kunkel D, Bienert M, Radbruch A, Burmester GR, Lauster R, Scheffold A and Buttgereit F. Membrane glucocorticoid receptors (mGCR) are expressed in normal human peripheral blood mononuclear cells and up-regulated after in vitro stimulation and in patients with rheumatoid arthritis. Faseb J 2004;18:70–80. 29. Mendelson CR. Role of transcription factors in fetal lung development and surfactant protein gene expression. Annu Rev Physiol 2000;62:75–915. 30. Cole TJ, Blendy JA, Monaghan AP, Krieglstein K, Schmid W, Aguzzi A, Fantuzzi G, Hummler E, Unsicker K and Schutz G. Targeted disruption of the glucocorticoid receptor gene blocks adrenergeic chromafﬁn cell development and severely retards lung maturation. Genes Dev 1995;9:1–14. 31. Brewer JA, Kanagawa O, Sleckman BP and Muglia LJ. Thymocyte apoptosis induced by T cell activation is mediated by glucocorticoids in vivo. J Immunol 2002;169:1837–1843. 32. Cole TJ, Solomon NM, Van Driel R, Monk JA, Bird D, Richardson SJ, Dilley RJ and Hooper SB. Altered epithelial cell proportions in the fetal lung of glucocorticoid receptor null mice. Am J Respir Cell Mol Biol 2004;30:613–619. 33. Muglia LJ, Bae DS, Brown TT, Vogt SK, Alvarez JG, Sunday ME and Majzoub JA. Proliferation and differentiation defects during lung development in corticotropinreleasing hormone-deﬁcient mice. Am J Respir Cell Mol Biol 1999;20:181–188. 34. Reichardt HM, Kaestner KH, Tuckermann J, Kretz O, Wessely O, Bock R, Gass P, Schmid W, Herrlich P, Angel P and Schutz G. DNA binding of the glucocorticoid receptor is not essential for survival. Cell 1998;93:531–541. 35. Kaplan F, Comber J, Sladek R, Hudson TJ, Muglia LJ, Macrae T, Gagnon S, Asada M, Brewer JA and Sweezey NB. The growth factor midkine is modulated by both glucocorticoid and retinoid in fetal lung development. Am J Respir Cell Mol Biol 2003;28:33–41. 36. Lyons CA and Garite TJ. Corticosteroids and fetal pulmonary maturity. Clin Obstet Gynecol 2002;45:35–41. 37. Newnham JP, Moss TJ, Nitsos I and Sloboda DM. Antenatal corticosteroids: the good, the bad and the unknown. Curr Opin Obstet Gynecol 2002;14:607–612. 38. Hanson M. Birth weight and the fetal origins of adult disease. Pediatr Res 2002;52:473–474. 39. Elenkov IJ and Chrousos GP. Stress hormones, proinﬂammatory and antiinﬂammatory cytokines, and autoimmunity. Ann NY Acad Sci 2002;966:290–303. 40. De Kloet ER. Hormones and the stressed brain. Ann NY Acad Sci 2004;1018:1–15. 41. McEwen BS. Protection and damage from acute and chronic stress: allostasis and allostatic overload and relevance to the pathophysiology of psychiatric disorders. Ann NY Acad Sci 2004;1032:1–7. 42. van Steensel B, van Binnendijk EP, Hornsby CD, van der Voort HT, Krozowski ZS, Rde Kloet E and van Driel R. Partial colocalization of glucocorticoid and mineralocorticoid receptors in discrete compartments in nuclei of rat hippocampus neurons. J Cell Sci 1996;109(Pt 4):787–792. 43. Helm KA, Han JS and Gallagher M. Effects of cholinergic lesions produced by infusions of 192 IgG-saporin on glucocorticoid receptor mRNA expression in hippocampus and medial prefrontal cortex of the rat. Neuroscience 2002;115:765–774. 44. de Kloet RE, Oitzl MS and Joels M. Stress and cognition: are corticosteroids good gus or bad guys? Trends Neurosci 1999;22:422–426.

295 45. Oitzl MS, de Kloet ER, Joels M, Schmid W and Cole TJ. Spatial learning deﬁcits in mice with a targeted glucocorticoid receptor gene disruption. Eur J Neurosci 1997;9:2284–2296. 46. Wei Q, Lu XY, Liu L, Schafer G, Shieh KR, Burke S, Robinson TE, Watson SJ, Seasholtz AF and Akil H. Glucocorticoid receptor overexpression in forebrain: a mouse model of increased emotional lability. Proc Natl Acad Sci USA 2004;101:11851–11856. 47. De Kloet RE, Vreugdenhil E, Oitzl MS and Joels M. Brain corticosteroid receptor balance in health and disease. Endocrine Rev 1998;19:269–301. 48. Lupien SJ and McEwen BS. The acute effects of corticosteroids on cognition: integration of animal and human model studies. Brain Res Rev 1997;24:1–27. 49. de Kloet RE. Stress in the brain. Eur J Pharmacol 2000;405:187–198. 50. Gass P, Reichardt HM, Strekalova T, Henn F and Tronche F. Mice with targeted mutations of glucocorticoid and mineralocorticoid receptors: models for depression and anxiety? Physiol Behav 2001;73:811–825. 51. Sandeep TC and Walker BR. Pathophysiology of modulation of local glucocorticoid levels by 11 b-hydroxysteroid dehydrogenases. Trends Endocrinol Metab 2001;12:446–453. 52. Bohn MC, Howard E, Vielkind U and Krozowski Z. Glial cells express both mineralocorticoid and glucocorticoid receptors. J Steroid Biochem Mol Biol 1991;40:105–111. 53. Yau JLW and Seckl JR. Corticosteroids and the brain. Curr Opin Endocrinol Diabetes 1995;2:239–247. 54. Guyton AC, Hall JE. The adrenocortical hormones. In: Textbook of Medical Physiology, 9th edn, W.B. Saunders Company, Philadelphia, USA, 1996, pp. 957–983. 55. Heim C, Ehlert U and Hellhammer DH. The potential role of hypocortisolism in the pathophysiology of stress-related bodily disorders. Psychoneuroendocrinology 2000;25:1–35. 56. Herman JP, Schafer MK, Young EA, Thompson R, Douglass J, Akil H and Watson SJ. Evidence for hippocampal regulation of neuroendocrine neurons of the hypothalamic–pituitary–adrenocortical axis. J Neurosci 1989;9:3072–3082. 57. Tronche F, Kellendonk C, Kretz O, Gass P, Anlag K, Orban PC, Bock R, Klein R and Schutz G. Disruption of the glucocorticoid receptor gene in the nervous system results in reduced anxiety. Nat Genet 1999;23:99–103. 58. Harris HJ, Kotelevtsev Y, Mullins JJ, Seckl JR and Holmes MC. Intracellular regeneration of glucocorticoids by 11 b-hydroxysteroid dehydrogenase (11 b-HSD)-1 plays a key role in regulation of the hypothalamic–pituitary–adrenal axis: analysis of 11 b-HSD-1-deﬁcient mice. Endocrinology 2001;142:114–120. 59. Stafford JM, Wilkinson JC, Beechem JM and Granner DK. Accessory factors facilitate the binding of glucocorticoid receptor to the phosphoenolpyruvate carboxykinase gene promoter. J Biol Chem 2001;276:39885–39891. 60. Bianchi M, Meng C and Ivashkiv LB. Inhibition of IL-2-induced Jak-STAT signaling by glucocorticoids. Proc Natl Acad Sci USA 2000;97:9573–9578. 61. Barnes PJ and Adcock I. Anti-inﬂammatory actions of steroids: molecular mechanisms. Trends Pharmacol Sci 1993;14:436–441. 62. Homo-Delarche F, Fitzpatrick F, Christeff N, Nunez EA, Bach JF and Dardenne M. Sex steroids, glucocorticoids, stress and autoimmunity. J Steroid Biochem Mol Biol 1991;40:619–637. 63. Thompson EB. Apoptosis and steroid hormones. Mol Endocrinol 1994;8:665–671.

296 64. Slocumb CH, Polley HF, Hench PS and Kendall EC. Effects of cortisone and ACTH on patients with rheumatoid arthritis. Mayo Clin Proc 1950;25:476–478. 65. Fantuzzi G and Ghezzi P. Glucocorticoids as cytokine inhibitors: role in endocrine control and therapy of inﬂammatory diseases. Mediat Inﬂammation 1993;2:263–270. 66. Goulding NJaG PM. Glucocorticoids, lipocortins and the immune response. Curr Opin Immunol 1993;5:108–113. 67. Waage A, Slupphaug G and Shalaby R. Glucocorticoids inhibit the production of IL6 from monocytes, endothelial cells and ﬁbroblasts. Eur J Immunol 1990;20:2439–2443. 68. Arzt E, Sauer J, Pollmacher T, Labeur M, Holsboer F, Reul JM and Stalla GK. Glucocorticoids suppress interleukin-1 receptor antagonist synthesis following induction by endotoxin. Endocrinology 1994;134:672–677. 69. Wiegers GJ, Croiset G, Reul JM, Holsboer F and Rde Kloet E. Differential effects of corticosteroids on rat peripheral blood T-lymphocyte mitogenesis in vivo and in vitro. Am J Physiol 1993;265:E825–E830. 70. Barber AE, Coyle SM, Marano MA, Fischer E, Calvano SE, Fong Y, Moldawer LL and Lowry SF. Glucocorticoid therapy alters hormonal and cytokine responses to endotoxin in man. J Immunol 1993;150:1999–2006. 71. Sapolsky R, Rivier C, Yamamoto G, Plotsky P and Vale W. Interleukin-1 stimulates the secretion of hypothalamic corticotropin-releasing factor. Science 1987;238:522–524. 72. Chrousos GP. The hypothalamic–pituitary–adrenal axis and immune-mediated inﬂammation. N Engl J Med 1995;332:1351–1362. 73. Chen WF, Scollay R, Clark-Lewis I and Shortman K. The size of functional T-lymphocyte pools within thymic medullary and cortical cell subsets. Thymus 1983;5:179–195. 74. Webster JI, Tonelli L and Sternberg EM. Neuroendocrine regulation of immunity. Annu Rev Immunol 2002;20:125–163. 75. Stocklin E, Wissler M, Gouilleux F and Groner B. Functional interactions between Stat5 and the glucocorticoid receptor. Nature 1996;383:726–728. 76. Takeda T, Kurachi H, Yamamoto T, Nishio Y, Nakatsuji Y, Morishige K, Miyake A and Murata Y. Crosstalk between the interleukin-6 (IL-6)-JAK-STAT and the glucocorticoid-nuclear receptor pathway: synergistic activation of IL-6 response element by IL-6 and glucocorticoid. J Endocrinol 1998;159:323–330. 77. Moser M, De T, Smedt T, Sornasse F, Tielemans AA, Chentouﬁ Muraille E, Van Mechelen M, Urbain J and Leo O. Glucocorticoids down-regulate dendritic cell function in vitro and in vivo. Eur J Immunol 1995;25:2818–2824. 78. Caldenhoven E, Liden J, Wissink S, Van de Stolpe A, Raaijmakers J, Koenderman L, Okret S, Gustafsson JA and Van der Saag PT. Negative cross-talk between RelA and the glucocorticoid receptor: a possible mechanism for the antiinﬂammatory action of glucocorticoids. Mol Endocrinol 1995;9:401–412. 79. Herrlich P. Cross-talk between glucocorticoid receptor and AP-1. Oncogene 2001;20:2465–2475. 80. Schule R, Rangarajan P, Kliewer S, Ransone LJ, Bolado J, Yang N, Verma IM and Evans RM. Functional antagonism between oncoprotein c-Jun and the glucocorticoid receptor. Cell 1990;62:1217–1226. 81. Wei P, Inamdar N and Vedeckis WV. Transrepression of c-Jun gene expression by the glucocorticoid receptor requires both AP-1 sites in the c-Jun promoter. Mol Endocrinol 1998;12:1322–1333. 82. Philips M. Maira A, Mullick A, Chamberland M, Lesage S, Hugo P and Drouin J. Antagonism between Nur77 and glucocorticoid receptor for control of transcription. Mol Cell Biol 1997;17:5952–5959.

297 83. Malkoski SP and Dorin RI. Composite glucocorticoid regulation at a functionally deﬁned negative glucocorticoid response element of the human corticotropin-releasing hormone gene. Mol Endocrinol 1999;13:1629–1644. 84. Caelles C, Gonzalez-Sancho JM and Munoz A. Nuclear hormone receptor antagonism with AP-1 by inhibition of the JNK pathway. Genes Dev 1997;11:3351–3364. 85. Baldwin Jr. AS. The NF-kappa B and I kappa B proteins: new discoveries and insights. Annu Rev Immunol 1996;14:649–683. 86. Heck S, Bender K, Kullmann M, Gottlicher M, Herrlich P and Cato AC. I kappaB alpha-independent downregulation of NF-kappaB activity by glucocorticoid receptor. EMBO J 1997;16:4698–4707. 87. McKay LI and Cidlowski JA. Molecular control of immune/inﬂammatory responses: interactions between nuclear factor-kappa B and steroid receptor-signaling pathways. Endocrine Rev 1999;20:435–459. 88. Liden J, Delaunay F, Rafter I, Gustafsson J and Okret S. A new function for the C-terminal zinc ﬁnger of the glucocorticoid receptor. Repression of RelA transactivation. J Biol Chem 1997;272:21467–21472. 89. Barton K, Muthusamy N, Chanyangam M, Fischer C, Clendenin C and Leiden JM. Defective thymocyte proliferation and IL-2 production in transgenic mice expressing a dominant-negative form of CREB. Nature 1996;379:81–85. 90. Perkins ND, Felzien LK, Betts JC, Leung K, Beach DH and Nabel GJ. Regulation of NF-kappaB by cyclin-dependent kinases associated with the p300 coactivator. Science 1997;275:523–527. 91. Avots A, Buttmann Chuvpilo MS, Escher C, Smola U, Bannister AJ, Rapp UR, Kouzarides T and Serﬂing E. CBP/p300 integrates Raf/Rac-signaling pathways in the transcriptional induction of NF-ATc during T cell activation. Immunity 1999;10:515–524. 92. Chakravarti D, LaMorte VJ, Nelson MC, Nakajima T, Schulman IG, Juguilon H, Montminy M and Evans RM. Role of CBP/P300 in nuclear receptor signaling. Nature 1996;383:99–103. 93. Sheppard KA, Phelps KM, Williams AJ, Thanos D, Glass CK, Rosenfeld MG, Gerritsen ME and Collins T. Nuclear integration of glucocorticoid receptor and nuclear factor-kappaB signaling by CREB-binding protein and steroid receptor coactivator-1. J Biol Chem 1998;273:29291–29294. 94. Liden J, Rafter I, Truss M, Gustafsson JA and Okret S. Glucocorticoid effects on NFkappaB binding in the transcription of the ICAM-1 gene. Biochem Biophys Res Commun 2000;273:1008–1014. 95. Nissen RM and Yamamoto KR. The glucocorticoid receptor inhibits NFkappaB by interfering with serine-2 phosphorylation of the RNA polymerase II carboxy-terminal domain. Genes Dev 2000;14:2314–2329. 96. McKay LI and Cidlowski JA. CBP (CREB binding protein) integrates NF-kappaB (nuclear factor-kappaB) and glucocorticoid receptor physical interactions and antagonism. Mol Endocrinol 2000;14:1222–1234. 97. De Bosscher K, Vanden Berghe W, Vermeulen L, Plaisance S, Boone E and Haegeman G. Glucocorticoids repress NF-kappaB-driven genes by disturbing the interaction of p65 with the basal transcription machinery, irrespective of coactivator levels in the cell. Proc Natl Acad Sci USA 2000;97:3919–3924. 98. De Bosscher K, Vanden Berghe W and Haegeman G. Glucocorticoid repression of AP-1 is not mediated by competition for nuclear coactivators. Mol Endocrinol 2001;15:219–227.

298 99. Scollay R and Shortman K. Thymocyte subpopulations: an experimental review, including ﬂow cytometric cross-correlations between the major murine thymocyte markers. Thymus 1983;5:245–295. 100. Iwata M, Hanaoka S and Sato K. Rescue of thymocytes and T cell hybridomas from glucocorticoid-induced apoptosis by stimulation via the T cell receptor/CD3 complex: a possible in vitro model for positive selection of the T cell repertoire. Eur J Immunol 1991;21:643–648. 101. D’Adamio F, Zollo O, Moraca R, Ayroldi E, Bruscoli S, Bartoli A, Cannarile L, Migliorati G and Riccardi C. A new dexamethasone-induced gene of the leucine zipper family protects T lymphocytes from TCR/CD3-activated cell death. Immunity 1997;7:803–812. 102. Vacchio MS, Papadopoulos V and Ashwell JD. Steroid production in the thymus: implications for thymocyte selection. J Exp Med 1994;179:1835–1846. 103. Lechner O, Wiegers GJ, Oliveira-Dos-Santos AJ, Dietrich H, Recheis H, Waterman M, Boyd R and Wick G. Glucocorticoid production in the murine thymus. Eur J Immunol 2000;30:337–346. 104. Pazirandeh A, Jondal M and Okret S. Conditional expression of a glucocorticoid receptor transgene in thymocytes reveals a role for thymic-derived glucocorticoids in thymopoiesis in vivo. Endocrinology 2005;146:2501–2507. 105. King LB, Vacchio MS, Dixon K, Hunziker R, Margulies DH and Ashwell JD. A targeted glucocorticoid receptor antisense transgene increases thymocyte apoptosis and alters thymocyte development. Immunity 1995;3:647–656. 106. Ashwell JD, Lu FW and Vacchio MS. Glucocorticoids in T cell development and function. Annu Rev Immunol 2000;18:309–345. 107. Tolosa E, King LB and Ashwell JD. Thymocyte glucocorticoid resistance alters positive selection and inhibits autoimmunity and lymphoproliferative disease in MRL-lpr/lpr mice. Immunity 1998;8:67–76. 108. Lu FW, Yasutomo K, Goodman GB, McHeyzer-Williams LJ, McHeyzer-Williams MG, Germain RN and Ashwell JD. Thymocyte resistance to glucocorticoids leads to antigen-speciﬁc unresponsiveness due to ‘‘holes’’ in the T cell repertoire. Immunity 2000;12:183–192. 109. Sacedon R, Vicente A, Varas A, Morale MC, Barden N, Marchetti B and Zapata AG. Partial blockade of T-cell differentiation during ontogeny and marked alterations of the thymic microenvironment in transgenic mice with impaired glucocorticoid receptor function. J Neuroimmunol 1999;98:157–167. 110. Pazirandeh A, Xue Y, Prestegaard T, Jondal M and Okret S. Effects of altered glucocorticoid sensitivity in the T cell lineage on thymocyte and T cell homeostasis. Faseb J 2002;16:727–729. 111. Jondal M, Okret S and McConkey D. Killing of immature CD4+ CD8+ thymocytes in vivo by anti-CD3 or 50 -(N-ethyl)-carboxamide adenosine is blocked by glucocorticoid receptor antagonist RU-486. Eur J Immunol 1993;23:1246–1250. 112. Xue Y, Murdjeva M, Okret S, McConkey D, Kiuossis D and Jondal M. Inhibition of I-Ad-, but not Db-restricted peptide-induced thymic apoptosis by glucocorticoid receptor antagonist RU486 in T cell receptor transgenic mice. Eur J Immunol 1996;26:428–434. 113. Purton JF, Boyd RL, Cole TJ and Godfrey DI. Intrathymic T cell development and selection proceeds normally in the absence of glucocorticoid receptor signaling. Immunity 2000;13:179–186.

299 114. Purton JF, Zhan Y, Liddicoat DR, Hardy CL, Lew AM, Cole TJ and Godfrey DI. Glucocorticoid receptor deﬁcient thymic and peripheral T cells develop normally in adult mice. Eur J Immunol 2002;32:3546–3555. 115. Brewer JA, Khor B, Vogt SK, Muglia LM, Fujiwara H, Haegele KE, Sleckman BP and Muglia LJ. T-cell glucocorticoid receptor is required to suppress COX-2-mediated lethal immune activation. Nat Med 2003;9:1318–1322. 116. Baumann S, Dostert A, Novac N, Bauer A, Schmid W, Fas SC, Krueger A, Heinzel T, Kirchhoff S, Schutz G and Krammer PH. Glucocorticoids inhibit activation-induced cell death (AICD) via direct DNA-dependent repression of the CD95 ligand gene by a glucocorticoid receptor dimmer. Blood 2005;106:617–625. 117. Ashwell JD, Vacchio MS and Galon J. Do glucocorticoids participate in thymocyte development? Immunol Today 2000;21:644–646. 118. Godfrey DI, Purton JF, Boyd RL and Cole TJ. Stress-free T-cell development: glucocorticoids are not obligatory. Immunol Today 2000;21:606–611. 119. Godfrey DI, Purton JF, Boyd RL and Cole TJ. Glucocorticoids and the thymus: the view from the middle of the road. Trends Immunol 2001;22:243. 120. Jondal M, Pazirandeh A and Okret S. A role for glucocorticoids in the thymus. Trends Immunol 2001;22:185–186. 121. Jondal M, Pazirandeh A and Okret S. Different roles for glucocorticoids in thymocyte homeostasis. Trends Immunol 2004;25:595–600. 122. Franchimont D, Galon J, Vacchio MS, Fan S, Visconti R, Frucht DM, Geenen V, Chrousos GP, Ashwell JD and O’Shea JJ. Positive effects of glucocorticoids on T cell function by up-regulation of IL-7 receptor alpha. J Immunol 2002;168:2212–2218. 123. Bledsoe RK, Montana VG, Stanley TB, Delves CJ, Apolito CJ, McKee DD, Consler TG, Parks DJ, Stewart EL, Willson TM, Lambert MH, Moore JT, Pearce KH and Xu HE. Crystal structure of the glucocorticoid receptor ligand binding domain reveals a novel mode of receptor dimerization and coactivator recognition. Cell 2002;110:93–105. 124. Miner JN. Designer glucocorticoids. Biochem Pharmacol 2002;64:355–361. 125. Buttgereit F, Burmester GR and Lipworth BJ. Optimised glucocorticoid therapy: the sharpening of an old spear. Lancet 2005;365:801–803. 126. Link JT, Sorensen BK, Patel J, Emery M, Grynfarb M and Goos-Nilsson A. Discovery of novel nonsteroidal glucocorticoid receptor modulators. Bioorg Med Chem Lett 2004;14:2209–2212. 127. Miner JN, Tyree C, Hu J, Berger E, Marschke K, Nakane M, Coghlan MJ, Clemm D, Lane B and Rosen J. A nonsteroidal glucocorticoid receptor antagonist. Mol Endocrinol 2003;17:117–127. 128. Baulieu EE. RU 486 (mifepristone). A short overview of its mechanisms of action and clinical uses at the end of 1996. Ann NY Acad Sci 1997;828:47–58. 129. Link JT, Sorensen BK, Patel J, Arendsen D, Li G, Swanson S, Nguyen B, Emery M, Grynfarb M and Goos-Nilsson A. Optimization and metabolic stabilization of a class of nonsteroidal glucocorticoid modulators. Bioorg Med Chem Lett 2004;14:4169–4172. 130. Link JT, Sorensen B, Patel J, Grynfarb M, Goos-Nilsson A, Wang J, Fung S, Wilcox D, Zinker B, Nguyen P, Hickman B, Schmidt JM, Swanson S, Tian Z, Reisch TJ, Rotert G, Du J, Lane B, von Geldern TW and Jacobson PB. Antidiabetic activity of passive nonsteroidal glucocorticoid receptor modulators. J Med Chem 2005;48:5295–5304.

300 131. Schacke H, Schottelius A, Docke WD, Strehlke P, Jaroch S, Schmees N, Rehwinkel H, Hennekes H and Asadullah K. Dissociation of transactivation from transrepression by a selective glucocorticoid receptor agonist leads to separation of therapeutic effects from side effects. Proc Natl Acad Sci USA 2004;101:227–232. 132. Coghlan MJ, Jacobson PB, Lane B, Nakane M, Lin CW, Elmore SW, Kym PR, Luly JR, Carter GW, Turner R, Tyree CM, Hu J, Elgort M, Rosen J and Miner JN. A novel antiinﬂammatory maintains glucocorticoid efﬁcacy with reduced side effects. Mol Endocrinol 2003;17:860–869. 133. Zhi L, Ringgenberg JD, Edwards JP, Tegley CM, West SJ, Pio B, Motamedi M, Jones TK, Marschke KB, Mais DE and Schrader WT. Development of progesterone receptor antagonists from 1,2-dihydrochromeno[3,4-f]quinoline agonist pharmacophore. Bioorg Med Chem Lett 2003;13:2075–2078. 134. Kym PR, Kort ME, Coghlan MJ, Moore JL, Tang R, Ratajczyk JD, Larson DP, Elmore SW, Pratt JK, Stashko MA, Falls HD, Lin CW, Nakane M, Miller L, Tyree CM, Miner JN, Jacobson PB, Wilcox DM, Nguyen P and Lane BC. Nonsteroidal selective glucocorticoid modulators: the effect of C-10 substitution on receptor selectivity and functional potency of 5-allyl-2,5-dihydro-2,2,4-trimethyl-1H-[1]benzopyrano[3,4-f]quinolines. J Med Chem 2003;46:1016–1030. 135. Elmore SW, Pratt JK, Coghlan MJ, Mao Y, Green BE, Anderson DD, Stashko MA, Lin CW, Falls D, Nakane M, Miller L, Tyree CM, Miner JN and Lane B. Differentiation of in vitro transcriptional repression and activation proﬁles of selective glucocorticoid modulators. Bioorg Med Chem Lett 2004;14:1721–1727. 136. Webb P, Nguyen P and Kushner PJ. Differential SERM effects on corepressor binding dictate ERalpha activity in vivo. J Biol Chem 2003;278:6912–6920. 137. Yoshihara HA and Scanlan TS. Selective thyroid hormone receptor modulators. Curr Top Med Chem 2003;3:1601–1616. 138. Lim W, Nguyen NH, Yang HY, Scanlan TS and Furlow JD. A thyroid hormone antagonist that inhibits thyroid hormone action in vivo. J Biol Chem 2002;277:35664–35670. 139. Shah N and Scanlan TS. Design and evaluation of novel nonsteroidal dissociating glucocorticoid receptor ligands. Bioorg Med Chem Lett 2004;14:5199–5203. 140. Williams SP and Sigler PB. Atomic structure of progesterone complexed with its receptor. Nature 1998;393:392–396. 141. Rogatsky I, Wang JC, Derynck MK, Nonaka DF, hodabakhsh DBK, Haqq CM, Darimont BD, Garabedian MJ and Yamamoto KR. Target-speciﬁc utilization of transcriptional regulatory surfaces by the glucocorticoid receptor. Proc Natl Acad Sci USA 2003;100:13845–13850. 142. Norris JD, Chang C and McDonnell DP. Estrogen receptor–cofactor interactions as targets for novel drug discovery. Ernst Schering Res Found Workshop 2001:181–201. 143. van Beeren HC, Bakker O and Wiersinga WM. Desethylamiodarone interferes with the binding of co-activator GRIP-1 to the beta 1-thyroid hormone receptor. FEBS Lett 2000;481:213–216. 144. Reichardt HM, Tuckermann JP, Gottlicher M, Vujic M, Weih F, Angel P, Herrlich P and Schutz G. Repression of inﬂammatory responses in the absence of DNA binding by the glucocorticoid receptor. EMBO J 2001;20:7168–7173.

301

Recent developments in biodegradable synthetic polymers Pathiraja Gunatillake, Roshan Mayadunne and Raju Adhikari PolyNovo Biomaterials Pty Ltd, Bag 10, Clayton South, Bayview Avenue, Clayton 3169, Australia Abstract. This chapter reviews recent developments in biodegradable synthetic polymers focusing on tailoring polymer structures to meet material speciﬁcation for emerging applications such as tissue engineered products and therapies. Major classes and new families of synthetic polymers are discussed with regard to synthesis, properties and biodegradability, and known degradation modes and products are summarized based on studies reported during the past 10–15 years. Polyesters and their copolymers, polyurethanes, polyphosphazenes, polyanhydrides, polycarbonates, polyesteramides and recently developed injectable polymer systems based on polypropylenefumarates, polyurethanes and acrylate/urethane systems are reviewed. Polyesters such as polyglycolides, polylactides and their copolymers still remain as the major class of synthetic biodegradable polymers with products in clinical use. Although various copolymerization methods have addressed needs of different applications, release of acidic degradation products, processing difﬁculties and limited range of mechanical properties remains as major disadvantages of this family of polymers. Injectable polymers based on urethane and urethane/acrylate have shown great promise in developing delivery systems for tissue engineered products and therapies. Keywords: biodegradable polymers; polyurethanes; polyesters; tissue engineering; biocompatibility; biodegradation; injectable polymers; synthesis; mechanical properties; orthopedics; polyphosphazenes; polyanhydrides; polycarbonates; poly(ortho esters); copolymers

Introduction A major drive for continued research to develop biodegradable synthetic polymers is the need for new materials with properties tailored to meet the biochemical and biomechanical requirements in the emerging technologies such as tissue engineering, regenerative medicine, gene therapy, novel drug delivery systems and implantable devices. Over the past 25 years, signiﬁcant efforts have been devoted to the development of synthetic biomaterials, and a vast majority of these efforts have been focused on identifying ‘‘off the shelf’’ polymers that were biologically inert and stable in biological environments. Polysiloxanes, polyurethanes, polyesters and polyoleﬁns are among the few families of synthetic polymers that are currently used in devices and prostheses implanted to help support the functions of organs and biological tissues. Recently, there has been a major shift in approach to repair/regenerate damaged tissues and organs. Instead of implanting permanent devices, a much better outcome could be expected for the patient if a temporary Corresponding author: Tel: +61-3-9545-2501.

E-mail: [email protected] (P. Gunatillake). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12009-8

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

302 support is provided using a biodegradable polymer until the body can regenerate the damaged tissues and organs. A major challenge to the biomaterials researcher is the design of new materials that meets a number of demanding requirements in emerging technologies. The new material must support the tissue regeneration process while providing mechanical support and eventually degrades to non-toxic products with little or no harm to the body. Additionally, the polymers may be used as a cell delivery system using minimally invasive procedures in the development of advanced tissue engineered products and therapies. Accordingly, there is a need for new materials custom designed to ﬁt varying needs of physicochemical properties necessary for the development of new products and therapies. The main aim of this review is to provide the readers with an update on recent developments in different classes of synthetic biodegradable polymers. While providing a brief introduction to chemistry and properties of major classes of synthetic biodegradable polymers, the review will focus on various copolymerization strategies, and new classes of biodegradable polymers reported in the literature over the last 10–15 years. Polyesters A number of families of synthetic biodegradable polymers are available for selection for a speciﬁc application. Among these poly(a-hydroxy acid esters) are the most widely investigated. Many review articles [1–7] have described the synthesis and properties of poly(glycolic acid) (PGA), poly(lactic acid) and their copolymers, the key members in this family. These polymers continue to be the most widely used synthetic biodegradable polymers in clinical applications. The major applications include resorbable sutures, drug delivery systems and orthopedic ﬁxation devices, such as pins, rods and screws [7]. Despite some disadvantages, these polymers remain attractive because of their ease of degradation by hydrolysis of ester linkages and the degradation products are resorbed through metabolic pathways in some cases. They also have the potential to tailor the structure to alter degradation rates and mechanical properties by using various copolymerization strategies. Poly(a-hydroxy esters) Synthesis and properties: Poly(glycolic acid), poly(lactic acid) and their copolymers are among the most widely studied biodegradable polymers in the polyester family. Several excellent review articles have discussed synthesis, properties and biodegradation of poly(a-hydroxy esters) [1–7], and readers are referred to these reviews for more details. PGA is a rigid thermoplastic material with high crystallinity (46–50%). The glass transition and melting temperatures of PGA are 361C and 2251C, respectively. Because of high crystallinity, PGA is not soluble in most organic

303 solvents; the exceptions are highly ﬂuorinated organic solvents such as hexaﬂuoro isopropanol [7]. Although common processing techniques, such as extrusion, injection and compression molding can be used to fabricate PGA into various forms, its high sensitivity to hydrolytic degradation requires careful control of processing conditions [8,9]. Porous scaffolds and foams can also be fabricated from PGA, but the properties and degradation characteristics are affected by the type of processing technique. Solvent casting, particulate leaching method and compression molding are used to fabricate PGA based implants. The preferred method for preparing high molecular weight PGA is ringopening polymerization of glycolide (Fig. 1), the cyclic dimer of glycolic acid [10,11], and both solution and melt polymerization methods can be used. The common catalysts used include organo tin, antimony or zinc. If stannous octoate is used, temperature of approximately 1751C is required for a period of 2–6 h for polymerization. Although it is possible to synthesize these polymers by acid-catalyzed polycondensation of respective acids, the resulting polymers generally have a low molecular weight, broad molecular weight distribution and consequently poor mechanical properties [12]. The attractiveness of PGA as a biodegradable polymer in medical applications is that its degradation product glycolic acid is a natural metabolite. A major application of PGA is in resorbable sutures (Dexon, American Cyanamide Co.). Numerous studies [13,14] have established a simple degradation mechanism via homogeneous erosion. The degradation process occurs in two stages, the ﬁrst involves the diffusion of water into the amorphous regions of the matrix and simple hydrolytic chain scission of the ester groups. O CH

O n

R

io sat en

O HO

CH R

R =H: Glycolic acid R = CH3: Lactic Acid

C

c oly

C

d on

n

low molecular weight PGA & PLA

p

O azeotropic dehydration polycondensation

OH cy cl i

O

CH R

sa tio n

O C

R

g-o Rin

O

ing pen

po

ion sat eri ly m

n

High molecular weight PGA & PLA

O C

C

R

O

Fig. 1. Synthetic routes to poly(glycolic acid) and poly(lactic acid).

304 The second stage of degradation involves largely the crystalline areas of the polymer, which becomes predominant when the majority of the amorphous regions have been eroded. In a study of Dexon sutures in vitro, the ﬁrst stage degradation predominates during the ﬁrst 21 days and a further 28 days for the degradation of the crystalline regions. After 49 days, the reported weight loss was around 42% with complete loss of mechanical properties. Because of the bulk degradation of PGA, there is a sudden loss of mechanical properties. Although the degradation product glycolic acid is resorbable at high concentrations, they can cause an increase of localized acid concentration resulting in tissue damage. The ultimate fate of glycolic acid in vivo is considered to be the conversion into carbon dioxide and water, with removal from the body via the respiratory system [15]. However, Hollinger [16] has suggested that only lactic acid (LA) follows this pathway, and that glycolic acid is converted into glyoxylate (by glycolate oxidase), which is then transferred into glycine after reacting with glycine transaminase. Poly(lactic acid) is present in three isomeric forms D(), L(+) and racemic (D,L), and the polymers are usually abbreviated to indicate the chirality. Poly(L)LA and poly(D)LA are semicrystalline solids, with similar rates of hydrolytic degradation as PGA. PLA is more hydrophobic than PGA, and is more resistant to hydrolytic attack than PGA. For most applications, the (L) isomer of LA is chosen because it is preferentially metabolized in the body. P(L)LA, poly(lactic–glycolic acid) (PLGA) copolymers and PGA are among the few biodegradable polymers with Food and Drug Administration (FDA) approval for human clinical use. Both polycondensation and ring-opening polymerization methods (Fig. 1) are used to prepare polylactic acid, although the ring-opening polymerization of lactides is the preferred method as it provides high molecular weight polymer with good mechanical properties [1]. Mechanical properties of LA polymers can be varied to a large degree ranging from soft and elastic materials to stiff and high-strength materials by controlling the degree of crystallinity, molecular weight and forming stereo complexation of enantiomeric lactic acids. Semicrystalline PLA has an approximate tensile strength of 50–70 MPa, tensile modulus of 3 GPa, ﬂexural modulus of 5 GPa and an elongation at break of 4% [17–19]. Superior mechanical properties have been achieved by stereo complexation of enantiomeric PLAs, and the improvement is ascribed to the formation of stereocomplex crystallites that acts as intermolecular cross-links [20]. The solubility of LA polymers is highly dependent on the molar mass, degree of crystallinity and other comonomer units present in the polymer. Chlorinated or ﬂuorinated solvents, dioxane, dioxalane and furane are good solvents for enantiomerically pure PLA. Poly(rac-lactide) and poly(mesolactide) are soluble in a range of other organic solvents such as acetone, pyridine, ethylacetate, dimethylsulfoxide, dimethylformamide, etc., in addition

305 to the previously mentioned solvents. So¨derga˚rd and Stolt [1] have recently reviewed the properties of LA polymers, and readers are referred to this article for more details. Copolymers of a-hydroxy acids Synthesis and properties: To address the material needs in drug delivery systems, tissue engineering and regenerative medicine, researchers have focused on developing various copolymer systems of a-hydroxy acids with other monomers. In addition to the copolymers of lactic and glycolic acids, the most frequently used other monomers for copolymerization include e-caprolactone, d-valerolactone, trimethylene carbonate (TMC) and 1,5dioxepan-2-one. The full range of copolymers of LA and glycolic acid has been investigated. The two main series are those of (L)LA/GA and (DL)LA/GA. Gilding and Reed [21] have shown that compositions in the 25–75% range for (L)LA/ GA and 0–70% for the (DL)LA/GA are amorphous. For the (L)LA/GA copolymers, resistance to hydrolysis is more pronounced at either end of the copolymers compositions range [21–24]. The 70/30 GA/LA has the highest water uptake, hence the most readily degradable in the series. In another study, Miller et al. [22] have shown that the 50/50 copolymer was the most unstable with respect to hydrolysis. However, it is generally accepted that intermediate copolymers are very much more unstable than the homopolymers. The ﬁrst commercial use of this copolymer range was the suture material Vicryl (Ethicon Inc.), which is composed of 8% (L)LA and 92% GA. The main application of (D,L-LA/GA) copolymer has been in the ﬁeld of controlled drug release. The degree of crystallinity and melting temperature of PLA polymers can be reduced by random copolymerization with other monomers which disrupts the crystallization ability of poly(lactide) segments. For example, a 50/ 50 copolymer of L-lactic acid and e-caprolactone has shown to be largely amorphous with a common glass transition temperature of 151C [25]. The average LA sequence length has been reported to have a large inﬂuence on both thermal and mechanical properties of the copolymers [26,27]. Likewise, copolymerization with 1,5-dioxepan-2-one [28] and TMC [29] lowers the crystallinity of PLA and consequently lower melting points. Copolymerization with poly(ethylene glycol) (PEG) has been investigated to increase hydrophilicity of PLGA polymers, primarily targeting drug delivery applications. The choice of PEG as a precursor is largely due to its good biocompatibility and hydrophilicity. Chen et al. [30] and Kwon et al. have reported [31] the synthesis of thermo-sensitive triblock copolymers for protein delivery. PLGA–PEG–PLGA triblock copolymers can be made water soluble by using appropriate molecular weight PEG, and Choi et al. [32] have demonstrated that such copolymers can be used for sustained delivery of

306 peptides. Other studies have shown the usefulness of PLGA–PEG copolymers for stent-based controlled delivery of angiostatin [33], as excipients to enhance the gene transfection of various cationic vector systems [34], and delivery of plasmid TGF-beta 1 for diabetic wound healing [35]. Recently, Kumar et al. [36] and Huh et al. [37] have reviewed PLGA–PEG block copolymers for drug delivery applications, and these reviews cover a range of linear and star block copolymers based on PEG with PLA, PDLA and PCL. In an attempt to improve cell attachment to PLGA copolymers, Yoon et al. [38] modiﬁed PLGA by attaching cell-adhesive peptides such as (R: Argenine; G: glycine; D: aspartic acid) RGD peptide and demonstrated signiﬁcant improvement in cell attachment compared to unmodiﬁed polymer. The same group also investigated the effect of immobilization of hyaluronic acid (HA) and poly(L-lysine) to promote the regeneration of cartilage tissue [39,40]. Breitenbach et al. [41] have synthesized poly(vinyl alcohol) (PVA)-based branched polyesters bearing PLGA side chains, which are covalently attached via the hydroxyl groups in PVA. The hydrophilicity of the branched polymer system can be varied by varying the length of PLGA chains and exhibited different release proﬁles compared with unmodiﬁed polymer. An added advantage of these polymer systems is the apparent change in mode of degradation from bulk to surface erosion [42,43]. Further modiﬁcation of these polymers by attaching sulfobutyl groups can create negatively charged polymers [44], while attachment of amino groups such as dimethylamino amine, etc. can generate positively charged copolymers. Ouchi and Ohya [45] have reported the synthesis of random and block copolymers of depsipeptide and L-lactide with reactive (ionic) side groups, comb-type PLA and branched PLA and PLA-grafted polysaccharides and PLA with terminal saccharide residues. These polymers have the potential to be used in drug delivery applications. In summary, various copolymerization approaches have been used successfully to alter the mechanical properties and degradation rates to suit different applications of which drug delivery is the major application area of focus for copolymers of PLGA. Table 1 lists several commercially signiﬁcant copolymers and recently developed copolymers with their potential applications. Biodegradation and biocompatibility of poly(a-hydroxy esters): The degradation of PLA, PGA and PLA/PGA copolymers generally involves random hydrolysis of their ester bonds. PLA degrades to form LA, which is normally present in the body. This acid then enters tricarboxylic acid cycle and is excreted as water and carbon dioxide. No signiﬁcant amounts of accumulation of degradation products of PLA have been reported in any of the vital organs [24]. Carbon13 labeled PLA has demonstrated little radioactivity in feces or urine indicating that most of the degradation products are released through respiration. It is also reported that in addition to hydrolysis PGA is

307 Table 1. Properties of copolymers based on a-hydroxy acids. Copolymer

Major effect of the incorporation of comonomer

Degradation characteristics

Commercial products

Refs

Poly(lactic-co-glycolic acid) (PLGA)

Control degradation rates and mechanical properties

Composition dependent

[3,46,47]

Poly(glycolic acid-cotrimethylene carbonate) (PGA/TMC) and PGA/ TMC/dioxane Poly(glycolic acid-cocaprolactone)P(GA/CL)

To reduce rigidity of PGA and to reduce degradation rate

Half of mechanical strength is lost in 2 weeks

Vicryls (90% GA and 10% LA), Vicryl Rapids (irradiated) Panacryls, Polysorbs Biosyns

Improve processability, degradation rates and tensile strength Increase degradation rate and processability Improve processability, degradation rates and tensile strength Increase hydrophilicity/ degradation rate, drug compatibility Increase hydrophilicity and alter degradation mode. Can be modiﬁed to have cationic and anionic groups to accommodate different drugs

120 days (in suture form) for complete degradation Half of mechanical strength is lost in 2 weeks Half of mechanical strength is lost in 8 weeks

Monocryls (GA/ CL:75/25)

[50]

Monosyns

[3]

Used in veterinary applications

[51]

Dependent on the PEG content

Drug, peptide delivery

[31–37]

Dependent on PVA molecular weight and PLGA chain length

Drug delivery

[41–44]

P(GA/TMC/CL)

P(LA/CL)

PLGA and PEG block copolymers

PVA–PLGA graft copolymers

[48,49]

also broken down by certain enzymes, especially those with esterase activity [52]. Glycolic acid also can be excreted by urine. The rate of degradation, however, is determined by factors, such as conﬁgurational structure, copolymer ratio, crystallinity, molecular weight, morphology, stresses, amount of residual monomer, porosity and site of implantation. Both in vitro and in vivo studies have been carried out to ascertain the biocompatibility of PLA and PGA. Many studies suggest that these polymers are sufﬁciently biocompatible [16,53] although certain studies [54–57] suggest otherwise. Recent studies have shown that porous PLA–PGA scaffolds may be the cause of signiﬁcant systemic or local reactions, or may promote adverse responses during the tissue repair process. PLA–PGA copolymers used

308 in bone repair applications have shown to be biocompatible, non-toxic and non-inﬂammatory [16,53]. Since PLA–PGA have been used successfully in clinical use as sutures, their use in ﬁxation devices or replacement implants in musculoskeletal tissues may be considered safe. Concerns about the biocompatibility of these materials have been raised when PLA and PGA produced toxic solutions probably as a result of acidic degradation products [58]. This is a major concern in orthopedic applications where implants with considerable size would be required, which may result in release of degradation products with high local acid concentrations. Another concern is the release of small particles during degradation, which can trigger an inﬂammatory response. It has been shown that as the material degrades the small particles that break off are phagocytized by macrophages and multinucleated giant cells [59]. It was also noted that no adverse biological responses occur especially if the material volume is relatively small. In clinical studies where PGA was used as fracture ﬁxation, foreign-body responses or osteolytic reactions have been reported [60–63].

Other polyesters Polylactones Synthesis and properties: Poly(caprolactone) (PCL) is the most widely studied in this family [64–65]. PCL is a semicrystalline polymer with a glass transition temperature of about–601C. The polymer has a low melting temperature (59–641C) and is compatible with a range of other polymers. PCL has relatively low tensile strength (23 MPa) but extremely high elongation at break (>700%). PCL degrades at a much lower rate than PLA and is a useful base polymer for developing long-term, implantable drug delivery systems. PCL is prepared by the ring-opening polymerization of the cyclic monomer e-caprolactone. Catalysts such as stannous octoate are used to catalyze the polymerization and low molecular weight alcohols can be used as initiator which can also be used to control the molecular weight of the polymer [66,67]. The other polyesters in the lactone family include poly(3-hydroxybutyrate) (PHB), poly(3-hydroxyvalerate) and poly(valerolactone). Among these PHB is the most investigated for biomedical applications. The polymer can be synthesized by ring-opening polymerization or by microbial methods. PHB is a stiff and brittle polymer with tensile strength of about 40 MPa [68]. Biodegradation and biocompatibility of polylactones: The homopolymer has a degradation time of the order of 2–3 years [64,69,70]. PCL with an initial average molecular weight of 50,000 takes about 3 years for complete degradation in vitro [71]. The rate of hydrolysis can be altered by copolymerization with other lactones, for example, a copolymer of caprolactone and valerolactone degrades more readily [72]. Copolymers of e-caprolactone with DL-lactide have been synthesized to yield materials with more rapid

309 degradation rates (e.g., a commercial suture MONOCRYL, Ethicon) [70]. PCL is considered a non-toxic and a tissue compatible material [69]. Blends with other polymers and block copolymers and low molecular weight polyols and macromers based on caprolactone backbone are a few of the possible strategies to explore this class of polymers for various applications. Poly(ortho esters) Synthesis and properties: Poly(ortho esters) are another class of synthetic biodegradable polymers developed and investigated for applications such as drug delivery in ocular, burns treatment, management of post-operative pain and orthopedic applications. Heller et al. [73,74] have reported on the development of several families of poly(ortho esters) with different mechanical properties and degradation rates. The polymers are relatively easily synthesized by reacting a diol and a diketene and proceeds virtually instantaneously after addition of a trace amount of an acidic catalyst [75]. Polyols based on lactides can be copolymerised to increase degradation rates. Poly(ortho ester)s degrade by surface erosion, and the incorporation of lactide units allows the control of degradation rate. The degradation of the lactide segment releases lactic acid, which catalyze the degradation of the ortho ester groups [75]. Block copolymers of poly(ortho esters) and PEG can form micelles useful in drug delivery applications. Biocompatibility and degradation: Biocompatibility of different poly(ortho esters) has been demonstrated. For example, in ophthalmic drug delivery tests, the polymer was well tolerated and showed no signiﬁcant inﬂammatory reaction [76]. Recently, a GMP toxicology study on one poly(ortho ester) was concluded for products under development for post-surgical pain treatment, osteoarthritis and ophthalmic disease. In orthopedic applications, preliminary in vivo studies have shown that poly(ortho ester) to increase bone growth in comparison to PLGA [77]. Polyanhydrides Synthesis and properties: Polyanhydrides are one of the most extensively studied [78–82] classes of biodegradable polymers with demonstrated biocompatibility and excellent controlled release characteristics. Polyanhydride synthesis, properties and biodegradation have been reviewed recently [78–80]. Langer [83] in 1980 was the ﬁrst to exploit hydrolytic instability of aliphatic polyanhydrides for sustained release of drugs. Owing to the hydrophobic nature, polyanhydrides degrade by surface erosion [81] that makes them very attractive for controlled-release applications. Polyanhydride-based drug delivery systems have been utilized clinically [84]. Diacids or diacid chlorides are the most commonly used monomers for the synthesis of polyanhydrides [78,79]. Melt condensation, ring-opening polymerization, interfacial condensation and dehydrochlorination have been

310 used for polyanhydride synthesis [85,86]. Figure 2 illustrates the dehydration of a mixture of diacids by melt polycondensation [87]. The dicarboxylic acid monomers are converted into the mixed anhydride of acetic acid by reﬂux in excess acetic anhydride. High molecular weight polymers are prepared by melt polycondensation of prepolymer in vacuum under nitrogen sweep. Typical temperatures employed in melt polycondensation are in the range 150–2001C. Catalysts such as cadmium acetate, earth metal oxides and diethyl zinc/water, have been used to catalyze polycondensation reaction. For thermally sensitive monomers, solution polymerization using a range of solvents has been reported [88–90]. Figure 3 shows several common diacid monomers used in the synthesis of polyanhydrides. Langer and coworkers [87,88] have synthesized polyanhydrides from sebasic acid and bis(p-carboxyphenoxy)methane for drug delivery applications. This polyanhydride is used to deliver carmustine, an anticancer drug, to sites in the brain where a tumor has been removed. The degradation products are

n HOOC

(CH2)8

COOH

+

m HOOC

(CH2)14

COOH

catalyst O O

C

O (CH2)8

O

O

C

O

C

(CH2)14

C

n

m

Fig. 2. Polyanhydrides from dehydration of dicarboxylic acids.

O HO

O n

OH

n=4: Adipic acid n=8: Sebasic acid

O HO

n=1: p-carboxyphenoxy acetic acid (CPA) n=4: 5-(p-carboxyphenoxy)valeric acid n=7: 8-(p-carboxyphenoxy)octanoic acid O

O n

O

O

OH

n=1: bis(p-carboxyphenoxy)methane(CPM) n=31,3-bis(p-carboxyphenoxy)propane

OH

HO O Funaric acid

meta=isophtahalic acid para=terephthalic acid

Fig. 3. Dicarboxycylic acid monomers useful in preparing biodegradable polyanhy-

drides.

311 non-toxic and have controlled surface erosion degradation mechanism that allows delivery of drugs at a known rate. Aliphatic polyanhydrides such as those based on sebasic acid or adipic acid are brittle, crystalline and soluble in common organic solvents and generally have fast degradation rates [91,92]. Aromatic polyanhydrides are crystalline with high melting temperatures (>2001C), insoluble in most organic solvents and degrade slowly [93]. Various copolymerization approaches using mixtures of different diacids as well as the incorporation of fatty acids, polyether segments, photo cross-linkable groups and blends of polyanhydrides have been employed to modify degradation rates and mechanical properties [78]. Polyanhydrides have limited mechanical properties that restrict their use in load-bearing applications such as in orthopedics. For example poly[1,6bis(carboxyphenoxy) hexane] has a Young’s modulus of 1.3 MPa [94, 95], which is well below the modulus of human cortical bone (40–60 MPa). To combine good mechanical properties of polyimides with surface-eroding characteristics of polyanhydrides, poly(anhydrides-co-imides) have been developed [96,97], particularly for orthopedic applications. Examples include poly[trimellitylimidoglycine-co-bis(carboxyphenoxy)hexane] and poly[pyromellitylimidoalanine-co-1,6-bis(carbophenoxy)-hexane] [97,98]. These poly(anhydride-co-imides) have signiﬁcantly improved mechanical properties, particularly compressive strengths. Materials with compressive strengths in the 50–60 MPa range has been reported for poly(anhydridesco-imides) based on succinic acid trimellitylimidoglycine and trimellitylimidoalanine [97]. The degradation of these copolymers occurred via hydrolysis of anhydride bonds, followed by the hydrolysis of imide bonds. Photo cross-linkable polyanhydrides have also been developed for use in orthopedic applications, particularly focusing on achieving high mechanical strength [90]. The systems developed are based on dimethacrylated anhydrides. For example, dimethacrylate macromers based on sebacic acid and 1,6-bis(p-carboxyphenoxy)hexane have been reported [90,99]. Both ultraviolet (UV) and visible light curing methods have been investigated with these macromonomers. The most effective means of photopolymerization of these macromonomers was found to be 1.0 wt% camphorquinone and 1.0 wt% ethyl-4-N,N-dimethyl aminobenzoate with 150 mW/cm2 UV power source. Combination of redox type and visible initiation has provided means of achieving efﬁcient curing of thick samples. Depending on the monomers used, the mechanical properties as well as degradation time can be varied. Compressive strengths of 30–40 MPa, and tensile strengths of 15–27 MPa, similar to those of cancelleous bone, have been reported [100]. Biocompatibility and biodegradation of polyanhydrides: Polyanhydrides are biocompatible [101], have well-deﬁned degradation characteristics and have been used clinically in drug delivery systems [94]. Polyanhydrides degrade by hydrolysis of the anhydride linkage and generally undergo a linear mass loss

312 during erosion. The hydrolytic degradation rates can be altered by simple changes in the polymer backbone structure by choosing the appropriate diacid monomers. Poly(sebasic acid) degrades quickly (about 54 days in saline), while poly(1,6-bis(-p-carboxyphenoxy)hexane degrades much more slowly (estimated 1 year). Accordingly, combinations of different amounts of these monomers would result in polymer with degradation properties customdesigned for a speciﬁc application [102]. Minimal inﬂammatory responses to sebacic acid/1,3-bis(p-carboxyphenoxy)propane (SA/CPP) systems have been reported when implanted subcutaneously in rats up to 28 weeks. Loose vascularized tissue had grown into the implant at 28 weeks, with no evidence of ﬁbrous capsule formation [101]. No data have been reported about polymer sterilizability and heat generation during polymerization. A 12-week study using 2–3 mm diameter full thickness defect in the distal femur of rabbits showed good tolerance of the SA/CPP polymer system and osseous tissue in the outer zone of some implants [101].

Polycarbonates Synthesis and properties: Polycarbonates are another class of synthetic polymers explored as biodegradable polymers. Aliphatic polycarbonates such as poly(trimethylene carbonate) degrade under physiological conditions. Most aliphatic polycarbonates become extremely soft in the temperature range 40–601C and weak mechanical properties make them less attractive for use in most applications. One advantage of polycarbonates is their degradation products, most cases the corresponding diols are less acidic than those produced by degradation of polyesters such as poly(lactic acid). Recent studies [103–108] have focused on developing aliphatic polycarbonates with functional groups which allow modiﬁcation of degradation rate and mechanical properties, targeting drug delivery and tissue engineering applications. Polycarbonates are generally prepared by reacting diol compounds with cyclic carbonates or with phosgene (Fig. 4). Several studies have focused on strategies to increase degradation rates by incorporating pendant hydroxyl, carboxyl and amino groups [108–110] to increase hydrophilic character of the polymer. A series of amphiphilic graft polymers of poly(2,2-dimethyltrimethylene carbonate) (PDTC) and poly-a-b-(N-2-hydroxyethyl)-L-aspartamide (PHEA) have been reported [108,109], to increase the hydrophilicity and degradation rate. The hydrophilicity of graft polymer increased with increasing HEA content, and consequently the water absorption level of the polymer. Cholesteryl end-capped polycarbonates have also been reported [110,111] by ring-opening polymerization without catalyst in different molecular weight range. Such polymers are expected to have promising applications in tissue engineering.

313 O HO

CH2 CH2

O

C NH

CH C

CH2

C O

O

OH

O

O R phosgene

CH2Cl2 O CH2 CH2

C NH

O CH C

CH2

O

C O

O

O R

n

Fig. 4. Synthetic route to prepare polycarbonates.

Biocompatibility and biodegradation: Tyrosine-based polycarbonates have been reported as other promising degradable polymers for use in orthopedic applications [99,112–114]. These polymers posses three potentially hydrolyzable bonds: amide, carbonate and ester. Studies have shown [99] that the carbonate group hydrolyzes at a faster rate than the ester group, and the amide bond is not labile in vitro. Since the hydrolysis of the carbonate groups yields two alcohols and carbon dioxide, the problem of acid bursting seen in polyesters is alleviated. By variation of the structure of the pendant R group, polymers with different mechanical properties, degradation rates as well as cellular response could be prepared. Polycarbonate having an ethyl ester pendant group has shown to be strongly osteoconductive and good bone apposition, and possesses sufﬁcient mechanical properties for load bearing bone ﬁxations. In vivo studies have demonstrated that the polymer was biocompatible and promoted signiﬁcant bone growth [99,114]. Polyesteramides Synthesis and properties: Polyesteramides have both amides as well as ester linkages which attributes amphiphilicity and biodegradability [115–118]. They are designed to couple the excellent mechanical properties of polyamides and the biodegradability of polyester [119]. Owing to the polar nature of amide groups and their ability to form hydrogen bonds, these polymers exhibit good thermal and mechanical properties even at low molecular weight. On the other hand, hydrolytically degradable ester bond provides

314 biodegradability to the polymer. Another advantage of this class of polymers is the ability to incorporate a-amino acids, which provides sites for enzyme-induced biodegradation [117]. Polycondensation of monomers with carboxyl, alcohol and amino functional groups are used to prepare polyesteramides. The availability of a range of suitable monomers, such as hydroxy acids, dicarboxylic acids, amino acids, diamines, amino alcohols and diols allows the preparation of a range of copolymers with different mechanical properties and biodegradability. Examples of biodegradable polyesteramides reported in the literature include polyesteramides based on e-caprolactone and 11-aminoundecanoic acid [120], e-caprolactone, 11-aminoundecanoic acid and ethylene glycol [121], adipic acid, caprolactam and 1-4 butanediol, and their branched polymers with glycerol as branching agent [120–127]. In addition, many amino acid and amino alcohol-based polyesteramides have also been reported in the literature [128–133]. Kise et al. [134] have reported synthesis and biodegradability of a series of novel polyesteramides based on alpha-amino acid, 2aminoethanol and dicarboxylic acid chlorides. Synthesis of biodegradable polyesteramide microspheres based on e-caprolactone and 11-amino undecanoic acid and their degradation properties have been reported by Qian et al. [135]. Degradable polyesteramides ﬁbers based on 11-aminodecanoic acid have also been prepared by melt spinning. These ﬁbers exhibited very high tensile strength (140 MPa) [136]. A biologically safe anti-corrosive coating based on a polyesteramide from Pongamia glabara oil has also been reported [137]. The solubility of these polymers is dependent on amino acid residues and they are generally soluble in highly polar solvents like formic acid and triﬂuroethanol and insoluble in non-polar solvents such as ethylacetate. Polyesteramide generally containing alpha-amino acid are soluble in chloroform [128,129], whereas those containing Gly residues are soluble in aprotic polar solvents such as DMF and DMSO. Biodegradability: The ester group is the more readily hydrolyzable linkage in these polymers and amide linkages are not easily hydrolyzable due to strong hydrogen bonding and the associated crystallinity. Accordingly, polyesteramides with high level of amide linkages are extremely slow to degrade. The biodegradability of amino acid-based polyesteramides is depended on the speciﬁcity of enzymes to amino acid derivatives. The degradability of these polymer was reported low when L-Ala and L-Val was introduced into the polymers [134]. Kim et al. [138] have reported the relationship between hydrophilicity and biodegradability of polyesteramides. Branched polyesteramides degraded slowly in PBS and hydrolysis was reported primarily on ester bonds. Branching of polyesteramide have been reported to substantially enhance hydrolysis both in alkaline and in PBS solution [139].

315 Polyphosphazenes Synthesis and properties: Polyphosphazenes are a relatively new class of biodegradable polymers, distinct from other classes of biodegradable polymers due to their synthetic ﬂexibility and versatile adaptability for applications. Polyphosphazenes are high molecular weight, essentially linear polymers with an inorganic backbone consisting of alternating phosphorous and nitrogen atoms bearing two side groups attached to each phosphorous atom. R N

P R

n R = alkoxy, aryloxy or amino groups (I)

Although the early attempts to synthesize polyphosphazene dates back to 1895 by Stokes [140], the ﬁrst successful synthesis of poly(dichlorophosphazene) was reported by Allcock and Kugel [141] in 1965. Different polyphosphazenes are synthesized by means of macromolecular substitution reactions carried out on a reactive polymeric intermediate, poly(dichlorophosphazene), (NPCl2)n. This intermediate is prepared by thermal ring-opening polymerization of hexachlorotriphosphazene. The poly(dichlorophosphazene) is hydrolytically unstable due to the high reactivity of P–Cl bonds. Allcock’s group utilized the high reactivity of the P–Cl bond to synthesize a range of hydrolytically stable poly(organophosphazenes) by replacing chlorine atoms with alkoxide or aryloxide [142], primary or secondary amines [143] or organometallic reagents [144]. The polyphosphazenes consist of over 700 different polymers with the general structure (I) [145]. The substituents on phosphorous inﬂuence the properties of these polymers and their typical properties include biocompatibility, ﬂexibility, high dipole moment, broad range of glass transition temperature, chemical inertness, elastomeric properties, ﬂame-retardancy, mechanical strength and solvent permeability. It has been reported recently that by substituting with appropriate side groups, polyphosphazenes could be rendered biodegradable. Examples of such side groups include amines, amino acid esters, glucosyl, glyceryl, lactate or imidazolyl units [146–148]. Among these, aminoacid ester substituted polymers have been the most widely investigated as biodegradable polymers. Phosphazenes with amino acid esters (II) [149–151] and imidazole (III) have shown excellent hydrolytic degradability. The hydrolytic stability can be

316 modulated by cosubstituting with less hydrolytically sensitive groups on the polymer backbone [144]. N N

HNCHRCOOR' N

N P N

P HNCHRCOOR' n (II)

N

n

(III)

Alkoxy-substituted polyphosphazenes are also shown to be hydrolytically unstable. Examples include, glyceryl-substituted polyphosphazene (III) [152], glucosyl and methyl amino-cosubstituted polyphosphazene [153] as well as esters of glycolic or LA-substituted polyphosphazenes [154]. OH OH OCH2CHCH2 N

P OCH2CHCH2 OH OH

n

(VI)

Glycerol-substituted polyphosphazenes (VI) can be cross-linked by reacting with cross-linking agents such as adipoyl chloride and hexamethylenediisocyanate. The glycolic or lactic acid-substituted polymers hydrolyze much faster than the homopolymers PLA and PGA. Although many studies have focused on investigating the effect of various substituents on hydrolytic stability, thermal and other properties, there is hardly any information on mechanical properties of biodegradable polyphosphazenes reported in the literature. The most widely investigated area of application is controlled drug release, and readers are referred to a recent review article for details [155]. Numerous degradation studies carried out have shown that hydrolysis of polyphosphazenes leads to compound(s) derived from the pendant groups, and phosphate and ammonia due to backbone degradation. The rate of degradation as well as the physico-chemical properties of these polymers can be tuned by incorporation of appropriate ratios of different pendant group. By careful controlling the nature and composition of side group substituents, it is possible to control the rate of degradation of polyphosphazenes over periods of hours, days, months or years. The inﬂuence of different amine

317 groups on the hydrolytic degradation of polyphosphazenes was extensively studied by Allcock et al. [156]. Majority of the aminated polymers undergo faster degradation in acidic pH compared to physiological or basic pH. Biocompatibility and biodegradation: Laurencin et al. [157] have investigated methylphenoxy and either imidazolyl or ethylglycinate-substituted polyphosphazenes for skeletal tissue regeneration. Both materials supported the growth of MC3T3-E1, an osteogenic cell line. Increase in imidazolyl side groups resulted in a reduction in cell attachment and growth on the polymer surface and an increase in the rate of degradation of the polymer. In contrast, substitution with ethylglycinato group favored increased cell adhesion and growth accompanied with an increase in the rate of degradation. In another study, Laurencin et al. [158] reported that porous matrices of poly[(50% ethylglycinato) (50% p-methylphenoxy) phosphazene] with pore sizes of 150–250 mm are good substrates for osteoblast-like cell attachment and growth. Polyurethanes Synthesis and properties: Thermoplastic polyurethanes (TPUs) represent a major class of synthetic polymers that have been evaluated for a variety of biomedical applications. Conventional TPUs are prepared from three monomers, a diisocyanate, a diol or diamine chain extender and a long-chain diol. These monomers react to form linear, segmented copolymers consisting of alternating hard and soft segments, which are characteristic structural features of conventional TPUs. The hard segment (HS) is composed of the alternating diisocyanate and chain extender molecule, whereas the soft segment (SS) is formed from the long-chain linear diol. Scheme 1 illustrates the general structure of TPU. Owing to thermodynamic incompatibility of hard and soft segments, TPUs exhibit two-phase morphology. The HSs aggregate to form microdomains resulting in structure consisting of glassy or semicrystalline domains and rubbery SSs aggregate to form soft domains, which are mostly amorphous. TPU can also be prepared by reacting equi-molar amounts of a linear diol and a diisocyanate, and this strategy has been used in designing biodegradable polyurethanes. By choosing two SS polyols with glass transitions above and below ambient temperature, TPU elastomers with good mechanical strength can also be prepared using this approach. The interest in polyurethanes for biomedical applications is due to their excellent mechanical properties, good biocompatibility and structural versatility achievable to tailor polymer structure to meet the needs of a wide variety of medical implant applications. Polyurethanes are currently used in applications such as cardiac pace makers and vascular grafts. Most of the research on biomedical polyurethanes was focused on improving biocompatibility and stability as well as designing low modulus elastomers for applications such as synthetic heart valves. These research efforts have resulted

318

OCN

CH2

n

NCO

HOCH2CH2CH2CH2OH

O HO

CH2)5 C

O O

(CH2)p n

O

C

(CH2)5

OH m

Poly(caprolactone) diol

HS

Urethane linkage

SS

H

O

N

C

O

Scheme I. Example of an aliphatic TPU.

in the development of siloxane-based polyurethanes (Elast-EonTM), which have greater in vivo stability than conventional polyetherurethanes such as those based on poly(tetramethylene oxide) (PTMO) [159]. The variety of chemical functionality that can be built into the polymer chain allows the design of polyurethanes that are degradable in the biological environment. In designing biodegradable polyurethanes, the chemical structure of the diisocyanate and polyol play a pivotal role. By the appropriate choice of these compounds and relative proportions, polyurethanes with a range of mechanical properties as well as biodegradation characteristics to suit different applications can be formulated. Figure 5 shows a few examples of diisocyanates useful in formulating biodegradable polyurethanes. Although diisocyanates are toxic compounds, once incorporated into a polyurethane, the resulting polyurethane after hydrolytic degradation does not release the diisocyanate, but the corresponding diamine. Accordingly, the choice of the diisocyanate is largely governed by the toxicity of the corresponding diamine. Diisocyanates such as 4,40 -methylene diphenyl diisocyanate (MDI) and toluene diisocyanate (TDI), commonly used in many industrial polyurethane formulations, are not used in formulating biodegradable polyurethanes because of the high toxicity of the corresponding diamines. Aliphatic diisocyanates such as ethyl lysine diisocyanate (ELDI), methyl lysine diisocyanate (MLDI), hexamethylene diisocyanate (HDI) and 1,4-butanediisocyanate (BDI) are among the most suited in formulating biodegradable polyurethanes. Polyester polyols are the preferred choice for obvious reasons. The ester linkages are susceptible to hydrolytic degradation and certain enzymes also can render ester linkages to degrade. The SS diols used in biodegradable

319 OCN

(CH2)4 CH

NCO

COOR Ethyl 2,6-diisocyanatohexanoate (R=Ethyl) and Methyl 2,6-diisocyanatohexanoate (R=Methyl) OCN

(CH2)4

NCO

1,4-Diisocyanatobutane OCN

(CH2)6 NCO

1,6-Diisocyanatohexane CH3 OCN

CH2CCH2CHCH2CH2 CH3

NCO

CH3

2,2,4-Trimethylhexamethylenediisocyanate

Fig. 5. Diisocyanates used in biodegradable polyurethane formulations.

polyurethanes are polyesters, and the exceptions are Pluronics, as these polyether polyols are used as copolyols to increase hydrophilicity. In this review, properties of biodegradable polyurethanes are provided based on the type of diisocyanate used on the polyurethane formulation. 1,6-Hexamethylenediisocyanate-based polyurethanes: HDI is the most widely investigated diisocyanate in formulating biodegradable polyurethanes. The choice of this diisocyanate is largely due to the relative non-toxic nature [160] of the corresponding diamine 1,6-hexanediamine. The symmetrical molecular structure also leads to strong intermolecular attractions through hydrogen bonding resulting in elastomers with high strength. Elastomers with ultimate tensile strength over 60 MPa and elongation of 580% have been reported for HDI-based polyurethanes [161,162]. HDI-based TPUs display a tendency to cold draw and ﬁbers with high strength can be drawn from HDIbased polyurethanes. Structure property relationship of HDI-based polyurethanes with a range of linear polyester diols has been reported, and among these polycaprolactone is the most widely investigated SS diol. Gorona et al. [163] synthesized a series of polymers based on poly(e-caprolactone)diol (e-PCL) with molecular weight in the range 1,080–5,800. HDI was used as a chain extender in preparing these polymers, but elastomers with good mechanical strength (UTS 30 MPa) and elongation at break (980%) were obtained by this approach. Many researchers have investigated the effect of incorporating mixtures of SS polyols, in most cases two polyols, to achieve different mechanical properties and to improve biocompatibility as well as to alter degradation behavior of polyurethanes. Saad et al. [164] have prepared a series of

320 copolymers containing PCL and poly(R-3-hydroxybutyrate)diols which are linked by HDI. These copolymers differ from the conventional segmented polyurethanes in that the diisocyanate is used as a linker without a chain extender. In these copolymers, PHB due to its crystalline nature plays a role of the HS, and copolymers containing 20 wt% PHB have shown tensile strength up to 27 MPa and elongation at break 890%. Increasing the PHB content decreased the strength of the copolymer. A similar study reported by Cohn et al. [165] investigated triblock copolymers prepared by chain extending linear triblock copolymer diol e-PCL-PEO-e-PCL with HDI. The water uptake as well as the rate of degradation increased with the increase in PEO content in the polyurethane and its molecular weight. Gorona et al. [163,166,167] have investigated the effect of incorporating hydrophilic polyols poly(ethylene oxide) (PEO), poly(ethylene–propylene–ethylene)oxide (PEO-PPO-PEO) diol and hydrophobic e-PCL on properties of biodegradable polyurethanes. The HSs in these polyurethanes were 1,6-hexamethylenediisocyanate chain extended with either 1,4-butanediol or 2-amino-1-butanol. Increasing the PEO content resulted in higher water absorption; a polyurethane based on a 50/50 mixture of PEO (MW 2,000) and e-PCL (MW 530) absorbed 212% water, compared with 2% water absorption for PCL-based polymer. The amount of PEO, and its molecular weight signiﬁcantly affected the mechanical properties, water absorption, calciﬁcation and hydrolytic degradation. Linear diols of triblock copolymers based on LA and ethylene glycol, chain extended with HDI produce poly(ether-ester) urethanes with different degradation rates [168]. Polyurethane based on copolymer diol with a higher PLA content was more hydrophobic, degraded slowly in in vitro, compared to the copolymer with high PEG content. Kylma¨ et al. [169] have employed a melt processing method to prepare polyester urethane blends to investigate the effect of blending poly(lactic acid-co-e-caprolactone-urethane) [P(LAco-CL)] on properties and morphology of lactic acid-based amorphous poly(ester-urethane)s. The copolymer polyols with different ratios of LA and CL were used as the rubber to modify the properties of more rigid PLAbased polyurethane. The polyester urethanes were prepared by chain extending the corresponding linear diols with HDI. The incorporation of the rubbery polyurethane, resulted in toughening of the more brittle PLA-based polyurethanes, and polyurethane based on a linear copolymer diol of LA and CL (70/30) with 20% loading in the blend produced a polymer with elongation approaching 100%. However, the strength of the materials was signiﬁcantly compromised. Gorona et al. [170] have investigated the effect of PCL molecular weight (530–2,000 range), catalyst and chain extender on properties of polyurethanes based on HDI and isophorone diisocyanates. The highest strength (63 MPa tensile strength) was observed for polyurethane based on PCL with a molecular weight of 530. The type of catalyst, chain extender and PCL

321 molecular weight had a signiﬁcant effect on mechanical properties of the polyurethane. 1,4-Butanediisocyanate-based polyurethanes: Similar to HDI, BDI due to its symmetrical molecular structure produce polyurethanes with good mechanical properties. Furthermore, the degradation product 1,4-butanediamine (putrescine) is a naturally occurring non-toxic compound. Guan et al. [171] synthesized a family of polyester urethaneureas from poly(caprolactone) diol, BDI, lysine ethyl ester or putrescine as chain extender. Flexible elastomers with elongation at break of 660–895% and tensile strength from 9 to 29 were produced and lysine chain extended polyurethaneureas were generally weaker materials compared with those based on putrescine. Spaans et al. [172] synthesized high tensile strength (35 MPa) polyurethanes from BDI and e-PCL. De Groot et al. [173] prepared polyurethanes based on BDI and copoly(L-lactide/e-caprolactone) using prepolymer method. Chain extension of isocyanate-terminated prepolymer with butanediamine was not possible due to the susceptibility of lactide bonds to aminolysis. Chain extension with 1,4-butanediol produced polyurethanes with poor mechanical properties, presumably due to trans-esteriﬁcation. This problem was avoided by chain extending the copolymer diol with an isocyanate-terminated block, and polyurethane with tensile strength of 45 MPa was obtained. Lysine diisocyanate-based polyurethanes: Lysine diisocyanate (LDI) is another diisocyanate that has received recent attention from researchers for developing biodegradable polyurethanes. Polyurethanes based on LDI when degraded release lysine, a non-toxic amino acid presents in proteins such as collagen, as one of the main degradation products. Lysine diisocyanate is not commercially available (being developed by Kyowa Hakko Kogyo Co., Ltd.) but can be prepared from L-lysine monohydrochloride [174,175]. Both ELDI and MLDI can be prepared. Storey et al. [176] have prepared poly(ester urethane) networks from LDI and a series of polyester triols based on DL-lactide, g-caprolactone and their copolymers. Networks based on poly(DL-lactide) were rigid (T g ¼ 601C) with ultimate tensile strengths of 40–70 MPa, whereas those based on caprolactone triols were low modulus elastomers with tensile strengths of 1–4 MPa. Networks based on copolymers were more elastomeric (elongation up to 600%) with compressive strengths between 3 and 25 MPa. Hydrolytic degradation under simulated physiological conditions were dependent on the type of triol and DL-lactide-based networks were the most resistant with no degradation observed for 60 days, caprolactone-based triol networks were resistant up to 40 days, whereas the high lactide-based copolymer networks were the least resistant and substantial degradation observed in about 3 days. Bruin et al. [174] have reported on the synthesis of degradable polyurethane networks based on star-shaped polyester prepolymers. The star prepolymers were prepared from myoinisitol, a pentahydroxy sugar molecule by ring-opening copolymerization of L-lactide or glycolide with caprolactone.

322 The prepolymers were cross-linked using 2,6-diisocyanatohexanoate. The degradation products of these PU networks are considered non-toxic. The resulting network polymers were elastomeric with elongation in the range 300–500% and tensile strengths varying between 8 and 40 MPa depending on the branch length etc. Preliminary experiments in guinea pigs have shown that the polyurethanes biodegrade when implanted subcutaneously. Zang et al. [177] have developed a peptide-based polyurethane scaffold for tissue engineering. LDI was reacted ﬁrst with glycerol to form a prepolymer, which upon reaction with water produced a cross-linked porous sponge due to liberation of carbon dioxide. Initial cell growth studies with rabbit bone marrow stromal cells (MSC) have shown that the polymer matrix supported cell growth and was phenotypically similar to those grown on tissue culture polystyrene. Hirt et al. [178] and De Groot et al. [179] reported on the synthesis and properties of degradable polyurethanes based on LDI, 2,2,4-triethylhexamethylene diisocyanate and a number of polyester and copolyester polyols such as Diorezs, caprolactone, ethylene glycol copolymers, and polyhydroxy butyrate and valerate copolymers. The polyurethanes ranged from elastomers with elongations at break as high as 780%, but with low tensile strengths (5.8–8.1 MPa). Saad et al. [180] reported on the cell and tissue interaction of four such polymers prepared from 2,2,4-trimethylhexamethylene diisocyanate and 2,6-diisocyanato methyl caproate, and polyols, a,odihydroxy-poly(R-3-hydroxybutyrate-co-(R)-3-hydroxyvalerate)-block-ethylene glycol and two commercial diols, Diorezs and PCL-diol. In vitro studies indicated that these polyesterurethanes did not activate macrophages and showed good level of cell adhesion and growth, which were also conﬁrmed by in vivo results. The incorporation of a chain extender with a hydrolytically labile linkage has been investigated to increase the degradation rates of polyurethanes as well as to modify the properties. Skarja and Woodhouse [181,182] have investigated the effect of an amino acid (phenylalanine)-based chain extender on polyurethanes prepared from methyl lysine diisocyanate (MLDI), polycaprolactone and PEO. Their results showed that PEO-based polyurethanes were generally weaker but PCL-based materials were relatively strong. However, no results were reported on the degradation of these polyurethanes. Gunatillake et al. [183,184] have prepared a series of diol chain extenders based on a-hydroxy acids and ethylene glycol. Polyurethanes based on these chain extenders degraded faster than those based on conventional chain extenders such as 1,4-butanediol and ethylene glycol. Dahiyal et al. [185] have incorporated chain extender with phosphate ester linkages. Examples include bis(2-hydroxyethyl)phosphate and bis(2-hydroxyhexyl)phosphate. One advantage claimed for this polyurethane is the ability to covalently attach drugs as pendent to the polymer chain via the phosphorous.

323 Gogolewsky and Pennings [186,187] have reported on a design of an artiﬁcial skin composed of polylactide/polyurethane mixtures where the PU was non-degradable. In vivo studies with guinea pigs showed that the artiﬁcial skin adhered to wound well, and protected from ﬂuid loss and infections up to 40 days exhibiting potential as a skin substitute. Micro-porous polyurethane amide and polyurethane–urea scaffolds have been evaluated by Spaans et al. [188] for repair and replacement of knee-joint meniscus. The SSs in these polyurethanes were based on 50/50 l-lactide/PCL and chain extenders were adipic acid and water, the reaction of latter with BDI provided carbon dioxide to produce porous scaffolds. Salt crystals were also added to produce porous structure, and the addition of surfactants combined with ultrasonic waves regulated the pore structure. Porous scaffolds with porosity of 70 to 80% were achieved by this technique. These scaffolds exhibited tearing problems during suturing [189], which was partly circumvented by using a different suturing system. A meniscal replica implanted contained only ﬁbro-cartilage after 18 weeks and decreased the degradation of the articular cartilage. Gunatillake et al. [190–192] have developed polyurethane prepolymers that can be cross-linked to form both rigid and elastomeric compositions useful in a range of biomedical applications including scaffolds in tissue engineering. The differential reactivity of the isocyanate functional groups in diisocyanates, such as LDI, is used to prepare prepolymers that are liquid at and above ambient temperatures by reacting with multi-hydroxy functional core molecules, such as pentaerythritol. Under controlled reaction conditions, star/hyperbranched polyols with isocyanate end-functional groups can be prepared. Reaction of prepolymer with an appropriate hydroxyl compounds, biodegradable polyurethane networks can be prepared. With the appropriate choice of precursors, polyurethane with high mechanical properties can be prepared. Polyurethanes based on other diisocyanates: TDI, 2,2,4-trimethylhexamethylene diisocyanate, 1,10 -methylene-bis(4-isocyanatocyclohexane) (HMDI) and isoporone diisocyanate are among the other diisocyanates investigated, although to a lesser extent, in formulating biodegradable polyurethanes. Storey et al. [193] have prepared a series of polyurethane networks by cross-linking D,L-lactide, glycolide, e-caprolactone and trimethylene carbonate copolyester triols with TDI. Amorphous networks with tensile strength approaching 50 MPa were obtained with e-PCL as SS polyol. Glycolide containing PU networks showed the fastest degradation rate. Bogdanov et al. [194] investigated the effect of e-PCL molecular weight (2,000, 4,000 and 7,300) on properties of HMDI-based polyurethanes, and showed that straininduced crystallization occurred when stretched above 100%. In another study, Lee et al. [195] prepared a series of polyurethanes based on HMDI and a mixture of poly(butylenes succinate) (PBS) and poly(ethylene glycol) SS polyols. Elastomers with elongation at break in the range 160–230% and tensile strength 2.0–2.4 kg/mm2 were observed and DSC results showed that

324 PBS segments formed crystalline domains. Incorporation of PEG increased the degradation rate. Biocompatibility and biodegradation: Both in vitro and in vivo studies have indicated that the biocompatibility of biodegradable polyurethanes is generally very favorable. Animal studies have demonstrated rapid cell in-growth with no adverse tissue reactions. However, the mechanism(s) of degradation, effect of degradation product, their toxicity and how those products are removed from the body are not clearly understood. Tissue in-growth and degradation of two biodegradable polyurethanes based on e-PCL/L-LA (MW 2,000), BDI and chain extender 1,4-butanediol (BDO) were investigated by subcutaneous implantation in rats [196]. The polymers were fabricated as foams with different porosities using the salt leaching method. In the foam with high porosity, complete tissue in-growth was observed, before polymer degradation, whereas the foam with lower macroporosity, the polymer degraded before complete tissue in-growth. Authors have indicated that fewer interconnected pores in the latter as the primary reason for low tissue in-growth. In another study [197], similar polyurethanes foams were implanted in the avascular region of canine lateral menisci. The study demonstrated that the polymer implants did not inhibit the healing process, and the scaffold became intensively integrated with the host meniscal tissue. Polyurethane networks based on LDI and poly(glycolide-co-g-caprolactone) macrodiol was evaluated by Bruin et al. [198] as two-layer artiﬁcial skin. The degradation of the skin in vivo was faster than that in in vitro. Subcutaneous implantation in guinea pigs showed that the porous polyurethane networks allowed rapid cell in-growth, degraded almost completely 4–8 weeks after implantation and evoked no adverse tissue reactions. Grad et al. [163,199] have studied the chondrocyte attachment, growth and maintenance on biodegradable aliphatic polyurethanes based on e-PCL and pluronics [163] in vitro for 42 days. The results demonstrated that porous scaffolds supported chondrocyte attachment and production of extracellular matrix proteins. However, with prolonged time in culture, the diffusion of large amount of matrix molecules into the culture medium and the cell dedifferentiation were noticed. In another study [198], the degradation of these polyurethanes was demonstrated by in vitro hydrolytic experiments (371C, pH 7.4 buffer). At 42 weeks in buffer, 2% weight loss and molecular weight reduction of 15–80% were observed and the extent of degradation was largely dependent on the polymer composition and the hydrophilic segment (Pluronic) content. In vitro degradation studies [200] of a family of poly(ester urethane)s (DegrapolTM) have shown that (PBS buffer at 371C and 701C) the rate of degradation is predominantly controlled by the number of easily hydrolyzable linkages. Degrapol block copolymers are built from HSs of poly [3-R-hydroxybutyrate)-co-(3-R-hydroxyvalerate)] (PHBV) and copolyester SSs ethylene glycol, poly(glycolic acid)diol and e-PCL. The study concluded

325 that the degradation rate is determined by the quantity and distribution of weak links in Degrapol; high glycolate-containing polymers degraded rapidly. Borkenhagen et al. [201] investigated a series of Degrapol polyester urethanes as nerve guidance channels (NGC). Nerve regeneration along an 8mm gap transected in the sciatic nerve of rats was demonstrated with NGC fabricated from Degrapol. Over a 24-week period weight losses of 33–88% were observed for polymer with different amounts of PHBV. Inﬂammatory reactions associated with polymer degradation did not interfere with nerve regeneration. Guan et al. [171] have synthesized a family of polyesterurethaneureas from poly(caprolactone) diol, BDI, lysine ethyl ester or putrescine as chain extender, and investigated the toxic effects of their in vitro degradation products. The polymers degraded with >50% mass loss in buffer, and endothelial cells cultured for 4 days with medium containing degradation products showed no toxic effects; surface modiﬁcation with RGD peptide further enhanced cell adhesion. Biodegradable polyurethanes exhibit good biocompatibility as demonstrated in a number of studies summarized above. Numerous in vitro degradation studies have demonstrated the degradation of these materials. In most polyurethane compositions, the nature of the SS forming polyol governs the degradation rates, although the diisocyanate structure and relative proportions of soft and hard segments also inﬂuence the degradation rates. What is lacking are detailed studies to understand the exact nature of the degradation products, and how those products are resorbed/released from the body. Injectable biodegradable polymers The design of synthetic polymer systems as injectable liquids, gels or pastes has received considerable research interests because of the potential opportunities to develop advanced therapies and products for repairing damaged tissues or organs. These injectable polymer systems have the advantage of employing minimally invasive procedures such as arthroscopic delivery to provide support while the damaged tissues are regenerated. Additionally they may also be useful for delivering cells, growth factors or other promoters to accelerate the tissue regeneration process. The biodegradable synthetic polymers described previously in this review have been experimented for use as prefabricated scaffolds with some success for tissue engineering application, but they have many limitations to adopt as injectable polymer systems that cure in situ or by an external trigger. Precursors developed based on the monomers used in many synthetic biodegradable polymers have the potential to be developed as injectable polymers. Figure 6 illustrates few examples of such precursors reported in the literature. This section reviews some of the recently reported synthetic injectable polymer systems with potential for use in advanced medical implants and tissue engineering technologies.

326 O

O HO

C

CH2)5

(CH2)p

O

OH

(CH2)5

C

O

m

n Poly(caprolactone) diol

(p = 2, 4 etc)

O OH

O O

HO O

CH3

CH3 n

Poly(propylenefumarate) diol CH3 HO

CH

O C

(CH2CH2O)

O

n

O

CH3

C

CH

O

H m

n Poly(ethyleneglycol-co-lactic acid) diol O HO

CH2)4

O

O (CH2)4

C

C

H

O n

Poly(tetramethyleneadipate) diol O C

O

C

R CH

O

H

m 4

Star polyols of glycolides (R=H) and lactides (R=CH3)

Fig. 6. Examples of polyols used in biodegradable polyurethanes.

Urethane-based injectable polymers Synthesis and properties: Gunatillake et al. [190–192,202–204] have developed polyurethane prepolymers that can be cross-linked to form both rigid and elastomeric materials (NovoSorbTM) useful in a range of biomedical applications including scaffolds for tissue engineering. The differential reactivity of the isocyanate functional groups in diisocyanates, such as LDI, is used to prepare prepolymers that are liquids at and above ambient temperatures by reacting with multi-hydroxy functional core molecules, such as pentaerythritol. Under controlled reaction conditions, star/hyperbranched prepolymer with isocyanate end-functional groups can be prepared (Scheme 2). For example, reacting a diisocyanate with a core molecule, such as pentaerythritol,

327 CH2OH

HOCH2

HOCH2

OCN

+

C

(CH2)4

NCO

CH

COOC2H5

CH2OH

lysine diisocyanate

Pentaerythritol

OCN

CH

(CH2)4

COOC2H5 OCN

CH

(CH2)4

NHCOOCH2

CH2OOCNH (CH2)4

NCO

COOC2H5

C NHCOOCH2

CH

CH2OOCNH (CH2)4

COOC2H5

CH

NCO

COOC2H5 Prepolymer

Scheme II. Synthetic route to prepare isocyanate end-functional prepolymer (A).

glucose or glycerol produces isocyanate end-functional prepolymers. A typical example of a prepolymer (Prepolymer A) is shown in Scheme 2. The second component (Prepolymer B) is usually a polyester polyol and suitable polyols include polycaprolactone, poly(ortho esters), poly(glycolic acid), polylactic acid and their copolymers (Fig. 6). The polyol component may be modiﬁed by adding a second polyol to alter hydrophilic/hydrophobic characteristics. Reaction of prepolymer A with B (along with other additives if needed) in appropriate proportions produces a cross-linked polymer network. Reaction of prepolymer with an appropriate hydroxyl compounds, biodegradable polyurethane networks can be prepared. With the appropriate choice of precursors, polyurethane with high mechanical strength can be prepared. For example, materials with compressive strength of 260 MPa and compressive modulus over 2 GPa have been reported [192]. These degradable polymers were developed as two-part systems with options to incorporate cells or other biological components to promote cell growth and to polymerize in situ. Porous solid polymers can be obtained by reacting with a cross-linker such as water, which generates carbon dioxide during curing. The polymer compositions can be formulated to cure with temperature rise controlled not to exceed body temperature [192]. Both in vitro and in vivo studies have demonstrated the biocompatibility and degradability of these polymers [205,206]. Free radically polymerizable injectable polymers Synthetic polymers: Free radically polymerizable synthetic polymers have found diverse applications in medicine ranging from simple externally applied scaffolds to internally placed bone cements. Most potential synthetic biomaterials developed for use in this area have injectable consistency, with or without cellular material and cured in a fraction of a second by free radical

328 polymerization. The macromer or prepolymers are designed to either cure with photo or redox initiation. The focus of this section will be on free radically cross-linkable biodegradable material that can be resorbed by the body over a period of time. Some have the potential of inclusion of biological material including laboratory-cultured cells that have the necessary phenotype to generate the required tissue. The reader is directed to brief review articles covering literature up to 2001 by Mikos et al. on ‘‘Photoinitiated polymerization of biomaterial’’ [207] and on ‘‘New directions in photopolymerizable biomaterials’’ by Anseth et al. [208]. Our group, initially at the Commonwealth Scientiﬁc and Industrial Research Organisation (CSIRO) and now at Polynovo Biomaterials, have been interested in the development of an in-situ curing polymer that a surgeon could quite rapidly and easily inject to a defective site by arthroscopic means, shaped to ﬁt irregular defectives sites before photo cured to a hard material when desired. We have called the method ‘‘cure on demand’’ and patented recently [209]. The core of these multi-arm polymers is designed to be low molecular weight lactic/glycolic homo or copolymers and the terminal ends functionalized with free radically cross-linkable groups. They can be tailored to degrade from very fast to very slow by changing the composition of the polymer chains and depending on the application at hand. Curing is achieved with visible or UV light, is very rapid and typically takes 20–60 s to obtain solid material. The attractions of cure on demand system are the injectability of polymers, making way to surgeons to perform minimally invasive procedures, the ability to cure under physiological conditions (wet environment and 371C) and when desired, leaving ample time for the surgeon to repair irregular sites and cure when desired with minimal heat generation. More recently, we have synthesized a variety of different polymers and polymer types, tested in vitro and in vivo and formulated a library of polymers that are suitable for a variety of applications. These applications vary from orthopedic, cartilage, scaffolds, to matrices for the delivery of biological materials. While initially the main aim was to formulate the material to have an injectable consistency, these prepolymers can be made to putty-like material that cures on demand, to prefabricated ﬁlms with thickness as low as 100 mm, 3D structures for tissue engineering, etc. O

O

O

O O

m

O

PEG-LA-DMA

O O

n

O m O

O

O n PEGDMA

Delivery of chondrocytes for cartilage repair is reported to be successful with PEG-based macromonomers. A group led by Kristi Anseth is a

329 signiﬁcant contributor to this area and have investigated the suitability of PEG-based materials for chondrocyte delivery [210]. A blend of linear PEG, chain extended with LA units and end capped with methacrylate groups with poly(ethylene glycol) dimethacrylate (PEGDMA) were used in this investigation. The chondrocytes are encapsulated in a 10–15% mixture of this blend in PBS, photo cross-linked to a hydrogel upon delivery. While the LA links in the polymer are introduced to enhance the degradability of the matrix, the latter dimethacrylate gives the strength by increased cross-link density and hence the mechanical strength, the cross-link density and the rate of degradation is controlled by the relative proportions of the two-blended materials. Chondrocytes encapsulated in these gels were shown to be viable for 6 weeks and produce cartilaginous tissue rich in glycosaminoglycans and collagen. The degradation of the blend with 15% PEGDMA was shown to match closely with the rate at which the cartilaginous tissue in growth. They have also examined the effect of many other factors that dictate survival of cells in the encapsulant. In one study [211] cross-link density of the blend is varied to investigate the effect of and correlate the rate of degradation of the encapsulant to the evolution of the extracellular matrix components. The degree of cross-linking also affects the mechanical strength of the scaffold material and inﬂuence the morphology of the chondrocytes encapsulated in an implant under compressive strain. The study highlights the importance of the optimization of the cross-link density to achieve ideal non-cell deformations and heterogeneity in designing encapsulant materials. Another study [212] has shown that an increase in the thickness of the scaffold material from 2 to 8 mm, a typical defect size in a cartilage, does not compromise the biochemical content of the cellular matrix produced. Cell studies have been conducted using these PEG hydrogels [213–216]. Mallapragada et al. [217] have synthesized and characterized copolymers of poly(ethylene glycol) methylether methacrylate (PEGMEM) and 2-(diethylamino)ethyl metacrylate (DEAEM) for biomedical applications including gene delivery. At a ratio of 30:70, PEGMEM:DEAEM, the copolymer is claimed to be water soluble and potentially injectable to the defect site. The cell viability studies carried out have shown the copolymer to be less toxic than the homopolymer of DEAEM. Cross-linked PVA derivatives (Fig. 7) have been used as hydrogels for cell delivery in many tissue-engineering studies. Preparation of PVA-based hydrogels are generally achieved by functionalization of the many pendant O O

+ OH

O

O

Fig. 7. Acrylate functionalization of PVA.

O

O OH

330 hydroxyl groups with glycidyl methacrylate (West) or methacrylamidoacetaldehyde dimethylacetal (Anseth). The ﬁnal cross-link density required to give the needed mechanical strength is adjusted by varying the degree of functionalization of the backbone of the PVA polymer. The efﬁciency by which these polymers can be cross-linked with the photo initiators and characterization of the ﬁnal hydrogels are well documented in the literature. West et al. [218] have seeded PVA hydrogels with human dermal ﬁbroblasts, cultured, and evaluated for cell viability, proliferation and extracellular matrix production. Viability of cells for 2 weeks with homogeneous cell proliferation and the production of extracellular matrix proteins were observed. Modiﬁed with RGD peptides these cross-linkable polymers were found to support increased attachment and proliferation of ﬁbroblasts in a dose-dependent manner. West et al has reviewed the advantages of photo polymerizable hydrogels, the photo initiators and the materials in current use. Anseth et al. [219] have led a study to design a PVA-based polymeric carrier for chondrocytes in cartilage repair. The key components of the investigated carrier were methacrylate functionalized PVA and chondroitin sulfate, cross-linked by UV light. Both the ingredients were tested individually and in a number of combinations to produce interpenetrating network structures and found to have various swelling and compressive strengths depending on the cross-link densities of the two components. While the crosslinked homopolymer chondroitin sulfate degraded completely in 20 h in the presence of chondroitinase ABC, the 50:50 blend with PVA-MA gave only a mass loss of 25% during the same period. Cell encapsulation work attempted with the same polymers in the presence of a cross-linkable RGD sequence demonstrated cell viability for 3 days. Detailed evaluation of the characterization [220] and degradation behavior of these types of hydrogels have been evaluated and compared to theoretical approaches. Results demonstrate that the theoretical predictions correlated well with mass loss data obtained and with increased cross-link density the complete degradation of the hydrogels increased. The experimental data were later compared to a bulk degradation mathematical model [221]. The group led by Kristi Anseth [222] has developed scaffolds having high porosity (80%) for bone tissue engineering. In this study [223], a few of cured linear oligomers of LA with ethylene glycol as the core, functionalized with terminal acrylate groups have been evaluated for bone formation. The porous scaffolds were fabricated using a particulate leaching technique with sodium chloride as the porogen. The scaffolds in vitro in PBS at 371C degraded with a 30–60% mass loss in 8 weeks. An in vivo rat model [222] based on these polymers to study bone formation in a critical-sized cranial defect with or without adsorbed osteoinductive growth factors has revealed the polymers to degrade in approximately 8 months. While more bone in growth was observed when growth factors were included in the implants compared to

331 untreated sites, while those that did not have growth factors (only polymer) were primarily ﬁlled with ﬁbrous tissue with mild inﬂammation after 9 weeks. O HO

O

OH

O O

n

poly(propylene fumarate)

O

O O

O

O

O

O O poly(propylene fumarate) diacrylate

Free radically cross-linkable poly(propylene fumarate)-based networks [224–230] have been extensively investigated for in vitro biocompatibility, degradability and cell viability [231,232]. These materials are evaluated as potential matrices to deliver cells in orthopedic applications. Developed as a blend that can be injected, the formulation is composed of a linear macromonomer, a diacrylate functionalized oligomer that is cured with a free radical UV initiator, bis(2,4,6-trimethyl benzoyl) phosphine oxide, at 365 nm. The linear macromonomer is an alternating copolymers of fumaric acid and propylene glycol having number average molecular weight set at 1,700 and 2,600 g/mol while the free radically cross-linkable oligomer is composed of the same fumarate core, end capped with two polypropylene glycol and two terminal acrylate units. The network structures are produced by free radical polymerization of the unsaturated groups within the repeating fumarate units and the terminal acrylate functions. The mechanical properties of the crosslinked polymers, curing proﬁles, their solution viscosities and heat generation during curing have been separately evaluated [233,234]. Inclusion of cross-linkers in the blend in varied proportions is used to control the cross-link density of the network and is shown to have an inﬂuence on cell viability/attachment. Fibroblast attachment is observed to be greatest to networks with the highest double bond conversions and increases with the increase of the acrylate-based cross-linker. This highlights the fact that the central fumarate double bond is less reactive than the acrylate group and participates in a lesser extent in the network formation. While the leachables from the non-cross-linked polymer mixtures were found to be cytotoxic, the accelerated degradation products (1 N NaOH, 601C) and each of the components of the oligomers, fumaric acid, acrylic acid, etc. separately was found to be non-cytotoxic at low concentrations. Two other cross-linking agents N-vinyl pyrrolidone (NVP), PEG-diacrylate (PEG-DA), have been studied earlier by the same group [235,236]. The latter cross-linker is used to formulate an injectable biodegradable hydrogel, together with oligo[poly(ethylene glycol)fumarate] (OPF), which has been evaluated for cell viability using MSC from rats [237]. The cells are shown to be viable after 2 and 24 h with 80% viable in 24 h for less than 25% (w/v) concentration of OPF oligomer. Although high viability is achieved at these

332 OPF levels, the general trend is decreased viability with increased fumarate oligomer in the blend. Of the two PEG-DA used, Mn 575 and 3,400, signiﬁcant increased cell viability is observed for the higher molecular weight PEG-DA. The effect of leachables from the hydrogels have separately been tested for cell viability and once cross-linked these leachable products have been shown to have minimal adverse effects on the viability of MSC with 90% cell viability. The redox free radical initiator system [238] used for these experiments is also shown to have minimal interference with cells, and each component, tested separately, is shown to be non-cytotoxic. Recently, novel biodegradable amino acid containing anhydride polymers have been developed for orthopedic applications [239]. These oligomers are based on methacrylated aminocarproyl maleamic acid, methacrylated alanyl maleamic acid, triethylene glycol dimethacrylate, etc. to afford novel blends. The compressive strengths for these blends varied from 31 to 114 MPa. The changes in mechanical strength during degradation have also been evaluated. O R

O

O

O O

O O

n=2,3,4

poly(CL/TMC) diacrylate

Liquid acrylate-endcapped biodegradable copolymers have been prepared by Matsuda et al. [240]. The copolymers of e-caprolactone and TMC are prepared by ring-opening polymerization in the presence of tin(II)2-ethyl hexanoate as catalyst in toluene and the resulting copolymer puriﬁed by precipitation. Trimethylene glycol and di, tri and tetra functional PEG derivatives are used as initiators. The hydroxyl-terminated copolymers are subsequently functionalized with acryloyl chloride to afford the cross-linkable prepolymers. Matsuda et al. in their investigation have shown that these copolymers cured with the radical initiator camphorquinone (and dimethylaminoethyl methacrylate as sensitizer) preferentially degrade by surface erosion except when PEG is used as the core initiator molecule. In this event the polymer absorb water due to the hydrophilic nature, swells and undergoes surface as well as bulk erosion as expected. Fabrication of sterolithographic microstructured architectures through rapid liquid to solid transformations using a moving UV-light pen and a computer-aided design program has been demonstrated in a subsequent communication [241]. In vivo subcutaneous implants in rats and micro-needle-structured surfaces loaded with antiinﬂammatory drugs have conﬁrmed the earlier observation of the degradation to be by surface erosion when the core was triethylene glycol. An initiator is a key component in a free radically cross-linkable formulation. The initiator molecule by interaction with photons in the case of

333 photo initiation or chemically with the interaction of two species (redox initiators) provides the initially radicals for the terminal double bonds to cross-link together. A wide range of initiators are commercially available for the curing of acrylate type formulations, often developed for coating applications, however, their use in biomedical applications is limited due to uncertainties on how high energy radicals would interact with cellular membranes. Several groups have initiated these studies to demonstrate that some initiators have no adverse effect on cell populations and cause minimal oxidative damage and subsequent cell death. Elisseeff and co-workers [242] have investigated the varied cytocompatibility of a number of free radical initiators on six cell lines. Cellular proliferation rates, growth kinetics and detailed comparison of available photo initiators are described. In another publication, Anseth and associates [243] have investigated the cytocompatibility of UV and visible light photo-initiating systems on NIH/3T3 ﬁbroblasts. Both studies have concluded that the commercial UV initiator Irgacure 2959 is well tolerated by many cell types and is suited to biomedical corrective therapies. Chemically modified natural polymers: HA is known to be involved in developmental events in the morphogenesis of many embryonic organs such as in cell proliferation in limb development, tendon regeneration and fetal wound repair. It is also known to be a major component of the cardiac jelly during the development of the heart, is non-thrombogenic and nonimmunogenic [244]. Combined with these known facts, its relative abundance from natural resources and ease with which it can be modiﬁed to a useful biological material, modiﬁed HA is regarded to be one of the key chemically modiﬁed natural candidate polymers that have found use in the area of biomedical implants. HA is a polysaccharide, composed of repeating units of N-acetyl glucosamine, with many un-functionalized hydroxyl groups. Hence, modiﬁcation of naturally occurring HA is mostly achieved by functionalizing the hydroxyl groups with the (meth) acrylate group. Many different methods are used to achieve this, but the most common ones being the use of methacrylic acid or methacrylic anhydride in the presence of a base. Following the reaction the methacrylated HA (MA-HA) is often washed, precipitated and dialyzed before use. Applications that have used modiﬁed HA vary from implantable bone, cartilage substitutes, scaffolds for would healing to HA gels being used as matrices for tissue engineering of heart valves. In one of the recent studies carried out by Anseth and co-workers [244] investigated the use of modiﬁed HA hydrogel scaffolds as a biological carrier for valvular interstitial cells (VIC). VICs are said to resemble myoﬁbroblasts, which are reported to play and important role in tissue remodeling. In this study, various molecular weights of HA were tested and the degradation products of HA hydrogel and the starting macromers were observed to signiﬁcantly increase VIC proliferation, the lower molecular

334 weights exhibiting the greatest stimulation compared to the controls. Addition of low molecular weight HA degradation products to VIC cultures were observed to increase by four- and two-fold in matrix and elastin production, respectively. In addition it was shown that VICs encapsulated with HA hydrogels remained viable with signiﬁcant elastin production in 6 weeks making way to better understanding of the relationship between HA and VICs. A new class of biodegradable hydrogels composed of a blend of functionalized poly(D,L)-lactic acid and hydrophilic dextran segments has been synthesized [245]. The unsaturated vinyl groups are separately introduced to the poly(D,L-lactic acid) (PDLLA) segments and dextran by acrolylation in the presence of a base. The mixture is cross-linked in DMSO using a free radical initiator, 5% w/w 2,2-dimethoxy 2-phenyl acetophenone. Many blends with different weight ratios have been prepared, cross-linked and characterized. The swelling ratios for the cross-linked blends have been evaluated and the various blended hydrogels are claimed to have a wide range of water absorption difﬁcult to achieve with pure hydrophilic gels. HA is known to participate in the differentiation, proliferation and migration of cells during wound healing and hence the attraction in its use in applications in simple and superﬁcial wound applications. The relative abundance of HA from natural resources and the ease with which it can be modiﬁed to materials that are of value has made modiﬁed HA one of the very attractive implantable biomaterials. Concluding remarks Among the many families of synthetic biodegradable polymers explored for various biomedical applications polyglycolides, polylactides and their copolymers remain the most widely investigated. Several products based on these polymers are in clinical use currently. The biodegradability of these polymers as well as the bioresorption of degradation products remains as the main attraction to this family of polymers for further exploitation for use in emerging technologies in the biomedical ﬁeld. Research efforts over the last two decades have focused on various copolymerization approaches to overcome some of the disadvantages of these polymers. Incorporation of hydrophilic segments to alter mechanical properties as well as to tailor degradation time to suit different applications have produced polymer systems for drug delivery applications and some tissue engineering technologies. Among other classes of biodegradable polymers, biodegradable polyurethanes offer many advantages, which include inherent good biocompatibility, processing versatility and combinations of structure variations to tailor properties ranging from soft elastomers to rigid materials. These features provide many opportunities to tailor materials to meet the needs of many emerging technologies. Polyphosphazenes and polyanhydrides are two other classes of polymers with useful properties for drug delivery applications.

335 The development of synthetic injectable polymers systems with capabilities to deliver cell and biological molecules still remains a major area of challenge for biomaterials researchers. Recent developments in urethane-based systems and free radically polymerizable precursors based on polyesters and urethanes have shown very promising results and these two systems have the greatest potential to overcome the many challenges in the delivery of cells and biological molecules, such as growth factors. These polymer systems offer the combination of versatility in tailoring structure to meet mechanical property speciﬁcations and degradation times as well as the options to cure in situ or on demand for ﬂexibility in delivery employing minimally invasive surgical procedures. References 1. 2. 3. 4.

5. 6. 7.

8.

9.

10.

11.

12. 13.

So¨derga˚rd A and Stolt M. Properties of lactic acid polymers and their correlation with composition. Prog Polym Sci 2002;27:1123–1163. Okada M. Chemical synthesis of biodegradable polymers. Prog Polym Sci 2002;27:87–133. Ueda H and Tabata Y. Polyhydroxyalkanoate derivatives in current clinical applications and trials. Adv Drug Deliver Rev 2003;55:501–518. Chu C-C and Park J. Biodegradable polymeric biomaterials: an updated overview. In: Biomaterials, Park JB and Bronzino JD (eds), CRC Press, Boca Raton, FL, 2003, pp. 95–115. Albertsson A-C and Varma IK. Aliphatic polyesters: synthesis, properties and applications. Adv Polym Sci 2002;157:1–40. Tsuji H and Ikada Y. Physical properties of polylacides. Curr Trend Polym Sci 1999;4:27–46. Amass W, Amass A and Tighe B. A review of biodegradable polymers: uses, current developments in the synthesis and characterisation of biodegradable polyesters, blends of biodegradable polymers and recent advances in biodegradation studies. Polym Int 1998;47:89–144. Mikos AG, Temenoff JS. Formation of highly porous biodegradable scaffolds for tissue engineering. Electron J Biotechnol 2003;3:114–119 (available on http://www.ejb. orgcontent/vol3/issue2/full/5). Jenn AC, Peter SJ and Mikos AG. Preparation and use of porous poly a-hydroxyester scaffolds for bone tissue engineering. In: Tissue Engineering Methods and Protocols, Morhgan JR and Yarmush ML (eds), Totowa, NJ, Humana Press, 2000, pp. 133–140. Hollinger JO, Jamiolkoski DD and Shalaby SW. Bone repair and a unique class of biodegradable polymers: the polyesters. In: Biomedical Applications of Synthetic Biodegradable Polymers, Hollinger JO (ed), Boca Raton, FL, CRC Press, 1997, pp. 197–222. Sawhney AS and Drumheller PD. Polymer synthesis. In: Frontiers in Tissue Engineering, Patrick CW, Mikos Ag and McIntire LV (eds), New York, Pergamon, 1998, pp. 83–106. Agrawal CM, Athanasiu KA and Heckman JD. Biodegradable PLA/PGA polymers for tissue engineering in orthopaedics. Mater Sci Forum 1997;250:115–128. Chu CC. An in-vitro study of the effect of buffer on degradation of poly(glycolic acid) sutures. J Biomed Mater Res 1981;15:19–27.

336 14. Chu CC. The in-vitro degradation of poly(glycolic acid) sutures-effect of pH. J Biomed Mater Res 1981;15:795–804. 15. Gilding DK. In: Biodegradable Polymers. Biocompatibility of Clinical Implant Materials, Vol. II, Williams DF (ed), Boca Raton, FL, CRC Press, 1981, pp. 209–232. 16. Hollinger JO. Preliminary report on osteogenic potential of a biodegradable copolymer of polylactide (PLA) and polyglycolide (PGA). J Biomed Mater Res 1983;17:71–82. 17. Grijpma DW, Nijehuis AJ, van Wijk PGT and Pennings AJ. High impact strength aspolymerized PLLA. Polym Bull 1992;29:571–578. 18. Fambri L, Pegoretti A, Fenner R, Incardona SD and Migliaresi C. Biodegradable ﬁbres of poly(L-lactic acid) produced by melt spinning. Polymer 1997;38(1):79–85. 19. To¨ma¨la¨ P. Biodegradable self-reinforced composite materials; manufacturing structure and mechanical properties. Clin Mater 1992;10:29–34. 20. Jacobson S and Fritz HG. Plasticizing polylactide – the effect of different plasticizers on the mechanical properties. Polym Eng Sci 1999;39(7):1303–1310. 21. Gilding DK and Reed AM. Biodegradable polymers for use in surgery-polyglycolic/ poly(lactic acid) homo-and copolymers: 1. Polymer 1979;20:1459–1464. 22. Miller RA, Brady JM and Cutright DE. Degradation rates of oral resorbable implants (polylactates and polyglycolates: rate modiﬁcation with changes in PLA/PGA copolymer ratios. J Biomed Mater Res 1977;11:711–719. 23. Reed AM and Gilding DK. Biodegradable polymers for use in surgery-poly(glycolic)/ poly(lactic acid) homo and copolymers 2. In-vitro degradation. Polymer 1981;22: 494–504. 24. Vert M, Christel P, Chabot F, Leary J, Hastings GW, Ducheyne P (eds), Macromolecular Biomaterials, Boca Raton, FL, CRC Press, 1984, pp. 119–142. 25. Grijpma DW, Zondervan GJ and Pennings AJ. High molecular weight copolymers of L-lactic and e-caprolactone as biodegradable elastomeric implant materials. Polym Bull 1991;25:327–333. 26. Grijpma DW and Pennings AJ. Polymerization temperature effects on the properties of L-lactic and e-caprolactone. Polym Bull 1992;25:335–341. 27. Kricheldorf HR and Kreiser I. Polylactones 13. Transesteriﬁcation of poly(L-lactic acid) with poly(glycolic acid), poly(e-propiolactone), and poly(e-caprolactone). J Macromol Sci Chem 1987;24A(11):1345–1356. 28. Gruvega˚rd M, Lindberg T and Albertsson A-C. Random copolymers of 1,5-dioxepan2-one. J Macromol Sci Pure Appl Chem A 1988;35(6):885–902. 29. Buchholz B. Analysis and characterisation of resorbable DL-lactide-trimetylene carbonate copolyesters. J Mater Sci: Mater Med 1993;4:381–388. 30. Chen S, Pieper R, Webster DC and Singh J. Triblock copolymers: synthesis, characterization, and delivery of a model protein. Int J Pharm 2005;288:207–218. 31. Kwon YM and Kim SW. Biodegradable triblock copolymer microspheres based on thermosensitive sol–gel transition. Pharm Res 2004;21:339–343. 32. Choi S, Kwon YM and Kim SW. Control of blood glucose by novel GLP-1 delivery using biodegradable triblock copolymer PLGA–PEG–PLGA in type 2 diabetic rats. Pharm Res 2004;21:827–831. 33. Ganaha F, Kao EY, Wong H, Elkins CJ, Lee J, Modanlou S, Rhee C, Kuo MD, Yuksel E, Cifra PN, Waugh JM and Dake MD. Stent-based controlled release of intravascular angiostatin to limit plaque progression and in-stent restenosis. J Vasc Interv Radiol 2004;15:601–608. 34. Jeong JH, Kim SW and Park TG. Biodegradable triblock copolymer of PLGA–PEG–PLGA enhances gene transfection efﬁciency. Pharm Res 2004;21:50–54.

337 35. Lee PY, Li Z and Huang L. Thermosensitive hydrogel as a TGF-beta gene delivery vehicle enhances diabetic would healing. Pharm Res 2003;20:1995–2000. 36. Kumar N, Ravikumar MNV and Domb AJ. Biodegradable block copolymers. Adv Drug Deliv Rev 2001;53:23–44. 37. Huh KM, Cho YW, Park K. PLGA–PEG block copolymers for drug formulations. Drug Deliv Technol 2003;3(5):42, 44–49. 38. Yoon JJ, Soon SH, Doo SL and Park TG. Immobilisation of cell adhesive RGD peptide onto the surface of highly porous biodegradable polymer scaffolds fabricated by a gas forming/salt leaching method. Biomaterials 2004;25:5613–5620. 39. Yoo HS, Soon SH, Doo SL and Park TG. Hyaluronic acid modiﬁed biodegradable scaffolds for cartilage tissue engineering. Biomaterials 2005;26:1925–1933. 40. Chun KW, Yoo HS, Yoon JJ and Park TG. Biodegradable PLGA micro carriers for injectable delivery of chondrocytes: effect of surface modiﬁcation on cell attachment and function. Biotechnol Prog 2004;20:1797–1801. 41. Breitenbach A and Kissel T. Biodegradable comb polyesters: part I. Synthesis, characterisation and structural analysis of poly(lactide) and poly(lactic-co-glycolide) grafted onto water soluble poly(vinyl alcohol) as backbone. Polymer 1998;39: 3261–3271. 42. Breitenbach A, Pistel KF and Kissel T. Biodegradable comb polyesters: part II. Erosion and release properties of poly(vinyl alcohol)-graft-poly(lactic-co-glycolic acid). Polymer 2000;41:4781–4792. 43. Pistel KF, Breitenbach A, Zange-Volland R and Kissel T. Brush-like branched biodegradable polyesters: part III. Protein release from microspheres of poly(vinyl alcohol)-graft-poly(lactic-co-glycolic acid). J Control Release 2001;73:7–20. 44. Jung T, Breitenbach A and Kissel T. Sulfobutylated poly(vinyl alcohol)-graft-poly(lactide-co-glycolide)s facilitate the preparation of small negatively charged biodegradable nanospheres. J Control Release 2000;67:157–169. 45. Ouchi T and Ohya Y. Design of lactide copolymers as biomaterials. J Polym Sci Part A Polym Chem 2004;42:453–462. 46. Craig PH, Kelly RJ, Higgin GA. Arch Surg 1970;100:486. 47. Blomstedt B, Jacobson S. Acta Chir Scand 1977;143:259. 48. Katz A, Mukherjee DP, Kaganov AL, Gordon S. Surg Gynecol Obstet 1985;161:213. 49. Rodeheaver GT, Beltran KA, Green CW, Fraulkner BC, Stiles BM, Stanimir GW, Traeland H, Fried GM, Brown HC, Edlich RF. J Long Term Med Implants 1996;6:181. 50. Bezwada RS, Jamiolkowski DD, Lee I-Y, Agrawal V, Persivale J, Trenka-Benthin S, Erneta M, Suryadevara J and Liu S. Monocryl suture, a new ultra-pliable absorbable monoﬁlament suture. Biomaterials 1995;16:1141–1148. 51. Tomihata K, Suzuki M, Oka T and Ikeda Y. A new resorbable monoﬁlament suture. Polym Deg Stab 1998;59:13–18. 52. Willaim DF and Mort E. Enzyme-accelerated hydrolysis of polyglycolic acid. J Bioeng 1977;1:231–238. 53. Nelson JF, Stanford HG and Cutright DE. Evaluation and comparison of biodegradable substances as osteogenic agents. Oral Surg 1977;43:836–843. 54. Schakenraad JM, Nieuwenhues P, Molenaar I, Helder J, Dykstra PJ and Feijen J. Invivo and in-vitro degradation of glycine/DL-lactic copolymers. J Biomed Mater Res 1989;23:1271–1288. 55. vanSliedregt A, vanBlitterswijk CA, Hesseling SC, Grote JJ and deGroot K. The effect of molecular weight of polylactic acid on in-vivo biocompatibility. Adv Biomater 1990;9:207–212.

338 56. vanSliedregt A, Radder AM, deGroot K and vanBlitterswijk CA. In-vitro biocompatibility testing of polylactides. Part I: Proliferation of different cell types. J Mater Sci: Mater Med 1992;3:365–370. 57. Verheyen CCPM, deWijn JR, vanBlitterswijk CA, Rozing PM and de Groot K. Examination of efferent lymph nodes after 2 years of transcortical implantation of poly(L-lactide) containing plugs: a case report. J Biomed Mater Res 1993;27:1115–1118. 58. Tayler MS, Daniels AU, Andriano KP and Heller J. Six bioabsorbable polymers: invitro acute toxicity of accumulated degradation products. J Appl Biomater 1994;5:151–157. 59. Gibbons DK. Tissue response to resorbable synthetic polymers. In: Degradation Phenomena on Polymeric Biomaterials, Plank H, Dauner M and Renardy M (eds), New York, Springer, 1992, pp. 97–104. 60. Bo¨stman OM. Osteolytic changes accompanying degradation of absorbable fracture ﬁxation implants. J Bone Joint Surg 1991;73B:679–682. 61. Bo¨stman OM. Intense granulomatous inﬂammatory lesions associated with absorbable internal ﬁxation devices made of polyglycolide in ankle fractures. Clin Orthop 1992;278:178–199. 62. Bo¨stman OM, Pa¨ivaa¨rinta U, Partio E, Vasenius J, Manniner M and Rokkanen P. Degradation and tissue replacement of an absorbable polyglycolide screw in the ﬁxation of rabbit osteomies. J Bone Joint Surg 1992;74A:1021–1031. 63. Bo¨stman OM, Partio E, Hirvensalo E and Rokannen P. Foreign-body reaction to polyglycolide screws. Acta Orthop Scand 1992;63:173–176. 64. Holland SJ, Tighe BJ. Biodegradable polymers. In: Advances in Pharmaceutical Sciences, Vol. 6, London, Academic Press, 1992, pp. 101–104. 65. Hyashi T. Biodegradable polymers for biomedical applications. Prog Polym Sci 1994;19:663–702. 66. In’tVeld PJA, Velner EM, Van DeWhite P, Hamhuis J, Dijkstra PJ and Feijen J. Melt block copolymerisation of e-caprolactone and L-lactic acid. J Polym Sci Part I Polym Chem 1997;35:219–226. 67. Storey RF and Tayler AE. Effect of stanneous octoate on the composition, molecular weight and molecular weight distribution of ethylene-glycol-initiated poly(e-caprolactone). J Macromol Sci Pure Appl Chem A 1998;A35:750. 68. Hori Y, Yamaguchi A and Hagiwara T. Chemical synthesis of high molecular weight poly(3-hydroxybutyrate-co-4-hydroxybutyrate). Polymer 1995;36(24):4703–4705. 69. Kronenthal RL. Biodegradable polymers in medicine and surgery. Polym Sci Technol 1975;8:119–137. 70. Middleton JC and Tipton AJ. Synthetic biodegradable polymers as orthopaedic devices. Biomaterials 2000;21:2335–2346. 71. Gabelnick HL. Long acting steroid contraception. In: Advances in Human Fertility and Reproductive Endocrinology, Vol. 3, Mishell Jr. DR (ed), New York, Raven Press, 1983, pp. 149–173. 72. Pitt CG, Marks TA, Schindler A, Naltrexone Willette RE, Barnet G (eds), National Institute on Drug Abuse Research Monogram, Vol. 28, 1981, pp. 232–253. 73. Heller J, Barr J, Ng SY, Schwach-Abdellauoi K and Gurny R. Poly(ortho esters); synthesis, characterisation, properties and uses. Adv Drug Deliv Rev 2002;54:1015–1039. 74. Heller J, Barr J, Ng SY, Schwach-Abdellauoi K, Emmahl S, Rothen-Weinhold A and Gurny R. Poly(ortho esters) – their development and some recent applications. Eur J Pharmaceut Biopharmaceut 2000;50:121–128.

339 75. NG SY, Vandamme T, Tayler MS and Heller J. Synthesis and erosion studies of selfcatalyzed poly(ortho esters). Macromolecules 1977;30:770–772. 76. Einmahl S, Behan-Cohen F, Tabatabay C, Savoldelli M, D’Hermies F, Chauvaud D, Heller J and Gurny R. Visocus bioerodible poly(ortho ester) as a new biomaterial for intraocular applications. J Biomed Mater Res 2000;50(4):556–573. 77. Andriano KP, Tabata Y, Ikada Y and Heller J. In vitro and in vivo comparison of bulk and surface hydrolysis in absorbable polymer scaffolds for tissue engineering. J Biomed Mater Res (Appl Biomater) 1999;48:602–612. 78. Kumar N, Langer R and Domb A. Polyanhydrides: an overview. Adv Drug Deliv Rev 2002;54:889–910. 79. Go¨pferich A and Tessmar J. Polyanhydride degradation and erosion. Adv Drug Deliv Rev 2002;54:911–931. 80. Katti DS, Lakshmi S, Langer R and Laurencin CT. Adv Drug Deliv Rev 2002;54:933–961. 81. Kohn J and Langer R. Bioresorbable and bioerodible materials. In: An Introduction to Materials in Medicine, Ratner BD, Hoffman AS, Schoen FJ and Lemon JE (eds), Santiago, Academic Press, 1997, pp. 65–73. 82. Ashammakhi N and Rokkanen P. Absorbable polyglycolide devices in trauma and bone surgery. Biomaterials 1997;18:3–9. 83. Rosen HB, Chang J, Wnek GE, Linhardt R and Langer R. Bioerodible polyanhydrides for controlled drug delivery. Biomaterials 1983;4:131–133. 84. Brem H, Piantadosi K, Burger PC, Walker M, Selker R, Vick NA, Black K, Sisti M, Brem G, Mohr G, Muller P, Morawetz R and Schold SC. Lancet 1995;345:1008. 85. Leong KW, Domb A and Langer A. Polyanhydride. In: Encyclopaedia of Polymer Science and Technology, Mark HF and Kroschwitz JI (eds), New York, Wiley, 1989, p. 648. 86. Domb AJ, Amselem S, Sha J and Maniar M. Polyanhydride – synthesis and characterization. Adv Polym Sci 1993;107:93–141. 87. Domb AJ and Langer R. Polyanhydrides I: Preparation of high molecular weight polyanhydrides. J Polym Sci Part A Polym Chem 1987;25:3373–3386. 88. Domb AJ, Ron E and Langer R. (Polyanhydride), One step polymerization using phosgene or diphosgene as coupling agents. Macromolecules 1988;21:1925–1929. 89. Leong KW, Simonte V, Langer R. Synthesis of polyanhydride: melt condensation dehydrochlorination and dehydration coupling. Macromolecules 1987:20:705–712. 90. Burkoth AK and Anseth KS. A review of photo-cross linked polyanhydrides: in-situ forming degradable networks. Biomaterials 2000;21:2395–2404. 91. Albertsson AC and Eklund M. Short methylene segment cross links in degradable aliphatic polyanhydride: network formation, characterisation, and degradation. J Appl Polym Sci Polym Chem 1996;34:1395–1404. 92. Davies MC, Khan MA, Domb A, Langer R, Watts JF and Paul Aj. The analysis of the surface chemical structure of biomedical aliphatic polyanhydrides using XPS and time of ﬂight SIMS. J Appl Polym Sci 1991;42:1597–1605. 93. Conix A. Aromatic polyanhydrides: a new class of high melting ﬁbre-forming polymers. J Polym Sci Part A 1958;29:343–353. 94. Leong KW, Brott BC and Langer R. Bioerodible polyanhydrides as drug-carrier matrices, I: Characterisation, degradation and release characteristics. J Biomed Mater Res 1985;19:941–955. 95. Uhrich KE, Thomas TT, Laurencin CT and Langer R. In-vitro degradation characteristics of poly(anhydride-amide) containing trimellitylimodoglycine. J Appl Polym Sci 1997;63:1401–1411.

340 96. Uhrich KE, Thomas TT, Laurencin CT and Langer R. Synthesis and characterisation of degradable polyanhydrides. Macromolecules 1995;28:2148–2193. 97. Attawia MA, Uhrich KE, Botchwey E, Fan M, Langer R and Laurencin CT. Cytotoxicity testing of poly(anhydride) for orthopaedic applications. J Biomed Mater Res 1995;29:1233–1240. 98. Seidel JO, Uhrich KE, Laurencin CT and Langer R. Erosion of poly(anhydride-coimides): a preliminary mechanistic study. J Appl Polym Sci 1996;62:1277–1283. 99. Muggli DS, Burkoth AK, Keyser SA, Lee HR and Anseth KS. Reaction behaviour of biodegradable, photo cross-linkable polyanhydrides. Macromolecules 1998;31:4120–4125. 100. Anseth KS, Svaldi DC, Laurencin CT and Langer R. Photo-polymerisation of novel degradable networks for orthopaedic applications. In: Photo-polymerisation, ACS Symposium Series 673, Scranton A, Bowman V and Peiffer R (eds), Washington, DC, American Chemical Society, 1997, pp. 189–202. 101. Laurencin CT, Peirrie-Jacques HM and Langer R. Toxicology and biocompatibility considerations in the evaluation of polymeric materials for biomedical applications. Clin Lab Med 1990;10:549–570. 102. Temenoff JS and Mikos AG. Injectable biodegradable materials for orthopaedic tissue engineering. Biomaterials 2000;21:2405–2412. 103. Rokici G. Prog Polym Sci 2000;25:259. 104. Pego AP, Van LM, Brouwer JA, Van LA, van Wachem PB, Poot AA, Grijpma DW and Feijen JJ. Biomed Mater Res 2003;A67:1044. 105. Zhu KJ, Hendren RW, Jensen K and Pitt CG. Macromolecules 1991;67A:1044. 106. Matsuda T, Kwon K and Kidoaki S. Biomacromolecules 2004;5:295. 107. Yu C, Zhang L and Shen Z. J Mol Catal A Chem 2004;212:365. 108. Ariga T, Takata T and Endo T. Macromolecules 2002;35:746. 109. Wang H, Dong JH and Qiu KY. J Polym Sci Part A Polym Chem 1998;36:695. 110. Rokikicki G. Prog Polym Sci 2000;25:259. 111. Wan T, Zou T, Cheng SX and Zhuo RX. Biomacromolecules 2005;6:524–529. 112. Tangpasuthadol V, Pendharkar SM and Kohn J. Hydrolytic degradation of tyrosinederived polycarbonates, a class of new biomaterials. Part I: Study of model compounds. Biomaterials 2000;21:2371–2378. 113. Tangpasuthadol V, Pendharkar SM, Peterson RC and Kohn J. Hydrolytic degradation of tyrosine-derived polycarbonates, a class of new biomaterials. Part II: Study of model compounds. Biomaterials 2000;21:2379–2387. 114. Pulapura S and Kohn J. Tyrosine derived polycarbonates: backbone modiﬁed pseudopoly(amino acids) designed for biomedical applications. Biopolymers 1992;32:411–417. 115. Arvanitoxannis I, Nakayama A, Kawasaki N and Yamamoto N. Angew Makrom Chem 1994;222:111. 116. Goodman I and Vachon RN. Polym J 1984;20(6):529. 117. Fan Y, Kobayashi M and Kise H. J Polym Sci 2001;AI39:1318. 118. Touminnen J and Seppala JV. Macromolecules 2000;33:3530. 119. Okada M. Chemical synthesis of biodegradable polymers. Prog Polym Sci 2002;27:87–133. 120. Qain ZY, Li S, He Y and Liu XB. Hydrolytic degradation study of biodegradable polyesteramidecopolymers based on e-caprolactone and 11-aminoundecanoic acid. Biomaterials 2004;25:1975–1981. 121. Qian ZY, Li S, He Y and Liu XB. Thermal and hydrolytic degradation behaviour of degradable poly(esteramideco-polymers based on e-caprolactone and aminoundecanoic acid and poly(ethylene glycol). Polym Degrad Stabil 2004;84:41–47.

341 122. Qian ZY, He Y and Liu XB. Structure and property study of degradable polyesteramide ﬁber processing and alkaline degradation behaviour. Polym Degrad Stabil 2004;83:127–132. 123. Qian ZY, Li S and Liu XB. Synthesis and thermal degradation of biodegradable polyesteramide based on e-caprolactone and 11-aminoundecanoic acid. Polym Degrad Stabil 2003;81:279–286. 124. Qian ZY, Li S and Liu XB. Synthesis characterization and in vitro degradation of biodegradable polyesteramide based on lactic acid. Colloid Polym Sci 2003;281:869–875. 125. He Y, Qian ZY and Liu XB. Alkaline degradation behaviour of polyesteramide ﬁbres: surface erosion. Colloid Polym Sci 2004;282:972–978. 126. Qian ZY, Li S, He Y and Liu XB. Preparation of biodegradable polyesteramide microspheres. Colloid Polym Sci 2004;282:1083–1088. 127. Zou YB, Wang L and Liu X. Stabilisation and mechanical properties of biodegradable polyesteramide and its ﬁlled composites. Polym Degrad Stabil 2004;83:87–92. 128. Arabuli N, Tsitlanadze G, Edilashvili L, Kharadze D, Goguadze T, Beridze T, Gomurashvili Z and Katsarava R. Macromol Chem Phys 1994;195:2279. 129. Katsarava R, Beridze V, Arabuli N, Kharadze D, Chu CC and Won CY. J Polym Sci Part A Polym Chem 1999;37:391. 130. Suarez N, Brocchini S and Kohn J. Biomaterials 1998;19:2347. 131. Koyama E, Sanda F and Endo T. J Polym Sci Part A Polym Chem 1997;35:345. 132. Koyama E, Sanda F and Endo T. J Polym Sci Part A Polym Chem 1997;35:2925. 133. Villuendas I, Iribarren JI and Munoz-Guerra S. Macromolecules 1999;32:8015. 134. Kise H, Kobayashi M and Fan Y. J Polym Sci Part A Polym Chem 2001;39:1318–1328. 135. Liu X, Zhang H, He Y, Li S and Qian Z. Colloid Polym Sci 2004;282:1083–1088. 136. Qian Z, He Y, Zou Y, Li S and Liu X. Polym Degrad Stabil 2004;83:127–132. 137. Ahmad S, Ashraf SM, Naqvi F, Yadav S and Hasnat A. Prog Org Coat 2003;47:95–102. 138. Park C, Kim Y, Yoo TY and Im SS. J Appl Polym Sci 2003;90:2708–2714. 139. Liu X, Zhang H, He Yi and Li S. Polym Degrad Stabil 2005;88:309–316. 140. Stokes HN. On the chloronitrides of phosphorous. Am Chem J 1895;17:275–290. 141. Allcock HR and Kugel RL. Synthesis of high polymeric alkoxy and aryloxy phosphonitriles. J Am Chem Soc 1965;87:4216–4217. 142. Allcock HR, Kugel RL and Valan KJ. Phosphonitrile compounds VI. High molecular weight polyoxy- and aryloxy phosphazenes. Inorg Chem 1966;5:1709–1715. 143. Allcock HR, Kugel RL and Valan KJ. Phosphonitrile compounds VI. High molecular weight polyoxy- and aryloxy phosphazenes. Inorg Chem 1966;5:1716–1718. 144. Allcock HR and Chu CTW. Reaction of phenyl lithium with polydichlorophosphazene. Macromolecules 1979;12:551–555. 145 Mark JE, Allcock HR, and West R. Inorganic Polymers, New Jersey, Prentice-Hall, 1992, pp. 61–140. 146. Allcock HR. Inorganic–organic polymers as route to biodegradable materials. Macromol Symp 1999;144:33–46. 147. Behravesh E, Yasko AW, Engle PS and Mikos AG. Synthetic biodegradable polymers for orthopaedic applications. Clin Ortho Relat Res 1999;367S:118–185. 148. Qui LY and Zhu KJ. Novel biodegradable polyphosphazenes containing glycine ethyl ester and benzyl ester of amino acethydroxamic acid as co-substituents: synthesis, characterization and degradation properties. J Appl Polym Sci 2000;77:2955–2987. 149. Scopelianos AG. Polyphosphazenes as new biomaterials. In: Biomedical Polymers Designed-to-degrade Systems, Shalaby SW (ed), New York, Hanser, 1994, pp. 153–171.

342 150. Allcock HR, Fuller TJ, Mack DP, Matsumura K and Smeltz KM. Synthesis of poly[(aminoacid alkyl ester)phosphazenes]. Macromolecules 1977;10:824–830. 151. Allcock HR, Pucher SR and Scopelianos AG. Poly[(aminoacid ester) phosphazenes]: synthesis, crystallinity and hydrolytic sensitivity in solution and the solid state. Macromolecules 1994;27:1071–1075. 152. Allcock HR and Kwon S. Glyceryl polyphosphazenes: synthesis, properties and hydrolysis. Macromolecules 1988;21:1980–1985. 153. Allcock HR, Pucher SR and Scopelianos AG. Polyphosphazenes with glucosyl and methyl amino, triﬂuroethoxy, phenoxy or (methoxyethoxy) ethoxy side groups. Macromolecules 1991;24:23–34. 154. Allcock HR, Pucher SR and Scopelianos AG. Synthesis of poly(organophosphazenes) with glycolic acid ester and lactic acid ester side groups: prototypes for new bioerodible polymers. Macromolecules 1994;27:1–4. 155. Lakshmi S, Katti DS and Laurencin CT. Biodegradable polyphosphazenes for drug delivery applications. Adv Drug Deliv Rev 2003;55:467–482. 156. Allcock HR, Fuller TJ and Matsumura K. Hydrolysis pathways for aminophosphazenes. Inorg Chem 1982;21:515–521. 157. Laurencin CT, Maria EN, Elgendy HM, El-Amin SF, Allcock HR, Pucher SW and Ambrosio AA. Use of polyphosphazenes for skeletal tissue regeneration. J Biomed Mater Res 1993;27:963–973. 158. Laurencin CT, El-Amim SF, Ibim SE, Willoughby DA, Attawia M, Allcock HR and Ambrosio AA. A highly porous 3-dimensional polyphosphazene polymer matrix for skeletal tissue regeneration. J Biomed Mater Res 1996;30:133–138. 159. Gunatillake PA, Meijs GF and McCarthy SJ. Developments in design and synthesis of biostable polyurethanes. In: Biomedical Applications of Polyurethanes, Vermette P, Griesser HJ, Laroche G and Guidoin R (eds), Georgetown, Landes Biosciences, 2001, pp. 160–170. 160. Tuominen J, Kylma J, Kapanen A, Venelampi O, Ita¨vaara and Seppa¨la¨ J. Biodegradation of lactic acid based polymers under controlled composting conditions and evaluation of the ecotoxicological impact. Biomacromolecules 2002;3:445–455. 161. Vande Velde K and Kiekens P. Biopolymers: overview of several properties and consequences on their applications. Polym Test 2002;21:433–442. 162. Gorona K and Gogolewski S. Novel biodegradable polyurethanes for medical applications. ASTM Spec Tech Publ 2000;STP1396:39–57. 163. Gorona K and Gogolewski S. Biodegradable polyurethanes for implants. II. In-vitro degradation and calciﬁcation of materials from poly(e-caprolactone)-poly(ethylene oxide) diols and various chain extenders. J Biomed Mater Res 2002;60:592–606. 164. Saad GR, Lee YJ and Seliger H. Synthesis and characterisation of biodegradable poly(esterurethanes) based on bacterial poly(R-3-hydroxybutyrate). J Appl Polym Sci 2002;83:703–718. 165. Cohn D, Stern T, Gonza´les MF and Epstein J. Biodegradable poly(ethylene oxide)/ poly(e-caprolactone multiblock copolymers. J Biomed Mater Res 2002;59:273–281. 166. Gorona K and Gogolewski S. Novel biodegradable polyurethanes for medical applications. In: Synthetic Biodegradable Polymers for Implants, Vol. STP 1936, Agrawal CM, Parr JE and Lin St (eds), Philadelphia, American Society of Testing and Materials, 2000, pp. 39–57. 167. Gorona K and Gogolewski S. In vitro degradation of novel medical biodegradable aliphatic polyurethanes based on e-caprolactone and Pluronics with various hydrophilicities. Polym Deg Stab 2002;75:113–122.

343 168. Izhar U, Schwalb H, Borman JB, Hellener GR, Hotoveli-Salomon A, Marom G, Stern T and Cohn D. Novel synthetic selectively degradable vascular prostheses: a preliminary implantation study. J Surg Res 2001;95:152–160. 169. Kylma¨ J, Hiljanen-Vainio M and Seppa¨la¨ J. Miscibility, morphology and mechanical properties of rubber-modiﬁed biodegradable poly(ester-urethanes). J Appl Polym Sci 2000;76:1074–1084. 170. Gorona K, Polowinski S and Gogolewski S. Synthesis and characterisation of biodegradable poly(e-caprolactone urethane)s, I. Effect of polyol molecular weight, catalyst, and chain extender on the molecular weight and physical characteristics. J Polym Sci Part A Polym Chem 2002;40:156–170. 171. Guan J, Sacks MS, Beckman E and Wagner W. Synthesis, characterisation, and cytocompatibility of elastomeric, biodegradable poly(ester-urethane)ureas based on poly(caprolactone) and putrescine. J Biomed Mater Res 2002;61:493–503. 172. Spaans CJ, De Goot JH, Belgraver VW and Pennings AJ. A new biomedical polyurethane with high modulus based on 1,4-butanediisocyanate and e-caprolactone. J Mater Sci: Mater Med 1998;9(12):675–678. 173. De Groot JH, Spaans CJ, Dekens FG and Pennings AJ. On the role of aminolysis and transesteriﬁcation in the synthesis of e-caprolactone and L-lactide based polyurethanes. Polym Bull 1998;41:299–306. 174. Bruin P, Venstra GJ, Nijenhuis AJ and Pennings AJ. Design and synthesis of biodegradable poly(ester-urethane) elastomer networks composed of non-toxic building blocks. Makromol Chem Rapid Commun 1988;9:589–594. 175. Storey RF, Wiggins JS, Mauritz KA and Puckett AD. Bioresorbable composites. II: Nontoxic, L-lysine-based poly(ester urethane) matrix composites. Polym Compos 1993;14:17. 176. Storey RF, Wiggins JS and Puckett AD. Hydrolyzable poly(ester urethane) networks from L-lysine diisocyanate and D,L-lactide/e-caprolactone homo and copolyester triols. J Polym Sci Part A Polym Chem 1994;32:2345–2363. 177. Zang JY, Beckman EJ, Piesco NP and Agrawal S. A new peptide-based urethane prepolymer: synthesis, biodegradation, and potential to support cell growth in-vitro. Biomaterials 2000;21:1247–1258. 178. Hirt TD, Neuenschwander P and Suter UW. Synthesis of degradable, biocompatible, and tough block copolyesterurethanes. Macromol Chem Phys 1996;197:4253–4268. 179. De Groot JH, Nijenhuis AJ, Bruin P, Pennings AJ, Veth RPH, Klompmaker J and Jansen HWB. Use of porous biodegradable implants in meniscus reconstruction. 1. Preparation of porous biodegradable copolyurethanes for the reconstruction of meniscus lesions. Colloid Polym Sci 1990;268:1073–1081. 180. Saad B, Hirt TD, Welti M, Uhlschmid GK, Neuenschwander P and Suter UW. Development of degradable polyurethanes for medical applications: in-vitro and in-vivo evaluation. J Biomed Mater Res 1997;36:65–74. 181. Skarja GA and Woodhouse KA. Synthesis and characterisation of degradable polyurethane elastomers containing an amino acid-based chain extender. J Biomater Sci Polym Ed 1998;9:271–295. 182. Skarja GA and Woodhouse KA. Structure–property relationships of degradable polyurethane elastomers containing an amino acid-based chain extender. J Appl Polym Sci 2000;75:1522–1534. 183. Gunatillake P, Adhikari R, Moore T. Biodegradable polyurethanes and polyurethaneureas. PCT Patent Application, 2005. 184. Moore T, Adhikari R, Gunatillake P, Barton A, Biodegradable polyurethane. Annual Meeting of the Australian Society of Biomaterials, 2004.

344 185. Dahiyal BI, Posadas EM, Hirosue S, Hostin E and Leong KW. Degradable biomaterials with elastomeric characteristics and drug-carrier function. React Polym 1995;25:101–109. 186. Gogollewski S and Pennings AJ. Biodegradable materials of polylactides 4, Porous biomedical materials based on mixtures of polylactides and polyurethanes. Makromol Chem Rapid Commun 1982;3:839–845. 187. Gogollewski S and Pennings AJ. An artiﬁcial skin based on biodegradable mixtures of polylactides and polyurethanes for full thickness skin wound covering. Makromol Chem Rapid Commun 1983;4:675–680. 188. Spaans CJ, Belgraver VW, Rienstra O, De Groot JH, Veth RPH and Pennings Aj. Solvent-free fabrication of micro-porous polyurethane amide and polyurethane–urea scaffolds for repair and replacement of the knee-joint meniscus. Biomaterials 2000;21:2453–2460. 189. De Groot JH, De Vrije R, Pennings AJ, Klompmaker J, Veth RPH and Jansen HWB. Use of porous polyurethanes for meniscal reconstruction and meniscal prostheses. Biomaterials 1996;17:163–173. 190. Gunatillake PA, Adhikari R. Biodegradable polyurethane/urea compositions. PCT Patent Application PCT/AU2003/000935, 2003. 191. Gunatillake PA, Adhikari R, Le TPT, Danon SJ, Seymour K, Bean P, Thissen H, Werkmeister JA, Ramshaw JAM, White JF, Glattauer V and Tebb T. Injectable biodegradable polyurethanes for tissue engineering. Seventh World Biomaterial Congress, Sydney, Vol. 703, 2004. 192. Adhikari R, Gunatillake V, Mayadunne RTA, Houshyar S, Karunaratne O, Grifﬁths IM. Injectable biodegradable polyurethanes for orthopaedic application. 30th Meeting, Society for Biomaterials, Memphis, TN, USA, 2005, p. 442. 193. Storey R and Hickey TP. Degradable polyurethane networks based on D,L-lactide, ecaprolactone, and trimethylene carbonate homopolymer and copolymer triols. Polymer 1994;35(4):831–838. 194. Bogdanov B, Toncheva V and Schacht E. Synthesis and characterisation of poly(esterurethanes). Macromol Symp 2000;152:117–126. 195. Lee S, Yu SC and Lee YS. Degradable polyurethanes containing poly(butylene succinate) and poly(ethylene glycol). Polym Deg Stab 2001;72:81–87. 196. Van Tienen TG, Heijkants GJC, Buma P, de Groot JH, Pennings AJ and Veth RPH. Tissue in growth and degradation of two biodegradable porous polymers with different porosities and pore size. Biomaterials 2002;23:1731–1738. 197. Tienen TG, Heijkants RGJC, Buma P, De Groot JH, Pennings AJ and Veth RPH. A porous polymer scaffold for meniscal lesion repair – a study in dogs. Biomaterials 2003;24:2541–2548. 198. Bruin P, Smedinga J, Pennings AJ and Jokman MF. Biodegradable lysine diisocyanatebased poly(glycolide-co-e-caprolactone)-urethane network in artiﬁcial skin. Biomaterials 1990;11:191–295. 199. Grad S, Kupcsik L, Gorona K, Gogolewski S and Alini M. The use of biodegradable polyurethane scaffolds for cartilage tissue engineering: potential and limitations. Biomaterials 2003;24:5163–5171. 200. Lendlein A, Colussi M, Neuenschwander P and Suter UW. Hydrolytic degradation of phase-segregated multiblock copoly(ester urethanes)s containing weak links. Macromol Chem Phys 2001;202:2702–2771. 201. Borkenhagen M, Stoll RC, Neuenschwander P, Suter U and Aebischer P. In vivo performance of a new biodegradable polyester urethane system used as a nerve guidance channel. Biomaterials 1998;19:2155–2165.

345 202. Adhikari R, Gunatillake PA, Le TPT, Danon SJ, Seymour K, Bean P, Thissen H, Werkmeister JA, Ramshaw JAM, White JF, Glattauer V and Tebb T. Injectable biodegradable polyurethanes for tissue engineering: effect of phosphorylcholine. Seventh World Biomaterial Congress, Sydney, Vol. 621, 2004. 203. Adhikari R, Gunatillake PA, Le TPT, Danon SJ, Seymour K, Bean P, Thissen H, Werkmeister JA, Ramshaw JAM, White JF, Glattauer V and Tebb T. Effect of sulfonium zwitterions on properties and degradation of biodegradable polyurethanes. Seventh World Biomaterial Congress, Sydney, Vol. 518, 2004. 204. Adhikari R, Gunatillake PA, Le TPT, Danon SJ, Seymour K, Bean P, Thissen H, Werkmeister JA, Ramshaw JAM, White JF, Glattauer V and Tebb T. Injectable biodegradable polyurethanes for cartilage repair: evaluation of biocompatibility and biodegradability. Seventh World Biomaterial Congress, Sydney, Vol. 1618, 2004. 205. John RF, Gunatillake PA, Raju A, Jerome W, John R. Preliminary in vivo evaluation of a polymeric adhesive for meniscal defect repair. World Biomaterial Congress, Sydney, Vol. 333, 2004. 206. White JF, Glattauer JF, Tebb TA, Gunatillake T, Adhikari R, Mayadunne R, Ramshaw JAM, Werkmeister JA. Regeneration of articular cartilage using beads and polymer gel delivery systems. Seventh World Biomaterial Congress, Sydney, May 2004. 207. P Fisher J, Dean D, Engel PS and Mikos AG. Photoinitiated polymerization of biomaterials. Ann Rev Mater Res 2001;31:171–181. 208. Anseth KS and Burdick JA. New directions in photopolymerizable biomaterials. MRS Bull 2002;27:130–136. 209. Mayadunne R, Gunatillake P, Adhikari A. Biocompatible polymer compositions for dual or multi-stage curing. Australian Provisional Patent Application 2004901111, 2004. 210. Bryant SJ, Bender RJ, Durand KL and Anseth KS. Encapsulating chondrocytes in degrading PEG hydrogels with high modulus: engineering gel structural changes to facilitate cartilaginous tissue production. Biotechnol Bioeng 2004;86:747–755. 211. Bryant SJ, Anseth KS, Lee DA and Bader DL. Crosslinking density inﬂuences the morphology of chondrocytes photoencapsulated in PEG hydrogels during the application of compressive strain. J Orthopaed Res 2004;22:1143–1149. 212. Bryant S and Anseth KS. The effects of scaffold thickness on tissue engineering cartilage in photocrosslinked poly(ethylene oxide) hydrogels. Biomaterials 2001;22:619–626. 213. Bryant SJ and Anseth KS. Hydrogel properties inﬂuence ECM production by chondrocytes photoencapsulated in poly(ethylene glycol) hydrogels. J Biomed Mater Res 2002;59:63–72. 214. Bryant SJ and Anseth KS. Controlling the spatial distribution of ECM components in degradable PEG hydrogels for tissue engineering cartilage. J Biomed Mater Res, Part A 2003;64A:70–79. 215. Burdick JA, Mason MN, Hinman AD, Thorne K and Anseth KS. Delivery of osteoinductive growth factors from degradable PEG hydrogels inﬂuences osteoblast differentiation and mineralization. J Control Release 2002;83:53–63. 216. Nuttelman CR, Tripodi MC and Anseth KS. Synthetic hydrogel niches that promote hMSC viability. Matrix Biol 2005;24:208–218. 217. Anderson BC and Mallapragada SK. Synthesis and characterization of injectable, water-soluble copolymers of tertiary amine methacrylates and poly(ethylene glycol) containing methacrylates. Biomaterials 2002;23:4345–4352. 218. Schmedlen RH, Masters KS and West JL. Photopolymerizable hydrogels for tissue engineering applications. Biomaterials 2002;23:4325–4332.

346 219. Bryant SJ, Davis-Arehart KA, Luo N, Shoemaker RK, Arthur JA and Anseth KS. Synthesis and characterization of photopolymerized multifunctional hydrogels: watersoluble poly(vinyl alcohol) and chondroitin sulfate macromers for chondrocyte encapsulation. Macromolecules 2004;37:6726–6733. 220. Martens P and Anseth KS. Characterization of hydrogels formed from acrylate modiﬁed poly(vinyl alcohol) macromers. Polymer 2000;41:7715–7722. 221. Martens PJ, Bowman CN and Anseth KS. Degradable networks formed from multifunctional poly(vinyl alcohol) macromers: comparison of results from a generalized bulk-degradation model for polymer networks and experimental data. Polymer 2004;45:3377–3387. 222. Burdick JA, Frankel D, Dernell WS and Anseth KS. An initial investigation of photocurable three-dimensional lactic acid based scaffolds in a critical-sized cranial defect. Biomaterials 2003;24:1613–1620. 223. Burdick JA, Philpott LM and Anseth KS. Synthesis and characterization of tetra functional lactic acid oligomers: a potential in situ forming degradable orthopaedic biomaterial. J Polym Sci Part A Polym Chem 2001;39:683–692. 224. Timmer MD, Shin H, Horch RA, Ambrose CG and Mikos AG. In vitro cytotoxicity of injectable and biodegradable poly(propylene fumarate)-based networks: unreacted macromers, cross-linked networks, and degradation products. Biomacromolecules 2003;4:1026–1033. 225. Timmer MD, Carter C, Ambrose CG and Mikos AG. Fabrication of poly(propylene fumarate)-based orthopaedic implants by photo-crosslinking through transparent silicone molds. Biomaterials 2003;24:4707–4714. 226. Timmer MD, Ambrose CG and Mikos AG. Evaluation of thermal- and photo-crosslinked biodegradable poly(propylene fumarate)-based networks. J Biomed Mater Res, Part A 2003;66A:811–818. 227. Fisher JP, Timmer MD, Holland TA, Dean D, Engel PS and Mikos AG. Photoinitiated cross-linking of the biodegradable polyester poly(propylene fumarate). Part I. Determination of network structure. Biomacromolecules 2003;4:1327–1334. 228. Fisher JP, Holland TA, Dean D and Mikos AG. Photoinitiated crosslinking of the biodegradable polyester poly(propylene fumarate). Part II. In vitro degradation. Biomacromolecules 2003;4:1335–1342. 229. Hedberg EL, Shih CK, Lemoine JJ, Timmer MD, Liebschner MAK, Jansen JA and Mikos AG. In vitro degradation of porous poly(propylene fumarate)/poly(-lactic-coglycolic acid) composite scaffolds. Biomaterials 2005;26:3215–3225. 230. Hedberg EL, Kroese-Deutman HC, Shih CK, Crowther RS, Carney DH, Mikos AG and Jansen JA. In vivo degradation of porous poly(propylene fumarate)/poly(-lactic-coglycolic acid) composite scaffolds. Biomaterials 2005;26:4616–4623. 231. Payne RG, McGonigle JS, Yaszemski MJ, Yasko AW and Mikos AG. Development of an injectable, in situ crosslinkable, degradable polymeric carrier for osteogenic cell populations. Part 2. Viability of encapsulated marrow stromal osteoblasts cultured on crosslinking poly(propylene fumarate). Biomaterials 2002;23:4373–4380. 232. Payne RG, McGonigle JS, Yaszemski MJ, Yasko AW and Mikos AG. Development of an injectable, in situ crosslinkable, degradable polymeric carrier for osteogenic cell populations. Part 3. Proliferation and differentiation of encapsulated marrow stromal osteoblasts cultured on crosslinking poly(propylene fumarate). Biomaterials 2002;23:4381–4387. 233. Timmer MD, Jo S, Wang C, Ambrose CG and Mikos AG. Characterization of the cross-linked structure of fumarate-based degradable polymer networks. Macromolecules 2002;35:4373–4379.

347 234. Timmer MD, Ambrose CG and Mikos AG. In vitro degradation of polymeric networks of poly(propylene fumarate) and the crosslinking macromer poly(propylene fumarate)diacrylate. Biomaterials 2002;24:571–577. 235. Shin H, Temenoff JS and Mikos AG. In vitro cytotoxicity of unsatd. oligo[poly(ethylene glycol) fumarate] macromers and their crosslinked hydrogels. Biomacromolecules 2003;4:552–560. 236. Shin H, Quinten Ruhe P, Mikos AG and Jansen JA. In vivo bone and soft tissue response to injectable, biodegradable oligo (poly(ethylene glycol) fumarate) hydrogels. Biomaterials 2003;24:3201–3211. 237. Temenoff JS, Park H, Jabbari E, Conway DE, Shefﬁeld TL, Ambrose CG and Mikos AG. Thermally crosslinked oligo(polyethylene glycol fumarate) hydrogels support osteogenic differentiation of encapsulated marrow stromal cells in vitro. Biomacromolecules 2004;5:5–10. 238. Temenoff JS, Shin H, Conway DE, Engel PS and Mikos AG. In vitro cytotoxicity of redox radical initiators for crosslinking of oligo(poly(ethylene glycol) fumarate macromers. Biomacromolecules 2003;4:1605–1613. 239. Xie D, Chung I, Puckett AD and Mays JW. Novel biodegradable amino acid containing anhydrided oligomers for orthopedic applications. J App Poly Sci 2005;96:1979–1984. 240. Mizutani M and Matsuda T. Liquid acrylate-endcapped biodegradable poly(e-caprolatone-co-trimethylene carbonate). I. Preparation and visible light-induced photocuring characteristics. J Biomed Mater Res 2002;62:387–394. 241. Matsuda T and Mizutani M. Liquid acrylate-endcapped biodegradable poly(e-caprolatone-co-trimethylene carbonate). II. Computer aided strereolithographic microarchtectural surface photoconstructs. J Biomed Mater Res 2002;62:395–403. 242. Williams CG, Malik AN, Kim TK, Manson PN and Elisseeff JH. Variable cytocompatibility of six cell-lines with photoinitiators used for polymerizing hydrogels and cell encapsulation. Biomaterials 2005;26:1211–1218. 243. Bryant SJ, Nuttelman CR and Anseth KS. Cytocompatibility of UV and visible light photoinitiating systems on cultured NIH/3T3 ﬁbroblasts in vitro. J Biomater Sci Polym Ed 2000;11:439–457. 244. Masters KS, Shah DN, Leinwand LA and Anseth KS. Crosslinked hyaluronan scaffolds as a biologically active carrier for valvular interstitial cells. Biomaterials 2005;26:2517–2525. 245. Zang Y, Won C-Y and Chu C-C. Synthesis and characterisation of biodegradble network hydrogels having both hydrophobic and hydrophilic components with controlled swelling behaviour. J Polym Sci A Polym Chem 1999;37:4554–4569.

349

Bleaching with lignin-oxidizing enzymes Pratima Bajpai, Aradhna Anand and Pramod K. Bajpai Thapar Centre for Industrial Research & Development, Patiala, India Abstract. General concern about the environmental impact of chlorine bleaching efﬂuents has led to a trend towards elementary chlorine-free or totally chlorine free bleaching methods. Considerable interest has been focused on the use of biotechnology in pulp bleaching, as large number of microbes and the enzymes produced by them are known to be capable of preferential degradation of native lignin and complete degradation of wood. Enzymes of the hemicellulolytic type, particularly xylan-attacking enzymes xylanases are now used commercially in the mills for pulp treatment and subsequent incorporation into bleach sequences. Certain white-rot fungi can delignify Kraft pulps increasing their brightness and their responsiveness to brightening with chemicals. The fungal treatments are too slow but the enzymes produced from the fungi can also delignify pulps and these enzymatic processes are likely to be easier to optimize and apply than the fungal treatments. This article presents an overview of the developments in the application of lignin-oxidizing enzymes in bleaching of chemical pulps. The present knowledge of the mechanisms on the action of enzymes as well as the practical results and advantages obtained on the laboratory and industrial scale are discussed. Keywords: biobleaching; lignin-oxidizing enzymes; white-rot fungi; lignin peroxidase; laccase; manganese peroxidase.

Introduction Organochlorine compounds have been the matter of concern in the pulp and paper industry. These compounds are produced mainly by the reactions between residual lignin present in wood ﬁbers and the chlorine used for bleaching. Some of these compounds are found to be toxic, mutagenic, persistent, bioaccumulating and to cause harm to biological systems. Because of concerns about the short- and long-term environmental effects of chlorinated organic compounds, governments in many countries have imposed limits on their discharge. The implementation of non-chlorine bleaching technologies, e.g., use of oxygen, hydrogen peroxide, ozone and enzymes is being examined to reduce or eliminate the use of chlorine in the bleach plant [1]. Use of biotechnology in pulp bleaching has attracted considerable attention and has achieved interesting results in recent years. Biobleaching is one of the promising alternatives for eliminating chlorine-based chemicals in pulp bleaching process. The aims of the enzymatic treatment depend on the actual mill conditions and may be related to environmental demands, reduction of chemical costs or Corresponding author: Tel: +91-175-2393570. Fax: +91-175-2364012, 2365522.

E-mail: [email protected], [email protected] (P. Bajpai). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12010-4

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

350 maintenance/improvement of the product quality. It is being studied in several laboratories all over the world. The most commonly used enzymes in bleaching are hydrolytic enzymes such as xylanases. Xylanase applications are often referred to as prebleaching or bleach boosting because the nature of the effect is to enhance the effect of bleaching chemicals rather than to remove lignin directly [2–7]. The enzyme does not attack the lignin-based chromophores but rather the xylan network by which the residual lignin particles are surrounded and trapped. A limited hydrolysis of the xylan network is often sufﬁcient to facilitate the subsequent chemical attack on lignin with various bleaching chemicals, without sacriﬁcing the yield. Mill scale experience has demonstrated that xylanase pretreatment usually results up to 20–25% savings in bleaching chemicals with simultaneous reduction of pollutant emissions [8–13]. The most common problems with xylanase treatment cited in a mill survey have been corrosion of equipment and maintenance of the brown stock residence time [8]. Corrosion of mild steel facilities, due to sulfuric acid used to reduce the pulp pH, has been encountered in several mills. The brown stock residence time must be maintained for as long as possible but usually at least 1–2 h to obtain the maximum beneﬁts of enzyme treatment. This sometimes means that the mills must maintain the storage tower nearly full, which curtails its ability to act as a buffer between the pulping mill and the bleach plant. Other problems reported with enzyme treatment included difﬁculties in application and in bleach plant control. These relate to subtle action of enzymes, which is not easily observed on-line or in rapid testing. A decreased tear strength and pitch formation was also reported in some mills [8]. Thus, the effect of xylanase prebleaching is limited due to its indirect mode of action and other problems. Therefore, the enzyme needs to be more effective than the xylanase. Many of the fungi that attack wood in nature are able to degrade lignin. So, there appears to be an opportunity to exploit them for direct attack on lignin in pulps, an attack that can be speciﬁc and effective even if the amount of residual lignin is low. The organisms that most effectively biodegrade lignin are the Basidiomycete fungi. Lignin-oxidizing species are called whiterot fungi, because they typically turn wood white as it decays. This effect is another indication that these fungi may be useful in pulp bleaching [14–27]. But a serious shortcoming of the fungal bleaching process is the long incubation time required for contact with the biomass. Typical contact periods range from 5 to 14 days for both hardwood and softwood pulps. The reaction time, required at its current stage of development, makes it practically not feasible and economically unattractive [18,28–32]. There is a need to identify/ develop fast growing white-rot fungal cultures, which could do the job in less time. As the fungal treatments are too slow, the potential candidates for deligniﬁcation must selectively decompose the lignin in the ﬁber and likely be easier to optimize and apply than the fungal treatments. Signiﬁcant efforts

351 have been made to study the potential of enzymes produced by white-rot fungi for bleaching of chemical pulps. Several reports in the literature suggest that the enzymes lignin peroxidases (LiP), manganese peroxidases (MnP) and laccases (L) could prove useful in bleaching of pulps [19–20,33–90]. Lignin-oxidizing enzymes White-rot fungi are the main producers of lignin-oxidizing enzymes (Table 1). These fungi secrete a number of oxidative enzymes and some hitherto unknown substances (mediators) into their environment together effecting a slow but continuous degradation. The most important lignin-oxidizing enzymes are Table 1. Important white-rot fungi producing ligninolytic enzymes. White-rot fungi

Ligninolytic enzymes

Coriolus versicolor

Lignin peroxidase, manganese peroxidase

Phlebia radiata Pleurotus ostreatus Pleurotus saju-caju Pleurotus sanguineus Cyathus bulleri Coriolus pruinosum

Lignin peroxidase, manganese peroxidase

Phanerochate chrysosporium Ceriporiopsis subvermispora

Manganese peroxidase, laccase

Dichomitus squalens Lentinus edodes Panus tigrinus Rigidoporus lignosus Polyporus varius

Lignin peroxidase

Bjerkandera adusta Daedaleopsis confragora Pycnoporus cinnabarinus

Laccase

Qudemansiella radicata

Lignin peroxidase, laccase

Pleurotus florida Polyporus pletensis Polyporus brumalis Phlebia tremellosus Phlebia ochraceofulva Bjerkandera adusta Pleurotus eryngii Pleurotus ostreatus Coriolus versicolor Pleurotus sajor-caju

Aryl alcohol oxidase, veratryl alcohol oxidase

352 lignin peroxidases, manganese peroxidases and laccases. Lignin peroxidase and manganese peroxidase appear to constitute a major component of the ligninolytic system. The crystallographic structure of lignin peroxidase from Phanerochaete chrysosporium has been studied at different resolutions [91–92]. The model comprises all 343 amino acids, one heme molecule and three sugar residues. The crystal structure reveals that the enzyme consists mostly of helical folds with separate domains on either side of the catalytic heme. The enzyme also contains four disulﬁde bridges formed by the eight cysteine residues and two structural calcium ions, which appear to be important for maintaining the integrity of the active site [92]. Manganese peroxidase presents a good sequence homology with lignin peroxidase. However, except for preliminary information [93] no detailed reports on the structure of manganese peroxidase have yet been published. Presently, an attempt to crystallize laccase is in progress at University of Georgia, Athens (personal communication). The lignin peroxidases are able to catalyze the oxidation of non-phenolic aromatic rings in lignin into cation radicals in the presence of hydrogen peroxide. Manganese-dependent peroxidases oxidize phenolic units in lignin. These require Mn2+, which is oxidized to Mn3+ in the presence of chelators and hydrogen peroxide. Mn3+ is the real oxidizing agent attacking the lignin molecule. Laccase uses molecular oxygen as a co-substrate. The enzyme oxidizes phenolic subunits in lignin and simultaneously reduces oxygen to water. The substrate range of laccases can be extended to non-phenolic subunits by adding readily oxidized substrates. Possible mechanisms Lignin peroxidase oxidizes its substrate by two consecutive one-electron oxidation steps, with intermediate cation-radical formation. Due to its high redox potential, lignin peroxidase can also oxidize non-phenolic methoxysubstituted lignin subunits. The enzyme can depolymerize dilute solutions of lignin, oxidize and degrade a variety of dimers and oligomers structurally related to lignin in vitro, and catalyze the production of activated oxygen species [94–96]. Lignin peroxidase has a similar catalytic cycle to that of the horseradish peroxidase. The ﬁve-redox states of lignin peroxidase have been characterized, and its catalytic cycle has been investigated by various groups of scientists [97,98]. Lignin peroxidase catalyzes a large variety of reactions, e.g., cleavage of b-0-4 ether bonds and Ca–Cb bonds in dimeric lignin model compounds the basis for the depolymerization reactions catalyzed by lignin peroxidase. The enzyme also catalyzes decarboxylation of phenyl-acetic acids, oxidation of aromatic Ca-alcohols to Ca-oxocompounds, hydroxylation, quinone formation and aromatic ring opening [96,99–100]. Studies of lignin peroxidase oxidation of dihydropteroate synthase (DHPS) conﬁrm the involvement of lignin peroxidase in the initial degradation of lignin

353 [100–101]. It has been shown that veratryl alcohol is oxidized to a cation radical capable of participating in intermolecular charge transfer reactions [102]. Based on this property of the cation radicals, it was proposed that the cation radicals of veratryl alcohol, the product of lignin peroxidase catalysis, might mediate in the oxidation of lignin. In addition, these radicals may assist in the reaction of lignin peroxidase compound II with the reducing agent and thereby maintain the active peroxidase cycle. The reduction of lignin peroxidase II to native enzyme results in the release of two substrate cation radicals (Fig. 1a). The mechanism by which substrate cation radicals promote reduction of compound II is not clear. The cation radicals may form a pathway of charge transfer away from the vicinity of the porphyrin so that the chance of radical generation in the vicinity of the active site of lignin peroxidase is minimized. This is unlikely to happen with phenoxy radicals, which are incapable of participating in charge transfer reactions. Their formation would also result in lignin peroxidase inhibition (Fig. 1b). Results from different laboratories have shown that veratryl alcohol can act both as a stabilizer of lignin peroxidase and as a charge transfer mediator. Lignin peroxidase has a preference for phenolic compounds as compared to veratryl alcohol. Thus, in order to achieve maximum enzyme turnover and oxidation of non-phenolic compounds, the presence of phenolic compounds in the vicinity of lignin peroxidase needs to be low [102]. Moreover, oxidation of phenolic compounds causes inhibition of lignin peroxidase [103]. This inhibition can be overcome by lowering the H2O2 level relative to phenol and enzyme levels. Although much of the research has focused on lignin peroxidases, these enzymes are not necessarily involved in lignin degradation and may not be secreted by all lignin-oxidizing fungi [46]. Coriolus versicolor produces laccases as well as lignin- and manganese-dependent peroxidases; however, Archibald [104] found that lignin peroxidases secreted by C. versicolor did not appear to play an important role in lignin degradation. This enzyme could not be detected in pulp bleaching culture and, moreover, addition of an excess of lignin peroxidase inhibitor did not interfere with bleaching [104]. Manganese peroxidase acts exclusively as a phenol oxidase on phenolic substrates using Mn2+/Mn3+ as an intermediate redox couple. Several lines of evidence suggest that manganese peroxidase is a key enzyme in fungal bleaching: (1) catalase, an enzyme that destroys H2O2 inhibits the bleaching, (2) mutants of C. versicolor deﬁcient in manganese peroxidase do not bleach, and bleaching activity is partially restored by addition of manganese peroxidase, (3) isolated manganese peroxidase produces partial deligniﬁction of pulps when supplied with H2O2, Mn2+ and chelator [105]. Manganese peroxidase has a catalytic cycle very similar to that of lignin peroxidase. However, in this case, Mn2+ readily reduces compound II to the native enzyme to complete the catalytic cycle [106] (Fig. 2). This enzyme is involved in the oxidation of phenols and phenolic lignin structures. It oxidizes Mn2+ to

354

Fig. 1. Catalytic cycle of lignin peroxidase showing the oxidation of (a) Veratryl

alcohol (substrate modiﬁed compound II) and (b) phenolic compounds (enzyme feed back control) VA-veratryl alcohol; X-phenolic compound.

Mn3+ in the presence of a proper chelating agent and Mn3+ must form a complex with the chelators before it oxidizes phenolic substrates [107]. Organic acids are good chelators and white-rot fungi are producers of oxalic acid, malonic acid, pyruvic acid and malic acid. Mn3+/oxalate and Mn3+/ malonate form very stable chelators, which probably also function in vivo. Malonate facilitates Mn3+ dissociation from the enzyme and has a relatively low Mn2+ binding constant [107]. The puriﬁed manganese peroxidase has

355

Fig. 2. Catalytic cycle of manganese peroxidase and its ﬁve oxidation state ROOH H2O2, AH2 phenolic substrate.

Fig. 3. Oxidative pathway for catalytic action of laccase on lignin.

been reported to depolymerize DHP and also to degrade high molecular mass chlorolignins [108]. Mn3+ complex can oxidize phenolic lignin substructures by acting as a mediator between the enzyme and the polymer, and leads to the formation of phenoxy radicals as intermediates [109]. Subsequently, Ca–Cb cleavage or alkyl–phenyl cleavage would yield depolymerized fragments including quinones and hydroxyquinones. Figure 3 shows the oxidative pathway for catalytic action of manganese peroxidase on lignin.

356 Laccase appears to play an important role in degradation of lignin. Both constitutive and induced forms of laccases are known [110–113]. All laccasses are glycoproteins and they generally contain four copper ions [114,115]. These are distributed among three different binding sites and each copper ion appears to play an important role in the catalytic mechanism. Laccase is a true phenol oxidase with broad speciﬁcity toward aromatic compound containing hydroxyl and amine groups. The enzyme oxidizes phenols and phenolic substructures by one electron abstraction with formation of radicals that can either repolymerize or lead to depolymerization [96]. Figure 4 shows the oxidative pathway for catalytic action of laccase on lignin. In an effort to elucidate the role of phenolic and non-phenolic lignin subunits in a laccase mediator (LM) system, vanillyl alcohol was oxidized with laccase in the presence and absence of the mediator 1-hydroxybenzotriazol (HBT) [80]. Furthermore, the role of phenolic, aliphatic hydroxyl, and carboxylic acid moieties in lignin degradation was elucidated by selectively blocking them. The modiﬁed samples were then subjected to laccase and laccase HBT treatments. On the basis of this data it was possible to establish the role of this mediator. HBT mediates the oxidation of lignin by inducing side-chain oxidation and oxygen-addition products rather than oxidative coupling reactions. Jurasek et al. [81] attempted to develop a model of the three-dimensional structure of lignin to provide a framework for interpretation and prediction of interactions between the enzymes and lignin on the molecular level. Inspection of the model of ligniﬁed secondary wall showed the accessibility restrictions upon lignin-oxidizing enzymes, and points toward a concept of an enzyme factory outside the cell wall, producing chemicals, which have access to and break down the lignin and allow its fragments to leave the network. Once out of the wall, these fragments may interact with enzymes directly, resulting in further reduction of the fragment size and ﬁnally mineralization. Studies of side-chain cleavage and ring opening of lignin model compounds have shown that both laccase and lignin peroxidase that catalyze one-electron oxidation of

Fig. 4. Oxidative pathway for catalytic action of manganese peroxidase on lignin.

357 either phenolic or non-phenolic compounds, are involved in the initial degradation of lignin substructure model compounds [96]. Phenolic and nonphenolic (permethylated) synthetic [14C] lignins were depolymerized by Trametes villosa laccase in the presence of a radical mediator, 1-HBT [78]. Gel permeation chromatography of the treated lignins showed that approximately 10% of their substructures were cleaved. The system also cleaved a bO-4-linked model compound, 1-(4-ethoxy-3-methoxy-ring-[14C]phenyl)-2-(2methoxyphenoxy)-propane-1,3-diol, and a b-1-linked model, 1, 2-bis-(3-methoxy-4-[14C]methoxyphenyl)-propane-1,3-diol, which represents non-phenolic substructures in lignin. High performance liquid chromatography of products from the oxidized models showed that they were produced in sufﬁcient yields to account for the ability of laccase/HBT to depolymerize non-phenolic lignin. Laccase has also been shown to catalyze the cleavage of aromatic rings in a similar way to lignin peroxidase [116]. Laccase catalyzes demethoxylation reactions of terminal phenolic units. It can also degrade b-dimers and b-0-4 dimers via Ca oxidation alkyl–aryl cleavage and Ca–Cb cleavage [116]. It was observed that laccase from Ceriporiopsis subvermispora transformed 4,6-ditert-butylguaiacol into a ring opening product, the muconolactone derivative 2,4-di-(tert-butyl), 4-(methoxy-carbonyl methyl)-2-buten-4-olide. This work clearly demonstrated that 18O incorporated into muconolactone was from 18 O2 but not from H218O. It is obvious from the results that laccase, like LiP can cleave aromatic rings. Until 1990, laccase had been considered to be able to degrade only phenolic lignin model compounds [99]. However, Bourbonnais and Paice [117] reported that oxidation of non-phenolic lignin substructures by laccase from Trametes versicolor took place in the presence of a suitable redox mediator i.e., the dye 2,20 -azinobis (3-ethylbenzthiazoline-6-sulfonate) (ABTS). Laccase along with ABTS has also been shown to also delignify Kraft pulp [34]. Comparative studies were carried out on the kinetics and mechanism of pulp biobleaching with laccase-mediator system (LMS) with two different mediators, 1-HBT and N-hydroxyacetanilide (NHA) [58,85]. The kinetic studies on Kappa number reduction and dioxygen uptake suggest that a very fast rate of deligniﬁcation with NHA at the beginning of the process is the result of fast formation of the oxidized mediator species. However, a very slow deligniﬁcation rate after the initial phase (0.5–1 h) could be caused by low stability of the mediator species. After the reaction time of 2 h, the degree of deligniﬁcation is higher when HBT is used as mediator. In contrast to the deligniﬁcation with NHA, the formation of the oxidized mediator species is the ratedetermining step of the pulp biobleaching with dioxygen in the LMS using HBT as mediator. Increase in temperature increases the rate of chemical reactions, but decreases the laccase stability. Increasing oxygen pressure improves the efﬁciency of deligniﬁcation due to better penetration of the reagents, but does not affect the rate of chemical reactions. The presence of the mediator prevented repolymerization of Kraft lignin by T. versicolor laccase [44]. Other phenolic reagents have also been proposed and found to work as mediators

358 [43]. The brightening and deligniﬁcation of Kraft pulp has been shown with mediator 1-HBT [35–36]. Moreover, the production of a fungal metabolite, which acts as a physiological lacasse/redox mediator, has been shown to be produced by the white-rot fungus Pycnoporus cinnabarinus in University of Georgia (personal communication). The mechanism of enzymatic oxidation of HBT, violuric acid (VA), and NHA and three N-OH compounds was studied with seven fungal laccases [79]. The oxidation had a bell-shaped pH-activity proﬁle with an optimal pH ranging from 4 to 7. The oxidation rate was found to be dependent on the redox potential difference between the N-OH substrate and laccase. A laccase with a higher redox potential or an N-OH compound with a lower redox potential tended to have a higher oxidation rate. Similar to the enzymatic oxidation of phenols, phenoxazines, phenothiazines, and other redox-active compounds, an ‘‘outer-sphere’’ type of single-electron transfer from the substrate to laccase and proton release are speculated to be involved in the rate-limiting step for N-OH oxidation. Paice et al. [45] have suggested that Cellobiose quinone oxidoreductase (CBQase), a quinone-reducing enzyme also plays a number of roles in deligniﬁcation of pulps. Studies with CBQase have shown that the enzyme can reduce reaction products of laccase or peroxidases such as quinones, radicals and Mn2+ [118]. Concurrently, cellobionate, a potential chelator of Mn3+ is formed. However, cellobionate is less efﬁcient than malonate or oxalate as a chelator in the manganese peroxidase catalytic cycle, probably because it has a higher binding constant for Mn2+. Performance of lignin-oxidizing enzymes in bleaching Several patents have been ﬁled concerning the use of ligninases for bleaching of Kraft pulps [35,36,119–125]. One of these makes use of a chemically modiﬁed lignin peroxidase [125]. In addition, an European patent application exists for the production of recombinant ligninase [120]. Arbeloa et al. [33] used lignin peroxidase enzyme from P. chrysosporium to improve the bleachability of hardwood and softwood Kraft pulps. Enzyme treatment prior to chemical bleaching increased brightness and decreased lignin content in the pulp. The ﬁnal brightness of pulp was found to be higher by about 0.8–0.9 points as compared to control. Egan [126] has reported kappa number reductions of 24 and 26% after treatment with ligninase 118 from P. chrysosporium followed by alkaline extraction. He reported that pulp viscosity did not change but did not report any brightness data. Other researchers have not been able to achieve much bleaching effects from either pure or crude enzyme preparations from P. chrysosporium. At Paprican, manganese peroxidase from T. versicolor was examined for bleaching of Kraft pulp [45,46,105]. Studies on the effect of enzyme treatment on deligniﬁcation of hardwood Kraft pulp showed about 14% reduction in kappa number. The deligniﬁcation was found to be accompanied by the release of

359 methanol from phenolic methoxyl groups in Kraft pulp lignin. With softwood Kraft pulp, deligniﬁcation was observed over a wide range of initial lignin contents (kappa number) (Table 2). Only the pulp of lowest initial kappa number gave a higher brightness after enzyme treatment. Subsequent treatment with alkaline hydrogen peroxide resulted in pulps up to 7–8 points brighter than those obtained without enzyme (Table 3). In the absence of added manganese and malonate buffer, comparable brightness gain of 6 points could be achieved. Moreira et al. [63] isolated the MnP of Bjerkadera and tested in vitro with eucalyptus oxygen-deligniﬁed Kraft pulp (ODKP) based on measuring the reduction in kappa number as an indicator of lignin oxidation. The MnP preparation applied at 60 U/g pulp for 6 h caused a signiﬁcant decrease of 11–13% in the kappa number in the ODKP under optimal conditions compared to parallel-incubated controls lacking enzyme. Hatakka et al. [127] reported degradation of wheat straw with enzymes from P. chrysosporium and C. subvermispora. Enzymes from both the fungi decreased the amount of lignin and hemicellulose and increased the relative amount of cellulose. Bermek et al. [128] have reported that a combination of manganese peroxidase and xylanase has pulp-bleaching effects that are far superior to those of the individual enzymes used sequentially. The bleaching effect was a Table 2. Effect of manganese peroxidase treatment on softwood Kraft pulpsa. Enzyme treatment

Initial kappa number

Final kappa number

Methanol (mg/l)

Final brightness (% ISO)

Control Manganese peroxidase

10.8

9.7 7.9

0 8.3

35.4 38.6

Control Manganese peroxidase

12.4

10.7 9.6

0 9.0

37.1 38.1

Control Manganese peroxidase

16.2

15.0 13.1

2.5 14.2

33.0 28.1

Control Manganese peroxidase

21.2

18.5 17.3

2.3 16.1

31.2 25.3

Treatments with manganese peroxidase enzyme was run for 24 h with 1 U/ml of enzyme, manganese sulfate (0.5 mM), glucose (10 mM) and glucose oxidase (0.025 U/ml) in sodium malonate buffer (50 mM, pH 4.5). Control contained all components except manganese peroxidase. Source: Based on data from Paice et al. [45]. a Softwood pulp was deligniﬁed to the indicated initial lignin contents by oxygen deligniﬁcation or by alkaline extraction with oxygen (Eo).

360 Table 3. Effect of manganese peroxidase treatment on QP bleaching of softwood Kraft pulpa. Enzyme treatment

NaOH charge Kappa in the number peroxide after QP stage (%)

Brightness (% ISO)

Viscosity (mPa s)

H 2O 2 consumed (%)

Control Manganese peroxidase

2.5

8.5 5.2

71.8 79.0

20.7 20.9

1.29 1.60

Control Manganese peroxidase

2.0

8.1 5.6

70.7 78.4

21.4 19.5

1.19 1.49

Control Manganese peroxidase

1.5

8.3 5.7

69.4 77.3

22.8 20.6

1.04 1.34

Treatment with manganese peroxidase enzyme was run for 24 h with 1 U/ml of enzyme, manganese sulfate (0.5 mM), glucose (10 mM) and glucose oxidase (0.025 U/ml) in sodium malonate buffer (50 mM, pH 4.5). Control contained all components except manganese peroxidase. Source: Based on data from Paice et al. [46]. a Softwood pulp was used after being oxygen deligniﬁed to kappa number 15.0.

synergistic action between the enzymes rather than an addition of the brightening abilities of each individual enzyme. Bermek and Eriksson [73] have reported the effect of unsaturated fatty acids; thiol-containing compounds and various other organic compounds in pulp bleaching experiments with MnP. Thiol-containing compounds did not improve the pulp bleaching effect by MnP. Some unsaturated fatty acids, linoleic and linolenic acids provided a better pulp bleaching effect than Tween 80. A combination of Tween 80 and a carboxylic acid resulted in higher pulp brightness than that obtained with Tween 80 alone. Laccase mediators, 3-hydroxy-1, 2,3-benzotriazin-4 (3 H)-one, could also enhance the MnP pulp bleaching effect. Ducka and Pekarovicova [51] used crude ligninases from P. chrysosporium for bleaching of softwood Kraft pulp. The pulp, after oxygen bleaching with kappa number 19.7 and 36.5% MgO brightness, was pretreated with ligninases (L) or xylanases (X) and bleached in QEOPDP sequence. The brightness of pulp bleached in this sequence was 87.8%, which is about 3.2 points higher than the control and is approximately the same as with the use of commercial xylanases. Martinez et al. [129] have reported deligniﬁcation of wheat straw with enzyme from Pleurotus eryngii. Enzyme treatment did not have any adverse effect on physical strength properties of the pulp. Kantelinen et al. [130] studied the capability of lignin-modifying enzymes of Phlebia radiata to improve the bleachability of Kraft and peroxyformic acid pulps. The effect of lignin modifying enzymes alone on Kraft pulps was

361 insigniﬁcant. Bleachability of the Kraft pine pulp was found to be improved only when laccase was used after hemicellulase treatment [130]. Kadimaliev et al. [77] have studied the Lignin consumption and synthesis of ligninolytic enzymes by the fungus Panus (Lentinus) tigrinus cultivated on solid phase (modiﬁed and unmodiﬁed birch and pine sawdusts). The fungus grew better and consumed more readily the birch lignin than the pine wood. Peroxidase activity was higher in the case of pine sawdust; laccase and ligninolytic activities, in the case of birch sawdust. Treatment with ammonia or sulfuric acid decreased lignin consumption by the fungus cultivated on either medium. Modiﬁcation of sawdust by ultrasound increased lignin consumption and may be recommended for accelerating biodegradation of lignocellulose substrates. Niku-Paavola et al. [131] treated oxygen deligniﬁed pine Kraft pulps with lignin modifying enzymes – laccase, manganese-dependent peroxidases, lignin peroxidase – and xylanase in order to improve bleachability. Pulps were treated either with a puriﬁed single enzyme or alternatively enzymes were successively added in different combinations. The residual pulps were deligniﬁed with alkaline hydrogen peroxide and analyzed for kappa number and brightness. Lignin modifying enzymes did not improve the bleachability when acting alone. Xylanase treatment increased the brightness by 1–2.5 ISO units and xylanase combined with lignin modifying enzymes increased the brightness by a further one unit. By extending the alkaline peroxide step, the ﬁnal brightness increased in all samples, whereas the relative difference between reference and enzyme-treated samples remained constant [131]. Machii et al. [72] have examined the biobleaching of manganese-less oxygen-deligniﬁed hardwood Kraft pulp (E-OKP) by the whiterot fungi P. sordida YK-624 and P. chrysosporium in the solid-state fermentation system. P. sordida YK-624 possessed a higher brightening activity than P. chrysosporium, increasing pulp brightness by 13.4 points after 7 days of treatment. In these fermentation systems, lignin peroxidase (LiP) activity was detected as the principle ligninolytic enzyme, and manganese peroxidase and laccase activities were scarcely detected over the course of treatment of EOKP by either fungus. Moreover, a linear relationship between brightness increase and cumulative LiP activity was observed under all tested culture conditions with P. sordida YK-624 and P. chrysosporium. These results indicated that LiP is involved in the brightening of E-OKP by both white-rot fungi. Milagres et al. [132] studied the effect of number of stages on the treatment of hardwood Kraft pulp with xylanase (X) alone and sequentially, with xylanases (X) and laccases (L). They also studied the effect of various orders of sequential treatment with these enzyme preparations. It was noted that a multi-stage process achieved greater deligniﬁcation than a one-stage process. About 23% and 11.2% deligniﬁcation was obtained using the sequence X-X-X-L and L-L-L-X, respectively, whereas about 7.9% and 5% deligniﬁcation was obtained with the sequence X-L and L-X, respectively. Sequence order L-L-X, L-X-L did not have much effect on the lignin content

362 (12.9% deligniﬁcation). The absolute number of X stages was found to have a greater impact. Three laccases, a natural form and two recombinant forms obtained from two different expression hosts, were characterized and compared for paper pulp bleaching [75]. Laccase from P. cinnabarinus, a wellknown lignolytic fungus, was selected as a reference for this study. The corresponding recombinant laccases were produced in Aspergillus oryzae and A. niger hosts using the lacI gene from P. cinnabarinus to develop a production process without using the expensive laccase inducers required by the native source. It was observed that treatment of wheat straw Kraft pulp using laccases expressed in P. cinnabarinus or A. niger with 1-hydroxybenzotriazole as redox mediator achieved a deligniﬁcation close to 75%, whereas the recombinant laccase from A. oryzae was not able to delignify pulp. Vaheri and Piirainen [124] showed that the oxidizing enzyme laccase can be used in conjunction with manganese ions to reduce the consumption of chlorine chemic als applied in the later stages of bleaching. Viikari et al. [133] were unable to demonstrate any bleaching effects on pine Kraft pulp with ligninases and oxidases from the white-rot fungus P. radiata or puriﬁed ligninase from P. chrysosporium. When the ligninases were applied after treatment with hemicellulases, there were slight reductions in kappa number. However, these were within the accuracy limits of the analytical method. An extensive deligniﬁcation and brightening observed with the fungus were not achieved with the isolated manganese peroxidase [105] or lignin peroxidase and laccase [33,130]. These results clearly show that single enzymes are not able to mimic the complete biological system. Small improvements can be achieved by addition of low molecular weight aromatic compounds like veratryl alcohol or other substances such as ABTS and Remazol blue [117,122]. Lignozym GmbH Germany, worked with enzymes plus chemical mediators, which create a redox system throughout the pulp-treatment period [35–40]. Their idea was to ﬁnd a system, which is a good mimic of the natural situation. Starting in 1987 with the enzyme mediator concept, Lignozym has now improved the performance of the mediator system for the laccase of T. versicolor by changing and further ﬁne-tuning the chemical nature of the component. The mediator is HBT) [39–40]. The treatment of pulp with laccase alone does not result in any degradation of lignin but just in a structural change or repolymerization, the LMS causes a signiﬁcant kappa number reduction. According to the present understanding, the laccase while oxidizing the chemical mediator is generating a strongly oxidized co-mediator, which is the real bleaching agent (Fig. 5). The LMS provides a broad ﬂexibility with respect to the pulp substrate, the technical requirements for application and the ﬁnal quality of the pulp. LMS is compatible with all other bleaching sequences. The performance of the LMS has been proven in pilot-plant trials. The summary of the results is presented in Table 4. A degree of deligniﬁcation of >50% could be obtained in a single step, even if the mediator

363

Fig. 5. Possible mechanism of laccase and mediator action on lignin.

dosage is reduced by factor 0.6. The technology is now ready for large-scale application. Herpoel et al. [134] reported enzymatic deligniﬁcation of wheat straw pulp by a sequential xylanase and LM treatments. It was found that sequential treatment of wheat straw chemical pulp with xylanase and laccase followed by alkaline extraction lowered the kappa number by about 60%. The enzymatic treatment reduced the chlorine consumption and resulted in higher ﬁnal brightness. Poppius-Levlin et al. [55] have reported 60% reduction in kappa number after LM treatment and alkaline extraction. Pulp yield and viscosity were found to be high with negligible changes in hemicellulose and hexenuronic acid. Variations in treatment temperature, laccase charge and mediator charges are found to have a pronounced effect on lignin reactions; their effect on pulp carbohydrates, however, was insigniﬁcant. Sealey et al. [135] reported that with oxygen reinforced alkaline extraction; laccase-HBT biobleaching could obtain over 70% deligniﬁcation in one stage. Further treatment of pulp with another laccase HBT stage increased the deligniﬁcation to about 80%. It seems that laccase HBT biobleaching is capable of reacting with the last vestiges of residual lignin, which are typically very unreactive. Sigoillot et al. [75] have studied the ability of feruloyl esterase (from A. niger) and Mn2+ oxidizing peroxidases (from P. chrysosporium and P. eryngii) to decrease the ﬁnal lignin content of ﬂax pulp. Laccase from P. cinnabarinus

364 Table 4. Summary of results from the pilot plant trial with laccase-mediator system. Sequence

Pulp

Dosage of Degree of enzyme/ mediator deligniﬁcation (kg/TP) (%)

Max. brightness (% ISO)

L-E-Q-P L-E-L-E-Q-P L-E-Q-(P)

A A B

2/13 2 2/2 8 2/8

56.6 50.6/67.7 44.2

76.5 82.7

Parameter

L stage

E stage

Q stage

P stage

Consistency (%) Temperature (oC) pH Residence time (min) Pressure (bar) Dosage

10

10

5

10

45

60

60

75

4.5 120

11.5 60

5 30

11.2 210

— 0.2% DTPA

— 3% peroxide

Conditions:

2 — Enzyme: 2 kg/ NaOH T Mediator: variable

Source: Based on data from Call and Mucke [41].

(without mediator) also caused a slight improvement of pulp brightness that was increased in the presence of aryl-alcohol oxidase. However, the best results were obtained when the laccase treatment was performed in the presence of a mediator, 1-HBT, enabling strong deligniﬁcation of pulps. The enzymatic removal of lignin resulted in high ﬁnal brightness values that are difﬁcult to attain by chemical bleaching of this type of pulp. Lignozym has introduced a new mediator, NHA that is biodegradable and has been claimed to be cost-effective [136]. Biobleaching with NHA also allows the enzyme to maintain about 80% of its original activity after treatment for an hour, whereas biobleaching with HBT causes a severe loss of enzyme activity. In January 1995, Wacker Chemie GmbH had taken the worldwide exclusive license of the Lignozym technology for its commercialization. The next steps planned were full-scale mill trials and setting up of production facility for both the mediator and the enzyme [41]. Fu et al. [137] bleached Eucalyptus urophylla Kraft pulp with laccase in the presence of NHA and obtained 43% reduction in kappa number after alkali extraction. The addition of surfactant improved the dissolution of lignin, and hence improved the pulp brightness and also the activity of the laccase. The effectiveness of HBT and NHA in LMS has been conﬁrmed [56]. Higher levels of deligniﬁcation were achieved with HBT compared to NHA. The

365 beneﬁts of oxidative reinforced alkaline stages were demonstrated. Nuclear magnetic resonance (NMR) analysis of lignin samples demonstrated the advantages of LMS/HBT versus LMS/NHA systems. Residual lignin isolated following LMS/HBT treatment was more enriched in lignin carboxylic acid than that from LMS/NHA. Geng et al. [70] synthesized and investigated a number of hydroxamic acids as LMs for pulp bleaching. As compared with NHA, one of the most effective LMs reported so far, N-(4-cyanophenyl)acetohydroxamic acid (NCPA), resulted in the highest brightness and lowest kappa number of hardwood Kraft pulp of all the LMs studied. The bleaching efﬁcacy of a laccase/7cyano-4-hydroxy-2H-1, 4-benzoxazin-3-one system was also comparable with that of a laccase/NHA system. The effect of ABTS on laccase catalyzed oxidation of Kraft pulp was reported by Bourbonnais and Paice [34]. It was found that demethylation (diagnostic for deligniﬁcation) was enhanced and kappa number was decreased with ABTS. Bourbonnais et al. [44] compared the activities of induced laccases I and II (from T. versicolor) with and without ABTS, on hardwood and softwood Kraft pulps. It was noted that in the absence of ABTS, the two puriﬁed laccases were not able to reduce the kappa number of either pulp, but both produced small amounts of methanol with the hardwood pulp only. When ABTS was present, the two enzymes were equally effective in delignifying and demethylating either pulp. The enhancement of deligniﬁcation by the dye ABTS is not fully understood, it may be that the ABTS radical cation acts as an electron carrier between residual lignin in the Kraft pulp ﬁber wall and the large laccase molecule. Signiﬁcant differences in reactivity between various fungal laccases for pulp deligniﬁcation in the presence of mediators ABTS and HBT were found [138]. Arias et al. [76] have demonstrated that application of the laccase from Streptomyces cyaneus in the presence of ABTS to biobleaching of eucalyptus Kraft pulps resulted in a signiﬁcant decrease in the kappa number (2.3 U) and an important increase in the brightness (2.2%, as determined by the International Standard Organization test) of pulps, showing the suitability of laccases produced by Streptomycetes for industrial purposes. A comparison of T. versicolor laccase with various mediators, including ABTS, HBT, Remazol blue, nitroso-napthols and phenothiazines has shown that HBT gave the most extensive deligniﬁcation, but deactivated the enzyme and, therefore, required a higher enzyme dosage. Poppius-Levlin et al. [139] subjected three different chemical pulps, i.e., a pine Kraft pulp, a two-stage oxygen-deligniﬁed pine Kraft pulp, and birch formic acid/peroxyformic acid (MILOX) pulps to HBT- and ABTS-mediated laccase treatments. HBT was found to be more effective than ABTS in the presence of laccase in deligniﬁcation and gave higher pulp brightness. All the laccase/HBT-treated pulps showed an increased response to alkaline hydrogen peroxide bleaching, and as a consequence oxygen-deligniﬁed pulp and

366 MILOX pulp reached full brightness in one and two stages, respectively. Even the pine Kraft pulp reached a ﬁnal brightness of 83%. Moreover, a xylanase stage before or together with laccase/HBT slightly improved the effect of laccase/HBT and gave a higher ﬁnal brightness after peroxide bleaching than without the xylanase treatment. Kandioller et al. [68] used HBT, NHA, VA and potassium-octacyanomolybdate (IV) (PCM) as mediators in combination with the laccase (L) of T. versicolor at various dosages to deligniﬁy and bleach the pulps. Kappa number reductions between 5.6% and 64.3%, depending on pulp type, enzyme and mediator charge, were obtained following alkaline extraction (E). On all pulps, VA was the most efﬁcient mediator in terms of kappa number reduction. In terms of brightness gain after L-E treatment, VA was most efﬁcient on bagasse soda pulp (up to 1.0 point) and hardwood sulﬁte dissolving pulp (up to 7.0 points). HBT was the most efﬁcient mediator in terms of brightness boost on bagasse soda pulp (4.0 points) and hardwood soda-AQ pulp (1.5 points), whereas VA was most efﬁcient on hardwood sulﬁte dissolving pulp (1.1 points). Chlorine dioxide savings were achieved on all the three pulps: hardwood sulﬁte dissolving pulp (55% savings), hardwood soda-AQ pulp (50%) and bagasse soda pulp (50%). Camarero et al. [84] have compared three fungal laccases (from P. cinnabarinus, T. versicolor and P. eryngii) and two mediators, 2,20 -ABTS and 1HBT. P. cinnabarinus and T. versicolor laccases in the presence of HBT gave the best results in terms of high brightness and low lignin content (kappa number). The former laccase also resulted in the best preservation of cellulose and the largest removal of residual lignin as revealed by analytical pyrrolysis, and was selected for subsequent totally chlorine free (TCF) bleaching. Up to 90% deligniﬁcation and strong brightness increase were attained after an LM treatment followed by H2O2 bleaching. This TCF sequence was further improved by applying H2O2 under pressurized O2. Chandra et al. [140] have reported that the bleaching of high kappa Kraft pulps with an LMS provided 42.6–61.1% deligniﬁcation following an E+P stage when VA was used as the mediator. The pulp yield after the (E+P) treatment was +99.9%. A comparison of mediator efﬁciency indicated that VA was a superior reagent for LM deligniﬁcation of high lignin content pulps in comparison to N-hydroxybenzotriazole or N-acetyl-N-phenyl hydroxylamine. Molecular modeling of these three mediators indicates an elevated impaired electron density for VA over the other mediators, perhaps accounting for its improved performance. Structural analysis of the residual lignin after the laccase treatments indicated that the laccase-stage oxidizes primarily the phenolic structures of lignin. Full sequence ECF bleaching of high- and low-kappa SW Kraft pulps after an L(EoP) or L-L(EoP) indicated the pulps could be readily bleached to +85 Tappi brightness. Recently, Call et al. [67] made an evaluation of the chlorine dioxide saving potential of various enzymatic treatments. Treatments included an extended

367 hydrolase-mediated oxidation system (HOS), and an improved laccase oxidation system (LASox). Before chlorine dioxide sequences, these enzymatic treatments, attained a kappa reduction of 40%, with only a small decrease in viscosity. After the chlorine dioxide sequences, the 88.5% ISO brightness target was easily reached. Strength properties of untreated and enzymetreated pulps were equivalent. A saving of around half of the chlorine dioxide charge at good strength properties, and at a comparable cost, was seen to be possible. One disadvantage was some loss in hemicellulose content (2%), probably due to impurities in the hydrolase (lipase) formulation used. Nevertheless, this enzyme-based approach seems to have high potential for deligniﬁcation and chlorine dioxide saving. In Paprican, the use of transition metal complexes as catalytic mediators in the enzymatic deligniﬁcation and bleaching of Kraft pulps has been investigated [65,141]. An oxygen-deligniﬁed softwood Kraft pulp was treated with laccase in the presence of a transition metal complex potassium octacynomolybdate. At all charges of mediator, pulp deligniﬁcation exceeded that of the control pulps. Pulp viscosity did however suffer a slight loss at the highest dosage of mediator. Treatment of hardwood Kraft pulp at the same reaction conditions produced similar results. The molybdenum mediator can be recycled after pulp deligniﬁcation and reused with the same efﬁciency as a fresh solution of mediator. LMS is also found to be effective in removing hexenuronic acid from Kraft pulp [142]. Jose et al. [82] have applied a series of manganese substituted polyoxotungstates, [SiW11MnIII(H2O)O39]5– and [PW11MnIII(H2O)O39]4–, and [SiW11VVO 5– as catalysts for the oxygen deligniﬁcation of unbleached eucalyptus 40] Kraft pulp with laccase of T. versicolor. Unlike the modest results obtained in the LMS at 45–601C (lignin oxidation and catalyst re-oxidation occurred at the same stage), a sustainable deligniﬁcation with removal of about 50% of residual lignin was achieved with [SiW11MnIII(H2O)O39]5– and [SiW11VVO40]5– when the Kraft pulp treatment was carried out with polyoxometalate at 1101C (lignin oxidation stage) and with laccase at 451C (catalyst re-oxidation stage) in separate stages. The use of [PW11MnIII(H2O)O39]4– in this multi-stage process was limited by the low re-oxidation rate with laccase. The best selectivity on the pulp deligniﬁcation was found with polyoxoanion [SiW11MnIII (H2O)O39]5–, whereas [SiW11VVO40]5– was the most effective in the oxidative deligniﬁcation. The inﬂuence of process factors on the poly oxo metallates (POMs) re-oxidation, such as the amount of laccase, oxygen pressure and temperature was studied. UV–Vis and 51V NMR studies indicated that poly oxo metallates (POMs) maintained stable after redox turnovers during the pulp deligniﬁcation. Ehara et al. [143] investigated the role of Tween 80 in the biobleaching of unbleached hardwood Kraft pulp with manganese peroxidase. The effect of the surfactant on the brightness of the pulp and the manganese peroxidase activity was compared with the effect of two other surfactants (Tween 20).

368 Tween 80 and Tween 20 both dispersed degraded lignin and activated manganese peroxidase, although these effects did not explain the signiﬁcantly higher brightness increase achieved with Tween 80. Vares et al. [144] studied deligniﬁcation of semi-chemical wheat straw pulp with LMS and manganese peroxidase in combination with Tween 80. Differences between the enzymatically treated pulp and untreated control were rather small. However, some favourable or prominent effects were caused by Tween 80 alone or in combination with manganese peroxidase and by laccase-HBT treatment followed by a chelating step with ethylene diamine tetra-acetic acid. Ana et al. [83] have studied the catalytic oxygen bleaching of Kraft pulp using the heteropolyanion [SiW11MnIII(H2O)O39]5– (SiW11MnIII) and laccase of T. versicolor. The oxidation of the residual lignin in pulp with the heteropolyanion was followed by the catalyst re-oxidation by laccase in a separate stage. The alternative treatment in a multi-stage process with SiW11MnIII at 1101C and with laccase at 451C allowed a sustainable eucalyptus Kraft pulp deligniﬁcation with removal of about 50% of the residual lignin with minimal polysaccharides damage. Effect of lignin-oxidizing enzymes on pulp properties, yield and efﬂuent quality Lignin-oxidizing enzymes due to the speciﬁc action on lignin do not affect the viscosity, strength properties and yield of the pulp. Efﬂuent properties namely AOX, color, etc., are also expected to improve due to reduced chlorine consumption required to achieve a given brightness. Bourbonnais and Paice [34] have reported that the treatment of Kraft pulp with laccase and ABTS did not affect the pulp viscosity and zero span breaking length. In a pilot-plant trial with LMS, the strength properties and viscosity of the pulp with L-E-Q-P sequence were not found to be affected [39–40]. Three enzyme preparations (crude cellulase, laccase and proteinase) were evaluated for their potential to improve the papermaking properties of mechanical pulp [81]. After treating a long ﬁber-rich fraction of the pulp with enzyme, the ﬁbers were recombined with untreated ﬁnes for handsheet making and testing. None of the enzymes altered the retention of ﬁnes or the consolidation of the furnish mix during handsheet formation. All three enzymes increased tensile stiffness index, which is a measure of the initial resistance of the handsheets to strain. Only the laccase preparation, an enzyme that modiﬁes pulp lignin, consistently increased ﬁber-bonding to enhance other strength properties of the handsheets. Bourbonnais and Paice [52] reported the effect of laccase-ABTS treatment on viscosity and strength properties of an oxygen deligniﬁed softwood Kraft pulp. However, the tear index was found to be decreased by enzyme-mediator action. Egan [126] reported that pulp viscosity did not change after treatment of pulp with Ligninase 118. Camarero et al. [145] have reported that when highquality ﬂax pulp was bleached in a TCF sequence using an LMS, the pulp properties obtained by using this system, could not be attained by conventional

369 TCF bleaching of ﬂax pulp. Arbeloa et al. [33] reported that when softwood TMP was treated with lignin peroxidase enzyme, some of the strength properties, viz. burst index, tear index and breaking length slightly increased. Kondo et al. [146] have reported that manganese peroxide treatment is quite effective for improving the strength properties of softwood and hardwood bleached Kraft pulps and deinked pulp. In L-E-L-E sequence, the total viscosity loss was 19% compared with the initial pulp. This viscosity loss was found to be comparable with that found following alkaline extraction alone. The viscosity of laccase-ABTS-treated sulﬁte pulp following the hot alkaline extraction was higher as compared to that of untreated pulp because of hemicellulose removal. Herpoel et al. [134] reported higher tear strength when wheat straw chemical pulp was treated with xylanase and laccase followed by alkaline extraction. Not much information is available on the impact of ligninoxidizing enzymes on efﬂuent quality and AOX discharges. Vaheri and Mikki [123] reported that treatment of pulp with laccase produced an efﬂuent with lower content of chloroorganics. Advantages Lignin-oxidizing enzymes are highly speciﬁc toward lignin; there is no damage or loss of cellulose. This results in better strength and yield of bleached pulp. As compared to oxygen deligniﬁcation, treatment with lignin-oxidizing enzymes results in more removal of lignin. This translates into substantial savings of energy and bleaching chemicals, which in turn would lead to lower pollution load. Limitations and future prospects Lignin-oxidizing enzymes are not currently available in sufﬁcient quantity for mill trials and scale-up of enzyme production from fungal cultures is costly. Cloning of genes for lignin-oxidizing enzymes has been reported and may provide an alternative production route. The LMs are expensive and alternatives are less effective. There are two ways of improving LMS for pulp bleaching. One is to discover new laccases that have extraordinarily high redox potential, the other is to ﬁnd a very effective LM. Since the LMs reported so far have very wide structural variations, it should be possible to discover inexpensive and more effective LMs than what has been done to date. Chelating agents for Mn3+ in the manganese peroxidase reaction may also be a signiﬁcant cost item. Manganese peroxidase requires hydrogen peroxide but are inactivated by concentrations above about 0.1 mmol/l. Even when hydrogen peroxide is kept below 0.1 mmol/l, manganese peroxidase becomes inactive relatively rapidly. Experience with xylanase and with other enzymes has shown that enzymes can be successfully introduced in the plant. Thus, oxidative enzymes that can be regarded as catalysts for oxygen- and

370 hydrogen peroxide-driven deligniﬁcation, may also ﬁnd a place in the bleach plant in coming years. References 1. 2. 3.

4. 5.

6.

7.

8.

9. 10.

11. 12. 13.

14. 15. 16. 17.

Bajpai P and Bajpai PK. In: Advanced Biochemical Engineering/Biotechnology, Vol. 56, Schepper T (ed), Berlin, Springer, 1996, p. 1. Viikari L, Kantelinen A, Sundquist J and Linko M. Xylanases in bleaching: from an idea to industry. FEMS Microb Rev 1994;13:335–350. Bajpai P. Microbial xylanolytic enzyme system. In: Advances in Applied Microbiology, Vol. 43, Nidelman S and Laskin A (eds), New York, Academic Press, 1997, pp. 141–194. Sudha B, Veeramani H and Sumathi S. Bleaching of bagasse pulp with enzyme pretreatment. Water Sci Technol 2003;47(10):163–168. Roncero MB, Torres AL, Colom JF and Vidal T. TCF bleaching of wheat straw pulp using ozone and xylanase. Part B: kinetic studies. Bioresource Technol 2003;87(3):315–323. Roncero MB, Torres AL, Colom JF and Vidal T. The effect of xylanase on lignocellulosic components during the bleaching of wood pulps. Bioresource Technol 2005;96(1):21–30. Kansoh AL and Nagieb ZA. Xylanase and mannanase enzymes from Streptomyces galbus NR and their use in biobleaching of softwood kraft pulp. Antonie Van Leeuwenhoek 2004;85(2):103–114. Tolan JS, Olson D and Dines RE. Survey of mill usage of xylanase. In: Enzymes for Pulp and Paper Processing, Jeffries TW and Viikari L (eds), ACS Symposium Series, Vol. 655, American Chemical Society, Washington, DC, 1996, pp. 25–35. Turner JC, Skerker PS, Burns BJ, Howard JC, Alonso MA and Andres JL. Bleaching with enzymes instead of chlorine: mill trials. Tappi J 1992;75(12):83–89. Jean P, Hamilton J and Senior DJ. Mill trial experiences with xylanase: AOX and chemical reductions. In: Preprints of the 80th Annual Meeting Technical Section, 1–2 February, Montreal, Canada, 1994, pp. A229–A233. Scott BP, Young F and Paice MG. Mill scale enzyme treatment of a softwood Kraft pulp prior to bleaching. Pulp Paper Can 1993;94(3):57–61. Yee Y and Tolan JS. Three years experience running enzymes continuously to enhance bleaching at Weyerhaeuser Prince Albert. Pulp Paper Can 1997;98(10):T370–T375. Paice M, Renaud S, Bourbonnais R, Labonte S and Berry R. The effect of xylanase on Kraft pulp bleaching yield. In: 9th International Conference on Biotechnology in Pulp and Paper Industry, 10–14 October, Durban, South Africa, 2004. Kirk TK and Yang HH. Partial deligniﬁcation of unbleached kraft pulp with ligninolytic fungi. Biotechnol Lett 1979;1:347–352. Pellinen J, Abuhasan J, Joyce TW and Chang HM. Biological deligniﬁcation of pulp by Phanerochaete chrysosporium. J Biotechnol 1989;10:161–170. Tran AV and Chambers RP. Deligniﬁcation of an unbleached hardwood pulp by Phanerochaete chrysosporium. Appl Microbiol Biotechnol 1987;25:484–490. Paice MG, Jurasek L, Ho C, Bourbonnais R and Archibald FS. Direct biological bleaching of hardwood kraft pulp with the fungus Coriolus versicolor. Tappi J 1989;72(5):217–221.

371 18. Reid ID, Paice MG, Ho C and Jurasek L. Biological bleaching of softwood kraft pulp with the fungus Trametes versicolor. Tappi J 1990;73(8):149–153. 19. Reid ID and Paice MG. Biological bleaching of Kraft pulps by white-rot fungi and their enzymes. FEMS Microbiol Rev 1994;13:369–376. 20. Reid ID and Paice MG. Effect of residual lignin type and amount on bleaching of Kraft pulp by Trametes versicolor. Appl Environ Microbiol 1994;60(5):1395–1400. 21. Addleman K and Archibald FS. Kraft pulp bleaching and deligniﬁcation by dikaryons and monokaryons of Trametes versicolor. Appl Environ Microbiol 1993;59:266–273. 22. Fujita K, Kondo R and Sakai K. Biobleaching of softwood kraft pulp with white-rot fungus IZU-154. Tappi J 1993;76(1):81–84. 23. Nishida T, Katagiri N and Tsutsumi Y. New analysis of lignin-degrading enzymes related to biobleaching of kraft pulp by white-rot fungi. In: 6th International Conference on Biotechnology in Pulp and Paper Industry, Vienna, Austria, 1995. 24. Tsuchikawa K, Kondo R and Sakai K. Application of ligninolytic enzymes to bleaching of kraft pulp II: totally chlorine-free bleaching process with the introduction of enzyme treatment with crude enzymes secreted from Phanerochaete sordida YK-624. Jpn Tappi J 1995;49:1332–1337. 25. Ishimura D, Kondo R, Sakai K and Hirai H. Biobleaching of kraft pulp with mutants from white-rot fungus Phanerochaete sordida YK-624. In: Proceedings on the 7th International Conference on Biotechnology in Pulp and Paper Industry, Vol. B, Vancouver, BC, Canada, 1998, p. B237. 26. Geng X and Li K. Degradation of non-phenolic lignin by the white-rot fungus Pycnoporus cinnabarinus. Appl Microbiol Biotechnol 2002;60(3):342–346. 27. Kluczek-Turpeinen B, Tuomela M, Hatakka A and Hofrichter M. Lignin degradation in a compost environment by the deuteromycete Paecilomyces inﬂatus. Appl Microbiol Biotechnol 2003;61(4):374–379. 28. Fujita K, Kondo R, Sakai K, Kashino Y, Nishida T and Takahara Y. Biobleaching of kraft pulp using white-rot fungus IZU-154. Tappi J 1991;74(11):123–127. 29. Kirkpatrick N, Reid ID, Ziomek E, Ho C and Paice MG. Relationship between fungal biomass production and the brightening of hardwood kraft pulp by Coriolus versicolor. Appl Environ Microbiol 1989;55:1147–1152. 30. Kirkpatrick N, Reid ID, Ziomek E and Paice MG. Biological bleaching of hardwood kraft pulp using Trametes versicolor immobilized in polyurethane foam. Appl Environ Microbiol 1990;33:105–108. 31. Kirkpatrick N, Ziomek E and Reid ID. Effect of increased oxygen availability on the biological bleaching of hardwood kraft pulp by Coriolus versicolor. In: Applications of Biotechnology in Pulp and Paper Manufacture, Kirk TK and Chang HM (eds), MA, Butterworth-Heinemann, 1990, pp. 137–143. 32. Kirkpatrick N, Reid ID, Ziomek E and Paice MG. Physiology of hardwood kraft pulp bleaching by Coriolus versicolor and use of foam immobilization for the production of mycelium-free bleached pulps. In: Biotechnology in Pulp and Paper Manufacture, Kirk TK and Chang HM (eds), MA, Butterworth-Heinemann, 1990, pp. 131–136. 33. Arbeloa M, De Leseleuc J, Goma G and Pommier JC. An evaluation of the potential of lignin peroxidases to improve pulps. Tappi J 1992;75(3):215–221. 34. Bourbonnais R and Paice MG. Demethylation and deligniﬁcation of Kraft pulp by Trametes versicolor laccase in the presence of 2,20 -azinobis-3-ethylbenzthiazoline-6-sulphonate. Appl Microbiol Biotechnol 1992;36:823–827. 35. Call HP. Multi Component Bleaching System WO 94/29425, 1994.

372 36. Call HP. Process for Modifying, Breaking Down or Bleaching Lignin, Materials Containing Lignin or Like Substances (PCT) World Patent Application WO 94/29510, 1994. 37. Call HP and Mu¨cke I. Enzymatic bleaching of pulps with the laccase-mediator-system In: Pulping Conference AlChE session, San Diego, CA, USA, 1994, pp. 38–52. 38. Call HP and Mu¨cke I. Further improvements of the laccase-mediator-system (LMS) for enzymatic deligniﬁcation and results from large scale trials, In: International NonChlorine Bleaching Conference, 5–9 March, Amelia Island, FL, USA, 1995, p. 16. 39. Call HP and Mu¨cke I. The laccase-mediator-system (LMS). Biotechnology in the pulp and paper industry: recent advances in applied and fundamental research. In: Proceedings of the 6th International Conference on Biotechnology in Pulp and Paper Industry, Srebotnik E and Messner K (eds), Vienna, Austria, 1995, pp. 27–32. 40. Call HP and Mu¨cke I. State of the art of enzyme bleaching and disclosure of a breakthrough process. In: International Non-Chlorine Bleaching Conference, Amelia Island, Fl, USA, 1994. 41. Call HP and Mu¨cke I. History, overview and applications of mediated ligninolytic systems, especially laccase-mediator-systems (lignozym-process). J Biotechnol 1997;53:163–202. 42. Call HP. New enzymatically mediated deligniﬁcation and bleaching systems. In: 10th International Symposium on wood and pulping chemistry 10th Biennial ISWPC Main Symposium, Vol. 1, 7–10 June, Yokohama, Japan, 1999, p. 540. 43. Call HP. Process for producing cellulose from lignin containing raw materials using an enzyme or microorganism while monitoring and maintaining the redox potential. U.S. Pat 1993;5:203–964. 44. Bourbonnais R, Paice MG, Reid ID, Lanthier P and Yaguchi M. Lignin oxidation by laccase isozymes from Trametes versicolor and role of the mediator 2, 20 -azinobis (3ethylbenzothiazoline-6-sulfonate) in Kraft lignin depolymerization. Appl Environ Microbiol 1995;61(5):1876–1880. 45. Paice MG, Bourbonnais R and Reid ID. Bleaching Kraft pulps with oxidative enzymes and alkaline hydrogen peroxide. Tappi J 1995;78(9):161–170. 46. Paice MG, Bourbonnais R, Reid ID, Archibald FS and Jurasek L. Oxidative bleaching enzymes. J Pulp Pap Sci 1995;21:J280–J284. 47. Kondo R, Harazono K and Sakai K. Bleaching of hardwood Kraft pulp with manganese peroxidase secreted from Phanerochaete chrysosporium. Appl Environ Microbiol 1994;60:4359–4363. 48. Kondo R, Kurashiki K and Sakai K. In vitro bleaching of hardwood Kraft pulp by extra cellular enzymes secreted from white-rot fungi in a cultivation system using a membrane ﬁlter. Appl Environ Microbiol 1994;60:921–926. 49. Harazono K, Kondo R and Sakai K. Bleaching of hardwood Kraft pulp with manganese peroxidase from Phanerochaete sordida YK-624 without addition of MnSO4. Appl Environ Microbiol 1996;62(3):913–917. 50. Kondo R, Harazono K, Tsuchikawa K and Sakai K. Biological bleaching of Kraft pulp with lignin-degrading enzymes. In: Enzymes in Pulp and Paper Processing, Jeffries TW and Viikari L (eds), ACS Symposium Series, Vol. 655, American Chemical Society, Washington, DC, 1996, pp. 228–240. 51. Ducka I and Pekarovicova A. Ligninases in bleaching of softwood Kraft pulp. In: 6th International Conference on Biotechnology in Pulp and Paper Industry, 11–15 June, Vienna, Austria, 1995.

373 52. Bourbonnais R and Paice MG. Enzymatic deligniﬁcation of Kraft pulp using laccase and a mediator. Tappi J 1996;76(6):199–204. 53. Kondo R, Hirai H, Harazono K and Sakai K. Biobleaching of Kraft pulp with lignindegrading fungi and their enzymes. In: Proceedings of the 6th International Conference on Biotechnology in Pulp and Paper Industry and Recent Advances in Applied and Fundamental Research, Vienna, Austria, 1995, p. 33. 54. Nelson PJ, Chin CWJ, Viikari L and Tenkanen M. The use of a laccase-mediator stage in bleaching of eucalypt Kraft pulps. In: International Pulp Bleaching Conference Book 2, 1–5 June, Helsinki, Finland, 1998, p. 349. 55. Poppius-Levlin K, Tamminen T, Rajanen K, Jaaskelainen AS and Viikari L. Suitability of laccase-mediator systems for hardwood Kraft pulp deligniﬁcation. In: Proceedings on the 10th International Symposium on Wood and Pulping Chemistry – 10th Biennial ISWPC – Main Symposium, Vol. 1, 7–10 June, Yokohama, Japan, 1999, p. 556. 56. Chakar FS and Ragauskas AJ. The effects of oxidative alkaline extraction stages after laccase HBT and laccase NHAA treatments – an NMR study of residual lignins. J Wood Chem Technol 2000;20(2):169–184. 57. Li K, Eriksson, K-EL and Bermek H. A new technique for screening of laccase-mediators. In: Biological Science Symposium, 19–23 October, San Francisco, CA, USA, 1997, pp. 349–353. 58. Balakshin M, Chen CL, Gratzl JS, Kirkman AG and Jakob H. Biobleaching of pulp with dioxygen in the laccase-mediator system, Part I.: kinetics of deligniﬁcation. Holzforschung 2000;54(4):390–396. 59. Niku-Paavola ML, Tamminen T, Hortling B, Viikari L and Poppius-Levlin K. Reactions of laccase towards high and low molecular mass lignin. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 56. 60. Kruus K and Viikari L. Comparison of various mediators for laccases. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 117. 61. Kandioller G and Christov L. Efﬁciency of Trametes versicolor laccase-mediator systems in pulp deligniﬁcation and bleaching. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 223. 62. Surma-Slusarska B and Leks-Stepien J. The effect of Cerrena unicolor and Trametes versicolor laccases on the TCF bleaching of Birch and Pine Kraft pulps. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 228. 63. Moreira MT, Feijoo G, Maceiras A and Lema JM. Inﬂuence of operational parameters on the in vitro deligniﬁcation of eucalyptus Kraft pulp by manganese peroxidase. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 230. 64. Moreira MT, Sierra-Alvarez R, Lema JM, Feijoo G and Field JA. Oxidation of lignin in eucalyptus kraft pulp by manganese peroxidase from Bjerkandera sp. strain BOS55. Bioresource Technol 2001;78(1):71–79. 65. Paice M, Bourbonnais R, Renaud S, Amann M, Candussio A, Rochefort D, Leech D, Labonte S and Sacciadis G. Laccase/mediator catalyzed deligniﬁcation: trials with new mediators. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 48.

374 66. Eriksson K-EL and Li K. The potential of laccase/mediator systems in bleaching of wood pulp. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 53. 67. Call H-P, Call S, Garcia-Lindgren C and Marklund A. Extended lab trials: combined enzymatic deligniﬁcation and bleaching systems. In: 9th International Conference on Biotechnology in Pulp and Paper Industry, 10–14 October, Durban, South Africa, 2004. 68. Kandioller G and Christopher L. Deligniﬁcation efﬁciency of Trametes versicolor laccase in combination with various mediators. In: 9th International Conference on Biotechnology in Pulp and Paper Industry, 10–14 October, Durban, South Africa, 2004. 69. Poppius-Levlin K and Tamminen T. Potential and limitations of enzymes in pulp bleaching. In: 9th International Conference on Biotechnology in Pulp and Paper Industry, 10–14 October, Durban, South Africa, 2004. 70. Geng X, Li K and Xu F. Investigation of hydroxamic acids as laccase-mediators for pulp bleaching. Appl Microbiol Biotechnol 2004;64(4):493–496. 71. Sigoillot C, Record E, Belle V, Robert JL, Levasseur A, Punt PJ, Van Den Hondel CA, Fournel A, Sigoillot JC and Asther M. Natural and recombinant fungal laccases for paper pulp bleaching. Appl Microbiol Biotechnol 2004;64(3):346–352. 72. Machii Y, Hirai H and Nishida T. Lignin peroxidase is involved in the biobleaching of manganese-less oxygen-deligniﬁed hardwood kraft pulp by white-rot fungi in the solidfermentation system. FEMS Microbiol Lett 2004;233(2):283–287. 73. Bermek H, Li K and Eriksson KE. Studies on mediators of manganese peroxidase for bleaching of wood pulps. Bioresource Technol 2002;85(3):249–252. 74. Wong KK, Richardson JD and Mansﬁeld SD. Enzymatic treatment of mechanical pulp ﬁbers for improving papermaking properties. Biotechnol Prog 2000;16(6):1025–1029. 75. Sigoillot C, Camarero S, Vidal T, Record E, Asther M, Perez-Boada M, Martinez MJ, Sigoillot JC, Asther M, Colom JF and Martinez AT. Comparison of different fungal enzymes for bleaching high-quality paper pulps. J Biotechnol 2005;115(4):333–343. 76. Arias ME, Arenas M, Rodriguez J, Soliveri J, Ball AS and Hernandez M. Kraft pulp biobleaching and mediated oxidation of a nonphenolic substrate by laccase from Streptomyces cyaneus CECT 3335. Appl Environ Microbiol 2003;69(4):1953–1958. 77. Kadimaliev DA, Revin VV, Atykian NA and Samuilov VD. Effect of wood modiﬁcation on lignin consumption and synthesis of lignolytic enzymes by the fungus Panus (Lentinus) tigrinus. Prikl Biokhim Mikrobiol 2003;39(5):555–560. 78. Srebotnik E and Hammel KE. Degradation of nonphenolic lignin by the laccase/1hydroxybenzotriazole system. J Biotechnol 2000;81(2):179–188. 79. Xu F, Kulys JJ, Duke K, Li K, Krikstopaitis K, Deussen HJ, Abbate E, Galinyte V and Schneider P. Redox chemistry in laccase-catalyzed oxidation of N-hydroxy compounds. Appl Environ Microbiol 2000;66(5):2052–2056. 80. Crestini C, Jurasek L and Argyropoulos DS. On the mechanism of the laccase-mediator system in the oxidation of lignin. Chemistry 2003;9(21):5371–5378. 81. Jurasek L, Archibald FS, Bourbonnais R, Paice MG and Reed ID. Prospects for redox enzymes to enhance Kraft pulp bleaching. In: Proceedings on Biological Sciences Symposium, 3–6 October, Minneapolis, MN, 1994, p. 239. 82. Jose AFG, Ana PMT, Dmitr YVE and Ana MBX. Oxygen bleaching of kraft pulp with polyoxometalates and laccase applying a novel multi-stage process. J Mol Catal B: Enzym 2005;33(3–6):57–64. 83. Ana PMT, Jose AFG, Armindo RG, Dmitr YVE and Ana MRBX. A novel approach for the oxidative catalysis employing polyoxometalate–laccase system: application to

375

84.

85.

86. 87.

88.

89.

90.

91. 92. 93.

94.

95. 96. 97. 98.

99. 100.

the oxygen bleaching of kraft pulp. Short Commun Catal Commun 2004;5(9): 485–489. Camarero S, Garcı´ a O, Vidal T, Colom J, Del Rı´ o JC, Gutie´rrez A, Gras JM, Monje R, Martı´ nez MJ and Martı´ nez AT. Efﬁcient bleaching of non-wood high-quality paper pulp using laccase-mediator system. Enzyme and Microb Technol 2004;35(2–3):113–120. Balakshin M, Chen CL, Gratzl JS, Kirkman AG and Jakob H. Biobleaching of pulp with dioxygen in laccase-mediator system – effect of variables on the reaction kinetics. J Mol Catal B: Enzym 2001;16(3–4):205–215. Palonen H and Viikari L. Role of oxidative enzymatic treatments on enzymatic hydrolysis of softwood. Biotechnol Bioeng 2004;86(5):550–557. Leonowicz A, Cho NS, Luterek J, Wilkolazka A, Wojtas-Wasilewsk AM, Matuszewska A, Hofrichter M, Wesenberg D and Rogalski J. Fungal laccase: properties and activity on lignin. J Basic Microbiol 2001;41(3–4):185–227. Ramos J, Rojas T, Navarro F, Davalos F, Sanjuan R, Rutiaga J and Young RA. Enzymatic and fungal treatments on sugarcane bagasse for the production of mechanical pulps. J Agr Food Chem 2004;52(16):5057–5062. Gronqvist S, Viikari L, Niku-Paavola ML, Orlandi M, Canevali C and Buchert J. Oxidation of milled wood lignin with laccase, tyrosinase and horseradish peroxidase. Appl Microbiol Biotechnol 2005;67(4):489–494. Li K, Horanyi PS, Collins R, Phillips RS and Eriksson KL. Investigation of the role of 3-hydroxyanthranilic acid in the degradation of lignin by white-rot fungus Pycnoporus cinnabarinus. Enzyme Microb Technol 2001;28(4–5):301–307. Edwards SL, Raag R, Wariishi H, Gold MH and Poulos TL. Crystal structure of lignin peroxidase. Proc Natl Acad Sci 1993;90. Pulos TL, Ewards SL, Wriishi H and Gld MH. Crystallographic reﬁnement of lignin peroxidase at 2 A˚. J Biol Chem 1993;268(6):4429–4434. Sundaramoorthy S, Kishi K, Gold MH and Poulos TL. Preliminary crystallographic analysis of manganese peroxidase from Phanerochaete chrysosporium. J Mol Biol 1994;238(5):845–856. Barr DP, Shah MM, Grover TA and Aust SD. Production of hydroxyl radical by lignin peroxidase from Phanerochaete chrysosporium. Arch Biochem Biophys 1992;298:480–485. Hammel KE and Moen MA. Depolymerization of a synthetic lignin in vitro by lignin peroxidase. Enzyme Microb Technol 1991;13:15–18. Higuchi T. Biodegradation mechanism of lignin by white-rot basidiomycetes. J Biotechnol 1993;30(1):1–11. Cai D and Tein M. On the reactions of lignin peroxidase compounds III (isozyme H8. Biochem Biophys Res Commun 1989;162:464–470. Cai D and Tein M. Kinetic studies on the formation and decomposition of compound II and III. Reactions of lignin peroxidase with hydrogen peroxide. J Biol Chem 1992;267:11149–11155. Higuchi T. Lignin biochemistry: biosynthesis and biodegradation. Wood Sci Technol 1990;24:23–63. Higuchi T Mechanism of lignin degradation by lignin peroxidase and laccase of white-rot fungi. In: Biogenesis and biodegradation of plant cell polymers, Lenis NG and Paice MG (eds), ACS Symposium No. 399, American Chemical Society, 1989, pp. 482–502.

376 101. Umezawa T and Higuchi T. Cleavage of aromatic ring and b-4-O-bond of synthetic lignin (DHP) by lignin peroxidase. FEBS Lett 1989;242:325–330. 102. Harvey PJ, Gilardi G- F, Goble ML and Palmer JM. Charge transfer reactions and feed back control of lignin peroxidase by phenolic compounds: signiﬁcance in lignin degradation. J Biotechnol 1993;30(1):57–63. 103. Harvey PJ and Palmer JM. Oxidation of phenolic compounds by ligninase. J Biotechnol 1990;13:169–175. 104. Archibald FS. Lignin peroxidase is not important in biological bleaching and deligniﬁcation of Kraft brown stock by Trametes verisicolor. Appl Environ Microbiol 1992;58:3101–3109. 105. Paice MG, Reid ID, Bourbonnais R, Archibald FS and Jurasek L. Manganese peroxidase produced by Trametes versicolor during pulp bleaching, demethylates and deligniﬁes Kraft pulp. Appl Environ Microbiol 1993;59:260–265. 106. Gold MH, Akileswaran L, Wariishi H, Mino Y and Loehr TM. Spectral characterization of manganese-peroxidase an extra cellular heme enzyme from Phanerochaete chrysosporium. In: Lignin Enzymatic and Microbial Degradation Symposium Intern INRA, Odier E (ed), Paris, 1987, p. 113. 107. Wariishi H, Valli K and Gold MH. Manganese(II) oxidation by lignin peroxidase from the basidiomycete Phanerochaete chrysosporium. Kinetic mechanism and role of chelators. J Biol Chem 1992;267:23688–23699. 108. Wariishi H, Valli K and Gold MH. In vitro depolymerization of lignin by manganese peroxidase of Phanerochaete chrysosporium. Biochem Biophys Res Commun 1991;176:269–275. 109. Gogld MH, Wariishi H and Walli K. Extra cellular peroxidases involved in lignin degradation by the white-rot basidiomycete Phanerochaete chrysosporium. ACS Symp Ser 1989;389:127. 110. Eggert C, Temp V and Eriksson K- EL. The ligninolytic system of the white-rot fungus Pycnoporus cinnabarinus: Puriﬁcation and characterization of the laccase. Appl Environ Microbiol 1996;62:1151–1158. 111. Garzillo AMV, Dipaolo S, Burla G and Buonocore V. Differently induced extra cellular phenol oxidases from Pleurotus ostreatus. Phytochemistry 1992;31:3685–3690. 112. Vasdev K and Kuhad RC. Decolourization of poly R-478 (polyvinyl amine sulfonate anthrapyridone) by Cyathns bulleri. Folia Microbiol 1994;39(1):61–70. 113. White NA and Body L. Differential extra cellular enzyme production in colonies of Coriolus versicolor, Phlebia radiata and Phlebia rufa: effect of gaseous regime. J Gen Microbiol 1992;138(12):2589–2595. 114. Reinhammer B. Laccase. In: Copper Proteins and Copper Enzymes, Lontie R (ed), Boca Raton, CRC, 1984, p. 1. 115. Sariaslani FS. Microbial enzymes for oxidation of organic molecules. Crit Rev Biotechnol 1989;9:171–257. 116. Kawai S, Umezawa T, Shimada M and Higuchi T. Aromatic ring cleavage of 4,6di(tert-butyl) guaiacol, phenolic lignin model compound by laccase of Coriolus versicolor. FEBS Lett 1988;236:309–311. 117. Bourbonnais R and Paice MG. Oxidation of non-phenolic substrates. An expanded role for laccase in lignin biodegradation. FEBS Lett 1990;267:99–102. 118. Bao WL, Fukushima Y, Jensen KA and Moen MA. Oxidative degradation of nonphenolic lignin during lipid peroxidation by fungal manganese peroxidase. FEBS Lett 1994;354:297–300.

377 119. Farrell R. Use of rldmtm 1-6 and other Ligninolytic Enzymes (PCT) International Appl. WO 87/00, 1987, p. 564. 120. Farrell RL, Gelep P, Anillouis A, Javaherian K, Malone TE, Rusche JR, Sadownick BA and Jackson JA. Sequencing and Expression of Ligninase cDNA of Phanerochaete chrysosporium, EP 0216080, 1987. 121. Farrell RL, Kirk TK and Tien M. Novel Enzymes for Degradation of Lignin, WO 87/ 00550, 1987. 122. Olsen WL, Slocomb JP, Gallagher HP and Kathleen BA. Enzymatic Deligniﬁcation of Lignocellulosic Material, EP 0,345,715 A1, 1989. 123. Vaheri M and Miiki K. Redox Enzyme Treatment in Multistage Bleaching of Pulp. EP 0,408,803 A1, 1991. 124. Vaheri M and Piirainen O. Bleaching of Pulp in Presence of Oxidizing Enzyme and Transition Metal Compound, WO 92/09741, 1992. 125. Gysin B and Griessmann T. Bleaching Wood Pulp with Enzymes, EP 0418201 A2, 1991. 126. Egan M. Proceedings of the 2nd Annual Pulp and Paper Chemical Outlook Conf, 12–13 November, Montreal, Corpus Information Services Ltd. Montreal, 1985. 127. Hatakka AI, Bocchini P, Vares T and Galletti GC. Production of lignin-degrading enzymes on solid straw medium by Phanerochaete chrysosporium and Ceriporiopsis subvermispora and degradation of the lignocellulosic substrate, 1997 Biological Sciences Symposium, 19–23 October, San Francisco, CA, USA, 1997, pp. 19–23. 128. Bermek H, Li K and Eriksson K- EL. Pulp bleaching with manganese peroxidase and xylanase: a synergistic effect. Tappi J 2000;83(10):69. 129. Martinez AT, Camarero S, Ruiz-Duenas FJ, Heinﬂing A and Martinez MJ. Studies on microbial and enzymatic applications in paper pulp manufacturing from non-woody plants based on white-rot fungi from the genus Pleurotus. In: 2000 Pulping Process and Product Quality Conference, 5–8 November, Boston, MA, USA, 2000, pp. 10pp. 130. Kantelinen A, Hortling BO, Ranua M and Viikari L. Effects of fungal and enzymatic treatments on isolated lignins and pulp bleachability. Holzforschung 1993;47:29–35. 131. Niku-Paavola ML, Ranua M, Suurnakki A and Kantelinen A. Effects of lignin-modifying enzymes on pine Kraft pulp. Bioresource Technol 1994;50:73–77. 132. Milagres AMF, Medeiros MB and Borges LA. Sequential treatment of eucalyptus Kraft pulp with Penicillium janthinellium xylanase and Pleurotus ostreatus laccase. In: 6th International Conference on Biotechnology in Pulp and Paper Industry, 11–15 June, Vienna, Austria, 1995. 133. Viikari L, Tenkanen M, Buchert J, Ratto M, Bailey M, Siika-aho M and Linko M. Hemicellulases for industrial applications. In: Bioconversion of Forest and Agricultural Plant Residues, Saddler JN (ed), Wallingford, CAB International, 1993, pp. 131–182. 134. Herpoel I, Jeller H, Fang G, Petit-Conil M, Bourbonnair R, Robert J- L, Asther M and Sigoillot J- C. Efﬁcient enzymatic deligniﬁcation of wheat straw pulp by a sequential xylanase–laccase mediator treatment. J Pulp Pap Sci 2002;28(3):67–71. 135. Sealey JE, Ragaukas AJ and Runge TM. Biobleaching of Kraft pulps with laccase and hydroxybenzotriazole. Proceedings of the Tappi Biological Science Symposium, 1997, pp. 339–342. 136. Amann A. The Lignozym Process Coming Closer to the Mill, ISWPC Montreal F41–F4-5, 1997. 137. Fu S, Zhan H and Yu H. Preliminary study on biobleaching of Eucalyptus urophylla Kraft pulp with laccase-mediator system. China Pulp Pap 2000;19(2):8–15.

378 138. Bourbonnais R, Leech D, Paice MG and Freiermuth B. Reactivity and mechanism of laccase-mediators for pulp deligniﬁcation. Proceedings on Tappi Biological Science Symposium, 1997, pp. 335–338. 139. Poppius-Levlin K, Wang W, Ranua M, Niku-Paavola M-L, Viikari L. Biobleaching of chemical pulps by laccase/mediator systems. In: 1997 Biological sciences Symposium, San Francisco, CA, USA, 19–23 October, 1997, pp. 327–333. 140. Chandra RP, Chakar FS, Allison L, Kim DH, Ragauskas AJ and Elder T. Delving into the fundamental LMS deligniﬁcation of high kappa Kraft pulps. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 54. 141. Bourbonnais R, Rochefort D, Paice MG, Renaud S and Leech D. Transition metal complexes: a new class of laccase-mediators for pulp bleaching. Tappi J 2000;83(10):68. 142. Fagerstro¨m R, Tenkanen M, Kruus K and Buchert J. Removal of hexenuronic acid side groups from Kraft pulp by laccase/mediator treatment. In: 8th International Conference on Biotechnology in Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, pp. 225–230. 143. Ehara K, Tsutsumi Y and Nishida T. Role of Tween 80 in biobleaching of unbleached hardwood Kraft pulp with manganese peroxidase. J Wood Sci 2000;46(2):137–142. 144. Vares T, Almondros G, Galletti GC, Hatakka A, Dorado J, Bocchini P and Martinez AT. Effect of ligninolytic enzymes and mediators on paper making properties and chemical composition of semi-chemical wheat straw pulp. In: 1997 Biological Sciences Symposium, 19–23 October, San Francisco, CA, USA, 1997, pp. 405–412. 145. Camarero S, Garcia O, Vidal T, Colom J, Del Rio JC, Gutirrez A, Gras JM, Monje R, Martinez MJ and Martinez AT. Efﬁcient bleaching of non-wood high-quality paper pulp using laccase-mediator system. Enzyme Microbiol Technol 2004;35(2–3):113–120. 146. Kondo R, Tsuchikawa K and Sakai K. Application of manganese peroxidase to modiﬁcation of ﬁbers. In: 8th International Conference on Biotechnology Pulp and Paper Industry, 4–8 June, Helsinki, Finland, 2001, p. 70.

379

Clinical trial registries and clinical trial result posting: new paradigm for medical writers MaryAnn Foote Vice President, Medical Writing, Abraxis Bioscience, Los Angeles, CA 90049 Abstract. Clinical trial registries and posting of clinical trial results have recently become standard procedures for drug development. Several groups, including journal editors and professional trade organizations have called for legislation or have mandated terms or both for the public disclosure of current trials and the results of the clinical trials within a short timeframe after the trial has ended. Keywords: FDAMA 113, ICMJE, Legislation, PhRMA, Publication, WHO

Introduction In 1997, the United States Food and Drug Administration (FDA) passed a law requiring the listing of clinical trials for the treatment of serious or lifethreatening conditions. The law is part of the Food and Drug Administration and Modernization Act, Section 113, otherwise known as FDAMA 113. On 29 February 2000, a Web site, clinicaltrials.gov was started by the National Library of Medicine (NLM), and in March 2002, the FDA provided guidance on information that would be required for listing a trial in the database. The International Committee of Medical Journal Editors (ICMJE) decided that their journals will publish papers of clinical trial results only if the trial was posted at the time of start of enrollment on a publicly accessible Web site, such as clinicaltrials.gov. Other groups have entered into the discussion, including World Health Organisation (WHO) and Pharmaceutical Researchers and Manufacturers of America (PhRMA). Clinical trial databases are now an integral part of drug development. How did the journal editors, legislators, and drug sponsors arrive at this point? This paper summarizes what is known about the issue as of October 2005, but the processes, legislation, and corporate commitments have changed and probably will continue to change too rapidly to be completely current. How did we get here? Authorship, accountability, and the business of clinical trials The scientiﬁc method consists of several steps: observation, questioning, formulation of a hypothesis, testing and possibly retesting of the hypothesis, Corresponding author: Tel: +310-882-0171.

E-mail: [email protected] (M. Foote). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12011-6

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

380 and publication or communication of the results. ‘‘Science’’ is not complete until the results of an experiment/trial, the hypothesis testing whether successful or not, are published or otherwise presented to other scientists. When results of trials are published or disseminated through meeting abstracts and presentations, other scientists are able to repeat the work, verify its relevance and importance, and move science forward to the next level. Scientists are not the only people who beneﬁt from publication of clinical trial results: patients and their physicians can beneﬁt from reading about current medical research, particularly for rare or life-threatening diseases. Controversy over the writing and publishing of clinical trial results has been ongoing since the early 1990s in the United States. The newest facet of the controversy involves the transparency of clinical trials and their availability to a large patient base, i.e., not just patients who are being treated by clinical researchers in large hospital settings, as well as transparency of clinical trial results, i.e., not just ﬁndings that can be used in marketing and promotion. The posting of clinical trials and clinical trial results can be seen as the latest step in the controversy concerning transparency of scientiﬁc work. This topic has been reviewed in detail in a number of publications and editorials (for reviews, see [1–3]). The initial concern was with authorship and the idea that the author is responsible for the contents of the scientiﬁc paper. The earliest endeavor to deﬁne authorship and discourage guest authorship can be attributed to the ICJME who developed the Uniform Requirements that were ﬁrst published in 1991 [4]. A revised version was published in 1997 [5]; in the later version, the ﬁrst attempts were made to link authorship and accountability for the data in the paper. Before the release of the Uniform Requirements, it was routine practice to add the name of the laboratory head or section head to all papers, a practice known as ‘‘guest’’ authorship. The tradition of guest authorship has not disappeared, however, and remains controversial. Numerous articles have been published concerning the use of ‘‘ghostwriters,’’ particularly writers supported by drug sponsors [6–8]. Ghostwriters are not named authors and often are not acknowledged. Opponents of the use of medical writers who assist in the preparation of manuscripts based on clinical study data believe that ghostwriters manipulate the data and that authors are often coerced into agreeing with the data in the paper [7,8]. Proponents of the use of medical writers to assist in the preparation of clinical trial papers believe that writers who contribute substantially to the writing or editing of a manuscript should be acknowledged with their permission and with disclosure of any pertinent professional or ﬁnancial relationships [9]. In September 2001, 11 journals simultaneously published a paper by Davidoff et al. [10] calling for more involvement of physicians and academic centers in the research required for gaining marketing approval for drugs. At

381 issue was the journal editors’ uneasiness about the ability of drug sponsors to satisfactorily collect and honestly report on the efﬁcacy and, particularly, the safety of their product candidates. In response to Davidoff et al., and growing public sentiment about how drug sponsors ran clinical trials and reported the data, PhRMA issued guidelines on the conduct of clinical trials and publication of the results [11]. In these guidelines, PhRMA members committed to the timely publication or communication of clinical trial results, but did not commit to publishing results of all trials, unless they were medically important. The guidelines allowed for delay to protect intellectual property and for drug sponsors to review all publications before submission to a journal. PhRMA issued a news release stating that PhRMA member companies will post their studies to clinicaltrials.com as of July 1, 2005 [12]. Under this policy, member companies will post, on a voluntary basis, information about all new hypothesis-testing clinical trials by this date and will post ongoing hypothesistesting trials by September 13, 2005. PhRMA also joined with three other international pharmaceutical associations (European Federation of Pharmaceutical Industries and Associations, the International Federation of Pharmaceutical Manufacturers and Associations, and the Japanese Pharmaceutical Manufacturers Association) in agreeing on voluntary principles for disclosing information about clinical trials [13]. PhRMA, WHO, and ICMJE do not agree on what data need to be posted at the start of a clinical trial (Table 1). More controversy: Unpublished results and public health The 2005 Clinical Trial Congress reported that most drug sponsors publish only clinical trial data that can be used in marketing and promotion and that most clinical trial data are seen by regulatory agencies only [14]. In September 2004, Merck voluntary stopped a clinical trial (the APPROVe trial for the treatment of adenomatous polyps) and removed its Cox-2 inhibitor, Vioxxs (roecoximib), from the market because of possible links to increased heart attacks and strokes in patients taking the drug [15]. A meta-analysis of the results of another Vioxx trial (VIGOR, Vioxx Gastrointestinal Outcomes Research) released in 2000 [16] noted an increased risk of cardiovascular events compared with naproxen [17]. Merck argued that the meta-analysis of the rofecoxib data suggested an increased risk for cardiovascular events for patients who received the drug compared with patients who received naproxen and was not a comparison of rofecoxib vs. placebo. The controversy continues in the courts and in the popular press. Legislation Proposals for mandatory registration of clinical trials were made and strongly supported by various groups (e.g., journal editors, consumer

382 Table 1. Comparison of items required for disclosure. The major stakeholders agree on 15 of the 20 items. Most drug sponsors agree that the remaining ﬁve items can be disclosed at a time later than the start of the trial to protect their proprietary interests. WHO is attempting to broker an agreement between drug sponsors and the ICMJE. Disclosure requirements

ICMJE

WHO

PhRMA

1. Unique trial number 2. Registration date 3. Secondary IDs 4. Funding source(s) 5. Primary sponsor 6. Secondary sponsor 7. Responsible contact person 8. Research contact person 9. Brief title of study 10. Ofﬁcial (reg) title 11. IRB review 12. Condition 13. Intervention(s) 14. Key inc/exc 15. Study type 16. Start date 17. Target sample size 18. Recruitment status 19. Primary outcome 20. Secondary outcome

x x x x x x x x x x x x x x x x x x x x

x x x x x x x x x ?* x x ?* x x x ?* x ?* ?*

x x x x x x x x x 0 x x 0 x x x 0 x 0 0

Numbers 10, 13, 17, 19, and 20 are still in negotiations; possibly added by sponsor at later date. ICMJE, International Committee of Medical Journal Editors; PhRMA, Pharmaceutical Research and Manufacturers Association; WHO, World Health Organisation.

advocacy groups) because of concerns that drug sponsors were concealing negative data and concerns about the protection of patients who enroll in clinical trials of investigational products. Another argument in favor of mandatory clinical trial registries stressed the availability of current information concerning clinical trials and their results to researchers, treating physicians, and potential study subjects. Legislation in the United States In October 2004, the Fair Access to Clinical Trials Act was introduced to the United States Senate with a similar bill introduced into the United States House of Representatives [18]. FDAMA 113 requires the Secretary of Health and Human Services (HHS) ‘‘to establish, maintain and operate a data bank of information on clinical trials for drugs for serious or life-threatening diseases and conditions.’’

383 FDAMA 113 provides for an exemption from inclusion of a sponsor’s clinical trial information in the Registry if such disclosure would substantially interfere with the sponsor’s timely enrollment of subjects in the study. In those cases, a sponsor must certify to HHS (through the FDA) that the respective study meets this exemption. HHS reserves the right to reject such certiﬁcation and require inclusion of the clinical trial information in the Registry. Legislation in Europe Legislation in Europe set up a database, EuroPharm, that provides a range of publicly available information on medicinal products with marketing authorization in the European Community and the European Community has discussed the issue of transparency in reporting clinical trial data. The legal basis for EuroPharm was set out in Regulation No. 726/2004 and cover all medicinal products. The EuroPharm database is accessible only to European regulatory authorities at the time of trial initiation, somewhat similar to the conﬁdentiality of marketing applications submitted to the FDA (i.e., New Drug Application (NDA) or Common Technical Document (CTD)). After approval of the product in the European Community, some data will be made publicly accessible, again mirroring the process with the FDA who release some efﬁcacy and safety data at the time it grants marketing approval. Another database in the European Union is European clinical trials database (EudraCT). Draft guidance was issued concerning what data would be considered useful in EudraCT [19]. The information in EudraCT would provide information such as the disease or condition studied, phase of study, objectives of the study, and primary and secondary endpoints. Notiﬁcation of enrolling clinical trials As discussed earlier, clinicaltrials.gov appears to be the preferred Web site for posting clinical trials. This Web site is operated by the US NIH and it was originally established to notify US citizens about federally funded clinical research [20]. Clinicaltrials.gov posts trials funded by US federal agencies or funded by drug sponsors who have submitted clinical trial protocols under the Investigational New Drug (IND) regulations as well as international trials. Many drug sponsors have established publicly accessible databases for clinical trials of their drugs and can be found by a Web search for the company in question. The information that is generally supplied in the registry includes study title, rationale, objective, phase; study design, treatment schedules, locations of study sites; statistical methods; study period (i.e., when study started and when study is anticipated to ﬁnish); and study population or demographics. Once the trial is completed and the data

384 analyzed, results for primary and secondary endpoints, adverse events, and references to publications in the medical literature will be added. Posting of clinical trial results: the second part of the controversy The ICJME, American Medical Association (AMA), and some US legislators advocated mandatory registration of clinical trials, and the ICJME announced in September 2004, that starting July 2005, they will not consider publication of results of any clinical trial that was not registered in a publicly accessible database as of that date [21]. The editors at the British Medical Journal, a member journal of the ICJME, issued its own statement concerning clinical trial registries [22]. The primary difference between the two statements is the recognition of appropriate trial registries. While neither statement endorses a given registry, ICJME requires trials to be listed on the registry operated by the National Institutes of Health in the United States (US NIH) (www.clinicaltrials.gov) and the British Medical Journal cites this registry and also a registry operated by Current Science Ltd (www. cursci.co.uk), a British publishing group that manages BioMed Central, which provides for immediate free access to peer-reviewed research. Other journals have voiced support for clinical trial registries, but it remains to be seen if they will mandate registration as a prerequisite to publication. The ICJME policy states that all trials that commence enrollment as of July 1, 2005 must be posted before patient recruitment starts; trials that were started before this date can be registered retrospectively, but must be done before September 15, 2005 if the researchers wish to have their work published in a ICJME journal. The NIH announced a voluntary new policy on 3 February 2005 that was designed to accelerate public access to published results of NIH-funded research [20]. The NIH policy asks its scientists to publicly release their manuscripts as soon as possible, and within 12 months of publication, and to use PubMed Central, the NIH National Library of Medicine Web site (www.pubmedcentral.nih.gov). The NIH had originally stipulated 6 months, but in concession to publishers who need to maintain income from subscriptions or from on-line sale of single articles, the release time was lengthened to 12 months. Possible problems Most researchers and drug sponsors agree on easy access to clinical trials of new and innovative therapies and free and easy access to clinical trial results. However, the rush to establish databases has left a number of issues that require resolution to protect patients and drug sponsors [23]. Patients who are interested in clinical trials may delay treatment as the ‘‘shop’’ for trials, without understanding the science behind the drug.

385 Patients may not understand the intricacies of drug development and may not understand why a development of a seemingly promising therapy is halted. Drug sponsors may be reluctant to develop novel therapeutics if they are required to post all studies on databases. Increased product information, particularly for products that discontinue development, may lead to increased lawsuits. Posting of results will in itself increase the amount of information available to litigators to use in selecting and prosecuting lawsuits. Discussion and conclusions The transparency of posting clinical trial protocols and clinical trial results will make the process of drug development transparent to the public, and to drug sponsor’s competitors. Because all drug sponsors will be required to post clinical trial protocols and their results within a narrow timeframe, companies that understand their data quickly and can capitalize on the knowledge to move their products forward will succeed. It is possible that clinical research will become more efﬁcient under the new guidelines. The public should become more aware of the nature of clinical research. In the case of Vioxx, the safety monitoring committee began monitoring the potential for cardiovascular risk at their ﬁrst meeting. The committee did not halt the trial immediately because a deﬁnitive answer was needed for the research question. When the risk of cardiovascular complications met the test of statistical signiﬁcance, the trial was stopped and drug voluntarily withdrawn by the manufacturer. The action, or perceived inaction, of Merck in waiting until statistical signiﬁcance was conﬁrmed may not be clear to the public. It is necessary to remind the public that all ethical drugs have side effects. To withdraw a medicine too fast and without full understanding jeopardizes patients who beneﬁt from the medicine. In the situation of the NIH, taxpayers fund the research; thus, it is reasonable to allow open access to publication of data from NIH-sponsored trials. The time frame for release of NIH-sponsored data is shorter than the time frame, generally, for commercial drug sponsors. Publishers will need to ﬁnd ways to maintain income from sale of reprints. Most drug sponsors are not concerned with posting of clinical trial protocols and many drug sponsors have long had Web sites or have used clinicaltrials.com for the posting of open clinical trials as a means of recruiting study subjects. What may be more of a concern is the plethora of clinical trial registries that drug sponsors will need to populate and verify on an ongoing basis. References 1. 2.

Foote MA. Guidelines and policies for medical writers. DIA Forum 2004;40:36–39. Foote MA. Guidelines and policies for medical writers in the biotech industry: an update on the controversy. Biotech Ann Rev 2004;10:259–264.

386 3. 4. 5. 6. 7. 8. 9. 10.

11.

12.

13.

14.

15. 16.

17. 18.

19.

20. 21. 22. 23.

Foote MA. Review of current authorship guidelines and the controversy regarding publication of clinical trial data. Biotech Ann Rev 2003;9:303–313. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. N Engl J Med 1991;324:424–428. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. N Engl J Med 1997;336:309–315. Reidenberg JW. Unmasking ghost writers. Clin Pharmacol Ther 2001;70:208–209. Flanagin A and Rennie D. Acknowledging ghosts. JAMA 1995;273:73. Rennie D and Flanagin A. Authorship! authorship! guests, ghosts, grafters, and the two-sided coin. JAMA 1994;271:469–471. Hamilton CW and Royer MW. For the AMWA Task Force on the Contributions of Medical Writers to Scientiﬁc Publications. AMWA J 2003;18:13–16. Davidoff F, DeAngelis CD, Drazen JM, Nicholls MG, Hoey J, Hojgaard L, Horton R, Kotzin S, Nyelenna M, Overbeke AJ, Sox HC, Van Der Weyden MB and Wilkes MS. Sponsorship, authorship, and accountability. N Engl J Med 2001;345:825–826. Pharmaceutical Research and Manufacturers Association. PhRMA adopts principles for conduct of clinical trials and communication of clinical trial results. http://www. phrma.org/mediaroom/press/releases/20.06.2002.427.cfm, 2002 (accessed 28 May 2003). Pharmaceutical Research and Manufacturers Association. Pharmaceutical companies to make more information available about clinical trials. http://www.phrma.org/mediaroom/ press/releases/06.01.2005.1112.cfm, 2005 (accessed 13 March 2005). Pharmaceutical Research and Manufacturers Association. International alliance of pharmaceutical associations agrees on principles for disclosing information. http://www.phrma. org/mediaroom/press/releases/06.01.2005.1114.cfm, 2005 (accessed 13 March 2005). Bio-IT World. At clinical trials congress, news about industry-run trial registries.http:// www.bio-itworld.com/news/030205_report7668.html?action=print (accessed 10 March 2005). Merck & Co. Merck announces voluntary worldwide withdrawal of VIOXXs. http:// www.vioxx.com/rofecoxib/vioxx/consumer/index/jsp (accessed 17 March 2005). Bombardier C, Laine L, Reicin A, Shapiro D, Bugos-Vargas R, Davis B, Day R, Ferraz MB, Hawkey CJ, Hochberg MC, Kvien TK and Schnitzer TJ, VIGOR Study Group. Comparison of upper gastrointestinal toxicity of recoxib and naproxen in patients with rheumatoid arthritis. N Engl J Med 2000;343:1520–1528. Juni P, Nartey L, Reichenbach S, Sterchi R, Dieppe PA and Egger M. Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. Lancet 2004;364:2021–2029. Fair Access to Clinical Trials Act of 2004; House of Representatives, H R 5252, 108th Congress, 2d Sess; http://olpa.od.nih.gov/tracking/house%5Fbills/session2/ default.asp (accessed 13 March 2005). European Commission. Enterprise and Industry Directorate-General. Guideline on the data ﬁelds from the European clinical trials database (EudraCT) that may be included in the European database on Medicinal Products. http://pharmacos.eudra.org/F2/pharmacos/ docs/Doc2005/03_05/Draft_guid_EudraCT_data_20050303.pdf (accessed 17 March 2005). National Institutes of Health. http://www.nih.gov/news/pr/feb2005/od-03.htm (accessed 14 March 2005). DeAngelis C, Drazen JM, Frizelle FA, et al. Clinical trial registration: a statement from the International Committee of Journal Editors. Lancet 2004;364:911–912. Abbasi K. Compulsory registration of clinical trials. Br Med J 2004;329:637–638. Yingling G, All MB. Thompson Guide to Good Clinical Practice, Vol. 12, June 2005 Boston, Massachussetts.

387

Medical writing departments in biopharma companies: how to establish a department MaryAnn Foote1, and Karen Soskin2 1

Vice President, Medical Writing, Abraxis Bioscience, Los Angeles, CA 90049 Quality Standards and Training, Global Pharmaceutical Regulatory Affairs, Abbott Laboratories, Abbott Park, IL, USA 2

Abstract. Medical writers have important roles in preparing the documentation for approval for marketing of new products, writing manuscripts for publication, and other nonclinical, clinical, and promotional materials. Medical writing departments can be organized in different ways to accommodate the needs of the company. When organizing a new department or when determining metric for an existing department, it is important to understand what medical writers in the biopharma industry do, how they are recruited and trained, and how metrics are developed. Keywords: human resources, measures of work, organizational structure, technical writing.

Introduction Medical writing, which is a different career than technical writing, recently has become a career path of great interest to many people. While the Web site for the Bureau of Labor Statistics [1] seems to make little distinction among the types of writers, it does report that opportunities are good for writers with training in a specialized ﬁeld and that demand for writers with expertise in medicine is expected to increase because of the continuing expansion of scientiﬁc and technical information and the need to communicate it to others. Developments and discoveries in science and technology generate demand for people to interpret technical information for a more general audience. Because medical writers, particularly medical writers in the biopharma industry, often are actively involved in many aspects of drug discovery, development, and commercialization, it is not unusual for a writer to move into other areas within the industry. Writers are skilled at organizing data and documents, facile with numbers and words, and both independent and team oriented. Often writers lend their talent of precision work to quality checks in conjunction with their overarching knowledge of a product to bolster the storyline of a document. Some writers serve as stewards of the critical internal review process. In combining independence and team strength, writers offer an intense blend that can bring greater predictability to drug-development deliverables (e.g., regulatory documents). Over time, many biopharma writers move beyond the limits of their writer roles within a company and Corresponding author: Tel: +1-310-882-0171.

E-mail: [email protected] (M. Foote). BIOTECHNOLOGY ANNUAL REVIEW VOLUME 12 ISSN 1387-2656 DOI: 10.1016/S1387-2656(06)12012-8

r 2006 ELSEVIER B.V. ALL RIGHTS RESERVED

388 become leaders in development or clinical operations, data management, or marketing. Given that medical writers are in great demand and of high value to the biopharma industry, we thought it would be useful to discuss the different roles and responsibilities of medical writers and strategies for their hiring and training. Experience suggests that many recruiters are not familiar with medical writing as a career choice and need coaching in the proﬁle of an ideal candidate. Executives at small biopharma companies may ﬁnd the information particularly relevant when a medical writing department needs to be developed. With the current biopharma focus on scientiﬁc and operational improvement, companies of all sizes continue to evaluate and change their organizations. Managers of established medical writing department may be interested in how we have addressed some issues common in the successful management of a department. This paper is based on more than 50 cumulative years of experience in the pharmaceutical and biotechnology ﬁelds and is based on a presentation given in May 2004 [2]. We also engaged some of our colleagues in discussions of various aspects of managing a medical writing department. Job descriptions ‘‘Medical writing’’ is a generic term and medical writers are employed in a variety of settings [3,4]. The types of documents usually done by medical writers are varied (Table 1) [5]. In general, medical writers write or edit biomedical communications or teach medical writing to others. Many write the regulatory documents required for clinical study and approval of new Table 1. Types of documents written by medical writers in the biopharma industry.

Investigator brochures Investigational new drug application (IND) component parts and updates, including nonclinical sections Protocols Clinical study reports, including safety narratives Fast-track applications Brieﬁng documents for regulatory agencies Summary documents for common technical document Responses to regulatory authority questions Labeling/package inserts Meeting abstracts/posters/slide presentations Journal articles/review articles/book chapters Summaries of investigator meetings Patient education materials Corporate annual reports Patent applications

389 drugs and devices (e.g., investigator brochures, protocols, clinical study reports, common technical document (CTD)) or write manuscripts based on clinical trial results. Other medical writers write patient information leaﬂets or nonclincal manuscripts or textbooks. Given that medical writers do many things, it is necessary for a company to determine precisely why a medical writer is needed so that a satisfactory job description can be written. It is imperative that the company understands the job it wishes to ﬁll. Is the need for someone to design, implement, and monitor clinical trials or is the immediate need for someone to make sense of the mass of data that has accumulated from the clinical trial? For the former position, a company may be best served by hiring someone with a PhD, Pharm D, or MD degree with previous clinical trial experience; for the latter position, a skilled, trained medical writer would be needed. If the need is for formatting and submitting electronic ﬁling documents without need for summarizing, writing, and editing, a document specialist hire would be better. Most companies require written job descriptions with clear responsibilities and objectives of the position. The job description should be used in recruiting and the recruiter should be familiar with the role a medical writing candidate will have at the company. Clear and precise information about what the position entails will undoubtedly save aggravation in the long run. Experience suggests that many recruiters are not familiar with medical writing as a career choice and that the recruiters need coaching in the proﬁle of an ideal candidate. An important consideration for companies just establishing a medical writing department is the concept of a job family. Small biopharma companies often staff their medical writing departments with one writer, a lessthan-optimal situation (discussed subsequently). As the department grows and the scope of the job and the experience of the staff increases, it may be useful to have various positions and grade levels that overlap (i.e., technical vs. managerial) or follow on (i.e., Medical Writer, Senior Writer). Having job families allows candidates to see how they ﬁt into the overall schema of the department and provides direction for promotion and advancement. We strongly urge against the concept of a department with one writer with one job description. A faithful, long-time employee may want and deserve to be promoted and the criteria for advancement should be readily apparent. If, by some chance, the department will for the foreseeable future have only one writer, the job description should note this and the hiring manager should be upfront about long-term prospects. In developing a medical writing department, it is important to address scope (i.e., what documents are the responsibility of the department) and the related challenge of how to handle the overﬂow of work. Most medical writing departments experience more demand for their work product than supply of writers or qualiﬁed candidates or both. Writing departments are typically staffed through a combination of

390 permanent employees, on-site contractors, off-site freelance writers, and assistance from contract research organizations (CROs). Medical writing leaders commonly and creatively juggle delegation to varied resources that are paid, housed, and incentivized differently. Many medical writing departments are responsible for creating and maintaining templates for clinical documentation, creating naming conventions for electronic folder structures, and creating and maintaing electronic common technical document (eCTD). Word processing is often part of a medical writing department’s responsibilities as certain electronic tasks beneﬁt from an understanding of content (bookmarking, hyperlinking, importing). Training on electronic document management for clinical research staff may also be handled by a medical writing department. Qualiﬁcations One of the questions most frequently asked of working medical writers is how one becomes a medical writer. A few programs are available that teach the rudiments of medical writing [3] and writers enter the ﬁeld through various other careers [6]. In perfect ‘‘Catch-22’’ scenario, most biopharma companies will not hire writers without a proven track record in medical writing. Generally, biopharma companies require a science background as part of the qualiﬁcations for the position of medical writer. Depending on the grade level of the job, a Bachelor’s degree in science may be sufﬁcient. Some companies require a Master’s degree or higher degree for more senior positions. While a science background is useful, it may be prudent to consider candidates with related clinical development or other work experience, particularly writing experience in laboratories, at CROs, or in academia. It also may be prudent to consider candidates who may not have the educational background of speciﬁc experience but have the ability to write clearly and are able to quickly and correctly organize large amounts of data into compelling scientiﬁc arguments. Knowledge of statistics or statistical analyses is very important. While medical writers usually do not need to be able to analyze raw data on their own, knowledge of means, medians, p values, and the like are rudimentary and should be considered the minimal amount of knowledge required for the position. An understanding of pharmacokinetics is useful and desirable in a medical writer. Computer skills are essential, including word processing, graphics, desktop publishing, and electronic publishing, depending upon the job roles and responsibilities. Potential writers should be able to work independently, but should also be able to work as part of a team, setting strategy, determining how to present data, and resolving opposing points of view among team members. In many situations, the medical writer also acts as the project manager, keeping the writing task and perhaps the overall ﬁling on track and in this situation, a more senior and experienced writer should be hired. Collaborative and

391 inﬂuencing skills and cross-functional effectiveness are often key to a medical writer’s success. Writers in the biopharma industry who have knowledge of drug regulations for ﬁlings, promotional activities, authorship, and other topics that are relevant to the roles and responsibilities of the job are the ideal candidates. It is critical for a medical writer to have at least a rudimentary understanding of the process of drug development from nonclinical to clinical stages and what documents are required at each stage and what safety issues must be addressed at each stage. (For a review of drug development, see Ref [7].) A writer new to the biopharma industry probably would not have this knowledge and skills and would, thus, not be a candidate for more senior-level positions of most biopharma companies. Some companies, in fact, will not hire inexperienced medical writers, preferring that someone else take on the laborious and costly process of training. Many biopharma companies ask for writing samples during the interview stage, particularly if the candidate is a serious consideration for the position. Most writers are well aware of the conﬁdentiality of their work and the fact that they cannot release even a section of a clinical study report to any one outside the company or a regulatory agency. Sometimes, a document can be ‘‘purged’’ of identifying information by Greeking (i.e., using random letters to replace letters in the name of a drug candidate, disease, or patient population), but such a tactic often makes the document difﬁcult to evaluate properly. Candidates can be asked to write a short 1- to 2-page summary of a topic of interest in lieu of supplying conﬁdential documents. Some companies administer a writing test. These tests tend to be more concerned with grammar and copyediting, although it is possible to supply raw data and hypothesis statement and ask the candidate to prepare tables and ﬁgures with accompanying appropriate text. Such tests often are ‘‘take home’’ tests, which can beg the question did the candidate seek assistance. A caveat concerning testing: all candidates should be tested if any candidates are tested to avoid the appearance of discrimination. Compensation The adage ‘‘you get what you pay for’’ is generally true with medical writers. Experienced freelance writers can command upwards of $200/h, but in-house writers are compensated differently. The salary for an in-house writer takes into consideration beneﬁts such as insurance and paid vacations. It is necessary, however, to research current salaries in your geographical region and within the industry to verify your competitiveness. Many industry guidebooks have entries for technical writers, which is not the same ﬁeld as medical writers. Entries concerning salaries for editors and journalists are not appropriate in most circumstances, although the entries for clinical scientists might be. The American Medical Writers Association

392 (AMWA) periodically polls its membership concerning salary. The most current survey suggests that the mean salary for medical writers employed full time was $74,016 in 2004 [8]. The survey showed that the mean income reported in the surveys for the year previous to the survey has increased steadily from 1989 ($38,887) to 1994 ($49,967) to 2002 ($67,351). Many companies include the cost of ‘‘perks’’ in their total compensation package, detailing the value of beneﬁts, such as health insurance, vision and dental plans, and 401K plans. (It is important to remember these add-ons. Sometimes freelance writers apply for in-house jobs and if they are experienced and in the $200/h range, they often expect to continue to receive this salary. The complete beneﬁt package for these writers should be provided in terms of how much money is involved.) Some companies offer stock options or shares, often to only executives but sometimes to all employees. Certainly, relocation and relocation allowances should be considered for writers at senior levels in the organization. Because writers are in high demand, it may be necessary to offer some extras.

Recruiting strategy Recruiting can be done in-house by the hiring manager or Human Resources Department of the biopharma company, or a search ﬁrm (i.e., recruiter or ‘‘headhunter’’) may be engaged. Because medical writers are in such demand and because it is imperative to hire the best ones to ensure that corporate goals will be met, recruiting must be focused to be efﬁcient. Given the nature of the work, potential medical writers may be found in the laboratories of your company or in postdoctoral programs of local universities. Again we stress, a job description and clear hiring requirements are needed before attempting to recruit writers. Do you need a scientist who can do ﬁrst drafts of regulatory documents that will be edited and formatted by experienced medical writers or do you need writers who can supervise other writers, set policies, and prepare corporate templates. Other good sources of writers are the American Medical Association (AMWA)(www.amwa.org) that mails a quarterly job sheet to its members, the Drug Information Association (DIA) (www.diahome.org), and the Council of Science Editors (CSE)(www.councilscienceeditors.org). Another excellent source of medical writers are freelance writers and writers at CROs that you use frequently. Both sets of writers have experience with your policies and procedures. In a sense, they have had a probationary period with your company so you should be aware of the quality of their work. Many freelance writers do not wish to be full-time employees; raiding the CRO for internal writers may lessen the CROs ability to deliver documents to you on time and with quality. In the latter case, be wary of ‘‘cannibalizing’’ your CRO.

393 Once potential candidates have been identiﬁed, a quick telephone screen usually can determine if the person has the right qualiﬁcations and potential for the position. Many people would like to be medical writers, but they do not understand the position, are not qualiﬁed, or both. After a telephone screen, if the candidate appears qualiﬁed and interested in the position and the location of the position, an on-site interview should be scheduled. The interview team should include representatives from the various functions within the department as well as other members of the team and other collaborative departments. For example, biostatisticians and medical monitors, as well as other writers in the department would be good choices for the interview team. Every company has rules and policies concerning interviewing and background checks. If company policy allows it, calls to former employers and coworkers and members of professional organizations may glean further information. Education and continuing education Recruiting and training medical writers is a difﬁcult, time-consuming, and expensive process. The hiring manager should not abandon good writers. It is incumbent on the manager and the biopharma company to provide growth opportunities; provide chances for development, including ‘‘stretch’’ assignments; and to provide training. Table 2. Taking into consideration that each study is unique and has its own set of interactions, complexity may be deﬁned. Quantiﬁable items that generally increase the complexity of a study: Increasing number of groups or cohorts in a trial Increasing number of protocol amendments Increasing number of countries in which study was conducted Increasing number of study centers Increasing number of interim looks at data for safety, interim analyses, or futility analyses Increasing number and type of analyses Increasing number of covariates Increasing number of efﬁcacy endpoints Other items that can increase the complexity of a study but which are not easily quantiﬁable: Study phase (complexity is often phase 1, phase 3, phase 2) Adverse event proﬁle of test article Need for pharmacokinetic, pharmacoeconomic, or clinical immunology components Number of substances tested (i.e., combination trials) Number of cycles of treatment (e.g., chemotherapy cycles) Type of comparator (i.e., placebo vs. active drug)

394 Table 3. Suggested ‘‘value’’ of some common documents a medical writer might prepare. Task

Unit

FTE

IND Investigator’s brochure (new) Investigator’s brochure (update) Clinical study report (abbreviated)a Clinical study report (phase 2) Clinical study report (pivotal phase 3) Summary documents for CTD Brieﬁng document Manuscript Poster Abstract

2 2 2 2–4 3 4 5 5 1 0.3 0.15

– – 0.1 0.27 – 0.36 0.44 – – – 0.03

a Clinical study reports can range from very abbreviated reports to pivotal phase 3 reports. CTD ¼ common technical document. FTE ¼ full-time writer equivalent.

Table 4. Potential objective and subjective metrics for document quality.

Number of typographical errors; errors in grammar, style, and cross-references; numerical errors in tables or text Number of errors in formatting (e.g., margins, headers/footers) Number of errors in organization (e.g., all elements required by template are present) Number of substantive questions received from a regulatory agency or substantive comments received from peer review Subjective metricsa Does clinical study report clearly and directly address the protocol-speciﬁed study objectives? Does the document support the target label? Overall clarity of the document, including arguments logical and complete, issues identiﬁed and addressed, terms and abbreviations used consistently, organization, and clarity in writing

a If objective review has corrected typographical and grammatical errors and other distractions, subjective review can be done faster.

AMWA offers excellent courses on basic writing and grammatical skills, while DIA training is useful for strengthening or enriching statistical, pharmacokinetic, and regulatory knowledge. Some biopharma companies require regular attendance, certiﬁcation, or both from AMWA. Other companies require attendance at DIA or other professional meetings. Another way to administer continuing education, besides attendance at meetings, is to provide training in-house. This approach often requires a dedicated senior staff member to devise and complete training. It is often difﬁcult to have full attendance even at in-house sessions because of ongoing deadlines that need to be met. One solution is to hold informal sessions where

395 Table 5. Potential process metrics for documents.

Number of drafts Time required to complete drafts Number of requests for additional tables/graphs/ﬁgures/data Time from analysis complete to ﬁnal clinical study report Time from start of clinical study report shell to completion of clinical study report shell Time from analysis complete to manuscript submitted Time from manuscript submission to acceptance Acceptance for publication

a speciﬁc topic, such as one section of a document, is discussed in detail and experiences shared. Headcount and metrics A good medical writing department is a tremendous resource to a biopharma company, allowing rapid writing of key regulatory documents and manuscripts based on trial results, ensuring consistency in formats and messages/terminology, etc. Once the value of a medical writing department is recognized, many people will enquire about the availability of a writer for their projects. While it is satisfying to expand the department’s list of services, it is important to maintain a good quality of life for the medical writers. Thus, it is important to determine how many medical writers are needed in a medical writing department and how much work each writer should complete each year. Budgeting purposes generally require some forward thinking and metrics can be used to help determine the size of a department. We are not aware of any hard and fast data that determine the number of writers needed. Much depends on the work to be done. Taking clinical study reports as an example, in general, the complexity of a study increases with many factors, including number of cohorts and number of covariates (Table 2). However, it is possible within an organization to set a minimum number of documents that must be completed in a year; this metric can be useful at review time to determine compensation and possible promotions. Table 3 is an attempt to assign a value to some common documents a medical writer might prepare and to suggest a full-time writer equivalent (FTE). Each medical writer in a department is assumed to complete 10 units per year. Another metric is based on full-time medical writer equivalent. Because metrics vary among companies and because the data are difﬁcult to obtain from companies with which we consulted, we do not have data for all categories. Other metrics pertain to the quality of reports and publications. Document quality is very difﬁcult to deﬁne or measure. The most easily measured aspects of quality (potential objective measures) are the least important ones (Table 4). On the other hand, the most important aspects (potential

396

Fig. 1. Model for separate clinical development and medical affairs medical writing

departments.

subjective measures) (Table 4) can only be assessed subjectively. To our knowledge, based on contacts in the biopharma industry and presentations at professional conferences, no industry-wide standard metrics for document quality have been established, but the topic is of great interest. An important issue in the context of a subjective review is whether the quality of the document can be separated from the quality of the study. For example, clinical study reports and manuscripts of excellent quality can be written based on results from studies with poor results. The converse is equally true. Should a clinical study report that does not support the target label be considered of poor quality if the study design was inappropriate for this purpose? Some tangible metrics can be listed to determine the quality of the process of preparing documents (Table 5). Structure of a medical writing department Medical writing groups can be found to be aligned with clinical development, biostatistics, regulatory affairs, and other departments. Some departments

397

Fig. 2. Model for single medical writing department with separate clinical development and medical affairs groups.

have a single ‘‘medical writer’’, a situation we discourage, who is usually a combination of secretary, technician, and editor than a true medical writer. The most common division is between clinical development and commercialization (i.e., medical affairs, scientiﬁc affairs, professional services). The terms ‘‘clinical development’’ and ‘‘medical affairs’’ are used in this paper to differentiate the two groups. The ﬁrst responsible for initiating trials and gaining marketing approval (investigational new drug application (IND) to common technical document (CTD) span), while the latter is responsible for phase 3b and 4 studies, clinical grants, and investigator-initiated studies. The structure of the medical writing department largely depends on the structure of the parent company and the work that the writers will be doing. One option (Fig. 1) calls for separate departments for clinical development and medical affairs. The advantages of a separate department model are that this model is simple and familiar and allows a focus on medical affairs for biopharma companies with marketed products. The disadvantages of the model include maintains two separate departments with similar responsibilities, allows

398

Fig. 3. Model for fully integrated medical writing department.

minimal ﬂexibility to shift resources, allows minimal capitalization on expertise gained in phase 1 to 3 studies to be transferred to phase 3b and 4 studies, may result in inconsistent processes within one biopharma company, creates competition between departments for hiring and rewards; and may result in inequity in grade levels and compensation. A second option has separate clinical development and medical affairs staff (Fig. 2) [9]. The advantage of this model is that it maintains a medical affairs focus. The disadvantages include that the model maintains essentially two separate departments, creates competition between groups for hiring and rewards, and allows minimal ﬂexibility to shift resources. A third model establishes a fully integrated department (Fig. 3). The advantages are that this model maximizes ﬂexibility in shifting resources, facilitates use of consistent processes, develops broader skill set and expertise within a therapeutic area, and allows writers to follow the entire development of a product from IND to commercialization. This model, however, does require

Table 6. Snapshot of the medical writing landscape. We used our contacts within the industry to try to understand some basic common areas among medical writing departments in six large biopharma companies. We ensured conﬁdentiality to the respondents. All companies interviewed required that medical writing be done by the medical writing department (i.e., centralization of writing resources); all companies required a scientiﬁc, rather than journalistic or English, degree; and all companies used an electronic submission system. No. 1

No. 2

No. 3

No. 4

No.5

No. 6

Reporting structure

Regulatory affairs

Medical affairs

Clinical operations and regulatory affairs

Clinical operations

Therapeutic area

Regulatory affairs

% of work outsourced No. review cycles per document % of writers with AMWA certiﬁcation

o25%

>50%

50%

50%

50%

50%

3 or more

3–5

2

3

3–5

2

>80%

o20%

o20%

40–60%

o20%

40–60%

399

400 almost daily prioritization of resources between clinical development and medical affairs projects. Summary Table 6 compares several aspects of medical writing. Medical writers in the biopharma industry often are responsible for and do more than simply writing or editing documents. Because most writers tend to be scientists with good organizational skills, they may be involved in writing, editing, developing templates and style guides, developing standards for electronic publishing, and coordinating paper-based and electronic-based review of documents. Good medical writers allow other team members to focus on their area of expertise, comfortable with the knowledge that the writer will do his or her part to ensure success of the project. Best practice is to delineate clearly the scope of writing services offered by the department. Stating expected timeframes for deliverables will help manage expectations. The range of documents prepared by a medical writing department is large and bears no or little relationship to the size of the company. Some major companies have medical writing departments that focus exclusively on clinical study reports, while some smaller companies have medical writing departments that produce nearly all regulatory and promotional documents and manuscripts. Acknowledgements Our thanks to Mary Black, Melissa Gannon, Sally McLeish, RN, Caryn Pacholski, Susan Whitt, and Jim Yuen for helpful discussions over the years. References 1. 2. 3. 4. 5. 6. 7. 8. 9.

Bureau of Labor. http://www.bls.gov/oco/pdf/ocos089.pdf. Accessed 3 May 2005. Foote MA. Recruiting and managing a writing team. Talk presented at Publication Planning 2004: Practical Strategies for Publication Excellence, Princeton, NJ, May 2004. Foote MA. Medical writing as a career choice. DIA Forum 2003;39:18–21. Foote MA. What’s in store for pharmaceutical medical writers? AMWA J 2000;15:7–8. Bernhardt SA. Improving document review practices at pharmaceutical companies. J Bus Tech Commun 2003;17:439–473. Berman SJ. Pharmaceutical medical writers: diverse paths and diverse skills. AMWA J 1999;14:6–7. Foote MA and Neumann TK. A primer on drug development for medical writers. AMWA J 2003;18:61–66. Gray T and Hamilton CW. Findings from the 2004 AMWA salary survey. AMWA J 2004;19:145–151. Purcell T. Using cross-functional teams to write combined clinical and statistical research reports: a communication specialist’s perspective. DIA J 1999;33:159–163.

401

Index of authors

Adhikari, R. 301 Aguilar, M.-I. 85 Alvarez-Lorenzo, C. 225 Anand, A. 349 Bajpai, P. 349 Bajpai, P.K. 349 Battad, J. 31 Cole, T.J. 269 Concheiro, A. 225 Devenish, R. 31 Fauconnot, L. 67 Fay, L.B. 67 Foote, M.A. 379, 387 Gerits, N. 153 Grigorov, M. 67 Gunatillake, P. 301

Gupta, M.N. 1 Jain, S. 1 Lee, T.-H. 85 Liu, Z.C. 137 Mayadunne, R. 301 Mikalsen, T. 153 Moens, U. 153 Mondal, K. 1 Mutch, D.M. 67 Prescott, M. 31 Rossjohn, J. 31 Soskin, K. 387 Swi Chang, T.M. 137 Teotia, S. 1 Wilmann, P. 31

403

Keyword index

activation-modulated drug delivery 226, 246 afﬁnity precipitation 4, 6–7, 11, 19–20 all-protein chromophore 56 all-protein chromophores 32 anaplastic lymphoma kinase 191 aqueous two-phase system 6 artiﬁcial cells 137, 141 artiﬁcial lipid membrane models 86, 88 atomic force microscopy 122

drug discovery 75 dual polarization interferometry 122

biobleaching 349, 357, 361, 363–365, 367 bioencapsulation 137 bioinformatics 67 biomathematics 72, 76, 80 bone marrow cells 137, 139–141, 143

FDAMA 113 379, 382–383 Feedback-regulated release 226 FLT-3 159, 180, 185, 189–190, 200 free-radically polymerizable injectable polymers, 327 poly(ethylene glycol) dimethacrylate, 329 PVA hydrogels, 330 poly(propylene fumarate), 331 natural polymers, 333 free-suspended lipid membranes 87 FRET 31, 33, 43–44, 46, 52–55 functional genomics 75, 78

c-Abl 154, 159, 174, 185, 191–192 c-Kit 154, 180, 185, 187–189, 192, 200 Cholesterol traps 256 chromophore 31–35, 37–42, 46–49, 52–53, 55–60 chromoprotein 31–32, 38, 56 clinical trials 153, 155–156, 159–162, 176–179, 181–185, 187–188, 191, 193, 196–197, 201–203, 205–206 coculture 137, 139–141, 147 coencapsulation 137, 141–144 competitive binding 246 computational lipidomics 70, 80 controlled release 227, 235 coupled plasmon-waveguide resonance 120 covalent imprinting 229–230 covalently immobilized lipid membranes 96 drug delivery 225–227, 232–233, 241–242, 246–247, 251, 258–259

EGFR 154, 157, 162, 174, 176–179, 182–183, 185 enantioselective release 235–236 encapsulation of liposomes in soft gels 98 encapsulation of liposomes in sol gels 101 expanded bed chromatography 4

gas chromatography 70 GFP 31–32, 34–35, 37–41, 43, 45, 52, 54, 56–59 glaucoma 242 glucocorticoid receptor 269–270, 272, 274, 286–288, 292 glucocorticoids 269–272, 274, 276, 278–288, 290–291 glucose traps 255 green ﬂuorescent protein (GFP) 31 Gunn rat 137–138, 143–144, 147 health management 67–68 hepatocytes 137–147

404 human resources 392 hydrolytically induced drug release 247 hyperbilirubinemia 138, 143, 147 ICMJE 379, 381 immobilisation of liposomes and proteoliposomes 98–99, 101 immune function 269 imprinted beads 257 imprinted hydrogels 232, 243–245, 248–249, 251, 253 imprinted particles 233–234 Imprinted soft contact lenses 241 imprinted traps 255 imprinting of peptides and proteins 238 imprinting with cyclodextrins 239 Imprinting without solvents 241 imprinting without template 229 inﬂammatory disease 269, 289–290, 292 injectable biodegradable polymers, 325 Urethane-based injectable polymers, 326 integrin-linked kinase 194, 204 kinase 153–162, 174–186, 188–195, 198–206 laccase 351–353, 355–358, 360–369 legislation 379, 381–383 ligand-binding domain 273–274, 292 lignin peroxidase 351–354, 356–358, 361–362, 369 lignin-oxidizing enzymes 349, 351, 356, 358, 368–369 lipidomics 72, 75–76, 80 macro-afﬁnity ligand facilitated three phase partitioning (MLFTPP) 4 manganese peroxidase 351–356, 358–359, 361–362, 367–369 MAP kinase 175, 198 mass spectrometry 70–71 measures of work 387

micro and nano-arrayed lipid bilayers 110 microencapsulation 146 MLFTPP 14–16, 19 molecular imprinting 225–227, 232, 238–240, 242–243, 246, 248, 254–255, 258–259 molecular imprinting in water 231, 258 mTOR 194–198 NIPA 250–251 non-covalent imprinting 228, 231, 257 nonreceptor tyrosine 153 nuclear magnetic resonance 70–71 nutrition 75–76 optical biosensors 117 organizational structure 387 PDGFR 183–190, 200, 202 PhRMA 379, 381 pH-sensitive imprinted system 251 PKC 174, 183, 185, 187, 189–190, 196, 201–203 poly(ortho esters), 309 synthesis and properties, 309 biodegradation, 309 polyanhydridess, 309 synthesis and properties, 309 biodegradation, 319 polycarbonates, 312 synthesis and properties, 313 biodegradation, 313 polyesteramides, 313 synthesis and properties, 313 biodegradation, 314 polyesters, 302 poly(a-hydroxy esters), 302 poly(glycolic acid), 302 poly(lactic acid), 304 synthesis and properties, 302, 307 copolymers of a-hydroxy acids, 305 biodegradability, 306

405 polylactones, 308 synthesis and properties, 308 biodegradation, 308 poly(3-hydroxy butyrate) (PHB), 308 polymer supported/lipid bilayers 103 polyphosphazenes, 315 synthesis and properties, 315 biodegradation, 317 polyurethanes, 317 1,6-hexamethylenediidocyanate-based polyurethanes, 319 lysine diisocyanate-based polyurethanes, 321 biocompatibility, 324 biodegradation, 324 principal component analysis 73 publication 380–381, 384–385 rate-programmed drug delivery 226, 233 rational design of imprinted systems 258 reconstitution of membrane proteins 114–116 red ﬂuorescent protein 31, 35 scanning probe microscopy 122 selective glucocorticoid receptor modulator 269, 286, 289, 292 Smart materials 225

stem cells 137, 140 steroid ligand 269 stimuli-sensitive polymers 249 supported lipid membranes 92, 119 surface plasmon resonance spectroscopy 117 syngeneic 137, 142 technical writing 387 Temperature-sensitive imprinted system 251 Tethered bilayer membranes 108 Timolol 242–245 TPP 13–16, 18–19 transplantation 137–138, 141–144, 146–147 type-2 diabetes 269, 288, 292 VEGFR 154, 161, 177, 180–186, 190, 200 viability maintenance 141–142, 147 white-rot fungi 349–351, 354, 361 WHO 379–385 xenogeneic 137–138, 142